JPH05346877A - Operation monitoring system for multi-task system - Google Patents

Abstract

PURPOSE:To prevent a malfunction of a watch dog timer caused by a task being in a waiting state, and also, to easily determine a limit time of an operation of the task at every task by executing the count by the watch dog timer with respect to the time when each task is actually in an executing state, in the multi-task system. CONSTITUTION:The system is provided with a task operation limit time table LMY for storing separately a limit time of an operation of each task, a watch dog timer circuit part TMR for counting an operation time of the task, and a task operation time counter table CNT for storing separately the counted operation time with regard to each task. In such a state, based on the task operation limit time table LMT and the task operation time counter table CNT, when the time when each task is operated actually exceeds the limit time of the operation of its task, abnormality of the system is detected and the system is initialized.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an operation monitoring system for a multitasking system, which is used to automatically detect an abnormality in the multitasking system by using a watchdog timer.

[0002]

2. Description of the Related Art Conventionally, a multi-task system has been widely used in which one CPU executes a plurality of tasks in a time-division manner. In a general multitasking system,
A watchdog timer monitors whether the system is operating normally. This is a method in which each task confirms the normal operation of the system by sending a live notification to the watchdog timer within a certain time.
If there is no notification of existence within a preset time, it is regarded as an abnormal operation of the system and the system is forcibly initialized.

[0003]

In the above-mentioned conventional example, the task needs to notify the watchdog timer of its existence within a preset time limit, but in the watchdog timer, the task is in the running state. The task operation time is counted regardless of whether the task is in the waiting state. Therefore, for example, there is a case where the system is erroneously initialized due to the fact that a certain task is in the I / O wait state and is operating normally, but the live notification cannot be made. Also, when setting the value of the predetermined time limit for waiting for the notification of existence from the task,
In addition to the task execution time, the time required for input / output and the time required for task switching must be taken into consideration, and a large value can be set with a large margin to prevent malfunction. Often, it was difficult to set meaningful values.

The present invention has been made in view of the above circumstances. An object of the present invention is to prevent a watchdog timer from malfunctioning due to a task in a waiting state and to limit a task operation time limit. It is to provide an operation monitoring method of a multitask system that can be easily determined for each task.

[0005]

In order to solve the above-mentioned problems, in a multitask system operation monitoring method according to the present invention, in a multitask system in which a plurality of tasks are simultaneously executed, A task operation time limit table LMT for individually storing the operation time limit, a watchdog timer circuit unit TMR for counting the operation time of the task, and the operation time counted for each task are individually stored. When the time during which each task actually operates based on the task operation time counter table CNT, the task operation time limit table LMT and the task operation time counter table CNT exceeds the operation time limit of the task. It is provided with an initialization means for detecting an abnormality of the system and initializing the system. It is an.

[0006]

In the present invention, when the time when each task actually operates exceeds the operation time limit of the task, the system abnormality is detected and the system is initialized. When setting the time limit for the operation of the task, it is not necessary to consider the I / O wait time that changes depending on the system operation status and the time required for task switching, and the watchdog timer waits for the notification of existence from the task. It is possible to set the time limit to a meaningful value related to the task execution time.

[0007]

1 is a block diagram showing the overall structure of the present invention. In this system, two tasks 1 and 2 are OS
Managed by (operating system),
The state is stored in the task control block (TCB). Each task issues a live notification to the watchdog timer circuit unit TMR. According to the operation flow shown in FIG. 2, the watchdog timer circuit unit TMR periodically refers to the task control block (TCB) and updates the counter corresponding to the task whose state is RUN (execution state). Then, the counter value is compared with the corresponding value on the task operation time limit table LMT, and if the time limit is exceeded, the system is forcibly reset.
Also, if there is a live notification from a certain task, the value of the counter corresponding to that task is reset. The watchdog timer circuit unit TMR recognizes the task being executed by referring to the task control block (TCB) that the OS (operating system) updates according to the execution state of the task, and recognizes the task operation time counter table C.
Update the counter corresponding to the running task on NT,
When the value of one or more counters in the counter table CNT exceeds the corresponding value in the task operation time limit table LMT, the system is initialized.

FIG. 3 exemplifies the state transitions of task 1 and task 2 over time, the timing of alive notification, and the timing of counting. In this example, the time limit of the task operation counted by the watchdog timer circuit unit TMR is 20 for task 1 and 30 for task 2. The operation in this case will be described in order. First, as shown in FIG. 4, the task control block (TCB) is updated by the OS (operating system), the state of task 1 becomes the RUN state, the state of task 2 becomes the READY state, The watchdog timer circuit unit TMR updates only the counter corresponding to the task 1 whose state is the RUN state in the task operation time counter table CNT, and when the task 1 enters the WAIT state, the counter of the task 1 becomes 10. . Next, as shown in FIG. 5, task 1 enters the WAIT state to wait for the completion of input / output from the I / O (input / output device), and task 2 outputs R.
When it goes to the UN state, the watchdog timer circuit section TM
R updates only the counter corresponding to the task 2 whose state is the RUN state in the task operation time counter table CNT. The counter corresponding to task 2 is 2
When it is counted up to 2, it is reset by the live notification from the task 2, becomes 0 again, and the task 2 becomes R.
The counter is incremented to 18 before returning to the EADY state. Then, as shown in FIG. 6, when the task 1 becomes the RUN state again, the watchdog timer circuit unit TMR updates only the counter corresponding to the task 1 whose state is the RUN state in the task operation time counter table CNT. .. When the counter corresponding to task 1 is counted up to 15, it is reset to 0 by the live notification from task 1.

[0009]

According to the present invention, in a multitasking system, a task in a waiting state is alive in the watchdog timer because the watchdog timer counts the time when each task is actually in the executing state. There is no inconvenience such as being reset accidentally by not notifying, and when setting the time limit of the operation of each task, it is not necessary to consider the operation of other tasks or the input / output wait time, This has an effect that the set value of the time limit for forcibly resetting can be easily determined based on the program of each task.

[Brief description of drawings]

FIG. 1 is a block diagram showing an overall configuration of the present invention.

FIG. 2 is a flowchart for explaining the operation of the present invention.

FIG. 3 is an explanatory diagram showing a state transition of a task according to the present invention.

FIG. 4 is an explanatory diagram showing a first operation during execution of the first task in the present invention.

FIG. 5 is an explanatory diagram showing an operation at the time of execution of a second task in the present invention.

FIG. 6 is an explanatory diagram showing a second operation when the first task according to the present invention is executed.

[Explanation of symbols]

 LMT task operation time limit table CNT Task operation time counter table TMR Watchdog timer circuit

Claims (1)

[Claims] 1. In a multitask system in which a plurality of tasks are executed simultaneously, a task operation time limit table for individually storing the operation time limit of each task and a task operation time limit table for counting the task operation time Based on the watchdog timer circuit section, the task operation time counter table for individually storing the operation time counted for each task, and the task operation time limit table and the task operation time counter table An operation monitoring method for a multi-task system, comprising: an initialization means for detecting a system abnormality and initializing the system when the operation time exceeds the operation time limit of the task.