JP2000105707A - System for detection of abnormal state of processing program - Google Patents

Abstract

PROBLEM TO BE SOLVED: To obtain a system for detection of an abnormal state of a processing program which can detect an abnormality even in case of event successive occurrence by finding a difference between the time stored in a start time entry and the current time, and judging an abnormality and outputting an alarm if the difference exceeds a certain time. SOLUTION: A constant-cycle monitoring process (140) which operates in certain cycles extracts (140-1) the contents of a start time entry of a process management table which are not 0 (i.e., not in an event waiting state) and calculates (140-2) the difference from the current time entry (151-d). Here, the calculated difference is the time when a program execution begins or the program is held in an execution state after being released from waiting. The time is compared (140-3) with a prescribed time, and if the former exceeds the latter, a fault is assumed to exist, so that an alarm is outputted (140-0). Therefore, the time when the processing program 101 is executed continuously is monitored and its abnormality can easily be detected.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a method for detecting an abnormality in a processing program, and more particularly to a method for detecting an abnormality in a processing program when a plurality of processing programs are executed in accordance with an execution priority.

[0002]

2. Description of the Related Art Conventionally, when detecting an in-loop state due to a logical abnormality of a processing program without using dedicated hardware such as a watchdog timer, a processing time for the processing program for one event is defined.
If the processing did not end within the specified time, it was regarded as abnormal.

FIG. 5 shows a time chart according to a conventional abnormality detection method. When the processing program is started 525 by the event occurrence 510 in the event waiting state 501, 1
If the processing is completed 530 within the specified time 560 due to the occurrence of the event twice, it is regarded as normal. But,
If a new event 515 occurs before the processing is completed 530, the processing program returns to the event waiting state 50.
The processing starts immediately and continues to operate without being 1, but such a case could not be recognized as abnormal.

In Japanese Patent Application Laid-Open No. 5-108418, it is necessary to terminate a program in response to one request, and to wait for an event without terminating the program and execute a process corresponding to the event. Abnormality cannot be detected for programs that are repeated.

[0005]

In the conventional abnormality detection method, the time-out check range is a check to determine whether the time from the start of processing 525 corresponding to the event to the completion 530 of the event processing is not completed within the specified time 560. Although it was possible to detect the in-loop state of the processing program for one request, as shown in FIG. 5, if the event 515 occurs continuously before the processing for one event 510 is completed 530, the processing program There was no way to detect the continuous execution status.

An object of the present invention is to use a dedicated hardware such as a watchdog timer or the like, and continuously generate an event 615 before the processing for one event 610 is completed 630 as shown in FIG. In this case, the time during which the processing program continues to be executed is set as a time-out check range 620, and whether the time is within a specified time 660 is checked. To provide a method for detecting an abnormal state of a processing program.

[0007]

In order to achieve the above-mentioned object, according to the present invention, there is provided a method for detecting an abnormality in a processing program, the method comprising the steps of: It has a process management table for each processing program, has a start time entry for storing the transition time to the execution state of the processing program in the process management table, stores the current time in the start time entry when the processing program is initially started, When the processing program transitions from the execution state to the event waiting state, clears the start time entry to 0, stores the event occurrence when an event occurs, and transitions from the event waiting state to the execution state due to the event occurrence The current time in the start time entry That.

The monitoring process polls the start time entry in the process management table at regular intervals, finds the difference between the time stored in the start time entry and the current time. The processing program is determined to be abnormal and outputs an alarm.

Another method for detecting an abnormal state of a processing program according to the present invention has a process management table for each activated processing program, an execution state counter for storing the execution state of the processing program in the process management table, and a processing program. When the processing program transits from the execution state to the event waiting state, the execution state counter is cleared to 0, and when an event occurs, the occurrence of the event is stored. When a transition from the event waiting state to the execution state occurs due to occurrence, 1 is stored in the execution state counter.

The monitoring process polls the execution status counter in the process management table at regular intervals. If the execution status counter is not 0, the counter is incremented.
When the execution state counter exceeds a specified value as a result of the increment, the processing program is determined to be abnormal and an alarm is output.

[0011]

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

FIG. 1 is a block diagram of a processing program abnormality detection method based on processing program start time information according to claim 1. In the operating system 100 shown in the figure, there is a process management table 150 secured for each program. In the present invention, a start time entry 151 for storing the execution start time of a process is provided in the process management table 150. I have. Further, an event flag entry 152 for storing flag information indicating occurrence of an event when the event occurs is provided.

In a program activation process 110 that operates when the processing program is activated, the current time storage means 11
According to 0-1, the execution start time is stored 151-c in the start time entry 151 of the process management table 150, and the processing program 101 is executed 110-2. When the processing program 101 notifies the operating system of the event wait 101-1 to the operating system 100-b, the event wait process 120 first determines the event flag information 120-1 and 152-a of the process management table 150, and if the flag is ON. If the event has already occurred, the event flag is set to OFF 120-5 and 152 without entering the waiting state.
-B and return to the processing program in the execution state 100-
c, if the flag is OFF, since no event has occurred, the start time clearing means 120-2 stores 0 in the start time entry of the process management table 151-b
Then, it becomes the event waiting 120-3.

In the event receiving process 130 which is activated by the occurrence 160 of the event in which the processing program 101 is waiting, the event flags are set to ON 130-1, 152.
-C, and an event waiting release 130-2 is performed. Thereby, the event wait 120-3 of the event wait processing 120 is performed.
Is released, and the current time is stored in the start time entry of the process management table by the current time storage means 120-4.
2-b, and return control to the processing program 100-c. In the processing program 101, processing 101 corresponding to the event
Perform -2. When the processing is completed, the event wait 101 is resumed.
-1 is notified 100-b to the operating system.

As a result, while the processing program 101 is waiting for an event, the content of the start time entry 151 of the process management table 150 becomes 0, and the processing program 10
When a loop occurs in the event handling process 101-2 of 1
When the event occurrences 160 continue, the content of the start time entry 151 of the process management table 150 is the time when the event wait was last released.

On the other hand, in the fixed period monitoring process 140 that operates at a fixed period, the contents of the start time entry in which the contents of the start time entry in the process management table are other than 0 (that is, not in the event waiting state) are extracted.
1, 151-d The difference from the current time is calculated 140-2. The difference calculated here is the time that the program has been in the execution state since the start of execution or the release of the wait, and the time is compared with the specified time.
-3, and if it exceeds, consider an abnormal state and output an alarm 140-4.

FIG. 2 is a block diagram of a processing program abnormality detection system using an execution state counter according to a second embodiment.
In the operating system 200 shown in the figure, there is a process management table 250 secured for each program, and in the present invention, the process management table 250 counts up at regular intervals after the processing program enters the execution state. An execution state counter 251 is provided.
Further, an event flag entry 252 for storing flag information indicating the occurrence of an event when an event has occurred is provided.

In the program starting process 210 that operates when the processing program is started, 1 is stored in the execution state counter 251 of the process management table 250.
Then, the processing program is executed 210-2. When the processing program 201 notifies the operating system of the event wait 201-b 200-b, the event wait processing 220 first determines the event flag information of the process management table 250 by 220-1 and 252.
-A, if the flag is ON, an event has already occurred and the event flag is set to OFF2 without waiting.
20-5, 252-b, return to the processing program in the execution state 200-c, and if the flag is OFF, since no event has occurred, 0 is stored in the execution state counter 251 of the process management table 250. 2,251-
Then, the process waits for an event 220-3.

In the event receiving process 230, which is executed by the processing program 201 upon occurrence of the waiting event 260, the event flags are set to ON 230-1, 252.
-C, and then cancel the event waiting 230-2. Thereby, the event waiting 220-3 of the event waiting process 220
Is released, and 1 is stored in the execution state counter 251 of the process management table 220-4, 251-a, then the event flag is turned OFF 220-5, 252-b, and control is returned to the processing program 200-c. The processing program 201 performs processing 201-2 corresponding to the event. When the processing is completed, the event waiting 201-1 is notified 200-b to the operating system again.

As a result, while the processing program 201 is waiting for an event, the content of the execution state counter 251 of the process management table 250 becomes 0, and the processing program 20
When a loop occurs in the event handling process 201-2 of 1,
When the event occurrences 260 continue, 0 is not set in the execution state counter 251 of the process management table 250.

On the other hand, in the fixed period monitoring process 240 operating at a fixed period, the contents of the execution state counter 251 in the process management table are other than 0 (that is, the contents of the execution state counter 251 are not in the event waiting state).
-1,251-d, add 1 to the extracted value 240
-2, and store 251-e in the execution state counter. It is determined 240-3 whether the value of the execution state counter to which 1 has been added does not exceed a specified value. If the execution state counter exceeds the specified value, it is determined that the processing program is in an abnormal state, and an alarm is generated. Output 240-4.

FIG. 3 shows a detailed flow of the event waiting process of the processing program abnormality detection system by the execution status counter according to the second embodiment. The event waiting process 300 shown in the figure first determines whether an event has already occurred based on the event flag information in the process management table.
If the event has occurred (flag is ON), the event flag is turned OFF 300-5 without changing the execution state counter of the process management table, and the process returns to the processing program 300-6. If no event has occurred (the flag is in the OFF state), the execution state counter of the process management table is set to 0 (300-2), and the process transitions to event waiting 300-3.

When an event occurs, event reception processing 3
At 50, an event flag indicating that an event has occurred is turned ON 350-1, and if the processing program is in an event waiting state, the event waiting is canceled 350-2. When the event wait is released, the execution state counter of the process management table is set to 1 (300-4), the event flag is turned off (300-3), and control is returned to the processing program (300-6). In this way, when events occur continuously, the execution state counter is not cleared to 0, and the execution state counter determines whether or not the processing program continues to operate. It is possible to make a judgment.

FIG. 4 shows a detailed flow of the periodic monitoring processing of the processing program abnormality detection method by the execution state counter according to the second aspect. Periodic monitoring processing 4 shown in the figure
00 has a loop structure for searching the entire process management table. First, the head of the process management table is set 400-6 as a search target. Next, a determination 400-1 is made as to whether all have been searched. When the search is completed, the periodic monitoring process ends. If the search is not completed, it is determined 400-2 whether the content of the execution status counter of the process management table to be searched is other than 0, and if it is 0, the program is determined to be in the waiting state. Is updated 400-7, and the process returns to the search completion determination processing 400-1 of the process management table.

If it is not 0, it is determined that the program is in the execution state, and 1 is added to the execution state counter.
-3, it is determined whether the value does not exceed the specified value 400
-4. If it is within the specified value, it is determined that the state is normal, and the search target of the process management table is updated.
7. Search completion judgment processing 40 of the process management table
Return to 0-1.

If the specified value is exceeded, it is determined that the state is abnormal, and an alarm is output 400-5. Thereafter, the search target of the process management table is updated 400-
7. Search completion judgment processing 40 of the process management table
Return to 0-1. This makes it possible to detect a processing program in an abnormal state on a process basis.

[0027]

According to the present invention, even if an event occurs continuously within the processing time for an event in a processing program without using dedicated hardware such as a watchdog timer,
By monitoring how long the processing program has been running,
An abnormal processing program can be easily detected.

[Brief description of the drawings]

FIG. 1 is a configuration diagram of a processing program abnormality detection method based on a start time according to the present invention.

FIG. 2 is a schematic configuration diagram of a processing program abnormality detection method using an execution state counter according to the present invention.

FIG. 3 is a detailed flowchart of an event waiting process according to the present invention.

FIG. 4 is a detailed flowchart of a periodic monitoring process according to the present invention.

FIG. 5 is a time chart at the time of continuous occurrence of an event by conventional monitoring.

FIG. 6 is a time chart when events occur continuously according to the present invention.

[Explanation of symbols]

100, 200 ... operating system, 100-
a, 100-b, 100-c, 200-a, 200-b,
200-c: Interface with operating system, 101, 201: Processing program, 110, 21
0: program start processing, 120, 220, 300: event waiting processing, 130, 230: event reception processing,
140, 240, 400 ... fixed period monitoring processing, 150, 2
50: Process management table, 151: Start time entry, 151-a, 151-b, 151-d, 153-1
... Start time entry operation, 152,252 ... Event flag, 152-a, 152-b, 152-c, 252
a, 252-b, 252-c ... event flag operation, 1
60, 260, 510, 515, 610, 615 ... event occurrence, 251 ... execution state counter, 251-a, 2
51-b, 251-c, 251-d, 251-e: execution state counter operation; 500, 600: event execution processing state; 501, 601: event waiting state;
0,620: Timeout check range, 525, 53
0,630: Completion of processing for event 540,640
... Start of processing corresponding to the event, 560, 660... Specified time, 650.

 ──────────────────────────────────────────────────続 き Continuing on the front page (72) Inventor Saburo Iijima 5-2-1 Omikacho, Hitachi City, Ibaraki Prefecture Inside Hitachi Process Computer Engineering Co., Ltd. (72) Ryoyoshi Yoshizawa 5-chome Omikacho, Hitachi City, Ibaraki Prefecture No. 1 Inside Hitachi, Ltd. Omika Plant (72) Inventor Teruyuki Yasunaga 5-2-1 Omika-cho, Hitachi City, Ibaraki Prefecture Hitachi Process Computer Engineering Co., Ltd. (72) Inventor Tadashi Kojima Omika, Hitachi City, Ibaraki Prefecture 5-2-1, Machimachi, Hitachi Process Computer Engineering Co., Ltd. F-term (reference) 5B042 GA23 JJ08 JJ13 JJ25 5B098 GC03 JJ01

Claims (2)

[Claims] 1. An operating system for executing a plurality of processing programs according to an execution priority,
The operating system has a process management table for each started processing program, has a start time entry in the process management table for storing a transition time to the execution state of the processing program, and sets the current time at the initial startup of the processing program. Means for storing the event in the start time entry, means for storing the occurrence of the event, and means for determining whether the event has already occurred when the processing program transitions from the execution state to the event waiting state. If so, the process returns to the processing program, and means for clearing the start time entry to 0 when no event has occurred, and storing the current time in the start time entry when transitioning from the event waiting state to the execution state due to the occurrence of the event. And a monitoring process that operates at regular intervals. Calculating a difference between the start time entry in the process management table and the current time, and when a predetermined time has elapsed, determining that the processing program is abnormal and outputting an alarm. Processing program abnormal state detection method to be executed. 2. An operating system for executing a plurality of processing programs according to an execution priority.
The operating system has a process management table for each activated processing program, has an execution status counter for storing the execution status of the processing program in the process management table, and stores 1 in the execution status counter when the processing program is initially activated. Means for storing the occurrence of an event, and means for determining whether an event has already occurred when the processing program transitions from the execution state to the event waiting state. Returning means for clearing the execution state counter to 0 when no event has occurred, means for storing 1 in the execution state counter when transitioning from the event waiting state to the execution state due to the occurrence of the event, An operation monitoring process is provided. In the monitoring process, the process management table is used. Means for incrementing the execution state counter when the execution state counter is not 0 and means for outputting an alarm when the execution state counter exceeds a prescribed value as a result of the increment and determining that the processing program is abnormal. An abnormal state detection method for a processing program.