Method of storing downloadable firmware on bulk media

• US 20070150790A1
• Filed: 12/27/2005
• Published: 06/28/2007
• Est. Priority Date: 12/27/2005
• Status: Active Grant

• Current Assignee: SanDisk Technologies LLC (Western Digital Corporation)
• Original Assignee: SanDisk Corporation (Western Digital Corporation)
• Inventors: Gross, Stephen, Mee, Bryan
• Granted Patent: US 7,546,515 B2
• Time in Patent Office: Days
• US Class Current: 714/763
• CPC Class Codes: G06F 11/1044 with specific ECC/EDC distribution

Abstract

A portion of data stored in a non-volatile memory may be found to be corrupted when it is read. Where parity data is generated from portions of data and the parity data is stored with the portions of data, the corrupted data may be reconstructed from the parity data and uncorrupted portions of data.

Claims

1. A method of generating parity bit groups, comprising:

organizing data in a non-volatile memory system logically in a multidimensional data group array, the data group array having first data group rows in a first dimension and second data group rows in a second dimension that is orthogonal to the first dimension;

generating first parity bit groups corresponding to the first data group rows;

generating second parity bit groups corresponding to the second data group rows;

the first parity bit groups, the second parity bit groups and data groups of the data group array each having an equal number of bits; and

associating logically the generated parity bit groups with the corresponding data group rows.

2. The method of claim 1, wherein the data comprise:

items of data, represented by one of bits or bytes.

3. The method of claim 1, wherein generating first and second parity bit groups comprises:

generating parity bits by performing logical operations on corresponding items of data of the data groups along the data group rows.

4. The method of claim 3, wherein

the logical operations are at least one of summations and exclusive ors.

5. The method of claim 1, further comprising:

generating a parity bit group of the first parity bit groups by performing logical operations on the corresponding parity bits of the first parity bit groups.

6. The method of claim 1, wherein

the organized data correspond to firmware of the non-volatile memory system.

7. A method of managing corrupted data in a non-volatile memory array, the method comprising:

organizing data in a non-volatile memory system logically in a data group array, the data group array having one or more data group rows;

generating one or more parity bit groups, corresponding to the one or more data group rows a parity bit group having a number of bits equal to the number of bits in a data group;

associating logically the one or more generated parity bit groups with the one or more corresponding data group rows;

using an error correction code to identify a corrupted data group within the data group array; and

using the one or more parity bit groups to correct the corrupted data group.

8. The method of claim 7, wherein using the error correction code comprises:

computing and appending the error correction code to the data group as part of writing the data group to the non-volatile memory array;

recomputing an error correction code of the data group as part of a reading of the data;

comparing the stored error correction code of the data group with the recomputed error correction code of the data group; and

identifying the data group as corrupted when the appended error correction code does not agree with the recomputed error correction code.

9. The method of claim 7, wherein using the one or more parity bit groups to correct the corrupted data group comprises:

doing at least one of reconstructing or replacing the corrupted data group.

10. The method of claim 9, wherein the reconstructing of the corrupted data group comprises:

searching for a data group row in which the corrupted data group is the only corrupted data group.

11. The method of claim 9, wherein the reconstructing of the corrupted data group comprises:

reading in the parity bit group associated with a data group row when the data group row is identified in which the corrupted data group is the only corrupted data group.

12. The method of claim 11, wherein the reading of the parity group comprises:

using an error correction code to detect whether the read parity group is corrupted.

13. The method of claim 12, wherein the reconstructing of the corrupted data group comprises:

searching for a different data group row in which the corrupted data group is the only corrupted data group when the error correction code detects that the parity group of the identified data group row is corrupted.

14. The method of claim 13, wherein the correcting of the corrupted data group comprises:

flagging the corrupted data group when the parity groups of every data group row, in which the corrupted data group is the only corrupted data group, are corrupted; and

using an error correction code to identify a different corrupted data group.

15. The method of claim 11, wherein the reconstructing of the corrupted data group comprises:

reconstructing of the items of data of the corrupted data group by performing logical operations on the corresponding items of data of the non-corrupted data groups of the identified data group row and the corresponding parity bits of the parity bit group of the identified data group row.

16. The method of claim 15, wherein the logical operations comprise at least one of summations and exclusive ors.

17. The method of claim 10, wherein the managing of the corrupted data groups comprises:

flagging a corrupted data group, when no data group row is found in which the corrupted data group is the only corrupted data group; and

using an error correction code to identify a different corrupted data group.

18. The method of claim 7 wherein the one or more data group rows include a first data group row that consists of a first number of data groups used to generate a first parity bit group and a second data group row that consists of a second number of data groups used to generate a second parity bit group, the first number and the second number not being equal.

19. The method of claim 7 wherein subsequent to the using an error correction code to identify a corrupted data group, correction of the data group by ECC is attempted and the correction using the one or more parity bit groups only occurs after correction of the data group by ECC is unsuccessful.

20. A method of correcting data that is read from a non-volatile memory array within a memory card that is removably connected to a host, comprising:

reading a data group from the non-volatile memory array;

applying an Error Correction Code (ECC) algorithm to the data group to detect errors and repair ECC-repairable errors in the data group; and

subsequently, if errors that are not ECC-repairable are detected in the data group, using one or more additional data groups and a parity group that is stored in the card to reconstruct data within the data group and substituting the reconstructed data for the data that contains errors.

21. The method of claim 20 wherein the parity group was previously stored in the non-volatile memory at a time of writing of the data group.

22. The method of claim 21 wherein the parity group was calculated by an Exclusive OR (XOR) operation on the data group and the one or more additional data groups.

23. The method of claim 20 wherein, if errors that are not ECC-repairable are detected, the locations of the errors are determined by the ECC algorithm and the locations are used to determine which data to reconstruct.

24. The method of claim 20 wherein the data group, the additional data groups and the parity bit groups are sectors of data.

25. The method of claim 24 wherein the data group and the additional data groups are stored in a page of the memory array.

26. The method of claim 20 wherein the data group and the additional data groups form a first row in a first direction, and the data group is also part of a second row in a second direction, the second direction being orthogonal to the first direction.

27. The method of claim 26 wherein the parity group was generated from the first row and a second parity group was generated from the second row, the parity group and the first row used to reconstruct data within the data group only when reconstruction of the data within the parity group using the second row and the second parity group is not successful.

28. A method of storing data in a non-volatile memory array that has a minimum unit of programming of a page, a page containing multiple addressable units of data, comprising:

deriving a parity group by performing a parity operation on a plurality of addressable units of data, the plurality of addressable units of data designated for storage in a page of the non-volatile memory, the parity group having the same size as an addressable unit of data;

storing the plurality of addressable units of data in the page; and

storing the parity group in the non-volatile memory array.

29. The method of claim 28 wherein the parity group is stored in the page with the plurality of addressable units of data.

30. The method of claim 28 wherein the parity group is stored in a portion of the non-volatile memory that is dedicated to storage of parity groups.

31. The method of claim 28 wherein the plurality of addressable units of data form less than a full page of data, further comprising:

storing the plurality of addressable units of data in the page with padding data bits;

subsequently replacing the padding data bits with additional data thereby corrupting one of the plurality of addressable units of data; and

subsequently recreating the one of the plurality of addressable units of data from the parity group and the other ones of the plurality of addressable units of data.

32. The method of claim 28 wherein the non-volatile memory array is in a removable memory card that is connected to a host and the plurality of addressable units of data are sent by the host to the memory card for storage.