Magnetic random access memory

Abstract

The invention provides a magnetic random access memory which can prolong the service life. The magnetic random access memory has an active area and a boot area, and has a reliability determination area to which new data different from old data stored in the reliability determination area is written when data is written in the active area, and a reliability determination area to which new data different from old data stored in the reliability determination area is written when data is written in the boot area. The data stored in the reliability determination area is read, and when the read data is different from the new data, the operation area and the other area are exchanged.

Description

Technical Field

[0001] The present invention relates to an MRAM (Magnetoresistive Random Access Memory). Background Art

[0002] Conventionally, a technique has been known in which the remaining service life of an MRAM is calculated based on statistical information, and writing to the MRAM is prohibited when the service life approaches. (For example, Patent Document 1).

[0003] Patent Document 1: Japanese Unexamined Patent Application Publication No. 2014-167809

[0004] In the conventional technique, writing cannot be performed on the MRAM when the service life approaches. Summary of the Invention

[0005] The present invention has been made in view of the above problems, and an object thereof is to provide a technique for extending the service life of an MRAM.

[0006] To achieve the above object, the present invention provides a magnetoresistive random access memory having a working area and a startup area, the working area having a reliability determination area. When data is written to the working area, the data is written to an area other than the reliability determination area within the working area, and new data different from the old data stored in the reliability determination area is written to the reliability determination area. The data stored in the reliability determination area is read, and when the read data is different from the new data, the working area and another area are swapped.

[0007] That is, in the MRAM, when any data is written to the working area, the data stored in the reliability determination area is rewritten. Therefore, if any area within the working area is compared with the reliability determination area, the reliability determination area has the highest usage frequency and reaches the service life earliest. It is possible to determine whether the service life has been reached by reading the data stored in the reliability determination area and determining whether the read data is consistent with the written data. Therefore, if the reliability determination area reaches the service life, by swapping the working area with another area, the working area can be used for a longer period. As a result, the service life of the MRAM can be extended. Brief Description of the Drawings

[0008] Figure 1 is a block diagram showing the structure of an IC chip.

[0009] Figure 2 is a flowchart of the service life management process.

[0010] Figure 3 It is a block diagram showing the structure of an IC chip in which the working area and the startup area are swapped.

[0011] Figure 4 It is a block diagram showing the structure of an IC chip.

[0012] Figure 5 It is a block diagram showing the structure of an IC chip in which the working area and the unused area are swapped.

[0013] Explanation of reference numerals:

[0014] 10 Processor, 21 Startup area, 22 Working area, 22a Area for reliability determination, 23 Unused area, 24 Non - used area. Detailed implementation manners

[0015] Here, the implementation manners of the present invention will be described in the following order.

[0016] (1) Structure of the IC chip

[0017] (2) Expiration date management process

[0018] (3) Other implementation manners, etc.

[0019] (1) Structure of the IC chip:

[0020] Figure 1 It is a block diagram showing the structure of the IC chip 1 including the MRAM (Magnetoresistive Random Access Memory) 20 related to the present invention. The IC chip 1 is a device provided as a package in which various circuits are formed, and is mounted on a computer included in a general - purpose computer, an in - vehicle device, etc. The IC chip 1 has a processor 10 and an MRAM 20, and the processor 10 realizes various functions by executing a program recorded in the MRAM 20. The processor 10 can be realized by various known circuits. For example, it may also have a storage device such as an SRAM (Static Random Access Memory).

[0021] Regions are pre - distinguished in the MRAM20. That is, the MRAM20 has a boot area 21 and a working area 22. In this embodiment, the boot area 21 is the storage area of the operating system. In this embodiment, when power is supplied to the IC chip 1, the processor 10 reads the operating system stored in the boot area 21 and executes the boot sequence. After the boot sequence ends, the processor 10 executes various processes under the execution of the operating system. In addition, in the processing of the operating system, it also includes the processing of application programs executed under the execution of the operating system.

[0022] The working area 22 is the storage area of the data processed in the processing of the operating system. That is, when data is stored in various processes under the execution of the operating system, the data is written into the working area 22.

[0023] In addition, in this embodiment, the working area 22 includes a reliability determination area 22a. In this embodiment, the reliability determination area 22a is an area with a capacity of 1 bit (1bit). Through the processing of the operating system, when data is written into the working area 22, new data different from the old data stored in the reliability determination area 22a is written into the reliability determination area 22a.

[0024] Specifically, when any data is written into an area other than the reliability determination area 22a of the working area 22, the reliability determination area 22a is rewritten. That is, when the old data stored in the reliability determination area 22a is 0, the new data 1 is written into the reliability determination area 22a. When the old data stored in the reliability determination area 22a is 1, the new data 0 is written into the reliability determination area 22a.

[0025] When the above - mentioned rewriting is performed, each time any data is written into the working area 22, the data in the reliability determination area 22a is rewritten. Therefore, the number of rewrites in the reliability determination area 22a is greater than or equal to the number of rewrites in any area within the working area 22. Therefore, the number of rewrites in the reliability determination area 22a is the maximum value of the number of rewrites in the working area 22.

[0026] Therefore, the reliability determination area 22a reaches the end - of - life first within the working area 22. Therefore, when the reliability determination area 22a reaches the end - of - life, if the working area 22 is swapped with other areas that have not reached the end - of - life, the service life of the working area 22 can be extended.

[0027] Therefore, in this embodiment, after new data is written into the reliability determination area 22a, a read is performed. If the read - out data is different from the new data, the working area 22 is swapped with other areas.

[0028] In addition, in the present embodiment, the other area is the startup area 21. That is, writing to the startup area 21 is performed when writing the operating system, but in most cases, no writing or almost no writing is performed thereafter. Therefore, generally, the number of write operations in the startup area 21 is significantly less than that in the working area 22. Even if the operating system is rewritten or corrected, generally, the number of write operations in the startup area 21 is less than that in the working area 22.

[0029] Therefore, when the data read from the reliability determination area 22a is different from the written data, by swapping the working area 22 and the startup area 21, the effective usable period of the MRAM 20 can be extended.

[0030] (2) Lifetime management process:

[0031] Next, the lifetime management process executed in the IC chip 1 will be described. When power is supplied to the IC chip 1, the processor 10 reads the operating system with reference to the startup area 21 (step S100). That is, the data of the operating system stored in the startup area 21 is loaded into the memory in the processor 10, and the execution of the operating system is started. In this state, the processor 10 can execute any process that can be executed under the execution of the operating system.

[0032] The processor 10 executes any process at an arbitrary timing and under the execution of the operating system (step S105). When executing any process, the processor 10 determines whether data is written to the working area 22 through this process (step S110). In step S110, when it is not determined that data is written to the working area 22, the processor 10 repeats the processes after step S105.

[0033] In step S110, when it is determined that data is written to the working area 22, the processor 10 rewrites the reliability determination area 22a (step S115). That is, the old data stored in the reliability determination area 22a is overwritten with new data. In addition, the processor 10 writes data to the working area 22 (step S120). That is, the data to be written through the process of step S105 is written to the area other than the reliability determination area 22a in the working area 22.

[0034] Next, the processor 10 reads the data in the reliability determination area 22a (step S125) and determines whether the data is normal (step S130). That is, when the read data is consistent with the data after being rewritten in step S115, it is determined that the data is normal. When the data is inconsistent, it is not determined that the data is normal.

[0035] In step S130, when it is determined that the data is normal, the processor 10 repeats the processing after step S105. In step S130, when it is not determined that the data is normal, the processor 10 swaps the working area 22 and the startup area 21 (step S135). That is, the processor 10 transfers the data of the operating system stored in the startup area 21 to an area other than the reliability determination area 22a in the working area 22. In addition, this swapping can be performed by various methods, and it can also be performed while backing up the data to an idle area of the MRAM 20, a backup area in an external memory of the MRAM 20, etc.

[0036] Then, the processor 10 executes the processing after step S100 and starts again from the reading of the operating system. That is, the IC chip 1 is restarted. However, after the restart, the area that was the startup area 21 before the restart is used as the working area 22, and the area that was the working area 22 before the restart is used as the startup area 21. In addition, if the processing in step S105 can be interrupted, the data of both the startup area 21 and the working area 22 can be backed up to other areas, etc., and the data of the startup area 21 and the data of the working area 22 can be swapped, and the processing in step S105 can be started again after the swapping is completed, then the restart can also be omitted.

[0037] Figure 3 Indicates Figure 1 The state after swapping the MRAM 20 shown. Since the area that was the startup area 21 before the restart becomes the working area 22, a part of it is used as the reliability determination area 22a. In addition, the area that was the working area 22 before the restart becomes the startup area 21. However, since the area 21a that was used as the reliability determination area 22a before the restart has reached the usage limit of the rewrite count, it is excluded from the startup area 21 and not used.

[0038] According to the above structure, after using the working area 22 until the usage limit of the reliability determination area 22a, the working area 22 can be further used until the usage limit of another reliability determination area 22a. As a result, compared with the case where no swapping is performed, the usage limit of the MRAM 20 can be extended to about twice.

[0039] (3) Other embodiments, etc.:

[0040] The above embodiments are examples for implementing the present invention, and various other embodiments can also be adopted. For example, the usage method of the MRAM can be various, and it can also be configured not as a chip integrated with a circuit for implementing functions other than the MRAM, such as a processor, but as a single-chip MRAM. In addition, a chip having an MRAM may also include a circuit other than a processor.

[0041] In addition, the area swapped with the working area is not limited to the startup area. For example, it can also be a structure in which the working area and the unused area of the MRAM are swapped. Figure 4 It is a block diagram showing a structure in which an unused area 23 is ensured outside the startup area 21 and the working area 22 after the manufacture of the IC chip 1. In Figure 4 the example shown, an unused area 23 with the same capacity as the working area 22 is ensured in advance.

[0042] In this structure, the expiration date of the MRAM is also managed through the Figure 2 processing shown. However, in step S135, a swap between the working area 22 and the unused area 23 is performed. Specifically, in step S135, the processor 10 transfers the data other than the reliability determination area 22a stored in the working area 22 to the unused area 23. When the transfer is performed, the area that was the unused area 23 becomes the working area 22 as shown in Figure 5 . In addition, the processor 10 ensures a reliability determination area 22a within the new working area 22. And the area that was the working area 22 before the transfer is set as the unused area 24. The processor 10 does not use the unused area 24 hereafter.

[0043] According to the above structure, after using the working area 22 until the expiration date of the reliability determination area 22a, the working area 22 can be further used until the expiration date of another reliability determination area 22a. As a result, compared with the case where no swap is performed, the expiration date of the MRAM 20 can be extended to approximately twice. In addition, since the startup area 21 is not swapped, even if the area used as the working area 22 before the swap reaches the expiration date, the data in the startup area 21 will not be affected.

[0044] The MRAM is a magnetic random access memory, and any memory that stores information through a change in the magnetization state can be used. That is, in the MRAM, compared with other known memories, it has the characteristics that the data retention guarantee period is relatively long, but the guaranteed number of data rewrites is relatively small. In this way, even if the characteristic of the MRAM is that the guaranteed number of rewrites is relatively small, the expiration date can be extended by swapping the working area.

[0045] The startup area is used at startup, so it is used, for example, when starting a system that uses the MRAM, etc., but is not frequently accessed after startup. On the other hand, the working area is used after startup and is more frequently accessed than the startup area.

[0046] The area for reliability determination only needs to be an area where data is rewritten each time data is written to the working area. That is, as long as new data different from the old data stored in the reliability determination area is written to the reliability determination area each time data is written to the working area. In addition, in the above-described embodiment, the reliability determination area is a part of the area ensured within the working area, but the reliability determination area can also be ensured in an area outside the working area. In this case, when swapping the working area and other areas, the reliability determination area is newly ensured from the unused area.

[0047] When data is written to the working area, new data different from the old data stored in the reliability determination area is written to the reliability determination area. That is, in the reliability determination area, the stored old data is overwritten with different new data. By performing such rewriting, as long as the number of rewrites in the reliability determination area can be made equal to or greater than the maximum number of rewrites in the working area.

[0048] The old data and the new data only need to be different. As in the above-described embodiment, if the reliability determination area is 1 bit, either the old data or the new data is 1 and the other is 0. The reliability determination area can also be multiple bits. In this case, when rewriting the reliability determination area, it is preferable to rewrite all bits with different data.

[0049] In addition, the method of the present invention can also be applied as a program or a method. In addition, the above-described system, program, and method can be implemented either as a single device or as multiple devices, including various modes. In addition, changes such as part being software and part being hardware can be appropriately made. In addition, the invention is also established as a recording medium for a program of a control system. Of course, the recording medium for this program can be a magnetic recording medium or a semiconductor memory, and can be considered in exactly the same way for any recording medium developed in the future.

Claims

1. A magnetic random access memory having a working area and a start-up area, wherein: The magnetic random access memory has a reliability determination area. When data is written into the working area, new data different from old data stored in the reliability determination area is written into the reliability determination area. The data stored in the reliability determination area is read, and when the read data is different from the new data, the working area and the other area are exchanged.

2. The magnetic random access memory according to claim 1, wherein: The other area is the startup area.

3. The magnetic random access memory according to claim 1, wherein: The other area is an unused area of ​​the magnetic random access memory.

4. The magnetic random access memory according to claim 1, wherein: The boot area is a storage area of ​​the operating system. The work area is a storage area for data processed in the processing of the operating system.

Images