JPH10307737A - Device and method for detecting runaway of cpu - Google Patents

Abstract

PROBLEM TO BE SOLVED: To speedily perform the runaway detection of CPU without depending on the structure of software. SOLUTION: In this device having a bus error exception processing function and processing software from plural tasks different in processing time, a watchdog timer(WDT) 1 starts counting synchronously with WDT reset showing the start of task processing and outputs a bus error signal when a count value exceeds a prescribed threshold value. Then, a bus error timer 2 starts counting synchronously with the bus error signal and outputs an alarm signal showing the runaway of software when a count value exceeds a prescribed threshold value.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

[0001] 1. Field of the Invention [0002] The present invention relates to an apparatus and a method for detecting runaway of a CPU.

[0002]

2. Description of the Related Art Conventionally, a watchdog timer used for detecting runaway of a CPU requires a predetermined time, before or after a task is executed, regardless of the length of a task processing time by a CPU. Alternatively, it is common practice to prioritize tasks and execute the reset with the task with the lowest priority, assuming that the task with the lowest priority is performed while the task with the highest priority is executed a predetermined number of times. .

[0003]

As described above, the conventional C
In the PU runaway detection device, the time constant of the watchdog timer is lower in execution frequency than the task with the shortest processing time, but must be set to the task with the longest processing time that is always executed at random. Even if the CPU goes out of control while executing a short task, the runaway of the CPU is detected after the elapse of the processing time of the task with the longest processing time.
There is a possibility that detection of runaway of the CPU may be delayed.

[0004] This is because the watchdog timer counts up or down according to a hardware set value, and when it reaches the set value, a timeout occurs without notification to the CPU and an alarm or reset of the CPU is executed. That is because In other words, as long as the software is operating normally, the watchdog bearer must avoid timeouts, so it runs less frequently than tasks that take less time,
It is necessary to set the timing for the task with the longest processing time, which is always executed at random. Therefore, the timing of CPU runaway detection depends on software.

SUMMARY OF THE INVENTION The present invention has been made in view of the above-described problems, and has a CPU capable of quickly detecting a runaway of a CPU without depending on software.
To provide a runaway detection device and method.

[0006]

In order to achieve the above object, as a means related to the apparatus of the present invention, a CPU having a bus error exception handling function and processing software from a plurality of tasks having different processing times. A runaway detection device, which starts counting in synchronization with a watchdog timer reset indicating the start of task processing, and outputs a bus error signal when the count value exceeds a predetermined threshold value; and a bus error signal. Starts counting in synchronization with
A means having a bus error timer for outputting an alarm signal indicating that U has run away is employed. Further, as a means related to the method of the present invention, a runaway detection method of a CPU having a bus error exception handling function and processing software from a plurality of tasks having different processing times, wherein a predetermined time has elapsed from the start of the task processing. Means is employed in which, when a clock is counted and a predetermined time elapses after the clocking, a runaway of the CPU is determined when the task processing is not started based on the bus error exception processing.

[0007]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS Hereinafter, an embodiment of an apparatus and method for detecting runaway of a CPU according to the present invention will be described with reference to the drawings.

[1] Description of Configuration First, the functional configuration of the present embodiment will be described with reference to the block diagram shown in FIG. In the figure, a CPU 8 has a bus error exception handling function and operates according to software stored in a ROM (read only memory) 9. This bus error exception processing is performed by the bus error factor storage unit 11 stored in the memory 10, the watchdog timer bus error signal S4 generated by the watchdog timer 1 via the address bus S1 and the data bus S2 of the CPU 8, and the CPU 8 The bus error signal S10 generated due to the transfer error of the bus cycle is identified, and in the case of the watchdog timer bus error signal S4, the watchdog timer 1 is reset.

The watchdog timer 1 comprises a counter 3, a bus error generator 4, and a bus cycle status monitor 5. The counter 3 can arbitrarily set a time constant. The count value is reset based on the watchdog timer reset S3 driven by the CPU 8 after the execution of the task, and the watchdog timer reset S3 is input. Instead, when the count value reaches the set value, a timeout signal S6 is output.

The bus cycle state monitor 5 monitors a bus cycle control signal S5 output from the CPU 8 and
If the bus cycle of U8 is being executed, a bus cycle execution signal S7 and a bus error signal S10 indicating a transfer error such as no bus response from an external device are output. The bus error generator 4 receives the timeout signal S6,
The bus error signal S4 is output when the bus cycle of the CPU 8 is being executed by the bus cycle execution signal S7.
The bus error generating unit 4 holds the output of the watchdog timer bus error signal S4 until the bus cycle of U8 is executed.

The bus error timer 2 comprises a counter 6 and an alarm generator 7. The counter 6 can arbitrarily set a time constant. When the watchdog timer bus error signal S4 output from the watchdog timer 1 is input, the counter 6 starts counting, and when the watchdog timer reset S3 is input. The timer is reset, the watchdog timer reset S3 is not input, and when the count value reaches the set value, a timeout signal S8 is output. The alarm generator 7 outputs a time-out signal S8
Is input, an alarm signal S9 is output to the outside.

[2] Description of Operation Next, the operation of the present embodiment will be described with reference to FIGS. These figures show the operation timing of the CPU 8 which outputs the watchdog timer reset at the task change point. Of these figures, FIG. 2 is a timing chart when the software operates normally, and FIG.
FIG. 4 is a timing chart when a PU runs away. FIG.
3A shows the execution state of the task in the CPU 8, FIG. 3B shows the transition of the count value of the watchdog timer 1, and FIG. 3C shows the transition of the count value of the bus error timer 2. .

First, the operation of the present embodiment when the CPU is operating normally will be described with reference to FIG. The tasks A to D of the CPU 8 executed during the interval from time t0 to t1 in FIG. 2B have a shorter processing time than the time set in the watchdog timer 1, and the count value of the watchdog timer 1 Since the task ends before the time reaches the timeout value of the watchdog timer 1, the watchdog timer 1 is reset after the task is executed, and no bus error exception handling occurs.

A task E executed from time t1
Since the processing time is long, the count value of the watchdog timer 1 reaches the timeout value at time t2, the bus error signal becomes active, and the bus error timer 2 starts counting as shown in FIG.

At this time, the CPU 8 suspends the task E, enters a bus error exception process, and
When it is recognized that the bus error has occurred, the watchdog timer 1 is reset at time t3, and the suspended task E
Resume. Similarly, the bus error timer 2 is stopped and reset because the watchdog timer 1 is reset at time t3.

Next, referring to FIG. 3, an operation when the CPU goes out of control will be described. As shown in FIG. 3B, the tasks A to D of the CPU 8 executed during the period from time ta to time tb are shorter in processing time than the time set in the watchdog timer 1, The task ends before the count value of the watchdog timer 1 reaches the timeout value of the watchdog timer 1, and the reset of the watchdog timer 1 is performed after the task execution, so that bus error exception processing does not occur.

For example, task C executed from time tb
Assuming that the CPU has runaway during the execution of the operation, the task does not shift from the duster C to the next task until time tc when the count value of the watchdog timer 1 reaches the time-out value, and The error signal becomes active, and the bus error timer 2 starts counting.

At this time, since the CPU 8 cannot shift from the task C to the bus error exception processing, the reset of the watchdog timer 1 is not executed, and the bus error timer 2 sets the timeout value of the bus error timer 2 at time td. Reach. As a result, the alarm signal S9 is activated, and the runaway of the software is notified to the upper card or the operator.

[0019]

As described above, the C according to the present invention is used.
According to the apparatus and method for detecting a runaway of a PU, a runaway detection apparatus for a CPU having a bus error exception handling function and processing software from a plurality of tasks having different processing times, comprising: A watchdog timer that starts counting in synchronization with the timer reset and outputs a bus error signal when the count value exceeds a predetermined threshold value, and starts counting in synchronization with the bus error signal and the count value reaches a predetermined threshold value If the value is exceeded, CP
Since a bus error timer for outputting an alarm signal indicating that U has runaway is provided, the threshold value for the count value of the watchdog timer can be adjusted to the processing time of frequently executed tasks. Irrespective of the CPU, runaway detection of the CPU can be performed quickly.

[Brief description of the drawings]

FIG. 1 is a block diagram showing a functional configuration of an embodiment of an apparatus and method for detecting runaway of a CPU according to the present invention.

FIG. 2 is a timing chart in a case where software runs normally in an apparatus and a method for detecting runaway of a CPU according to an embodiment of the present invention;

FIG. 3 is a timing chart when software runs away in the embodiment of the apparatus and method for detecting runaway of a CPU according to the present invention;

[Explanation of symbols]

 1 Watchdog timer 2 Bus error timer 3 Counter 4 Bus error generator 5 Bus cycle status monitor 6 Counter 7 Alarm generator 8 CPU 9 ROM 10 ... Memory 11 Bus address storage unit S1 Address bus S2 Data bus S3 Watchdog timer reset S4 Watchdog timer bus error signal S5 Bus cycle control signal S6 Time-out signal S7 ... Bus cycle execution signal S8 ... Timeout signal S9 ... Alarm signal S10 ... Bus error signal

Claims (2)

[Claims] 1. A runaway detecting device for a CPU having a bus error exception handling function and processing software from a plurality of tasks having different processing times, wherein a watchdog timer reset (S3) indicating the start of task processing is provided. A watchdog timer (1) that starts counting in synchronization and outputs a bus error signal (S10) when the count value exceeds a predetermined threshold value, starts counting in synchronization with the bus error signal, and A bus error timer (2) for outputting an alarm signal (S9) indicating that the CPU has runaway when a predetermined threshold value is exceeded; and a runaway detection device for the CPU. 2. A runaway detection method for a CPU having a bus error exception handling function and processing software from a plurality of tasks having different processing times, comprising: measuring a predetermined time from the start of task processing; A runaway monitoring method for a CPU, comprising determining a runaway of a CPU when task processing is not started based on a bus error exception process after a predetermined time has elapsed.