A method for preventing system failure based on on-board software

Abstract

The application relates to a method for preventing system failure based on a satellite-borne software, which comprises the following steps: a target program comprises a plurality of parallel threads, the target program is copied into two parts, which are respectively used for executing an entity process and a virtual process; the entity process and the virtual process asynchronously execute the target program, and the virtual process executes at least one thread prior to the entity process; the virtual process synchronously executes a plurality of threads, marks the states of the executed threads, records an execution result if the state mark is a normal thread, and informs the entity process to execute the thread; if the state mark is an abnormal thread, the execution of the thread is ended, and the entity process is informed not to execute the thread. The application does not modify a source program and does not add new hardware support, builds asynchronous virtual fault-tolerant software, preforms the target program in advance, avoids the collapse of the entity process when the program is formally executed, and prevents the system failure of a spacecraft controller caused by the collapse of the program.

Claims

1. A method for preventing system failure based on onboard software, characterized in that: Includes the following steps: The target program includes several parallel threads. The target program is copied into two copies, one for execution by the physical process and the other by the virtual process. The physical process and the virtual process execute the target program asynchronously, and the virtual process executes at least one thread before the physical process. The virtual process executes multiple threads synchronously. The executing threads are marked with a status. If the status is marked as normal, the execution result is recorded and the entity process is notified to execute the thread. If the status is marked as abnormal, the execution of the thread is terminated and the entity process is notified not to execute the thread. Each of the aforementioned threads contains several function blocks connected in sequence; When a virtual process executes any thread, it calls and executes the function blocks in that thread in sequence. When a function block is executed, the status is marked as thread normal, and the next function block is called and executed. This continues until all function blocks of the thread have been called and executed, and the status of all function blocks is marked as thread normal. Then, the physical process is notified to execute the thread. When a virtual process executes a certain function block and its status is marked as thread abnormal, the virtual process rolls back to the recovery point, which is the previous function block, and notifies the physical process not to execute that thread.

2. The method for preventing system failure based on onboard software according to claim 1, characterized in that: When a virtual process marks a thread as abnormal, it restarts the thread. After restarting, it continues to call and execute the next function block from the recovery point until all function blocks of the thread have been called and executed, and the status of all function blocks is marked as normal. Then, it notifies the physical process to execute the thread.

3. The method for preventing system failure based on onboard software according to claim 2, characterized in that: A thread restarted in a virtual process will also have its corresponding thread restarted in the physical process.

4. The method for preventing system failure based on onboard software according to claim 1, characterized in that: If the entity process finishes executing a thread according to the notification from the virtual process, and has not yet received a notification from the virtual process for an executable thread, the entity process enters a waiting state until it receives a notification from the virtual process for an executable thread, at which point it executes the thread that was notified.

5. The method for preventing system failure based on onboard software according to claim 1, characterized in that: After executing the thread notified by the virtual process, the entity process compares the execution result with the execution result recorded by the virtual process. If the execution results are consistent, the entity process executes the next thread notified by the virtual process; otherwise, the entity process returns to the starting point of the thread.

6. The method for preventing system failure based on onboard software according to claim 5, characterized in that: When the physical process compares the execution result with the execution result recorded by the virtual process, and the execution results are inconsistent, the process rolls back to the beginning of the thread, restarts the thread, and re-executes the thread until the execution result is consistent with the execution result of the virtual process, thus completing the execution of the thread.

Images