Recovery from iSCSI corruption with RDMA ATP mechanism

1. A method comprising:

detecting and managing an error detected in an iSCSI (Internet Small Computer System Interface) PDU (Protocol Data Unit) by using a RDMA (Remote Direct Memory Access) dedicated receive error queue for error recovery.

2. The method according to claim 1 , wherein said error comprises data corruption in iSCSI PDU data, and wherein if the iSCSI PDU consumes WQEs (work queue elements) in a receive queue, the error is reported via at least one of a completion queue and an Asynchronous Event Queue.

3. The method according to claim 1 , wherein said error comprises header corruption in iSCSI PDU data.

4. The method according to claim 3 , further comprising reporting the error using an Asynchronous Event Notification mechanism.

5. The method according to claim 4 , further comprising using a RNIC (Remote-direct-memory-access-enabled Network Interface Controller) for each iSCSI connection which comprises an Error Recovery Queue having at least two receive queues, wherein one of the receive queues is used to receive iSCSI Control PDUs and another receive queue is used for error recovery to allow software to take out data from reassembly buffers.

6. The method according to claim 1 , further comprising performing sequencing validation of the iSCSI PDU.

7. A computer program product comprising:

instructions for detecting and managing an error detected in an iSCSI PDU by using a RDMA dedicated receive error queue for error recovery.

8. The computer program product according to claim 7 , wherein said error comprises data corruption in iSCSI PDU data, and wherein if the iSCSI PDU consumes WQEs in a receive queue, and the instructions include instructions for reporting the error via a completion queue.

9. The computer program product according to claim 7 , wherein said error comprises data corruption in iSCSI PDU data, and wherein if the iSCSI PDU consumes WQEs in a receive queue, and the instructions include instructions for reporting the error via an Asynchronous Event Queue.

10. The computer program product according to claim 7 , wherein said error comprises header corruption in iSCSI PDU data.

11. The computer program product according to claim 10 , wherein the instructions include instructions for reporting the error using an Asynchronous Event Notification mechanism.

12. The computer program product according to claim 11 , further comprising instructions for implementing a RNIC for each iSCSI connection, the RNIC comprising an Error Recovery Queue having at least two receive queues, wherein one of the receive queues is used to receive iSCSI Control PDUs and another receive queue is used for error recovery to allow software to take out data from reassembly buffers.

13. The computer program product according to claim 12 , further comprising instructions for operating the RNIC in streaming mode, wherein all received data is placed to posted WQEs of the Error Recovery Queue.

14. The computer program product according to claim 7 , further comprising instructions for performing sequencing validation of the iSCSI PDU.

15. A system comprising:

an RDMA Service Unit;

an RDMA Messaging Unit operative to process inbound and outgoing RDMA messages, and to use services provided by said RDMA Service Unit to perform direct placement and delivery operations; and

an iSCSI Messaging Unit operative to perform an iSCSI offload function and to process inbound and outgoing iSCSI PDUs, said system being adapted to detect and manage an error detected in an iSCSI PDU by using a RDMA dedicated receive error queue for error recovery.

16. The system according to claim 15 , wherein said error comprises data corruption in iSCSI PDU data, and wherein if the iSCSI PDU consumes WQEs in a receive queue, the system comprises at least one of a completion queue and an Asynchronous Event Queue through which the error is reported.

17. The system according to claim 15 , wherein said error comprises header corruption in iSCSI PDU data.

18. The system according to claim 17 , wherein the system comprises an Asynchronous Event Notification mechanism through which the error is reported.

19. The system according to claim 18 , further comprising a RNIC for each iSCSI connection, the RNIC comprising an Error Recovery Queue having at least two receive queues, wherein one of the receive queues is used to receive iSCSI Control PDUs and another receive queue is used for error recovery to allow software to take out data from reassembly buffers.

20. The system according to claim 15 , wherein the system is adapted to perform sequencing validation of the iSCSI PDU.