CN106874139A - It is a kind of for error checking and the decoding method corrected, device and processor - Google Patents

Abstract

This application discloses a kind of for error checking and the decoding method corrected, device and processor, wherein, the coding method includes：Storage control analyzes the significance level of the data when the write instruction of data is received；When the significance level of the data belongs to the first importance level, BCH and LDPC is respectively adopted the data are encoded, or use the LDPC of not higher than the first code check to encode the data；When the significance level of the data belongs to the second importance level, the data are encoded using the BCH or higher than the LDPC of first code check；The data after by coding are written in memory, and preserve the coding information of the data.Through the above way, code encoding/decoding mode can intelligently be configured for different pieces of information.

Description

A codec method, device and processor for error checking and correction

technical field

The present application relates to the field of storage technologies, in particular to a codec method, device and processor for error checking and correction.

Background technique

NAND memory is currently a widely used data storage tool. During the process of data transmission and storage inside the NAND memory, the data may be damaged due to hardware failure, software failure, hard disk error, etc. of the memory. Therefore, in order to ensure the integrity of user data, it is necessary to perform error checking and correction (Error Correcting Code, ECC for short) on the stored data.

Generally, when writing data into a NAND memory, the data needs to be encoded so that the encoded data can implement error checking and correction, and then the data that can implement error checking and correction can be written into the memory. When the data needs to be read out, the stored data is correspondingly decoded to realize error checking and correction, and then the accurate data obtained after the error checking and correction is output.

However, the existing memory always adopts a codec mode fixedly, and the codec mode cannot be flexibly configured for different data.

Contents of the invention

The present application provides a codec method, device and processor for error checking and correction, which can intelligently configure codec modes for different data.

The first aspect of the present application provides an encoding method for error checking and correction, including: when the storage controller receives a data write instruction, analyze the importance of the data; when the importance of the data belongs to the first When an important level, adopt BCH and low-density parity-check code LDPC to encode described data respectively, or adopt the LDPC that is not higher than the first coding rate to encode described data; When two important levels are used, the data is encoded by using the BCH or LDPC higher than the first code rate, wherein the importance of the first important level is higher than that of the second important level; and writing the coded data into the memory, and saving the coding information of the data, wherein the coding information of the data includes the coding method adopted by the data.

Wherein, the step of analyzing the importance of the data includes: analyzing at least one of the source, attribute and type of the data to obtain the importance of the data; when the importance exceeds the first determining that the importance of the data belongs to the first level; and when the value of the importance does not exceed the first preset value, determining that the importance of the data belongs to the second level.

Wherein, the second importance level includes a first importance sub-level and a second importance sub-level; when the value of the importance level does not exceed the first preset value, it is determined that the importance level of the data belongs to the second The step of the importance level includes: when the importance degree value does not exceed the first preset value and exceeds the second preset value, determining that the importance degree of the data belongs to the first importance sub-level; When the degree value does not exceed the second preset value, it is determined that the importance of the data belongs to the second important sub-level, wherein the first preset value is greater than the second preset value; and the When the importance of the data belongs to the second importance level, the step of encoding the data by using the BCH or LDPC higher than the first code rate includes: when the importance of the data belongs to the In the first important sub-level, the data is encoded by using BCH higher than the preset parity number or LDPC higher than the first code rate and lower than the second code rate, in the importance of the data When it belongs to the second important sub-level, the data is encoded by using BCH not higher than the preset parity number or LDPC not lower than the second code rate, wherein the first code The rate is higher than the second code rate.

Wherein, the memory is a NAND memory of SLC, MLC or TLC type; before the step of writing the encoded data into the memory and saving the encoding information of the data, the method also includes: Determine the type of the memory, and when the type of the memory is SLC, the value of q in the finite field GF (2^q) in LDPC decoding is 1; when the type of the memory is MLC , set the value of q to 2; when the type of the memory is TLC, set the value of q to 3; and the encoding information of the data further includes the value of q.

Wherein, the generation matrix of the BCH is selected according to the check digit of the BCH, and/or the generation matrix of the LDPC is selected according to the code rate of the LDPC.

The second aspect of the present application provides a decoding method for error checking and correction, including: when a memory controller receives a data read instruction, reads the data and the encoding information of the data from the memory, wherein , the encoding information includes the encoding method adopted by the data, the encoding method adopted by the data depends on the importance of the data, and when the importance of the data is the first importance level, BCH and low-density parity are used A combination of check code LDPC or an encoding method of LDPC not higher than the first code rate; when the importance of the data is the second important level, the BCH or LDPC higher than the first code rate is used encoding method, the importance of the first importance level is greater than the importance of the second importance level; determine the encoding method of the data according to the encoding information, and use the decoding method corresponding to the encoding method to encode the Data is decoded, wherein the decoding method includes a combination of BCH and LDPC, LDPC not higher than the first code rate, BCH or LDPC higher than the first code rate; and outputting the data obtained after the decoding .

Wherein, the second importance level includes a first important sublevel and a second important sublevel, and the importance of the first important sublevel is greater than that of the second important sublevel; the encoding method adopted by the data It also includes the encoding of BCH higher than the preset check digit or LDPC higher than the first code rate and lower than the second code rate when the importance of the data is the first important sub-level mode, when the importance of the data is at the second important sub-level, the encoding mode of BCH not higher than the preset check digit or LDPC not lower than the second code rate is adopted, wherein , the first code rate is higher than the second code rate; and the BCH or the LDPC decoding method higher than the first code rate includes BCH higher than the preset check digit, higher than the second code rate A decoding method of LDPC with a code rate lower than the second code rate, BCH with the preset parity number not higher than the preset check digit, or LDPC with the second code rate not lower than the second code rate.

Wherein, after the step of determining the encoding method of the data according to the encoding information, and decoding the data by using a decoding method corresponding to the encoding method, the method further includes detecting the BCH When the number of erroneous bits exceeds a preset bit value and/or the actual number of iterations of the LDPC exceeds a preset number, an alarm is sent out that there is a problem with the block storing the data in the memory.

Wherein, the memory is a NAND memory of SLC, MLC or TLC type; the encoding information of the data also includes the value of q, when the type of the memory is SLC, the value of q is 1, when the When the type of the memory is MLC, the value of q is 2; when the type of the memory is TLC, the value of q is 3; In the step of decoding data, if the LDPC decoding is performed, the value of q in the encoded information is used as the value of q in the finite field GF(2^q) in the LDPC decoding, to perform the LDPC decoding.

Wherein, the check matrix of the BCH is selected according to the number of check digits of the BCH, and/or the check matrix of the LDPC is selected according to the code rate of the LDPC.

The third aspect of the present application provides an encoding device for error checking and correction, including: an analysis module, configured to analyze the importance of the data when receiving a data write instruction; an encoding module, configured to When the importance of the data belongs to the first importance level, respectively adopt BCH and low density parity check code LDPC to encode the data, or adopt LDPC not higher than the first code rate to encode the data; When the importance of the data belongs to the second importance level, the data is encoded by using the BCH or LDPC higher than the first code rate, wherein the importance of the first importance level is higher than that of the second The importance of two important levels; and a writing module, configured to write the encoded data into the memory, and store the encoding information of the data, wherein the encoding information of the data includes the encoding information used by the data Encoding.

The fourth aspect of the present application provides a decoding device for error checking and correction, including: a reading module, used to read the data and the Data encoding information, wherein the encoding information includes the encoding method adopted by the data, the encoding method adopted by the data depends on the importance of the data, and when the importance of the data is the first importance level, A combination of BCH and low-density parity-check code LDPC or an LDPC encoding method not higher than the first code rate is used; when the importance of the data is the second important level, the BCH or higher than the first code rate is used In an LDPC encoding method with a code rate, the importance of the first importance level is greater than the importance of the second importance level; the decoding module is used to determine the encoding method of the data according to the encoding information, and adopt the same method as The decoding method corresponding to the encoding method decodes the data, wherein the decoding method includes a combination of BCH and LDPC, LDPC not higher than the first code rate, BCH or LDPC higher than the first code rate ; and an output module, configured to output the data obtained after the decoding.

A fifth aspect of the present application provides a codec processor for error checking and correction, including the above encoding device and/or the above decoding device.

In the above solution, the storage controller analyzes the importance of the data, and configures different encoding and decoding methods for the data according to different degrees of importance, that is, the more important the data, the stronger the encoding and decoding capabilities of error checking and correction. It realizes the intelligent configuration of memory for data encoding and decoding, and ensures the reliable and accurate storage of important data. At the same time, the storage controller selects encoding and decoding methods with low error checking and correcting capabilities for data with low importance, which reduces the required encoding capability and check code length, improves encoding efficiency and saves storage space.

Description of drawings

Fig. 1 is a flow chart of an embodiment of an error checking and correcting coding method used in the present application;

Fig. 2 is a flow chart of another embodiment of the encoding method used for error checking and correction in the present application;

Fig. 3 is a schematic diagram of encoding information stored in another embodiment of the encoding method for error checking and correction of the present application.

Fig. 4 is a flowchart of an embodiment of the error checking and correcting decoding method used in the present application;

Fig. 5 is a flow chart of another embodiment of the decoding method used for error checking and correction in the present application;

Fig. 6 is a schematic diagram of LLR probability during LDPC decoding in another embodiment of the decoding method for error checking and correction in the present application;

Fig. 7a is a schematic structural diagram of an embodiment of an encoding device used for error checking and correction in the present application;

Fig. 7b is a schematic structural diagram of an embodiment of a compiling module of an encoding device for error checking and correction according to the present application.

FIG. 8 is a schematic structural diagram of an embodiment of a decoding device used for error checking and correction in the present application.

detailed description

In the following description, for purposes of illustration rather than limitation, specific details, such as specific system architectures, interfaces, and techniques, are set forth in order to provide a thorough understanding of the present application. It will be apparent, however, to one skilled in the art that the present application may be practiced in other embodiments without these specific details. In other instances, detailed descriptions of well-known devices, circuits, and methods are omitted so as not to obscure the description of the present application with unnecessary detail.

Please refer to FIG. 1 . FIG. 1 is a flowchart of an embodiment of an encoding method for error checking and correction in this application. This implementation manner may be executed by a storage controller in a memory, and the method includes:

S11: When the storage controller receives a data write instruction, analyze the importance of the data.

The storage controller first divides the data to be written into importance levels, such as including the first importance level and the second importance level, and selects a corresponding encoding method according to the importance of the data. For example, the memory is used as the storage device of the computer, and when the storage controller receives a data write instruction issued by the computer file system, it receives the data and analyzes whether the importance of the data belongs to the first important level or the second important level.

Specifically, the storage controller may determine the importance of the parameter according to the analyzed data. For example, importance values corresponding to different data parameters are prestored in the memory. When receiving a data write instruction, the storage controller searches the above-mentioned pre-stored information for the importance value corresponding to the parameter of the data to obtain the data's importance value, and judges whether the data's importance value exceeds the first a preset value, when the importance degree value exceeds the first preset value, it is determined that the importance degree of the data belongs to the first importance level; and when the importance degree value does not exceed the first preset value, Determining the importance of the data belongs to the second importance level. Wherein, the parameters of the data may include at least one of the source, attribute and type of the data.

S12: When the importance of the data belongs to the first importance level, the storage controller uses BCH and LDPC to encode the data respectively, or uses LDPC not higher than the first code rate to encode the data; When the importance of the data belongs to the second importance level, the data is encoded by using the BCH or LDPC higher than the first code rate, wherein the importance of the first importance level is higher than the The importance of the second level of importance.

The storage controller of this application can adopt different encoding methods according to the importance of data. Among them, the encoding methods that can be used include BCH (from the abbreviation of Bose, Ray-Chaudhuri and Hocquenghem), Low Density Parity Check Code (English: Low Density Parity Check Code, LDPC for short), or a combination of BCH and LDPC. The order of the error checking and correcting capabilities of the above three encoding methods from strong to weak is: the combination of BCH and LDPC, LCPD, and BCH. Moreover, the storage controller can adaptively adopt LDPC with different code rates (English: code rate). For the original data with a certain code length, when the code rate of LDPC is higher, more check codes will be obtained after encoding, and the error detection and correction capabilities will be stronger.

In this embodiment, the storage controller adopts a coding method with stronger error checking and correcting capabilities for data with a higher importance level. Specifically, when the storage controller judges that the importance of the data to be written belongs to the first importance level, BCH and LDPC (that is, the aforementioned combination of BCH and LDPC) are used to encode the data respectively, wherein the LDPC The code rate can be lower than or higher than the first code rate, and the order of BCH and LDPC encoding is not limited, such as first performing BCH encoding and then performing LDPC encoding on the output of BCH encoding, or performing LDPC encoding first and then performing LDPC encoding on the output of BCH encoding The output of the LDPC code is BCH coded; or when the storage controller judges that the importance of the data to be written belongs to the first importance level, the data is coded with LDPC not higher than the first code rate. When the storage controller judges that the importance of the data to be written belongs to the second important level, the data is encoded by using the BCH, or the data is encoded by LDPC higher than the first code rate . In practical applications, the first code rate can be configured by the user and stored in the memory, and the first code rate is usually 0.2-0.7, specifically 0.2, 0.3, 0.5, 0.6 or 0.7. When using the above encoding method, for the data of the same importance level, the storage controller can use different code rates according to the importance value of the data. For example, the higher the importance value of the data, the lower the LDPC code rate it adopts. . In one embodiment, the memory can pre-store 16 code rates such as 0.5, 0.6, 0.650.7, 0.75, 0.8, 0.85, 0.9, 0.95, etc., and the storage controller selects one of the code rates according to its importance. Of course, the storage controller can also use the same code rate for all data of the same importance level.

S13: The storage controller writes the encoded data into the memory, and saves encoding information of the data, where the encoding information of the data includes an encoding mode adopted by the data.

After the data to be written is encoded according to the above-mentioned encoding method of S12, the obtained encoded data includes original data and a check code, and the encoded data is written into the memory. Moreover, in this embodiment, different encoding modes can be configured according to different data. In order to facilitate decoding, the storage controller also saves the encoding information of the data in the memory, or in a software layer such as a new technology file system (English: New Technology File System, referred to as: NTFS), or a non-sorted block image file system (English: Unsorted Block Image File System, referred to as: UBIFS). The storage controller may also establish a mapping relationship between the data and its encoding information in a list manner, so as to associate the data and its encoding information. When the encoded data needs to be decoded, the storage controller can read the corresponding encoding information to obtain its encoding mode, and decode it in a decoding mode corresponding to the encoding mode.

Please refer to FIG. 2 . FIG. 2 is a flow chart of another embodiment of an encoding method for error checking and correction in the present application. This method comprises the following steps:

S21: When the storage controller receives the data write instruction, analyze at least one of the source, attribute and type of the data to obtain an importance value of the data.

For example, the memory pre-stores importance values corresponding to different sources, attributes and types of the data. If the data source is an administrator account, its importance value is 80%; if the data source is a tourist account, its importance value is 30%; if the data type is a boot loader (English: boot loader), its importance value The value is 100%, and the data type is video or audio, then its importance value is 20%. The storage controller looks up the pre-stored information to obtain the value of the importance of the data. Wherein, when the data parameters analyzed by the storage controller include two or more types, the average value of the importance values corresponding to the two or more data parameters is used as the data importance value. If the data analyzed by the storage controller is the video of the administrator account, the value of the data importance is (80%+30%)/2=55%.

S22: When the importance degree value exceeds a first preset value, the storage controller determines that the importance degree of the data belongs to the first importance level; and when the importance degree value does not exceed the first preset value, The storage controller determines that the importance of the data belongs to the second importance level.

In this embodiment, the first preset value is used as the boundary of the importance level, and the data is divided into two main importance levels: the first importance level and the second importance level. Further, the second importance level in this embodiment may also include the first importance sub-level and the second importance sub-level. In the above S22, when the importance value does not exceed the first preset value, the step of determining that the importance of the data belongs to the second importance level includes: when the importance value does not exceed the first preset value When a preset value exceeds a second preset value, it is determined that the importance of the data belongs to the first sub-level of importance; when the value of the importance does not exceed the second preset value, it is determined that the data The degree of importance of belongs to the second importance sub-level, wherein the first preset value is greater than the second preset value. The first preset value and the second preset value can be set by the user. The first preset value may be 50% to 80%, and the second preset value may be 30% to 60%.

For example, the first preset value is 60%, and the second preset value is 40%. The storage controller obtains from S21 that the importance value of the data to be written is 55%, not exceeding 60% but exceeding 40%, so it is determined that the data belongs to the second important level and belongs to the first important sub-level in the second important level .

S23: When the importance of the data belongs to the first importance level, the storage controller uses BCH and LDPC to encode the data respectively, or uses LDPC not higher than the first code rate to encode the data; When the importance of the data belongs to the first important sub-level, the data is processed by using BCH higher than the preset check digit or LDPC higher than the first code rate and lower than the second code rate encoding, when the importance of the data belongs to the second important sub-level, the data is processed by using BCH not higher than the preset parity number or LDPC not lower than the second code rate encoding, wherein the importance of the first importance level is higher than that of the second importance level, and the first code rate is higher than the second code rate.

Based on the previous implementation manner, the storage controller can also adaptively adopt BCHs with different check digits. The more check digits the BCH uses, the stronger its error detection and correction capabilities will be. The preset check digit can be 50bit to 500bit. The first code rate usually takes a value of 0.2-0.7, and the second code rate usually takes a value of 0.7-0.95.

For example, the default check bit is 80 bits, the first code rate is 0.6, and the second code rate is 0.8. When the data to be written belongs to the first important level, the storage controller uses BCH and LDPC to encode the data respectively, or uses LDPC with a code rate not higher than 0.6 to encode the data; When the degree of importance belongs to the first important sub-level, the data is encoded using a BCH higher than 80 bit check bits, or using LDPC with a code rate higher than 0.6 and lower than 0.8 to encode the data; When the importance of the data belongs to the second important sub-level, the data is encoded by using BCH with a parity bit not higher than 80 bits, or by LDPC with a code rate not lower than 0.8. When using the above encoding method, the storage controller can specifically use different code rates and different check digits according to the importance of the data, such as the higher the importance of the data, the lower the code rate of LDPC it adopts, and The more check digits the BCH adopts.

Wherein, the generation matrix of the BCH is selected according to the check digit of the BCH, and/or the generation matrix of the LDPC is selected according to the code rate of the LDPC. For example, the memory may pre-store a plurality of BCH generator matrices G1 corresponding to different parity digits, and a plurality of LDPC generator matrices G2 corresponding to different code rates. When the storage controller determines the encoding method for the data to be written, it can read the corresponding BCH generation matrix G1 to perform corresponding BCH encoding, and/or read Take the generator matrix G2 of LDPC for corresponding LDPC encoding. Further, after the storage controller reads the generated matrix, it can analyze whether there is a rule in the generated matrix, and if there is a rule, divide the generated matrix into several small matrices, and use the several small matrices to encode the data, so as to Improve coding efficiency.

Optionally, due to different encoding modes, the number of check code bits obtained is different. Generally, the stronger the error correction capability of the encoding mode, the more check code bits. Therefore, the storage controller can also check the size of the storage space of the check code configured by the memory itself when determining the encoding method according to the degree of importance, and estimate whether the storage space of the check code is sufficient to store the selected encoding method to be written. The check code obtained by encoding the data, if it can, then determine that the encoding method is used for encoding, otherwise, increase the code rate of the LDPC and/or reduce the check digit of the BCH to ensure the storage space of the check code The size matches.

S24: The storage controller determines the type of the memory, and when the type of the memory is SLC, sets the value of q in the finite field GF(2^q) in LDPC decoding to 1; when the memory When the type of memory is MLC, the value of q is 2; when the type of the memory is TLC, the value of q is 3.

In this embodiment, the memory may be a single-level cell (English: Single-Level Cell, referred to as: SLC), a multi-level cell (English: Multi-Level Cell, referred to as: MLC) or a triple-level cell (English: Trinary-Level_Cell , referred to as: TLC) type NAND memory, the NAND memory specifically can be Raw Nand, or solid state hard drive (English: Solid State Drives, referred to as: SSD) and the like. Since the SLC type stores 1 bit for each unit, it needs to be divided into 2 ¹ = 2 voltage state representations; the MLC type stores 2 bits for each unit, and needs to be divided into 2 ² = 4 voltage state representations; the TLC type stores 2 bits for each unit 3 bits need to be divided into 2 ³ =8 voltage state representations. Correspondingly, SLC, MLC, and TLC type memories have different and sequentially increasing requirements for error checking and correction. Therefore, for the error checking and correction requirements of different types of memory, the memory controller can predetermine the type of the memory, and determine the q in the finite field GF(2^q) in the LDPC decoding by the type of the memory Specifically, if the SLC type, set q to 1, if the MLC type, set q to 2, and if the TLC type, set q to 3.

It can be understood that S24 may not be performed after S21-23 is performed, and in other embodiments, the storage controller may perform it at any time between receiving the data writing instruction and step S25.

S25: The storage controller writes the encoded data into the memory, and saves the encoding information of the data, where the encoding information of the data includes the encoding method adopted by the data and the value of q .

For example, the encoding method in the encoding information can be represented by a code group consisting of: "whether to use LDPC", "code rate of LDPC", "whether to use BCH", "check digit of BCH". Wherein, whether to adopt LDPC and whether to adopt BCH can be represented by but not limited to 1 bit, for example, 0 means yes, and 1 means no. The value of q can be represented by but not limited to 1bit or 2bit, for example, 01 represents that q is 1, 10 represents that q is 2, and 11 represents that q is 3. The code rate of LDPC and the check digit of BCH can be represented by but not limited to 2bit-6bit, for example, 4bit. Optionally, the code group used to record the encoding information can also reserve 1 bit-2 bits for extension.

It can be understood that the code group used to record the coding information can be adaptively set according to the actual coding method. For example, if in one embodiment, the encoding method adopted by the storage controller is only a combination of BCH and LDPC, or LDPC with different code rates, then the code group used to record the encoding information can be as shown in Figure 3 The format, that is, does not include the item "whether to use LDPC" and the item "BCH check digit", and the encoding information is 1 byte per page.

When the storage controller receives the instruction to read the data, it can determine the encoding method adopted by the data according to the encoding information, and decode it according to the corresponding decoding method.

In the above solution, the storage controller analyzes the importance of data and configures different encoding methods for the data according to different degrees of importance, that is, the more important the data, the stronger the encoding method for error checking and correcting capabilities is, and the data encoding by the memory is realized. The adaptive configuration of the mode ensures the reliable and accurate storage of important data. At the same time, due to the encoding method with stronger error checking and correcting capability, the required encoding capability is stronger and the required check code length is longer. The storage controller selects a coding method with low error checking and correcting capabilities for data with low importance, which reduces its required coding capability and check code length, improves coding efficiency and saves storage space.

Please refer to FIG. 4 . FIG. 4 is a flowchart of an embodiment of a decoding method for error checking and correction in this application. In this implementation method, the data decoded by the storage controller may be the data obtained by encoding in the above implementation manner. The method includes the following steps:

S41: The storage controller reads the data and the encoding information of the data from the memory when receiving a data read instruction.

For example, the memory is used as the storage device of the computer. When the storage controller receives the data reading instruction issued by the computer file system, it obtains the storage address carried in the reading instruction, and reads the data from the storage address of the memory. data, and the coded information that has a mapping relationship with the data is found by looking up the table.

Wherein, the encoding information includes the encoding method adopted by the data, and the encoding method adopted by the data depends on the importance of the data. When the importance of the data is the first importance level, BCH and low-density A combination of parity check code LDPC or an encoding method of LDPC not higher than the first code rate; when the importance of the data is the second important level, the BCH or LDPC higher than the first code rate is used In a coding manner, the importance degree of the first importance level is greater than the importance degree of the second importance level. For example, the data and encoding information are the encoded data and stored encoding information in the embodiment shown in FIG. 1 .

S42: The storage controller determines the encoding method of the data according to the encoding information, and uses a decoding method corresponding to the encoding method to decode the data, where the decoding method includes a combination of BCH and LDPC, not LDPC and BCH higher than the first code rate or LDPC higher than the first code rate.

Specifically, when the storage controller determines that the encoding method of the data is a combination of BCH and LDPC, the corresponding decoding method is a combination of BCH and LDPC, that is, the data is decoded by using BCH and LDPC respectively. Wherein, the order of performing BCH and LDPC decoding is not limited, for example, BCH decoding may be performed first and then LDPC decoding is performed on the output of BCH decoding, or LDPC decoding may be performed first and then BCH decoding may be performed on the output of LDPC decoding. When the storage controller determines that the encoding method of the data is LDPC encoding with a code rate not higher than the first code rate, the corresponding decoding method is LDPC decoding with a code rate not higher than the first code rate. When the storage controller determines that the encoding method of the data is BCH encoding, its corresponding decoding method is BCH decoding not higher than the first code rate. When the storage controller determines that the encoding method of the data is LDPC encoding higher than the first code rate, its corresponding decoding method is LDPC decoding higher than the first code rate.

Wherein, the encoding information obtained in S41 includes the code rate of LDPC, and the storage controller uses the code rate in the encoding information as the code rate during LDPC decoding.

S43: The storage controller outputs the data obtained after performing the decoding.

Since the decoding method in this embodiment corresponds to the encoding method of the data, a codec method with higher error checking and correction capabilities is selected for data with a high degree of importance, and a codec method with lower error detection and correction capabilities is selected for data with a lower degree of importance. , which realizes the self-adaptive configuration of the memory for data encoding and decoding, and ensures the reliable and accurate storage of important data. At the same time, the decoding capability required for data with low importance is also reduced, and the decoding efficiency is improved.

Please refer to FIG. 5 . FIG. 5 is a flow chart of another embodiment of a decoding method for error checking and correction in the present application. This method comprises the following steps:

S51: When the storage controller receives a data read instruction, read the data and the encoding information of the data from the memory.

Wherein, the encoding information includes the encoding method adopted by the data, and the encoding method adopted by the data depends on the importance of the data as the first importance level or the second importance level, and the second importance level includes the first The most important child and the second most important child. When the importance of the data is the first importance level, a combination of BCH and low-density parity check code LDPC or an encoding method of LDPC not higher than the first code rate is used; when the importance of the data is the first The first important sub-level sometimes adopts a BCH encoding method higher than the preset check digit or an LDPC encoding method higher than the first code rate and lower than the second code rate, when the importance of the data is the The second important sub-level sometimes adopts the BCH encoding method not higher than the preset parity number or the LDPC encoding method not lower than the second code rate, and the importance of the first important level is greater than that of the second important level. The importance of two important levels, the importance of the first important sub-level is greater than the importance of the second important sub-level, and the first code rate is higher than the second code rate. For example, the data and coded information are the data encoded by the embodiment shown in Figure 2 and the stored coded information

S52: The storage controller determines the encoding method of the data according to the encoding information, and uses a decoding method corresponding to the encoding method to decode the data, where the decoding method includes a combination of BCH and LDPC, not LDPC higher than the first code rate, BCH higher than the preset check digit, LDPC higher than the first code rate and lower than the second code rate, and not higher than the preset check digit A decoding method of BCH or LDPC with a code rate not lower than the second code rate.

Based on the embodiment shown in Figure 4, the BCH in this embodiment or the LDPC decoding method higher than the first code rate includes BCH higher than the preset check digit, higher than the first code rate and lower than A decoding method of LDPC with a second code rate, BCH with a check digit not higher than the preset check digit, or LDPC with a code rate not lower than the second code rate. Specifically, when the storage controller determines that the encoding method of the data is BCH encoding with a higher than the preset check digit, the corresponding decoding method is BCH decoding with a higher than the preset check digit. When the storage controller determines that the encoding method of the data is LDPC encoding that is higher than the first code rate and lower than the second code rate, its corresponding decoding method is higher than the first code rate and lower than LDPC decoding of the second code rate. When the storage controller determines that the encoding method of the data is BCH encoding not higher than the preset check digit, the corresponding decoding method is BCH decoding not higher than the preset check digit. When the memory controller determines that the encoding method of the data is LDPC encoding with a code rate not lower than the second code rate, the corresponding decoding method is LDPC decoding with a code rate not lower than the second code rate.

Wherein, the check matrix of the BCH is selected according to the number of check digits of the BCH, and/or the check matrix of the LDPC is selected according to the code rate of the LDPC. For example, the memory may pre-store a plurality of BCH parity check matrices H1 corresponding to different parity digits, and a plurality of LDPC parity check matrices H2 corresponding to different code rates. When the storage controller determines the decoding mode for the data to be written, it can read the corresponding BCH parity check matrix H1 to perform corresponding BCH decoding corresponding to the determined code rate and/or parity digits, and/or Read the check matrix H2 of LDPC to perform corresponding LDPC decoding.

S53: When the storage controller detects that the number of erroneous bits of the BCH exceeds a preset bit value and/or the actual number of iterations of the LDPC exceeds a preset number of times, it issues an alarm that there is a problem with the block storing the data in the memory .

In this embodiment, the storage controller may also give an early warning to the block storing the data. When the storage controller performs BCH decoding, it detects whether the number of erroneous bits of the BCH exceeds the preset bit value. If it performs LDPC decoding, it detects whether the actual number of iterations of the LDPC exceeds the preset number of times. If one of them is exceeded, an alarm is issued to indicate that there is a problem with the block storing the data in the memory, for example, an alarm message containing the number of error bits of the BCH or the actual number of iterations of LDPC is sent to the upper computer system. Wherein, the preset bit value and the preset number of times can be set by the user, and the preset bit value can specifically be but not limited to any bit value in 1/1000 to 5/100 of the data amount of the data, and the preset number of times It can be but not limited to 2-10 times.

Wherein, the memory may be a NAND memory of SLC, MLC or TLC type, specifically, the NAND memory may be Raw Nand, SSD or the like. The encoding information of the data also includes the value of q, when the type of the memory is SLC, the value of q is 1, when the type of the memory is MLC, the value of q is 2, when When the type of the memory is TLC, the value of q is 3. In the S53 step, if the LDPC decoding is performed, the value of q in the encoded information is used as the value of q in the finite field GF(2^q) in the LDPC decoding, to The LDPC decoding is performed.

S54: The storage controller outputs the data obtained after performing the decoding.

The specific encoding and decoding of the LDPC in the above embodiments will be described below.

1) LDPC code:

Generating matrices corresponding to different code rates are stored in the memory, and the memory controller acquires corresponding generating matrices G according to the determined LDPC code rates, and the coded data is obtained by the formula v=u·G. Among them, v is the encoded codeword, including information code and check code; u is the original data to be written.

In addition, the encoding of BCH is similar to the above, that is, the encoded data is also obtained from v=u·G. But the implementation of BCH and LDPC is different. In this application, the value of q in the finite field GF(2^q) in BCH coding can also be configured. Generally speaking, in order to improve the ability to resist burst errors, the value of q is related to the type of NAND. If it is an SLC type, set q to a value of 1, if it is an MLC type, set q to a value of 2, and if it is a TLC type, set q to a value of 3, that is, use non-binary encoding. Of course, for any type of NAND, q can be 1, that is, binary LDPC encoding is used.

2) LDPC decoding:

a. Initialization: For each variable node (v=1...N), the log-likelihood ratio (English: Log-Likelihood Ratio, abbreviated: LLR) message is set to:

b. Variable node update: each variable node has dv input messages {U _vp }p=1...dv∈R ^q , where R ^q represents a set formed with q as the unit;

c. Check node update: For each variable node, it can be considered that there are dc input messages {U _cp }p=1...dc∈R ^q , where R ^q represents a set composed of units of q;

d. Judging whether the decoded codeword is correct according to the check matrix H, if it is correct, stop the iteration; otherwise, continue to iterate and decode until the number of iterations reaches the set maximum number of times, if it is still not correctly decoded at this time code, the decoding fails. The parity check matrix H is dual to the generator matrix G when the data is encoded, that is, G·H=0.

In the above LDPC decoding process, the probability of a certain point is obtained through LLR to determine the value represented by its voltage, that is, as shown in Figure 6, point t is closer to the voltage range of 11, that is, the probability of point t being 11 obtained through LLR is the largest , so the value represented by the voltage at point t is 11, where P(v) in FIG. 6 represents probability, and Vth represents voltage. For example, the initial LLR for b ₀ =1 can be calculated by the following formula:

Please refer to FIG. 7a, which is a schematic structural diagram of an embodiment of an encoding device for error checking and correction in the present application. The encoding device 700 is used to encode data and write it into a memory. The encoding device 700 includes an analyzing module 710 , an encoding module 720 and a writing module 730 .

The analysis module 710 is configured to analyze the importance of the data when receiving the data write instruction.

The coding module 720 is configured to respectively use BCH and low density parity check code LDPC to code the data when the importance of the data belongs to the first important level, or use LDPC not higher than the first code rate to code the data encoding the data; when the importance of the data belongs to the second importance level, the data is encoded by using the BCH or LDPC higher than the first code rate, wherein the first importance level The importance degree is higher than the importance degree of the second importance level.

The writing module 730 is configured to write the encoded data into the memory, and save the encoding information of the data, wherein the encoding information of the data includes the encoding method adopted by the data.

Optionally, the analysis module 710 performing the step of analyzing the importance of the data includes: analyzing at least one of the source, attribute, and type of the data to obtain an importance value of the data; when the importance When the degree value exceeds the first preset value, it is determined that the importance of the data belongs to the first importance level; and when the importance degree value does not exceed the first preset value, it is determined that the importance of the data belongs to the first level Second level.

Optionally, the second importance level includes a first importance sub-level and a second importance sub-level. The analysis module 710 performing the step of determining that the importance of the data belongs to the second importance level when the importance level value does not exceed the first preset value includes: when the importance level value does not exceed the preset value When the first preset value exceeds the second preset value, it is determined that the importance of the data belongs to the first importance sub-level; when the importance does not exceed the second preset value, it is determined that the The importance of the data belongs to the second importance sub-level, wherein the first preset value is greater than the second preset value. When the encoding module 720 performs the step of encoding the data using the BCH or LDPC higher than the first code rate when the importance of the data belongs to the second importance level, the step includes: When the importance level belongs to the first important sub-level, the data is encoded by using BCH higher than the preset parity number or LDPC higher than the first code rate and lower than the second code rate, and in the When the importance of the data belongs to the second important sub-level, the data is encoded by using BCH not higher than the preset check digit or LDPC not lower than the second code rate, wherein, The first code rate is higher than the second code rate.

Optionally, the memory is an SLC, MLC or TLC type NAND memory. The encoding device may also include a value module, configured to determine the type of the memory, and when the type of the memory is SLC, obtain the value of the q in the finite field GF(2^q) in the LDPC decoding is 1; when the type of the memory is MLC, the value of q is 2; when the type of the memory is TLC, the value of q is 3. Moreover, the encoding information of the data saved by the writing module 730 also includes the value of q.

Optionally, the generation matrix of the BCH is selected according to the parity number of the BCH, and/or the generation matrix of the LDPC is selected according to the code rate of the LDPC.

In an application embodiment, the encoding module 720 may include a BCH encoder 721 , an LDPC encoder 722 , a selector 723 and a controller 724 as shown in FIG. 7 b . The controller 724 is used to determine the encoding method according to the importance of the data obtained by the analysis module 710 , and send the signal whether to bypass the BCH to the selector 723 , and send the code rate signal of LDPC to the LDPC encoder 722 . The selector 723 is used to determine whether the data needs to be BCH encoded according to the bypass signal. If necessary, the output of the BCH encoder 721 is selected to be connected to the input of the LDPC encoder 722. If not, the original data is not encoded by the BCH The encoder 721 is directly input to the LDPC encoder 722 . The BCH encoder 721 is used to BCH encode the input. The LDPC encoder 722 is configured to determine a code rate according to the coderate signal, and perform LDPC encoding on the input according to the determined code rate.

Please refer to FIG. 8 . FIG. 8 is a schematic structural diagram of an embodiment of a decoding device for error checking and correction in the present application. The decoding device 800 is used for reading data from a memory and decoding it. The decoding device 800 includes a reading module 810 , a decoding module 820 and an output module 830 .

The reading module 810 is used to read the data and the coding information of the data from the memory when the storage controller receives the data reading instruction, wherein the coding information includes the coding method adopted by the data, The encoding method adopted by the data depends on the importance of the data. When the importance of the data is at the first importance level, the combination of BCH and low-density parity-check code LDPC is adopted or the code rate is not higher than the first code rate. The LDPC encoding method; when the importance of the data is the second importance level, the BCH or the LDPC encoding method higher than the first code rate is adopted, and the importance of the first importance level is greater than the State the importance of the second level of importance.

The decoding module 820 is configured to determine the encoding method of the data according to the encoding information, and decode the data by using a decoding method corresponding to the encoding method, wherein the decoding method includes a combination of BCH and LDPC, or a combination of BCH and LDPC. LDPC and BCH higher than the first code rate or LDPC higher than the first code rate.

The output module 830 is configured to output the data obtained after the decoding.

Optionally, the second importance level includes a first importance sub-level and a second importance sub-level, and the importance degree of the first importance sub-level is greater than that of the second importance sub-level. The encoding method adopted by the data also includes BCH which is higher than the preset check digit or higher than the first code rate and lower than the first code rate when the importance of the data is the first important sub-level. The encoding method of LDPC with two code rates, when the importance of the data is the second important sub-level, the BCH that is not higher than the preset check digit or not lower than the second code rate is adopted The LDPC encoding method, wherein the first code rate is higher than the second code rate. The BCH or the LDPC decoding method higher than the first code rate includes BCH higher than the preset check digit, LDPC higher than the first code rate and lower than the second code rate, not higher than the The decoding method of the BCH with the preset parity number or the LDPC with the second code rate not lower than the second code rate.

Optionally, the decoding device 800 may also include an alarm module, which is used to send out a Alerts that there is a problem with the block in which the data is stored.

Optionally, the memory is an SLC, MLC or TLC type NAND memory. The encoding information of the data also includes the value of q, when the type of the memory is SLC, the value of q is 1, when the type of the memory is MLC, the value of q is 2, when When the type of the memory is TLC, the value of q is 3. In the step of decoding the data using the decoding method corresponding to the encoding method, if the decoding module 820 performs the LDPC decoding, the decoding module 820 is specifically used to convert the The value of q is used as the value of q in the finite field GF(2^q) in the LDPC decoding to perform the LDPC decoding.

Optionally, the check matrix of the BCH is selected according to the number of check digits of the BCH, and/or the check matrix of the LDPC is selected according to the code rate of the LDPC.

Each module of the above-mentioned encoding device is specifically used to execute the corresponding steps in the above-mentioned encoding method, and each module of the above-mentioned decoding device is specifically used to execute the corresponding steps in the above-mentioned decoding method. For details, please refer to the relevant description above, and details are not repeated here.

The present application also provides an implementation manner of a codec processor for error checking and correction, the codec processor may be set in a memory, and the codec processor may include the encoding device in the above embodiments and/or The decoding device in the above embodiments. For example, in one embodiment, the memory includes encoding means and decoding means. The encoding device is the encoding device in the above embodiments. The decoding means is the decoding means in the above embodiments.

In the above solution, the storage controller analyzes the importance of data and configures different encoding methods for the data according to different degrees of importance, that is, the more important the data, the stronger the encoding method for error checking and correcting capabilities is, and the data encoding by the memory is realized. The adaptive configuration of the mode ensures the reliable and accurate storage of important data. At the same time, the storage controller selects a coding method with low error checking and correction capabilities for data with low importance, which reduces the required coding capability and check code length, improves coding efficiency and saves storage space.

In the several implementation manners provided in this application, it should be understood that the disclosed systems, devices and methods may be implemented in other ways. For example, the device implementations described above are only illustrative. For example, the division of the modules or units is only a logical function division. In actual implementation, there may be other division methods. For example, multiple units or components can be Incorporation may either be integrated into another system, or some features may be omitted, or not implemented. In another point, the mutual coupling or direct coupling or communication connection shown or discussed may be through some interfaces, and the indirect coupling or communication connection of devices or units may be in electrical, mechanical or other forms.

The units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or may be distributed to multiple network units. Part or all of the units can be selected according to actual needs to achieve the purpose of the solution of this embodiment.

In addition, each functional unit in each embodiment of the present application may be integrated into one processing unit, each unit may exist separately physically, or two or more units may be integrated into one unit. The above-mentioned integrated units can be implemented in the form of hardware or in the form of software functional units.

If the integrated unit is realized in the form of a software function unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application is essentially or part of the contribution to the prior art or all or part of the technical solution can be embodied in the form of a software product, and the computer software product is stored in a storage medium , including several instructions to make a computer device (which may be a personal computer, a server, or a network device, etc.) or a processor (processor) execute all or part of the steps of the methods described in various embodiments of the present application. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), magnetic disk or optical disk and other media that can store program codes. .

Claims (13)

1. it is a kind of for error checking and correct coding method, it is characterised in that including：

Storage control analyzes the important journey of the data when the write instruction of data is received Degree；

When the significance level of the data belongs to the first importance level, BCH and LDPC is respectively adopted The data are encoded, or the LDPC of use not higher than the first code check is carried out to the data Coding；When the significance level of the data belongs to the second importance level, using the BCH or LDPC higher than first code check is encoded to the data, wherein, described first is important Significance level of the significance level of level higher than second importance level；And

The data after by coding are written in memory, and preserve the coding letter of the data Breath, wherein, the coding information of the data includes the coded system that the data are used.

2. it is according to claim 1 for error checking and correct coding method, its feature It is, the step of the significance level of the analysis data, including：

At least one in source, attribute and the type of the data is analyzed, the data are obtained Importance value；

When the importance value is more than the first preset value, the significance level category of the data is determined In the first importance level；And when the importance value is no more than first preset value, it is determined that The significance level of the data belongs to the second importance level.

3. it is according to claim 2 for error checking and correct coding method, its feature It is that second importance level includes the first important sub- level and the second important sub- level；

It is described when the importance value be no more than first preset value when, determine the data The step of significance level belongs to the second importance level includes：

When the importance value is no more than first preset value and during more than the second preset value, really The significance level of the fixed data belongs to the described first important sub- level；When the importance value does not surpass When crossing second preset value, determine that the significance level of the data belongs to the described second important son Level, wherein, first preset value is more than second preset value；And

When the significance level in the data belongs to the second importance level, using the BCH or The step of person encodes higher than the LDPC of first code check to the data, including：

When the significance level of the data belongs to the first important sub- level, using higher than default school Test digit BCH or higher than first code check and less than the second code check LDPC to the number According to being encoded, when the significance level of the data belongs to the second important sub- level, using not Higher than it is described it is default verification digit BCH or be not less than the LDPC of second code check to described Data are encoded, wherein, first code check is higher than second code check.

4. according to any one of claims 1 to 3 for error checking and correct coding staff Method, it is characterised in that the memory is stored for the NAND of SLC, MLC or TLC type Device；

It is described by coding after the data be written to memory, and preserve the volume of the data Before the step of code information, methods described also includes：

Determine the type of the memory, and when the type of the memory is SLC, by LDPC The q values in finite field gf (2^q) in decoding are 1；When the type of the memory It is 2 by the q values during for MLC；When the type of the memory is TLC, by institute It is 3 to state q values；And

The coding information of the data also value including the q.

5. according to any one of claims 1 to 3 for error checking and correct coding staff Method, it is characterised in that the generator matrix of the BCH is carried out according to the verification digit of the BCH Selection, and/or, the generator matrix of the LDPC is selected according to the code check of the LDPC.

6. it is a kind of for error checking and correct coding/decoding method, it is characterised in that including：

Storage control reads the data when the reading for receiving data is instructed from memory And the coding information of the data, wherein, the coding information includes the coding that the data are used Mode, the coded system that the data are used depends on the significance level of the data, in the number According to significance level be the first importance level when, using the combination of BCH and LDPC or not higher than the The coded system of the LDPC of one code check；When the significance level of the data is the second importance level, The coded system of the LDPC using the BCH or higher than first code check, described first Significance level of the significance level of importance level more than second importance level；

The coded system of the data is determined according to the coding information, and is used and the coding staff The corresponding decoding process of formula is decoded to the data, wherein, the decoding process includes BCH With the combination of LDPC, it is not higher than LDPC, BCH of the first code check or higher than described first yard The LDPC of rate；And

Output carries out the data obtained after the decoding.

7. it is according to claim 6 for error checking and correct coding/decoding method, its feature It is that second importance level includes the first important sub- level and the second important sub- level, first weight The significance level of sub- level is wanted more than the significance level of the described second important sub- level；

The significance level that the coded system that the data are used is additionally included in the data is described Used during one important sub- level higher than it is default verification digit BCH or higher than first code check And the coded system of the LDPC less than the second code check, it is described the in the significance level of the data The BCH of the not higher than described default verification digit used during two important sub- levels is not less than described The coded system of the LDPC of the second code check, wherein, first code check is higher than second code check； And

The BCH includes higher than default higher than the LDPC decoding processes of first code check Verify the BCH of digit, the LDPC higher than first code check and less than the second code check, not higher than The decoding side of the BCH or the LDPC for being not less than second code check of the default verification digit Formula.

8. it is according to claim 6 for error checking and correct coding/decoding method, its feature It is in the coded system that the data are determined according to the coding information, and to use and institute After stating the step of the corresponding decoding process of coded system is decoded to the data, methods described Also include：

The error bit number for detecting the BCH exceedes predetermined bit value and/or the LDPC When actual iterations exceedes preset times, send and the block of the data is stored in the memory go out The alarm of existing problem.

9. according to claim 6 or 7 for error checking and correct coding/decoding method, its It is characterised by, the memory is the nand memory of SLC, MLC or TLC type；Institute State the coding information also value including q of data, when the type of the memory is SLC, institute It is 1 to state q values, and when the type of the memory is MLC, the q values are 2, work as institute When the type for stating memory is TLC, the q values are 3；And

The data are decoded using decoding process corresponding with the coded system described The step of in, if carrying out the LDPC decodings, by taking for the q in the coding information It is worth the value of the q in the finite field gf (2^q) in being decoded as the LDPC, to carry out institute State LDPC decodings.

10. according to claim 6 or 7 for error checking and correct coding/decoding method, its It is characterised by, the check matrix of the BCH is selected according to the verification digit of the BCH, And/or, the check matrix of the LDPC is selected according to the code check of the LDPC.

11. it is a kind of for error checking and correct code device, it is characterised in that including：

Analysis module, for when the write instruction of data is received, analyzing the important of the data Degree；

Coding module, for when the significance level of the data belongs to the first importance level, adopting respectively The data are encoded with BCH and LDPC, or using the LDPC of not higher than the first code check The data are encoded；When the significance level of the data belongs to the second importance level, use The BCH is encoded higher than the LDPC of first code check to the data, wherein, Significance level of the significance level of first importance level higher than second importance level；And

Writing module, is written in memory for the data after by coding, and preserves described The coding information of data, wherein, the coding information of the data includes the coding that the data are used Mode.

12. it is a kind of for error checking and correct decoding apparatus, it is characterised in that including：

Read module, for storage control when the reading for receiving data is instructed, from memory The middle coding information for reading the data and the data, wherein, the coding information includes described The coded system that data are used, the coded system that the data are used depends on the important of the data Degree, when the significance level of the data is the first importance level, using BCH's and LDPC With reference to or the not higher than coded system of the LDPC of the first code check；It is in the significance level of the data During the second importance level, the coding of the LDPC using the BCH or higher than first code check Mode, the significance level of the significance level more than second importance level of first importance level；

Decoder module, for determining the coded system of the data according to the coding information, and adopts The data are decoded with decoding process corresponding with the coded system, wherein, the solution The combination of code mode including BCH and LDPC, not higher than LDPC, BCH of the first code check or LDPC of the person higher than first code check；And

Output module, for exporting the data for obtained after the decoding.

13. it is a kind of for error checking and correct codec processor, it is characterised in that including power Profit requires the decoding apparatus described in code device and/or claim 12 described in 11.