WO2021208341A1 - Method and system for detecting and recovering memory bit flipping in secondary power equipment - Google Patents

Abstract

A method and system for detecting and recovering memory bit flipping in secondary power equipment, related to the technical field of memory error correction. The technical problem to be solved is equipment malfunctioning or result abnormality caused by memory bit flipping in secondary power equipment in the prior art. The method comprises the following steps: performing an ECC verification code calculation according to a preset ECC segment length with respect to an application running area when an application is loading, acquiring an ECC verification code for segment data of the application when the application is loading; performing an ECC verification code calculation according to the preset ECC segment length with respect to the segment data of the application when the application is running, acquiring an ECC check code for the segment data of the application when the application is running; comparing the ECC verification code with the ECC check code; and, if the segment data of the application when the application is running is determined, on the basis of the comparison result, to have experienced a single bit error, correcting the bit where the error occurred.

Description

Method and system for detecting and restoring memory bit flip of power secondary equipment Technical field

The invention relates to a method and system for detecting and restoring the memory bit flip of a power secondary device, and belongs to the technical field of memory error correction.

Background technique

The secondary devices of modern power systems are mostly embedded devices, composed of a large number of chips. One of the core components of the memory chip has millions of memory units, and each unit can store "0" or "1", etc. information. Since the memory is used to temporarily store the program and data being executed, once there is a data error inside it, it will affect the normal operation of the program, and can cause the entire system to fail in severe cases. Therefore, its reliability and fault tolerance have always been hot issues in the industry. .

The industry has studied semiconductor and memory technology for many years and found that the causes of memory errors are mainly divided into hard errors and soft errors. Among them, hard errors can occur repeatedly, mainly due to hardware damage to the memory unit in the memory chip or external wiring errors. Hard errors can generally only be solved by replacing the hardware. Soft errors occur randomly and can be recovered after the program is reloaded. A soft error is usually an unexpected rollover of 1-bit data (from 0->1 or 1->0), and the probability of occurrence is high.

Soft errors are usually caused by single event effects. Single Event Effects (SEEs) refer to radiation particles (heavy particles, protons, neutrons, X-rays, gamma rays and alpha particles, etc.) from cosmic radiation and ground radiation environment that cause integrated circuits and even electronic devices. The phenomenon of severe damage. When the radiant particles collide with the silicon material, direct or indirect ionization will generate additional radiation to induce electron-hole pairs, and the electric field in the reverse-biased depletion layer of the device can separate these electron-hole pairs and drift through them. The process effectively collects them, which results in the accumulation of extra charge in the sensitive areas of the device. When the accumulated charge is large enough, a large voltage transient pulse will be generated, which can temporarily reverse the voltage of the sensitive node of the circuit. In combinational circuits, such voltage transient pulses are called single event transients (SETs). SETs can propagate along the circuit to the storage unit, such as the storage unit of Static Random Access Memory (SRAM). If the appropriate conditions are met, SETs can cause the memory cell to capture incorrect timing information and cause bit changes.

Error Correcting Code (ECC) technology can be used to solve the problem of soft errors caused by 1-bit flipping in memory and Flash devices. This technology appeared after the "parity check" technology, which is a more advanced means of checking and correcting storage errors, and has been widely used in workstations and server products. ECC technology is to store a code for data encryption in an extra place on the data bit. When the data is written into the memory, the corresponding ECC code is also saved; when the data just stored is re-read, the saved ECC code It will be compared with the ECC code calculated in real time when the data is read. If the two codes are not the same, they will be decoded to determine which bit of the data is incorrect. Then this error bit will be discarded and the memory controller will release the correct data. If the same error data is read out again, the correction process will be executed again.

In recent years, some embedded processors and digital signal processor chips (DSP) have begun to add hardware ECC error detection and correction technologies, such as Ti’s multi-core ARM processor AM572x, Ti’s DSP chip C665X, and Xilinx’s latest Fully programmable SOC chip UltraScale MPSoC, etc. However, most bit flip detection and recovery solutions require the support of hardware units, and they all implement bit flip error protection from a hardware perspective. Because hardware ECC error detection and error correction requires the integration of hardware error detection and error correction control modules in the processor's memory controller, which leads to a significant increase in chip cost, it is currently widely used in processors and RAM devices in power secondary equipment. Both do not support the hardware ECC detection and recovery function, and the problem that the protection bit can not be flipped from the hardware level.

Summary of the invention

In view of the shortcomings of the prior art, the purpose of the present invention is to provide a method and system for detecting and restoring the memory bit flip of a power secondary device, so as to solve the abnormal function or result of the power secondary device caused by the memory bit flip in the prior art. Technical issues.

In order to solve the above technical problems, the technical solutions adopted by the present invention are:

A method for detecting and restoring the memory bit flip of a power secondary device includes the following steps:

Calculate the ECC check code of the application running area when the application is loaded according to the preset ECC segment length, and obtain the ECC check code of the segment data of the application when the application is loaded;

Perform ECC check code calculation on the segment data of the application program when the application program is running according to the preset ECC segment length, and obtain the ECC check code of the segment data of the application program when the application program is running;

Compare the ECC check code with the ECC check code;

If it is determined according to the comparison result that a single-bit error occurs in the segment data of the application program when the application program is running, correct the bit where the error occurred.

Further, after obtaining the ECC check code of the segment data of the application program when the application program is loaded, the method further includes:

Store the obtained ECC check code and the storage quantity is not less than three copies;

Compare the ECC check code with the ECC check code, including:

Compare the stored ECC check codes one by one;

If at least two comparison results of the stored ECC check codes are consistent, the ECC check code with the same comparison results is extracted and compared with the ECC check code.

Further, comparing the ECC check code with the ECC check code also includes: extracting the ECC check code with the same comparison result, and comparing and replacing the ECC check code with the inconsistent result.

Further, the comparison of the ECC check code with the ECC check code also includes:

If the stored ECC check codes are inconsistent with each other, the redundant backup recovery mechanism is used to restore the code segment of the application when the application is running.

Further, if it is determined according to the comparison result that there are no less than two bits in the segment data of the application program when the application program is running, the redundant backup recovery mechanism is used to restore the code segment of the application program when the application program is running.

Further, the use of a redundant backup and recovery mechanism to restore the code segment of the application when the application is running includes:

Obtain the code segment of the application program when the application program is loaded and calculate the CRC check code to obtain the CRC check code of the code segment;

Calculate the CRC check code of the code segment of the application program when the application is running, and obtain the CRC check code of the code segment;

Compare the CRC check code with the CRC check code;

If the comparison result between the CRC check code and the CRC check code is inconsistent, the code segment corresponding to the CRC check code is extracted, and the code segment of the application program is covered when the application program is running.

Further, after obtaining the code segment of the application program when the application program is loaded, the method further includes: compressing the obtained code segment;

Before extracting the code segment corresponding to the CRC check code, it also includes: decompressing the compressed code segment corresponding to the CRC check code.

Further, after obtaining the code segment of the application program and its CRC check code when the application program is loaded, it further includes:

Store the obtained code segment and its CRC check code and the number of storage is not less than two copies;

Extract the code segment corresponding to the CRC check code, including:

Online verification of the stored CRC check code;

If the online check result of the stored CRC check code is always consistent, extract the code segment corresponding to the CRC check code whose online check result is always consistent.

Further, extracting the code segment corresponding to the CRC check code also includes: extracting the CRC check code and its code segment whose online check result is always consistent, and the CRC check code and its code whose online check result is not always consistent Segment to replace.

Further, performing online verification on the stored CRC check code includes: comparing the current state of the stored CRC check code with the previous state at a preset time interval.

Further, it also includes: detecting and restoring the pre-registered key data in the application;

The key data detection and recovery method includes:

Extract the memory address of the key data based on the pre-registration information of the key data;

Obtain the memory data in the memory address when the application program is loaded and calculate the CRC check code to obtain the CRC check code;

Calculate the CRC check code for the memory data in the memory address of the key data when the application is running, and obtain the CRC check code;

Compare the CRC check code with the CRC check code;

If the comparison result between the CRC check code and the CRC check code is inconsistent, the memory data corresponding to the CRC check code is extracted, and the memory data in the memory address is overwritten when the application program is running.

Further, after obtaining the memory data and the CRC check code in the memory address when the application program is loaded, the method further includes: storing the obtained memory data and the CRC check code and the storage quantity is not less than two copies;

Extract the memory data corresponding to the CRC check code, including:

Online verification of the stored CRC check code;

If the online check result of the stored CRC check code is always consistent, extract the memory data corresponding to the CRC check code whose online check result is always consistent.

Further, extracting the memory data corresponding to the CRC check code also includes: extracting the CRC check code and its memory data whose online check results are always consistent, and the CRC check code and its memory whose online check results are not always consistent. Data row replacement.

To achieve the above objective, the present invention also provides a system for detecting and restoring the memory bit flip of a power secondary device, which includes: a single-bit error recovery module, the single-bit error recovery module includes:

ECC check code calculation sub-module: used to calculate the ECC check code of the application running area when the application is loaded according to the preset ECC segment length to obtain the ECC check code of the segment data of the application when the application is loaded, and Perform ECC check code calculation on the segment data of the application when the application is running according to the preset ECC segment length to obtain the ECC check code of the segment data of the application when the application is running;

ECC check code storage sub-module: After obtaining the ECC check code of the segment data of the application program when the application program is loaded, it is used to store the obtained ECC check code and the storage quantity is not less than three;

ECC check code comparison sub-module: used to compare the stored ECC check codes one by one. If at least two comparison results of the stored ECC check codes are consistent, extract the ECC check with the same comparison result The code is compared with the ECC check code;

ECC check code replacement sub-module: used to extract ECC check codes with consistent comparison results, and compare and replace ECC check codes with inconsistent results;

Single-bit error correction sub-module: If a single-bit error occurs in the segment data of the application program when the application program is running according to the comparison result, it is used to correct the bit where the error occurs.

Further, it also includes a redundancy backup and recovery module for recovering the code segment of the application program when the application program is running by using a redundancy backup and recovery mechanism, and the redundancy backup recovery module includes:

Code segment acquisition sub-module: used to acquire the code segment of the application when the application is loaded;

CRC check code calculation sub-module: used to calculate the CRC check code of the code segment of the application program when the application program is loaded to obtain the CRC check code of the code segment, and perform CRC check on the code segment of the application program when the application program is running Check code calculation to obtain the CRC check code of the code segment;

CRC check code storage sub-module: After obtaining the code segment and CRC check code of the application program when the application is loaded, it is used to store the obtained code segment and its CRC check code and the storage quantity is not less than two copies. ；

CRC check code online check sub-module: used to check the stored CRC check code online;

CRC check code comparison sub-module: used to compare CRC check code with CRC check code;

Code segment coverage sub-module: If the online check result of the stored CRC check code is always consistent, it is used to extract the code segment corresponding to the CRC check code whose online check result is always consistent. Segment to cover;

CRC check code replacement sub-module: used to extract the CRC check code and its code segment whose online check result is always consistent, and replace the CRC check code and its code segment whose online check result is not always consistent;

Code segment compression sub-module: used to compress the obtained code segment after obtaining the code segment of the application program when the application is loaded, and the code segment corresponding to the CRC check code before extracting the code segment corresponding to the CRC check code. The compressed code segment is decompressed.

Further, it also includes a critical data detection and recovery module, and the critical data detection and recovery module includes:

Memory address extraction sub-module: used to extract the memory address of the key data based on the pre-registered information of the key data;

Memory data acquisition sub-module: used to acquire the memory data in the memory address when the application program is loaded;

The CRC check code calculation sub-module is also used to perform CRC check code calculation on the memory data in the memory address when the application program is loaded to obtain the CRC check code, and to calculate the key data in the memory address when the application program is running CRC check code calculation is performed on the memory data to obtain the CRC check code;

The CRC check code storage sub-module is also used to store the obtained memory data and its CRC check code after obtaining the memory data and its CRC check code in the memory address when the application program is loaded. Not less than two copies;

The code segment coverage sub-module is also used to extract the memory data corresponding to the CRC check code whose online check result is always consistent if the online check result of the stored CRC check code is always consistent, and the application program is running. Overwrite the memory data in the memory address;

The CRC check code replacement sub-module is also used to extract the CRC check code whose online check result is always consistent and its memory data, and replace the CRC check code whose online check result is not always consistent and its memory data row.

To achieve the above objective, the present invention also provides a computer processing control device, including:

Memory: used to store instructions;

The processor is configured to operate according to the instructions to execute the steps of the method for detecting and restoring the memory bit flip of a power secondary device provided by the present invention.

To achieve the above objective, the present invention also provides a computer-readable storage medium on which a computer program is stored. When the program is executed by a processor, the method for detecting and restoring the memory bit flip of a power secondary device provided by the present invention is implemented. step.

Compared with the prior art, the present invention achieves the beneficial effects: the method and system of the present invention comprehensively use two mechanisms of ECC error detection and correction and redundant backup and recovery to realize the rapid positioning and recovery of single-bit errors and multi-bit errors, respectively. Error recovery, taking into account both the efficiency of error correction and the ability of error correction, makes up for the insufficiency of simply using ECC hardware error detection and error correction that cannot correct more than two errors, which is beneficial to ensure the reliable operation of power secondary equipment. Since there is no need to detect and restore circuits inside the processor, it is not affected by changes in the processor's own process design, and achieves the goal of using the existing architecture and saving development costs. At the same time, it can be extended to processors of different architectures and has a wide range of applicability.

Description of the drawings

Figure 1 is a schematic diagram of the address allocation of the on-chip memory of the microcontroller in the system embodiment of the present invention;

2 is a schematic diagram of the backup process of the code segment of the application program in the system embodiment of the present invention;

Figure 3 is a schematic diagram of a backup process of key data in an embodiment of the system of the present invention;

FIG. 4 is a schematic diagram of the recovery process of the code segment of the application program in the system embodiment of the present invention;

Fig. 5 is a schematic diagram of the restoration process of key data in the system embodiment of the present invention.

Detailed ways

The present invention will be further described below in conjunction with the accompanying drawings. The following embodiments are only used to illustrate the technical solutions of the present invention more clearly, and cannot be used to limit the protection scope of the present invention.

The program segment data generally has continuous addresses, which meets the operating requirements of the ECC error detection and correction algorithm. Therefore, the software ECC error detection and correction algorithm can be used to quickly locate and restore the single bit flip error of the program segment data; also taking into account the occurrence of system memory The possibility of multi-bit data errors, in order to improve the error correction capability of the error detection and correction program, a redundant backup recovery mechanism is added on the basis of the ECC error detection and correction, and the high efficiency and redundancy of the integrated ECC error detection and correction mechanism The strong error correction capability of the backup and recovery mechanism can complement each other to further ensure the correctness of the program segment data, thereby ensuring the reliability of the system. Generally, the addresses of key data are not continuous, it is difficult to apply ECC to detect and correct errors, and the data volume of key data is generally not large. Simply using redundant backup and recovery mechanisms can meet the needs of most applications. Therefore, only redundant data is used for key data. I backup and restore an error detection and correction mechanism.

Based on the above technical ideas, specific embodiments of the present invention provide a method for detecting and restoring a memory bit flip of a power secondary device, which includes the following steps:

Step 1: Load the application program with the error detection and correction program, and calculate the ECC check code in the application program operating area according to the length of the ECC segment (the segment length is confirmed by the digits of the ECC code), in order to prevent the ECC check code from running When an error occurs during the process, three copies of the ECC check code are stored. The ECC check codes of all segments are stored in the memory reserved area in three copies for each segment according to the sequence of the segments.

Step 2: When the application program is loaded, the error detection and correction program separates the code segment of the application program according to different file types, compresses the code segment data and calculates the CRC check code on it to obtain the CRC check code, which will be compressed The code segment data and CRC check code should be backed up at least in two copies, stored in the system memory or external storage. The compressed data is mainly to save the memory overhead of redundant backup.

Step 3: After the application is started, the number of segments that need to be ECC checked for a single execution of the error detection and correction program, and the processing mode after an unrecoverable error occurs, are set to the error detection and correction program. The application program registers the key data that needs to be protected with the error detection and correction program through the interface provided by the error detection and correction program. The error detection and correction program obtains the memory address of these key data, backs up the memory data in these memory addresses, press Ha The hash table is serialized and stored sequentially. The serialized whole piece of content is stored in more than two copies, and the CRC check code is calculated separately for each piece of content to obtain the CRC check code.

Step 4: After the system is running normally, the error detection and correction program periodically performs ECC error detection and error correction on the original program segment data according to the segment length, and calculates the ECC check code for the program data of the segment length each time to obtain the ECC check code, and Use three pre-stored ECC check codes to comprehensively judge the ECC check codes. First, compare the three ECC check codes with each other. If two ECC check codes are the same, the ECC check code is considered correct; if there is an error in the ECC check code among the three pre-stored ECC check codes If the code is verified, the error ECC check code is restored to the correct value. Then, the correct ECC check code is compared with the ECC check code. If the comparison results are consistent, it is determined that no bit error has occurred in the segment of data, otherwise it is determined that a bit error has occurred in the segment of data. If a single-bit error occurs in the segment data of the application program when it is judged to run, the correct ECC check code will be used directly for error correction and recovery; in order to reduce the impact of error detection and correction on the system load, the error detection and correction program will always Only one piece of length data is detected during the second run, and the next piece of data is detected during the next run.

Step 5: If more than two bit inversion errors or three ECC check codes in the segment data of the application program are judged to run, the redundant comparison recovery process will be entered. First, perform online verification on the CRC check code backed up in step 2, that is, compare the current state of the CRC check code with the previous state at a preset time interval. If the CRC check code comparison result is always consistent , The CRC check code is determined to be correct, otherwise the CRC check code is determined to be wrong. For the wrong CRC check code and its corresponding code segment data backup, the correct CRC check code and its corresponding code segment data can be used Back up to replace it in time. Then, for two or more bit flip errors, use the code segment corresponding to the correct CRC check code to cover the code segment of the running application; for the case where all three ECC check codes are wrong, the runtime application CRC check code is calculated for the code segment of the code segment, and the CRC check code of the code segment is obtained, and then the CRC check code is compared with the correct CRC check code. If the comparison results are consistent, it is determined that no bit has occurred in the running application. Turnover error, if the comparison results are inconsistent, it is determined that a bit turning error occurs in the running application, and the code segment corresponding to the correct CRC check code is used to cover the code segment of the running application. When covering the code segment of the running application, the error detection and correction program calls the system interface to turn off the system program preemption function or puts the running program into a hibernation state, and then decompress the correct backup version and restore it to the corresponding memory. And restore the correct backup version, backup check code and ECC check code to the original wrong backup version, and finally enable the system program preemption function or restore program operation.

Step 6. After the system is running normally, the error checking and correcting program periodically checks the correctness of the registered key data, obtains the memory address of the key data by querying the hash table, and compares the contents of the memory address with the backup version. Comparison, that is, the CRC check code is calculated on the content in the memory address to obtain the CRC check code, and the CRC check code is compared with the correct CRC check code backed up in step 3. Because the CRC check code can be checked online Check, and replace the wrong CRC check code and its corresponding memory data with the correct CRC check code and its corresponding memory data at any time, thus ensuring the correctness of the CRC check code. If the CRC check code is found to be inconsistent with the correct CRC check code, the memory data corresponding to the correct CRC check code is extracted, and the memory data in the corresponding memory address is overwritten when the application program is running. During the overwriting operation, the error detection and correction program calls the system interface to close the system program preemption function or make the running program enter the dormant state, and then restore the correct backup version to the corresponding memory address, and at the same time, the correct backup check code Restore to the backup version where the error occurred, and finally enable the system program preemption function or restore the program operation.

In the method embodiment of the present invention, the error detection and correction program always runs in the background of the system, and the error detection and correction process of program segment data and key data is continuously performed. When performing error recovery operations, a sandbox-like mechanism is used to prevent the error recovery process from being abnormally interrupted or the execution of abnormal codes, causing the entire system to enter an abnormal state, thereby ensuring the correctness of the error recovery process and results. After the error detection and correction program detects and repairs the error, it informs the application program of the address of the error, the type of error, and the current total number of errors through the interface. When an unrecoverable error occurs, it can be handled according to the pre-defined behavior of the application Abnormalities, including but not limited to reloading applications, system exiting, restarting the system through the watchdog, etc.

The specific embodiment of the present invention provides a system for detecting and restoring the memory bit flip of a power secondary device, which is used to implement the aforementioned method of the invention. Code calculation, ECC error detection and correction process, backup and recovery of data content, and processing of unrecoverable errors and other functions. The number of segments checked by each ECC of the error detection and correction program can be flexibly configured by the application program, so that the system of the present invention can be conveniently applied to systems with different processor load levels, and the error detection is determined according to the actual system processor load level The frequency of error correction keeps the overall load of the system at a reasonable level.

The system of the present invention is based on a microcontroller-free operating system environment. Because there is no intervention of the cache memory (Cache) and the memory management unit (MMU), the logical address is consistent with the physical address, and the content obtained by the processor accessing a certain address is the physical The internal real version of the memory, rather than the temporary version stored in the Cache, is easier to implement the method of the present invention. In this embodiment, taking a microcontroller-based non-operating system environment as an example, the method of the present invention is implemented in the on-chip memory of the microcontroller to realize the function of online error detection and recovery of the on-chip memory of the microcontroller. Other operating environments can also apply the relevant methods of the present invention after adding some additional operations.

As shown in Figure 1, it is a schematic diagram of the address allocation of the on-chip memory of the microcontroller in the system embodiment of the present invention. There are three types of programs running on it: The BOOT startup program is used to initialize the operating environment of the microcontroller and the peripheral equipment of the processor. The entire system enters the initial operating state; the error detection and correction program is used to guide the application program and perform ECC error detection and correction of the application code, and backup and restore the application code and key data; the application program is used to implement the specific functions of the system . During program design, the memory between the low address of the memory reserved area (Resv_low) and the high address of the memory reserved area (Resv_high) is reserved for the error detection and correction program for separate use, and is used to store the segmented ECC check code of the application code , Application code and key data backup. Other programs cannot use the memory space of the memory reserved area. This can be achieved by specifying the program's operating address space to avoid the memory reserved area during the program compilation and linking phase.

In this embodiment, the segmented ECC check code of the code segment is stored in three copies, which are ECC code A, ECC code B, and ECC code C; the redundant backups of the code segment and key data are all backed up at least two copies, each of which is backup A and backup B.

The application program parameter setting, ECC check code calculation and program segment backup process are shown in Figure 2. The error detection and correction program calculates the ECC check code of the application code segment by segment according to the length of the ECC segment and stores it in the application code ECC area of the memory Resv area. In order to prevent the ECC check code from changing during operation, the ECC check code of each piece of data is stored in three copies, namely ECC code A, ECC code B, and ECC code C. After that, the error detection and correction program separates the code segment according to the application file type, compresses the code segment data and calculates the check code, and stores the compressed data and check code in the application code backup A and backup B in the Resv area of the memory , And then guide the application to run.

The key parameter setting, application key data backup process, as shown in Figure 3. After the application program runs, the number of segments required for ECC verification of the error detection and correction program at one time, the processing mode after an unrecoverable error occurs, and other parameters are set to the error detection and correction program. After that, the application program registers the key data that needs to be protected with the error detection and correction program through the interface provided by the error detection and correction program. The error detection and correction program obtains the memory addresses of all key data, backs up the content in these memory addresses, and presses the hash The (hash) table is serialized and stored in the key data backup area. The serialized whole piece of content is sequentially stored in data backup A and data backup B in the key data backup area. Each backup data is used to calculate the check code separately. Verify the correctness of the backup data and attach it to the end of the backup data.

The program segment error detection and recovery process of the application program is shown in Figure 4. After the system runs normally, the error detection and correction program periodically performs ECC error detection and correction on the original program segment data according to the segment length. The segment length is configured by the application program to the error detection and correction program in the previous step. Each time the ECC check code is calculated for the program data of the segment length, and the pre-stored ECC code A, ECC code B, and ECC code C are comprehensively judged. If the original data has a single-bit error, it will directly perform error correction and recovery; at the same time, if the pre-stored ECC check code has an error, the check code will be restored to the correct value. In order to reduce the influence of error detection and correction on the system load, the error detection and correction program only detects one segment of data each time it runs, and the next segment of data is tested during the next run. When a multi-bit error that cannot be recovered by ECC is detected or all ECC check codes are wrong, it will enter the redundant comparison recovery process. When all ECC check codes are detected to be wrong, first verify that the currently running program is correct. If it is not correct, check whether backup A and backup B are correct respectively. If the data of backup A is correct, restore backup A to the running code Segment, the system preemption function needs to be turned off when restoring to prevent the program from being preempted by the application while restoring the code segment and causing program operation problems; if the backup A is incorrect and the backup B is correct, restore the backup B to the running code segment, and also need to Backup B is restored to backup A to ensure that subsequent verification can continue correctly. If it is found that both backup A and backup B are incorrect, it means that a very serious error has been made and the error cannot be recovered. At this time, the error detection and correction program will process the system according to the processing method set in advance by the application, such as restarting System, let the application quit running, etc. The error detection and correction program will also record the detected errors in its own dedicated variables for the application program to refer to. When a multi-bit error that cannot be recovered by ECC is detected, it is only necessary to verify whether backup A and backup B are correct. There is no need to verify whether the currently running program is correct. The subsequent process is the same as the processing flow when all ECC check codes are wrong.

The application of critical data error detection and recovery process is shown in Figure 5. After the system is running normally, the error detection and correction program periodically checks the correctness of the registered key data, obtains the memory address of the data by querying the hash table, and compares the key data with the backup A. If they are consistent, it means no occurrence In case of errors, the error detection and correction procedures will directly return to the next operation. When the key data is inconsistent with the backup A, it means that one party has made an error. At this time, compare whether the key data is the same as the backup B. The same means that the key data and the backup B are correct. The content of the backup A is wrong, and the backup will be backed up. After B is restored to backup A, the error detection and correction procedure returns. If the key data is different from backup A and backup B, you need to verify whether backup A and backup B are correct. If the data of backup A is correct, restore backup A to the key data; if backup A is incorrect, backup B is correct , The backup B is restored to the key data, and the backup B needs to be restored to the backup A at the same time to ensure that the subsequent verification can continue correctly. The system preemption function needs to be turned off when recovering to prevent the program from being preempted by the application program while the code segment is being recovered. If both backup A and backup B are found to be incorrect, it means that a very serious error has been made and the error cannot be recovered. At this time, the error detection and correction program will process the system according to the processing method set in advance by the application, such as restarting System, let the application quit running, etc. The error detection and correction program will also record the detected errors in its own dedicated variables for the application program to refer to.

The system of the present invention is differentiated based on functional modules, mainly including:

(1) Single bit error recovery module:

ECC check code calculation sub-module: used to calculate the ECC check code of the application running area when the application is loaded according to the preset ECC segment length to obtain the ECC check code of the segment data of the application when the application is loaded, and Perform ECC check code calculation on the segment data of the application when the application is running according to the preset ECC segment length to obtain the ECC check code of the segment data of the application when the application is running;

ECC check code storage sub-module: After obtaining the ECC check code of the segment data of the application program when the application program is loaded, it is used to store the obtained ECC check code and the storage quantity is not less than three;

ECC check code comparison sub-module: used to compare the stored ECC check codes one by one. If at least two comparison results of the stored ECC check codes are consistent, extract the ECC check with the same comparison result The code is compared with the ECC check code;

ECC check code replacement sub-module: used to extract ECC check codes with consistent comparison results, and compare and replace ECC check codes with inconsistent results;

Single-bit error correction sub-module: If a single-bit error occurs in the segment data of the application program when the application program is running according to the comparison result, it is used to correct the bit where the error occurs.

(2) Redundant backup and recovery module

Code segment acquisition sub-module: used to acquire the code segment of the application when the application is loaded;

CRC check code calculation sub-module: used to calculate the CRC check code of the code segment of the application program when the application program is loaded to obtain the CRC check code of the code segment, and perform CRC check on the code segment of the application program when the application program is running Check code calculation to obtain the CRC check code of the code segment;

CRC check code storage sub-module: After obtaining the code segment and CRC check code of the application program when the application is loaded, it is used to store the obtained code segment and its CRC check code and the storage quantity is not less than two copies. ；

CRC check code online check sub-module: used to check the stored CRC check code online;

CRC check code comparison sub-module: used to compare CRC check code with CRC check code;

Code segment coverage sub-module: If the online check result of the stored CRC check code is always consistent, it is used to extract the code segment corresponding to the CRC check code whose online check result is always consistent, and the code of the application program when the application program is running Segment to cover;

CRC check code replacement sub-module: used to extract the CRC check code and its code segment whose online check result is always consistent, and replace the CRC check code and its code segment whose online check result is not always consistent;

Code segment compression sub-module: used to compress the obtained code segment after obtaining the code segment of the application program when the application is loaded, and the code segment corresponding to the CRC check code before extracting the code segment corresponding to the CRC check code. The compressed code segment is decompressed.

(3) Critical data detection and recovery module

Memory address extraction sub-module: used to extract the memory address of the key data based on the pre-registered information of the key data;

Memory data acquisition sub-module: used to acquire the memory data in the memory address when the application program is loaded;

The aforementioned CRC check code calculation sub-module is also used to calculate the CRC check code of the memory data in the memory address when the application program is loaded to obtain the CRC check code, and to calculate the key data in the memory address of the application program when the application is running. CRC check code calculation for memory data to obtain CRC check code;

The aforementioned CRC check code storage sub-module is also used to store the obtained memory data and its CRC check code after obtaining the memory data and its CRC check code in the memory address when the application program is loaded, and the storage quantity is different. Less than two servings;

The aforementioned code segment coverage sub-module is also used to extract the memory data corresponding to the CRC check code whose online check result is always consistent if the online check result of the stored CRC check code is always consistent, and to check the memory data when the application is running. The memory data in the address is overwritten;

The aforementioned CRC check code replacement submodule is also used to extract the CRC check code and its memory data whose online check results are always consistent, and replace the CRC check code and its memory data rows whose online check results are not always consistent.

The specific embodiment of the present invention also provides a computer processing control device, including:

Memory: used to store instructions;

The processor is configured to operate according to the instructions to execute the steps of the method for detecting and restoring the memory bit flip of a power secondary device provided by the present invention.

The specific embodiment of the present invention also provides a computer-readable storage medium on which a computer program is stored. When the program is executed by a processor, the steps of the method for detecting and restoring the memory bit flip of a power secondary device provided by the present invention are realized.

Those skilled in the art should understand that the embodiments of the present application can be provided as methods, systems, or computer program products. Therefore, this application may adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, this application may adopt the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program codes.

This application is described with reference to flowcharts and/or block diagrams of methods, devices (systems), and computer program products according to embodiments of this application. It should be understood that each process and/or block in the flowchart and/or block diagram, and the combination of processes and/or blocks in the flowchart and/or block diagram can be realized by computer program instructions. These computer program instructions can be provided to the processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing equipment to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing equipment are used to generate It is a device that realizes the functions specified in one process or multiple processes in the flowchart and/or one block or multiple blocks in the block diagram.

These computer program instructions can also be stored in a computer-readable memory that can guide a computer or other programmable data processing equipment to work in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture including the instruction device. The device implements the functions specified in one process or multiple processes in the flowchart and/or one block or multiple blocks in the block diagram.

These computer program instructions can also be loaded on a computer or other programmable data processing equipment, so that a series of operation steps are executed on the computer or other programmable equipment to produce computer-implemented processing, so as to execute on the computer or other programmable equipment. The instructions provide steps for implementing the functions specified in one process or multiple processes in the flowchart and/or one block or multiple blocks in the block diagram.

The above are only the preferred embodiments of the present invention. It should be pointed out that for those of ordinary skill in the art, without departing from the technical principles of the present invention, several improvements and modifications can be made. These improvements and modifications It should also be regarded as the protection scope of the present invention.

Claims (18)

1. A method for detecting and restoring a memory bit flip of a power secondary device, which is characterized in that it includes the following steps:

   Calculate the ECC check code of the application running area when the application is loaded according to the preset ECC segment length, and obtain the ECC check code of the segment data of the application when the application is loaded;

   Perform ECC check code calculation on the segment data of the application program when the application program is running according to the preset ECC segment length, and obtain the ECC check code of the segment data of the application program when the application program is running;

   Compare the ECC check code with the ECC check code;

   If it is determined according to the comparison result that a single-bit error occurs in the segment data of the application program when the application program is running, correct the bit where the error occurred.
2. The method for detecting and restoring a memory bit flip of a power secondary device according to claim 1, wherein after obtaining the ECC check code of the segment data of the application when the application is loaded, the method further comprises: checking the obtained ECC The code is stored and the storage quantity is not less than three copies;

   Compare the ECC check code with the ECC check code, including:

   Compare the stored ECC check codes one by one;

   If at least two comparison results of the stored ECC check codes are consistent, the ECC check code with the same comparison results is extracted and compared with the ECC check code.
3. The method for detecting and restoring a memory bit flip of a power secondary device according to claim 2, characterized in that, comparing the ECC check code with the ECC check code, further comprising: extracting the ECC check code with the same comparison result, Compare and replace ECC check codes with inconsistent results.
4. The method for detecting and restoring a memory bit flip of a power secondary device according to claim 2, wherein the comparison of the ECC check code with the ECC check code further comprises:

   If the stored ECC check codes are inconsistent with each other, the redundant backup recovery mechanism is used to restore the code segment of the application when the application is running.
5. The method for detecting and restoring a memory bit flip of a power secondary device according to claim 1, wherein if it is determined according to the comparison result that there are no less than two bits in the segment data of the application program when the application program is running, there are no less than two bits, The redundant backup and recovery mechanism is used to restore the code segment of the application when the application is running.
6. The method for detecting and restoring a memory bit flip of a power secondary device according to claim 4 or 5, wherein the use of a redundant backup and restoring mechanism to restore the code segment of the application program when the application program is running includes:

   Obtain the code segment of the application program when the application program is loaded and calculate the CRC check code to obtain the CRC check code of the code segment;

   Calculate the CRC check code of the code segment of the application program when the application is running, and obtain the CRC check code of the code segment;

   Compare the CRC check code with the CRC check code;

   If the comparison result between the CRC check code and the CRC check code is inconsistent, the code segment corresponding to the CRC check code is extracted, and the code segment of the application program is covered when the application program is running.
7. The method for detecting and restoring a memory bit flip of a power secondary device according to claim 6, wherein after obtaining the code segment of the application program when the application program is loaded, the method further comprises: compressing the obtained code segment;

   Before extracting the code segment corresponding to the CRC check code, it also includes: decompressing the compressed code segment corresponding to the CRC check code.
8. The method for detecting and restoring a memory bit flip of a power secondary device according to claim 6, characterized in that, after acquiring the code segment of the application program and its CRC check code when the application program is loaded, the method further comprises: checking the acquired code segment And its CRC check code is stored and the storage quantity is not less than two copies;

   Extract the code segment corresponding to the CRC check code, including:

   Online verification of the stored CRC check code;

   If the online check result of the stored CRC check code is always consistent, extract the code segment corresponding to the CRC check code whose online check result is always consistent.
9. The method for detecting and restoring a memory bit flip of a power secondary device according to claim 8, wherein extracting the code segment corresponding to the CRC check code further comprises: extracting the CRC check code whose online check result is always consistent, and The code segment replaces the CRC check code and its code segment whose online verification results are not always consistent.
10. The method for detecting and restoring a memory bit flip of a power secondary device according to claim 8, wherein the online verification of the stored CRC check code includes: checking the stored CRC at a preset time interval The current state of the code is compared with the previous state.
11. The method for detecting and restoring a memory bit flip of a power secondary device according to claim 1, further comprising: detecting and restoring key data pre-registered in an application;

    The key data detection and recovery method includes:

    Extract the memory address of the key data based on the pre-registration information of the key data;

    Obtain the memory data in the memory address when the application program is loaded and calculate the CRC check code to obtain the CRC check code;

    Calculate the CRC check code for the memory data in the memory address of the key data when the application is running, and obtain the CRC check code;

    Compare the CRC check code with the CRC check code;

    If the comparison result between the CRC check code and the CRC check code is inconsistent, the memory data corresponding to the CRC check code is extracted, and the memory data in the memory address is overwritten when the application program is running.
12. The method for detecting and restoring a memory bit flip of a power secondary device according to claim 11, wherein after acquiring the memory data and the CRC check code in the memory address when the application program is loaded, the method further comprises: acquiring The memory data and its CRC check code are stored, and the storage quantity is not less than two copies;

    Extract the memory data corresponding to the CRC check code, including:

    Online verification of the stored CRC check code;

    If the online check result of the stored CRC check code is always consistent, extract the memory data corresponding to the CRC check code whose online check result is always consistent.
13. The method for detecting and restoring a memory bit flip of a power secondary device according to claim 12, wherein extracting the memory data corresponding to the CRC check code further comprises: extracting the CRC check code whose online check result is always consistent, and The memory data is replaced with the CRC check code and the memory data row whose online check result is not always consistent.
14. A system for detecting and restoring memory bit flips of power secondary equipment, which is characterized by comprising: a single-bit error recovery module, and the single-bit error recovery module includes:

    ECC check code calculation sub-module: used to calculate the ECC check code of the application running area when the application is loaded according to the preset ECC segment length to obtain the ECC check code of the segment data of the application when the application is loaded, and Perform ECC check code calculation on the segment data of the application when the application is running according to the preset ECC segment length to obtain the ECC check code of the segment data of the application when the application is running;

    ECC check code storage sub-module: After obtaining the ECC check code of the segment data of the application program when the application program is loaded, it is used to store the obtained ECC check code and the storage quantity is not less than three;

    ECC check code comparison sub-module: used to compare the stored ECC check codes one by one. If at least two comparison results in the stored ECC check codes are consistent, extract the ECC check with the same comparison result The code is compared with the ECC check code; ECC check code replacement sub-module: used to extract the ECC check code with the same comparison result, and compare and replace the ECC check code with inconsistent results;

    Single-bit error correction sub-module: If a single-bit error occurs in the segment data of the application program when the application program is running according to the comparison result, it is used to correct the bit where the error occurs.
15. The system for detecting and restoring the memory bit flip of a power secondary device according to claim 14, characterized in that it further comprises a redundant backup and restoration module for restoring the code segment of the application when the application is running by using a redundant backup and restoration mechanism, The redundant backup and recovery module includes:

    Code segment acquisition sub-module: used to acquire the code segment of the application when the application is loaded;

    CRC check code calculation sub-module: used to calculate the CRC check code of the code segment of the application program when the application program is loaded to obtain the CRC check code of the code segment, and perform CRC check on the code segment of the application program when the application program is running Check code calculation to obtain the CRC check code of the code segment;

    CRC check code storage sub-module: after obtaining the code segment of the application program and its CRC check code when the application program is loaded, it is used to store the obtained code segment and its CRC check code and the number of storage is not less than two copies ；

    CRC check code online check sub-module: used to check the stored CRC check code online;

    CRC check code comparison sub-module: used to compare CRC check code with CRC check code;

    Code segment coverage sub-module: If the online check result of the stored CRC check code is always consistent, it is used to extract the code segment corresponding to the CRC check code whose online check result is always consistent, and the code of the application program when the application program is running Segment to cover;

    CRC check code replacement sub-module: used to extract the CRC check code and its code segment whose online check result is always consistent, and replace the CRC check code and its code segment whose online check result is not always consistent;

    Code segment compression sub-module: used to compress the obtained code segment after obtaining the code segment of the application program when the application is loaded, and the code segment corresponding to the CRC check code before extracting the code segment corresponding to the CRC check code. The compressed code segment is decompressed.
16. The system for detecting and restoring a memory bit flip of a power secondary device according to claim 15, characterized in that it further comprises a key data detection and restoration module, and the key data detection and restoration module comprises:

    Memory address extraction sub-module: used to extract the memory address of the key data based on the pre-registered information of the key data;

    Memory data acquisition sub-module: used to acquire the memory data in the memory address when the application program is loaded;

    The CRC check code calculation sub-module is also used to perform CRC check code calculation on the memory data in the memory address when the application program is loaded to obtain the CRC check code, and to calculate the key data in the memory address when the application program is running CRC check code calculation is performed on the memory data to obtain the CRC check code;

    The CRC check code storage sub-module is also used to store the obtained memory data and its CRC check code after obtaining the memory data and its CRC check code in the memory address when the application program is loaded. Not less than two copies;

    The code segment coverage sub-module is also used to extract the memory data corresponding to the CRC check code whose online check result is always consistent if the online check result of the stored CRC check code is always consistent, and the application program is running. Overwrite the memory data in the memory address;

    The CRC check code replacement sub-module is also used to extract the CRC check code whose online check result is always consistent and its memory data, and replace the CRC check code whose online check result is not always consistent and its memory data row.
17. The computer processing control device is characterized in that it includes:

    Memory: used to store instructions;

    Processor: configured to operate according to the instructions to execute the steps of the method described in any one of claims 1 to 13.
18. A computer-readable storage medium having a computer program stored thereon is characterized in that, when the program is executed by a processor, the steps of the method according to any one of claims 1 to 13 are realized.

Images