System and method for fault tolerant TCP offload

• US 7,224,692 B2
• Filed: 01/06/2003
• Issued: 05/29/2007
• Est. Priority Date: 09/04/2002
• Status: Expired due to Fees

• Current Assignee: Avago Technologies General IP PTE Limited (Broadcom, Inc.)
• Original Assignee: Broadcom Corporation (Broadcom, Inc.)
• Inventors: Fan, Kan Frankie
• Primary Examiner(s): Shah; Chirag G.
• Application Number:
  US10/337,029
• Publication Number:
  US 20040042412A1
• Time in Patent Office: 1,604 Days
• Field of Search: None
• US Class Current: 370/394
• CPC Class Codes: H04L 69/16 Implementation or adaptation of Internet protocol [IP], of transmission control protocol [TCP] or of user datagram protocol [UDP]

Abstract

Systems and methods that provide fault tolerant transmission control protocol (TCP) offloading are provided. In one example, a method that provides fault tolerant TCP offloading is provided. The method may include one or more of the following steps: receiving TCP segment via a TCP offload engine (TOE); calculating a TCP sequence number; writing a receive sequence record based upon at least the calculated TCP sequence number to a TCP sequence update queue in a host; and updating a first host variable with a value from the written receive sequence record.

Claims

1. A method for providing fault tolerant Transmission Control Protocol (TCP) offloading, the method comprising:

receiving a TCP segment via a TCP offload engine (TOE);

calculating a TCP sequence number;

updating a value of a first TOE variable with the calculated TCP sequence number;

writing a receive sequence record based upon at least the calculated TCP sequence number to a TCP sequence update queue in a host;

updating a first host variable with a value from the written receive sequence record; and

if the received TCP segment acknowledges previously sent data, then

updating a value of a second TOE variable with the calculated TCP sequence number,

writing a send sequence record based upon at least the updated value of the second TOE variable to the TCP sequence update queue, and

updating a value of a second host variable with a value from the written send sequence record.

2. A method according to claim 1, further comprising:

removing the written receive sequence record from the TCP sequence update queue, and wherein updating the first host variable comprises updating the first host variable with a value from the removed written receive sequence record.

3. The method according to claim 1, wherein the first host variable comprises a HOST_RCV_NEXT variable.

4. The method according to claim 1, wherein the first TOE variable comprises a RCV_NEXT variable.

5. The method according to claim 1, further comprising:

if the received TCP segment acknowledges previously sent data, then removing the written send sequence record from the TCP sequence update queue, and wherein updating the value of the second host variable comprises updating the value of the second host variable with a value from the removed written send sequence record.

6. The method according to claim 1, wherein the second TOE variable comprises a SND_UNA variable.

7. The method according to claim 1, wherein the second host variable comprises a HOST_SND_UNA variable.

8. The method according to claim 1, further comprising:

if a fault is detected in the TOE, then continuing a TCP connection in software using at least one of the first host variable, the second host variable and the TCP sequence update queue.

9. The method according to claim 1, further comprising:

monitoring the TCP sequence update queue to detect a hardware failure event generated by the TOE.

10. The method according to claim 9, further comprising:

if a hardware failure event generated by the TOE is detected, then updating TCP state variables using values reported by the TOE via the TCP sequence update queue.

11. The method according to claim 9, further comprising:

if a hardware failure event generated by the TOE is detected, then continuing a TCP connection in software.

12. The method according to claim 9, comprising:

if the TOE experiences hardware failure, then transferring all TCP, state variables to the host via the TCP sequence update queue.

13. The method according to claim 1, further comprising:

probing the TOE; and

if the TOE does not properly respond to the probing, then assigning the first host variable to the first TOE variable, assigning the second host variable to the second TOE variable, and simulating TCP processing.

14. The method according to claim 13, further comprising:

if the TOE does not properly respond to the probing, then continuing a TCP connection in software.

15. The method according to claim 13, wherein the TOE does not properly respond to the probing if a response is not received by the host within a particular period of time.

16. The method according to claim 13, wherein the TOE does not properly respond to the probing if a response is not received by the host.

17. The method according to claim 13, wherein the TOE does not properly respond to the probing if a response is not received by the host in an exspected manner.

18. The method according to claim 13, wherein probing the TOE comprises periodically probing the TOE.

19. The method according to claim 1, further comprising:

transferring at least a payload of the received TCP segment to a host buffer via a direct momory access (DMA) engine.

20. The method according to claim 17, wherein, after transferring at least the payload of the received TCP segment, then acknowledging receipt of the received TCP segment.

21. In a system for providing fault tolerant Transmission Control Protocol (TCP) offloading, the system having at least a TOE and a host, the TOE having at least a first TOE variable storage buffer and a second TOE variable storage buffer, the host having at least a first host variable storage buffer, a second host variable storage buffer and a TCP sequence update data structure, at least one circuit that operates to, at least:

receive a TCP segment at the TOE,

calculate a TCP sequence number based upon the received TCP segment,

write a receive sequence record based upon at least the calculated TCP sequence number to the TCP sequence update data structure,

update the first host variable storage buffer with a value from the written receive sequence record, and

if the received TCP segment acknowledges previously sent data, then;

update a value stored in the second TOE variable storage buffer with the calculated TCP sequence number,

write a send sequence record based upon at least the updated value stored in the second TOE variable storage buffer to the TCP sequence update data structure, and

update a value stored in the second host variable storage buffer with a value from the written send sequence record.

22. The system according to claim 21, wherein the at least one circuit operates to update the first TOE variable storage buffer with the calculated TCP sequence number.

23. The system according to claim 21, wherein at least a first portion of the at least one circuit is located in the TOE, and at least a second portion of the at least one circuit is located in the host.

24. The system according to claim 21, wherein the at least one circuit operates to determine whether the received TCP segment acknowledges previously sent data, and if the received TCP segment acknowledges previously sent data, then remove the written send sequence record from the TCP sequence update data structure and update the value of the second host variable with a value from the removed written send sequence record.

25. The system according to claim 21, wherein the at least one circuit operates to detect a fault in the TOE, and if a fault is detected in the TOE, then continue a TCP connection in software using at least one of the first host variable, the second host variable and the TCP sequence update data structure.

26. The system according to claim 21, wherein the at least one circuit operates to monitor the TCP sequence update data structure to detect a hardware failure event generated by the TOE.

27. The system according to claim 26, wherein if a hardware failure event generated by the TOE is detected, then the at least one circuit operates to update TCP state variables using values reported by the TOE via the TCP sequence update data structure.

28. The system according to claim 26, wherein if a hardware failure event generated by the TOE is detected, then the at least one circuit operates to continue a TCP connection in software.

29. The system according to claim 26, wherein if the TOE experiences hardware failure, then the at least one circuit operates to transfer all TCP state variables to the host via the TCP sequence update data structure.

30. The system according to claim 21, wherein the at least one circuit operates to probe the TOE, and if the TOE does not properly respond to the probing, then at least:

assign the first host variable to the first TOE variable,

assign the second host variable to the second TOE variable, and

simulate TCP processing.

31. The system according to claim 30, wherein if the TOE does not properly respond to the probing, then the at least one circuit operates to continue a TCP connection in software.

32. The system according to claim 30, wherein the at least one circuit operates to determine that the TOE does not properly respond to the probing if a response is not received by the host within a particular period of time.

33. The system according to claim 30, wherein the at least one circuit operates to determine that the TOE does not properly respond to the probing if a response is not received by the host.

34. The system according to claim 30, wherein the at least one circuit operates to determine that the TOE does not properly respond to the probing if a response is not received by the host in an expected manner.

35. The system according to claim 30, wherein the at least one circuit operates to probe the TOE by periodically probing the TOE.

36. The system according to claim 21, wherein the at least one circuit operates to transfer at least a payload of the received TCP segment to a host buffer via a direct memory access (DMA) engine.

37. The system according to claim 36, wherein after transferring at least the payload of the received TCP segment, the at least one circuit operates to acknowledge receipt of the received TCP segment.