KR860000793B1 - Initial program loading method - Google Patents

Abstract

No content.

Description

Initial program rod method

1 is a block diagram showing a general configuration of an electronic calculator system.

2 is a flowchart art when a hardware error occurs during program processing.

3 is a flowchart of an embodiment of the present invention.

4 is a schematic configuration diagram of a service processor.

The present invention relates to an automatic initial program load (hereinafter referred to as IPL) method in a full calculator system.

In general, the computer system has a configuration as shown in FIG. In Fig. 1, reference numeral 1 denotes a central processing unit, numeral 2 denotes an input field processor, numeral 3 a memory bus, and numeral 4 a main memory. Input and output equipment not shown in the drawing is enclosed with the extrusion force processing apparatus 2. (5) is a bus and (6) is a service processor (SVP). The bus 5 is used by the service processor 6 to start the central processing unit 1 or read its state. Denoted at 7 is a dispenser device managed with the service processor 6, and 8 is a flexible disk device for avoiding information.

Next, the operation of FIG. 1 will be described with reference to FIG.

In general, in the electronic calculator system, a case where the hardware error B occurs while the program processing A is being executed. In this case, when the central processing unit 1 is operable, it is necessary to perform the retry C of the minor loop at the hardware level or to request the software to execute a larger rotation intra C so as to avoid the system stop. Try to continue. However, if more than this is a fixed failure of the device or a pure failure or external noise longer than the period of the retry C, the operation of the central processing unit 1 becomes difficult, and congestion worsens the system. Decide yourself so as not to give a warning (check stop D).

However, when the check stop D is reached, the system in which the system takes precedence is often not only IPL (including initialization of charge hardware) by human hands, but also returns unless the main circuit is fixed. There are not a few cases.

In a system in which the system is not allowed to stop, especially in unmanned operation, it takes a long time to return, and there is a disadvantage that the operation rate of the system worsens.

The present invention has been made to eliminate the drawbacks of such conventional ones. In addition to the central processing unit, a service processor is installed to monitor the status of the central processing unit at all times, and when the check stop D is reached, the service processor attempts to continue the system by executing an IPL instead of a person, if specified. Doing.

That is, if the service processor repeatedly executes the IPL operation for a predetermined number of times and still does not recover, the service processor is set to check stop.

Next, an embodiment of the present invention will be described with reference to the flowchart art of FIG. 3, FIG. 1, and FIG. First, the function of the service processor 6 will be described with reference to FIG. The service processor 6 mainly includes the state monitoring mechanism of the calculator system, the function of the interface of the man machine, the start stop function of the system, and the contention control function of the memory bus.

In detail, these functions indicate whether the calculator system is in the RUN state, the program stop state, the check stop state, etc., and ② indicates the read / write function of the memory, CPU, It functions as a register of an IOP, inputs to a key board, and display of data. ③ is a function to control the system start stop by instructions such as StarArt / Stop / Reset / IPL from key board, and ④ is a function to control the contention of CPU and IOP memory bus.

The service processor according to the present invention has a function of a check IPL described later, and includes a check IPL switch 10 for performing the check IPL operation and a counter 11 for counting how many times the check IPL has been performed.

Next, the operation will be described.

The abnormality B occurs during the program processing A, and the retry C is attempted by the hardware and software as in the prior art, but also fails. As a result, the central processing unit 1 cannot recover by itself and stops checking. It reaches D. Next, the service processor 6, which always monitors the state of the central processing unit 1, detects that the central processing unit 1 has been checked and stopped. In addition, the service processor 6, when the switch 10, which can be assigned to the system, is set to the check IPL (referred to as the check IPL for resetting from the power return) command E, it is possible to provide information for later cause-cause investigation. The disk device 8 is evacuated. The service processor 6 initializes the entire system including the central processing unit 1 to perform the IPL startup H after a suitable delay (meaning to confirm the real stop and, for example, to cut the time duration of the noise). do.

Besides that, the central processing unit 1 moves to the IPL operation, executes the initial program, and starts the operation of the system. At that time, the service processor 6 ends the operation of the central processing unit with the information that the IPL is in the check stop D state, and enables processing suitable for the check IPL.

Of course, in the case of the fixed failure of the main circuit elements, the CPU does not return after the check processor D starts by the service processor 6 several times.

In this case, in order to avoid the defect that IPL is permanently repeated, the service processor 6 installs a counter 11 that counts the number of check IPLs, and enters a system stop when a certain counter (for example, 16) is reached. have. And although the said embodiment assumed the single central processing unit 1, even if it is a some central processing unit 1, this invention is applicable. In this case, even if one central processing unit 1 stops due to a fixed failure, the service processor 6 can activate the IP to the other central processing unit 1 to reduce the total stop of the system as much as possible.

As described above, according to the present invention, in addition to the conventional IPL, an SVP having a function of activating the IPL at check stop is added, and no special hardware needs to be added. In addition, the present invention has an effect that can greatly reduce the probability of system stop without building a complex redundant system.

Claims (2)

Apart from the central processing unit and input / output processing unit, a service processor is installed that can actively monitor the entire system, interface with the perlator, and perform regular data retrieval. An error occurs during operation of the calculator system. Even if the electronic calculator is stopped, if the automatic return designation is specified, the system state at the time of stopping by the service processor is retracted, and after a predetermined time, an initial program code is applied to the central processing unit and the initial program is executed for a prescribed number of times. Initial program load method, characterized in that the check stop state only when the code is not recovered even when the code is executed. An initial program rod method according to claim 1, comprising a switch for specifying an automatic return by the initial program rod and a counter for counting the number of initial program rods in the service fatigue processor.

Images