Method for providing high availability within a data processing system via a reconfigurable hashed storage subsystem

• US 6,823,471 B1
• Filed: 07/30/1999
• Issued: 11/23/2004
• Est. Priority Date: 07/30/1999
• Status: Expired due to Fees

• Current Assignee: International Business Machines Corporation
• Original Assignee: International Business Machines Corporation
• Inventors: Clark, Leo James, Dodson, John Steve, Arimilli, Ravi Kumar, Guthrie, Guy Lynn, Lewis, Jerry Don
• Primary Examiner(s): Beausoliel, Robert
• Assistant Examiner(s): MASKULINSKI, MICHAEL C
• Application Number:
  US09/364,281
• Time in Patent Office: 1,943 Days
• Field of Search: 714/10, 714/9, 714/7, 714/11, 714/13, 714/3
• US Class Current: 714/10
• CPC Class Codes: G06F 11/20 using active fault-masking, e.g. by switching out faulty elements or by switching in spare elements

Abstract

A processor includes execution resources, data storage, and an instruction sequencing unit, coupled to the execution resources and the data storage, that supplies instructions within the data storage to the execution resources. At least one of the execution resources, the data storage, and the instruction sequencing unit is implemented with a plurality of hardware partitions of like function for processing a respective one of a plurality of data streams. If an error is detected in a particular hardware partition, the data stream assigned to that hardware partition is reassigned to another of the plurality of hardware partitions, thus preventing an error in one of the hardware partitions from resulting in a catastrophic failure.

Claims

1. A processor, comprising:

execution resources;

data storage; and

an instruction sequencing unit, coupled to said execution resources and said data storage, that supplies instructions within said data storage to said execution resources;

wherein of said execution resources, said data storage, and said instruction sequencing unit, at least said execution resources are implemented with a plurality of hardware partitions of like function for processing a respective one of a plurality of data streams, and wherein said instruction sequencing unit includes a hashing circuit that assigns said plurality of data streams to said plurality of hardware partitions based upon an address hash of addresses associated with instructions within said plurality of data streams, said hash being selected by a hash selection circuit within said processor, and wherein if an error is detected in a particular hardware partition among said plurality of hardware partitions that is assigned a particular data stream among said plurality of data streams to process, said hashing selection circuit reassigns said particular data stream to at least one other of said plurality of hardware partitions by changing the address hash implemented by the hashing circuit.

2. The processor of claim 1, wherein said processor reassigns said particular data stream in response to detection of a double-bit ECC error.

3. The processor of claim 1, wherein said processor reassigns said particular data stream dynamically during operation of said processor.

4. The processor of claim 1, wherein said data storage and said execution resources are implemented with a same number of hardware partitions.

5. The processor of claim 1, wherein following reassignment of said particular data stream, said particular hardware partition is idle.

6. A data processing system, comprising:

at least one interconnect;

at least one memory coupled to said interconnect; and

at least one processor coupled to said interconnect, wherein said processing includes;

execution resources;

data storage; and

an instruction sequencing unit, coupled to said execution resources and said data storage, that supplies instructions within said data storage to said execution resources;

wherein of said execution resources, said data storage, and said instruction sequencing unit, at least said execution resources are implemented with a plurality of hardware partitions of like function for processing a respective one of a plurality of data streams, and wherein said instruction sequencing unit includes a hashing circuit that assigns said plurality of data streams to said plurality of hardware partitions based upon an address hash of addresses associated with instructions within said plurality of data streams, said hash being selected by a hash selection circuit within said processor, and wherein if an error is detected in a particular hardware partition among said plurality of hardware partitions that is assigned a particular data stream among said plurality of data streams to process, said hashing selection circuit reassigns said particular data stream to at least one other of said plurality of hardware partitions by changing the address hash implemented by the hashing circuit.

7. The data processing system of claim 6, wherein said processor reassigns said particular data stream in response to detection of a double-bit ECC error.

8. The data processing system of claim 6, wherein said processor reassigns said particular data stream dynamically during operation of said processor.

9. The data processing system of claim 6, wherein said data storage and said execution resources are implemented with a same number of hardware partitions.

10. The data processing system of claim 6, wherein following reassignment of said particular data stream, said particular hardware partition is idle.

11. A method of operating a processor, including execution resources, data storage, and an instruction sequencing unit, coupled to said execution resources and said data storage, that supplies instructions within said data storage to said execution resources, wherein of said execution resources, said data storage, and said instruction sequencing unit, at least said execution resources are implemented with a plurality of hardware partitions of like function, said method comprising:

assigning a plurality of data streams to said plurality of hardware partitions based upon an address hash of addresses associated with instructions within said plurality of data streams;

processing each of the plurality of data streams within a respective one of the plurality of hardware partitions of like function; and

in response to detection of an error in a particular hardware partition among said plurality of hardware partitions that is processing a particular data stream among said plurality of data streams, reassigning said particular data stream to at least one other of said plurality of hardware partitions by changing said address hash.

12. The method of claim 11, wherein reassigning said particular data stream comprises reassigning said particular data stream in response to detection of a double-bit ECC error.

13. The method of claim 11, wherein reassigning said particular data stream comprises reassigning said particular data stream dynamically during operation of said processor.

14. The method of claim 11, wherein processing each of a plurality of data streams comprises processing said plurality of data streams within a corresponding plurality of hardware partitions in said data storage and within a corresponding plurality of hardware partitions in said execution resources.

15. The method of claim 11, and further comprising following reassignment of said particular data stream, idling said particular hardware partition.