JP7825268B2 - Task monitoring system and task monitoring program - Google Patents

Description

The present invention relates to a task monitoring system and task monitoring program for monitoring whether each task is being executed correctly and without abnormalities in a computer that executes multiple tasks each having its own periodic timer, and is particularly suitable for embedded microcomputers.

Conventionally, computers (including microcomputers that consolidate functions such as calculation, storage, and input/output into a small number of integrated circuits; hereafter simply referred to as "computers") that execute tasks using multiple different periodic timers have been equipped with software or hardware monitoring systems that monitor whether each of the tasks is being executed reliably and without abnormalities, making it possible to identify abnormalities and take measures to return the system to a normal state.

A traditional example of such a monitoring system is a hardware-based monitoring system using a watchdog timer (hereinafter referred to as "WDT"), which has long been known.

This WDT-based monitoring system determines whether the system is continuing to operate normally by connecting the computer and WDT internally or externally, and using the fact that a signal (watchdog signal) from the computer, which is generated within a specified cycle upon completion of a specific task, is input to the WDT within a timeout period, preventing traps and resets.If the watchdog signal is not input after the timeout period has elapsed, the WDT will determine that an abnormality has occurred in the computer and send a reset signal to reset the computer and prevent abnormal computer operation.

In the case of a conventional WDT-based monitoring system, for example, as shown in the example of Figure 5, when three types of tasks with different periodic timers (e.g., 1 ms, 5 ms, 10 ms) are executed by interrupts with the shortest execution period, the highest priority, a watchdog signal is output for the task executed with the longest periodic timer (10 ms) for the minimum time (not including execution error), as shown in Figure 6, and monitoring is typically performed by an internal WDT monitoring system as shown in Figure 7 or an external WDT monitoring system as shown in Figure 8.

In other words, because conventional WDT-based monitoring systems focus only on the longest period timer, when operations are executed using high-priority interrupts with shortest period timers, even if a task executed with a shorter period timer than the task executed with the longest period timer is not executed during the specified number of interrupt processing executions, a signal from the computer (watchdog signal) is input to the WDT within the timeout period, so the WDT cannot reset the computer and may not be able to prevent abnormal operation of a task executed with a short period timer.

For example, Japanese Patent Laid-Open Publication No. 51-052256 (Patent Document 1) discloses a WDT monitoring system that estimates the interrupt priority of periodic timers and the maximum execution time considered for each periodic timer, monitors whether the overall maximum execution time exceeds a threshold with a margin, and can reliably reset the computer.

In addition, a method for preventing program runaway is disclosed, for example, in Japanese Patent Laid-Open Publication No. 62-298844 (Patent Document 2), in which identification information for a modularized program is stored when it is executed, and the program is terminated abnormally if the information is determined to be inconsistent (abnormal).

However, both of these methods focus on and monitor module processing, which is a collection of periodic timers and programs that have the lowest execution priority and can potentially have long periodic tasks. This creates the problem that it is not possible to confirm whether tasks that have higher execution priority and are executed by short-period timers are being executed individually each time.

Japanese Patent Application Publication No. 51-052256 Japanese Unexamined Patent Publication No. 62-298844

The objective of the present invention is to provide a monitoring system that, in a computer that executes multiple tasks with different periodic timers, can reliably monitor tasks executed with short periodic timers using a method different from conventional monitoring systems.

The task monitoring system of the present invention, developed to solve the above problems, comprises a computer that executes multiple tasks, each with its own periodic timer, and monitoring means provided inside or outside the computer. It is a monitoring system for verifying that each task is being executed without any abnormalities. The monitoring means has a management variable for each task that fluctuates depending on the task's execution status, and monitors the task based on the value of the management variable.

This invention makes it possible to provide a monitoring system for a computer equipped with multiple tasks each having its own periodic timer, which can monitor each task individually, regardless of the length of the period or the priority of the task, rather than managing the tasks based on time as in conventional methods.

In particular, if the control variable is set to an initial value and then varies for each task by being decremented when the task is executed and incremented when the task is completed, by varying the control variable at the timing of task execution and completion, it is possible to monitor whether each task is being executed and completed normally based on the value of the control variable.

Furthermore, when all of the tasks are monitored using the sum of the control variables, which is the total value of the control variables, it is possible to monitor all tasks more efficiently than using individual control variables.

Furthermore, if the computer is equipped with a DAC (digital-analog converter), the monitoring means is equipped with a comparator circuit and an open-collector circuit, the DAC (digital-analog converter) converts the control variable into an analog voltage and outputs it as a judgment signal, the comparator circuit compares the judgment signal with a threshold value to determine whether it is normal or abnormal, and the open-collector circuit outputs the judgment result, the output judgment result can be used, for example, by an external hardware check mechanism to reset the computer, stop an actuator, or notify of an error.

Furthermore, a task monitoring program for verifying that each task is being executed normally on a computer that executes multiple tasks each having its own periodic timer, includes the steps of: setting an initial value for a control variable that is prepared for each task and varies depending on the execution status of each task; varying the control variable for each task by decrementing it during execution and incrementing it at completion; comparing the control variable with a threshold value, or by comparing the sum of the control variables, which is the total value of the control variables, with a threshold value, to determine whether the execution of the task related to the control variable or all tasks is normal or abnormal; and outputting the results of the determination. This enables task monitoring to be performed on a program executed by a computer.

According to the present invention, in a computer that executes multiple tasks with individual periodic timers, tasks executed with short periodic timers can also be reliably monitored by using a management variable that changes depending on the execution status of each task.

In a preferred embodiment of the task monitoring system of the present invention, an execution schedule showing the case where all tasks in a computer that executes multiple tasks each having an individual periodic timer are executed normally. FIG. 2 is an explanatory diagram showing control of management variables of each task in the embodiment shown in FIG. 1 . 3 is an explanatory diagram showing an example of detecting an abnormality in the control of the management variables shown in FIG. 2, and whether each task is normally executed and completed. 2. A hardware interface circuit for processing the control variables as signals in the embodiment shown in FIG. An execution schedule that shows the execution of multiple tasks under the shortest ideal conditions in a computer that executes multiple tasks each having its own periodic timer. 1 is an execution schedule showing an example of WDT monitoring of the longest execution time in a computer that executes multiple tasks each having its own periodic timer, in a task monitoring system using a conventional WDT. FIG. 1 is an explanatory diagram of an external WDT monitoring method in a conventional WDT-based task monitoring system. FIG. 1 is an explanatory diagram of an internal WDT monitoring method in a task monitoring system using a conventional WDT.

The following describes a preferred embodiment of the present invention with reference to the drawings.

Figure 1 shows an execution schedule for a computer that executes tasks with individual and mutually different periodic timers (e.g., three periods: 1 ms, 5 ms, and 10 ms) in a preferred embodiment of the task monitoring system of the present invention, where all tasks are executed normally. In this embodiment, the execution status of each task is managed using a variable-based management method that does not rely on time management, and it is confirmed that each task is being executed without any abnormalities. This variableization of the execution status of each task is a feature of the present invention.

Table 1 below shows the execution status of each task in this embodiment, expressed as a control variable (2 bits, 4-value control), and its status. In Figure 1 above, the control variable for the task with a periodic timer of 1 ms is A, the control variable for the task with a periodic timer of 5 ms is B, and the control variable for the task with a periodic timer of 10 ms is C.

In this embodiment, as shown in Figure 1, the start and end of one control cycle for each task of a different periodic timer is monitored using the management variables shown in Table 1.

In more detail, as shown in Figure 2, the initial value of the control variable is first set to 2 (10), and then the control variable for each task is decremented (-1) when the task is executed and incremented (+1) when the task ends.

In this case, for a task with a specific periodic timer, the time between task execution and the next task execution is based on that periodic timer (black arrow), and the time between task execution and task completion is not specifically specified.

As such, the control variables A, B, and C each take on four values: 0, 1, 2, and 3, with the normal value being 1 or 2 (see Table 1 above).When a task is being executed normally, the values of control variables A, B, and C will fluctuate between 1 and 2 due to a decrement (-1) when the task is executed and an increment (+1) when the task ends.

In other words, in this embodiment, when the values of control variables A, B, and C are 1 or 2, it can be determined that the execution status of the task corresponding to each control variable is normal.

On the other hand, as shown in Figure 3, for each task, a decrement (-1) is made when the task is executed and an increment (+1) is made when the task ends (for the n-1th task, the nth task, etc.). If the n-1th task does not end or the next nth task is not executed, the addition or subtraction will cause the control variable to fall outside the normal range and become 0 or 3, so abnormal task execution can be confirmed by the value of the control variable.

In addition, by using the control variable sum X, which is the sum of the control variables A, B, and C for each of the three types of tasks shown in this embodiment, it is possible to check whether each task is running and completing normally.

We will explain how to determine whether the execution status of each task is normal or abnormal using control variables A, B, C and the sum of control variables X. Table 2 below shows the relationship between the values of control variables A, B, C and the sum of control variables X.

As such, the control variables A, B, and C each have four possible values: 0, 1, 2, and 3, with the normal value being 1 or 2 (see Table 1 above). Therefore, when the sum of these values is taken as the control variable sum X, a value between 3 and 6 falls within the normal range, and the execution status of each task can be determined to be normal.

However, if the control variable sum X is outside the normal range of 3 to 6, i.e., if the control variable sum X is 0 to 2 or 7 to 9, it is an abnormal value, and the execution status of each task can be determined to be abnormal.

In this embodiment, the control variables are shown for four states, but by increasing the number of state bit digits beyond these four states, it becomes possible to set more precise thresholds, thereby strengthening or mitigating abnormality detection.

The results of the judgment of the execution status of each task using the control variable or the sum of the control variables are output, so that if the judgment result is abnormal, it is possible to reset the computer, stop the actuator, or notify an error, for example.

Furthermore, Figure 4 shows a hardware interface circuit (analog monitoring circuit) for external monitoring using the control variables of each task in the embodiment shown in Figure 1 as signals. It has a comparator circuit (COMP), a latch circuit (self-holding circuit), and an open-collector circuit that process signals output from a DAC (digital-to-analog converter, hereafter simply referred to as "DAC") provided in the computer.

First, the control variable used in the embodiment shown in Figure 1 is converted to an analog voltage by the DAC and output as a judgment signal.

Table 3 below shows the correlation between digital values and analog voltage values when the control variable sum X, which is the sum of the control variables A, B, and C for the three tasks in the embodiment shown in Figure 1, is output as an analog voltage of 0 to 5 V by the DAC as a judgment signal. The DAC can also use the control variable sum X as it is, as 5 V x control variable sum X ÷ MAX value.

Then, the analog voltage value of the output judgment signal is compared with a threshold value by a comparator circuit (COMP) to determine whether it is normal or abnormal.

As shown in Table 3, in this embodiment, an upper threshold and a lower threshold are set, and the range between the upper threshold and the lower threshold is considered to be the normal range, while values above the upper threshold and below the lower threshold are considered to be abnormal values.

Specific numerical values are 3.61V as the upper threshold and 1.39V as the lower threshold, with each threshold being set to the midpoint between the normal range and the abnormal value, but are not limited to this.

At this time, the judgment result output from the comparator circuit (COMP) is input to a latch circuit (self-holding circuit), which allows the abnormal state to be maintained until an abnormal state cancellation signal is input.

Then, by outputting the judgment result from the comparator circuit to the outside via the open collector circuit, monitoring can be performed using a hardware check mechanism, such as a reset operation using the computer's reset circuit, a stop operation using the actuator drive circuit's stop circuit, or an alarm using a warning indicator light.

Note that in the present invention, the number of tasks and the number of individual periodic timers are not limited to three as shown in this embodiment, but can be any number.

The task monitoring system of the present invention can also be realized as a task monitoring program executed on a computer.

In other words, the task monitoring program of the present invention is a task monitoring program for verifying that each task, each having its own periodic timer, is being executed without abnormalities in a computer that executes multiple tasks. The program executes the following steps: setting an initial value for a management variable that is prepared for each task and varies depending on the execution status of each task; varying the management variable for each task by decrementing it during execution and incrementing it at completion; comparing the management variable with a threshold value, or by comparing the sum of the management variables, which is the total value of the management variables, with a threshold value, to determine whether the execution of the task related to the management variable or all tasks is normal or abnormal; and outputting the results of the determination.

As described above, according to the present invention, in a computer that executes multiple tasks with individual periodic timers, by using a management variable that changes depending on the execution status of each task, it is possible to reliably monitor tasks executed with short periodic timers.

Claims (3)

A monitoring system comprising a computer that executes a plurality of tasks, each having a periodic timer, and a monitoring means provided inside or outside the computer, for confirming that each of the tasks is being executed without an error,
the monitoring means has a control variable for each task that varies depending on the execution state of the task, and monitors the task based on the value of the control variable ;
A task monitoring system characterized in that the management variable is set to an initial value and then varies for each task by being decremented when the task is executed and incremented when the task is completed . the computer is equipped with a DAC (digital-to-analog converter);
the monitoring means includes a comparator circuit and an open collector circuit;
The task monitoring system according to claim 1, characterized in that the DAC (digital-analog converter) converts the management variable into an analog voltage and outputs it as a judgment signal, the comparator circuit compares the judgment signal with a threshold value to determine whether it is normal or abnormal, and the open collector circuit outputs the result of the judgment. 1. A task monitoring program for verifying that each task is being executed normally in a computer that executes a plurality of tasks each having an individual periodic timer, the program comprising:
setting an initial value of a control variable that is prepared for each task and that varies depending on the execution state of each task;
fluctuating the control variable by decrementing it when the task is executed and incrementing it when the task is completed;
a step of comparing the control variable with a threshold value to determine whether execution of a task related to the control variable is normal or abnormal;
outputting a result of the determination;
A task monitoring program that executes the above steps.