JPS6015749A - In-line tmp system - Google Patents

Abstract

PURPOSE:To shorten the maintenance time and to close the fault affecting range into a local area by executing periodically a diagnosis program of a CPU processing system while the system is working. CONSTITUTION:A process 1 which is driven with a timer interruption phenomenon 4 is performed together with an action to execute an in-line TMP module 3. The task 1 is started at fixed time intervals designated optionally, and an in-line TMP action 2 is started from the task 1. This action is started in response to each CPU to execute the module 3 and to perform diagnosis of the CPU. When a fault of the CPU is detected from the result of diagnosis, a system stop alarm mechanism is started to execute a system stop processing 5. Thus the operation of the system is stopped by a specific system wait code.

Description

[Detailed Description of the Invention] [Field of Application of the Invention] The present invention is directed to a CPU U I! 1J (data garbled due to arithmetic errors due to faults in the Al threads, which can lead to data set destruction by periodically diagnosing and detecting faults during system operation.

[Background of the invention]

(Conventionally, diagnosis of the CPU control system was performed at a time separate from system operation, with dedicated maintenance time set aside. As a result, failures that occurred in the system could not be discovered until the next maintenance time.) It had the disadvantage of being late.

(2) The conventional input/output time monitoring function has a drawback in that it only checks for failures in input/output equipment, but does not check for failures in the CPU.

[Purpose of the invention]

The present invention enables a diagnostic program, which was conventionally carried out as part of maintenance work, to be carried out periodically during system operation.
The purpose is to detect CPU corruption at an early stage and minimize the impact on the system, such as data set destruction.

[Summary of the invention]

As a method for early implementation of CPU control systems, the conventional input/output time monitoring function can be replaced with input/output equipment in order to be able to execute diagnostic blocks that were previously performed during maintenance work while the system is running. This was done by combining these methods with the idea of extending the system not only to the CPU but also to the CPU.

[Embodiments of the invention]

An example of contamination according to the present invention will be explained with reference to the drawings.

There are a process 1 that runs in response to a timer interrupt event 4 and an action that executes an inline TMP module (diagnostics program) 3.

Start task 1 at arbitrarily specified fixed time intervals.

Invoke inline IllΔIP action 2 from this task. Inline T IVi P action is each C
Inline TM is activated for each CPU.
Execute P module 3 and diagnose C)'U. When an abnormality in the CPU is detected as a result of this diagnosis, a system stop warning mechanism is activated, and by executing system stop processing 5, the system is stopped with a special return system wait code.

The test contents for the inline TMP module are (i) Operation test of the four arithmetic operation instructions (2) Operation test of multiple combinations of multi-instructions (3) Test of correct operation by creating a register conflict state with the Di instruction (4) It is possible to carry out diagnostic tests that have conventionally been carried out using general diagnostic programs, such as checking the ordering of stores to storage devices. You can change the test content.

The inline 'I' MP method is a function that diagnoses the CPU by regularly executing a diagnostic blockrum, which was conventionally performed during maintenance work, using a timer interrupt as a trigger during system operation. Therefore, an abnormality is detected at time intervals activated by a timer and the system is immediately stopped, which has the effect of localizing the range of influence of the failure.

〔Effect of the invention〕

According to the present invention, the same diagnostic program as that for diagnosing the CPU control system during regular maintenance work is executed. Therefore, since failures can be detected at time intervals activated by the timer, there is an effect that the range of influence of failures can be localized.

Furthermore, since it is a practical implementation of a diagnostic program, it has the effect of shortening maintenance time.

In other words, since this is done while the system is running, there is no need to set aside dedicated maintenance time.

[Brief explanation of the drawing]

The figure is a processing diagram of an inline TMP method according to an embodiment of the present invention. 1... Process started by timer, 2... Process that executes inline TMP, 3... Inline TMP module, 4... Timer interrupt event, 5... System stop alarm mechanism sounds and system stops. process.

Claims (1)

[Claims] 1. Timer interrupts at constant time intervals that can be arbitrarily specified, and tasks activated by these, and 4! rCP
Actions to be executed for U number of machines and inspection for each model W1
In the module that combines the test instructions you want to use,
This is an inline TMP method characterized by regularly executing this module using the above-mentioned timer as a trigger.