Memory system, controller and control method of memory

Patent metadata

Field	Value
Publication number	US-9304906-B2
Application number	US-201414200625-A
Country	US
Kind code	B2
Filing date	Mar 7, 2014
Priority date	Sep 10, 2013
Publication date	Apr 5, 2016
Grant date	Apr 5, 2016

Abstract

According to one embodiment, a memory system includes non-volatile memory, a block management table that stores whether data in the non-volatile memory is valid or invalid in a unit of cluster, and a controller configured to execute compaction. In the block management table, first information related to likelihood that valid data within the block is invalidated is registered for each of the blocks. The controller is configured to select a block to be a target of the compaction based on the first information and use the selected block to execute the compaction.

First claim

What is claimed is:

1. A memory system, comprising:
non-volatile memory that includes a plurality of blocks as a unit of data erase, each one of the plurality of blocks including a plurality of pages as a unit of data write and read;
a block management table that stores whether data in the non-volatile memory is valid or invalid in a unit of cluster, the cluster being natural number times as large as a unit of sector, the natural number being equal to or more than two, the unit of sector being a minimum access unit from a host, a size of the cluster being smaller than a size of the block; and
a controller configured to:
manage data stored in the non-volatile memory based on the block management table; and
execute compaction based on the block management table, the compaction being a process that reads out a valid cluster from the block and rewrites the valid cluster to an erased block, wherein

in the block management table, for each of the blocks, first information related to the likelihood that valid data within the block is invalidated is registered, and
the controller is configured to select a block to be a target of the compaction based on the first information and use the selected block to execute the compaction.

2. The memory system according to claim 1 , wherein
in the block management table, for each of the blocks, an immediately-previous invalid cluster count is registered as the first information, the immediately-previous invalid cluster count being a number of invalid clusters that become invalid within a predetermined time period immediately prior to the compaction in the block among the clusters, and
the controller is configured to select the block to be the target of compaction based on the immediately-previous invalid cluster count.

3. The memory system according to claim 2 , wherein
in the block management table, for each of the blocks, a valid cluster count is further registered, the valid cluster count being a number of clusters in which valid data is stored, and
the controller is configured to select the block to be the target of compaction based on both the valid cluster count and the immediately-previous invalid cluster count.

4. The memory system according to claim 1 , wherein
in the block management table, for each of the blocks, a short-term cluster count is registered as the first information, the short-term cluster count being a number of clusters in which a time period until validation time has expired is shorter than a predetermined time, and
the controller is configured to select the block to be the target of compaction based on the short-term cluster count.

5. The memory system according to claim 4 , wherein
in the block management table, for each of the blocks, a valid cluster count is further registered, the valid cluster count being a number of clusters in which valid data is stored within the block among the clusters, and
the controller is configured to select the block to be the target of compaction based on both the valid cluster count and the short-term cluster count.

6. The memory system according to claim 1 , wherein
in the block management table, for each of the blocks, a number of clusters that are included in the block and each have an elapsed time after invalidation within a predetermined time period is registered as the first information.

7. The memory system according to claim 2 , wherein
the controller includes a block manager configured to update the immediately-previous invalid cluster count on the block management table at a time of data transfer between the non-volatile memory and the host,
in the block management table, for each of the blocks, a valid cluster count is further registered, the valid cluster count being a number of clusters in which valid data is stored within the block, and
the block manager is configured to register a latest immediately-previous invalid cluster count in a latest block management table based on a comparison result between a valid cluster count registered on the block management table at compaction a predetermined times before and a valid cluster count registered in the latest block management table.

8. The memory system according to claim 7 , wherein
the block manager is configured to update the block management table in a case where data is invalidated due to overwriting of data corresponding to an LBA or in a case where data within the block is invalidated due to a Trim command.

9. The memory system according to claim 2 , wherein
the controller is configured to set a first block as the target of compaction, a number of the immediately-previous invalid clusters in the first block being smaller than a predetermined number.

10. The memory system according to claim 2 , wherein
the controller is configured to set a second block as the target of compaction, a max count of a consecutive number of the immediately-previous invalid clusters in the second block being smaller than a predetermined number.

11. A controller, comprising:
a manager configured to manage data stored in non-volatile memory based on a block management table, the block management table storing whether data in the non-volatile memory is valid or invalid in a unit of cluster, the cluster being natural number times as large as a unit of sector, the natural number being equal to or more than two, the unit of sector being a minimum access unit from a host, a size of the cluster being smaller than a size of the block, the block being a unit of data erase of the non-volatile memory; and
a compaction controller configured to execute compaction based on the block management table, the compaction being a process that reads out a valid cluster from the block and rewrites the valid cluster to an erased block, wherein
the non-volatile memory includes a plurality of blocks as a unit of data erase, each one of the plurality of blocks including a plurality of pages as a unit of data write and read, and
in the block management table, for each of the blocks, first information related to likelihood that valid data within the block is invalidated is registered, and
the compaction controller is configured to select a block to be a target of the compaction based on the first information and use the selected block to execute the compaction.

12. The controller according to claim 11 , wherein
in the block management table, for each of the blocks, an immediately-previous invalid cluster count is registered as the first information, the immediately-previous invalid cluster count being a number of invalid clusters that become invalid within a predetermined time period immediately prior to the compaction in the block among the clusters, and
the controller is configured to select the block to be the target of compaction based on the immediately-previous invalid cluster count.

13. The controller according to claim 11 , wherein
in the block management table, for each of the blocks, a valid cluster count is further registered, the valid cluster count being a number of clusters in which valid data is stored, and
the controller is configured to select the block to be the target of compaction based on both the valid cluster count and the immediately-previous invalid cluster count.

14. The controller according to claim 11 , wherein
in the block management table, for each of the blocks, a short-term cluster count is registered as the first information, the short-term cluster count being a number of clusters in which a time period until validation time has expired is shorter than a predetermined time, and
the controller is configured to select the block to be the target of compaction

Assignees

• Toshiba Kk

Inventors

• Matsudaira Hiroki
• Hirao Takashi
• Tran Aurelien Nam Phong

Classifications

• G06F12/0246 Primary

  in block erasable memory, e.g. flash memory · CPC title

• G06F13/1668

  Details of memory controller · CPC title

• G06F12/0292

  using tables or multilevel address translation means (G06F12/023 takes precedence; address translation in virtual memory systems G06F12/10) · CPC title

• G06F3/0688

  Non-volatile semiconductor memory arrays · CPC title

• G06F3/0628

  making use of a particular technique · CPC title