Method and apparatus for allocating storage in a flash memory

• US 5,860,082 A
• Filed: 03/28/1996
• Issued: 01/12/1999
• Est. Priority Date: 03/28/1996
• Status: Expired due to Term

• Current Assignee: Datalight Incorporated (Tuxera Oy)
• Original Assignee: Datalight Incorporated (Tuxera Oy)
• Inventors: Smith, Kevin B., Garvin, P. Keith
• Primary Examiner(s): Swann, Tod R.
• Assistant Examiner(s): Chow, Christopher S.
• Application Number:
  US08/623,661
• Time in Patent Office: 1,020 Days
• Field of Search: 711/103, 711/154, 711/156, 711/171, 711/202, 395/500
• US Class Current: 711/103
• CPC Class Codes: G06F 12/0246 in block erasable memory, e.g. flash memory
• CPC Class Codes: G06F 3/0608 Saving storage space on storage systems
• CPC Class Codes: G06F 3/0644 Management of space entities, e.g. partitions, extents, pools
• CPC Class Codes: G06F 3/0679 Non-volatile semiconductor memory device, e.g. flash memory, one time programmable memory [OTP]

Abstract

A method and apparatus are provided for operating a flash memory (24) as a client memory, wherein the flash memory comprises a partition (28) of physical data storage locations. The partition is divided into a plurality of units (30). Each unit is organized into a plurality of allocations (60) where each allocation includes a plurality of blocks (58) of physical data storage locations. The partition is organized into a plurality of regions (46) which define a client address space such that each region represents a portion of the client address space. Interfacing the client address space with the physical address space requires maintaining an allocation list (56) for each unit that includes an allocation (68) entry describing each allocation in the unit. Each allocation entry includes a first field (70) storing a client start address for the allocation in the client address space and a second field (72) storing an allocation length indicating the number of blocks of data in that allocation. A third field (74) stores a status of the allocation, i.e., either valid, pending, bad or discarded, and a fourth field (76) stores a physical start address for the allocation in the partition.

Claims

1. A method for allocating storage in an auxiliary storage medium, the auxiliary storage medium comprising at least one partition of physical address space, the method comprising:

(a) organizing the partition into a plurality of regions, each region defining a portion of client address space which represents a number of units of physical address space in the partition; and

(b) organizing each unit of physical address space into a plurality of allocations, each allocation including at least one block of physical data storage locations in the unit.

2. The method of claim 1, further comprising using an allocation list for each unit of the partition to interface the client address space defined by the regions and the physical address space of the partition, the allocation list including at least one allocation entry describing each allocation in the unit, each allocation entry including:

(i) a client start address for the allocation in the portion of client address space defined by the region representing the unit;

(ii) an allocation length indicating the number of blocks in the allocation;

(iii) a status of the allocation; and

(iv) a physical start address for the allocation in the unit.

3. The method of claim 2, wherein using the allocation list for each unit comprises writing data to the partition at a given client range, the given client range including a given client start address in the client address space and a given allocation length for the data, writing data to the partition at the given client range comprising:

(a) determining a region of client address space representing the units in which to write the data by determining if the given client range falls within the portion of client address space defined by the region;

(b) determining if there is available physical address space in the units represented by the region to write the data;

(c) writing the data to the units represented by the region having available physical address space, wherein writing the data creates a new allocation in at least one unit represented by the region, the new allocation having a physical start address in the unit, an allocation length and a status indicating that data is currently being written to the new allocation; and

(d) adding a new allocation entry to the allocation list that describes the new allocation, the new allocation entry including the given client start address, the allocation length of the new allocation, the status of the new allocation, and the physical start address of the new allocation.

4. The method of claim 3, further comprising reclaiming physical address space in the units represented by the region if there is not available physical address space in the units represented by the region to write the data.

5. The method of claim 3, wherein writing data to the units represented by the region further comprises discarding data in the given client range that was superseded when the data was written to the partition.

6. The method of claim 5, wherein writing data to the units represented by the region further comprises updating the status of the new allocation entry to indicate the data of the new allocation is valid after data that was superseded has been discarded.

7. The method of claim 3, wherein writing data to the units represented by the region further comprises:

(a) scanning the allocation list of a unit represented by the region to determine a first available physical start address in the unit to which data can be written;

(b) creating the new allocation in the unit by writing as much data as possible to the unit at the first available physical start address; and

(c) repeating (a)-(b) for each unit represented by the region until the data for the region has been written.

8. The method of claim 3, wherein a plurality of regions are selected in which to write the data using the given client range and wherein (a)-(d) are repeated for each region until all of the data has been written to the partition.

9. The method of claim 2, wherein using the allocation list further comprises reading data from the partition at a given client range, the given client range including a given client start address in the client address space and a given allocation length for the data, reading data from the partition at the given client range comprising:

(a) determining a region of client address space representing the units from which to read the data by determining if the given client range falls within the portion of client address space defined by the region;

(b) scanning the allocation list of a unit represented by the region to determine a physical start address and a length for at least a portion of the data to be read;

(c) reading the portion of data from the unit at the physical start address and length; and

(d) repeating (b)-(c) until as much data as possible has been read from the units represented by the region.

10. The method of claim 9, wherein scanning the allocation list of the unit further comprises:

(a) locating an allocation entry in the allocation list of the unit having a client range comprising a client start address and an allocation length that overlaps the given client range; and

(b) determining the physical start address and length for the portion of the data using the overlapping given client range and the client range of the allocation entry.

11. The method of claim 10, wherein a plurality of regions are selected from which to read the data using the given client range and wherein (a)-(d) are repeated for each region until all of the data has been read from the partition.

12. The method of claim 2, wherein using the allocation list for each unit further comprises reclaiming physical address space in the units represented by the region in a given client range, the given client range including a given client start address in the client address space and a given allocation length for the data, reclaiming physical address space comprising:

(a) allocating at least one of the units represented by the region as a spare unit;

(b) selecting a unit represented by the region to be reclaimed;

(c) for each allocation in the selected unit described by an allocation entry in the allocation list of the unit, copying valid data from the selected unit to the spare unit by;

(i) examining the allocation entry in the allocation list that describes the allocation;

(ii) if the allocation entry includes a status indicating that the data in the unit is valid, copying the data in the allocation to a new allocation in the spare unit having a new physical start address in the unit; and

(iii) adding a new allocation entry to the spare allocation list that includes (A) the client start address of the allocation in the selected unit from which the data is being copied;

(B) the allocation length of the allocation in the selected unit from which the data is being copied;

(C) the new physical start address of the new allocation in the spare unit; and

(D) a status indicating that the data copied to the spare unit is valid;

(d) updating a unit header in the selected unit to indicate a new spare unit;

(e) updating a unit header in the spare unit to indicate a valid unit;

(f) erasing the new spare unit; and

(g) updating a unit header in the selected unit to indicate a spare unit.

13. The method of claim 2, wherein using the allocation list for each unit further comprises discarding data in a given client range, the given client range including a given client start address in the client address space and a given allocation length for the data, discarding data in a given client range comprising:

(a) determining a region of client address space representing the units in which data is to be data by determining if the given client range falls within the portion of client address space defined by the region;

(b) scanning the allocation list of a unit represented by the region for an old allocation entry describing an old allocation having data to be discarded;

(c) determining if the data to be discarded is located at a beginning of the old allocation described by the old allocation entry, and if not, creating a new allocation entry for the existing data located at the beginning of the old allocation by adding the new allocation entry describing the new allocation to the allocation list of the unit, the new allocation entry including (i) the client start address of the old allocation;

(ii) a new allocation length for the new allocation;

(iii) the physical start address of the old allocation; and

(iv) a status indicating that the data is valid;

(d) determining if the data to be discarded is located at an end of the old allocation described by the old allocation entry, and if not, creating a new allocation entry for the existing data located at the end of the old allocation by adding the new allocation entry describing the new allocation to the allocation list of the unit, the new allocation entry including (i) a new client start address for the new allocation;

(ii) a new allocation length for the new allocation;

(iii) a new physical start address for the new allocation; and

(iv) a status indicating that the data is valid;

(e) updating the status of the old allocation entry to indicate that the allocation entry is no longer valid; and

(f) repeating (b)-(e) for each unit within the region until all of the data to be discarded in the given client range is discarded.

14. The method of claim 13, wherein a plurality of regions are selected in which to discard data using the given client range and wherein (a)-(f) are repeated for each region until all of the data in the given client range that was superseded is discarded from the partition.

15. The method of claim 2, wherein a computer readable medium is encoded to perform the function of using an allocation list for each unit of the partition to interface the client address space defined by the regions and the physical address space of the partition.

16. An apparatus for allocating storage of data amongst storage mediums in a computer system, the apparatus comprising:

(a) a processing unit;

(b) an auxiliary storage medium, the auxiliary storage medium comprising at least one partition, the partition organized into a plurality of regions, each region defining a portion of client address space which represents a number of units in the partition;

(c) a main storage medium coupled to the processing unit, the storage medium at least temporarily storing program instructions that control the processing unit;

execution of said program instructions on the processing unit comprising;

(i) organizing each unit into a plurality of allocations, each allocation including at least one block of physical data storage locations in the unit; and

(iii) transferring storage of data to and from the auxiliary storage medium by using an allocation list for each unit of the partition to translate the client address space defined by the regions and the physical address space of the partition, the allocation list including at least one allocation entry describing each allocation in the unit, each allocation entry including (A) a client start address for the allocation in the portion of client address space defined by the region representing the unit;

(B) an allocation length indicating the number of blocks in the allocation;

(C) a status of the allocation; and

(D) a physical start address for the allocation in the unit.

17. The apparatus of claim 16, wherein transferring storage of data to the auxiliary storage medium further comprises writing data to the partition of the auxiliary storage medium at a given client range, the given client range including a given client start address in the client address space and a given allocation length for the data, writing data to the partition at the given client range comprising:

(a) determining a region of client address space representing the units in which to write the data by determining if the given client range falls within the portion of client address space defined by the region;

(b) determining if there is available physical address space in the units represented by the region to write the data;

(c) writing the data to the units represented by the region having available physical address space, wherein writing the data creates a new allocation in at least one unit represented by the region, the new allocation having a physical start address in the unit, an allocation length and a status indicating that data is currently being written to the new allocation; and

(d) adding a new allocation entry to the allocation list that describes the new allocation, the new allocation entry including the given client start address, the allocation length of the new allocation, the status of the new allocation, and the physical start address of the new allocation.

18. The apparatus of claim 17, further comprising reclaiming physical address space in the units represented by the region if there is not available physical address space in the units represented by the region to write the data.

19. The apparatus of claim 17, wherein writing data to the units represented by the region further comprises discarding superseded data in the given client range.

20. The apparatus of claim 19, wherein writing data to the units represented by the region further comprises updating the status of the new allocation entry to indicate the data of the new allocation is valid after the superseded data has been discarded.

21. The apparatus of claim 17, wherein writing data to the units represented by the region further comprises:

(a) scanning the allocation list of a unit represented by the region to determine a first available physical start address in the unit to which data can be written;

(b) creating the new allocation in the unit by writing as much data as possible to the unit at the first available physical start address; and

(c) repeating (a)-(b) for each unit represented by the region until the data for the region has been written.

22. The apparatus of claim 17, wherein a plurality of regions are selected in which to write the data using the given client range and wherein (a)-(d) are repeated for each region until all of the data has been written to the partition.

23. The apparatus of claim 16, wherein transferring data from the auxiliary storage medium further comprises reading data from the partition of the auxiliary storage medium at a given client range, the given client range including a given client start address in the client address space and a given allocation length for the data, reading data to the partition at the given client range comprising:

(x) determining a region of client address space representing the units from which to read the data using the given client range;

(y) scanning the allocation list of a unit represented by the region to determine a physical start address and a length for at least a portion of the data to be read;

(z) reading the portion of data from the unit at the physical start address and length; and

(aa) repeating (b)-(c) until as much data as possible has been read from the units represented by the region.

24. The apparatus of claim 23, wherein scanning the allocation list of the unit further comprises:

(a) locating an allocation entry in the allocation list of the unit having a client range comprising a client start address and an allocation length that overlaps the given client range; and

(b) determining the physical start address and length for the at least a portion of the data using the overlapping given client range and the client range of the allocation entry.

25. The apparatus of claim 23, wherein a plurality of regions are selected from which to read the data using the given client range and wherein (a)-(d) are repeated for each region until all of the data has been read from the partition.

26. The apparatus of claim 16, wherein transferring data to and from the auxiliary storage medium further comprises reclaiming physical address space in the units represented by the region in a given client range, the given client range including a given client start address in the client address space and a given allocation length for the data, reclaiming physical address space comprising:

(a) allocating one of the units represented by the region as a spare unit;

(b) selecting a unit represented by the region to be reclaimed;

(c) for each allocation in the selected unit described by an allocation entry in the allocation list of the unit, copying valid data from the selected unit to the spare unit by;

(i) examining the allocation entry in the allocation list that describes the allocation;

(ii) if the allocation entry includes a status indicating that the data in the unit is valid, copying the data in the allocation to a new allocation in the spare unit having a new physical start address in the unit; and

(iii) adding a new allocation entry to the spare allocation list that includes (1) the client start address of the allocation in the selected unit from which the data is being copied;

(2) the allocation length of the allocation in the selected unit from which the data is being copied;

(3) the new physical start address of the new allocation in the spare unit; and

(4) a status indicating that the data copied to the spare unit is valid;

(d) updating a unit header in the selected unit to indicate a new spare unit;

(e) updating a unit header in the spare unit to indicate a valid unit;

(f) erasing the new spare unit; and

(g) updating a unit header in the selected unit to indicate a spare unit.

27. The apparatus of claim 16, wherein transferring data to and from the auxiliary storage medium further comprises discarding data in a given client range, the given client range including a given client start address in the client address space and a given allocation length for the data, discarding data in a given client range comprising:

(a) determining a region of client address space representing the units in which data is to be discarded by determining if the given client range falls within the portion of client address space defined by the region;

(b) scanning the allocation list of a unit represented by the region for an old allocation entry describing an old allocation having superseded data;

(c) determining if the superseded data is located at a beginning of the old allocation described by the old allocation entry, and if not, creating a new allocation entry for non-superseded data located at the beginning of the old allocation by adding the new allocation entry describing the new allocation to the allocation list of the unit, the new allocation entry including (A) the client start address of the old allocation;

(2) a new allocation length for the new allocation;

(3) the physical start address of the old allocation; and

(4) a status indicating that the non-superseded data is valid;

(d) determining if the superseded data is located at an end of the old allocation described by the old allocation entry, and if not, creating a new allocation entry for the data located at the end of the old allocation by adding the new allocation entry describing the new allocation to the allocation list of the unit, the new allocation entry including (A) a new client start address for the new allocation;

(2) a new allocation length for the new allocation;

(3) a new physical start address for the new allocation; and

(4) a status indicating that the non-superseded data is valid;

(e) updating the status of the old allocation entry to indicate that the allocation entry is no longer valid; and

(f) repeating (b)-(e) for each unit until all of the superseded data in the given client range that is located in the unit is discarded.

28. The apparatus of claim 27, wherein a plurality of regions are selected in which to discard superseded data using the given client range and wherein (a)-(f) are repeated for each region until all of the superseded data in the given client range is discarded from the partition.

29. A computer-readable medium having a computer-executable component for allocating storage in an auxiliary storage medium, wherein the computer-executable component allocates storage in the auxiliary storage medium by:

(a) establishing at least one partition of physical address space in the storage medium;

(b) dividing the partition into a plurality of units, each unit comprising a portion of physical address space in the partition; and

(c) defining a client address space by organizing the plurality of units into a plurality of regions, each region defining a portion of client address space which represents the portions of physical address space within the plurality of units into which the region is organized.

30. The computer-readable medium of claim 29, wherein the computer-executable component further allocates storage in the auxiliary storage medium by organizing each unit into a plurality of allocations, each allocation including at least one block of physical data storage locations in the unit.

31. The computer-readable medium of claim 30, wherein the computer-executable component uses an allocation list for each unit of the partition to interface the client address space defined by the regions to the physical address space of the partition, the allocation list including at least one allocation entry describing each allocation in the unit, each allocation entry including:

(i) a client start address for the allocation in the portion of client address space relative to the region in which the unit is organized;

(ii) an allocation length indicating the number of blocks in the allocation;

(iii) a status of the allocation; and

(iv) a physical start address for the allocation in the unit.

32. The computer-readable medium of claim 31, wherein the computer-executable component uses the allocation list for each unit of the partition write data to the partition at a given client range, the given client range including a given client start address in the client address space and a given allocation length for the data, wherein the computer-executable component uses the allocation list to write data by:

(a) determining a region of client address space representing the units in which to write the data using the given client range;

(b) determining if there is available physical address space in the units represented by the region to write the data;

(c) writing the data to the units represented by the region having available physical address space, wherein writing the data creates a new allocation in at least one unit represented by the region, the new allocation having a physical start address in the unit, an allocation length and a status indicating that data is currently being written to the new allocation; and

(d) adding a new allocation entry to the allocation list that describes the new allocation, the new allocation entry including the given client start address, the allocation length of the new allocation, the status of the new allocation, and the physical start address of the new allocation.

33. The computer-readable medium of claim 32, wherein writing data to the units represented by the region further comprises:

(a) scanning the allocation list of a unit represented by the region to determine a first available physical start address in the unit to which data can be written;

(b) creating the new allocation in the unit by writing at least a portion of the data to the unit at the first available physical start address; and

(c) repeating (a)-(b) for each unit represented by the region until the data for the region has been written.

34. The computer-readable medium of claim 33, wherein the computer-executable component further uses the allocation list to write data by selecting a plurality of regions representing units in which to write the data using the given client range and wherein (a)-(c) are repeated for each region until all of the data has been written to the partition.

35. The computer-readable medium of claim 31, wherein the computer-executable component further uses the allocation list to read data from the partition at a given client range, the given client range including a given client start address in the client address space and a given allocation length for the data, wherein the computer-executable component uses the allocation list to read data by:

(a) determining a region of client address space representing the units from which to read the data using the given client range;

(b) scanning the allocation list of a unit of the region to determine a physical start address and a length for at least a portion of the data to be read;

(c) reading the portion of data from the unit at the physical start address and length; and

(d) repeating (b)-(c) until as much data as possible has been read from the region.

36. The computer-readable medium of claim 35, wherein scanning the allocation list of the unit further comprises:

(a) locating an allocation entry in the allocation list of the unit having a client range comprising a client start address and an allocation length that overlaps the given client range; and

(b) determining the physical start address and length for the portion of data using the overlapping given client range and the client range of the allocation entry.

37. The computer-readable medium of claim 36, wherein the computer-executable component further uses the allocation list to read data by selecting a plurality of regions from which to read the data using the given client range and wherein (a)-(b) are repeated for each region until all