Browse Prior Art Database

Technique to Handle Page Faults in Logical Tasks

IP.com Disclosure Number: IPCOM000082924D
Original Publication Date: 1975-Mar-01
Included in the Prior Art Database: 2005-Feb-28
Document File: 1 page(s) / 13K

Publishing Venue

IBM

Related People

Winters, RM: AUTHOR

Abstract

Some application systems provide multitasking capabilities that are transparent to the operating system. This facility is common in large data base/data communication systems. The term, logical task, is used to refer to an application defined task since it is not an OS/VS task. The application system contains a dispatching facility that functions similarly to the OS/VS Dispatcher. This multitasking facility is efficient; however, a basic problem exists in a virtual operating system environment. If one logical task gets a page fault, the OS/VS task allotted to the application system is placed in the wait state while the missing page is retrieved. Thus all of the logical tasks are nondispatchable until the missing page request is satisfied. Page faults in the logical tasks can cripple the entire application system.

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 52% of the total text.

Page 1 of 1

Technique to Handle Page Faults in Logical Tasks

Some application systems provide multitasking capabilities that are transparent to the operating system. This facility is common in large data base/data communication systems. The term, logical task, is used to refer to an application defined task since it is not an OS/VS task. The application system contains a dispatching facility that functions similarly to the OS/VS Dispatcher. This multitasking facility is efficient; however, a basic problem exists in a virtual operating system environment. If one logical task gets a page fault, the OS/VS task allotted to the application system is placed in the wait state while the missing page is retrieved. Thus all of the logical tasks are nondispatchable until the missing page request is satisfied. Page faults in the logical tasks can cripple the entire application system. The page fault problem for application systems with this type of multitasking capability can be alleviated in the following manner.

If a logical task receives a page fault, the application system must intercept the OS/VS interrupt processing. If it does not, then the entire application system, which might contain 30 logical tasks, may be set nondispatchable while the missing page is retrieved. It is therefore desirable to intercept the OS/VS interrupt processing and maintain a multitasking environment in spite of paging activity.

First, there must be provided a routine in the application system that executes with a zero protect key. The routine is usually invoked via a user- defined Supervisor Call Instruction (SVC). The routine is required to turn on a flag in the OM VS Task Control Bl...