JPS6029983B2 - Failure diagnosis method for computer systems - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION The present invention relates to a fault diagnosis method in an electronic computer system.

Generally, a computer system has multiple data transfer devices (CH) connected to a central processing unit (CPU).
The number and physical location of this type of data transfer device differs depending on the system. It's on.

Therefore, when diagnosing a computer system using a diagnostic program that targets a system with the maximum configuration in which all data transfer devices are installed, if a certain data transfer device is not installed, the diagnosis Data scan-in/scan-out operations are not performed correctly, causing diagnostic problems. On the other hand, in diagnosis using a diagnostic program that targets only the central processing unit, the scope of diagnosis is limited to the central processing unit, and failures inside the data transfer device connected to the central processing unit cannot be detected. .

As a method to deal with these shortcomings, a diagnostic program that diagnoses only the central processing unit and the data transfer devices actually connected to the central processing unit is available, depending on the number and mounting positions of data transfer devices in the system. There is a method to prepare for each system. However, this method requires man-hours for system generation of the diagnostic program performed for each system, and man-hours required for modifying the diagnostic program caused by adding or discarding data transfer devices in each system. Furthermore, the complexity of managing each system becomes a problem. The present invention solves the problems of the prior art as described above, and can be applied to any implementation state of the data transfer device with a single diagnostic program. The present invention provides a diagnostic method that has the effect of preventing errors (such as incorrect indications, etc.) and reducing management complexity.

Simply put, the present invention configures a diagnostic program by a plurality of program modules each diagnosing a central processing unit and various devices that can be connected to the central processing unit. A device mounting status word is introduced to indicate whether each mounting position is actually mounted or not, and each module of the diagnostic program refers to the device mounting status word above before diagnosis. If the program module indicates that the device to be diagnosed by the program module is not installed, the module skips execution of the diagnosis and moves on to execution of the next module.

Next, one embodiment of the present invention will be described with reference to the drawings.

FIG. 1 shows a schematic configuration for realizing the diagnostic method of the present invention. In the figure, 10. ~107 are diagnostic target devices such as data transfer devices connected to the central processing unit 20';
, 107 are devices in which two central processing units are installed, and devices 102, 103, 104, and 106, indicated by dashed lines, are not installed. For convenience, the central processing unit 20 is shown in FIG.
The maximum number of devices that can be connected is eight, but of course this is just an example. 21 is the device 10;

~107 This is a device implementation status word register that indicates the installed/unimplemented status of the device. Each bit of this register has a one-to-one relationship with each device.
A bit corresponding to a certain device shows a logic value "1" when the device is installed, and shows a logic value "0" when the device is not installed. For example, one signal line is prepared between each bit in the device implementation status word register 21 and the corresponding device, and when the device is installed, the value is fixed to "-1" on the device side. If the device is not installed, the signal line may be characterized so that the corresponding signal line has a fixed value of "0". Furthermore, if there is no space for mounting and it is impossible to set up a signal line, an existing data bus or the like may be used. FIG. 2 shows the device 10. 107 shows the relationship between these devices and the device mounting state word register 21 when the devices are in the mounting state shown in FIG. 30 is a diagnostic program storage medium, and the device 10 is attached to this medium.

~107 diagnostic programs are stored divided into eight program modules #0 to #7, each of which targets each device for diagnosis. Similarly, a diagnostic program for the central processing unit 20 is also prepared in a modular configuration, but is omitted in FIG. Each module #0 of the diagnostic program
Information indicating which device each program module corresponds to is given to #7, for example, as a mask pattern for the device mounting status word. FIG. 3 shows a mask pattern given in advance to a program module that diagnoses the device 10o in FIG. ”. Now, when a failure or the like occurs in the system and diagnosis is started, the apparatus 10 reads the diagnostic program from the diagnostic program storage medium 30 in the order of, for example, modules #0 to #7 from the central processing unit 2 and makes the device 10 the object of diagnosis.

- Execute the diagnosis of 107. At this time, each module of the diagnostic program refers to the device implementation status register 21 before executing the diagnosis, and if the device implementation status word indicates that the device to be diagnosed by the corresponding module is not installed, the module is the target. Skips the diagnostic execution of the device and moves on to execution of the next module. FIG. 4 shows a flowchart in which each program module refers to the device implementation status word and determines whether or not to diagnose the corresponding device. In other words, each program module performs a logical OR between the device implementation status word and a mask pattern unique to each module given in advance, and if all bits are "1" as a result, the corresponding device is installed. The device is then diagnosed.
For example, if the contents of the device implementation status word register 21 are as shown in FIG. 2, and the mask pattern given to the program module # is as shown in FIG. 3, the result of the logical sum of the two is All bits are "1", and the module #0 corresponds to the corresponding device 10. It is determined that it is implemented. On the other hand, 10. is the device 102
, 103, 104, and 106, the contents of the program module 21 should be "0100010r. In this case, the logical sum with the mask pattern shown in FIG. 3 is "01111111," and the program module # 0 assumes that the corresponding device 10o is not installed, and moves the process to the next module #1.As described above, according to the present invention, a diagnostic program prepared for diagnosing an uninstalled device is executed. It is possible to prevent failures in diagnosis caused by this, and it is also possible to prevent a decrease in the diagnostic rate by targeting the entire device that makes up the system.Furthermore, the diagnostic program can automatically diagnose the installed device. Since the system skips to the next step, it is only necessary to prepare one type for the maximum configuration of the system, which reduces the amount of manpower required for management.

[Brief explanation of the drawing]

FIG. 1 is an overall configuration diagram of an embodiment of the present invention, and FIG.
FIG. 3 is a detailed diagram showing the relationship between the device to be diagnosed and the device mounting status register in the figure; FIG. 3 is a diagram showing an example of a mask pattern given to each module of the diagnostic program;
The figure is a diagram showing an example of the operation flow of a diagnostic program. 10. ~107...Diagnosis target device, 20...Central processing unit, 30...Diagnostic program storage medium, #0 to #7...
・Diagnostic program module. Figure 1 Figure 3 Figure 2 Figure 4

Claims (1)

[Claims] 1. In a computer system in which a plurality of devices such as data transfer devices are connected to a central processing unit, a plurality of diagnostic program modules each target a maximum number of devices that can be connected to the central processing unit. In addition to providing a module, information indicating whether each device to be diagnosed is actually installed or not is stored in a specific register in the central processing unit, and a certain diagnostic program module diagnoses the corresponding device. Before executing, the specific register is referred to, and if the contents indicate that the device to be diagnosed by the corresponding program module is not installed, the process is skipped to the next diagnostic program module. A fault diagnosis method characterized by: