JPS63129429A - Program abnormality processing system - Google Patents

Abstract

PURPOSE:To prevent compulsory interruption from being generated at the time of CT (Crises Time) when runaway happens, by providing an area in which a branching destination by every plural control programs at an abnormal time is stored and the area in which the state of the control program just before being branched is stored, in a table which manages the state of the control program. CONSTITUTION:In the table PIC (Program Index Chart) on an OS (Operating System) which manages the state of each program, the area RP in which the branching destination by every program at the abnormal time is stored, and the area PSW(S) in which the state of the program just before being branched is stored, are provided. In other words, the area 1 has a PI (Program Information), the area 2 the RP (Reentry Point), the area 3a PSW (Program Status Word), the area 4 an RP(S) (Saved Recentry Point), the area 5 the PSW(S) (Saved Program Status Word), and the area 6 other managing information. In such way, it is possible to perform the setting, the cancelation, and the movement of the branching destination when the program itself falls in the abnormality.

Description

DETAILED DESCRIPTION OF THE INVENTION (A) Industrial Application Field The present invention relates to recovery processing for programs in a multiprogramming system that have run into abnormal operations, and is directed to branching when each program itself has run into an abnormality. This invention relates to a program abnormality processing method that allows setting, cancellation, movement, etc. of a destination.

(b) Conventional technology Generally, in a multiprogramming system, multiple programs (UserProgram)
Although it appears that the programs corresponding to UP (hereinafter abbreviated as UP) are running in parallel, in reality, the programs controlling each Un (operating system
ng System (hereinafter abbreviated as O8) exists, and for each UP there are computer hardware such as CPU (Central Information Processing Unit) usage time, memory that physically stores data, input/output capo) ( Ilo), disks (files for storing data), etc., to achieve the maximum effect on the entire system? He is working to his full potential.

In a multiprogramming system in which a plurality of UPs are controlled by the O8, various methods have been proposed as countermeasures against abnormalities occurring in some UPs, that is, program abnormalities or runaways. An example thereof is JP-A-5'9-140568.

Problems that the e-1 invention attempts to solve In the conventional example described above, the following first-order problem occurs.

(2) A common countermeasure for O8 is to interrupt UP processing (referred to as abort), but it is difficult to recognize the individual circumstances of UP.

Therefore, the UP processing is interrupted (abor) as a UP abnormality.
Even if countermeasures other than t) are known, they are often interrupted uniformly, making it difficult to investigate the cause of the abnormality.

■ When interrupting an UP from the O8 side, the O8 also handles the return processing of hardware resources to eliminate subsequent conflicts, so the O8 must constantly monitor each UP itself.
This is one of the causes of the huge size of O8 and the wasteful use of hardware such as the CPU and memory.

■ The time when an interruption would cause a contradiction, for example, in the middle of a data chain connection (when Cr is in a difficult situation)
1ses Time (abbreviated as CT), and it is difficult to determine this time from O8.

Therefore, in order to prevent this, current systems are forced to perform special processing such as interrupt disabling processing.

Function The program abnormality handling method of the present invention provides multiprogramming and is capable of using a small-scale O8 in an office computer that operates on limited hardware.
) k can be realized.

(E) The present invention will be explained with reference to the drawings. Fig. 1 shows a program index chart (hereinafter abbreviated as PIC) employing the program abnormality processing method of the present invention, and Fig. 2 shows a computer using the same method. The relation diagram of O8 and UP in , Figure 3 is RP
E (Reentry Point Entry) and RT N (Return from Reentry
Figure 4 is a state diagram showing the movement of the PIC when a command is executed.

Next, to explain the operation of the method of the present invention, the PIC showing the table on O8 that manages the status of each program,
Area for storing the branch destination in the event of an error for each program (RP
) and the area that stores the state of the program immediately before branching (
PSW(S)) is provided.

In addition, Message Path (hereinafter referred to as MSG-P), which has existed for a long time as a means of transmitting information between UP members,
Message (hereinafter referred to as MSG) using
(abbreviated as ). (Figures 1 and 2) Area (1) in the figure is P I (Program Info
rmation ), (2) is RP (Reentry
Po1nt), (3) is PSW (Progr
amS status Word), (4) is RP
(S) (5avedRecentryPoint
), (5) is P S W (S) (5aved
Program3 tatus Word), (6)
is an area with other management data, (force is U P 1
, (8) indicates UP2.

First, upload the program where the error may occur.
I (7), and the program that restores the UPI (71) is U P 2 (8).

Also, the program U that you want to branch when an abnormality occurs
The address in P I (7) is RP (Reentr
yPoint). Next, the command issued in U P I (7) is shown below, and the flow is explained using Figure 3.
This is performed using P S (Reentry Point Set).

This command sets the address specified in RP (2) of PICI.

Command RP R (Reentry Point Re
5et), RP(2) of PICI is reset, and after the RPR command is issued, there is no possibility that the processing of %UPI(7) will be interrupted while the RPS command is issued.

Command RT N (Return from Ree
ntry Po1nt) is the program U P (2)
UPI (7), which branches to RP due to (8), returns to the original process after removing the cause of the program abnormality.

Command RP E (Reentry Point E
ntry ) specifies the name and numerical value of the program in which the program error has occurred. The designated program (UP 1 (7) in this embodiment) is transferred to UPI (R P C2 designated by the program itself) by this command RPE.

The program abnormality is restored to a normal state using the above-mentioned command, and this process will be explained below with reference to FIG.

It is assumed that the UPI performs the RPS process immediately after startup, and then a program error occurs and the program starts to run out of control, making it impossible to remove the program.

Along with this, the operator activates U P 2 (8) and issues an RPE command to U P 1 (7). At the same time, a numerical value (any one of 0, 1, 2, . . . ) is passed to kUPl (7) through M S G -path k. At this time, the movement of PICI is as shown in Figure 3, R
Since P is stored in RP (S) and the original RP is reset, RPE commands from other programs become invalid during abnormal processing. Further, the address or register value immediately before control is transferred to the RP is held in P S W (S). As control is transferred to the RP, the numerical value (0, 1) specified (by operator's operation) from UP2 (8) is transferred to U P 1 C7), which has been pulled out of the runaway state.
, 2, . . .), U P I (7) performs abort processing or debug processing.

For example, in the above process, if the value is 0'', the resource is returned.
bort (abort), UP2 (8) is notified, and UPI is abort L. As a result of the notification, "end"
leading to.

On the other hand, if the numerical value is not "O", proceed to the step if the numerical value is "1", and if it is "1", the debugger is called, the fault investigation and recovery of the program abnormality is carried out, and the return is made to the original process by RTN.The RTN process Then, the original RP and PSW are each k
Restored from RP (S), P 5W (S), DP
I(7) can continue the original processing.

Therefore, UPI (7) notifies UP2 that it has been restored. These numerical values and operation results are based on the above M
The information will be communicated through the SGP path, and countermeasures can be taken according to the results.

(G) Effects of the Invention According to the program abnormality processing method of the present invention, firstly, when the user program (UP) itself is abnormal, for example, when it runs out of control, a forced abort at CT (Cr1ses Time) time can be performed. The first and second prevention is to prevent the user program from using its own resources, such as Ilo, files, etc.
Processing such as returning data memory becomes possible, and with O8 there is no need to monitor or manage all resources as in the past, making it possible to make the OS more compact.

Third, forcibly interrupting other programs (abort)
Since it is not intended to be used by SEs (system engineers), it can be used even by general users, and the RPS can be set at the desired position and timing.
It is possible to deal with program abnormalities even if

Fourth, for the UP where the above-mentioned program abnormality has occurred, the numerical values are transferred from other UPs by the operator's operation, for example, RPE U
A simple interruption (abort) can be made by key-in operation of PI
), it is also possible to provide means for returning to normality or investigating the cause of the abnormality, such as starting a debugger or executing repair processing, and furthermore, it is possible to notify the program that issued the RPE command of the results. Workarounds can be taken sequentially, and even in the worst case, it is possible to abort the process.

[Brief explanation of the drawing]

Figure 1 is a block diagram of a program indexer (PIC) using the program abnormality method of the present invention, and Figure 2 is an explanation of the relationship between the operating system (O8) and user program (UP) in a conbeater using the same method. Figure 3 is an explanatory diagram of the execution of each command of reentry point entry (RPE) and return from reentry point (RTN) in the same method, and Figure 4
The figure is an explanatory diagram of the operation of the same system. (1)-PI area, (2)-RP area (3)-
psw area, (4)-RP (S) area, (5)-
PSW (S) area, (6)...area containing other management data.

Claims (1)

[Claims] 1. A control program that controls a plurality of programs manages storage means such as a CPU, a storage element, and a disk for storing files that constitute a computer, and handles abnormal program operations in response to abnormal program operations. An area for storing a branch destination in the event of an abnormality for each of the plurality of programs and an area for storing the state of the program immediately before branching are provided in a table for managing program states, and an abnormal program in the plurality of programs is processed. A program abnormality handling method featuring the following.