TWI550624B - Memory data control method - Google Patents

Description

Memory data management method

The present invention relates to a data management method, and more particularly to a related technique applied to a memory and extending its service life.

In recent years, consumer use of memory has become quite common, and the importance of memory for durability has increased. In the current memory (such as flash memory), when users use it for repeated reading. In the case of data, the number of errors in the data field often accumulates as the number of readings increases, when the number of error data fields in one of the pages of a memory block exceeds an error correction code ( ECC) When the maximum value of the error can be corrected, the page will no longer be used and the memory read/write life will be shortened rapidly.

For example, in China's invention patent No. I356419, "Method for Extending the Service Life of Multi-Level Cell Flash Memory", it mainly proposes a flash memory wrapping process to avoid excessive error of paging, which includes the following steps: a multi-level cell flash memory comprising at least one memory block; reading one of the data stored in the memory block; error correction of the data by an error correction code (ECC) unit; when one of the data is incorrect The number exceeds a preset value, erasing the data of the memory block; and rewriting the data to the memory block.

As can be seen from the above patent, the data error of the page of the memory block can be corrected by the error correction code, but the error correction code has a maximum value (preset value) of the correctable error, and when the number of data errors exceeds the representative value When the upper limit of the maximum value is used, the data of the memory block needs to be erased, and the data is rewritten to the memory block. However, in the prior art, when the memory block is erased, For the data, all the pages of the memory block must be erased, which is quite time consuming; or a block is specified first, then the data of the block is erased, and the data is rewritten into the block. It takes more time, and there is also a risk of destroying the original data when the erasure is performed, and the life of the memory is shortened; therefore, the above-mentioned prior art does have a need to propose a better solution.

In view of the above deficiencies of the prior art, the main object of the present invention is to provide a memory data management method, which can increase the memory space utilization rate and avoid the number of damages of the data field to be quickly accumulated to the limit to prolong the service life of the memory.

The main technical means to achieve the above purpose is to control the memory data control method, which is mainly connected by a controller to a memory, each page of the memory has a plurality of data fields and a plurality of reserved fields. And the controller performs the following steps: receiving a data and generating a system information, writing the data and the system information into the data field and the reserved field respectively; reading the system information to obtain Corruption information of a data field; corresponding to the address of a damaged data field to the address of a spare field in the reserved field according to the damage information; and updating the system information to record the damaged data field The address of the address and the address of the spare field in the reserved field and a corresponding address relationship thereof.

The above method mainly controls the memory by the controller, and each page of the memory has a plurality of data fields and reserved fields, wherein the reserved field is for writing system information, and the system information has a data field. The damage information, by reading the system information in the reserved field, to obtain the damage information of the data field, and obtaining the damaged address of the data field according to the content of the damaged information, when the damage of the data field continues If it rises, the address of the damaged data field is corresponding to the address of the spare field in the reserved field according to the damage information, and the system information is updated in time to evaluate the error correction capability of the memory, and the data column is avoided. The number of bits of damage is quickly accumulated to the limit, and most reserved fields are provided instead of the damaged data fields to achieve the purpose of extending the life of the memory.

A preferred embodiment of the memory data management method of the present invention, as shown in FIG. 1 , is mainly connected to a majority of memory (Block) by a controller (not shown), the memory having a majority of pages ( Page 0~N), wherein the pages have a plurality of data fields 10 and a plurality of reserved fields 20, wherein the reserved fields 20 are capable of storing a system information, an error correction code (ECC) information, A system management information or other information, and can be arbitrarily arranged by the designer to store the information to be stored in the reserved field 20; in this embodiment, a plurality of reserved fields 20 are used to replace a plurality of data fields. More than one damage data field 11 in 10.

It must be specially stated that the general data damage (error) can be corrected by using the error correction code (ECC), but the correction effect of the error correction code has a limit value. When the limit value is exceeded, the data damage cannot be repaired. The paging of the memory cannot be used. In order to extend the service life of the paging of the memory, a plurality of reserved fields 20 are used instead of the damaged data field 11 to increase the use of paging of the memory. Rate, it is better to avoid the number of damages in the data field 10 to quickly accumulate to the limit.

The controller is configured to control the memory, and in the embodiment, the controller is installed with a firmware, and the tough system is capable of recombining the memory when the memory is read or written into the memory. The data field 10 of the undamaged data field and the data field of the spare field 22 in the page of the memory, the reserved field 20 in the embodiment has a system information field 21 and a plurality of spare fields 22, The system information field 21 stores the above system information, the system information includes a damage information and an error correction code (ECC) information, and the spare fields 22 are used to replace the damage in the plurality of data fields 10, respectively. Data field 11.

When the controller receives more than one piece of data and correspondingly generates the system information, and writes the data to the data field 10 of the memory, and writes the system information into the reserved field 20 The system information field 21 obtains the damage information of the data field 10 by reading the system information in the system information field 21, and obtains the damaged address of the data field 10 according to the content of the damaged information, and Further, an error correction code (ECC) in the system information is read by the controller, and according to the content of the error correction code, the damaged address of the data field 10 and the number of damages are obtained.

To illustrate how to avoid the rapid accumulation of the damage of the data field 10 to the limit, please refer to FIG. 2, which includes a horizontal axis and a vertical axis, the horizontal axis represents the error correction capability of the memory, and the vertical axis represents a The number of damages, the number of damages of the memory in this embodiment may constitute a bit error rate. When the number of damages in the data field 10 of the memory continues to rise, the error rate also rises to one. The first preset value T1 is determined by the controller whether the number of damages in the damage information reaches the first preset value T1, and if so, the address of the damaged data field 11 is corresponding to the reservation according to the damage information. The address of one of the spare fields in the field 22 is replaced by the majority of the reserved field 22 instead of the damaged data field 11;

Then, the system information is updated in real time through the controller, and the address of the damaged data field 11 and the address of the spare field in the reserved field 22 and a corresponding address relationship thereof are recorded, and the number of damages is determined through the foregoing instant determination. Whether the first preset value T1 is reached, so as to prevent the damage quantity of the data field 10 from rapidly accumulating to a second preset value T2 representing a limit value, and extending the time to reach the second preset value T2 to increase the memory The error correction capability of the error correction code is used for the purpose of extending the service life of the memory; in this embodiment, the first preset value T1 and the second preset value T2 are respectively based on an error correction code of the controller. The error correction capability is set. When the first preset value T1 is 40 (50, 60, 70) and the second preset value is 80, the error correction capability of the controller's error correction code can be 100. It is between 40 and 80, between 50 and 80, between 60 and 80 or between 70 and 80.

To further illustrate the specific application mode of the memory data management method of the present invention, after the controller is connected to the memory, each page of the memory has a plurality of data fields 10 and a plurality of reserved fields 20, Referring to FIG. 3, the controller further performs the following steps: receiving a piece of data and generating a system information, and writing the piece of data and the system information into the data field 10 and the reserved field 20 respectively ( S31); reading the system information to obtain the damage information of the data field 10 (S32); according to the damage information, the address of the damaged data field 11 is corresponding to a spare field in the reserved field 20. Address (S33); and updating the system information to record the address of the damaged data field 11 and the address of the spare field in the reserved field 20 and a corresponding address relationship thereof (S34).

In this embodiment, when performing the foregoing steps, the controller can accept a read memory instruction, when the controller receives the read memory instruction, as shown in FIG. 4, and the controller performs a read. Step of taking the memory, the method further comprises the steps of: obtaining the damage information of the data field 10 and the corresponding address relationship according to the system information (S41); reading the uncorrupted data field in the memory page 10 Information and information on the alternate field (S42).

When performing the foregoing steps, the controller can further accept a write memory instruction. When the controller receives the write memory instruction, as shown in FIG. 5, and the controller executes a write memory. The method further includes the following steps: obtaining the damage information of the data field 10 and the address corresponding information according to the system information (S51); writing the data to the uncorrupted data field 10 of the memory page and The alternate field (S52).

In the above steps, the steps of "removing the damage information of the data field 10 and the address corresponding information (S41) (S51) according to the system information" are respectively performed, as shown in FIG. The following steps: reading an error correction code in the system information (S61); according to the content of the error correction code, obtaining the damaged address of the data field 10 and the number of damages (S62).

Instantly updating the system information through the controller, obtaining the damaged address and the damage quantity of the data field 10, and recording the address of the damaged data field 11 and the spare field in the reserved field 20. The address and its corresponding address relationship, when the above steps are performed to the step of "reading the system information to obtain the damage information of a data field 10 (S32)", as shown in FIG. 7, the method further includes the following steps: Determining whether a number of damages in the damage information reaches a preset value (S71); if the preset value is reached, performing "corresponding to the address of a damaged data field 11 according to the damage information to the reserved field Step 20: Address of a spare field (S33).

If the number of the damages reaches the preset value, the number of the damages of the data field 10 is quickly accumulated to the limit, so as to achieve the purpose of extending the service life of the memory. In this embodiment, the preset value includes a first pre-preparation. The value T1 and the second preset value T2 are set to determine whether the damaged quantity reaches the first preset value T1 through the foregoing instant, and if yes, the address of the damaged data field 11 is immediately mapped to the reserved field 20 for use. The address of the field 22 is to prevent the number of damages of the data field 10 from rapidly accumulating to a second preset value T2 representing a limit value, and it is possible to extend the time when the damage amount climbs to the second preset value T2 to actually increase the number of times. The error correction capability of the error correction code of the memory.

10 Data field 11 Damage data field 20 Reserved field 21 System information field 22 Reserved field

1 is a schematic diagram of a memory data space in accordance with a preferred embodiment of the present invention. 2 is a graph showing the extended service life of a preferred embodiment of the present invention. 3 is a flow chart of a memory data management method according to a preferred embodiment of the present invention. 4 is a flow chart of a read memory in accordance with a preferred embodiment of the present invention. Figure 5 is a flow chart of a write memory in accordance with a preferred embodiment of the present invention. 6 is a flow chart showing the number of corrupted addresses obtained in a preferred embodiment of the present invention. 7 is a flow chart of evaluating error correction capability in accordance with a preferred embodiment of the present invention.

10 Data field 11 Damage data field 20 Reserved field 21 System information field 22 Reserved field

Claims (8)

A memory data management method mainly includes a controller connected to a memory, each page of the memory has a plurality of data fields and a plurality of reserved fields, and the controller performs the following steps: receiving a And generating a system information, writing the data and the system information into the data field and the reserved field respectively; reading the system information to obtain the damage information of a data field; determining the damage information Whether the number of damages reaches a preset value; if the preset value is reached, the address of a damaged data field is mapped to the address of a spare field in the reserved field according to the damage information; The system information is updated to record the address of the damaged data field and the address of the spare field in the reserved field and a corresponding address relationship thereof. The memory data management method according to claim 1, further comprising the step of reading the memory: obtaining the damage information of the data field and the corresponding address relationship according to the system information; and reading the memory page Information on the undamaged data fields and information on the alternate fields. The memory data management method according to claim 2, the method further comprising: a step of writing to the memory: obtaining the damage information of the data field and the address corresponding information according to the system information; writing the data to the The undamaged data field in the memory page and the alternate field. The memory data management method according to claim 3, wherein when the foregoing step is performed to obtain the damage information of the data field and the address corresponding information according to the system information, the method further comprises the following steps: reading the system An error correction code in the information; According to the content of the error correction code, the damaged address of the data field and the number of damages are obtained. The memory data management method according to any one of claims 1 to 4, wherein, in the step of determining whether the damage amount in the damage information reaches a preset value, the preset value is based on an error of the controller The error correction capability of the correction code is set. The memory data management method according to claim 5, wherein the error correction capability of the error correction code of the controller is between 40 and 80 percent. The memory data management method of claim 6, wherein the reserved field has a system information field and the spare field, and the system information field stores the damage information and the error correction code, and the standby The field is used to replace the damaged data field. The memory data management method according to claim 7, wherein the controller includes a firmware for reorganizing the undamaged data field of the memory page when the step of reading or writing the memory is performed. The order of the data for this alternate field.