Syndrome based decoding of a misaligned codeword

HK40137536APending Publication Date: 2026-09-11RETYM INC
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
HK62026127146
Authority / Receiving Office
HK · HK
Patent Type
Applications
Current Assignee / Owner
Priority Date
2023-08-29
Filing Date
2026-08-06
Publication Date
2026-09-11

Smart Images

  • Figure 00000000_0000_ABST
    Figure 00000000_0000_ABST
Patent Text Reader

Abstract

A method for decoding includes, in a decoder 20 of an error correction code (ECC), receiving, in a buffer 24, a codeword 50 of the ECC containing a plurality of symbols, the codeword being misaligned with one or both boundaries of the buffer. In the buffer, one or more bits that do not belong to the codeword are zeroed. One or more syndromes are computed for the codeword across both the codeword and the zeroed bits. Based on the calculated syndromes, locations of one or more erroneous symbols are identified in the codeword, and the erroneous symbols are corrected in the identified locations to restore an erroneous version of the codeword. The method further includes compensating for misalignment between the codeword and the buffer during calculation of the syndrome or during identification of the location of the erroneous symbol.
Need to check novelty before this filing date? Find Prior Art

Description

(19) State Intellectual Property Office (12) Invention Patent Application (10) Application Publication Number (43) Application Publication Date (21) Application Number 202480052656.X (22) Application Date 2024.07.14 (30) Priority Data 63 / 579,301 2023.08.29 US (85) PCT International Application Entering National Phase Date 2026.02.10 (86) PCT International Application Application Data PCT / US2024 / 037958 2024.07.14 (87) PCT International Application Publication Data WO2025 / 048967 EN 2025.03.06 (71) Applicant: Retham Corporation Address: California, USA (72) Inventor: Ben Shilomcha Anholt (74) Patent Agency: Beijing Anxin Fangda Intellectual Property Agency Co., Ltd. 11262 Patent Attorney Lu Jianping Yang Mingzhao (51) Int.Cl. H03M 13 / 03 (2006.01) H03M 13 / 47 (2006.01) G11B 20 / 18 (2006.01) G11C 29 / 42 (2006.01) G06F 11 / 08 (2006.01) (54) Invention Title Syndrome-Based Decoding of Unaligned Codewords (57) Abstract A method for decoding includes: in an error-correcting code (ECC) decoder 20, receiving in a buffer 24 a codeword 50 of ECC containing multiple symbols, the codeword being unaligned with one or both boundaries of the buffer. In the buffer, one or more bits that do not belong to the codeword are set to zero. A syndrome or more is calculated for the codeword across both the codeword and the zeroed bits. Based on the calculated syndrome, the location of one or more erroneous symbols in the codeword is identified, and the erroneous symbols are corrected at the identified locations to recover an error-free version of the codeword. The method also includes compensating for misalignment between the codeword and the buffer during the calculation of the syndrome or during the identification of erroneous symbol locations.Claims 2 pages, Description 10 pages, Drawings 5 ​​pages, CN 121713387 A 2026.03.20 CN 1 21 71 33 87 A 1. A method for decoding, comprising: receiving, in a decoder of an error-correcting code (ECC), a codeword of an ECC containing a plurality of symbols in a buffer, the codeword being misaligned with one or both boundaries of the buffer; setting one or more bits in the buffer that do not belong to the codeword to zero; calculating one or more syndications for the codeword across both the codeword and the zeroed bits; identifying the positions of one or more erroneous symbols in the codeword based on the calculated syndications; and correcting the erroneous symbols at the identified positions to recover an error-free version of the codeword, the method further comprising compensating for the misalignment between the codeword and the buffer during the calculation of the syndications or during the identification of the positions of the erroneous symbols. 2. The method of claim 1, wherein zeroing the bit comprises zeroing a first number of symbols matching the trailing bit after the codeword in the buffer, thereby cyclically shifting the synod by the first number, and wherein identifying the position comprises (i) compensating for the misalignment by cyclically shifting the synod in the reverse direction by the first number, and then (ii) identifying the position based on the cyclically shifted synod. 3. The method of claim 1, wherein zeroing the bit comprises zeroing a first number of symbols matching the trailing bit after the codeword in the buffer, thereby cyclically shifting the synod by the first number, and wherein identifying the position comprises calculating the position based on the cyclically shifted synod, and then compensating for the misalignment by cyclically shifting the position in the reverse direction by the first number. 4. The method of any one of claims 1 to 3, wherein zeroing the bit comprises zeroing a second number of leading bits preceding the codeword in the buffer. 5. The method of any one of claims 1 to 3, wherein receiving the codeword comprises receiving the codeword in the form of one or more blocks having a common length, and wherein calculating the syndrome comprises iteratively calculating the syndrome over the blocks. 6. The method of claim 5, wherein iteratively calculating the syndrome comprises iteratively calculating a given syndrome over multiple blocks using a common combinational circuit. 7. The method of any one of claims 1 to 3, wherein the ECC comprises a cyclic code selected from a list including Bosch-Joheri-Hockvanheim (BCH) codes and Reed-Solomon codes.8. The method according to any one of claims 1-3, comprising: receiving (i) a low reliability bit (LRB) indicating a bit having a low reliability level in the codeword, and (ii) a shift value corresponding to the misalignment; calculating a test pattern from the codeword using the LRB; calculating a shifted LRB by cyclically shifting the LRB according to the shift value; deriving a corresponding test pattern syntactic from the syntactic by modifying the syntactic based on the shifted LRB; and hard decoding the test pattern using the test pattern syntactic to recover an error-free version of the codeword. 9. A decoder comprising: a buffer configured to receive a codeword containing an error-correcting code (ECC) of multiple symbols, the codeword being misaligned with one or both boundaries of the buffer; and circuitry configured to: zero one or more bits in the buffer that do not belong to the codeword; calculate one or more synods for the codeword across both the codeword and the zeroed bits; identify the positions of one or more erroneous symbols in the codeword based on the calculated synods; and correct the erroneous symbols at the identified positions to recover an error-free version of the codeword, the circuitry further configured to compensate for the misalignment between the codeword and the buffer during the calculation of the synods or during the identification of the positions of the erroneous symbols. 10. The decoder of claim 9, wherein the circuitry is configured to zero the bit by zeroing a first number of symbols matching the trailing bit after the codeword in the buffer, thereby cyclically shifting the synod by the first number, and to identify the position by: (i) compensating for misalignment by cyclically shifting the synod in reverse by the first number, and then (ii) identifying the position based on the cyclically shifted synod. 11. The decoder of claim 9, wherein the circuitry is configured to zero the bit by zeroing a first number of symbols matching the trailing bit after the codeword in the buffer, thereby cyclically shifting the synod by the first number; and to identify the position by: calculating the position based on the cyclically shifted synod, and then compensating for misalignment by cyclically shifting the position in reverse by the first number. 12. The decoder of any one of claims 9 to 11, wherein the circuitry is configured to zero the bit by zeroing a second number of leading bits preceding the codeword in the buffer.13. The decoder of any one of claims 9 to 11, wherein the circuitry is configured to receive the codeword in the form of one or more blocks having a common length, and to iteratively compute the syndrome on the blocks. 14. The decoder of claim 13, wherein the circuitry is configured to iteratively compute the syndrome by using a common combinational circuit on multiple blocks to iteratively compute a given syndrome. 15. The decoder of any one of claims 9 to 11, wherein the ECC comprises cyclic codes selected from a list including Bosch-Joherry-Hokvenheim (BCH) codes and Reed-Solomon codes. 16. The decoder according to any one of claims 9 to 11, wherein the circuitry is configured to: receive (i) a low reliability bit (LRB) indicating a bit having a low reliability level in the codeword, and (ii) a shift value corresponding to the misalignment; calculate a test pattern from the codeword using the LRB; calculate a shifted LRB by cyclically shifting the LRB according to the shift value; derive a corresponding test pattern syndic from the syndic by modifying the syndic based on the shifted LRB; and hard decode the test pattern using the test pattern syndic to recover an error-free version of the codeword. Claims 2 / 2 Page 3 CN 121713387 A Syndication-Based Decoding of Misaligned Codewords

[0001] Cross-Reference to Related Applications This application claims the benefit of U.S. Provisional Patent Application 63 / 579,301, filed August 29, 2023, the disclosure of which is incorporated herein by reference. Technical Field

[0002] The embodiments described herein generally relate to error-correcting codes, and particularly to methods and systems for syndrome-based decoding of unaligned codewords.

[0003] Background Various error-correcting codes, such as cyclic codes, such as Bose–Chaudhuri–Hocquenghem (BCH) codes and Reed-Solomon (RS) codes, can be decoded using syndrome-based decoding methods. BCH and RS codes have many applications, including high-speed optical networks. For example, an extended BCH (256,239) code is specified in Annex E of the ITU-T G.709.3 / Y.1331.3 standard published on June 22, 2018, which is used as a component code in a product-like code called Open Forward Error Correction (OFEC); this standard defines Flexible Optical Transport Network (OTN), also known as FlexO.For example, the Optical Interconnect Forum (OIF) also specifies an encoding scheme consistent with OFEC, which also uses BCH(256,239) as the component code.

[0004] Overview The embodiments described herein provide a method for decoding, which includes, in a decoder of an error-correcting code (ECC), receiving in a buffer a codeword of ECC containing multiple symbols, the codeword being misaligned with one or both boundaries of the buffer. In the buffer, one or more bits that do not belong to the codeword are set to zero. A synod is calculated for the codeword across both the codeword and the zeroed bits. Based on the calculated synod, the positions of one or more erroneous symbols in the codeword are identified, and the erroneous symbols are corrected at the identified positions to recover an error-free version of the codeword. The method also includes compensating for the misalignment between the codeword and the buffer during the calculation of the synod or during the identification of the positions of the erroneous symbols.

[0005] In some embodiments, zeroing bits includes zeroing a first number of symbols matching trailing bits after the codeword in the buffer, thereby cyclically shifting the synod by a first number, and identifying the position includes (i) compensating for misalignment by cyclically shifting the synod in reverse by a first number, and then (ii) identifying the position based on the cyclically shifted synod. In other embodiments, zeroing bits includes zeroing a first number of symbols matching trailing bits after the codeword in the buffer, thereby cyclically shifting the synod by a first number, and identifying the position by calculating the position based on the cyclically shifted synod, and then compensating for misalignment by cyclically shifting the position in reverse by a first number. In still other embodiments, zeroing bits includes zeroing a second number of leading bits preceding the codeword in the buffer.

[0006] In embodiments, receiving the codeword includes receiving the codeword in the form of one or more blocks having a common length, and calculating the synod includes iteratively calculating the synod over the blocks. In another embodiment, iteratively calculating the syndrome includes iteratively calculating a given syndrome using common combinational circuitry on multiple blocks. In yet another embodiment, the ECC includes a cyclic code selected from a list including Bosch-Joherry-Hockvanheim (BCH) codes and Reed-Solomon codes.

[0007] In some embodiments, the method includes: receiving (i) a low reliability bit (LRB) indicating a bit with a low reliability level in a codeword, and (ii) a shift value corresponding to the misalignment; calculating a test pattern from the codeword using the LRB; calculating a shifted LRB by cyclically shifting the LRB according to the shift value; modifying the synod based on the shifted LRB; deriving a corresponding test pattern synod from the synod; and hard decoding the test pattern using the test pattern synod to recover an error-free version of the codeword.

[0008] According to the embodiments described herein, a decoder including a buffer and circuitry is also provided. The buffer is configured to receive a codeword containing a plurality of symbols of error-correcting code (ECC) that is misaligned with one or both boundaries of the buffer. The circuit is configured to: zero out one or more bits in the buffer that do not belong to the codeword; calculate one or more syndromes for the codeword based on the calculated syndrome, spanning both the codeword and the zeroed bits; identify the positions of one or more erroneous symbols in the codeword; and correct the erroneous symbols at the identified positions to recover an error-free version of the codeword. The circuit is also configured to compensate for misalignment between the codeword and the buffer during the calculation of the syndrome or during the identification of the erroneous symbol positions.

[0009] These and other embodiments will be more fully understood from the following detailed description of their embodiments, taken in conjunction with the accompanying drawings, in which: Brief Description of the Drawings FIG1 is a block diagram schematically illustrating a synod-based error-correcting code (ECC) decoder for compensating for misaligned buffer codewords according to an embodiment described herein; FIG2A and FIG2B are flowcharts schematically illustrating methods for compensating for codeword misalignment during the computation of synods and the identification of the location of erroneous symbols according to an embodiment described herein; FIG3 is a block diagram schematically illustrating an ECC decoder according to an embodiment described herein that iteratively computes synods over multiple blocks of codewords and compensates for codeword misalignment; FIG4 is a block diagram schematically illustrating an ECC decoder 150 for decoding extended BCH codes according to an embodiment described herein; and FIG5 is a flowchart schematically illustrating a method for soft decoding of codewords misaligned with buffer boundaries according to an embodiment described herein. Detailed Description

[0010] Summary The embodiments described herein provide an improved method and circuitry for decoding BCH codes and other cyclic error-correcting codes, which achieves this by efficiently calculating the syndrome of codewords that are misaligned with buffer boundaries. For this purpose, buffer bits that do not belong to the codeword are masked to zero, and codeword misalignment is compensated for during the calculation of the syndrome or during the calculation of the position of the error symbol.

[0011] The following description primarily relates to cyclic codes defined over a finite field. In this context, the term "symbol" refers to an element of the finite field over which the underlying code is defined. Depending on the code, a symbol may have a single bit or multiple bits.

[0012] Error-correcting codes (ECCs) define a mapping between a data message and a corresponding codeword of the ECC. Each codeword comprises a sequence of symbols (e.g., bits) and carries the data message along with redundant information.

[0013] ECC codewords received over a noisy channel may contain erroneous symbols. The goal of the decoder is to recover the original message by identifying the ECC codewords based on the received codewords. For example, the identified codeword may be the closest codeword in the ECC to the received codeword, according to a suitable distance metric (e.g., Hamming distance or Euclidean distance). Specification 2 / 10 Page 5 CN 121713387 A

[0014] In this context and claims, the term "codeword" means (i) an ECC codeword or (ii) a codeword containing one or more error symbols.

[0015] In some embodiments, decoding a received codeword involves deriving one or more syndromes from the codeword. The received codeword is an ECC codeword when all syndromes are equal to zero. In other cases, based on the syndromes, the decoder calculates the position of the error symbol and the corresponding error value in the received codeword and corrects the received codeword accordingly.

[0016] Syndromes can be calculated in various ways. In one approach, the syndrome vector is calculated by multiplying the parity-check matrix of the underlying ECC by the received codeword. However, this approach is generally inefficient because it requires storing the complete parity-check matrix or dynamically calculating it. In another approach, the adjoint is computed by treating the sign of the received codeword as the coefficients of a polynomial and evaluating the polynomial at selected powers of primitive elements in an appropriate finite field.

[0017] The received codeword is typically temporarily stored in a buffer of the decoder. The buffer may contain the complete codeword or a portion of the codeword (e.g., when codewords are received in multiple blocks). In some implementations, the codeword (or a portion thereof) is misaligned with one or both boundaries of the buffer. This may occur, for example, when the codeword length is not a multiple of the buffer length, when the codeword length is less than the buffer length, when the input includes fields that are not part of the codeword, or due to the way the decoder synchronizes the reception of the codeword.

[0018] In principle, additional buffering and / or additional computation can be used to handle codeword misalignment, which typically results in a large silicon area and high power consumption.For example, when calculating the syndrome over multiple blocks of a codeword, different hardware circuits can be customized for the first block, the middle block, and the last block, as well as for different degrees of codeword misalignment.

[0019] In the disclosed embodiments, misalignment between the codeword and the buffer boundaries is compensated when calculating the syndrome or when identifying the location of an erroneous symbol. In a multi-block hardware implementation, due to the compensation for misalignment, common combinational logic circuit units can be used on all blocks, which reduces silicon area and power consumption compared to using different circuit units for different blocks.

[0020] Consider a decoding method comprising: receiving, in a decoder of an error-correcting code (ECC), an ECC codeword containing multiple symbols in a buffer, the codeword being misaligned with one or both boundaries of the buffer. One or more bits in the buffer that do not belong to the codeword are set to zero. For the codeword, multiple syndromes are calculated across both the codeword and the zeroed bits. Based on the calculated synod, the positions of one or more erroneous symbols in the codeword are identified, and the erroneous symbols at the identified positions are corrected to restore an error-free version of the codeword. The method also includes compensating for misalignment between the codeword and the buffer during the calculation of the synod or during the identification of the erroneous symbols.

[0021] In some embodiments, zeroing bits includes: zeroing a first number of symbols matching trailing bits after the codeword in the buffer, thereby cyclically shifting the synod by a first number. In one embodiment, the position is identified by: (i) compensating for misalignment by cyclically shifting the synod in the reverse direction by a first number, and then (ii) identifying the position based on the cyclically shifted synod. In another embodiment, the position is identified by: calculating the position based on the cyclically shifted synod, and then compensating for misalignment by cyclically shifting the position in the reverse direction by a first number. In yet another embodiment, zeroing bits includes: zeroing a second number of leading bits preceding the codeword in the buffer. In this embodiment, the synod is not cyclically shifted.

[0022] In an embodiment, codewords are received in the form of one or more blocks having a common length, and the syndrome is obtained by iterative computation on the blocks. Furthermore, a given syndrome is computed through multiple iterations using a common combinational circuit on multiple blocks. By sharing the same circuitry across all blocks, low hardware complexity and low power consumption are achieved.

[0023] The disclosed method is applicable to the decoding of cyclic codes, such as, for example, the decoding of Bosch-Joheri-Hokvenheim codes (BCH) and Reed-Solomon codes. The disclosed embodiments are also applicable to the decoding of codes associated with cyclic codes, such as, for example, extended cyclic codes.

[0024] In an embodiment, the decoder uses a soft decoding method to decode the received codewords.In this embodiment, the method includes: receiving, in addition to the codeword, (i) a low reliability bit (LRB) indicating a bit with a low reliability level in the codeword, and (ii) a shift value corresponding to the misalignment; calculating a shifted LRB by cyclically shifting the LRB according to the shift value; deriving a corresponding test mode syndrome from the syndrome by modifying the syndrome based on the shifted LRB; and decoding the test mode using the test mode syndrome to recover an error-free version of the codeword.

[0025] In the disclosed art, the received and stored codeword is misaligned with the decoder buffer. Bits in the buffer that do not belong to the codeword are set to zero, causing the syndrome to be cyclically shifted by a shift value equal to the number of trailing bits set to zero (or the number of symbols matching the trailing bits). The syndrome or error position is cyclically shifted in reverse to compensate for the codeword misalignment. In embodiments where codewords are received in multiple blocks, common combinational logic circuitry is used to cumulatively calculate the syndrome across all blocks, thereby achieving a small silicon area and low power consumption.

[0026] System Description Figure 1 is a block diagram schematically illustrating a syntactic ECC decoder 20 according to an embodiment described herein, which compensates for codeword misalignment in the buffer.

[0027] For example, decoder 20 may include a decoder for decoding cyclic codes such as BCH codes or RS codes. BCH codes can be used in a variety of applications, such as satellite and other communication systems, optical disc players, DVDs, disk drives, USB flash drives, solid-state drives, and two-dimensional barcodes. RS codes can be used in many applications, including consumer technologies such as mini-discs, CDs, DVDs, Blu-ray discs, QR codes, data matrices, data transmission technologies such as DSL and WiMAX, broadcasting systems such as satellite communications, DVB, and ATSC, and storage systems such as RAID 6.

[0028] Consider defining ECC on a Galois field, where "q" is a prime number and a prime power, where "q" is a positive integer. For primitive BCH codes, the codeword length "q" is given by: q = q ... In some embodiments, the decoder may compute the syndrome by treating the received codeword as a polynomial having coefficients in and evaluating on a specific power of the primitive . In the case of binary codes such as BCH codes, the coefficients of are bits with binary values ​​"0" and "1". The number of syndromes required to decode the primitive BCH code is given by d-1, where "d" represents the minimum distance of the basic code. The j-th syndrome is denoted as and given by: Equation 1: .

[0029] In some embodiments, the sum in Equation 1 may be divided into multiple partial sums, where each partial sum is computed using combinational logic within a computation cycle (e.g., within a single clock cycle).

[0030] In some embodiments, codewords represented by the given information are received in multiple blocks, each block comprising coefficients of the given information, satisfying the given information. Furthermore, it is assumed that the blocks are received in descending order of power. Thus, the first block contains coefficients corresponding to the highest power (see page 4 / 10 of specification 7 CN 121713387 A), and the last block contains coefficients corresponding to the lowest power. In some embodiments, the j-th adjoint can be iteratively computed on the block, wherein in the first iteration, the adjoint is given by: Equation 2: , and in subsequent iterations, the adjoint is updated by: Equation 3: .

[0031] Alternatively, the adjoint can be initialized to zero, and the adjoint can be iteratively computed using Equation 3, starting from the first block (corresponding to the highest power). In Equations 2 and 3, the values ​​in the summation depend on the number of iterations.

[0032] The codewords (or portions thereof) received for decoding are typically stored in a buffer before being processed. The following description primarily addresses the case where the buffered codeword (or a portion thereof) is misaligned with the buffer boundary. This means that the buffer contains one or more leading bits before the codeword, one or more trailing bits after the codeword, or both. It should be noted that, in embodiments, for RS codes, for low-complexity implementation, the number of trailing bits can be chosen as a multiple of the number of bits per symbol.

[0033] In some embodiments, to handle codeword misalignment, the decoder masks (forces zero) the buffered bits that do not belong to the codeword. However, such masking may cause cyclic shifts in the syntactic, and thus also cause cyclic shifts in the position of the erroneous symbol by the same number of bits. As will be described in detail below, in the disclosed embodiments, the decoder compensates for codeword misalignment in the syntactic that causes cyclic shifts.

[0034] Decoder 20 includes buffer 24 and circuitry 28, which includes a masking module 30, a syntactic calculator 34, an error calculator 38, an error corrector 42, and a timing module 46.

[0035] Timing module 46 controls the input and output of data into and out of buffer 24. In the example of FIG1, after the input cycle is completed, buffer 24 contains codeword 50 of the underlying ECC, a preamble bit 52 (denoted as "LDb"), and a trailing bit 54 (denoted as TRb). The preamble bit and the trailing bit are not part of codeword 50. In alternative embodiments, the buffer may contain only one of codeword 50 and preamble bit 52 and trailing bit 54. Although in the example of FIG1, buffer 24 contains the complete codeword, in alternative embodiments, the length of the buffer may be shorter than the length of the codeword. In such embodiments, the input is divided into multiple blocks having the same length as buffer 24. This configuration will be described in detail below with reference to FIG3.

[0036] In some embodiments, the decoder processes the contents of the buffer for decoding codeword 50.To ignore the leading and trailing bits in the buffer, the masking module 30 applies a mask value 58 to the buffer contents. The masking operation preserves the codeword value and sets leading and trailing bits that do not belong to the codeword to zero. In this context, the phrase "zero bits in the buffer" means that the relevant bits in the buffer are set to zero, or that the relevant bits are set to zero after being read from the buffer (without modifying the buffer contents). In an embodiment, the mask value 58 is generated by the timing module 46. In some embodiments, the mask value corresponding to the leading and trailing bits is constant over time and may be hardwired.

[0037] The syndrome calculator 34 calculates the syndrome associated with the codeword 50 by processing the entire buffer contents. For example, the syndrome calculation may be based on Equation 1 above, or on Equations 2 and 3. The effects of codeword misalignment will be further described below.

[0038] Error calculator 38 receives the syndrome calculated by syndrome calculator and, based on the syndrome, determines the number of error symbols in the codeword, the position of the error symbols in the codeword, and the error value using any suitable method. For example, the error calculator can use the Berlekamp-Massey algorithm to generate a position polynomial corresponding to the error symbols, where the roots of the polynomial correspond to the error positions.

[0039] Error corrector 42 receives the error positions (e.g., in the form of indices in the codeword) and error values ​​from the error calculator and corrects the codeword 50 accordingly. Typically, the error corrector outputs an error-free version of codeword 50 (or an appropriate indication of decoding failure). For example, a mis-correction event may occur when the underlying code is capable of correcting a small number of errors and the noise level causing errors in the codeword is high.

[0040] The correct syndrome value is not modified because the codeword 50 is misaligned in the buffer due to one or more zeroed leading bits 52 in the buffer 24. However, the misalignment of codeword 50 caused by setting D>0 trailing bits 54 to zero (in RS code, an integer D>0 symbols match the trailing bits, and setting them to zero means setting the integer D>0 symbols that match the trailing bits to zero) has the effect of causing a D-bit cyclic shift of the correct syndrome value.

[0041] In some embodiments, as described below with reference to Figures 2A and 2B, the decoder compensates for the codeword misalignment caused by the trailing bits during the calculation of the syndrome or during the identification of the position of the erroneous symbols. For this purpose, the timing module 46 provides a shift value D (equal to the number of trailing bits 54 or the number of trailing symbols) to the syndrome calculator or error calculator.

[0042] In some embodiments, the decoder 20 supports decoding of codeword 50 using a soft decoding method.In these embodiments, in addition to codeword 50, the decoder also receives soft information 64 (e.g., the reliability level of the symbols assigned to codeword 50 respectively). In some embodiments, the decoder additionally receives or computes one or more low reliability bits (LRBs), which indicate codeword symbols with the lowest reliability level. Furthermore, circuit 28 includes soft decoding logic 66, which processes the generation and decoding of test pattern 68 derived from codeword 50. An embodiment of implementing ECC decoding using soft decoding will be described in detail below with reference to FIG5.

[0043] Compensation method for codeword misalignment FIG2A and FIG2B are flowcharts schematically illustrating methods for compensating for codeword misalignment during the computation of syntactic expressions and during the identification of the location of erroneous symbols, according to the embodiments described herein.

[0044] In the description of the methods of FIG2A and FIG2B, it is assumed that the base codeword is misaligned with the buffer boundary due to one or more trailing bits following the codeword in the buffer. Since the trailing bit (or an integer number of symbols matching the trailing bit) is set to zero, the syndrome is cyclically shifted by a shift value D equal to the number of trailing bits (or the number of symbols matching the trailing bit), and the error position derived from the cyclically shifted syndrome is also cyclically shifted by the same shift value D.

[0045] The method of FIG2A begins with syndrome calculation step 80: the syndrome is calculated by syndrome calculator 34 for the zeroed leading bit 52 (if present), codeword 50, and zeroed trailing bit 54. As described above, the resulting syndrome is cyclically shifted by a non-zero shift value D, D equal to the number of trailing bits (or the number of symbols matching the trailing bit).

[0046] In compensation step 82, the syndrome calculator corrects the syndrome by cyclically shifting the syndrome in reverse by the shift value D. In error position calculation step 84, the error calculator identifies the correct position of the error symbol in the codeword based on the corrected syndrome of step 82. After step 84, the method terminates.

[0047] The method of FIG2B begins with the syndrome calculation step 90: the syndrome calculator 34 calculates the syndrome for the zeroed leading bit 52 (if present on page 6 / 10 of the specification, CN 121713387 A), the codeword 50, and the zeroed trailing bit 54. As described above, the resulting syndrome is cyclically shifted by a non-zero shift value D, where D is equal to the number of trailing bits (or the number of trailing symbols). Step 90 is substantially identical to step 80 of the method of FIG2A.

[0048] In the error location calculation step 92, the error calculator identifies the position of the error symbol in the codeword based on the cyclically shifted syndrome. In the position correction step 94, the error locator corrects the error position by cyclically shifting the error position in the reverse direction by the shift value D. After step 94, the method terminates.

[0049] Iterative Complementary Computation Figure 3 is a block diagram schematically illustrating an ECC decoder 100 according to an embodiment described herein, which iteratively computes the companions over multiple blocks of a codeword and compensates for codeword misalignment.

[0050] It should be noted that the method of Figure 3 focuses on iterative companionary computation, and for clarity, functions performed by decoder elements such as the error calculator 38, error corrector 48, timing 46, and soft decoding logic 66 of Figure 1 are omitted from Figure 3.

[0051] Decoder 100 receives input data in the form of a block sequence 104. For clarity, Figure 3 depicts only three blocks, including a first block 104A, an intermediate block 104B, and a last block 104C. In this example, codeword 50 spans all blocks, where the last block contains a trailing bit 54 of a number D, and the first block optionally contains one or more preamble bits 52.

[0052] The decoder 100 includes a buffer 120 and circuitry 122, which includes a masking module 124 and multiple iterative synod calculators 130. For clarity, Figure 3 depicts only two iterative synod calculators—130A for exponentiation and 130B for exponentiation, where for primitive BCH codes, j and i are in the range 1…d-1, and “d” denotes the minimum distance of the code. In a practical implementation, the decoder may include up to d-1 different iterative synod calculators. The calculation of error locations (by error locator 38) is based on (d-1) synods. Typically, for narrow primitive BCH codes, only synods for odd powers are directly calculated, while synods for even powers can be derived from synods for odd powers with low complexity. For RS codes and general BCH codes, the required number of synods is d-1 = 2t, where “t” represents the number of correctable errors.

[0053] In the example of FIG3, buffer 120 and block 104 share the same length of 16 bits (in Equations 2 and 3). Alternatively, other suitable buffer and block lengths may be used.

[0054] Blocks are received and buffered in multiple iterations. For each iteration, masking module 124 masks the contents of buffer 120 with a corresponding mask value. In this example, the mask value of the first block contains a “0” bit at the leading bit position and a “1” bit in the codeword portion; the mask value of the middle block contains only a “1” bit; and the mask value of the last block contains a “0” bit at the trailing bit position and a “1” bit in the codeword portion. The effect of masking over the entire block is to preserve the codeword and set the leading and trailing bits to zero.

[0055] Each iteration in the iterative synod calculators (e.g., 130A and 130B) computes a synod value for the specific power of the primitive over the received block. Different iterative syntactic calculators have the same structure, but their internal circuitry differs depending on the power.As an example of the syndic calculator 130, the iterative syndic calculator 130A will now be described in detail.

[0056] The syndic calculator 130A receives the masking result from the masking module 124 in each iteration (i.e., each block) and updates the incremental calculation of the syndic.

[0057] The iterative syndic calculator 130A includes a feedforward circuit 140, a bitwise logic XOR gate 142, a feedback circuit 146, and a syndic buffer 148. As given in Equation 3 above, the iterative syndic calculator iteratively calculates the syndic. Specifically, the syndic buffer holds the updated syndic value up to the current iteration, the feedforward circuit 140 calculates the sum of the syndic in the current iteration, the feedback circuit calculates the product, and the bitwise XOR gate calculates the "+" operation in Equation 3.

[0058] In some embodiments, the iterative syndic calculator uses combinational logic to calculate the updated syndic (based on Equation 3) in each iteration. For example, in an embodiment, since the powers of primitives are known and do not change over time, combinational logic including logic XOR gates can be used to implement the summation operation computed by the feedforward circuit and the product operation implemented by the feedback circuit.

[0059] Example ECC Decoder for Extended BCH Codes Figure 4 is a block diagram schematically illustrating an ECC decoder 150 for decoding extended BCH codes according to an embodiment described herein.

[0060] It should be noted that, for clarity, the functions performed by decoder elements such as the error calculator 38, error corrector 48, timing 46, and soft decoding logic 66 of Figure 1 are omitted from Figure 4.

[0061] The example code of Figure 4 includes the extended BCH(256,239) code referenced above, which can be corrected for up to t=2 bits and is used as a component code in higher-order class product codes (e.g., the OFEC code mentioned above). For low-latency decoding, multiple decoders can be executed in parallel. For example, OFEC codes use a parallelism factor of 16, and the entire decoder can typically run hundreds of BCH decoders in parallel. Therefore, it is important to implement a low-complexity and low-power decoder, which, as mentioned above, can be achieved by applying a common iterative syntactic calculator circuit unit across all codeword blocks.

[0062] Decoder 150 receives an extended BCH codeword 152, which includes a 255-bit BCH codeword 154 and a single trailing parity bit 158. The parity bit is computed over the entire BCH codeword 154. The decoder receives the extended BCH codeword in 16-bit blocks 162, including a first block 162A, a middle block 162B, and a last block 162C. Each of the first 15 blocks includes a 16-bit BCH codeword, while the last block includes a 15-bit BCH codeword and a trailing parity bit 158.

[0063] The ECC decoder 150 includes a 16-bit buffer 166 and circuitry 168. The buffer 166 receives and temporarily stores a 16-bit block at a time. The masking module 170 receives a mask value block by block (e.g., from a timing module (not shown)) and uses this mask value to mask the contents of the buffer. In this example, the first 15 blocks are masked with an all-"1" mask value, while the last block is masked with 15 bits of all "1"s and a single zero bit (corresponding to a parity bit).

[0064] To decode the extended BCH codeword, the decoder 150 needs to compute the syndrome over the 255-bit BCH codeword and the parity result over the entire 256-bit extended BCH codeword.

[0065] In principle, since the number of correctable errors in this example is t=2, and the base code is a primitive BCH code, it is necessary to compute the syndrome for the four powers of the primitives… Furthermore, since the syndromes of even powers can be derived from the corresponding syndromes of odd powers with low complexity, the decoder 150 of FIG4 includes only two iterative syndrome calculators 174A and 174B for iteratively calculating the sum, for example using Equation 3 above. The calculated syndrome is cyclically shifted by one element because the parity bit 158 ​​is set to zero in the last block. In some embodiments, as described above, compensation for the parity bit misalignment can be performed during the calculation of the syndrome or during the calculation of the error position.

[0066] To check the parity bit across the entire extended BCH codeword, the decoder 150 includes a full XOR gate 180 and a parity bit register 184 initialized to zero. For each received block, the full XOR gate applies a 16-bit XOR operation to the 16 bits of the block with the current parity value in the parity bit register, and updates the parity bit register with the 1-bit XOR result. After the end block, the decoder verifies whether the value of the parity bit register is equal to "0".

[0067] Modifications to soft decoding In some embodiments, the received codeword is accompanied by soft information that can be used to improve decoding capabilities. For example, in a soft decoding method known as the "chase decoding method", the decoder attempts to decode multiple different test patterns derived from the received codeword.

[0068] In some embodiments, the soft information received with the original codeword includes the reliability level of bits (or symbols) assigned to the codeword. The soft information may also include a low reliability bit (LRB), which indicates the bit position (or symbol position) in the received codeword with the lowest reliability level. (In an alternative embodiment, the LRB is calculated by the decoder.) The decoder generates the test pattern by flipping selected bits in the original codeword indicated by the LRB.Since the syndrome of the received codeword is known, the decoder can efficiently compute the syndrome of the test pattern by modifying the original syndrome calculated for the received codeword based on the position of the flipped bit. For example, if a "0" bit in the received codeword is flipped, the power of the primitive corresponding to the flipped bit (or the symbol containing the flipped bit) is added to the syndrome of the received codeword. Similarly, if a "1" bit in the codeword is flipped, the power of the primitive corresponding to the flipped bit (or the symbol containing the flipped bit) is removed from the syndrome of the received codeword. It should be noted that for binary codes or GF(2m) codes, the addition and removal (of primitive powers) are mathematically the same operation. In Chase decoding of RS codes, flipping a single bit modifies the coefficient of the corresponding primitive power, which is used to modify the syndrome.

[0069] In cases where the syndrome of a received codeword is cyclically shifted by a certain shift value due to the tail bit (or the number of symbols matching the tail bit) being zeroed out by the number of shift values, the decoder generates a syndrome of a test mode from the original syndrome of the received codeword based on a cyclically shifted version of the LRB that depends on the shift value. The decoder uses the syndrome calculated for the test mode to apply hard decoding to the test mode to find the error location in the test mode. In this context, the phrase "cyclically shifting the LRB according to the shift value" means that the index in the codeword indicated by the LRB is cyclically shifted according to the shift value. For example, for binary code, the LRB is cyclically shifted by the same number of tail bits. In an embodiment, for RS code, the shift value is equal to the number of symbols matching the tail bit, and the LRB is shifted by that shift value multiplied by the number of bits per symbol.

[0070] FIG5 is a flowchart schematically illustrating a method for soft decoding of a codeword that is not aligned with the buffer boundary according to an embodiment described herein.

[0071] This method can be performed, for example, by decoders 20, 100, or 150. The method will be described for binary BCH codes. However, the method is similarly applicable to other cyclic codes, such as RS codes.

[0072] The method begins in receiving step 200, where the decoder receives a codeword, a reliability level assigned to the codeword bits, and a corresponding LRB indicating one or more codeword bits with the lowest reliability level. The codeword is stored in a buffer and is unaligned due to multiple trailing bits (shift values). In primitive syndrome generation step 204, the decoder generates primitive syndromes for the received codeword, which are cyclically shifted by setting the unaligned trailing bits of the codeword to zero. In pattern generation step 208, the decoder derives a test pattern from the received codeword based on the received LRB by flipping one or more bits of the codeword indicated by the LRB.

[0073] In order to generate a syndrome for the test mode based on the corresponding cyclically shifted syndrome of the received codeword, the positions of the powers of the primitives should be correctly aligned. For this purpose, the decoder cyclically shifts the LRB by a shift value in LRB shifting step 212. In test mode syndrome calculation step 216, the decoder calculates the test mode syndrome by modifying the original syndrome of the received codeword based on the cyclically shifted LRB. The syndrome of the test mode is shifted by the same shift value as the syndrome of the received codeword.

[0074] In decoding step 220, the decoder decodes the test mode by finding the cyclically shifted error position based on the cyclically shifted syndrome of the test mode. In correction step 224, the decoder reverses the shift value of the error position and uses the cyclically shifted error position to correct the test mode. After step 224, the method terminates.

[0075] The method of FIG5 is given by way of example, and other suitable methods may also be used. For example, instead of calculating the error position of the shift as described in the method of FIG2B, the decoder may perform a reverse shift of the cyclic shift as described in the method of FIG2A.

[0076] The configurations of decoder 20 of FIG1, decoder 100 of FIG3, and decoder 150 of FIG4 are illustrative configurations depicted for clarity only. In alternative embodiments, any other suitable decoder configuration may also be used.

[0077] The different elements and components of decoders 20, 100, and 150 may be implemented using dedicated hardware or firmware, such as using hard-wired or programmable logic, for example, in an application-specific integrated circuit (ASIC) or a field-programmable gate array (FPGA). Additionally or alternatively, some functions of decoders 20, 100, and 150, such as the accompanying calculator 34, error calculator 38, error corrector 42, soft decoder logic 66, and iterative accompanying calculators 130 and 174 (e.g., 130A, 130B, 174A, and 174B), can be implemented in software and / or using a combination of hardware and software elements. For clarity, elements that are not essential for understanding the disclosed techniques are omitted from the figures.

[0078] In some embodiments, some functions of decoders 20, 100, and 150 (e.g., the accompanying calculator 34, error calculator 38, error corrector 42, soft decoder logic 66, and iterative accompanying calculators 130 and 174 (e.g., 130A, 130B, 174A, and 174B)) can be implemented in one or more programmable processors that perform the functions described herein in a software-programmed manner.For example, the software can be downloaded electronically to any processor via a network, or alternatively or additionally, the software can be provided and / or stored on a non-transitory tangible medium (such as magnetic storage, optical storage, or electronic storage).

[0079] The embodiments described above are given by way of example, and other suitable embodiments may also be used. For example, although the above embodiments primarily relate to BCH codes, the disclosed embodiments are similarly applicable to other cyclic codes, such as, for example, RS codes. A low-complexity approach in RS codes is to use unalignment that is a multiple of the symbol size. Furthermore, in RS Chase decoding, when the bits are flipped to derive the test pattern, the synod is modified according to the position and value of the bits in the corresponding RS symbol.

[0080] The disclosed embodiments are also applicable to shortened cyclic codes, some of which have had some symbols removed. In such embodiments, the cyclic shift of the synod and the error position is based on the index of the non-shortened code, since the shortened code is not a cyclic code.

[0081] It should be understood that the above embodiments are cited by way of example, and the appended claims are not limited to what has been specifically shown and described above. Conversely, the scope includes combinations and sub-combinations of the various features described above, as well as variations and modifications of these features, which would be apparent to a person skilled in the art upon reading the foregoing description and are not disclosed in the prior art. Documents incorporated herein by reference are considered part of this application, and the definitions in this specification should be considered only, unless any term is defined in these incorporated documents in a manner that conflicts to some extent with the definitions expressly or implicitly made in this specification. Specification 10 / 10 pages 13 CN 121713387 A Figure 1 Specification Drawings 1 / 5 pages 14 CN 121713387 A Figure 2A Figure 2B Specification Drawings 2 / 5 pages 15 CN 121713387 A Figure 3 Specification Drawings 3 / 5 pages 16 CN 121713387 A Figure 4 Specification Drawings 4 / 5 pages 17 CN 121713387 A Figure 5 Specification Drawings 5 / 5 pages 18 CN 121713387 A.

Claims

1. A method for decoding, comprising: In the error-correcting code (ECC) decoder, a codeword containing multiple symbols of ECC is received in a buffer, the codeword being misaligned with one or both boundaries of the buffer; In the buffer, one or more bits that do not belong to the codeword are set to zero; Calculate one or more syndromes for the codeword across both the codeword and the zeroed bits; Based on the calculated syntactic, identify the position of one or more erroneous symbols in the codeword; and Error symbols in the identified locations are corrected to recover an error-free version of the codeword. The method further includes compensating for misalignment between the codeword and the buffer during the calculation of the syntactic or during the identification of the location of the erroneous symbol.

2. The method according to claim 1, wherein, Setting the bit to zero includes setting a first number of symbols matching the trailing bit after the codeword in the buffer to zero, thereby cyclically shifting the syntagmatic expression by the first number, and wherein identifying the position includes (i) compensating for the misalignment by cyclically shifting the syntagmatic expression in reverse by the first number, and then (ii) identifying the position based on the cyclically shifted syntagmatic expression.

3. The method according to claim 1, wherein, Setting the bit to zero includes setting a first number of symbols matching the trailing bit after the codeword in the buffer to zero, thereby cyclically shifting the syntagmatic expression by the first number, and wherein identifying the position includes calculating the position based on the cyclically shifted syntagmatic expression, and then compensating for the misalignment by cyclically shifting the position in the reverse direction by the first number.

4. The method according to any one of claims 1 to 3, wherein, Setting the bits to zero includes setting a second number of leading bits to zero before the codeword in the buffer.

5. The method according to any one of claims 1 to 3, wherein, Receiving the codeword includes receiving the codeword in the form of one or more blocks having a common length, and wherein calculating the syndrome includes iteratively calculating the syndrome over the blocks.

6. The method according to claim 5, wherein, Iteratively calculating the syndrome involves using a common combinational circuit on multiple blocks to calculate the given syndrome through multiple iterations.

7. The method according to any one of claims 1 to 3, wherein, The ECC includes cyclic codes selected from a list that includes Bosch-Joherry-Hokvenheim (BCH) codes and Reid-Solomon codes.

8. The method according to any one of claims 1-3, comprising: Receive (i) a low reliability bit (LRB) indicating a bit with a low reliability level in the codeword, and (ii) a shift value corresponding to the misalignment; The test pattern is calculated from the codeword using LRB; The shifted LRB is calculated by cyclically shifting the LRB according to the shift value; The corresponding test mode syndrome is derived from the syndrome by modifying the syndrome based on the shifted LRB. and The test pattern is hardware decoded using the test pattern association method to recover an error-free version of the codeword.

9. A decoder, comprising: A buffer configured to receive a codeword containing multiple symbols of error-correcting code (ECC), the codeword being misaligned with one or both boundaries of the buffer; and The circuit is configured as follows: In the buffer, one or more bits that do not belong to the codeword are set to zero; Calculate one or more syndromes for the codeword across both the codeword and the zeroed bits; Based on the calculated syntactic, identify the position of one or more erroneous symbols in the codeword; and Error symbols in the identified locations are corrected to recover an error-free version of the codeword. The circuit is also configured to compensate for misalignment between the codeword and the buffer during the calculation of the syntactic or during the identification of the location of the erroneous symbol.

10. The decoder according to claim 9, wherein, The circuit is configured to zero out the bit by zeroing out the first number of symbols that match the trailing bit after the codeword in the buffer, thereby causing the synod to be cyclically shifted by the first number, and to identify the position by: (i) compensating for misalignment by cyclically shifting the synod in the reverse direction by the first number, and then (ii) identifying the position based on the cyclically shifted synod.

11. The decoder according to claim 9, wherein, The circuit is configured to zero out the bit by setting the first number of symbols matching the trailing bit after the codeword in the buffer, thereby causing the comorbid cyclic shift to the first number. The position is identified by calculating the position based on a cyclically shifted adjoint, and then compensating for misalignment by cyclically shifting the position in the reverse direction by the first amount.

12. The decoder according to any one of claims 9 to 11, wherein, The circuit is configured to zero the bit by setting a second number of leading bits preceding the codeword in the buffer to zero.

13. The decoder according to any one of claims 9 to 11, wherein, The circuit is configured to receive the codeword in the form of one or more blocks having a common length, and to iteratively compute the syndrome on the blocks.

14. The decoder according to claim 13, wherein, The circuit is configured to iteratively compute the given syndrome by using a common combinational circuit on multiple blocks to compute the syndrome in multiple iterations.

15. The decoder according to any one of claims 9 to 11, wherein, The ECC includes cyclic codes selected from a list that includes Bosch-Joherry-Hokvenheim (BCH) codes and Reid-Solomon codes.

16. The decoder according to any one of claims 9 to 11, wherein, The circuit is configured as follows: Receive (i) a low reliability bit (LRB) indicating a bit with a low reliability level in the codeword, and (ii) a shift value corresponding to the misalignment; The test pattern is calculated from the codeword using the LRB; The shifted LRB is calculated by cyclically shifting the LRB according to the shift value; The corresponding test mode syndrome is derived from the syndrome by modifying the syndrome based on the shifted LRB. and The test pattern is hardware decoded using the test pattern association method to recover an error-free version of the codeword.