JPS61112251A - Runaway detecting system - Google Patents

Abstract

PURPOSE:To detect quickly a defect of an RAM by having a jump to a prescribed address within an RAM area when an interruption is produced and executing interruption processing before a watchdog timer completes its counting action. CONSTITUTION:When a CPU has an interruption (110), a jump is given to a prescribed address within an RAM area before the interruption processing is carried out. A jump instruction to be given to another address with the RAM area is already set at the jump destination. The control is shifted (130) finally to the interruption processing after such RAM jump processing 120 is repeated several times. A normal state 135 is discriminated with the RAM if the interruption processing is reset. Otherwise a runaway signal is delivered 140 to discriminate a defective state 140 of the RAM.

Description

[Detailed Description of the Invention] [Technical Field of the Invention] The present invention provides a method for quickly detecting a RAM defect in order to reduce program runaway caused by a RAM defect when a program accesses a RAM area. This paper relates to a runaway detection method.

[Technical Background of the Invention] The main memory of a computer, for example in a microcomputer, is usually composed of a ROM that stores programs and a RAM that temporarily stores information. R.O.
The program stored in M may run out of control due to a partial defect in the ROM or various other causes, and may not perform the prescribed operation.

If a computer goes out of control, there is a risk of destroying important information stored in, for example, RAM, so it is necessary to detect the computer out of control as quickly as possible and take necessary measures.

Conventionally, in order to detect such runaway of a computer, a watchdog timer is provided, and this watchdog timer is reset at a predetermined time interval or less using a predetermined command of a program stored in a ROM to determine if the computer is operating normally. In some cases, the count time of the watchdog timer is always set to be less than or equal to a predetermined time. For example, if the computer goes out of control, the reset operation for the watchdog timer is not performed at a predetermined time interval or less, or is not performed at all, so that the counting time of the watchdog timer exceeds the predetermined time,
A signal identifying that a runaway has been detected is output from the watchdog timer.゛By the way, one of the reasons why a computer runs out of control is due to a defect in the RAM.
For example, this may be due to the jump destination address stored in RAM being destroyed.However, in the conventional method for detecting computer runaway as described above, such RAM failures cannot be detected. There are few systems that take into account the destruction of RAM contents caused by such problems, and there is a risk that computers will frequently run out of control due to this, and important information may be destroyed.

[Object of the Invention] The object of the present invention has been made in view of the above, and the object thereof is to provide a runaway detection method that promptly detects a defective RAM and reduces runaway of a computer. It's about doing.

[Summary of the Invention] In order to achieve the above object, the present invention has a ROM that stores a program and a RAM that temporarily stores information, and that is reset at predetermined time intervals or less by the program stored in the ROM. In a runaway detection method that determines that a program runaway has occurred when the count time of a watchdog timer reaches a predetermined time, when an interrupt occurs, the watchdog timer jumps to a predetermined address in the RAM area. The gist is to perform a predetermined process in the RAM area before the counting time reaches the predetermined time, detect that a predetermined processing condition is satisfied, and then start the interrupt process.

[Embodiments of the invention] Embodiments of the present invention will be described below with reference to the drawings.

FIG. 1 shows an example of a runaway detection device for implementing the runaway detection method of the present invention. In the same figure, CP
A ROM 3 that stores programs and a plurality of RAMs 5 that temporarily store information are connected to U1. In addition, CPU1 has a watchdog timer (WD lower) 7 is +
It is continued.

This watchdog timer 7, as shown in FIG.
The cpui is set by a reset signal Sr formed from a predetermined command read from a program stored in the ROM 3.

This predetermined instruction is suitably stored in the program and is adapted to reset the watchdog timer 7 at or below a predetermined time interval. Therefore, watchdog timer 7
The counting time TO is always less than the predetermined time Tr. However, due to, for example, a defective RAM, the CP
LI 1 is supplied with incorrect information or instructions and the cpui
If the clock runs out of control, the predetermined command is not read properly to reset the watchdog timer 7 at a time less than or equal to the predetermined time interval, so that the count time TO of the watchdog timer exceeds the predetermined time Tr. I end up. When the count time TO exceeds a predetermined time Tr, the watchdog timer 7 outputs a runaway detection signal Sd to the CPU 1 to reset the CPU 1 and stop the runaway.

FIG. 3 is a diagram for explaining the processing operation when trying to detect a defective RAM by using such a method of detecting a runaway of the CPU 1 using the double dog timer 7. This processing operation will be explained with reference to the flowchart in FIG. First, when the CPU 1 receives an interrupt (step 110), immediately before executing the interrupt operation, the CPU 1 executes an interrupt to address A in the RAM area, as shown in FIG. This address A stores an instruction to jump to address B in the RAM area, and this instruction similarly jumps to address B in the RAM area.
Jump to. Such a jump operation is repeated several times, for example, N times (N≧1) within the RAM area (
After step 120), the process jumps from the last address C to a predetermined address ITC for executing the above-mentioned interrupt processing, and executes this pruning operation. Note that the jump operation time in the RAM area from jumping into the RAM area to leaving it includes the time from when the watchdog timer was reset just before and started counting until jumping into the RAM area. However, it is necessary to set the number of jump instructions in the RAM area so that it is less than or equal to the predetermined time Tr.

In this way, even if the jump operation is repeated several times in the RAM area, that is, even if the jump operation is repeated, the final result will be detected within the predetermined time Tr before the watchdog timer outputs the runaway detection signal 3d. If it is possible to return to the predetermined address ITC for predistributed interrupts, it can be determined that the RAM is normal (
Steps 130, 135>. However, if there is a defect in the RAM and each jump operation stored in the RAM or its jump destination address is destroyed, the watchdog timer outputs the runaway detection signal Sd to the address ITC. Since it is not possible to return, the runaway detection signal Sd is output from the watchdog timer, and it is possible to detect that there is a defect in the RAM (steps 140 and 150).

Note that although the above embodiment describes the case of an interrupt, the case is not limited to this, and the same applies to the case of a fixed-cycle task.
You may also be able to return to the beginning of the task after jumping within the area.

Further, in the above embodiment, the jump operation was explained as an operation in the RAM 1 rear, but the jump operation is not limited to a plurality of times, but may be performed once. Furthermore, in addition to jump operations, for example, a predetermined arithmetic operation is performed, the result is checked, and if the result is correct, the return is made to the predetermined address for the interrupt, and if the result is not correct, the watchdog The runaway detection signal 3d may be output from the timer.

[Effects of the Invention] As explained above, according to the present invention, when an interrupt occurs, a jump is made to a predetermined address in the RAM area, and the RAM is read before the watchdog timer generates the runaway detection signal. Since the interrupt processing is avoided after the predetermined processing is performed in the area, the predetermined processing can be accomplished even if the contents of the RAM are destroyed due to a defect in the RAM. Therefore, a runaway detection signal is generated from the watchdog timer, and it is possible to identify that the RAM is defective, thereby reducing the running of the computer.

[Brief explanation of drawings]

FIG. 1 is a block diagram showing an example of a runaway detection device that implements the runaway detection method of the present invention, FIG. 2 is an operation waveform diagram of a watchdog timer used in the device shown in FIG. 1, and FIG. a diagram illustrating the location and flow of a jump instruction within a memory area illustrating the operation of the device shown in the figure;
FIG. 4 is a flowchart showing the operation of the apparatus shown in FIG. 1...CPU, 3...ROM, 5...RAM, 7
...Watchdog timer.

Claims (2)

[Claims] (1) It has a ROM that stores programs and a RAM that temporarily stores information, and the counting time of a watchdog timer that is reset at a predetermined time interval or less by the program stored in the ROM reaches a predetermined time. In a runaway detection method that sometimes determines that a program has runaway, when an interrupt occurs, the program jumps to a predetermined address in the RAM area, and before the count time of the watchdog timer reaches the predetermined time,
A runaway detection method characterized in that a preprogrammed process is executed within an area, and the interrupt process is started upon detecting that the process result satisfies a predetermined condition. (2) In the runaway detection method according to claim 1, the runaway detection is characterized in that at least one jump operation is executed in the RAM area as a process programmed in advance in the RAM area. method.