JPS60254338A - Abnormality detecting system of multiprocessor - Google Patents

Abstract

PURPOSE:To monitor a response time from a subprocessor by a counter and to detect a fault by providing a main processor with a counter to be started at the generation of a state informing request command from the main processor to the subprocessor. CONSTITUTION:The main processor MPU outputs a normal command to the subprocessor SPU and also outputs a state informing request command T so as to mix the command T in the normal command at a prescribed interval. At the output of the command T, the main processor MPU starts a timer counter. When receiving the command T, the subprocessor SPU returns a response T after a time Tans. Since a time Tout counted by the counter is set so as to be Tout>Tans when the subprocessor SPU operates normally, the counter is reset and the normal operation can be checked. If the response T is not returned withing the time Tout, the main processor MPU decides that the subprocessor SPU is failed.

Description

Detailed Description of the Invention [Field of Industrial Application] The present invention relates to a multiprocessor abnormality detection method capable of macroscopically detecting a failure occurring in a slave processor in a system composed of a master processor and a slave processor. .

Recently, a PO8 control processor (hereinafter referred to as SLP (Store Level Proc) used to control a POS (Point of 5ales) terminal device
essor). ) is composed of a general-purpose computer component (main processing unit (MPU)) and a sub-processing unit (SPU) that performs POS-specific control.

In such a configuration, when a failure occurs in the sub-processing device, it is necessary to be able to detect the failure in two ways as quickly as possible. This is the pos connected to it
This is a matter required to ensure normal operation of the terminal device.

The requirement is that macroscopic detection of failures occurring in sub-processing units is necessary in order to maintain normal operation of the entire system.

[Conventional technology]

In the past, failures occurring in the above-mentioned sub-processing devices have been detected at the hardware signal line level of the interface section provided between the main processing device and the sub-processing devices. That is, a response waiting timer is provided in the main memory to monitor the status of the sub-processing unit, and the timer is configured to respond to a response signal returned via a hardware signal line, and the response signal is set in the timer. If the request did not return within the specified time, it was determined that a failure had occurred in the subprocessing device.

However, as sub-processing devices have come to include processors, memories, etc., it has become insufficient to simply have a failure in the interface section as described above.

[Problem that the invention seeks to solve]

That is, when a sub-processing device is equipped with a processor or the like, it becomes necessary to perform macroscopic monitoring including its programs.

Therefore, in the past, a technique was used to monitor the time taken for all commands to complete.

Therefore, either the failure cannot be detected within a short period of time, or the monitoring interval cannot be freely set.

[Means for solving problems]

The present invention provides a multiprocessor abnormality detection method that can shorten the failure detection time and freely set the monitoring interval. In the system, a status notification request command for the sub-processing device is provided in a command issued from the main processing device to the sub-processing device, and a status notification request command for the sub-processing device is provided, The status notification request command is transferred mixed with the normal command of the sub-processing device, and the status notification response of the sub-processing device to the status notification request command is transferred to the sub-processing device within a predetermined time from the time of transfer of the status notification request command. When the sub-processing device is not found in various responses of the processing device, it is determined that an abnormality has occurred in the sub-processing device.

〔Example〕.

Hereinafter, the present invention will be described in detail with reference to the accompanying drawings.

FIG. 1 shows a system configuration for implementing the present invention.

In this figure, l is a processor SLP for controlling PO8, and this is a processor SLP for controlling PO8.
OS terminal devices 3t, 32, . . . are connected. The processor SLP includes a main processing unit Q p u and a sub-processing unit SP'U, both of which have a processor, memory, etc. therein. The processor SLP issues a normal command from the main processing unit MPU to the sub-processing unit SPU to request processing from the sub-processing unit SPU, and in response to the command, the sub-processing unit SPU polls the POS terminal device and sends a message to each PO8 terminal. It is configured to collect device data and return it to the main processing unit MPU in the form of a response, and is also configured to be able to issue a status notification request command from the main processing unit MPU in combination with the above-mentioned normal commands. The essential part of the present invention lies in what has been done. The main processing unit is provided with a counter that is activated when the status notification request command is issued, forming a characteristic part of the present invention.

With this configuration, the main processing unit MPU
For example, a normal command issued from a PO3 terminal device 3i (+ indicates multiple
Classify transaction data input from one of the s terminal devices),
After blocking (combining data into a certain amount of data), the sub-processing unit SPU, which is requested to send it back to the main processing unit MPU, processes the request, that is, polls the three POS terminal devices, and sends a certain amount of data. Even if the interval between the response collected and returned to the main processing unit MPU differs depending on the processing content, a delay in response due to such a difference in interval is normal processing in the sub-processing unit SPU, and The process takes a long time
It is possible to determine whether the problem is caused by any of the above, or whether the problem is caused by a failure occurring in the sub-processing unit SPU, which causes it to become unresponsive.

That is, according to the present invention, the main processing unit MPU issues normal commands to the sub-processing unit SPU, and also issues these jff
A status notification request command T is issued at predetermined intervals mixed with the i regular command (see FIG. 2).

When this status notification request command T is issued, the main processing bag W M
PU starts a timer. When the sub-processing unit SPU receives the command T, it returns a response T to the main processing unit after a time T ans. The time T ans required to return the response T is for comprehensively checking the inspection target including the software of the sub-processing unit SPU.

As long as the sub-processing unit SPU is operating normally, the time Tout measured by the counter is set so that Tout>Tan5, so the counter is always reset. Thereby, the normal operation of the sub-processing unit SPU can be confirmed.

Moreover, as shown in FIG. 2, the response T is the time To
If the response is not returned within ut, responses to other commands will not be returned at the same time, but since the response T that should be returned within time Tan5 is not returned to the main processing unit MPS, the sub-processing unit SPU can be determined to be a disability.

Since such judgment can be made within the shortest possible time, system environments where there is a possibility of losing a large amount of data if the abnormal character detection is delayed, such as the PO3 terminal using the main processing unit as described above, are This means that the functions of the present invention can be fully utilized in cases where it is necessary to accumulate data from devices.

By freely changing the interval at which the command T is issued, the intended effect can be best exhibited.

〔Effect of the invention〕

As explained above, according to the present invention, 1) the detection time of a failure occurring in the sub-processing device is shortened, and the harmful effects caused by the detection delay are eliminated as much as possible; Effects such as enhanced detection can be obtained by freely setting the detection interval.

[Brief explanation of drawings]

FIG. 1 is a diagram showing the system configuration for implementing the present invention, and FIG.
The figure is a diagram showing a signal system for explaining the present invention. In the figure, 1 is a PO8 control processor, 2 is a serial interface, 31. '32... is P. In the S terminal device, MPU is a main processing unit, and SPU is a sub-processing unit. Figure 1

Claims (1)

[Claims] In a system configured with a main processing unit and a sub-processing unit each having a processor, a status notification request command for the sub-processing unit is provided in a command issued from the main processing unit to the sub-processing unit. , transfer the status notification request command from the main processing device to the sub-processing device mixed with other normal commands, and respond to the status notification request command within a predetermined time from the time of transfer of the status notification request command. An abnormality detection method for a multiprocessor, characterized in that when a status notification response of the subprocessor is not found among various responses of the subprocessor, it is determined that an abnormality has occurred in the subprocessor. .