Finite state machine with a single process context for a RAID system

• US 7,219,353 B2
• Filed: 02/27/2001
• Issued: 05/15/2007
• Est. Priority Date: 02/27/2001
• Status: Active Grant

• Current Assignee: Avago Technologies International Sales Pte Limited (Broadcom, Inc.)
• Original Assignee: Broadcom Corporation (Broadcom, Inc.)
• Inventors: Arnott, Randy M., Noya, Eric S., Franklin, Chris R., Wong, Jeffrey T.
• Primary Examiner(s): Nguyen; Van H.
• Application Number:
  US09/794,879
• Publication Number:
  US 20020120789A1
• Time in Patent Office: 2,268 Days
• Field of Search: 709/325, 709/328, 711/114, 711/102, 711/100, 326/46, 713/1, 713/2, 714/6, 369/47.14, 719/310, 719325-327
• US Class Current: 719/325
• CPC Class Codes: G06F 3/0605 by facilitating the interaction with a user or administrator
• CPC Class Codes: G06F 3/0659 Command handling arrangements, e.g. command buffers, queues, command scheduling
• CPC Class Codes: G06F 3/0689 Disk arrays, e.g. RAID, JBOD
• CPC Class Codes: G06F 9/4498 Finite state machines

Abstract

A finite state machine (FSM) for a redundant array of independent disk includes a single process context that maintains an entire finite state required for input/output operations performed in a RAID system. The finite state is only updated in response to calls and call-backs. The call-backs can include procedure returns and interrupt signals. The call is received directly from an application program, and the call-backs are received from a driver and passed back directly to the application software by the finite state machine.

Claims

1. A finite state machine stored on a computer-readable medium for a redundant array of independent disks, the finite state machine comprising:

a single process context configured to maintain a plurality of entire finite states required for an input/output operation performed in the redundant array of independent disks;

means for initializing the finite states only in response to a call; and

means for updating the finite states only in response to call-backs;

wherein there is one entire finite state for each one of a plurality of concurrent I/O operations performed in the redundant array of independent disks.

2. The finite state machine of claim 1 wherein the single process context and the redundant array of independent disks operate independently of an operating system process context.

3. The finite state machine of claim 1 wherein the call-backs include procedure returns and interrupt signals.

4. The finite state machine of claim 1 wherein the single process context is external to an operating system process context.

5. The finite state machine of claim 1 where in the input/output operation specifies a multi-block operation.

6. The finite state machine of claim 5 wherein the multi-block operation includes more than two blocks.

7. The finite state machine of claim 1 further comprising:

a bus connecting the finite state machine to an external computer system via a software interface configured to translate a driver call into the call for the finite state machine; and

a hardware interface configured to translate the call into a driver call after the finite state is initialized.

8. The finite state machine of claim 1 further comprising:

a bus connecting the finite state machine to a context-less driver via a software interface configured to translate a driver call into the call for the finite state machine; and

a hardware interface configured to translate the call into a driver calls after the finite state is initialized.

9. A finite state machine stored on a computer-readable medium for a redundant array of independent disks, the finite state machine comprising:

a single process context configured to maintain an entire finite state required for an input/output operation performed in the redundant array of independent disks;

means for initializing the finite state only in response to a call; and

means for updating the finite state only in response to call-backs,

wherein the call is received by the finite state machine directly from application software, and the call-backs are received from a driver and passed back directly to the application software by the finite state machine.

10. The finite state machine of claim 9 wherein the driver is a context-less driver.

11. The finite state machine of claim 10 wherein the context-less driver further comprises:

a software interface configured to translate a driver call into the call for the finite state machine; and

a hardware interface configured to translate the call into a driver call after the finite state is initialized.

12. A finite state machine stored on a computer-readable medium for a redundant array of independent disks, the finite state machine comprising:

a single process context configured to maintain an entire finite state required for a multi-block input/output operation including more than two blocks performed in the redundant array of independent disks;

means for initializing the finite state only in response to a call; and

means for updating the finite state only in response to call-backs,

wherein the finite state machine enters a first state in response to receiving the call, enters a second state in response to a call-back indicating completion an operation on a first block, enters a third state while performing an operation on all but a last block, enters a third state when initiating an operation on the last block, and enters a fourth state upon completing the operation on the last block.

13. A computer-implemented method for operating a redundant array of independent disks, the method comprising:

providing a finite state machine for the redundant array of independent disks;

maintaining a plurality of entire finite states required for an input/output operation performed in the redundant array of independent disks in a single process context;

initializing the finite states only in response to a call; and

updating the finite states only in response to call-backs;

wherein there is one entire finite state for each one of a plurality of concurrent I/O operations performed in the redundant array of independent disks.

14. The method of claim 13 wherein the call-backs include procedure returns and interrupt signals.

15. The method of claim 13 wherein the single process context is external to an operating system process context.

16. The method of claim 13 wherein the input/output operation specifies a multi-block operation.

17. A computer-implemented method for operating a redundant array of independent disks, the method comprising:

providing a finite state machine for the redundant array of independent disks;

maintaining an entire finite state required for an input/output operation performed in the redundant array of independent disks in a single process context;

initializing the finite state only in response to a call; and

updating the finite state only in response to call-backs, wherein the call is received by the finite state machine directly from application software, and the call-backs are received from a driver and passed back directly to the application software by the finite state machine.

18. The method of claim 17 wherein the driver is a context-less driver.

19. The method of claim 18 further comprising:

translating, in a software interface, a driver call into the call; and

translating, in a hardware interface, the call into a driver call after the finite state is initialized.

20. A computer-implemented method for operating a redundant array of independent disks (RAID) in response to input/output calls from an external application, the method comprising:

providing an interface in the RAID that receives the input/output calls from the external application and transmits responses thereto;

providing a RAID finite state process operating entirely within the RAID, connected to said interface, that controls a plurality of concurrent input/output operations performed in the RAID without reliance on an external operating system;

operating the RAID finite state process based on calls and call-backs received from the external application through said interface; and

maintaining a plurality of entire finite states in the RAID finite state process, there being one entire finite state for each one of the plurality of concurrent I/O operations performed in the redundant array of independent disks.

21. The method of claim 20, wherein the step of operating the RAID finite state process is performed in response to an input/output call transmitted by a mini-port driver.

22. The method of claim 20, wherein the step of providing an interface in the RAID includes designating a register where the external application places RAID instructions.

23. The method of claim 22, further including the steps of:

placing a RAID instruction from the external application in said register;

signaling said finite state process that an instruction is available;

transmitting an acknowledgement signal to the external application; and

transmitting status information to the external application as the finite state process attempts to perform the RAID instruction.

24. The method of claim 20, wherein the step of providing an interface in the RAID includes providing an interface to receive instructions directly from the external application software rather than an operating system driver.