Method and apparatus for caching with variable size locking regions

• US 20030018855A1
• Filed: 05/08/2002
• Published: 01/23/2003
• Est. Priority Date: 07/16/2001
• Status: Active Grant

• Current Assignee: Oracle America, Inc. (Oracle Corporation)
• Original Assignee: Sun Microsystems Incorporated (Oracle Corporation)
• Inventors: Cohen, Earl T., McWilliams, Thomas M.
• Granted Patent: US 6,772,299 B2
• Time in Patent Office: Days
• US Class Current: 711/128
• CPC Class Codes: G06F 12/126 with special data handling, e.g. priority of data or instructions, handling errors or pinning

Abstract

A method of managing data in a cache memory includes mapping a member of a plurality of memory addresses in a main memory onto a first member of a plurality of cache lines, locking the first member of the plurality of cache lines creating a locked cache region and an unlocked cache region, remapping the member of the plurality of memory addresses from the first member of the plurality of cache lines onto a second member of the plurality of cache lines within the unlocked cache region, requesting data stored in the main memory, fetching the data from the locked cache region, if available in the locked cache region, fetching the data from the unlocked cache region, if not available in the locked cache region and available in the unlocked cache region, and fetching the data from the main memory, if not available in the locked cache region and not available in the unlocked cache region.

Claims

1. A method of managing data in a cache memory, comprising:

mapping a member of a plurality of memory addresses in a main memory onto a first member of a plurality of cache lines;

locking the first member of the plurality of cache lines creating a locked cache region and an unlocked cache region;

remapping the member of the plurality of memory addresses from the first member of the plurality of cache lines onto a second member of the plurality of cache lines within the unlocked cache region;

requesting data stored in the main memory;

fetching the data from the locked cache region, if available in the locked cache region;

fetching the data from the unlocked cache region, if not available in the locked cache region and available in the unlocked cache region; and

fetching the data from the main memory, if not available in the locked cache region and not available in the unlocked cache region.

2. The method of claim 1, the locked cache region comprising a lowest address space in the cache memory.

3. The method of claim 1, wherein a size of the locked cache region is adjustable.

4. The method of claim 1, wherein a size of the unlocked cache region is adjustable.

5. The method of claim 1, wherein the data is requested data by a member of a plurality of processors.

6. The method of claim 1, further comprising:

detecting whether the data is within the unlocked cache region, using a register that maintains a size of the unlocked cache region.

7. The method of claim 1, further comprising:

associating with the second member of the plurality of cache lines, a tag field to identify information stored in the second member of the plurality of cache lines.

8. The method of claim 7, the tag field comprising a sufficient width to support remapping of the member of the plurality of memory addresses onto a smallest possible portion of the unlocked cache region.

9. A method of managing data in a cache memory, comprising:

mapping a member of a plurality of memory addresses in a main memory onto a first member of a plurality of cache lines;

locking the first member of the plurality of cache lines creating a locked cache region and an unlocked cache region;

remapping the member of the plurality of memory addresses from the first member of the plurality of cache lines onto a second member of the plurality of cache lines within the unlocked cache region;

requesting data stored in the main memory;

fetching the data from the locked cache region, if available in the locked cache region;

fetching the data from the unlocked cache region, if not available in the locked cache region and available in the unlocked cache region;

fetching the data from the main memory, if not available in the locked cache region and not available in the unlocked cache region;

detecting whether the data is within the unlocked cache region, using a register that maintains a size of the unlocked cache region; and

associating with the second member of the plurality of cache lines, a tag field to identify information stored in the second member of the plurality of cache lines.

10. A cache memory management system, comprising:

a main memory comprising a plurality of memory addresses; and

a cache memory comprising a plurality of cache lines, the plurality of cache lines comprising a locked cache region and an unlocked cache region;

wherein a member of the plurality of memory addresses is mapped onto a first member of the plurality of cache lines;

wherein the member of the plurality of memory addresses is configured to remap from the first member of the plurality of cache lines onto a second member of the plurality of cache lines;

wherein the first member of the plurality of cache lines is within the locked cache region, and the second member of the plurality of cache lines is within the unlocked cache region.

11. The cache memory management system of claim 10, the locked cache region comprising the lowest address space in the cache memory.

12. The cache memory management system of claim 10, wherein a size of the locked cache region is adjustable.

13. The cache memory management system of claim 10, wherein a size of the unlocked cache region is adjustable.

14. The cache memory management system of claim 13, further comprising:

a register maintaining a size of the unlocked cache region.

15. The cache memory management system of claim 10, further comprising:

an address detection mechanism configured to detect whether a referenced member of the plurality of memory addresses is within the locked cache region.

16. The cache memory management system of claim 10, further comprising:

a tag field associated with the second member of the plurality of cache lines to identify information stored therein.

17. The cache memory management system of claim 16, the tag field comprising a sufficient width to support remapping the member of the plurality of memory addresses onto the second member of the plurality of cache lines.

20. The computer system of claim 19, the locked cache region comprising a lowest address space in the cache memory.

21. The computer system of claim 19, wherein a size of the locked cache region is adjustable.

22. The computer system of claim 19, wherein a size of the unlocked cache region is adjustable.

23. The computer system of claim 19, wherein the data is requested data by a member of a plurality of processors.

24. The computer system of claim 19, further comprising:

detecting whether the data is within the unlocked cache region, using a register that maintains a size of the unlocked cache region.

25. The computer system of claim 19, further comprising:

associating with the second member of the plurality of cache lines, a tag field to identify information stored in the second member of the plurality of cache lines.

26. The computer system of claim 25, the tag field comprising a sufficient width to support remapping of the member of the plurality of memory addresses onto a smallest possible portion of the unlocked cache region.

18. A cache memory management system, comprising:

a main memory comprising a plurality of memory addresses;

a cache memory comprising a plurality of cache lines, the cache memory comprising a locked cache region and an unlocked cache region;

a register maintaining a size of the unlocked cache region; and

an address detection mechanism configured to detect whether a referenced member of the plurality of memory addresses is within the locked cache region;

wherein a member of the plurality of memory addresses is mapped onto a first member of the plurality of cache lines;

wherein the member of the plurality of memory addresses is configured to remap from the first member of the plurality of cache lines onto a second member of the plurality of cache lines;

wherein the first member of the plurality of cache lines is within the locked cache region, and the second member of the plurality of cache lines is within the unlocked cache region.

19. A computer system for managing data in a cache memory, comprising:

a processor;

a memory;

software instructions stored in the memory to cause the computer system to perform;

mapping a member of a plurality of memory addresses in a main memory onto a first member of a plurality of cache lines;

locking the first member of the plurality of cache lines creating a locked cache region and an unlocked cache region;

remapping the member of the plurality of memory addresses from the first member of the plurality of cache lines onto a second member of the plurality of cache lines within the unlocked cache region;

requesting data stored in the main memory;

fetching the data from the locked cache region, if available in the locked cache region;

fetching the data from the unlocked cache region, if not available in the locked cache region and available in the unlocked cache region; and

fetching the data from the main memory, if not available in the locked cache region and not available in the unlocked cache region.

27. An apparatus managing data in a cache memory, comprising:

means for mapping a member of a plurality of memory addresses in a main memory onto a first member of a plurality of cache lines;

means for locking the first member of the plurality of cache lines creating a locked cache region and an unlocked cache region;

means for remapping the member of the plurality of memory addresses from the first member of the plurality of cache lines onto a second member of the plurality of cache lines within the unlocked cache region;

means for requesting data stored in the main memory;

means for fetching the data from the locked cache region, if available in the locked cache region;

means for fetching the data from the unlocked cache region, if not available in the locked cache region and available in the unlocked cache region; and

means for fetching the data from the main memory, if not available in the locked cache region and not available in the unlocked cache region.