Torn write mitigation

• US 9,448,896 B2
• Filed: 08/07/2013
• Issued: 09/20/2016
• Est. Priority Date: 08/07/2013
• Status: Active Grant

• Current Assignee: Seagate Technology LLC (Seagate Technology Holdings Plc)
• Original Assignee: Seagate Technology LLC (Seagate Technology Holdings Plc)
• Inventors: Gaertner, Mark Allen, Trantham, Jon, Krishnamurthy, Vidya, Faulhaber, Steve, Yang, Yong
• Primary Examiner(s): Kudirka, Joseph
• Application Number:
  US13/961,755
• Publication Number:
  US 20150046747A1
• Time in Patent Office: 1,140 Days
• US Class Current: 1/1
• CPC Class Codes: G06F 11/1441 Resetting or repowering
• CPC Class Codes: G06F 11/1666 where the redundant component is memory or memory area
• CPC Class Codes: G06F 11/2015 Redundant power supplies power supply failure G06F1/30

Abstract

Torn write mitigation circuitry determines if a write operation to memory is in progress at or about a time of power loss. In response to the write operation being in progress at or about the time of the power loss, the torn write mitigation circuitry causes torn write data and metadata to be stored to a non-volatile cache. The torn write data comprise data left in a degraded or uncorrectable state as a result of the loss of power. The metadata describe the torn write data.

Claims

1. A method, comprising:

detecting a power loss to a memory;

in response to a write operation to the memory being in progress at or about a time of the power loss,

determining if write data corresponding to torn write data are redundantly stored, the torn write data comprising data left in a degraded, unverified, or uncorrectable state as a result of the power loss;

storing the write data and metadata that describes the write data to a non-volatile cache in response to the write data not being redundantly stored; and

not storing the write data and the metadata to the non-volatile cache in response to the write data being redundantly stored.

2. The method of claim 1, wherein storing the write data comprises providing backup power while storing the write data to the non-volatile cache.

3. The method of claim 1, wherein the write data further comprises one or more sectors of data that were scheduled to be written to the memory and had not yet been written to the memory prior to the power loss.

4. The method of claim 1, wherein:

the memory comprises one or more of a rotating magnetic disk, magnetic tape, NOR flash, non-volatile static random access memory (NVSRAM), NAND flash, magnetoresistive random access memory (MRAM), phase change memory (PCM), spin torque random access memory (STRAM), resistive random access memory (RRAM), and battery backed RAM; and

the non-volatile cache comprises one or more of NOR flash, NVSRAM, NAND flash, MRAM, PCM, STRAM, RRAM and battery backed RAM.

5. The method of claim 1, further comprising reading the write data from the non-volatile cache and writing the write data to the memory after power is restored.

6. The method of claim 1, further comprising, after power is restored:

reading the metadata from the non-volatile cache;

identifying a location of the memory to restore the write data using the metadata; and

re-writing the write data to the location of the memory.

7. The method of claim 1, wherein storing the write data comprises identifying sectors that have been written to the memory but for which a feedback signal was not generated indicating that the sectors are recoverable.

8. The method of claim 7, wherein the feedback signal is based on servo sample processing that confirms drive heads of the memory are within track on cylinder limits.

9. The method of claim 1, wherein:

storing the metadata comprises storing information about skipped sectors of the memory; and

further comprising, after power is restored, restoring the write data from the non-volatile cache to the memory using the information about the skipped sectors.

10. The method of claim 1, wherein the memory comprises a magnetic disk and further comprising tracking positions of the magnetic disk during execution of a write command.

11. The method of claim 10, wherein tracking the positions comprises:

identifying a stop position where the write operation stops during the power loss; and

further comprising calculating an amount of the write data to store to the non-volatile cache using the stop position.

12. The method of claim 1, further comprising:

determining if a repair activity is active at the time of the power loss; and

if the repair activity is active, storing data associated with the repair activity to the non-volatile cache.

13. An apparatus, comprising:

torn write mitigation circuitry configured to

determine if a write operation to memory is in progress at a time of a power loss, and

in response to the write operation being in progress at the time of the power loss,

determine if write data corresponding to torn write data are redundantly stored, the torn write data comprising data left in a degraded, unverified, or uncorrectable state as a result of the power loss,

store the write data and metadata to a non-volatile cache in response to the write data not being redundantly stored, the metadata describing the write data, and

not store the write data and the metadata to the non-volatile cache in response to the write data being redundantly stored.

14. The apparatus of claim 13, wherein the metadata are stored in a header of the write data and a field of the metadata indicates an amount of the write data that have been stored in the non-volatile cache.

15. The apparatus of claim 13, wherein the write mitigation circuitry is further configured to encode at least one of the write data and the metadata using an error detection or error correction code before the write data and metadata are stored in the non-volatile cache.

16. A system, comprising:

a memory;

a non-volatile cache;

power loss detection circuitry configured to detect a loss of host-supplied power; and

torn write mitigation circuitry configured to;

determine if a write operation to the memory is in progress at or about the time of a power loss, and

in response to the write operation being in progress at the time of the power loss,

determine if write data corresponding to torn write data are redundantly stored, the torn write data comprising data left in a degraded, unverified, or uncorrectable state as a result of the power loss,

store the write data and metadata to a non-volatile cache in response to the write data not being redundantly stored, the metadata describing the write data, and

not store the write data and the metadata to the non-volatile cache in response to the write data being redundantly stored.

17. The system of claim 16, wherein the non-volatile cache is capable of retaining the write data without externally-supplied power for an extended period of time and is capable of being reliably written to at a rate such that storage of the write data to the non-volatile cache is completed within an available time after the power loss.

18. The system of claim 16, wherein:

the system is a solid state drive storage device;

the memory comprises multi-level cell (MLC) NAND flash; and

non-volatile cache comprises a portion of the MLC NAND flash operated in single level cell mode.

19. The system of claim 16, wherein:

the system is a hard disk drive;

the memory is a rotating magnetic disk; and

the non-volatile cache is an electronic flash memory.