Systems and methods for enabling threads to lock a stage prior to processing data

• US 8,584,126 B2
• Filed: 03/18/2008
• Issued: 11/12/2013
• Est. Priority Date: 03/23/2007
• Status: Active Grant

• Current Assignee: Vmware LLC (Broadcom, Inc.)
• Original Assignee: VMware, Inc. (Broadcom, Inc.)
• Inventors: Ang, Boon Seong, Lambeth, Andrew, Ramanan, Jyothir
• Primary Examiner(s): PETRANEK, JACOB ANDREW
• Application Number:
  US12/050,861
• Publication Number:
  US 20080235690A1
• Time in Patent Office: 2,065 Days
• Field of Search: 718/102
• US Class Current: 718/102
• CPC Class Codes: G06F 2209/548 Queue
• CPC Class Codes: G06F 9/526 Mutual exclusion algorithms
• CPC Class Codes: G06F 9/546 Message passing systems or structures, e.g. queues

Abstract

A system and method for maintaining processing order while permitting parallelism. Processing of a piece of work is divided into a plurality of stages. At each stage, a task advancing the work towards completion is performed. By performing processing as a sequence of tasks, processing can be done in parallel, with progress being made simultaneously on different pieces of work in different stages by a plurality of threads of execution.

Claims

1. A method for processing pieces of work in a plurality of stages, the method comprising:

taking a lock associated with a first stage;

performing a task associated with the first stage on a first piece of work;

determining if a lock associated with a second stage is available; and

if the lock associated with the second stage is not available, storing the first piece of work in a queue associated with the second stage and taking a second piece of work from a queue associated with the first stage and performing the task associated with the first stage on the second piece of work.

2. The method of claim 1, wherein determining if the lock associated with the second stage is available comprises:

attempting to take the lock associated with the second stage; and

determining if the attempt to take the lock associated with the second stage was successful.

3. The method of claim 1, further comprising:

if the lock associated with the second stage is available, taking a second piece of work from the queue associated with the second stage and performing the task associated with the second stage on the second piece of work.

4. The method of claim 3, further comprising:

taking a lock associated with a third stage;

performing a task associated with the third stage on the second piece of work; and

performing the task associated with the third stage on the first piece of work.

5. The method of claim 1, further comprising:

taking the lock associated with the second stage;

performing the task associated with the second stage on the first piece of work; and

performing the task associated with the second stage on the second piece of work.

6. The method of claim 1, further comprising determining if the queue associated with the second stage contains a second piece of work.

7. The method of claim 1, further comprising performing the task associated with the first stage on a second piece of work.

8. The method of claim 7, further comprising if the lock associated with the second stage is available, storing the second piece of work in a queue associated with a third stage.

9. The method of claim 7, further comprising if the lock associated with the second stage is not available, taking a lock associated with a third stage and performing a task associated with the third stage on the second piece of work.

10. A method for processing packets in a virtual switch connected to a plurality of virtual machines, the method comprising:

taking a lock associated with a first stage;

performing a task associated with the first stage on a first packet;

determining if a lock associated with a second stage is available; and

if the lock associated with the second stage is not available, storing the packet in a queue associated with the second stage and taking a second packet from a queue associated with the first stage and performing the task associated with the first stage on the second packet.

11. The method of claim 10, further comprising:

if the lock associated with the second stage is available, taking a second packet from the queue associated with the second stage; and

performing the task associated with the second stage on the second packet.

12. The method of claim 11, further comprising:

taking a lock associated with a third stage;

performing a task associated with the third stage on the second packet; and

performing the task associated with the third stage on the first packet.

13. The method of claim 10, further comprising:

taking the lock associated with the second stage;

performing the task associated with the second stage on the first packet; and

performing the task associated with the second stage on the second packet.

14. The method of claim 10, further comprising performing the task associated with the first stage on a second packet.

15. The method of claim 14, further comprising if the lock associated with the second stage is available, storing the second packet in a queue associated with a third stage.

16. The method of claim 14, further comprising if the lock associated with the second stage is not available, taking a lock associated with a third stage and performing a task associated with the third stage on the second packet.

17. The method of claim 10, wherein performing the task associated with the first stage comprises checking the first packet for errors.

18. The method of claim 10, wherein performing the task associated with the first stage comprises identifying a destination address of the first packet.

19. The method of claim 10, wherein performing the task associated with the first stage comprises determining a virtual switch port for transmission of the first packet.

20. The method of claim 10, wherein performing the task associated with the second stage comprises delivering the first packet to a virtual machine monitor.

21. A method for processing pieces of work in a plurality of stages, the method comprising:

taking a first lock associated with a first stage;

performing a task associated with the first stage on a first piece of work;

taking a second lock associated with the first stage;

releasing the first lock associated with the first stage;

determining if a lock associated with a second stage is available;

if the lock associated with the second stage is available, taking the lock associated with the second stage, releasing the second lock associated with the first stage, and performing a task associated with the second stage on the first piece of work; and

if the lock associated with the second stage is not available, storing the first piece of work in a queue associated with the second stage and releasing the second lock associated with the first stage.

22. A system for maintaining processing order while permitting parallelism, the system comprising:

a first queue associated with a first stage;

a first lock associated with the first stage;

a second lock associated with a second stage;

a first thread of execution assigned to a first piece of work, wherein the first thread of execution is configured to;

take the first lock;

perform a first task associated with the first stage on the first piece of work;

determine if the second lock is available;

if the second lock is available, take the second lock and perform a second task on the first piece of work; and

if the second lock is not available, store the first piece of work in the first queue; and

a second thread of execution assigned to a second piece of work, wherein the second thread of execution is configured to;

take the first lock;

perform the first task on the second piece of work;

determine if the second lock is available;

if the second lock is available, take the second lock and perform the second task on the second piece of work; and

if the second lock is not available, store the second piece of work in the first queue.

23. A non-transitory computer-readable storage medium comprising computer-readable instructions for processing pieces of work in a plurality of stages, the computer-readable storage medium causing one or more processors to perform the steps of:

taking a lock associated with a first stage;

performing a task associated with the first stage on a first piece of work;

determining if a lock associated with a second stage is available; and

if the lock associated with the second stage is not available, storing the first piece of work in a queue associated with the second stage and taking a second piece of work from a queue associated with the first stage and performing the task associated with the first stage on the second piece of work.

24. The non-transitory computer-readable storage medium of claim 23, wherein determining if the lock associated with the second stage is available comprises:

attempting to take the lock associated with the second stage; and

determining if the attempt to take the lock associated with the second stage was successful.

25. The non-transitory computer-readable storage medium of claim 23, wherein the computer-readable storage medium further causes the one or more processors to perform the steps of:

if the lock associated with the second stage is available, taking a second piece of work from the queue associated with the second stage; and

performing the task associated with the second stage on the second piece of work.

26. The non-transitory computer-readable storage medium of claim 25, wherein the computer-readable storage medium further causes the one or more processors to perform the steps of:

taking a lock associated with a third stage;

performing a task associated with the third stage on the second piece of work; and

performing the task associated with the third stage on the first piece of work.

27. The non-transitory computer-readable storage medium of claim 23, wherein the computer-readable storage medium further causes the one or more processors to perform the steps of:

taking the lock associated with the second stage;

performing the task associated with the second stage on the first piece of work; and

performing the task associated with the second stage on the second piece of work.

28. The non-transitory computer-readable storage medium of claim 23, wherein the computer-readable storage medium further causes the one or more processors to perform the step of determining if the queue associated with the second stage contains a second piece of work.

29. The non-transitory computer-readable storage medium of claim 23, wherein the computer-readable storage medium further causes the one or more processors to perform the step of performing the task associated with the first stage on a second piece of work.

30. The non-transitory computer-readable storage medium of claim 29, wherein the computer-readable medium further causes the one or more processors to perform the step of storing the second piece of work in a queue associated with a third stage if the lock associated with the second stage is available.

31. The non-transitory computer-readable storage medium of claim 29, wherein the computer-readable storage medium further causes the one or more processors to perform the steps of:

if the lock associated with the second stage is not available, taking a lock associated with a third stage; and

performing a task associated with the third stage on the second piece of work.

32. A non-transitory computer-readable storage medium comprising computer-readable instructions for processing packets in a virtual switch connected to a plurality of virtual machines, the computer-readable storage medium causing one or more processors to perform the steps of:

taking a lock associated with a first stage;

performing a task associated with the first stage on a first packet;

determining if a lock associated with a second stage is available;

if the lock associated with the second stage is available, taking the lock associated with the second stage, releasing the lock associated with the first stage, and performing a task associated with the second stage on the first packet; and

if the lock associated with the second stage is not available, storing the packet in a queue associated with the second stage and taking a second packet from a queue associated with the first stage and performing the task associated with the first stage on the second packet.

33. The non-transitory computer-readable storage medium of claim 32, wherein the computer-readable storage medium further causes the one or more processors to perform the steps of:

taking a second packet from the queue associated with the second stage if the lock associated with the second stage is available; and

performing the task associated with the second stage on the second packet.

34. The non-transitory computer-readable storage medium of claim 33, wherein the computer-readable storage medium further causes the one or more processors to perform the steps of:

taking a lock associated with a third stage;

performing a task associated with the third stage on the second packet; and

performing the task associated with the third stage on the first packet.

35. The non-transitory computer-readable storage medium of claim 32, wherein the computer-readable storage medium further causes the one or more processors to perform the steps of:

taking the lock associated with the second stage;

performing the task associated with the second stage on the first packet; and

performing the task associated with the second stage on the second packet.

36. The non-transitory computer-readable storage medium of claim 32, wherein the computer-readable storage medium further causes the one or more processors to perform the step of performing the task associated with the first stage on a second packet.

37. The non-transitory computer-readable storage medium of claim 36, wherein the computer-readable storage medium further causes the one or more processors to perform the step of storing the second packet in a queue associated with a third stage if the lock associated with the second stage is available.

38. The non-transitory computer-readable storage medium of claim 36, wherein the computer-readable storage medium further causes the one or more processors to perform the step of taking a lock associated with a third stage and performing a task associated with the third stage on the second packet if the lock associated with the second stage is not available.

39. The non-transitory computer-readable storage medium of claim 32, wherein performing the task associated with the first stage comprises checking the first packet for errors.

40. The non-transitory computer-readable storage medium of claim 32, wherein performing the task associated with the first stage comprises identifying a destination address of the first packet.

41. The non-transitory computer-readable storage medium of claim 32, wherein performing the task associated with the first stage comprises determining a virtual switch port for transmission of the first packet.

42. The non-transitory computer-readable storage medium of claim 32, wherein performing the task associated with the second stage comprises delivering the first packet to a virtual machine monitor.

43. A non-transitory computer-readable storage medium for processing pieces of work in a plurality of stages, the computer-readable storage medium causing one or more processors to perform the steps of:

taking a first lock associated with a first stage;

performing a task associated with the first stage on a first piece of work;

taking a second lock associated with the first stage;

releasing the first lock associated with the first stage;

determining if a lock associated with a second stage is available;

if the lock associated with the second stage is available, taking the lock associated with the second stage, releasing the second lock associated with the first stage, and performing a task associated with the second stage on the first piece of work; and

if the lock associated with the second stage is not available, storing the first piece of work in a queue associated with the second stage and releasing the second lock associated with the first stage.

44. A method for processing packets in a virtual switch connected to a plurality of virtual machines, the method comprising:

assigning a thread of execution to a first packet;

executing the thread of execution, wherein executing the thread of execution comprises;

taking a lock associated with a first stage;

performing a task associated with the first stage on the first packet;

determining if a lock associated with a second stage is available; and

if the lock associated with the second stage is available;

taking the lock associated with the second stage;

releasing the lock associated with the first stage;

determining if a second packet is in a queue associated with the second stage; and

if a second packet is in the queue associated with the second stage,

performing a task associated with the second stage on the second packet prior to performing the task associated with the second stage on the first packet.

45. The method of claim 44, wherein executing the thread of execution further comprises performing the task associated with the second stage on the first packet after performing the task associated with the second stage on the second packet.

46. The method of claim 45, wherein executing the thread of execution further comprises:

taking a lock associated with a third stage;

performing a task associated with the third stage on the second packet; and

performing the task associated with the third stage on the first packet.

47. A system comprising:

a virtual machine monitor comprising a virtual network interface controller, the virtual network interface controller comprising a first processing stage and a lock for the first processing stage;

a virtual network switch comprising a second processing stage and a lock for the second processing stage; and

at least one processor programmed to;

take the lock for the first processing stage;

perform a task associated with the first processing stage on a piece of work;

determine if the lock for the second stage is available; and

if the lock associated with the second stage is not available, store the piece of work in a queue associated with the second stage, take a second packet from a queue associated with the first stage, and perform the task associated with the first stage on the second packet.

48. A system comprising:

a virtual network switch comprising a first processing stage and a lock for the first processing stage;

a virtual machine monitor comprising a virtual network interface controller, the virtual network interface controller comprising a second processing stage and a lock for the second processing stage; and

at least one processor programmed to;

take the lock for the first processing stage;

perform a task associated with the first processing stage on a piece of work;

determine if the lock for the second stage is available; and

if the lock associated with the second stage is not available, store the piece of work in a queue associated with the second stage, take a second packet from a queue associated with the first stage, and perform the task associated with the first stage on the second packet.

49. A method for maintaining processing order while permitting parallelism, the method comprising:

assigning a first thread of execution to a first piece of work;

executing the first thread of execution, wherein executing the first thread of execution comprises;

taking a first lock associated with a first stage;

performing a first task associated with the first stage on the first piece of work;

determining if a second lock associated with a second stage is available;

if the second lock is available, taking the second lock and perform a second task on the first piece of work; and

if the second lock is not available, storing the first piece of work in a first queue associated with the first stage; and

assigning a second thread of execution to a second piece of work; and

executing the second thread of execution, wherein executing the second thread of execution comprises;

taking the first lock;

performing the first task on the second piece of work;

determining if the second lock is available;

if the second lock is available, taking the second lock and performing the second task on the second piece of work; and

if the second lock is not available, storing the second piece of work in the first queue.