System and method of handling erroneous data in computer systems

• US 7,925,931 B1
• Filed: 12/13/2006
• Issued: 04/12/2011
• Est. Priority Date: 12/13/2006
• Status: Active Grant

• Current Assignee: NVIDIA Corporation, Racing Optics Incorporated
• Original Assignee: NVIDIA Corporation
• Inventors: Currid, Andrew, Overby, Mark A., Smith, Michael John Sebastian
• Primary Examiner(s): Bonzo; Bryce P
• Application Number:
  US11/610,469
• Time in Patent Office: 1,581 Days
• Field of Search: 714/43
• US Class Current: 714/43
• CPC Class Codes: G06F 11/0745 in an input/output transactions management context input/output processing in general G06F13/00
• CPC Class Codes: G06F 11/0793 Remedial or corrective actions recovery from an exception in an instruction pipeline G06F9/3861; by retry G06F11/1402; for recovering from a failure of a protocol instance or entity H04L69/40

Abstract

Embodiments of the present invention provide a method for handling errors in data servers. Generally, embodiments of the invention enable a data packet that is marked as erroneous to be handled so that it is not committed to permanent storage. One or more components are configured to recognize a poisoned data indicator, and to respond to the indicator by taking programmed actions to delete the data, to stop the data from being transmitted, to notify upstream components, and to purge related data from downstream components.

Claims

1. A method for handling erroneous data in a computer system, comprising:

detecting an error in data by a first component in the computer system positioned in a data path between a source of the data and a target for the data, wherein the first component is able to detect the error without receiving any feedback from the target;

reporting the error to components upstream in the data path by setting one or more bits indicating the error in the data; and

halting transmission of the data to downstream components.

2. The method of claim 1, wherein:

the error is detected in a portion of data that is part of a data transaction including other portions already sent downstream; and

the method further comprises sending a translated error message to downstream components to abort handling the transaction.

3. The method of claim 1, wherein:

the error is detected in a portion of data that is buffered in the first component; and

the method further comprises deleting the buffered portion of data in which the error is detected.

4. The method of claim 1, wherein reporting the error to components upstream in the data path by setting one or more bits indicating erroneous data comprises setting a poisoned data indicator of a data protocols being used to transmit the data.

5. The method of claim 1, wherein reporting of the data error to upstream components is carried out by an interrupt message of the Hypertransport data protocol indicating an Advanced Error Reporting (AER) failure.

6. The method of claim 2, wherein the translated error message sent to downstream components is a cyclic redundancy check (CRC) message of the Serial Advanced Technology Attachment (SATA) storage protocol.

7. The method of claim 2, wherein the translated error message sent to downstream components is an error message of the Ethernet networking protocol.

8. The method of claim 2, wherein the translated error message sent to downstream components is a CRC message of the Universal Serial Bus (USB) data protocol.

9. The method of claim 1, wherein the first component is a central processing unit (CPU).

10. The method of claim 1, wherein the first component is a Serial Advanced Technology Attachment (SATA) controller.

14. The computer system of claim 10, wherein the component incorporating error-handling logic is a Universal Serial Bus (USB) controller.

15. The computer system of claim 10, wherein:

the data handling component detects the error in a portion of data that is part of a data transaction including other portions already sent downstream; and

the data handling component sends a translated error message downstream components to abort handling the transaction.

11. A computer system configured to handle erroneous data, the computer system comprising:

at least one component that is a source of data;

at least one component that is a destination for data;

at least one bus that transmits the data from the source components to the destination components; and

at least one data handling component incorporating error-handling logic configured to detect an error in data being transferred from a source of data to a destination for data, wherein the first component is able to the error without receiving any feedback from the target, report the data error to one or more components upstream between the data handling component and the source of data, and halt transmission of the data to one or more downstream components between the data handling component and the destination for data.

12. The computer system of claim 11, wherein the data handling component incorporating error-handling logic is a Serial Advanced Technology Attachment (SATA) controller.

13. The computer system of claim 12, wherein the destination component is a SATA drive.

16. A Serial Advanced Technology Attachment (SATA) controller, comprising:

a first interface for receiving data from an upstream device;

a second interface for transmitting data to an SATA drive; and

error-handling logic configured to independently detect an error, without receiving any feedback from the SATA drive, in data being transferred from the upstream device, report the data error to the upstream device, and halt transmission of the data to the SATA drive.

17. The controller of claim 16, wherein:

the data handling component detects the error in a portion of data that is part of a data transaction including other portions already sent downstream; and

the data handling component sends a translated error message to downstream components to abort handling the transaction.

18. The controller of claim 17, wherein:

the error is detected in a portion of data that is buffered in the first component; and

the error-handling logic is further configured to delete the buffered portion of data in which the error is detected.

19. The controller of claim 16, wherein the error-handling logic reports the error to the upstream component by setting a poisoned data indicator of the data protocol being used to transmit the data.