Transfer of error-analysis and statistical data in a fibre channel input/output system

• US 20040193968A1
• Filed: 03/14/2003
• Published: 09/30/2004
• Est. Priority Date: 03/14/2003
• Status: Active Grant

• Current Assignee: International Business Machines Corporation
• Original Assignee: International Business Machines Corporation
• Inventors: Kalos, Matthew J., Yudenfriend, Harry M., Flanagan, John R., Ricci, Louis W., Frazier, Giles R., Trowell, Kenneth M., Dugan, Robert J.
• Granted Patent: US 7,324,455 B2
• Time in Patent Office: Days
• US Class Current: 714/57
• CPC Class Codes: H04L 1/20 using signal quality detector
• CPC Class Codes: H04L 41/0677 Localisation of faults
• CPC Class Codes: H04L 43/0811 by checking connectivity

Abstract

A computer Input/Output system having a fabric, a control unit (CU) and a host computer including a channel, the channel having a channel port connected by a first link to a channel neighbor port of the fabric, and the control unit having a CU port connected by a second link to a CU neighbor port on the fabric. When an error is detected in the link or protocol between the channel and the control unit, error data are reported to the channel by the channel, the channel port, the channel neighbor port, the control unit, the CU port and the CU neighbor port and sent to the channel. The error data are provided to the host computer for analysis.

Claims

1. In a computer I/O system having a channel and a control unit, a method comprising:

detecting an error in a link between the channel and the control unit;

gathering error data in the channel and the control unit;

reporting the error data gathered in the control unit to the channel; and

making the error data reported by the control unit and gathered by the channel available to a host computer.

2. The method according to claim 1 wherein the error data gathered in the channel include port error data from a port connecting the channel to the link.

3. The method according to claim 1 wherein the error data gathered in the control unit include port error data from a port connecting the control unit to the link.

4. A computer I/O system comprising:

a host computer

a channel in said host computer;

a control unit;

a link between the channel and the control unit;

an error detector detecting an error in said link;

a Purge Path Extension (PPE) facility in said channel gathering error data in the channel;

a PPE facility in said control unit gathering error data in the control unit;

said PPE facility in said control unit reporting the error data gathered in the control unit and reporting them to the channel; and

said PPE facility in said channel making the error data reported by the control unit and gathered by the channel available to a host computer.

5. The computer I/O system according to claim 4 further comprising a port connecting said channel to said link and wherein the error data gathered in the channel include port error data from said port.

6. The computer I/O system according to claim 4 further comprising a port connecting said control unit to said link and wherein the error data gathered in the control unit include port error data from said port.

7. In a computer system having a fabric, a control unit (CU), and a host computer including a channel, the channel having a channel port connected by a first link to a channel neighbor port of the fabric, and the control unit having a CU port connected by a second link to a CU neighbor port on the fabric, the channel comprising:

a Purge Path Extended (PPE) facility in the channel transmitting a PPE request via said fabric to the CU when a transmission error is detected;

said PPE facility sending a Read Link Status (RLS) request to the channel neighbor port requesting error data kept by the channel neighbor port;

said PPE facility accepting Link Error Status Block (LESB) data sent by the channel neighbor port in response to said RLS request;

said PPE facility further accepting error data from the control unit sent in response to said PPE request;

storage in said PPE facility storing error data accepted in response to said RLS request and said PPE request; and

said PPE facility further providing an I/O interrupt to the host computer making available to the host computer error data in said storage.

8. The channel according to claim 7 wherein said PPE request includes a reason code for the detected transmission error.

9. The channel according to claim 7 wherein said PPE facility accepts and stores LESB data from the channel port.

10. The channel according to claim 7 wherein said PPE facility accepts and stores model dependent error data from the channel.

11. The channel according to claim 7 wherein said PPE facility accepts and stores LESB data from the CU neighbor port and the CU port and model dependent data transmitted from the control unit to the channel responsive to said PPE request.

12. In a computer system having a fabric, a control unit (CU), and a host computer including a channel, the channel having a channel port connected by a first link to a channel neighbor port of the fabric, and the control unit having a CU port connected by a second link to a CU neighbor port of said fabric, the fabric comprising:

a switch transferring a Purge Path Extended (PPE) request transmitted by the channel to the CU when a transmission error is detected;

the channel neighbor port of the fabric transmitting a Link Error Status Block (LESB) to the channel when it receives a Read Link Status (RLS) request from the channel, said LESB containing error data kept by the channel neighbor port; and

the CU neighbor port of the fabric transmitting a second LESB to the control unit when it receives a RLS request from the control unit, said second LESB containing error data kept by the CU neighbor port; and

said switch transferring a PPE response from the control unit to the channel in response to said PPE request, said PPE response containing the error data data of said second LESB.

13. The fabric according to claim 12 wherein said PPE response includes third LESB data from the CU port and model dependent data transmitted from the control unit to the channel responsive to said PPE request.

14. In a computer system having a fabric, a control unit (CU), and a host computer including a channel, the channel having a channel port connected by a first link to a channel neighbor port of the fabric, and the control unit having a CU port connected by a second link to a CU neighbor port of said fabric, the control unit comprising:

a Purge Path Extended (PPE) facility for receiving a PPE request via said fabric transmitted by the channel to the CU when a transmission error is detected;

said PPE facility sending a Read Link Status (RLS) request to the CU neighbor port of the fabric requesting error data kept by the CU neighbor port;

said PPE facility accepting Link Error Status Block (LESB) data sent by the CU neighbor port in response to said RLS request;

storage in said PPE facility storing error data accepted in response to said RLS request; and

said PPE facility further transmitting said LESB in a PPE response to the channel responsive to said PPE request.

15. The control unit according to claim 14 wherein said PPE response includes a reason code for the detected transmission error.

16. The control unit according to claim 14 wherein said PPE facility accepts and stores LESB data from the CU port.

17. The control unit according to claim 14 wherein said PPE facility accepts and stores model dependent error data from the control unit.

18. The channel according to claim 14 wherein said PPE facility accepts transfers the LESB data from the CU neighbor port and the CU port and model dependent data stored in the PPE facility of the control unit.

19. A computer system comprising:

a host computer having a channel;

a fabric;

a control unit (CU);

a channel port in said channel;

a channel neighbor port in said fabric;

a first link connecting said channel port and said channel neighbor port,

a CU neighbor port in said fabric;

a CU port in said control unit;

a second link connecting said CU neighbor port and said CU port;

a channel Purge Path Extended (PPE) facility in said channel transmitting a PPE request via said first link when a transmission error is detected;

said channel PPE facility sending a Read Link Status (RLS) request to the channel neighbor port requesting error data kept by the channel neighbor port;

said channel PPE facility accepting Link Error Status Block (LESB) data sent by said channel neighbor port in response to said RLS request;

a switch in said fabric transferring said PPE request transmitted by the channel to the CU via said second;

said control unit having a CU PPE facility sending a second RLS request responsive to the receipt of said PPE request, to the CU neighbor port requesting error data kept by the CU neighbor port;

said channel neighbor port transmitting a second LESB to the control unit when it receives said second RLS request from the control unit, said second LESB containing error data kept by said CU neighbor port;

said CU channel PPE facility further accepting error data from said CU neighbor port sent in response to said second PPE request;

CU storage in said CU PPE facility storing error data accepted in response to said second RLS request;

said CU PPE facility transmitting error data stored in said CU storage to said channel in a PPE response;

said switch transferring said PPE response from the control unit to the channel;

said channel PPE facility further accepting error data from said PPE response;

channel storage in said channel PPE facility storing error data accepted in response to said first RLS request and said PPE request; and

said channel PPE facility further providing an I/O interrupt to said host computer making available to the host computer, error data in said channel storage.

20. The computer system according to claim 19 wherein the PPE request includes a channel determined reason code for the detected transmission error, and the PPE response includes a control unit determined reason code for the transmission error.

21. The computer system according to claim 19 wherein said channel PPE facility accepts and stores LESB data from the channel port, and said CU PPE facility accepts and stores LESB data from the CU port.

22. The computer system according to claim 19 wherein said channel PPE facility accepts and stores LESB data from the channel port, and said CU PPE facility accepts and stores LESB data from the CU port.

23. The computer system according to claim 19 wherein said channel PPE facility accepts and stores LESB data from the CU neighbor port and the CU port and model dependent data transmitted from the control unit to the channel in said PPE response.

24. In a computer system having a fabric, a control unit (CU), a host computer having a channel, a channel port in said channel, a channel neighbor port in said fabric, a first link connecting said channel port and said channel neighbor port, a CU neighbor port in said fabric, a CU port in said control unit, and a second link connecting said CU neighbor port and said CU port, a method comprising;

transmitting a PPE request by a channel Purge Path Extended (PPE) facility in said channel via said first link when a transmission error is detected;

sending a Read Link Status (RLS) request by said channel PPE facility to the channel neighbor port requesting error data kept by the channel neighbor port;

accepting Link Error Status Block (LESB) data by said channel PPE facility sent by said channel neighbor port in response to said RLS request;

transferring by a switch in said fabric, said PPE request transmitted by the channel to the CU via said second link when a transmission error is detected;

responsive to the receipt of said PPE request, sending a second RLS request by a CU PPE facility in said control unit to the CU neighbor port requesting error data kept by the CU neighbor port;

transmitting a second LESB by said channel neighbor port to the control unit when it receives said second RLS request from the control unit, said second LESB containing error data kept by said CU neighbor port;

accepting by said CU channel PPE facility further error data from said CU neighbor port sent in response to said second PPE request;

storing in CU storage in said CU PPE facility, error data accepted in response to said second RLS request;

transmitting by said CU PPE facility, error data stored in said CU storage to said channel in a PPE response;

transferring by said switch, said PPE response from the control unit to the channel;

further accepting by said channel PPE facility, error data from said PPE response;

storing in channel storage in said channel PPE facility, error data accepted in response to said first RLS request and said PPE request; and

providing by said channel PPE facility, an I/O interrupt to said host computer and making available to the host computer, error data in said channel storage.

25. The computer system according to claim 24 comprising including is said PPE request, a channel determined reason code for the detected transmission error, and including is said PPE response, a control unit determined reason code for the transmission error.

26. The method according to claim 24 further comprising accepting and storing in said channel PPE facility, LESB data from the channel port, and accepting and storing in said CU PPE facility, LESB data from the CU port.

27. The method according to claim 24 further comprising accepting and storing in said channel PPE facility, LESB data from the channel port, and accepting and storing in said CU PPE facility, LESB data from the CU port.

28. The method according to claim 24 further comprising accepting and storing in said channel PPE facility, LESB data from the CU neighbor port and the CU port and model dependent data transmitted from the control unit to the channel in said PPE response.