JPS6340940A - Non-shared patching system for shared program - Google Patents

Abstract

PURPOSE:To execute a patch change of a shared program, to only a specific task, by storing the change contents in an address in a program on a main storage device, at the time loading a page containing an address of the shared program, onto an intrinsic virtual storage area. CONSTITUTION:First of all, a non-shared virtual storage area 21 for loading a shared program 5 is secured in an intrinsic virtual storage area 10 of a task 4. Subsequently, an address of a shared program store area 22 of the program 5 in a secondary storage device 6 is set to an external page table corresponding to the storage area 21. Simultaneously, by relocating a virtual address of the program 5 so as to be brought to addressing to the storage area 21, the storage area 22 in the device 6 is brought to mapping to the storage area 21. Thereafter, a real page RP1 of a main storage device 7 is allocated to the storage area 21 by a paging function, and a program text of the program 5 is page-loaded from the device 6. In such a case, the change contents are sent to the real page RP1 of the device 7 from a change contents store device 3.

Description

[Detailed Description of the Invention] [Field of Industrial Application] The present invention relates to a non-shared patch method for a non-shared program,
In particular, the present invention relates to a non-shared patch method for shared programs in a computer system that has shared programs shared among multiple tasks in a shared virtual storage area.

[Conventional technology]

Conventionally, in this type of patching method for shared programs, when patching a shared program, the following steps were taken: ■ Stop the schedule of the task in operation, change the corresponding location in the shared program's secondary storage device r, and update the shared virtual memory. ■ Storing the modified shared program in an area separate from the shared program storage area on the secondary storage device mapped to the shared virtual storage area, and linking this shared program to a specific task. [Problem to be solved by the invention] In the conventional shared program patching method described above, the shared program is executed on a private virtual storage area separate from the shared virtual storage area. When changing the patch, all running tasks are subject to change processing at the same time, and there is a drawback that changes cannot be made to only specific tasks.

In addition, the above disadvantages can be avoided if the changed shared program is linked to the private virtual storage area of a specific task, but the system resources required for link processing and the secondary storage VA'll that stores the changed shared program The disadvantage is that it requires an area of

An object of the present invention is to address the above-mentioned points and provide a non-shared batch method for shared programs that allows batch modification of shared programs only for specific tasks without using the area in the secondary storage device. Our goal is to provide the following.

[Means for solving problems]

The non-shared batch method for shared programs of the present invention provides, in a computer system having a shared program shared among multiple tasks in a shared virtual storage area, a shared program name of a shared program to be batch changed to a specific task; a change specification means for specifying an address within the program and contents of change to the address; 2
a mapping means for mapping a shared program storage area on the next storage device; and a change content storage means for storing the change content in the internal address.

(Operation) In the shared program non-shared patch method of the present invention, the change specification means specifies the shared program name of the shared program to be changed.

Specifies this in-program address and the changes to this address, and the mapping means uses private virtual storage and secondary storage to load the shared program into a specific task's private virtual storage that is different from the shared virtual storage. and the shared program storage area, and when the changed content storage means loads a page including the intra-program address of the shared program on the private virtual storage area, the changed content is stored in the intra-program address on the main storage device.

〔Example〕

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

FIG. 1 is a configuration diagram showing an embodiment of the present invention. In the non-shared batch method of the shared program of this embodiment, shared program change information 3 of the shared program 5 is sent to a specific task 4.
0, and the shared program 5 existing in the shared program storage area 22 in the secondary storage device 6.
The program text of task 4 is stored in the private virtual storage area 10 of task 4.
The mapping means 2 maps to a non-shared virtual storage area 21 in The main part thereof is composed of the change content storage means 3.

The shared program change information 30 includes a shared program name 31 of the shared program 5 to be changed, an intra-program address 32 of this shared program 5, and change details 33 for this intra-program address 32.

Next, the operation of the non-shared patch method of the shared program of this embodiment configured as described above will be explained.

Normally, as shown in FIG. 2, the shared program 5 stored in the shared program storage area 22 of the secondary storage device 6 is
Since it is shared by a plurality of tasks 40 and 41, it is mapped to the shared virtual storage area 20 that can be addressed by these tasks 40 and 41.

Here, as shown in FIG. 1, when a specific task 4 wants to receive a service in which the content of a certain address of the shared program 5 is changed, the change specifying means 1 selects the shared program name 31 to be changed and the address within this program. The shared program change information 30 consisting of an entry list having elements 32 and change contents 33 for this address 32 is given to the task 4 and linked to the management table of the task 4.

During dynamic linking processing when task 4 calls shared program 5 for the first time, search for shared program name 31 in shared program change information 30 specified by change specifying means 1, and if an entry matching shared program 5 exists, The mapping means 2 maps the shared program 5 to the non-shared virtual storage area 21 in the private virtual storage area 10 of the task 4 using the procedure described below.

First, a non-shared virtual storage area 21 for loading the shared program 5 is secured in the private virtual storage area 10 of the task 4. Next, the address of the shared program storage area 22 of the shared program 5 in the secondary storage device 6 is set in the external page table corresponding to the non-shared virtual storage area 21, and
The shared program storage area 22 in the secondary storage device 6 is mapped to the non-shared virtual storage area 21 by relocating the virtual address of the shared program 5 so that it is addressed to the non-shared virtual storage area 21.

In this way, the paging feature allows non-shared virtual storage 2
1 is assigned the real page RPI of the tα device 7,
The program text of the shared program 5 can be page loaded from the secondary storage device 6.

Note that at the same time as allocating the unique virtual storage area 10 at the start of execution of the task 4, the mapping means 2 performs the above-mentioned mapping process on the shared program 5 indicated by the shared program name 31 of the shared program change information i[i30. You can also do it.

When the task 4 refers to an address in the shared program 5 in the non-shared virtual storage area 21, if the virtual page VPI that includes this address has not yet been loaded on the main storage device 7, the real page RPI in the main storage device 7 , and page text is paged in from the external page API of the secondary storage device 6.

At this time, following the shared program name 31 of the shared program change information 30 specified by the change specifying means 1 (if the intra-program address 32 is included in the virtual page VPl, the program name 31 of the shared program change information 30 specified by the change specifying means 1) By storing the image of the changed content 33 following the intra-program address 32 at the corresponding address on the page RPI, the content change of the shared program 5 is completed.

Through the above processing, the task 4 can receive the changed service of the shared program 5, and the shared program 5 on the shared virtual storage area 20 can also be used during and after the changing process of the shared program 5 for the task 4. It is guaranteed that the other tasks 40 and 41 that share the shared program 5 will receive the service of the shared program 5 without being affected by changes in the shared program 5 to the task 4.

Note that if the real page RPI allocated to the virtual page VP1 in the non-shared virtual storage area 21 and in which the changed content 33 is stored needs to be released due to a high load on the real page resources of the main storage device 7, etc. One method is to once cancel the changed content 33 of the address 32 in the program, and then re-create the virtual page V including the address 32 in the program.
Change content storage means 3 when access to PI occurs
, or as another method, make this page an update page and use a storage area on the secondary storage device 6 that is different from the shared program storage area 22 in the secondary storage device 6 as a page-out area for the update page. When the virtual page VPI is accessed again, the page is paged in from the page-out area, and the changed content 33 of the address 32 in the program is stored without executing the process by the changed content storage means 3 again. Guarantee.

〔Effect of the invention〕

As explained above, the present invention allows patch changes to be made to a shared program shared by multiple tasks only to a specific task, and without increasing the area on the secondary storage device that stores this shared program. This has the effect of making it possible to implement it.

[Brief explanation of the drawing]

FIG. 1 is a block diagram showing an embodiment of the present invention, and FIG. 2 is a block diagram showing a shared program sharing method. In the figure, 1... Change specification means, 2... Mapping means, 3... Change content storage means, 4... Task, 5... Shared program, 6... Secondary storage device, 7 ...Main storage device, 10...Unique virtual storage area, 20...Shared virtual storage area, 21...Non-shared virtual storage area, 22...Shared program storage area, 30...Shared program Change information, 31...Shared program name, 32...Address within the program, 33...Change details, API...External page on secondary storage device, VPI...
- Virtual page in a non-shared virtual storage area, RPI...A real page on the main storage device.

Claims (1)

[Claims] In a computer system that has a shared program shared among multiple tasks in a shared virtual storage area, a specific task is provided with the shared program name of the shared program to be patched, the address within this program, and this shared program. a change specifying means for specifying address change contents; and a change specification means for loading the shared program into a private virtual storage area of the specific task that is different from the shared virtual storage area, on the private virtual storage area and the secondary storage device. a mapping means for mapping a shared program storage area; and storing the changed content in the program address on a main storage device when a page including the program address of the shared program is loaded onto the unique virtual storage area; A non-shared patch method for a shared program, characterized in that it has a change content storage means for storing changes.