JPH02224122A - Dynamic link system for memory patch data - Google Patents

Abstract

PURPOSE:To prevent the formation of a useless idle area in a memory by dynamically loading patch data and executing address resolution between a patch data area and an OS program in each loading. CONSTITUTION:The title system is constituted of a patch data registering func tion 10, a patch data loading function and a patch data address resolving func tion 30. Patch data are previously registered in a prepared patch data file 40, when necessary the patch data are loaded from the file 40 to a patch data area 60 and the address of patch data area 60 is written in the OS program 50 simultaneously with the loading of the patch data to the memory. Thereby, at the 2nd execution of the OS program or after, control can be directly trans ferred to the patch data in the area 60. Consequently, the generation of a useless idle area in the memory can be prevented.

Description

DETAILED DESCRIPTION OF THE INVENTION (Industrial Application Field) The present invention relates to a dynamic link method for memory batch data, and in particular to a method for dynamically linking memory batch data, particularly when a processing error is discovered in an operating system (O8) program in a computer system. The present invention relates to a dynamic linking method for memory batch data that prevents the generation of useless patch areas in memory batch procedures for program modification.

(Prior Art) Conventionally, when a processing error is discovered in an OS program in a computer system, the OS program cannot be easily corrected or replaced, so patch data has been used as a temporary correction method. create it, write it in the patch data area prepared in advance, and then
A branch instruction for branching to the batch data area and the address of the batch data area are written at the location in the S program where a processing error is detected. By doing this, when the OS program is running, control is transferred to the batch data area by a branch instruction, and after executing the patch data, the OS program is executed again.
Control is returned to the S program and the processing error in the OS program is corrected. The patch data created at this time is written directly into the patch data area, so the address of the patch data is determined at the time of creating the patch data. Therefore, the address of the batch data area is written along with a branch instruction to the location where a processing error is found in the OS program. Can be written.

(Problem to be Solved by the Invention) The conventional method of correcting processing errors in an OS program using the memory batch procedure described above has the problem that it ends up generating unnecessary memory, as shown below. was there.

This problem will be explained using FIGS. 5(a) to 5(c).

If a processing error is discovered in an OS program, the OS program can be corrected to perform the correct processing using the memory patch procedure described above, but the batch data area used at this time is only one in the O8. This area exists only for all OSs in the computer system.
It is shared by programs (this is done to save memory).

Now, as shown in Figure 5 (a), there are multiple OS programs (
When OS programs A, B, and C) use the batch data area, the batch data of each OS program will be mixed in the batch data area. In this state, when the batch data in the patch data area becomes unnecessary by replacing OS programs B and C with programs that have corrected processing errors (OS programs B' and C'), ), the area occupied by batch data that is no longer needed remains as an empty area.

Although this free area can be reused as an area for newly created patch data, the used areas before and after the free area cannot be destroyed, so the length of the newly created batch data is the free area. If the length is larger than the length of , it is necessary to create new patch data by dividing it into two parts.

However, creating batch data in two parts increases the possibility of mistakes, so normally, as shown in Figure 5 (C), the empty area is made unusable and the unused area at the back is used. ing. If such a procedure is repeated, a problem arises in that unusable free areas are scattered in the patch data area in a moth-eaten state, resulting in an increase in wasted memory. This is because the conventional memory batch procedure directly writes the address of the batch data area in the OS program along with a branch instruction to transfer control to the batch data area, which creates an empty area in the patch data area. This is a problem that occurs because patch data cannot be moved even if the patch data is not moved.

In this way, in the conventional method for correcting processing errors in OS programs using memory batch procedures in computer systems, as described above, the addresses of patch data in the patch data area are brought into the OS program. Even if unused areas are scattered throughout the area, the batch data cannot be moved, so there is a problem that the batch data area becomes unusable and memory waste increases. Ta.

(Means for Solving the Problems) The dynamic linking method for memory batch data according to the present invention provides a memory patching procedure that is performed to correct a processing error when a processing error is discovered. When registering patch data in the patch data file, the operating system (O
8) Includes a patch data registration function that embeds a cut command to transfer control to the patch data load function and a patch data ID given to the patch data in a specified location of the 10g, and this patch data registration function allows
When the cut command embedded in the OS program is executed, this is recognized as a request to load patch data, and the patch data corresponding to the patch data ID specified after the cut command is loaded into a patch data file. It includes a patch data loading function that searches for the patch data ID from within the system using the patch data ID as a key, and loads the found patch data into the batch data area, which is a memory area dedicated to writing patch data.
When the program is executed, the cut instruction in the OS program is rewritten to a normal branch instruction, and at the same time, the loading address of the loaded patch data is set in place of the patch data ID, so that the control is directly transferred to the patch data area. Contains data address resolution functionality.

(Example) Embodiments will be described below with reference to the drawings.

FIG. 1 is an overall configuration diagram showing an embodiment of the present invention.

Referring to FIG. 1, the patch data dynamic link system is comprised of a patch data registration function 10, a patch data load function 20, and a patch data address resolution function 30.

FIG. 2 is a detailed explanatory diagram of the patch data registration function 10.

Patch data group 80 input by the user from the input device 90
is composed of patch data 81, a patch data ID 82 unique within the system, and an address 83 of a cut point set in the OS program 50.

In the patch data registration function 10, the input patch data 81 is registered in the patch data file 40.
The patch data registration function 10 stores the address 41 in the patch data file 40 of the registered patch data 81,
A set of patch data IDs 82 is set in a mapping table 70 created in the main memory.

Finally, the patch data registration function 10 embeds the cut command 51 and patch data ID 82 in the position in the OS program 50 that corresponds to the specified cut lower address 83.

FIG. 3 is a detailed explanatory diagram of the patch data loading function 20.

After the processing of the patch data registration function 10 described above is completed, as a result of executing the cut command 51 incorporated by the patch data registration function 10 when actually executing the OS program 50, a patch data load request is detected and batch data loading is performed. Control is transferred to function 20. The patch data load function 20 reads the patch data ID 82 set after the cut command 51 in the OS program 50,
Search the mapping table 70 created in the main memory using the patch data ID 82 as a key,
Address 4 in patch data file 40 where patch data 81 corresponding to patch data ID 82 is registered
1 can be found.

Then, the target patch data 81 is transferred from the patch data file to the next free area in the patch data area 60 prepared in advance on the main memory. At this time, the address 61 in the patch data area 60 loaded with the patch data 81 is determined by the patch address resolution function 30.
will be notified.

FIG. 4 is a detailed explanatory diagram of the address resolution function 30 for resolving the address between the patch data 81 and the OS program 50.

The batch data load function 20 notifies the address resolution function 30 of the load address 61 in the patch data area 60 of the patch data 81 loaded and the address 83 in the OS program 50 where the cut instruction exists. The address resolution function 30 changes the cutting instruction 5 provided in the OS program 50 to a conventional branch instruction 52, and also changes the batch data ID 82.
The load address 61 in the patch data area 20 is written to the location where is written.

By doing this, from the second time onwards, this OS Grodarum 50
When executing the patch data load request is not detected and O
Control is transferred directly from the S Grodarum 50 to the patch data area 60, and the patch data 81 loaded there
can be executed.

As mentioned above, the embodiment has been explained, but as in this method, by dynamically loading patch data and resolving the address between the patch data area and the OS program each time, the patch data area is placed at the beginning. Since it can be used sequentially from the beginning, the problem of generation of wasteful empty areas in the patch data area in the conventional 0810 gram correction method using the memory batch procedure is solved.

(Effects of the Invention) As explained above, the present invention is a correction method using a memory patch procedure for correcting processing errors in an OS program, in which patch data is registered in advance in a prepared patch data file, and patch data is registered in advance when necessary. By loading the patch data from that file onto the patch data area, it is possible to avoid the wasteful empty area of memory that conventionally occurred in the patch data area, and
By writing the address of the patch data area in the OS program at the same time as loading the batch data into memory, it is possible to transfer control directly to the patch data in the patch data area when the OS program is executed for the second time or later. This has an excellent effect.

[Brief explanation of the drawing]

FIG. 1 is an overall configuration diagram showing an embodiment of the memory batch data dynamic link method according to the present invention, and FIG. 2 is a detailed explanatory diagram of the patch data registration function in this embodiment.
Fig. 3 is a detailed explanatory diagram of the patch data loading function of this embodiment, Fig. 4 is a detailed explanatory diagram of the patch address resolution function of this embodiment, and Fig. 5 is an explanation of problems in the conventional memory batch procedure. It is a diagram. 10...Batch data registration function, 20...Patch data load function, 30...Patch address resolution function,
40...Patch data registration file, 41...Address within the file, 50...OS program, 51...
・Cut instruction, 52... Branch instruction, 60... Patch data area, 61... Patch data load address, 70... Mapping table, 80... Patch data group, 81... Batch data , 82...Batch data ID, 83...Address within the OS program, 90
...Input device.

Claims (1)

[Claims] When a processing error is discovered, in the memory patch procedure performed to correct the error, patch data, which is a group of correct processing instructions, is registered in the patch data file, and at the same time, the specified operating system (OS) program The patch data registration function includes a cut command for transferring control to the patch data load function and a patch data registration function that embeds the patch data ID given to the patch data.
When a cut command embedded in a program is executed, this is recognized as a request to load patch data, and the patch data corresponding to the patch data ID specified after the cut command is loaded from within the patch data file. It includes a patch data loading function that searches using the patch data ID as a key and loads the found patch data into a patch data area that is a memory area dedicated to writing patch data, and when the OS program is executed for the second time or later, the OS Includes a patch data address resolution function that rewrites the cut instruction in the program to a normal branch instruction and at the same time sets the loading address of the loaded patch data instead of the patch data ID to directly transfer control to the patch data area. A dynamic link method for memory batch data characterized by the following.