JPH05241872A - Abnormality detection system for task - Google Patents

Abstract

PURPOSE:To easily detect the abnormality of a task without providing an exclu sive abnormality detection function on the task side and to recover from the fault at an early stage when a certain task is fallen into an indefinite loop state by a logic abnormality or a hardware abnormality in the task control by a time division OS. CONSTITUTION:In a software where a program (task) performing plural different processings operates in a time division, a task identification function identifying a task in operation when the processing within a certain task is fallen into an indefinite loop by a logic abnormality or a hardware abnormality, an address detection function detecting the operation address of the task identified in this task identification function, an address decision function deciding that the address detected by this address detection function has only changes within a stipulated range (the task does not move), a state recognition function recognizing the abnormal state by the decision of this address decision function and a restart function changing the operation address of the task to the head of the task when the abnormal state is recognized by this state recognition function.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a task abnormality detection method, and more particularly to a task abnormality detection method when a plurality of tasks are arbitrated and operating in time division in any software / firmware.

[0002]

2. Description of the Related Art A program (task) that performs a plurality of different processes is developed for the purpose of sequentially operating tasks in a time-sharing manner (operating system (O
When operating under S)), the task time-sharing operation is as shown in FIGS.

As shown in FIG. 2, when the task A operates for the time T, the OS forcibly cancels the processing of the task A and operates the task B regardless of the operating state of the task A.
When the OS terminates the processing of task A, it saves the operating environment such as the aborted program address and processing data as shown in FIG. 3, and is unconditionally preserved when the task A is operated next time. Restore the operating environment and continue processing. In this way, the task is controlled by the OS
→ B → C → A operates in this order.

[0004]

[Problems to be Solved by the Invention] Time division O in the prior art
In the task control by S, even if one task enters an infinite loop state due to a logical abnormality or a hardware abnormality, the operation moves to another task in time sharing, and if the entire program is viewed from the outside, the task Abnormalities cannot be easily detected and failures cannot be recovered early.

The present invention has been made in view of the above circumstances, and therefore an object of the present invention is to provide a novel task abnormality detection system which can solve the above problems inherent in the prior art. Especially.

[0006]

In order to achieve the above object, the task abnormality detection method according to the present invention is a task in which a program (task) performing a plurality of different processes operates in a time-division manner. Task discrimination function that discriminates the task that is operating when the processing of is in an infinite loop due to a logical abnormality or hardware abnormality, and an address detection function that detects the operating address of the task discriminated by this task discrimination function, An address determination function that determines that the address detected by this address detection function has only a change within a specified range (task is not running), and a state recognition function that recognizes an abnormal state by the determination of this address determination function, A lister that changes the operation address of the task to the beginning of the task when the status recognition function recognizes it in an abnormal state. Constituted by a door function.

That is, according to the present invention, only the program address is extracted from the data group defining the operating environment of the task to be saved when the operating task is switched, and the OS is provided with the function of monitoring. ..

This program address indicates the movement of the task at regular time intervals. Therefore, when there is almost no change in this program address, the task may not be operating (in an infinite loop state).

The time during which the OS permits one task to perform continuous operation is T, and the address section in which there is no data to be processed by the task is α bytes, and its operating time is β.
In this case, the difference between the task discrimination function for discriminating the task, the address detection function for extracting the program address, and the maximum and minimum values of the program address extracted T / β times in the past is less than α bytes. The task can be considered to be stopped (infinite loop operation) due to some failure by the address determination function.

This is recognized as an abnormal state by the state recognition function, and the OS restarts the task by the restart function that clears the operating environment data returned by the OS and changes the program address to the head position of the task. Also, an alarm output function notifies the outside of the abnormal state and the abnormal place.

[0011]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS The preferred embodiment of the present invention will be described below in detail with reference to the drawings.

FIG. 1 is a diagram showing a processing flow of an OS incorporating a function for realizing the present invention.

FIG. 2 is a block diagram showing the operation of each task under normal conditions.

In FIG. 2, when the OS switches the operation of task A to task B, the operating environment of task A at that time is saved (FIG. 3). At this time, the function for realizing the present invention is incorporated (FIG. 1).

In the OS, an area for recording a program address for judging the task operation for each task and a counter area for counting how long the task is in a stopped state are provided in advance.

In the switching processing, the switched task is discriminated, and if the difference between the program address recorded at the previous switching and the program address at this time is equal to or more than a specified value, the program address is unconditionally recorded. , Clear the counter to zero. If it is less than the specified value, the program address is not recorded and the counter is incremented by +1 instead. When this counter reaches the specified value, the stored operating environment is cleared and the program address to be operated next is changed to the beginning of the task.

As a result, when the task is operated next time, only the task is restarted, and the stopped state can be recovered. Further, it is possible to recognize the abnormal state by notifying the alarm to the outside.

[0018]

As described above, according to the present invention,
By detecting an abnormal stop state of a task operating under the time-sharing OS inside the OS, it is possible to restart only that task and return to normal processing in a short time as a whole. ..

According to the present invention, since it is not necessary for each task side to have a dedicated abnormality detecting function, the effect that the processing can be efficiently advanced is obtained.

[Brief description of drawings]

FIG. 1 is a diagram showing an operation of an OS incorporating a function for realizing the present invention.

FIG. 2 is a diagram showing the movement of each task that operates based on a time-division OS.

FIG. 3 is a diagram showing an operation of a conventional time division control OS.

Claims (2)

[Claims] 1. In software in which a program (task) that performs a plurality of different processes operates in a time-sharing manner, a task that is operating when a process in a task falls into an infinite loop due to a logical abnormality or a hardware abnormality is executed. The task discrimination function for discriminating, the address detection function for detecting the operation address of the task discriminated by the task discrimination function, and the address detected by the address detection function have only the change within the specified range (the task is not moving. ), An address judgment function that recognizes an abnormal state by the judgment of the address judgment function, and if the abnormal state is recognized by the state recognition function, the operation address of the task is changed to the head of the task. A task abnormality detection method characterized by having a restart function. 2. The task abnormality detection method according to claim 1, further comprising an alarm output function for notifying an abnormal state and an abnormal portion to the outside when the abnormal state is recognized by the state recognition function.