JPS59180736A - Program linking system - Google Patents

Abstract

PURPOSE:To make the physical structure and logical calling structure of a program independent, by solving the corresponding between a load module and the program at the time of compiling. CONSTITUTION:When a load module is called, the calling side program obtains an objective routine address based on the address of the called load module control unit and that of a control section table. When the load module is not loaded into a virtual storage device at this time, the control is handed over to a dispatching routine and, after the load module is loaded into the virtual storage device, the objective routine is called. When the processing is a static connection, the objective routine is directly called, because the routine is already loaded. Moreover, it is possible to realize the same function as overlay by utilizing the counter of the load module control unit.

Description

DETAILED DESCRIPTION OF THE INVENTION [Technical Field of the Invention] The present invention relates to a program linkage system that allows the physical structure and logical call structure of a program to be independent.

[Conventional technology and problems]

In the prior art, calling relationships between programs could not be managed independently. For this reason, tuning that changes the load module configuration (ttLnin!1
), a modification of the program was required. Furthermore, in a program with a dynamic link structure, the loading timing of a load module and the program linkage corresponded too closely.

[Purpose of the invention]

The present invention is based on the above consideration, and the present invention is a program linkage system that makes it possible to make the physical structure of the program and the logical call structure independent by resolving the correspondence between the load module and the program at the time of compiling. The purpose is to provide a method.

[Structure of the invention]

Therefore, the program linkage method of the present invention provides a load module φ address entry field, a load module name entry field, and a process specification entry field.
A load module table having multiple module control units and a dummy module for loading a specified load module onto a virtual storage device and connecting the load module to a corresponding load module control unit are prepared in advance. , when opening the above load module table on a virtual storage device, if the processing specification field of the load module control unit specifies static binding,
Load the load module specified in the load module name entry field of the load module control unit into the virtual storage device and enter the loading address in the load module address entry field of the load module control unit. If the processing specification field of the load module control unit specifies dynamic binding, the loading address of the dummy module loaded on the virtual storage device is used as the load address of the load module control unit. It is characterized by filling in the module address entry field.

[Embodiments of the invention]

Hereinafter, the present invention will be explained with reference to the drawings.

FIG. 1 explains the linkage in the present invention. In Figure 1, LMT is a load module table,
LMCU represents a load module control unit, LM represents a load module, and CT represents a C5ECT (control section) table. Load module table LMT contains multiple load module control units LMCU
Each load module control unit LMCU
has a load module address entry field, a load module name entry field, a process specification entry field, and a count field, and the load module address indicates the start address of the corresponding load module. The load module LM has a C5ECT table CT and several routines (control sections), and a C5ECT table CT and several routines (control sections).
The CT table CT has a plurality of C8ECT address entry columns. The contents of the C5ECT address entry field indicate the start address of the corresponding routine.

FIG. 2 shows the processing when opening the load module table LMT. First, load module table LM
Load T onto virtual storage. Next, it is checked whether the processing specification field of the first load module control unit LMCU indicates static coupling. If T (true), the load module specified by the contents of the load module name entry field is loaded, and this address is written in the load module address entry field. If F (false), enter the dispatcher load module (dummy) in the load module address entry field.
module) address.

This external processing is performed on all load module control units L.
Let's talk about MCU.

Next, calling a load module will be explained.

It is assumed that the calling program knows the address of the load module control unit to be called and the address of the C5ECT table (both relative addresses). The address of the target routine is the Roth-Mosier control unit L.
This can be determined by tracing the MCU address→load module LM address→C5ECT address.

Control passes from the yell to the dispatching routine,
After the dispatching routine loads the called load module onto the virtual storage device, it calls the target routine. Static binding processes are already loaded, so they are called directly. Well, the above-mentioned load
The procedure for calling a module is called the procedure ■. FIG. 3 is a flowchart showing the processing of the dispatching routine. When returning, it returns directly to the caller.

A function similar to an overlay can also be realized using the counter of the load module control unit LMCU.

FIG. 4 shows the calling side when implementing the overlay function. In this case, a procedure to call the deletion procedure is placed after the procedure (2). Figure 5 shows the load module table L when realizing the overlay function.
This is a turtle showing the configuration of MT. In this case, the load
A common counter is provided in the module table LMT. FIG. 6 shows a dispatching routine for implementing the overlay function.

This dispatching routine first increments the common counter by 1, reads the load module, and transfers the value of the common counter to the corresponding load module control unit LMC.
Set the value of U's counter, and finally transfer control to the target routine. FIG. 7 shows the processing performed by the deletion procedure in FIG. Erase from above and connect the dummy module to the load module control unit LMCU of the erased load 0 module.

Furthermore, if recursion is allowed, the calling procedure is as shown in FIG. In this case, a counter 2 is added to the linkage section, and the same value is entered into the counter and counter 2 at the time of loading. This calling procedure checks whether the load module containing the target routine has been loaded or not. If T (true), the contents of counter 2 are decremented by 1, and then the procedure (■) is performed, and if F (false), In this case, immediately perform the procedure ■, add 1 to the value of counter 2, check whether the counter of the corresponding load module is smaller than counter 2, and if T (true), perform the deletion procedure, and then False), do not perform the deletion procedure. This allows overlay to be achieved. Furthermore, rgcttγ5ion may also be present.

Furthermore, as shown in FIG. 9, it is also possible to delete a part first by specifying deletion at a branch point.

A calling procedure such as that shown in FIG. 10 can be used instead of the calling procedure shown in FIG. This calling procedure is for counter 2.
This corresponds to the case where the common counter (separate from the common counter in Figure 6) is present without having a common counter (different from the common counter in Figure 6), the common counter is incremented by 1 if it has been loaded, and the common counter is incremented by -1 when returning, so that "common counter = counter". This erases the load module from the virtual storage device.

〔Effect of the invention〕

As is clear from the above explanation, according to the present invention, after the production program with the load module configuration is completed,
This can be achieved by tuning with less man-hours.

[Brief explanation of the drawing]

Figure 1 is a diagram explaining linkage in the present invention, Figure 2 is a diagram showing processing when opening a load module table,
Figure 3 is a diagram explaining an example of a dispatching routine, Figure 4 is a diagram explaining the caller when implementing an overlay function, and Figure 5 is a diagram explaining the structure of a load module table when implementing an overlay function. , Figure 6 is a diagram showing an example of a dispatching routine to implement the overlay function, Figure 7 is a diagram showing processing according to the deletion procedure in Figure 4, and Figure 8 is a diagram showing a case where rectLrsitrL is allowed. FIG. 9 is a diagram showing branching, and FIG. 10 is a diagram showing the other side of the calling procedure when τecu and rzion are allowed. LMT...Load module table, LMCU...Load module control unit, LM...Load module, CT...C3ECT (control sequence) table. Patent Applicant: Fujitsu Ltd. Representative Patent Attorney Kyotani Shibu Sai 8 Figure 203

Claims (1)

[Claims] A load module table that has multiple load module control units including a load module address entry field, a load module name entry field, and a processing specification entry field, and a load module table that loads a specified load module onto a virtual storage device. When preparing a dummy module in advance to connect the load module to the corresponding load module control unit and opening the above load module table on the virtual storage device, the processing specification entry field for the load module control unit is If static binding is specified, the load module specified in the load module name field of the load module control unit is loaded onto the virtual storage device, and the loading address is sent to the load module control unit. Enter the loading address of the dummy 〇 module loaded on the virtual storage device if the processing specification entry field of the load module control unit specifies dynamic binding. A program linkage method characterized in that the following is entered in a load module Φ address entry column of the load module control unit.