# US 2022/0012180 A1 · Memory system for meta data management and operating method of memory system

> Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt

- Number: US 2022/0012180 A1
- Document ID: US20220012180A1
- Status: Published application
- Kind: A1
- Application number: 17/036,960
- Filed: 2020-09-29
- Published: 2022-01-13
- Earliest claimed: 2020-07-07
- Assignees: SK hynix Inc.
- Applicants: SK hynix Inc.
- Inventors: Ju Hyun KIM; Jin Yeong KIM
- Web page: https://patentgrep.com/patent/US20220012180A1

## CPC

- G06F 12/0238: {Memory management in non-volatile memory, e.g. resistive RAM or ferroelectric memory}
- G06F 12/0891: using clearing, invalidating or resetting means
- G06F 12/0868: Data transfer between cache memory and other subsystems, e.g. storage devices or host systems
- G06F 2212/1032: Reliability improvement, data loss prevention, degraded operation etc
- G06F 2212/1044: Space efficiency improvement
- G06F 2212/214: Solid state disk
- G06F 2212/7201: Logical to physical mapping or translation of blocks or pages
- G06F 2212/7203: Temporary buffering, e.g. using volatile buffer or dedicated buffer blocks
- G06F 2212/7204: Capacity control, e.g. partitioning, end-of-life degradation
- G06F 2212/7208: Multiple device management, e.g. distributing data over multiple flash devices
- Y02D 10/00: Energy efficient computing, e.g. low power processors, power management or thermal management
- G06F: ELECTRIC DIGITAL DATA PROCESSING (computer systems based on specific computational models G06N)
- Y02D: CLIMATE CHANGE MITIGATION TECHNOLOGIES IN INFORMATION AND COMMUNICATION TECHNOLOGIES [ICT], I.E. INFORMATION AND COMMUNICATION TECHNOLOGIES AIMING AT THE REDUCTION OF THEIR OWN ENERGY USE

## Abstract

A memory system comprising: a controller generates meta data in accordance with normal data being stored in a non-volatile memory device, and a buffer memory stores multiple meta slices configuring the meta data, the controller classifies an updated slice of the multiple meta slices as a first dirty slice, classifies a flushed slice of the first dirty slices as a second dirty slice, classifies a flushed slice of the second dirty slices as the meta slice and classifies an updated slice of the second dirty slices as a third dirty slice, classifies a flushed slice of the third dirty slices as the second dirty slice, and permits an update of each of the first to third dirty slices in a section in which a flush operation for each of the first to third dirty slices is performed.

## Claims

### Claim 1

A memory system comprising:
a non-volatile memory device;
a controller suitable for generating meta data in accordance with normal data being stored in the non-volatile memory device; and
a buffer memory suitable for storing multiple meta slices configuring the meta data,
wherein the controller
classifies each updated slice of the multiple meta slices as a first dirty slice using first state information of the corresponding updated slice,
classifies each flushed slice of the first dirty slices as a second dirty slice by using the first and second state information of the corresponding flushed slice,
classifies each flushed slice of the second dirty slices as the meta slice using the second state information of the corresponding flushed slice,
classifies each updated slice of the second dirty slices as a third dirty slice using the first state information of the updated slice,
classifies each flushed slice of the third dirty slices as the second dirty slice using the first state information of the flushed slice, and
enables update of each of the first to third dirty slices while flushing each of the first to third dirty slices,
wherein the first and second state information for each of the multiple meta slices is stored in the buffer memory.

### Claim 2 (depends on 1)

The memory system of claim 1, wherein the controller
generates first journal data comprising information on the update of the first to third dirty slices and stores the first journal data in the buffer memory,
flushes at least one of the first to third dirty slices when the at least one dirty slice is present in a state in which the first journal data has a size smaller than a set size,
flushes the first journal data by changing the first journal data into journal retention data when the first journal data has the set size, and
generates second journal data comprising update information on the first to third dirty slices after the flushing of the first journal retention data is started, and stores the second journal data in the buffer memory.

### Claim 3 (depends on 2)

The memory system of claim 2, wherein the second journal data is stored in a different location in the buffer memory than the location at which the journal retention data is stored in the buffer memory.

### Claim 4 (depends on 2)

The memory system of claim 2,
wherein the controller flushes at least one of the first to third dirty slices by writing at least one of the first to third dirty slices, and
wherein the controller flushes the journal retention data by writing the journal retention data in the non-volatile memory device.

### Claim 5 (depends on 2)

The memory system of claim 2, wherein the controller flushes at least one of the first to third dirty slices in a state in which the first journal data has a size smaller than a set size by:
moving the at least one of the first to third dirty slices to a reserved space of the buffer memory, and
writing, when flushing the journal retention data, the at least one of the first to third dirty slices moved to the reserved space together with the journal retention data in the non-volatile memory device.

### Claim 6 (depends on 5)

The memory system of claim 5, wherein the controller moves the at least one of the first to third dirty slices by:
splitting the at least one of the first to third dirty slices by a size unit, and
moving the split slices at set times.

### Claim 7 (depends on 6)

The memory system of claim 6, wherein the controller
checks a size of the first journal data at the time at which moving of the at least one of the first to third dirty slices to the reserved space is started, and
determines the size unit and the set times based on a result of the check.

### Claim 8 (depends on 2)

The memory system of claim 2,
wherein the controller:
searches the multiple meta slices for the first to third dirty slices in a round robin manner, and
flushes the first to third dirty slices retrieved in the search.

### Claim 9 (depends on 8)

The memory system of claim 8, wherein the controller searches for the first to third dirty slices from a next meta slice of a previously retrieved meta slice, when the second journal data is generated after the flushing of the journal retention data is started.

### Claim 10 (depends on 1)

The memory system of claim 1,
wherein the controller:
sets initial values of the first and second state information to “0” and “0”, respectively,
classifies each of the first dirty slices by setting a value of the first state information of the updated slice to “1”,
classifies each of the second dirty slices by setting values of the first and second state information of the flushed slice to “0” and “1”, respectively,
classifies each of the meta slices by setting a value of the second state information of the flushed slice to “0”,
classifies each of the third dirty slices by setting a value of the first state information of the updated slice to “1”, and
classifies each of the second dirty slices by setting a value of the first state information of the flushed slice to “0.”

### Claim 11

An operating method of a memory system comprising a non-volatile memory device and a buffer memory, the operating method comprising:
generating meta data configured with multiple meta slices in accordance with normal data being stored in the non-volatile memory device;
storing the meta data and the first and second state information in the buffer memory;
classifying, in a first classifying operation, each updated slice of the multiple meta slices as a first dirty slice using first state information of the corresponding updated slice;
classifying, in a second classifying operation, each flushed slice of the first dirty slices as a second dirty slice using the first information and second state information of the corresponding flushed slice;
classifying, in a third classifying operation, each flushed slice of the second dirty slices as the meta slice using the second state information of the corresponding flushed slice;
classifying, in a fourth classifying operation, each updated slice of the second dirty slices as a third dirty slice using the first state information of the corresponding updated slice; and
classifying, in a fifth classifying operation, each flushed slice of the third dirty slices as the second dirty slice using the first state information of the corresponding flushed slice,
wherein update of each of the first to third dirty slices is enabled while flushing each of the first to third dirty slices.

### Claim 12 (depends on 11)

The operating method of claim 11, further comprising:
generating first journal data comprising information of the update of the first to third dirty slices and storing the first journal data in the buffer memory;
flushing, in a first flush operation, at least one of the first to third dirty slices when the at least one dirty slice is present in a state in which the first journal data has a size smaller than a set size;
flushing, in a second flush operation, the first journal data by changing the first journal data to journal retention data when the first journal data has the set size; and
generating second journal data comprising update information corresponding to the first to third dirty slices after the flushing of the first journal retention data is started, and storing the second journal data in the buffer memory.

### Claim 13 (depends on 12)

The operating method of claim 12, wherein the second journal data is stored in a different location in the buffer memory than the location at which the journal retention data is stored in the buffer memory.

### Claim 14 (depends on 12)

The operating method of claim 12, wherein:
the first flush operation comprises writing any one of the first to third dirty slices in the non-volatile memory device, and
the second flush operation comprises writing the journal retention data in the non-volatile memory device.

### Claim 15 (depends on 12)

The operating method of claim 12, wherein the first flush operation comprises:
moving the at least one of the first to third dirty slices to a reserved space of the buffer memory, and
writing, during the second flush operation, the at least one of the first to third dirty slices moved to the reserved space together with the journal retention data in the non-volatile memory device.

### Claim 16 (depends on 15)

The operating method of claim 15, wherein the moving of the at least one of the first to third dirty slices includes:
splitting the at least one of the first to third dirty slices by a size unit, and
moving the split slices at set times.

### Claim 17 (depends on 16)

The operating method of claim 16, further comprising:
checking a size of the first journal data at the time at which moving of the at least one of the first to third dirty slices to the reserved space is started, and
determining the size unit and the timing based on a result of the checking.

### Claim 18 (depends on 12)

The operating method of claim 12,
further comprising searching the multiple meta slices for the first to third dirty slices in a round robin manner, and
wherein in the first flush operation is performed on the searched first to third dirty slices.

### Claim 19 (depends on 18)

The operating method of claim 18, wherein the first to third dirty slices are searched from a meta slice next to one among the first to third dirty slices that is previously searched when the second journal data is generated after the flushing of the journal retention data is started.

### Claim 20 (depends on 11)

The operating method of claim 11, further comprising:
setting initial values of the first and second state information to “0” and “0”, respectively, wherein
in the first classification operation, the first dirty slices are classified by setting a value of the first state information of the updated slices among the multiple meta slices to “1”,
in the second classification operation, the second dirty slices are classified by setting values of the first and second state information of the flushed slices among the first dirty slices to “0” and “1”, respectively,
in the third classification operation, the meta slices are classified by setting a value of the second state information of the flushed slices among the second dirty slices to “0”,
in the fourth classification operation, the third dirty slices are classified by setting a value of the first state information of the updated slices among the second dirty slices to “1”, and
in the fifth classification operation, the second dirty slices are classified by setting a value of the first state information of the flushed slices among the third dirty slices to “0.”
