JPS5856153A - Subroutine return system - Google Patents

Abstract

PURPOSE:To increase a program running speed by selecting one of exits on the basis of the processing result of a subroutine, and making a data return. CONSTITUTION:A branch to a subroutine CHECK2 is caused by an instruction BAL and data to be checked is compared with normal character check data; when the checked data is the normal character data, a return from the subroutine to a normal character data return address X is made and a branch to character data processing A is caused by an instruction B. When not, it is compared with normal pattern check data; when it coincides with the normal pattern data, a return to a normal pattern data return address X+1 is made and a branch to pattern data processing B is caused by the instruction B. Similarly, a comparison with special order check data is made and when coincidence with the special order is not obtained, a return to an error data return address X+3 is made to cause a branch to error processing D.

Description

DETAILED DESCRIPTION OF THE INVENTION The present invention relates to a subroutine return method when branching from a main program to a subroutine and returning to the main program again.

Internal control programs such as displays have subroutines in addition to the main program, and when specific processing is to be performed, the main program branches to the subroutine, performs the processing, and returns to the main program when the processing is completed. I often do this.

Conventionally, when branching to a subroutine and returning to the main program, the exit of the subroutine is usually fixed to one, and the return is generally made to return to the next stage of the main program where it was interrupted. there were.

FIG. 1 illustrates a conventional subroutine return method. The figure shows a subroutine CHEC that performs data judgment.
The return method when returning to the main program from K1 is shown, (α) is the flowchart in this case, (b
) indicates the corresponding command. By the way, CHEC
Kl means a subroutine that needs to set return information. The return information is set in a register R, not shown, and is "0#" for normal character data, "1" for normal pattern data, "2#" for special order, and "3#" for error data.

In Fig. 1, the subroutine CH is first commanded by BAL.
The process branches to ECK1, data is judged, and the judgment result is set in register RO as return information. Next, data "0" is set in a register R1 (not shown) by an instruction, and the contents of the register RO and the contents of the register R1 are compared by the next instruction C. If the comparison results are equal, the next instruction B branches to processing A, and character data processing is performed. On the other hand, if the result of the comparison is not equal, data @1# is set in register R1 in the next instruction without branching, and the contents of both registers are compared again in the next instruction C. From now on, repeat these steps to create register R.
1 is sequentially loaded with ``2#'' and ``3'' and compared with the register R. If they are equal, the process branches to corresponding processes C and D, and if they are not equal, the process proceeds to the next step.

In the conventional subroutine return method, since the return information determination routine as described above is executed, the number of program execution steps when returning to the main program is large. In the example shown in FIG. 1, the maximum number of steps is 4 for error data, and the number of running steps until error data is determined is (10 steps + step of subroutine CHECK1). The reason for this large number of steps is that the exit of a subroutine is fixed to one, and the processing result of the subroutine is set in a register as return information, and when returning to the main program, the return information is stored in the calling program. This is because the process begins only after the determination routine is executed.

The present invention attempts to solve the drawbacks of the prior art, and its purpose is to eliminate the need for return information, and therefore to create a return information setting routine within a subroutine and a return information determination routine in a calling program. It is an object of the present invention to provide a method that does not require a routine, greatly reduces the number of program running steps, and can improve the program running speed.

Hereinafter, the present invention will be described in detail with reference to Examples.

FIG. 2 shows an embodiment of the subroutine return method of the present invention. This figure shows a return method when returning to the main program from the subroutine CHECK2 for determining data, where (α) is a flowchart in this case, and (b) is a corresponding instruction. Here, CHECK2 indicates a subroutine that does not require setting return information.

In FIG. 2, the subroutine CH is first commanded by BAL.
Branches to ECK2 and first compares the data to be checked with the normal character check data. If the result of the comparison is normal character data, the subroutine returns to the normal character data return address X, and then the data is returned by instruction B. Branches to character data processing A. If the result of the comparison is that the character data is not normal, then the checked data is compared with the normal pattern check data. If the result of the comparison is normal pattern data, the subroutine returns to the normal pattern data return address X+1 and branches to pattern data processing B in response to instruction B. In the same way, all comparisons with the special order check data are performed in the subroutine, and if the result of the comparison is that it is a special order, the process branches to order processing C by returning to the special order return address X+2 and confirming that it is not a special order. In this case, the program branches to error processing by returning to the error data return address X+3.

FIG. 6 explains in more detail the subroutine return method illustrated in FIG. 2.

In the figure, the left side shows the main program, and the part inside the chain line on the right side shows a subroutine for data determination.

In FIG. 5, the data to be checked is set in register RO. In addition, normal character check data, normal pattern check data, and special order check data are sequentially set in register R1, and register RO
is compared with the checked data. On the other hand, in the return address stack area, the return address X is stored by the first branch instruction MAL, and the return address X is read by looking at the stack pointer each time a comparison is performed.

As a result of the first comparison, if the checked data is equal to the normal character check data, the return address is not rewritten and therefore address X is specified, which causes the main program to branch to character data processing. result,
If the checked data is equal to the normal pattern data, the return address is replaced, 1 is added to indicate address X+1, and the program branches to pattern data processing.

Similarly, if the comparison result is equal to the special order check data, address X+2 is specified as the return address and the process branches to order processing; if the comparison result is not equal to the special order check data, address X+5 is specified as the return address. Branch to error handling.

As described above, according to the method of the present invention, when returning from a subroutine to the main program, the exit of the subroutine is not fixed to one, but the exits are provided as many as the conditions of the processing result in the subroutine, and the calling program When returning to subroutine y, the return is made to one of multiple exits based on the processing result in subroutine y, so unlike the conventional method, the return information setting routine within the subroutine and the return at the caller are required. Since the information determination routine is no longer necessary, the number of program running steps is significantly reduced.

In the examples shown in FIGS. 2 and 6, the number of running steps is always (2) regardless of the determination result in the subroutine.
step + step of subroutine CHECK2), and the number of steps is clearly smaller than in the conventional method illustrated in FIG. Furthermore, subroutine CHECK1
and CHECK2, subroutine CHE
Since there is no need to set return information in CK2, the number of steps is smaller than that in subroutine CHIICK1. Therefore, according to the method of the present invention, the number of running steps when returning from a subroutine to the main program is significantly reduced compared to the case of the conventional method described above.

As described above, the subroutine return method of the present invention improves the program running speed and is extremely effective for internal control programs of display devices that require high-speed processing.

[Brief explanation of the drawing]

The first diagram shows the conventional subroutine return method.
FIG. 2 is a diagram showing an embodiment of the subroutine return method of the present invention, and FIG. 3 is a flowchart showing the subroutine return method shown in FIG. 2 in more detail. A, B, C, I)...processing, X, X+1. X+2.
X+5-...Returning Atto Patent Applicant Fujitsu Ltd. Representative Patent Attorney Tama Mushi Iridama Department (3 people) Figure 1 (α) Figure 2 (α) (b) (b)

Claims (1)

[Claims] In a program having a subroutine that is executed by branching from the main program, the number of exits for returning from the subroutine to the main program is equal to the number of conditions of processing results in the subroutine, and when returning from the subroutine to the calling program, A subroutine return method characterized by selecting one of a plurality of exits and returning based on the processing result of the subroutine.