JPH01258029A - Data processor - Google Patents

Abstract

PURPOSE:To provide a dynamic connection system with rapid performance at the time of execution similarly to a static connection system by executing program connecting edition during the period of execution without including dynamic connection processing in individual programs. CONSTITUTION:A LINK mechanism 5 retrieves a name code string coincident with a name code string of an NAMF part 17 out of name code strings stored in a name storing part 8 in an MAP table 7 and fetches an address in a storage device 2 which corresponds to the name code string from a loading address part 9. The address is set up in an ADRF part 16 and informed to an instruction executing mechanism 4. Consequently, the mechanism 4 knows the address of a subprogram 18 to be called by a CALL instruction 11 and continues program calling processing. During the execution of the instruction 11, the connection edition with the program 18 to be called in executed. Thus, connecting edition can be executed during the period of execution without including dynamic connection processing directly in the program.

Description

DETAILED DESCRIPTION OF THE INVENTION [Field of Industrial Application] The present invention relates to a data processing device that, when a program for the data processing device is created in multiple parts, is combined and edited during execution of the program.

[Conventional technology]

There are two main methods for combining and editing multiple programs that have been created separately:

The first method is to process a plurality of created programs using a combination editing program called a linkage editor prior to their execution. In the following, this first method will be referred to as the static binding method, whereas the second method will be referred to as 1, e.g. Madnick, S.E.

Donovan, J. J., “Operating S.
ystams”, McGraw-HiLL, 1
974 pp 173-179. As shown in
This is a method of combining and editing multiple programs while they are running. In the following, this second method will be referred to as a dynamic combination method.

[Problem to be solved by the invention]

Compared to the dynamic binding method, the static binding method essentially has the following drawbacks.

(1) If one of the multiple programs that have been jointly edited is modified, the entire program must be jointly edited again.

(2) The same program may be included in one group of programs created by combined editing and another group of programs created by similar combined editing.

As programs become larger and more complex, a dynamic linkage method that does not have the drawbacks mentioned above is desired. Dynamic link processing, that is, at the time when the link optical program is actually called, it is determined whether the program itself has been linked and edited (step 41), and if the link has been edited, the entrance address of the called program is retrieved from the storage area. It is necessary to incorporate code that will either branch (steps 42 and 43) or issue a LINK macro to request the management program (operating system) to search for the call destination and branch if the combined edit is not completed (step 44). there were.

In addition, when there is a module that provides a general-purpose function and it is used in multiple different programs, the above static combination method completes the combination and completes the whole before executing each program, so the general-purpose function is not provided. A copy of the module was created and combined into each program. This method not only increases the required capacity of the storage device that stores the program, but also increases the amount of effort required to modify the general-purpose function providing module. On the other hand, in the conventional dynamic linkage method, the basic unit of linkage is a subprogram, so
There were some aspects that did not match the data abstraction method that encapsulates data and multiple operations on it.

The first object of the present invention is to perform combination editing of programs during execution without incorporating dynamic combination processing into individual programs as described above, thereby improving the dynamic combination method.
An object of the present invention is to provide a data processing device that provides the same simplicity of description and high speed of execution as the static binding method.

A second object of the present invention is to make it possible to dynamically combine a module, that is, a data area, and a plurality of subprograms as a unit, thereby avoiding the problem of multiple copies of modules that occur when combining them before execution. The purpose of this invention is to provide a processing device.

[Means to solve the problem]

In order to achieve the above object, the first feature of the present invention is: (al) A value indicating that the combined edit is not completed (hereinafter, this value is referred to as an uncombined value) or a value indicating the address of the combining destination. fields to hold (below).

This field is called the ADRF part), and the field that holds the symbol string indicating the name of the link destination (hereinafter, this field is called the NAMF part) is a pair of data format data. .

(b2) An instruction to call a program by referring to data in the data format of (al) above (hereinafter, this instruction will be referred to as a CALL instruction), and an instruction to call a program by referring to data in the data format of (al) above. An instruction to access an area (hereinafter, this instruction will be referred to as an IACS instruction).

(cl) A table that records the addresses of all programs and data areas on a storage device, along with symbol strings indicating their names (hereinafter, this table will be referred to as a MAP table).

(di) A mechanism for performing combined editing using the data in (al) and the table in (cl) above when executing the instruction in (b) above (hereinafter, this mechanism will be referred to as the LINK mechanism).

The reason is that we have prepared the following.

In order to achieve the above object, the second feature of the present invention is to make modules reentrant and connectable during program execution, and for this purpose includes the following.

(a2) Instructions for referencing an external module, that is, a group of instructions for calling a subprogram in the external module and accessing data in the data area of the external module.

(b2) Activated module management table: The following information is arranged.

1. Symbol string indicating the name of the activated module.

2. Starting address of the code area of the activated module.

3. Starting address of the data area of the activated module.

Note that this table is created for each task.

(c2) Data format as the operand of the instruction in (a2) above: includes the following information fields.

1. A symbol string indicating the name of the referenced module.

2. In the case of an operand of a subprogram call instruction, the offset from the beginning of the module's code area to the subprogram entry point; in the case of a data access instruction, the offset from the beginning of the module's data area to the data.

3. A value indicating uncombined before the combination is completed, and a target subprogram entry point address or data address after the combination is completed.

(d2) Module management/coupling mechanism.

[Effect]

The above (bl) CALL L instruction and DAC3 instruction refer to the data in the above (al) data format during their execution. Here, if the value of the ADRF section (al) is an unbound value, the central processing unit activates the LINK mechanism (dl). The LINK mechanism is the NAMF of (al) above.
The symbol string that is the value of the section is searched in the MAP table of (cl), the address of the corresponding program or data area on the storage device is obtained, and the address is set in the ADRF section of (al).

Furthermore, the central processing unit uses the address obtained by the LINK mechanism to
Continue execution of ALL and DACS instructions.

In addition, in the above, if the value of the A D RF section in (al) above is not an unbound value and a correct address is set, the central processing unit can perform CA without starting the T, I N K mechanism.
Execute the LL instruction DΔC instruction.

According to the above configuration and operation, each program to be combined and edited can be edited by using only the data format in (al) above and the instructions in (a) above, without directly incorporating dynamic combination processing into the program. Combined editing can be done during runtime.

Further, in the field 3 of (c2) above, a value indicating unbound is set in the initial state. The central processing unit is as described above (
When executing the instruction in a2), it is checked whether the field (C2)-3 above in the operand of that instruction has a value indicating unbound, and if it is a value indicating unbound, the module management/coupling mechanism in (d2) above is executed. Start. The module management/coupling mechanism searches for the first address of the code area or the first address of the data area of the module using the symbol string of the module name in (C2)-1 above and the activated module management table in (b2) above, and searches for the first address of the code area or data area of the module, After determining the target address by adding the offset value of 2 and setting it in the field (C2)-3, the process returns to the predetermined command operation.

Through the above operations, the module combination is completed when the instruction is executed.

〔Example〕

An embodiment of the present invention will be described below with reference to FIG.

The data processing device is composed of a central processing unit 1, a storage device 2, and a system bus 3 that connects them. On the storage device 2 are two programs (M
AIN program 10 and SUB program 18) and MA
A MAIN data area 14 used by the IN program is located there.

The MAIN program 10 includes a CALL instruction 11 for calling a SUB program 18.

This CA 1. The L instruction 11 consists of an instruction code 12 and an operand section 13. The instruction code 12 is a command to the instruction execution mechanism 4 inside the central processing unit 1, and the operand section 13 is used to store data during instruction execution. It indicates the location. Here, the address of the link data 15 in the MAIN data area 14 is set in the operand section 12 of the 0ΔLL instruction 11. The link data 15 is composed of an AD RF section]6 which is scheduled to hold the address of the SUB program 18 to be called, and a NAMF section 17 where the name symbol string of the SUB program 18 is set. Here, in a state where the CALL instruction 11 has never been executed, a value indicating that the address has not been set = an unbound value is set in the AD RI section 16.

On the other hand, the central processing unit 1 includes an instruction execution mechanism 4, and further includes a MAP table 7 connected to a LINKfi structure 5 through an internal bus 6.

The MAP table 7 includes a name symbol section 8 that holds the name symbol strings of all programs and data areas on the storage device 2, and records the addresses of each program and data area on the storage device 2. It consists of a load address section 9.

Now, when the CALL instruction ]-1 in the MΔTN program lo is executed for the first time, the instruction execution mechanism 4 analyzes the instruction code 12, interprets it as a program call instruction, and instructs the process shown in the flowchart of FIG. The execution unit 4 executes, in this case, the A of the link data 15 indicated by the operand part 13 of the CAI, L instruction 11.
I) Since the RF section 16 has an uncoupled value in the initial state, the INK machine 5 in the instruction execution mechanism 14 is activated. (
Step 23) The LINK mechanism 5 uses the link data 15 indicated by the operand section 13 of the CALL instruction 11 and the MAP table 7 to perform the processing shown in the flowchart of FIG. In other words, search for a name symbol string in the name symbol section 8 of the 1MAP table 7 that matches the name 7 symbol string in the NAMF section 17 of the link data 15 (step 32).
Step 33) of extracting the address on the storage device 2 corresponding to the name symbol string from the load address field 9 of the MAP table 7, setting that address in the ADRF field of the link data 15, and at the same time informing the instruction execution mechanism 4 of that address. The address is notified and the operation of the LINK mechanism 5 is ended.

As a result, the instruction execution mechanism 4 learns the address of the SUB program 18 that is the program call destination of the CALL instruction 11, and continues the program call process.

As described above, while the CALL instruction 11 is being executed, editing is performed in conjunction with the called SUB program 18.

On the other hand, as mentioned above, the A D RF section 1 of the link data 15
After the address of SUB program 18 is set in 6, CA
When the LL instruction 11 is executed, the LINK mechanism 5 is called without being activated, as shown in the flowchart of FIG.

This example describes calling a program, but
In the case of accessing the data area, the operation of the LINK mechanism is similar, and only the final operation of the data access command differs.

Also, in this example, LINK411! I) RF
When the part 16 is an unbound value, the instruction execution mechanism 4
There is also a method of starting LINKfi structure 5 above.

Next, another embodiment of the present invention will be described with reference to FIG.

In this figure, a SUB subprogram 69 in a SUBMOD module 66 is called from a MAIN module 51 using operand data 53 and a CALL instruction 58.

CALL, the instruction 58 is located in the MA I N-11-joule code area 57, and the instruction code section 5 indicates the type of instruction.
9 and an operand section 60 indicating operand data 53 in the MAIN module data area 52. On the other hand, the operand data 53 includes a reference destination module name symbol string 56 indicating the reference destination module name, and a target S tJ from the beginning of the reference destination SURMOD module code area 68.
An offset section 55 that holds the offset m to the entry point of the B subprogram 69. and a pointer section 5 that holds a value indicating uncombined in the initial state and holds the entry point address of the target SUB subprogram 69 after combining.
Consists of 4.

The CALL instruction 58 is executed by the central processing unit, and at this time the central processing unit checks whether the pointer portion 54 of the operand data 53 indicated by the operand pointer 58 is an unbound value.

If it is not an unbound value, the value held by the pointer section 54 is interpreted as the subprogram entry point address, and the call is made immediately. On the other hand, if the pointer part is an unbound value, the calling operation is interrupted and the module management/coupling mechanism is activated.

The module management/coupling device end performs the processing shown in FIG. Search module data 62 (step 71)
. If there is matching started module data 62 here, and if the original instruction is a subprogram call instruction (step 74), the value of the code area pointer 64 and the value of the offset part 55 of the operand data 53 are added and the pointer If the original instruction is a data access instruction, the value of the data area pointer 63 and the value of the offset section 55 are added and set in the pointer section 54, and the respective instruction operations are restarted.

In the case of FIG. 5, since it is a subprogram call instruction, the sum of the value of the code area head pointer 64 and the value of the offset section 55 is set in the pointer section 54. Furthermore, if there is no activated module data 62 with a symbol string matching the reference module name symbol string 66 in the activated module management table 61, a new data area and code area are secured in the storage area, After loading the code of the module corresponding to the code area and adding new activated module data 62 having a symbol string indicating the data area address, code area address, and module name to the activated module management table 61, the operand The value of the pointer section 54 of the data 53 is set.

Note that although Figure 5 shows the case of subprogram calling, operand data 3 in Figure 5 also applies to data access.
Modules can be combined at runtime by using data in the same format as .

〔Effect of the invention〕

As described above, there are the following two effects when the dynamic bonding force formula is adopted for program bonding editing in a data processing device, which is the first feature of the present invention.

(1) It is possible to omit the need to incorporate dynamic linkage processing into individual programs.

(2) In multiple executions of instructions that require combined editing,
Combined editing is performed only when the first instruction is executed, and from the second time onwards, the instruction execution becomes faster.

Also, the second aspect of the present invention. According to this feature, modules are combined at runtime, thereby avoiding the problem of multiple copies of modules that would occur if they were combined before execution.

In particular, in the present invention, the pointer to the module's data area and the pointer to the code area are stored separately in the activated module management table, so that the parts of the module's components that do not change due to program execution are treated as the code area, and the parts that change. can be managed separately as a data area. For this reason, in a multitasking environment where one central processing unit executes multiple programs in parallel by time-sharing, an activated module management table is created for each task (program) and a data Oi area is secured individually. This allows the code area to be shared by all tasks, which has the effect of reducing the amount of memory required for program execution.

[Brief explanation of the drawing]

FIG. 1 is an overall explanatory diagram of one embodiment of the present invention, and FIG.
A flowchart showing an outline of the process when the instruction execution mechanism in the figure executes the CAL L instruction.
Flowchart showing e1 work of INK organization, No. 4
The figure is a flowchart showing the dynamic linkage process incorporated into the program in the conventional e+ linkage method, Figure 5 is an overall explanatory diagram of another embodiment of the present invention, and Figure 6 is a schematic process of the module management/linkage mechanism. This is a flowchart. 1...Central processing unit, 2...Storage device η, 3...System bus, 4...Instruction execution mechanism, 5...LINK
Mechanism, 6... Internal bus, 7... MAP table, 8
...Name symbol part, 9...Load address part, 10・MA
IN program, 11...CAT, L instruction, +2
...Instruction code, 13...Operand part, 14...M
ΔZN data area, 15...link data, +6...
Ar) RF section, 17...NAMF section, 18...S tJ
B program. 2nd part 3rd cause 4th part 6th part

Claims (1)

[Claims] 1. The storage device comprises a storage device in which programs and data are stored and a central processing unit, and a plurality of programs and data areas used by each program are each given an identification name. A symbol indicating the respective identification names of all programs and data areas on the storage device in a data processing device that uses each storage address on the storage device to call the program and access the data area. Data that includes a correspondence table between columns and storage addresses on the storage device, and pairs storage addresses on the storage device of programs or data areas with symbol strings indicating their identification names as data handled by the data processing device. format and a field holding a storage address on a storage device in the data format, each of which has a value meaning unset, and an instruction to call a program using the data format and an instruction to call a program using the data format. and an instruction for accessing a data area, and identifies a program or data area in the data format when a field to hold a storage address on a storage device in the data format is not set when the instructions are executed. A data processing device characterized in that the storage address of a program or data area is obtained using the value of a field holding a symbol string indicating a name and the correspondence table, and the program is executed. 2. In the data processing device according to claim 1, the program or data area on the storage device obtained by a correspondence table between the symbol string indicating the identification name of the program or data area and the storage address on the storage device. A data processing device characterized in that by setting a storage address on a storage device in the data format in a field to be held, the search of the correspondence table is omitted when the instruction is executed for the second and subsequent times. 3.Equipped with a storage device that stores programs and data and a central processing unit that executes the programs, one module is composed of multiple subprograms and the data area that they use, and by combining multiple modules. 1. A data processing device constituting an entire program, characterized in that module combination is executed using one machine language instruction. 4. In the data processing device according to claim 3,
An information data processing device comprising machine instructions that simultaneously call a subprogram in one module to a subprogram in another module and combine those modules. 5. A data processing apparatus according to claim 3, comprising a machine instruction for simultaneously referencing a data area of another module from a subprogram in one module and coupling those modules.