Method and apparatus for propagating error status over an ECC protected channel

• US 6,622,268 B2
• Filed: 11/29/2000
• Issued: 09/16/2003
• Est. Priority Date: 11/29/2000
• Status: Expired due to Term

• Current Assignee: Intel Corporation
• Original Assignee: Intel Corporation
• Inventors: Holman, Thomas J.
• Primary Examiner(s): DILDINE JR, R STEPHEN
• Application Number:
  US09/725,221
• Publication Number:
  US 20020138803A1
• Time in Patent Office: 1,021 Days
• Field of Search: 714/50, 714/701, 714/703-704, 714/708, 714/753, 714/765, 714/785
• US Class Current: 714/701
• CPC Class Codes: G11B 20/18 Error detection or correction; Testing , e.g. of drop-outs
• CPC Class Codes: H03M 13/01 Coding theory basic assumptions; Coding bounds; Error probability evaluation methods; Channel models; Simulation or testing of codes
• CPC Class Codes: H03M 13/03 Error detection or forward error correction by redundancy in data representation, i.e. code words containing more digits than the source words

Abstract

System and methods for propagating error status over an error checking and correcting (ECC) protected channel. A first device receives data and an error status associated with the data. The first device generates check bits for the data based on a first ECC code and combines the check bits with the data to form one or more code words. The first device sends the code words across the channel where the first device inserts a triple error into a nibble of at least one codeword sent if the error status indicated an uncorrectable error. A second device connected to the channel receives the code words sent across the channel. The second device detects triple errors within a nibble of any code word and any other single error in the code word using a second ECC code, where the second ECC code is the first ECC code with columns for check bits inserted.

Claims

1. A system for propagating error status over an error checking and correcting (ECC) protected channel comprising:

a first device, the first device receiving data and an error status associated with the data, the first device generating check bits for the data based on a first ECC code and combining the check bits with the data to form at least one codeword;

a channel operably connected to the first device, the first device sending the at least one codeword across the channel, the first device inserting a triple error into a nibble of one at least one codeword sent if the error status indicated an uncorrectable error; and

a second device operably connected to the channel, the second device receiving the at least one codeword sent across the channel, the second device detecting triple errors within a nibble of the at least one codeword and any other single error in the at least one codeword using a second ECC code.

2. The system according to claim 1, wherein the other single error comprises an error introduced by the channel.

3. The system according to claim 1, wherein the first device includes a controller, the controller receiving the error status.

4. The system according to claim 1, wherein the first device includes a code generator, the code generator generating check bits for the data based on the first ECC code and combining the check bits with the data to form the at least one codeword.

5. The system according to claim 4, wherein the check bits are located in the at least one codeword to allow for single 4 bit nibble error detection (S4ED).

6. The system according to claim 1, wherein the second ECC code comprises the first ECC code with additional columns inserted denoting designated positions of check bits, the check bit positions providing for single 4 bit nibble error detection (S4ED).

7. The system according to claim 1, wherein a check bit is generated for each row of the first ECC code, the data comprising at least one data word, each check bit comprising the dot product of a row of the first ECC code and a data word.

8. The system according to claim 1, wherein the first device includes an error injection circuit, the error injection circuit injecting a triple error into a nibble of the at least one codeword if the error status indicated an uncorrectable error.

9. The system according to claim 1, wherein the second device includes a syndrome processor, the syndrome processor receiving the at least one codeword and re-computing the check bits on the data using the second ECC code, the second device generating a syndrome by comparing the check bits in the at least one codeword with the re-computed check bits.

10. The system according to claim 9, wherein the syndrome comprises the dot product of the second ECC code AND each at least one codeword.

11. The system according to claim 9, further comprising an error classifier, the error classifier classifying errors in the received at least one codeword based on the syndrome.

12. The system according to claim 11, wherein the classification comprises one of no error, correctable error, and uncorrectable error.

13. The system according to claim 1, wherein the error status indicates one of no error, correctable error, and uncorrectable error.

14. The system according to claim 1, wherein the first ECC code comprises a matrix.

15. A method for propagating error status over an error checking and correcting (ECC) protected channel comprising:

receiving data and an error status associated with the data at a first device;

generating check bits for the data based on a first ECC code and combining the check bits with the data to form at least one codeword at the first device;

sending the at least one codeword across a channel, the first device inserting a triple error into a nibble of at least one codeword sent if the error status indicated an uncorrectable error; and

receiving the at least one codeword sent across the channel by a second device, the second device detecting any triple errors within a nibble of the at least one codeword and any other single error in the at least one codeword using a second ECC code.

16. The method according to claim 15, wherein the other single error comprises an error introduced by the channel.

17. The method according to claim 15, further comprising placing the check bits in the at least one codeword to allow for single 4 bit nibble error detection (S4ED).

18. The method according to claim 15, wherein the second ECC code comprises the first ECC code with additional columns inserted denoting designated positions of check bits, the check bit positions providing for single 4 bit nibble error detection (S4ED).

19. The method according to claim 15, further comprising generating a check bit for each row of the first ECC code, the data comprising at least one data word, each check bit comprising the dot product of a row of the first ECC code and a data word.

20. The method according to claim 15, further comprising recomputing the check bits on the data using the second ECC code by the second device, the second device generating a syndrome by comparing the check bits in the at least one codeword with the re-computed check bits.

21. The method according to claim 20, wherein the syndrome comprises the dot product of the second ECC code AND each at least one codeword.

22. The method according to claim 20, further comprising using the syndrome to classify errors in the received at least one codeword.

23. The method according to claim 20, further comprising determining a weight of nibbles of the syndrome where the weight is the number of 1'"'"'s in the syndrome, the classification being based on the weight.

24. The method according to claim 22, wherein the classification comprises one of no error, correctable error, and uncorrectable error.

25. The method according to claim 15, wherein the error status indicates one of no error, correctable error, and uncorrectable error.

26. An article comprising a storage medium containing instructions stored therein, the instructions when executed causing a processing device to perform:

receiving data and an error status associated with the data;

generating check bits for the data based on an ECC code and combining the check bits with the data to form at least one codeword; and

sending the at least one codeword across a channel, the processing device inserting a triple error into a nibble of the at least one codeword sent if the error status indicated an uncorrectable error.

27. The article according to claim 26, wherein the error status indicates one of no error, correctable error, and uncorrectable error.

28. The article according to claim 26, wherein the ECC code comprises a matrix.

29. The article according to claim 26, wherein a check bit is generated for each row of the ECC code, the data comprising at least one data word, each check bit comprising the dot product of a row of the ECC code and a data word.