Operating methods of a memory controller and a memory system including the memory controller
By employing adaptive flushing operations and leveraging the collaborative work of the error correction code engine, fault analysis module, and flushing management module, fault addresses in the storage system are identified and corrected, thus resolving the issue of insufficient accuracy in flushing operations and improving the reliability and stability of the system.
Patent Information
- Application Number
- CN202610293723.8
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Priority Date
- 2025-12-16
- Filing Date
- 2026-03-11
- Publication Date
- 2026-11-17
AI Technical Summary
Existing storage systems suffer from insufficient accuracy in recording erroneous information during flushing operations, leading to decreased system stability and additional system interruptions.
An adaptive scrubbing operation is adopted, which uses the error correction code engine, fault analysis module and scrubbing management module to work together to identify fault addresses and execute adaptive and specified scrubbing operations, thereby improving the accuracy of the scrubbing operation.
By detecting errors that could develop into uncorrectable errors, the reliability and stability of the storage system are improved, preventing additional system interruptions.
Smart Images

Figure CN123114845A_ABST
Abstract
Description
Cross-references to related applications
[0001] This patent application claims the priority of U.S. Provisional Application No. 63 / 807,211, filed May 16, 2025, and U.S. Patent Application Serial No. 19 / 422,422, filed December 16, 2025, which are incorporated herein by reference in their entirety. Technical Field
[0002] Various embodiments of this disclosure relate to semiconductor design techniques, and more specifically, to a storage system for performing a flushing operation. Background Technology
[0003] The storage system can periodically perform error checking and flushing (ECS) operations (hereinafter referred to as "flushing operations"), in which data is read from each storage area of the storage device, errors in the read data are checked and corrected, and the corrected data is written back to the corresponding storage area. The storage system can perform error logging operations to record error information obtained as a result of the flushing operations. Based on the error logging information, the storage system can detect errors that may develop into uncorrectable errors (UEs), thereby providing optimized reliability, availability, and serviceability (RAS) operations.
[0004] If the accuracy of the error log information obtained from the flushing operation is reduced, additional system interruptions may occur even after the RAS operation is performed, which could lead to decreased system stability. Therefore, various methods are being investigated to improve the accuracy of the flushing operation. Summary of the Invention
[0005] Embodiments of this disclosure relate to a memory controller and storage system capable of performing adaptive flushing operations.
[0006] According to one embodiment of this disclosure, a memory controller includes: an error correction code (ECC) engine configured to check and correct errors in data output from a storage device and generate error information; a fault analysis module configured to analyze an error log based on the error information, generate and collect fault addresses, and generate fault mode information by analyzing the relationships between the collected fault addresses; and a flush management module configured to control the storage device to perform a first flush operation and a second flush operation, and to control the second flush operation to be performed on one or more adjacent addresses specified within a fault boundary based on the fault addresses generated according to the first flush operation.
[0007] According to one embodiment of the present disclosure, a method of operating a storage system includes: performing a first flushing operation on a target row among a plurality of rows of the storage device and a target column among a plurality of columns of the storage device; and performing a second flushing operation on one or more adjacent rows specified within a fault boundary based on the target row, according to the result of the first flushing operation.
[0008] According to one embodiment of the present disclosure, a method of operating a storage system includes: sequentially performing a first flushing operation on a plurality of rows of a storage device at predetermined intervals; and performing a second flushing operation on one or more adjacent rows specified within a fault boundary based on a faulty row detected as a first flushing operation.
[0009] According to embodiments of this disclosure, a storage system can accurately identify the type of fault by performing additional specified flushing operations on adjacent rows within the fault boundary, based on faulty rows detected by on-demand flushing operations or inspection flushing operations.
[0010] According to embodiments of this disclosure, a storage system can provide optimized RAS operation by detecting errors that may develop into uncorrectable errors (UEs) that cannot be detected by conventional methods.
[0011] Furthermore, according to embodiments of this disclosure, the storage system can improve system stability by preventing additional system interruptions.
[0012] These and other features and advantages of the embodiments of this disclosure will become apparent to those skilled in the art from the following detailed description taken in conjunction with the accompanying drawings. Attached Figure Description
[0013] Figure 1 This is a block diagram illustrating a storage system according to an embodiment of the present disclosure.
[0014] Figure 2 It is shown Figure 1 A detailed block diagram of a portion of the memory controller shown.
[0015] Figure 3A and Figure 3B It is shown Figure 1 A diagram of a storage cell array.
[0016] Figures 4A to 4E This is a diagram illustrating the types of faults that may occur within a fault boundary according to an embodiment of the present disclosure.
[0017] Figure 5 This is a flowchart describing an adaptive scrubbing operation according to an embodiment of the present disclosure.
[0018] Figure 6This is a flowchart describing an on-demand scrubbing operation according to an embodiment of the present disclosure.
[0019] Figure 7 This is a flowchart describing an inspection and scrubbing operation according to an embodiment of the present disclosure.
[0020] Figure 8 and Figure 9 This is a flowchart used to describe a specified scrubbing operation according to embodiments of the present disclosure.
[0021] Figure 10 This is a block diagram illustrating a storage system including a storage module according to an embodiment of the present disclosure.
[0022] Figure 11 This is a block diagram illustrating a storage system including stacked storage devices according to an embodiment of the present disclosure.
[0023] Figure 12 This is a block diagram illustrating a mobile system including a storage device according to an embodiment of the present disclosure. Detailed Implementation
[0024] Various embodiments of the present disclosure will now be described in more detail with reference to the accompanying drawings. However, the embodiments of the present disclosure may take different forms and should not be construed as limited to those described herein. Rather, these embodiments are provided to make the present disclosure exhaustive and complete, and to fully convey the scope of the disclosure to those skilled in the art. Throughout the present disclosure, the same reference numerals refer to the same parts in the various figures and embodiments of the present disclosure.
[0025] It should be understood that when an element is referred to as being "coupled" or "connected" to another element, this may mean that the two elements are directly coupled, or that the two elements are electrically connected to each other through another circuit or element interposed therebetween. It should also be understood that, as used in this specification, the terms "comprising," "including," "having," etc., specify the presence of the stated features, quantities, steps, operations, elements, components, and / or combinations thereof, but do not exclude the presence or addition of one or more other features, quantities, steps, operations, elements, components, and / or combinations thereof. In this disclosure, unless the context clearly specifies otherwise, the singular form is also intended to include the plural form.
[0026] Figure 1 This is a block diagram illustrating a storage system according to an embodiment of the present disclosure.
[0027] Reference Figure 1The storage system 10 may include a memory controller 100 and a storage device 200. The storage system 10 can store data under the control of a host device 20 (e.g., a mobile phone, smartphone, MP3 player, laptop, desktop computer, game console, television, tablet computer, or in-vehicle infotainment system). The host device 20 may be an external device of the storage system 10.
[0028] The memory controller 100 can control the operation of the storage system 10 and control the data transfer between the host 20 and the storage device 200. The memory controller 100 can generate a command / address signal C / A based on a request REQ from the host 20 and provide the generated command / address signal C / A to the storage device 200. The memory controller 100 can provide the data DIO corresponding to the request REQ from the host 200 to the storage device 200 and provide the data DIO read from the storage device 200 to the host 200. For example, during a write operation, the memory controller 100 can provide the storage device 200 with a write command, address, and data. During a read operation, the memory controller 100 can provide the storage device 200 with a read command and address and provide the data read from the storage device 200 to the host 200.
[0029] Storage device 200 can store data DIO. Storage device 200 can operate under the control of memory controller 100. Storage device 200 may include a memory cell array 210, which includes a plurality of memory cells for storing data. Storage device 200 may include dynamic random access memory (DRAM), which includes dynamic memory cells. In one embodiment, storage device 200 may be double data rate synchronous dynamic random access memory (DDR SDRAM), low power double data rate (LPDDR) SDRAM, graphics double data rate (GDDR) SDRAM, Rambus dynamic random access memory (RDRAM), or others. According to one embodiment, storage device 200 may include an on-chip error correction code (ECC) engine.
[0030] Storage device 200 is configured to receive a command / address signal C / A from memory controller 100 to access a region selected from memory cell array 210. That is, storage device 200 can perform operations instructed by a command on the region selected by address. For example, storage device 200 can perform a write operation (e.g., a programming operation) to write data DIO to the region selected by address. During a read operation, storage device 200 can read data DIO from the region selected by address.
[0031] The memory controller 100 may include a host interface 110, a control module 120, an error correction code (ECC) engine 130, a storage interface 140, a fault analysis module 150, a flush management module 160, and a bus 170.
[0032] The host interface 110 may be an interface for communication between the host 20 and the memory controller 100. The host interface 110 may receive requests (REQ) from the host 20, receive data (DIO) read from the storage device 200 through the storage interface 140, and transmit the received data (DIO) to the host 20.
[0033] Control module 120 can receive requests (REQs) from host 20 via host interface 110. Control module 120 can control each component of memory controller 100 according to the requests (REQs). Control module 120 can generate various commands (e.g., activation commands, precharge commands, read commands, write commands, etc.) and addresses according to the requests (REQs). For example, control module 120 can generate an activation address along with an activation command, and generate a read or write address along with a read command or write command. Hereinafter, the commands and addresses generated in response to requests (REQs) from host 20 will be referred to as host-requested commands and addresses.
[0034] The control module 120 can set the order in which instructions in the request REQ from the host 20 are sent to the storage device 200, and generate commands to be applied to the storage device 200 according to a predetermined order of operations. To improve the performance of the storage device 200, the control module 120 can change the order in which request REQs are received from the host 20 and the order in which operations are sent to the storage device 200. For example, even if the host 20 requests a read operation from the storage device 200 first and then a write operation, the control module 120 can adjust the order so that the write operation is performed before the read operation.
[0035] ECC engine 130 can check and correct errors in the data DIO read from storage device 200, and generate error information and corrected data. When the number of error bits in the data DIO exceeds the error correction capability of ECC engine 130, control module 120 can notify host 20 that an uncorrectable error UE has occurred.
[0036] Storage interface 140 can be configured to communicate with storage device 200. Storage interface 140 can send command / address signals (C / A) and data DIOs to storage device 200, and receive data DIOs read from storage device 200. For example, storage interface 140 can provide a command / address signal (C / A) corresponding to a command and address generated by control module 120 to storage device 200. Furthermore, storage interface 140 can provide data DIOs corresponding to a request (REQ) provided from host interface 110 to storage device 200.
[0037] The fault analysis module 150 can generate error recording information (e.g., error log) based on the error information generated by the ECC engine 130, and generate a fault address by analyzing the error log.
[0038] The flushing management module 160 can control the storage device 200 to perform adaptive flushing operations, including a first flushing operation and a second flushing operation. In one embodiment of this disclosure, based on the fault address detected by the first flushing operation, the flushing management module 160 can control the storage device 200 to perform a second flushing operation on adjacent addresses specified within the fault boundary. It should be noted that the adjacent address may correspond to an address used to specify one or more rows (hereinafter referred to as "adjacent rows") adjacent to the fault row (or fault line) specified by the fault address. Details of the adaptive flushing operation according to an embodiment of this disclosure will be described later.
[0039] Furthermore, the memory controller 100 can transmit data between the host interface 110, control module 120, ECC engine 130, storage interface 140, fault analysis module 150, and flush management module 160 via bus 170. According to one embodiment, the host interface 110, control module 120, ECC engine 130, storage interface 140, fault analysis module 150, and flush management module 160 can communicate independently of each other without using bus 170. For example, the fault analysis module 150 and host interface 110 can communicate directly without using bus 170; the host interface 110 and storage interface 140 can also communicate directly without using bus 170.
[0040] Figure 2 It is shown Figure 1 A detailed block diagram of a portion of the memory controller shown.
[0041] Reference Figure 2 The diagram shows the control module 120, ECC engine 130, fault analysis module 150, and flush management module 160 of the memory controller 100.
[0042] ECC engine 130 can be checked from Figure 1The storage device 200 generates an error signal ERR based on errors in the data DIO read, and can generate error-corrected data by correcting the errors in the data DIO. Furthermore, the ECC engine 130 can also generate a decoding status signal DS based on the error correction result. The error signal ERR includes information about whether an error occurred in the read data DIO and the number of error bits, while the decoding status signal DS includes information about whether the error is a correctable error (CE), an uncorrectable error (UE), or no error (NE). The error signal ERR and the decoding status signal DS can be collectively referred to as "error information." During the read operation, the error-corrected data can be... Figure 1 The host interface 110 is provided to the host 20. During the adaptive flushing operation, the corrected data can be provided to the storage device 200 through the storage interface 140.
[0043] During a write operation, the ECC engine 130 can generate parity bits using data provided from the host 20, and can provide data DIO, including data and parity bits, to the storage device 200 through the storage interface 140.
[0044] The fault analysis module 150 can generate an error log by accumulating error signals ERR generated during the error checking operation, and can detect fault addresses FADD by analyzing the error log. The fault analysis module 150 can generate an error log by accumulating error signals ERR for a single row based on command / address signals C / A, and can analyze the error log to determine whether the corresponding row is a faulty row, thereby detecting fault addresses FADD. For example, when the accumulated error count in a row exceeds a threshold, the fault analysis module 150 can determine that the row is a faulty row and can generate a fault address FADD corresponding to that row. In one embodiment of this disclosure, the fault address FADD may include the row address specifying the faulty row. In the following description, the fault address FADD will be described as a faulty row address. However, this disclosure is not limited thereto; for example, the fault address FADD may include both the row address specifying the faulty row and the column address specifying the faulty column.
[0045] The fault analysis module 150 can collect fault row addresses FADD during adaptive flushing operations and analyze the faults of the memory device 200 based on the collected fault row addresses FADD to generate fault mode information F_MD. The fault analysis module 150 may include multiple address registers, and can sequentially store the fault row addresses FADD in the address registers each time they are detected. The fault analysis module 150 can analyze the relationships between the stored fault row addresses FADD to generate the fault mode information F_MD. The fault analysis module 150 can provide the fault row addresses FADD to the flushing management module 160 and can provide the fault mode information F_MD to the control module 120. The fault analysis module 150 can also be referred to as a fault analysis device and can be located outside the memory controller 100 to provide the fault row addresses FADD and fault mode information F_MD to the memory controller 100.
[0046] The flushing management module 160 can generate a flushing command ECS_CMD instructing a first flushing operation, and a flushing address ECS_ADD specifying the memory cells to be flushed. The flushing address ECS_ADD can be either a row address or a column address. When the fault analysis module 150 detects a faulty row address FADD through the first flushing operation, the flushing management module 160 can generate the flushing command ECS_CMD and the flushing address ECS_ADD to perform a second flushing operation on adjacent row addresses specified within the fault boundary based on the faulty row address FADD. Here, the adjacent row address can be an address specifying one or more adjacent rows adjacent to the faulty row corresponding to the faulty row address FADD. The flushing command ECS_CMD and the flushing address ECS_ADD generated by the flushing management module 160 can be provided to the control module 120.
[0047] In one embodiment of this disclosure, the first scrubbing operation may include at least one of an inspection scrubbing operation and an on-demand scrubbing operation, while the second scrubbing operation may include a designated scrubbing operation.
[0048] A patrol flushing operation is performed to proactively detect and correct errors by patrolling storage units not accessed by host 20. The patrol flushing operation can be performed as a background operation. During the patrol flushing operation, flushing can be performed on all storage units of storage device 200 within a predetermined time period (e.g., 24 hours). Each flushing operation may include a read operation for reading data from storage device 200, an error checking operation for checking and correcting errors in the read data, and a write operation for writing the corrected data back to storage device 200 (hereinafter referred to as a "write-back operation"). The flushing management module 160 can perform the patrol flushing operation by generating flushing commands ECS_CMD and flushing addresses ECS_ADD, the number of times as required to check for errors in all storage units of storage device 200 within the predetermined time period. During each flushing operation, the flushing management module 160 can generate a flushing command ECS_CMD for instruction read and write-back operations, and a flushing address ECS_ADD for specifying the memory cell to be flushed. The ECC engine 130 can perform error checking and correction operations on the data DIO read from the storage device 200 according to the flushing command ECS_CMD and the flushing address ECS_ADD.
[0049] An on-demand scrubbing operation is performed to correct errors in a storage unit (hereinafter referred to as the "target storage unit") accessed according to a request REQ from host 20. During the on-demand scrubbing operation, a scrubbing operation can be performed on the target storage unit. For example, during a read operation on the target storage unit according to a request REQ from host 20, an on-demand scrubbing operation can be performed by writing the data DIO corrected by ECC engine 130 back to the target storage unit. The scrubbing management module 160 can perform the on-demand scrubbing operation by generating a scrubbing command ECS_CMD that writes the corrected data DIO to the target storage unit and specifying the scrubbing address ECS_ADD of the target storage unit. It is worth noting that the scrubbing management module 160 may receive a command / address signal C / A in order to perform the on-demand scrubbing operation.
[0050] The specified scrubbing operation is executed to check for errors in memory cells connected to one or more adjacent rows specified within the fault boundary, based on faulty rows detected as a result of a routine scrubbing operation or an on-demand scrubbing operation. During the specified scrubbing operation, scrubbing operations can be performed on memory cells connected to adjacent rows. Each scrubbing operation may include a read operation for reading data from storage device 200, an error checking operation for checking and correcting errors in the read data, and a write-back operation for writing the corrected data back to storage device 200. The scrubbing management module 160 can execute the specified scrubbing operation by setting the fault boundary based on criteria described later and generating scrubbing commands ECS_CMD and scrubbing addresses ECS_ADD as many times as required to check for errors in memory cells connected to adjacent rows specified within the fault boundary.
[0051] In addition to the fault line address FADD, the flush management module 160 can also receive command / address signals C / A and decode status signals DS to control the read operations, error checking operations, and write-back operations of the first flush operation and the second flush operation.
[0052] The control module 120 may include a reliability, availability, and serviceability (RAS) manager 122 and a command / address processor 124.
[0053] The RAS manager 122 can perform RAS operations based on the fault mode information F_MD to improve the reliability of the storage device 200. In one embodiment, when the fault mode information F_MD is output from the fault analysis module 150, the control module 120 can store the fault mode information F_MD as repair information. For example, the control module 120 can store the repair information in a non-volatile memory (e.g., an array of fuses) so that the stored repair information is retained even when the storage system 10 is powered off. The RAS manager 122 can perform RAS operations based on the stored repair information upon power-up.
[0054] For example, RAS manager 122 can generate commands and addresses for instructing storage device 200 to perform RAS operations, or it can request host 20 to perform RAS operations. RAS manager 122 can request storage device 200 to perform post-packaging repair or error correction operations based on fault mode information F_MD associated with a specific row. RAS manager 122 can request host 20 to perform page detachment or row remapping operations based on fault mode information F_MD associated with two or more rows to exclude the use of two or more rows. RAS manager 122 can also request host 20 to perform storage device standby / migration operations to replace a storage device based on fault mode information F_MD indicating that multiple unexplained defects have occurred in a specific storage device among multiple storage devices, or determine that the corresponding storage device 200 is no longer available.
[0055] Command / address processor 124 can generate host-requested commands and addresses based on the request REQ from host 20. Command / address processor 124 can receive a flush command ECS_CMD and a flush address ECS_ADD from the flush management module 160. Although not shown, in addition to the flush command ECS_CMD and the flush address ECS_ADD, command / address processor 124 can also generate a refresh command for a refresh operation. Regardless of the request REQ from host 20, commands and addresses generated internally for internal operations (such as flush or refresh operations) can be defined as commands and addresses for internal requests. Command / address processor 124 can generate command / address signals C / A by scheduling the generation of host-requested commands and addresses as well as commands and addresses for internal requests. Command / address signals C / A can be provided to storage device 200 through storage interface 140.
[0056] Figure 2 An example of a RAS manager 122 included in a control module 120 is shown; however, the proposed disclosure is not limited thereto, and the RAS manager 122 may be configured separately from the control module 120. Furthermore, while the components of the control module 120 required in one embodiment of this disclosure are as follows... Figure 2 As shown, however, control module 120 may include additional components as required in other embodiments.
[0057] The following description will refer to the accompanying drawings. Figure 1 The structure of the storage cell array 210 and the fault boundary according to an embodiment of the present disclosure.
[0058] Figure 3A and Figure 3B It is used to describe Figure 1 A diagram of a storage cell array.
[0059] Reference Figure 3AThe storage cell array 210 may include multiple cell blocks MB, which are arranged in an array along a first direction X1 (hereinafter referred to as "row direction X1") and a second direction Y1 intersecting the row direction X1 (hereinafter referred to as "column direction Y1"). Each cell block MB may include connections to multiple word lines ( Figure 3B WL) and multiple bit lines ( Figure 3B Multiple storage units between BL) Figure 3B The cell blocks MB arranged along the row direction X1 can be divided into bit lines BL, while the cell blocks MB arranged along the column direction Y1 can be divided into word lines WL. Sub-word line driver regions SWB can be arranged between cell blocks MB in the row direction X1. Multiple sub-word line drivers ( Figure 4B The SWD (sub-word line drive area) can be set in the sub-word line drive area SWB. The bit line sense amplifier area BLSAB can be arranged between the cell blocks MB in the column direction Y1. Multiple bit line sense amplifiers ( Figure 4B The BLSA (Bit Line Sensing Amplifier) can be located in the BLSAB region. In one embodiment of this disclosure, a "cell block" can be defined as a group of memory cells that share a word line WL and a bit line BL and are arranged in the same manner.
[0060] For reference, in order to reduce the propagation delay of word line voltage as the number of memory cells connected to the word line increases and the distance between word lines decreases, a main word line can be divided into multiple (e.g., eight) sub-word lines, which are driven by sub-word line drivers. Hereinafter, the word line WL mentioned in this disclosure may correspond to a known sub-word line, and the sub-word line driver may be referred to as a word line driver.
[0061] In one embodiment, when a group of cell blocks arranged along the row direction X1 among multiple cell blocks MB is defined as a cell group MG, n or more adjacent cell groups (where n is an integer greater than or equal to 2) in the column direction Y1 can be set as fault boundaries. For example, when n is set to 2, the flushing management module 160 can set the cell group containing the faulty row corresponding to the faulty row address, as well as the cell group adjacent to it in the column direction Y1 (or the direction opposite to column direction Y1), as a fault boundary. In the following embodiments, the case where two adjacent cell groups in the column direction Y1 are set as fault boundaries FB will be described as an example.
[0062] Reference Figure 3B This shows the target Figure 3A The fault boundary FB includes a portion of the adjacent 2×2 cell blocks MB00 to MB11, specifically the region MA.
[0063] In row direction X1, the squares between adjacent cell blocks MB00 and MB01 or MB10 and MB11 represent sub-word line drivers (SWDs), while the lines extending to the left and right sides of the SWDs can represent word lines (WLs). In reality, there are many more sub-word line drivers (SWDs) and word lines, but for simplicity... Figure 3B The diagram only shows a portion of the line.
[0064] Each of the cell blocks MB00 to MB11 may include odd-numbered word lines (hereinafter referred to as "first word line WLO") and even-numbered word lines (hereinafter referred to as "second word line WLE") extending along the row direction X1 and alternating with each other in the column direction Y1. In odd-numbered cell blocks MB00 and MB10, the first word line WLO may share a sub-word line driver SWD with adjacent cell blocks in the row direction X1, while the second word line WLE may share a sub-word line driver SWD with adjacent cell blocks in the opposite direction X2. Conversely, in even-numbered cell blocks MB01 and MB11, the second word line WLE may share a sub-word line driver SWD with adjacent cell blocks in the row direction X1, while the first word line WLO may share a sub-word line driver SWD with adjacent cell blocks in the direction X2. That is, since two adjacent cell blocks in the row direction X1 share a sub-word line driver SWD, a sub-word line driver SWD can be assigned to two adjacent cell blocks in the row direction X1 or X2.
[0065] Each of the cell blocks MB00 to MB11 may include a first bit line BLU and a second bit line BLL extending along the column direction Y1 and alternately arranged in the row direction X1. The first bit line BLU may share a bit line sense amplifier BLSA with an adjacent cell block (not shown) in the column direction Y1, while the second bit line BLL may share a bit line sense amplifier BLSA with an adjacent cell block (not shown) in the opposite direction Y2. That is, since two adjacent cell blocks in the column direction Y1 share a bit line sense amplifier BLSA, one bit line sense amplifier BLSA can be assigned to two adjacent cell blocks in the column direction Y1 or Y2.
[0066] Multiple analyses have shown that most faults in current DRAM devices exhibit a tendency for bounded faults to occur in adjacent cell blocks, which is believed to be related to defect patterns caused by the layered architecture of DRAM devices.
[0067] Figures 4A to 4E This is a diagram illustrating the types of faults that may occur within a fault boundary according to an embodiment of the present disclosure.
[0068] Figures 4A to 4EThe diagram illustrates two master word lines arranged in each cell block, along with eight sub-word lines assigned to each master word line. For example, in the upper cell blocks MB00 and MB01, sixteen sub-word lines (rows SWL1 to SWL16) extend along the row direction and alternate along the column direction, with eight sub-word lines assigned to each of the two master word lines MWL1 and MWL2. In the lower cell blocks MB10 and MB11, sixteen sub-word lines (rows SWL17 to SWL32) extend along the row direction and alternate along the column direction, with eight sub-word lines assigned to each of the two master word lines MWL3 and MWL4. Sixteen sub-word line drivers SWD can be alternately arranged in the column direction on both sides of each cell block MB00 to MB11 to drive the corresponding sub-word lines. For example, eight sub-word line drivers on one side of each cell block drive odd-numbered sub-word lines, while eight sub-word line drivers on the opposite side drive even-numbered sub-word lines. Of the sixteen sub-word line drivers SWD arranged on both sides of each unit block, the upper eight sub-word line drivers are connected to the upper master word line (e.g., MWL1 or MWL3), while the lower eight sub-word line drivers are connected to the lower master word line (e.g., MWL2 or MWL4).
[0069] Figure 4A The diagram shows that one of the eight subword line drivers (indicated by a black box) between adjacent cell blocks MB00 and MB01 in the row direction is defective. In this case, the cumulative error count generated from the data output from subword line SWL3 (indicated by a dashed line) connected to the two arms of the defective subword line driver may exceed a threshold, causing subword line SWL3 to be detected as a faulty row.
[0070] Figure 4B The diagram shows that all sub-word line drivers connected to the upper master word line MWL1 (indicated by black boxes) among multiple sub-word line drivers between adjacent cell blocks MB00 and MB01 in the row direction are defective. In this case, the cumulative error count generated from the data output from the sub-word lines connected to the two arms of the defective sub-word line drivers (i.e., the odd-numbered sub-word lines SWL1, SWL3, SWL5, and SWL7 (indicated by dashed lines) connected to the upper master word line MWL1 of cell blocks MB00 and MB01) may exceed the threshold, causing sub-word lines SWL1, SWL3, SWL5, and SWL7 to be detected as faulty rows.
[0071] Figure 4CThe diagram shows a defective upper master word line MWL1 between adjacent cell blocks MB00 and MB01 in the row direction. In this case, the cumulative error count generated from the data output by all sub-word lines SWL1 to SWL8 (indicated by dashed lines) connected to the upper master word line MWL1 of cell blocks MB00 and MB01 may exceed a threshold, causing sub-word lines SWL1 to SWL8 to be detected as faulty rows.
[0072] Figure 4D The diagram illustrates a defective sub-word line driver with the same serial number among the eight sub-word line drivers connected to the upper master word line MWL1 and the lower master word line MWL2, respectively, between adjacent cell blocks MB00 and MB01 in the row direction (e.g., the third sub-word line in the sub-word line sequence of upper master word line MLW1 and the third sub-word line in the sub-word line sequence of upper master word line MLW2). In this case, the cumulative error count generated from the data output by sub-word lines SWL3 and SWL11 (indicated by dashed lines) with the same serial number as the defective sub-word line drivers connected to the two arms of the defective sub-word line drivers of cell blocks MB00 and MB01 may exceed a threshold, causing sub-word lines SWL3 and SWL11 to be detected as faulty rows.
[0073] Figure 4E The diagram illustrates that among multiple sub-word line drivers connected to upper main word lines MWL1 and MWL3 and lower main word lines MWL2 and MWL4 respectively between cell blocks MB00 and MB01 and between cell blocks MB10 and MB11, sub-word line drivers with the same serial number (e.g., the third serial number) are defective. In this case, the cumulative error count generated from the data output by sub-word lines SWL3, SWL11, SWL19, and SWL27 (indicated by dashed lines) with the same serial number as the defective sub-word line drivers connected to the two arms of the defective sub-word line drivers in all adjacent cell blocks MB00 to MB11 may exceed a threshold, causing sub-word lines SWL3, SWL11, SWL19, and SWL27 to be detected as faulty lines.
[0074] In one embodiment, the scrubbing management module 160 can set two adjacent cell groups in the column direction as fault boundaries (FBs), and perform adaptive scrubbing operations on the faulty rows and adjacent rows within the set fault boundaries (FBs). For example, as Figure 4A As shown, when sub-word line SWL3, which is connected to both arms of the defective sub-word line driver, is detected as a faulty row based on the result of the first scrubbing operation, the scrubbing management module 160 can perform a specified scrubbing operation on all or part of the adjacent rows (e.g., SWL1 to SWL8, SWL11, SWL19, and SWL27) arranged in two adjacent adjacent cell groups in the column direction, such as... Figures 4B to 4E As shown.
[0075] However, this disclosure is not limited thereto, and adjacent rows can be specified based on the type of fault that may occur within the fault boundary FB. For example, when a sub-word line SWL3 connected to the two arms of a defective sub-word line driver is detected as a faulty row based on the result of a first brushing operation, the brushing management module 160 can perform a specified brushing operation on all adjacent rows SWL1 to SWL16 arranged in cell blocks MB00 and MB01 in the same row direction.
[0076] The fault analysis module 150 can generate fault mode information F_MD by defining the fault type in the fault line within the fault boundary FB based on the result analysis of the adaptive flushing operation. For example, when an error occurs only in the data output from a single subword line (see...). Figure 4A The fault analysis module 150 can determine that a sub-word line driver driving a single sub-word line is defective. This occurs when an error occurs in the data output from an odd-numbered (or even-numbered) sub-word line connected to the main word line (see [link to main word line]). Figure 4B The fault analysis module 150 can determine that the odd-numbered (or even-numbered) sub-word line driver connected to the main word line is defective. In this case, the fault may be caused by a fault occurring in a contact or common signal line shared by the odd-numbered (or even-numbered) sub-word line drivers. When errors occur in the data output from almost all sub-word lines connected to a main word line (see...) Figure 4C The fault analysis module 150 can determine that the main word line itself is defective.
[0077] Furthermore, when an error occurs in the data output from a sub-word line with the same sequence number that is connected to two adjacent main word lines (see [link to main word line]),... Figure 4D The fault analysis module 150 can determine that the sub-word line driver driving the corresponding sub-word line is defective. In this case, the fault may be caused by a fault occurring in the contacts shared by the sub-word line drivers. When an error occurs in the data output from a sub-word line with the same sequence number among the sub-word lines connected to each of the main word lines (see...), the fault may occur when... Figure 4E The fault analysis module 150 can determine that the sub-word line driver driving the corresponding sub-word line is defective. In this case, the fault may be caused by a fault occurring in the signal path commonly supplied to the sub-word line driver.
[0078] Furthermore, this disclosure is not limited to this. The fault analysis module 150 can analyze multiple fault types and generate fault mode information F_MD based on the results of adaptive flushing operations performed on fault lines and adjacent lines within a defined fault boundary FB.
[0079] In the following text, refer to Figure 5The adaptive scrubbing operation according to an embodiment of the present disclosure will be described in detail below.
[0080] Figure 5 This is a flowchart describing an adaptive scrubbing operation according to an embodiment of the present disclosure.
[0081] Reference Figure 5 First, upon power-up, the storage system 10 can perform reliability, availability, and serviceability (RAS) operations to improve the reliability of the storage device 200 (in S110). For example, the storage system 10 can perform post-packaging repair operations, page removal operations, row remapping operations, or storage bank standby / migration operations.
[0082] The storage system 10 can perform initialization operations (in S120). During the initialization operation, the storage system 10 can perform the initialization sequence defined in the JEDEC standard. For example, the storage system 10 can perform internal circuit initialization operations such as voltage level stabilization and clock stabilization between the memory controller 100 and the storage device 200. Furthermore, the storage system 10 can perform initialization operations on the storage device 200, such as Mode Register Set (MRS) configuration, ZQ calibration, and impedance matching. Although the RAS operation (in S110) and the initialization operation (in S120) are performed in... Figure 5 The details are described separately, but according to the embodiment, the RAS operation (in S110) can be performed during the initialization operation (in S120).
[0083] Subsequently, the flush management module 160 of the memory controller 100 can determine whether there is a planned (or scheduled) designated flush operation (in S130). If the faulty row address FADD detected in the previous operation has not been flushed before power-down, the planned designated flush operation can be performed on the faulty row address FADD. However, this disclosure is not limited to this; after initialization, the planned designated flush operation can also be performed based on the stored repair information.
[0084] When a planned specified flush operation exists ("Yes" in S130), the flush management module 160 can perform the specified flush operation on the faulty row address FADD detected during the previous operation (in S142). The flush management module 160 can set a fault boundary FB and, based on the faulty row, generate flush commands ECS_CMD and flush addresses ECS_ADD multiple times, the number of times corresponding to the number of error checks performed on the memory cells connected to the specified adjacent rows within the fault boundary FB. For example, as... Figure 4AAs shown, when the fault row address FADD corresponding to sub-word line SWL3 is stored, the flush management module 160 can generate a flush command ECS_CMD and a flush address ECS_ADD to perform a specified flush operation on adjacent rows SWL1 to SWL8, SWL11, SWL19 and SWL27 arranged in two adjacent cell groups in the column direction, such as... Figure 4B – Figure 4E As shown. Furthermore, the specified scrubbing operation performed during initialization corresponds to the first type of specified scrubbing operation; its detailed process will be referred to... Figure 8 Describe it.
[0085] The fault analysis module 150 can generate fault mode information F_MD (in S144) by analyzing the fault type for fault lines detected as a result of a specified brushing operation.
[0086] After the planned flushing operation (in S142) and fault analysis (in S144) are completed, the RAS manager 122 can selectively execute RAS operations based on the fault mode information F_MD (in S146) to improve the reliability of the storage device 200. For example, the RAS manager 122 can control the storage system to perform post-packaging repair operations, page removal operations, row remapping operations, or storage bank standby / migration operations. The RAS operation (in S146) can be performed during runtime according to product specifications, or it can be omitted if runtime execution is impractical. Figure 5 In the diagram, the RAS operation (in S146) is shown with a dashed line to indicate that it is optional.
[0087] Subsequently, storage system 10 can perform normal operations such as read or write operations in response to a request REQ from host 20.
[0088] The flushing management module 160 can perform an on-demand flushing operation on the data DIO read from the target storage unit of the storage device 200 during a read operation (in S150). The ECC engine 130 can check for errors in the data DIO read from the target storage unit, generate an error signal ERR, and correct the errors to generate error-corrected data. The error-corrected data can then be provided to the host 20 via the host interface 110. At this point, the flushing management module 160 can perform the on-demand flushing operation by writing the error-corrected data back to the target storage unit. The detailed process of the on-demand flushing operation will be described in [reference needed]. Figure 6 The fault analysis module 150 can generate a fault row address FADD for the target memory cell based on the error signal ERR generated during the error checking operation on the data DIO.
[0089] The scrubbing management module 160 can perform a routine scrubbing operation on all storage cells of the storage device 200 at predetermined intervals (e.g., every 24 hours) (in S160). The scrubbing management module 160 can perform the scrubbing operation, including a read operation for reading data from the storage cells, an error checking and correction operation for the read data, and a write-back operation for writing the corrected data back to the storage device 200 line by line. (See reference...) Figure 7 This describes the detailed process of the inspection and scrubbing operation performed on each row. The fault analysis module 150 can generate a fault row address FADD corresponding to a certain row based on the error signal ERR generated during the inspection and scrubbing operation.
[0090] When a faulty row address FADD is detected as a result of an on-demand or inspection-based flushing operation, and a specified flushing operation is requested ("Yes" in S170), the flushing management module 160 can perform the specified flushing operation on the detected faulty row address FADD (in S182), and the fault analysis module 150 can analyze the fault type in the faulty row based on the result of the specified flushing operation to generate fault mode information F_MD (in S184). Thereafter, the RAS manager 122 can selectively execute RAS operations based on the fault mode information F_MD to improve the reliability of the storage device 200 (in S186). For example, the RAS manager 122 can control the storage system to perform post-packaging repair operations, page detachment operations, row remapping operations, or storage bank standby / migration operations. The RAS operation (in S186) can be performed during runtime according to product specifications, or it can be omitted if runtime execution is impractical. Figure 5 In S186, the RAS operation is also shown with a dashed line to indicate that it is optional.
[0091] The above operations (S150 to S186) can be repeated until the planned power-down ("No" in S190) occurs. That is, when a faulty line address FADD is detected as a result of an on-demand flushing operation (S150) or a patrol flushing operation (S160), and a specified flushing operation is requested ("Yes" in S170), the specified flushing operation can be performed on the detected faulty line address FADD (in S182), and the corresponding RAS operation can be selectively performed (in S186).
[0092] When the planned power-down occurs ("Yes" in S190), the RAS manager 122 can perform RAS operations based on the fault mode information F_MD (in S192). Therefore, the operation of the storage system 10 can be completed.
[0093] Depending on the product specifications, it is possible to perform only one of the following operations: on-demand brushing or inspection brushing, and not both. For example, in a product configured to perform only on-demand brushing, when a faulty row is detected as a result of the on-demand brushing operation, additional specified brushing operations can be performed on adjacent rows within the fault boundary based on the detected faulty row. Alternatively, in a product configured to perform only inspection brushing, when a faulty row is detected as a result of the inspection brushing operation, additional specified brushing operations can be performed on adjacent rows within the fault boundary based on the detected faulty row.
[0094] As described above, in the embodiments of this disclosure, when a faulty row is detected as a result of an on-demand scrubbing operation or a routine scrubbing operation, a specified scrubbing operation can be performed on adjacent rows within the fault boundary based on the detected faulty row. Therefore, the fault type can be accurately determined, and errors that may develop into uncorrectable errors (UEs) can be identified based on the fault type, thereby achieving optimized RAS operation.
[0095] Figure 6 This is a flowchart describing an on-demand scrubbing operation according to an embodiment of the present disclosure.
[0096] Reference Figure 6 This illustrates the on-demand flushing operation performed during a read operation (in... Figure 5 (S150 in the middle).
[0097] During a read operation, the command / address processor 124 can set the row address and column address according to the request REQ from the host 20 to specify the target memory cell to be read (in S210). Hereinafter, the row address and column address used to specify the target memory cell are referred to as the target row address RA and the target column address CA, respectively. The storage device 200 can read data DIO from the target memory cell corresponding to the target row address RA and the target column address CA (in S220).
[0098] If the read data DIO has no errors (NE) ("Yes" in S230), the read data DIO is provided to the host 20, and the process can end.
[0099] If a correctable error (CE) is detected in the read data DIO ("No" in S230 and "Yes" in S240), the ECC engine 130 can correct the error in the read data DIO. The flush management module 160 generates a flush command ECS_CMD and a flush address ECS_ADD for the instruction write-back operation, and accordingly, the corrected data can be written back to the target storage unit (in S242). The fault analysis module 150 can log the error based on the error signal ERR provided from the ECC engine 130 (in S244).
[0100] Normally, the ECC engine 130 of the memory controller 100 can correct multi-bit errors or symbol-level errors. When two adjacent bit errors occur in a single row, they can be treated as single-bit errors. However, when a correctable multi-bit error (CEm) occurs in a row, excluding two adjacent bit errors, the row can be considered a potentially faulty row. In one embodiment, the process can end when the correctable error (CE) is a single-bit error or a two adjacent bit error ("No" in S250). On the other hand, when the correctable error (CE) is a multi-bit error (CEm), excluding two adjacent bit errors ("Yes" in S250), the fault analysis module 150 can generate a faulty row address FADD corresponding to the target row address RA and request the flush management module 160 to perform a first-type (or non-fatal-type) specified flush operation on the faulty row address FADD (in S260). Then, the on-demand flushing process can be completed.
[0101] If an uncorrectable error (UE) is detected in the read data DIO ("No (UE)" in S240), the ECC engine 130 may poison the read data DIO (in S272). In this case, the ECC engine 130 may perform data poisoning by providing the host 20 with a specific pattern indicating that the read data DIO is corrupted, rather than the read data DIO itself. Typically, notifying the host 20 that data has been poisoned during the on-demand scrubbing operation can improve reliability. However, in one embodiment, the notification operation may be omitted during the on-demand scrubbing operation (in S272) and may be performed during a designated scrubbing operation of the second type (or fatal type) described later.
[0102] The fault analysis module 150 can record errors based on the error signal ERR provided from the ECC engine 130 (in S274). The fault analysis module 150 can generate a faulty row address FADD corresponding to the target row address RA and request the flush control module 160 to perform a second type of specified flush operation on the faulty row address FADD (in S280). The on-demand flushing process can then end. During the second type of specified flush operation described below, a data poisoning notification can be issued.
[0103] The first type of specified scrubbing operation is defined as non-fatal, while the second type is defined as fatal. The second type of specified scrubbing operation is performed under more severe conditions than the first type, and may omit the write-back operation in CE case, or in UE case of poisoning, while simultaneously performing a notification operation to host 2 of the poisoning error. Figure 8 The first type of specified scrubbing operation is described in the document, while Figure 9The second type of specified scrubbing operation is described in the document.
[0104] Figure 7 This is a flowchart describing an inspection and scrubbing operation according to an embodiment of the present disclosure.
[0105] Reference Figure 7 This illustrates a patrol flushing operation performed on storage cells connected to a specific row (in... Figure 5 (S160).
[0106] The flush management module 160 can generate a flush address ECS_ADD specifying the cell to be flushed, and a flush command ECS_CMD for instruction read operations. At this time, the flush management module 160 generates a flush address ECS_ADD that includes a row address RA specifying a particular row and a column address CA initialized to zero (in S310). The storage device 200 can read data DIO from the storage cell corresponding to the flush address ECS_ADD (in S320).
[0107] If the read data DIO has no errors (NE) ("Yes" in S330), the flush management module 160 can increment the value of column address CA by +1 (in S350). A flush operation can be performed on the next column address CA until column address CA reaches its maximum value ("No" in S340). The specified flush process can end when column address CA reaches its maximum value ("Yes" in S340).
[0108] If a correctable error (CE) is detected in the read data DIO ("No" in S330 and "Yes" in S360), the ECC engine 130 can correct the error in the read data DIO (in S362). The flush management module 160 generates a flush command ECS_CMD and a flush address ECS_ADD for the instruction write-back operation, and accordingly, the corrected data can be written back to the memory unit (in S362). The fault analysis module 150 can increment the error counter based on the error signal ERR provided from the ECC engine 130 (in S364).
[0109] At this point, if the correctable error (CE) is a single-bit error or an error between two adjacent bits, or if the cumulative error count (#CE) in the currently flushed row is below a threshold (e.g., 2) ("No" in S370), the flush management module 160 can increment the value of column address CA by +1 (in S350). A flush operation can be performed on the next column address CA until column address CA reaches its maximum value ("No" in S340). The specified flushing process can end when column address CA reaches its maximum value ("Yes" in S340).
[0110] On the other hand, if the correctable error (CE) is a multi-bit error (CEm) except for two adjacent bit errors, or if the cumulative error count (#CE) in the currently flushed row exceeds the threshold ("Yes" in S370), then the fault analysis module 150 can generate a faulty row address FADD corresponding to the row address RA, and the flush control module 160 is requested to perform a first type of specified flush operation on the faulty row address FADD (in S380), and the process can end.
[0111] If an uncorrectable error (UE) is detected in the read data DIO ("No (UE)" in S360), the ECC engine 130 can poison the read data DIO (in S392). In this case, since the inspection flushing operation is performed independently of the request REQ of the host 20, no notification is provided to the host 20. The fault analysis module 150 can log the error (in S394), generate a faulty line address FADD corresponding to the line address RA, and then request the flushing control module 160 to perform a first-type specified flushing operation on the faulty line address FADD (in S380), and the process can end.
[0112] Figure 8 This is a flowchart describing a specified scrubbing operation of a first type according to an embodiment of the present disclosure.
[0113] Reference Figure 8 The scrubbing management module 160 can identify adjacent rows to be scrubbed within the fault boundary (in S410) based on the fault row corresponding to the fault row address FADD. For example, when sub-word line SWL3 is detected as a fault row (see... Figure 4A The adjacent rows SWL1 to SWL8, SWL11, SWL19 and SWL27 can be identified as needing to be scrubbed.
[0114] First, a flushing operation can be performed on the memory cell connected to one of the identified adjacent rows SWL1 to SWL8, SWL11, SWL19 and SWL27 (e.g., SWL1).
[0115] The flush management module 160 can set the row address RA to specify the adjacent row SWL1 and initialize the column address CA to 0 (in S420). The flush management module 160 can provide a flush address ECS_ADD (including the row address RA and the column address CA) and a flush command ECS_CMD for instruction read operations. The storage device 200 can read data DIO from the storage cell corresponding to the flush address ECS_ADD according to the flush command ECS_CMD (in S430). That is, the storage device 200 reads data DIO from the storage cell connected to the first column of the adjacent row SWL1.
[0116] If the read data DIO has no error (NE) ("Yes" in S440), the flush management module 160 can increment the value of column address CA by +1 ("No" in S450 and S460).
[0117] If a correctable error (CE) is detected in the read data DIO ("No" in S440 and "Yes" in S470), the ECC engine 130 can correct the error in the read data DIO, and the corrected data can be written back to the storage unit (in S472). The fault analysis module 150 can increment the error counter based on the error signal ERR from the ECC engine 130 (in S474).
[0118] At this point, if the correctable error (CE) is a single bit error or an adjacent two bit error, or if the cumulative error count (#CE) in the currently flushed row is below the threshold ("No" in S476), the flush management module 160 can increment the value of column address CA by +1 ("No" in S450 and S460).
[0119] When the value of column address CA increases by 1 (in S460), storage device 200 can read data DIO from the storage cell connected to the second column of the adjacent row SWL1 (in S430). The above operation can be repeated for the remaining columns of the adjacent row SWL1 until column address CA reaches its maximum value ("No" in S450) (S440 to S476).
[0120] When the column address CA reaches its maximum value ("Yes" in S450), the flush management module 160 can change the value of the row address RA to specify the next adjacent row SWL5 and initialize the column address CA to 0 ("No" in S492 and S494). Therefore, the storage device 200 can read the data DIO from the storage cell connected to the first column of the next adjacent row SWL5 (in S430).
[0121] On the other hand, if the correctable error (CE) is a multi-bit error (CEm) but not an adjacent two-bit error, or if the cumulative error count (#CE) in the currently flushed row exceeds the threshold ("Yes" in S476), then the fault analysis module 150 can generate a fault row address FADD corresponding to the adjacent row SWL1 (in S478). Then, the flush management module 160 can change the value of the row address RA to specify the next adjacent row SWL5 and initialize the column address CA to 0 ("No" in S492 and S494). Therefore, the storage device 200 can suspend the flush operation on the adjacent row SWL1 and read the data DIO from the storage cell connected to the first column of the next adjacent row SWL5 (in S430).
[0122] If an uncorrectable error (UE) is detected in the read data DIO ("No (UE)" in S470), the ECC engine 130 can poison the read data DIO (in S482). In this case, since the specified flushing operation is performed independently of the request REQ of the host 20, no notification is provided to the host 20. The fault analysis module 150 can log the error (in S484) and generate a faulty row address FADD corresponding to the adjacent row SWL1 (in S478). Then, the flushing management module 160 changes the value of the row address RA to specify the next adjacent row SWL5 and initializes the column address CA to 0 ("No" in S492 and S494). Therefore, the storage device 200 can suspend the flushing operation on the adjacent row SWL1 and read the data DIO from the storage cell connected to the first column of the next adjacent row SWL5 (in S430).
[0123] The process can end once the specified scrubbing operation ("Yes" in S492) has been performed on all identified adjacent rows SWL1 to SWL8, SWL11, SWL19 and SWL27.
[0124] Subsequently, the fault analysis module 150 can generate fault mode information F_MD by analyzing the fault line addresses FADD detected as a result of the adaptive flushing operation to determine the fault types between fault lines.
[0125] Figure 9 This is a flowchart describing a second type of specified scrubbing operation according to an embodiment of the present disclosure.
[0126] Reference Figure 9 The flushing management module 160 can identify adjacent rows within the fault boundary to be flushed (in S510) based on the fault row corresponding to the fault row address FADD. For example, when adjacent rows SWL1 to SWL8, SWL11, SWL19 and SWL27 can be identified as needing to be flushed, a flushing operation can be performed on the memory cell connected to one of the adjacent rows SWL1 to SWL8, SWL11, SWL19 and SWL27 (e.g., SWL1).
[0127] The flush management module 160 can set the row address RA to specify the adjacent row SWL1 and initialize the column address CA to 0 (in S520). The flush management module 160 can provide the flush address ECS_ADD (including the row address RA and the column address CA) and the flush command ECS_CMD for instruction read operations. The storage device 200 reads the data DIO from the storage cell corresponding to the flush address ECS_ADD according to the flush command ECS_CMD (in S530). That is, the storage device 200 reads the data DIO from the storage cell connected to the first column of the adjacent row SWL1.
[0128] If the read data DIO has no error (NE) ("Yes" in S540), the flush management module 160 can increment the value of column address CA by +1 ("No" in S550 and S560).
[0129] If a correctable error (CE) is detected in the read data DIO ("No" in S540 and "Yes" in S570), the fault analysis module 150 can increment the error counter based on the error signal ERR provided from the ECC engine 130 (in S572). At this time, since the second type of specified flushing operation is performed again on data for which an uncorrectable error (UE) has occurred during the on-demand flushing operation, the error correction operation and write-back operation can be omitted because they are considered unnecessary.
[0130] If the correctable error (CE) is a single bit error or an adjacent two bit error, or if the cumulative error count (#CE) in the currently flushed row is below the threshold ("No" in S574), then the flush management module 160 can increment the value of column address CA by +1 ("No" in S550 and S560).
[0131] When the value of column address CA increases to 1 (in S560), storage device 200 can read data DIO from the storage cell connected to the second column of the adjacent row SWL1 (in S530). The above operation can be repeated for the remaining columns of the adjacent row SWL1 until column address CA reaches its maximum value ("No" in S550) (S540 to S576).
[0132] When the column address CA reaches its maximum value ("Yes" in S550), the flush management module 160 can change the value of the row address RA to specify the next adjacent row SWL5 and initialize the column address CA to 0 ("No" in S592 and S594). Therefore, the storage device 200 can read the data DIO from the storage cell connected to the first column of the next adjacent row SWL5 (in S530).
[0133] On the other hand, if the correctable error (CE) is a multi-bit error (CEm) excluding adjacent two-bit errors, or if the cumulative error count (#CE) in the currently flushed row exceeds the threshold ("Yes" in S574), then the fault analysis module 150 can generate a fault row address FADD corresponding to the adjacent row SWL1 (in S576). Then, the flush management module 160 can change the value of the row address RA to specify the next adjacent row SWL5 and initialize the column address CA to 0 ("No" in S592 and S594). Therefore, the storage device 200 can suspend the flushing operation on the adjacent row SWL1 and read the data DIO from the storage cell connected to the first column of the next adjacent row SWL5 (in S530).
[0134] If an uncorrectable error (UE) is detected in the read data DIO ("No (UE)" in S570), the fault analysis module 150 can record the error (in S582). Since the second type of designated flushing operation is performed again on data for which an uncorrectable error (UE) has already occurred during the on-demand flushing operation, the read data DIO will not be poisoned. The fault analysis module 150 can generate a fault row address FADD corresponding to the adjacent row SWL1 (in S576). Then, the flushing management module 160 can change the value of the row address RA to specify the next adjacent row SWL5 and initialize the column address CA to 0 ("No" in S592 and S594). Therefore, the storage device 200 can suspend the flushing operation on the adjacent row SWL1 and read the data DIO from the storage cell connected to the first column of the next adjacent row SWL5 (in S530).
[0135] Once the specified flushing operation is performed on all identified adjacent rows SWL1 to SWL8, SWL11, SWL19, and SWL27 ("Yes" in S592), a notification operation omitted during the on-demand flushing operation can be performed (in S596). That is, in the second type of specified flushing operation, regardless of the generation of the faulty row address FADD, a notification operation can be performed before the process ends to notify of data poisoning and a fatal error. Therefore, the reliability of the storage system can be improved. After this, the process can terminate.
[0136] The fault analysis module 150 can generate fault mode information F_MD by analyzing the fault line address FADD as a result of the adaptive flushing operation to determine the fault type between fault lines.
[0137] Figure 10 This is a block diagram illustrating a storage system including a storage module according to an embodiment of the present disclosure.
[0138] Reference Figure 10The storage system 2000 may include a storage module 2100 and a memory controller 2200.
[0139] The memory controller 2200 can control the overall operation of the storage system 2000 and control the data transmission between the host 2300 and the storage module 2100. The memory controller 2200 can generate a command / address signal C / A based on a request REQ from the host 2300, provide the command / address signal C / A to the storage module 2100, provide the data DIO corresponding to the request REQ from the host 2300 to the storage module 2100, and provide the data DIO read from the storage module 2100 to the host 2300.
[0140] Storage module 2100 may include a plurality of storage devices 2100_1 to 2100_10. In one embodiment, each of storage devices 2100_1 to 2100_10 may correspond to Figure 1 Storage device 200 as described herein.
[0141] In one embodiment, the memory controller 2200 may correspond to Figure 1 The memory controller 100 described herein. The memory controller 2200 may include: an error correction code (ECC) engine configured to check and correct errors in data output from the storage device and generate error information; a fault analysis module configured to analyze an error log generated based on the error information to generate and collect fault addresses, and to generate fault mode information by analyzing the relationships between the collected fault addresses; and a scrubbing management module configured to control the storage device to perform a first scrubbing operation and a second scrubbing operation, and to control the second scrubbing operation to perform on one or more adjacent addresses specified within the fault boundary based on the fault addresses generated according to the first scrubbing operation. As described above, in one embodiment of this disclosure, when a fault line is detected as a result of an on-demand scrubbing operation or a patrol scrubbing operation, additional specified scrubbing operations can be performed on adjacent lines specified within the fault boundary with reference to the detected fault line. Therefore, the fault type can be accurately determined, and errors that may develop into uncorrectable errors (UEs) can be identified based on the fault type, thereby achieving optimized RAS operation.
[0142] Figure 11 This is a block diagram illustrating a storage system including stacked storage devices according to an embodiment of the present disclosure.
[0143] Reference Figure 11 The storage system 3000 may include a packaging substrate 3140, an interposer 3130, a stacked storage device 3110, and a processor 3120.
[0144] The packaging substrate 3140 may include a printed circuit board (PCB). The packaging substrate 3140 may be electrically connected to an external system board, motherboard, or module board via bumps.
[0145] Interposer 3130 may be formed on package substrate 3140. Interposer 3130 may be a silicon substrate in which only wiring is formed.
[0146] The stacked storage device 3110 and processor 3120 can be formed on the interposer layer 3130. The stacked storage device 3110 and processor 3120 can be disposed on the interposer layer 3130 with spacing between them. Although Figure 11 A stacked storage device 3110 is shown, but embodiments of the present disclosure are not limited thereto, and one or more stacked storage devices may be formed on the interposer layer 3130 around the processor 3120.
[0147] Processor 3120 may include a memory controller (MC) 3121 and a physical interface circuit (PHY) 3122. The memory controller 3121 may be configured to control the stacked storage device 3110. The physical interface circuit 3122 may interface between the memory controller 3121 and the stacked storage device 3110. The physical interface circuit 3122 may be an interface circuit that converts signals transmitted from the memory controller 3121 into signals suitable for use in the stacked storage device 3110, and outputs signals transmitted from the stacked storage device 3110 into signals suitable for use in the memory controller 3121. Processor 3120 may be one of a variety of processors, such as a microprocessor unit (MPU), a central processing unit (CPU), a general-purpose processing unit (GPU), and a host processing unit (HPU).
[0148] The stacked memory device 3110 may include a lower chip 3114 and one or more upper chips 3112_0 to 3112_3 vertically stacked on an interposer layer 3130. An example of the stacked memory device 3110 formed by stacking multiple chips as described above is a high-bandwidth memory (HBM). Through-electrodes (TSVs) are formed between the lower chip 3114 and the upper chips 3112_0 to 3112_3, through which signals (i.e., commands, addresses, and data) can be transmitted between the chips.
[0149] The lower chip 3114 may include a physical interface circuit (PHY) 3116 for interfacing with the memory controller 3121.
[0150] In one embodiment, at least one of the memory controller 3121 and the lower chip 3114 may correspond to Figure 1The memory controller 100 described herein, and each of the upper chips 3112_0 to 3112_3 may correspond to Figure 1 The storage device 200 described herein. At least one of the memory controller 3121 and the lower chip 3114 may include: an error correction code (ECC) engine configured to check and correct errors in data output from the storage device and generate error information; a fault analysis module configured to analyze an error log generated based on the error information to generate and collect fault addresses, and to generate fault mode information by analyzing the relationships between the collected fault addresses; and a scrubbing management module configured to control the storage device to perform a first scrubbing operation and a second scrubbing operation, and to control the second scrubbing operation to be performed on one or more adjacent addresses specified within the fault boundary based on the fault addresses generated according to the first scrubbing operation. As described above, in one embodiment of this disclosure, when a fault line is detected as a result of an on-demand scrubbing operation or a patrol scrubbing operation, additional specified scrubbing operations can be performed on adjacent lines specified within the fault boundary with reference to the detected fault line. Therefore, the fault type can be accurately determined, and errors that may develop into uncorrectable errors (UEs) can be identified based on the fault type, thereby achieving optimized RAS operation.
[0151] Figure 12 This is a block diagram illustrating a mobile system including a storage device according to an embodiment of the present disclosure.
[0152] Reference Figure 12 The mobile system 4000 may include an application processor (AP) 4100, a storage device 4200, a network device 4300, a memory device 4400, and a user interface 4500.
[0153] The application processor 4100 can drive components, operating systems (OS), or user programs included in the mobile system 4000. For example, the application processor 4100 can be provided as a system-on-a-chip (SoC).
[0154] Storage device 4200 may function as main memory, operational memory, buffer memory, or cache of mobile system 4000. Storage device 4200 may include volatile random access memory (e.g., DRAM, SDRAM, DDR SDRAM, DDR2 SDRAM, DDR3 SDRAM, Low Power Double Data Rate 3 Synchronous Dynamic Random Access Memory (LPDDR3 SDRAM)) or non-volatile random access memory (e.g., phase-change RAM (PRAM), resistive random access memory (ReRAM), magnetoresistive random access memory (MRAM), ferroelectric RAM (FRAM), etc.). According to one embodiment, storage device 4200 may be configured as a reference. Figure 10 The storage module 2100.
[0155] Network device 4300 can communicate with external devices. For example, network device 4300 can support wireless communication, such as Code Division Multiple Access (CDMA), Global System for Mobile Communications (GSM), Wideband CDMA (WCDMA), CDMA-2000, Time Division Multiple Access (TDMA), Long Term Evolution (LTE), Global Microwave Access Interoperability (WiMAX), Wireless Local Area Network (WLAN), Ultra Wideband (UWB), Bluetooth, Wi-Fi, etc. For example, network device 4300 may be included in application processor 4100.
[0156] Memory device 4400 can store data. For example, memory device 4400 can store data received from application processor 4100. Alternatively, memory device 4400 can transfer stored data to application processor 4100. For example, memory device 4400 can be implemented as a non-volatile semiconductor memory device, such as phase-change RAM (PRAM), magnetic RAM (MRAM), resistive RAM (RRAM), NAND flash memory, and 3D NAND flash memory.
[0157] In one embodiment, the application processor 4100 and the storage device 4200 may respectively correspond to Figure 1 The memory controller 100 and storage device 200 are shown. The application processor 4100 may include: an error correction code (ECC) engine configured to check and correct errors in data output from the storage device and generate error information; a fault analysis module configured to analyze an error log generated based on the error information to generate and collect fault addresses, and to generate fault mode information by analyzing the relationships between the collected fault addresses; and a scrubbing management module configured to control the storage device to perform a first scrubbing operation and a second scrubbing operation, and to control the second scrubbing operation to be performed on one or more adjacent addresses specified within the fault boundary based on the fault addresses generated according to the first scrubbing operation. As described above, in one embodiment of this disclosure, when a fault line is detected as a result of an on-demand scrubbing operation or a patrol scrubbing operation, additional specified scrubbing operations can be performed on adjacent lines specified within the fault boundary with reference to the detected fault line. Therefore, the fault type can be accurately determined, and errors that may develop into uncorrectable errors (UEs) can be identified based on the fault type, thereby achieving optimized RAS operation.
[0158] While this disclosure has been described with reference to specific embodiments, it will be apparent to those skilled in the art that various changes and modifications can be made without departing from the spirit and scope of this disclosure as defined by the following claims. Furthermore, these embodiments can be combined to form other embodiments.
Claims
1. A memory controller, comprising: The ECC engine checks and corrects errors in the data output from the storage device and generates error messages. ECC stands for Error Correction Code. The fault analysis module includes: analyzing error logs based on the error information; generating and collecting fault addresses; and generating fault mode information by analyzing the relationships between the collected fault addresses; and A flushing management module, which: controls the storage device to perform a first flushing operation and a second flushing operation; and controls the second flushing operation to be performed based on the fault address generated according to the first flushing operation for one or more adjacent addresses specified within the fault boundary.
2. The memory controller as described in claim 1, in, The storage device includes a plurality of cell blocks arranged in an array along the row and column directions, and sharing a sub-word line driver in the row direction and a bit line sense amplifier in the column direction; as well as The brushing management module sets the fault boundary, which includes n adjacent unit groups in the column direction, where n is an integer greater than or equal to 2. Each unit group is formed by unit blocks arranged along the row direction among the plurality of unit blocks.
3. The memory controller as claimed in claim 1, wherein, When the fault address specifies the k-th sub-word of the main word line, the flushing management module generates one or more adjacent addresses, where k is an integer, and the one or more adjacent addresses include at least one of the following: Sub-word lines located within the same main word line as the k-th sub-word line, The kth sub-word adjacent to the main word line, and The kth sub-word line of all main word lines within the fault boundary.
4. The memory controller as claimed in claim 3, wherein, The brushing management module generates one or more adjacent addresses by specifying a sub-word line that is arranged in the same unit block as the k-th sub-word line.
5. The memory controller as claimed in claim 3, wherein, The fault analysis module: When only the kth sub-word line fails, generate the fault mode information for the sub-word line driver that drives the kth sub-word line; When an odd or even number of sub-word lines within the same main word line fail, fault mode information is generated for the common signal lines associated with the main word line. When the kth sub-word line of the adjacent main word line fails, fault mode information is generated for the contact shared by the kth sub-word line; as well as When the kth sub-word line of all main word lines fails, fault mode information is generated for the signal path that is commonly provided to the sub-word line driver, which drives the kth sub-word line.
6. The memory controller of claim 1, further comprising: The RAS manager instructs the storage device or requests the host to perform operations to improve the reliability of the storage device based on the failure mode information. RAS stands for Reliability, Availability and Serviceability.
7. The memory controller as claimed in claim 1, wherein, The first scrubbing operation includes at least one of the following: A sequential flushing operation performed on all rows of the storage device at predetermined intervals, and an on-demand flushing operation performed during a read operation on a target row in response to a request from the host.
8. The memory controller as claimed in claim 1, wherein, The second scrubbing operation includes: A specified flushing operation performed sequentially on one or more rows specified by one or more adjacent addresses.
9. A method of operating a storage system, comprising: A first flushing operation is performed on a target row among multiple rows of the storage device and a target column among multiple columns of the storage device; as well as Based on the result of the first scrubbing operation, a second scrubbing operation is performed on one or more adjacent rows specified within the fault boundary of the target row.
10. The operating method as described in claim 9, further comprising: Fault mode information is generated by analyzing the relationship between the target row and the faulty row detected as a result of the second scrubbing operation; as well as Based on the fault mode information, the storage device is instructed or the host is requested to perform an operation to improve the reliability of the storage device.
11. The operating method as described in claim 9, further comprising: When the storage device comprises a plurality of cell blocks arranged in an array along the row and column directions, a sub-word line driver is shared in the row direction and a bit line sense amplifier is shared in the column direction; The fault boundary is defined, comprising n adjacent cell groups in the column direction, where n is an integer greater than or equal to 2, and each cell group is formed by cell blocks arranged along the row direction from the plurality of cell blocks.
12. The operating method as described in claim 9, wherein, The first scrubbing operation includes: Read data from the target storage unit specified by the target row and the target column; When the read data contains correctable multi-bit errors other than errors in adjacent two-bit pairs, a request is made to perform the second flushing operation of the first type on the one or more adjacent rows; and When the read data contains uncorrectable errors, a second type of second scrubbing operation is requested for the one or more adjacent rows.
13. The operating method as described in claim 12, wherein, The second scrubbing operation of the first type and the second type includes: Identify one or more adjacent rows within the fault boundary based on the target row; and The following process is repeated for the one or more adjacent rows by sequentially reading data from the storage cell of one of the adjacent rows: when the read data contains the correctable multi-bit error other than the adjacent two-bit error, or when the cumulative error count of a row exceeds a threshold, or when the read data contains the uncorrectable error, the row is identified as a faulty row.
14. The method of operation as described in claim 13, further comprising a second scrubbing operation for the first type: When an error in the read data is correctable, the read data is corrected, and the corrected read data is written back to the storage unit; and When the error in the read data is uncorrectable, the read data is marked as poisoned.
15. The method of operation as described in claim 13, further comprising the second scrubbing operation for the second type: After repeating the process on one or more adjacent rows, the host is notified that poisoned data has been detected.
16. The operating method as described in claim 9, further comprising: A third flushing operation is performed sequentially on the plurality of rows of the storage device at predetermined intervals; as well as Based on the faulty rows detected by the third scrubbing operation, the second scrubbing operation is performed on one or more adjacent rows specified within the fault boundary.
17. The operating method as described in claim 9, further comprising: Upon power-up, based on pre-stored fault lines, the second scrubbing operation is performed on one or more adjacent lines specified within the fault boundary.
18. A method of operating a storage system, comprising: The first flushing operation is performed sequentially on multiple rows of the storage device at predetermined intervals; as well as Based on the faulty rows detected as a result of the first scrubbing operation, a second scrubbing operation is performed on one or more adjacent rows specified within the fault boundary.
19. The operating method as described in claim 18, further comprising: Fault mode information is generated by analyzing the relationship between the result of the first scrubbing operation and the fault lines detected as the result of the second scrubbing operation. as well as Based on the fault mode information, the storage device is instructed or the host is requested to perform an operation to improve the reliability of the storage device.
20. The operating method as described in claim 18, further comprising: When the storage device comprises a plurality of cell blocks arranged in an array along the row and column directions, a sub-word line driver is shared in the row direction and a bit line sense amplifier is shared in the column direction; The fault boundary is defined, comprising n adjacent cell groups in the column direction, where n is an integer greater than or equal to 2, and each cell group is formed by cell blocks arranged along the row direction from the plurality of cell blocks.
21. The operating method as described in claim 18, wherein, The first scrubbing operation includes: Data is read sequentially from the storage unit connected to a selected row among the plurality of rows; When the read data contains correctable multi-bit errors other than adjacent two-bit errors, or when the accumulated error count exceeds a threshold, a request is made to perform the second flushing operation on the one or more adjacent rows; and When the read data contains uncorrectable errors, a request is made to perform the second scrubbing operation on the one or more adjacent rows.
22. The operating method as described in claim 18, wherein, The second scrubbing operation includes: Identify one or more adjacent rows within the fault boundary based on the faulty row; and The following process is repeated for the one or more adjacent rows by sequentially reading data from the storage cell of one of the adjacent rows: if the read data contains correctable multi-bit errors other than adjacent two-bit errors or the accumulated error count exceeds a threshold, or when the read data contains uncorrectable errors, a row is identified as a faulty row.
23. The operating method as described in claim 22, further comprising: When the error in the read data is correctable, the read data is corrected, and the corrected read data is written back to the storage unit; as well as When the errors in the read data are uncorrectable, the read data is marked as poisoned.
24. The operating method as described in claim 18, further comprising: In response to a request from the host, a third flush operation is performed on the target row and target column during the read operation; as well as Based on the target row detected by the third scrubbing operation, the second scrubbing operation is performed on one or more adjacent rows specified within the fault boundary.
25. The operating method as described in claim 18, wherein Upon power-up, based on pre-stored fault lines, the second scrubbing operation is performed on one or more adjacent lines specified within the fault boundary.
