JPS6084651A - Control system of input and output device - Google Patents

Abstract

PURPOSE:To cope with the fault occurrence of an input/output device effectively by providing an input/output controller and a fault counting part between a host device and the input/output device. CONSTITUTION:A central processing unit CPU is connected to plural input/output devices I/O through the input/output controller IOC. The input/output controller IOC has a polling device PL and a control processor muP. The fault monitoring device EDT in the input/output controller IOC generates an error state when informed of fault occurrence and sends it to a fault value counting part ECN. The fault value counting part ECN consists of memories MM-A- MM-C, a large/small judging circuit CP, and a +1 adding circuit AD. The fault occurrence information from the fault monitoring device EDT is summed up in the memory MM-A successively and when the cumulated value attains to a specific value, a fault information signal is sent out to the central processing unit CPU to send out a control stop signal to the input/output controller IOC.

Description

DETAILED DESCRIPTION OF THE INVENTION (1) Technical Field of the Invention The present invention relates to a method for controlling an input/output device in an information processing system that can effectively deal with the occurrence of a failure in the input/output device.

2) Prior art and problems Conventionally, when multiple input/output devices are controlled by a host device such as a central processing unit, a failure occurring in an input/output device is immediately reported to the central processing unit, and the central processing unit collects the status. conduct. An interrupt is generated to the central processing unit, and an operation called logging based on the error status begins. This method is simple, but if a failure persists in the input/output device, the central processing unit will only perform logging processing and will not be able to process information because it will generate interrupts, for example, 5 to 6 times per second. Become.

(3) Purpose of the Invention The purpose of the present invention is to provide a control method for input/output devices in an information processing system that improves the above-mentioned drawbacks and makes it possible to effectively deal with failures in the input/output devices. .

(4) Structure of the Invention The structure of the present invention to achieve the above-mentioned object is that a plurality of input/output devices are connected to a higher-level device, and a failure occurring in an input/output device is notified to the higher-level device and logically disconnected. In an input/output device control method, an input/output control device and a fault counting unit are provided between the host device and the input/output device, and the input/output control device monitors the status of the input/output device by polling and detects a fault when a fault occurs. It is equipped with a fault monitoring device that sends the notification to the fault value counting unit, and a means for logically disconnecting the faulty input/output device based on the notification from the fault value counting unit and automatically reconnecting it when the cause of the fault is removed. , the fault value counting unit is equipped with means for receiving fault notifications from each input/output device, and means for notifying the fault monitoring device when a fault has occurred in the same person's output device a predetermined number of times or more; That's true.

(5) Embodiment of the invention FIG. 1 is a diagram showing the configuration of an embodiment of the invention, in which a central processing unit CPU is connected to a plurality of input/output devices I/O via an input/output control device OC. ing.

The input/output control device IOC includes a polling device PL and a control processor μP, and inquires of the input/output device I/O as to the presence or absence of transmitted/received data by specifying an I/O number in a well-known manner. If a failure occurs within an I/O, that I/O will notify the IOC when polled. The IOC is equipped with a fault monitoring device ECN, and when the above-mentioned fault occurrence response is received, the "error status" is displayed.
is created and notified to the fault value counting unit ECN. Since there are some failures that can be automatically recovered in a short time, the μP repeatedly polls the I/O from the polling device PL.

When the failure continues, notification to the failure value counter ECN continues. The failure value counting unit ECN includes memories MM-A, MM-B, and MM-.
C, a size discrimination circuit CP, and a +1 addition circuit AD. The signal a1 from the fault monitoring device EDT is a fault occurrence notification and is stored in a predetermined area of the memory MM-A corresponding to the input/output device. Memory MM-C is used to store an initial value of zero in memory MM-A. A predetermined value is set in advance in the memory MM-B, and a magnitude determination circuit CP determines whether the stored value in the memory MM-A exceeds the preset value. If the value is not exceeded and the signal a1 from the fault monitoring device EDT is repeated, the value stored in the corresponding area of the memory MM-A is updated by the +1 addition circuit AD. If the judgment of the magnitude judgment circuit CP exceeds the limit, it is notified to the central processing unit CPU as an abnormal failure notification signal a2, and to the input/output control device IOC as a control stop signal b of the input/output device. Inside the IOC, there is an element such as a gate GT that logically connects and disconnects the I/O, so when signal b is notified, the GT is closed and the I/O that has been notified of the failure is logically disconnected. Separate to. The CPU performs logging only once after being notified of the occurrence of an I/O failure. Next, the IOC μP repeatedly polls the I/O from the polling device PL and monitors the response. That is, even if normal data is transmitted from the I/O, it is not accepted and only information regarding the failure is transmitted and received. When the cause of the failure is removed manually or automatically, the I/O notifies the old normal value in response to polling. The IOC notifies the failure value counter ECN of the failure release signal cl, resets the memory MM-A, and at the same time notifies the central processing unit CPU of I/O recovery with the signal c2, and starts the I/O by command. Make a request. The central processing unit performs I/O with a predetermined command.
Start up and return to the original state.

FIG. 2 shows an operation flowchart of the IOC information processing device that executes the above operations.

(6) Effects of the Invention In this way, according to the present invention, even if a central processing unit is connected to a large number of I/Os, it can operate only to log the failure status when an I/O failure occurs. The original operation of the central processing unit can be continued. Since most of the information regarding failures is processed by the IOC, the operating efficiency of the system is high.

[Brief explanation of drawings]

FIG. 1 is a diagram showing the configuration of an embodiment of the present invention, and FIG. 2 is a flowchart showing the operation of an information processing device included in an input/output control device that executes the operations shown in FIG. CPU...Central processing unit IOC...Input/output control device I/O...Input/output device PL...Polling device EDT...Fault monitoring device ECN...Fault value counting unit GT...Gate μP...191 people for information processor patents Fujitsu Limited agent Patent attorney Eitsugu Suzuki

Claims (1)

[Scope of Claim] In an input/output device control method in which a plurality of input/output devices are connected to a host device and a failure occurring in the input/output device is notified to the host and logically disconnected, the host device and the input/output device The input/output control device includes an input/output control device and a fault counting section, and the input/output control device includes a fault monitoring device that monitors the status of the input/output device by polling and sends a notification to the fault value counting section when a fault occurs. , means for logically disconnecting a faulty input/output device based on a notification from the fault value subunit and automatically reconnecting it when the cause of the fault has been removed; A control method for an input/output device, comprising: means for counting failure notifications; and means for notifying the failure monitoring device when a failure has occurred in the same input/output device a predetermined number of times or more.