Direct data file storage in flash memories

• US 7,877,539 B2
• Filed: 02/16/2005
• Issued: 01/25/2011
• Est. Priority Date: 02/16/2005
• Status: Active Grant

• Current Assignee: SanDisk Technologies LLC (Western Digital Corporation)
• Original Assignee: SanDisk Corporation (Western Digital Corporation)
• Inventors: Sinclair, Alan W., Smith, Peter J.
• Primary Examiner(s): Chery; Mardochee
• Assistant Examiner(s): Chrzanowski; Matthew R
• Application Number:
  US11/060,249
• Publication Number:
  US 20060184720A1
• Time in Patent Office: 2,169 Days
• Field of Search: 711/103
• US Class Current: 711/103
• CPC Class Codes: G06F 12/0246 in block erasable memory, e.g. flash memory
• CPC Class Codes: G06F 2212/7201 Logical to physical mapping or translation of blocks or pages
• CPC Class Codes: G06F 2212/7205 Cleaning, compaction, garbage collection, erase control

Abstract

Host system data files are written directly to a large erase block flash memory system with a unique identification of each file and offsets of data within the file but without the use of any intermediate logical addresses or a virtual address space for the memory. Directory information of where the files are stored in the memory is maintained within the memory system by its controller, rather than by the host. The file based interface between the host and memory systems allows the memory system controller to utilize the data storage blocks within the memory with increased efficiency.

Claims

1. A method of transferring data between a host system and a re-programmable non-volatile semiconductor mass storage system having memory cells organized into blocks of memory cells that are erasable together and which individually store multiple units of data in multiple pages, an individual block being erased prior to new data being written therein, wherein:

a plurality of files containing varying amounts of data are generated by the host system,

the host system identifies the individual files of data generated by unique file identifiers and offsets of data within individual files, and sends data of the individual files and their said file identifiers and offsets to the mass storage system,

the mass storage system writes at least one unit at a time of individual ones of the files of data received from the host system into pages of at least one previously erased block of memory cells until the data received from the host system have been written, wherein an individual unit of data contains more than 512 bytes, and

the mass storage system directly translates the file identifiers and offsets received from the host into physical addresses of blocks and pages of memory cells in which data of the identified files are written without the use of any intermediate logical address conversion.

2. The method of claim 1, wherein the individual unit of data contains at least 1024 bytes.

3. The method of claim 2, wherein at least 8 units of data are written in parallel into individual blocks of the mass storage system.

4. The method of claim 1, wherein the memory cells in the blocks of file mass storage system are arranged in a plurality of rows and writing data includes writing at least two units of data per row within said at least one previously erased block of memory cells.

5. The method of claim 4, wherein writing data additionally includes writing the units of data to one row of the memory cells within said at least one previously erased block of memory cells at a time in sequence from one end of the block to another end thereof.

6. The method of claim 1, wherein the mass storage system translates file identifiers and offsets received from the host into physical addresses of blocks of memory cells by maintaining a directory of corresponding addresses of mass storage system blocks and the unique file identifiers and offsets.

7. The method of claim 1, wherein writing the received data includes concurrently writing at least one unit of the received host file data into each of at least two previously erased blocks of memory cells.

8. The method of claim 7, additionally comprising linking blocks together as metablocks into which the received host file data are written.

9. The method of claim 1, wherein the memory cells are organized according to a NAND architecture.

10. The method of claim 1, wherein the mass storage system is contained within an enclosure having external contacts adapted to connect with a host that generates the host data files.

11. The method of claim 1, wherein the mass storage system contains a number of memory cells sufficient to store at least 256 megabytes of data.

12. The method of claim 1, wherein the plurality of files are generated by the host system from application programs running on the host system.

13. The method of claim 1, additionally wherein

the host generates the data of a number of files opened at one time that is less than a preset number N allowed by mass storage system, and

the mass storage system writes data of the number of opened files received from the host into a corresponding number of opened blocks of memory cells that are individually dedicated to receive and store the data of only one of the number of files opened at one time.

14. The method of claim 13, additionally wherein the host generates a command to close one of the number of opened files, and the mass storage system responds to the command to close one of the number of opened files by garbage collecting at least one of the blocks dedicated to receive and store the data of the file to be closed.

15. The method of claim 14, wherein the mass storage system garbage collects at least one of the blocks dedicated to receive and store the data of the closed file if said at least one of the blocks is only partially filled with valid data of the file to be close.

16. The method of claim 1, additionally wherein the translation of the file identifiers and offsets received from the host into physical addresses of blocks and pages of memory cells in which data of the identified files are written is maintained in the mass storage system as a separate index for each file that contains at least one record of a group of data of a file having continuous offset addresses that is written into a specified block of memory cells, the record including an address of the data group within the specified block of memory cells and a corresponding offset address of the data of the file stored at said specified block address.

17. The method of claim 16, wherein the offset and block addresses are maintained with a granularity of one byte.

18. In a flash memory system having a memory with a data storage capacity of at least 256 megabytes and a controller of the memory, the memory being organized with storage elements thereof being grouped into blocks of elements that are erasable together, the controller operates according to a method that comprises:

receiving host generated data files of varying amounts of data to be stored in the memory from outside of the memory system, data of individual received data files being identified by a unique file identifier and offsets of data within the file,

identifying storage element blocks that are erased, storing one or more units of data of the received data files in the identified blocks at a time, the units of data individually containing more than 512 bytes of received data, and

maintaining within the memory, address translation records that directly map the blocks and storage elements within the blocks in which data of the individual data files are stored to the unique file identifier and offsets of the data within the file, without the use of any intermediate logical address conversion.

19. The method of claim 18, wherein the units of data individually contain at least 1024 bytes of received host file data.

20. The method of claim 18, wherein storing the received data files includes writing one unit at a time of the received host file data into each of at least two erased blocks of memory cells.

21. The method of claim 20, which additionally comprises logically linking blocks together as metablocks into which the received host file data are stored.

22. The method of claim 18, wherein the controller operating method occurs in the flash memory system contained within an enclosure having external contacts adapted to connect with a host that generates the host files.

23. A mass storage memory system, comprising:

re-programmable non-volatile memory cells formed on at least one semiconductor substrate, the memory cells being arranged in a plurality of blocks of memory cells that are erasable together and which are connected together in a plurality of series strings of memory cells that share conductive word lines extending across them to define rows of memory cells, the memory cells in individual rows storing at least 1024 bytes of data and the blocks individually containing at least eight rows of memory cells,

a controller including a microprocessor that is adapted to receive files of varying amounts of data that are individually identified by unique file identifiers and offsets of data within the file, and which causes a received data file to be stored in one or more blocks of memory cells, and

an address translation index maintained by the controller that identifies the blocks and rows of memory cells in which the received data are stored by a direct correspondence with unique file identifications and offsets of data within the files in which data of the identified files are written without the use of any intermediate logical address conversion.

24. The memory system of claim 23, wherein the controller links at least some of the blocks of memory cells together to form a plurality of metablocks that individually contain two or more blocks, and wherein the controller causes data of a received data file to be written in parallel into rows of the blocks of at least one metablock.

25. The memory system of claim 23, wherein the memory system is contained within a hand held enclosure having external contacts adapted to be removably connected with a host that generates the files of data.