Control logic design to support USB cache offload

• US 9,286,203 B2
• Filed: 01/07/2013
• Issued: 03/15/2016
• Est. Priority Date: 01/07/2013
• Status: Active Grant

• Current Assignee: Avago Technologies International Sales Pte Limited (Broadcom, Inc.)
• Original Assignee: Avago Technologies General IP PTE Limited (Broadcom, Inc.)
• Inventors: Abraham, Moby J., Anupindi, Lakshmana M., Skinner, R. Brian B., Rizzo, James A., Jander, Mark J.
• Primary Examiner(s): BRADLEY, MATTHEW A
• Application Number:
  US13/735,720
• Publication Number:
  US 20140195718A1
• Time in Patent Office: 1,163 Days
• Field of Search: 711/114
• US Class Current: 1/1
• CPC Class Codes: G06F 1/30 Means for acting in the event of power-supply failure or interruption, e.g. power-supply fluctuations for resetting only G06F1/24
• CPC Class Codes: G06F 11/00 Error detection; Error correction; Monitoring error detection, correction or monitoring in information storage based on relative movement between record carrier and transducer G11B20/18; monitoring, i.e. supervising the progress of recording or reproducing G11B27/36; in static stores G11C29/00
• CPC Class Codes: G06F 11/1441 Resetting or repowering
• CPC Class Codes: G06F 11/2015 Redundant power supplies power supply failure G06F1/30
• CPC Class Codes: G06F 12/0246 in block erasable memory, e.g. flash memory
• CPC Class Codes: G06F 12/0804 with main memory updating G06F12/0806 takes precedence
• CPC Class Codes: G06F 9/00 Arrangements for program control, e.g. control units program control for peripheral devices G06F13/10

Abstract

A redundant array of independent drives controller and board controlled cache off-loading during a power failure is described. Methods associated with the use of the redundant array of independent drives controller and board for controlled cache off-loading during a power failure are also described.

Claims

1. A method operable on a Redundant Array of Independent Drives (RAID) controller that is operable to communicatively couple to a host system through a Peripheral Component Interconnect Express (PCIe) connector, the method comprising:

monitoring a power supply from the PCIe connector to the RAID controller;

identifying when a power failure from the power supply to the RAID controller has taken place;

enabling an alternate power source for the RAID controller once the power failure is identified;

identifying whether a cache offload is required from cache on the RAID controller to a non-volatile storage device at a time of the power failure;

preparing the RAID controller for the cache offload during the power failure;

off-loading the cache from the RAID controller during the power failure;

determining whether the power supply from the PCIe connector to the RAID controller has been restored in response to completing the cache offload from the RAID controller;

booting the RAID controller if the power supply has been restored; and

disabling the alternate power source to power down the RAID controller if the power supply has not been restored.

2. The method of claim 1, further comprising:

initiating a power loss latch mode once the power failure has been identified.

3. The method of claim 2, further comprising:

monitoring if the power loss latch mode has been asserted and enabling an alternate clock source for the RAID controller.

16. The method of claim 1 further comprising:

blocking a reset signal from the PCIe connector to the RAID controller while the cache offload is in progress.

4. A Redundant Array of Independent Drives (RAID) storage controller operable to communicatively couple to a host system through a Peripheral Component Interconnect Express (PCIe) connector to control Input/Output (I/O) operations to a plurality of storage devices, the RAID controller comprising:

a cache operable to cache data of write I/O requests to the storage devices;

power control logic;

an alternate power supply communicatively coupled to the power control logic that is operable to supply power to the RAID storage controller; and

a Universal Serial Bus (USB) interface operable to communicatively couple to a nonvolatile storage device,

wherein the power control logic is further operable to detect a loss of power from the PCIe connector for the RAID storage controller, and responsive to the loss of power, is further operable to;

to activate the alternate power supply;

to initialize the USB interface using power from the alternate power supply;

to perform a cache offload to transfer the cached data to the nonvolatile storage device through the USB interface;

determine whether power has returned to the PCIe connector;

boot the RAID controller if the power has returned; and

disable the alternate power supply to power down the RAID controller if the power has not returned.

5. The RAID storage controller of claim 4, further comprising:

boot control logic communicatively coupled to the power control logic and operable to detect a reset signal from the PCIe connector for the RAID storage controller, to determine whether the cache offload is in progress, and to delay an application of the reset signal to the RAID storage controller until the cache offload is complete.

6. The RAID storage controller of claim 4, wherein:

the power control logic is further operable to power down modules of the RAID storage controller that are unnecessary for the cache offload in response to the loss of power.

7. The RAID storage controller of claim 4, further comprising:

boot control logic communicatively coupled to the power control logic,

wherein the power control logic is further operable to direct the boot control logic to reinitialize the RAID storage controller after the power has returned to the PCIe connector,

wherein the boot control logic is operable to restore the data to the cache memory from the nonvolatile storage device through the USB interface when the RAID storage controller reinitializes.

8. A method operable in a Redundant Array of Independent Drives (RAID) storage controller, the RAID storage controller being communicatively coupled to a host system through a Peripheral Component Interconnect Express (PCIe) connector to control Input/Output (I/O) operations of the host system to a plurality of storage devices, the method comprising:

caching, in a cache memory of the RAID storage controller, data of write I/O requests directed to the plurality of storage devices;

detecting a loss of power from the PCIe connector; and

in response to detecting the loss of power;

activating an alternate power supply;

initializing a Universal Serial Bus (USB) interface using power from the alternate power supply;

performing a cache offload by transferring data from the cache memory to a nonvolatile storage device through the USB interface;

determining whether power has returned to the PCIe connector;

booting the RAID controller if the power has returned; and

disabling the alternate power supply to power down the RAID controller if the power has not returned.

9. The method of claim 8, further comprising:

detecting a reset signal from the PCIe connector for the RAID storage controller;

determining whether the cache offload is in progress; and

delaying an application of the reset signal to the RAID storage controller until the cache offload is complete.

10. The method of claim 8, further comprising:

powering down modules of the RAID storage controller that are unnecessary for the cache offload in response to detecting the loss of power.

11. The method of claim 8, further comprising:

reinitializing the RAID storage controller after the power has returned to the PCIe connector; and

restoring the data to the cache memory from the nonvolatile storage device through the USB interface when the RAID storage controller reinitializes.

12. A non-transitory computer readable medium for a Redundant Array of Independent Drives (RAID) storage controller, the RAID storage controller being communicatively coupled to a host system through a Peripheral Component Interconnect Express (PCIe) connector to control Input/Output (I/O) operations of the host system to a plurality of storage devices, the computer readable medium comprising instructions that, when directed by processor of the RAID storage controller, direct the RAID storage controller to:

cache data of write I/O requests directed to the plurality of storage devices;

detect a loss of power from the PCIe connector; and

in response to the loss of power;

activate an alternate power supply;

initialize a Universal Serial Bus (USB) interface using power from the alternate power supply;

perform a cache offload to transfer data from the cache memory to a nonvolatile storage device through the USB interface;

determine whether power has returned to the PCIe connector;

boot the RAID controller if the power has returned; and

disable the alternate power supply to power down the RAID controller if the power has not returned.

13. The computer readable medium of claim 12, further comprising instructions that direct the RAID storage controller to:

detect a reset signal from the PCIe connector for the RAID storage controller;

determine whether the cache offload is in progress; and

delay an application of the reset signal to the RAID storage controller until the cache offload is complete.

14. The computer readable medium of claim 12, further comprising instructions that direct the RAID storage controller to:

power down modules of the RAID storage controller that are unnecessary for the cache offload in response to the loss of power.

15. The computer readable medium of claim 12, further comprising instructions that direct the RAID storage controller to:

reinitialize the RAID storage controller after the power has returned to the PCIe connector; and

restore the data to the cache memory from the nonvolatile storage device through the USB interface when the RAID storage controller reinitializes.

17. A method operable on a Redundant Array of Independent Drives (RAID) controller that is operable to communicatively couple to a host system through a Peripheral Component Interconnect Express (PCIe) connector, the method comprising:

monitoring power supplied from the PCIe connector to the RAID controller;

determining that a power failure from the PCIe connector to the RAID controller has occurred;

determining whether an I/O data cache on the RAID controller is dirty responsive to the power failure; and

responsive to determining that the I/O data cache is dirty;

activating a backup power supply to power the RAID controller;

performing a cache offload from the RAID controller to non-volatile memory;

determine whether power has returned to the PCIe connector;

boot the RAID controller if the power has returned; and

disable the alternate power supply to power down the RAID controller if the power has not returned.

18. The method of claim 17 further comprising:

masking a reset signal from the PCIe connector to the RAID controller while the cache offload is in progress.