Browse Prior Art Database

Variable-Sized System Queue Area in a Virtual Memory System

IP.com Disclosure Number: IPCOM000061353D
Original Publication Date: 1986-Jul-01
Included in the Prior Art Database: 2005-Mar-09
Document File: 1 page(s) / 13K

Publishing Venue

IBM

Related People

Lang, TG: AUTHOR [+3]

Abstract

A method is described which, in response to varying load on the Virtual Memory (VM) system, changes the amount of memory which can be addressed with the same address regardless of whether virtual address translation is enabled in the system. When the operating system of the Virtual Resource Manager (VRM) of the VM system requires an area of memory, known, for example, as System Queue Area (SQA) for allocating control blocks to represent resources, the SQA must have the attribute that its virtual address equals its real address, so that the control blocks may be accessed regardless of whether virtual address translation is enabled. In response to changes in system load, the size of SQA required may change drastically.

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 51% of the total text.

Page 1 of 1

Variable-Sized System Queue Area in a Virtual Memory System

A method is described which, in response to varying load on the Virtual Memory (VM) system, changes the amount of memory which can be addressed with the same address regardless of whether virtual address translation is enabled in the system. When the operating system of the Virtual Resource Manager (VRM) of the VM system requires an area of memory, known, for example, as System Queue Area (SQA) for allocating control blocks to represent resources, the SQA must have the attribute that its virtual address equals its real address, so that the control blocks may be accessed regardless of whether virtual address translation is enabled. In response to changes in system load, the size of SQA required may change drastically. It is acceptable for the solution of this problem to cause the program requesting a control block to be forced to wait while the SQA is expanded. In accordance with the new method, to satisfy the virtual=real (V=R) requirement, the VRM reserves the entire range of virtual addresses, for example, from 0 to 16 megabytes for SQA and code and data which is permanently pinned in memory. When the system is IPLed (initial program loaded), the size of V=R pinned memory for the VRM is known, and this amount of space is reserved, then one page of memory (2048 bytes) is assigned to SQA. Normally, this would leave subsequent virtual addresses available to be assigned to pageable virtual memory. Instead, the first pageable address is assigned, starting at the 16-megabyte boundary, leaving a "hole" in the virtual address space. When the existing SQA is consumed, a page of...