DOCSLIB.ORG

Segmentation

Total Page：16

File Type：pdf, Size：1020Kb

Download full-text PDF Read full-text
Segmentation Base and Bounds virtual memory physical memory 0 Address Translation code 6250 (base) data Arvind Krishnamurthy bound Spring 2004 stack 6250+bound Each program loaded into contiguous regions of physical memory. Address Translation Recap Base and Bounds (contd.) n Goal: memory protection n Built in Cray-1 n Translate every memory reference to the actual physical n Hardware cost: two address bound registers, adder, n Programmable: relies on a memory address translation comparator è fast table virtual address > n On a context switch: n On process switch, switch the translation table error save/restore base, bound n Install translations and let the program run + base registers n Who installs translations? Software n What are the pros/cons of n Not user level software è need to distinguish between user and this approach? kernel code è need for protected kernel mode physical address n Hardware support for kernel mode: bit in a “processor status word” n When set, allows all kinds of protected operations n In kernel mode, all memory references are physical addresses Memory Protection Segmentation n Motivation virtual physical address address n separate the virtual address space into several segments so that we can Translation Box share some of them if necessary CPU n also allow holes in the address space (MMU) physical memory n A segment is a region of logically contiguous memory n Main idea: generalize base and bounds by allowing a table of base&bound pairs Data read or write (untranslated) (assume 2 bit segment ID, 12 bit segment offset) virtual segment # physical segment start segment size n This lecture: how to implement the translation? Should be fast n Options code (00) 0x4000 0x700 n Base and Bounds data (01) 0x0000 0x500 n Segmentation - (10) 0 0 n Paging stack (11) 0x2000 0x1000 n Multilevel translation 1 Segmentation example Object file format (assume 2 bit segment ID, 12 bit segment offset) n v-segment # p-segment segment physical memory Notice: start size n Segmentation table performs the task of runtime relocation code (00) 0x4000 0x700 n Loader’s task is simple; linker still needs to perform static relocation data (01) 0x0000 0x500 0 - (10) 0 0 stack (11) 0x2000 0x1000 n 500 Standard file format: ELF, COFF virtual memory type “man a.out” to see detail n magic number 0 2000 n the header information 700 n a list of segments: 3000 1000 (a) size needed for BSS segment (uninitialized variables) 1500 (b) data segment (with initialized global and static variables) (c) text segment (including executable instructions) 3000 4000 4700 n optional relocation information n optional symbol table and line number information 4000 Segmentation example (cont’d) Object file format (cont’d) Virtual memory for strlen(x) char chArray[40]; static double x; physical memory for strlen(x) Runtime segments: Main: 240 store 1108, r2 int y = 13; 244 store pc+8, r31 static long z = 2001; BSS segment: chArray, x x: 108 a b c \0 248 jump 360 data segment: y, z … 24c main () { code segment: all the machine instructions … int i = 3, j, *ip; stack segment: local variables Main: 4240 store 1108, r2 heap segment: dynamic memory allocation strlen: 360 loadbyte (r2), r3 4244 store pc+8, r31 … ip = malloc(sizeof(i)); 4248 jump 360 The NOFF object file contains: 420 jump (r31) chArray[5] = i; 424c code segment … y = 2.0 * z; … data segment with initial values (initData) } strlen: 4360 loadbyte (r2), r3 BSS segment with size only (uninitData) x: 1108 a b c \0 … … 0 ------------------------> increasing offset 4420 jump (r31) … Header text (code) initialized data Segmentation Implementation Address space in Nachos Virtual address 0xfff high address error stack segment offset > n Have a table of (seg, size) segment n Protection: each entry has seg size n (nil,read,write) data uninitialized data (Heap + BSS) . n On a context switch: segment ---------------------- . save/restore the table or a initialized data pointer to the table in kernel main memory memory n Have we addressed all of text + the requirements with this segment approach? (code) physical address 0x000 low address address entry point for __start translation 2 Paging How many PTEs do we need ? n Motivations n Worst case for 32-bit address machine 20 12 n both branch & bounds and segmentation still require fancy memory n # of processes ´ 2 (if page size is 4096 = 2 bytes) management (e.g., first fit, best fit, re-shuffling to coalesce free fragments if no single free space is big enough for a new segment) n What about 64-bit address machine? n can we find something simple and easy 52 n # of processes ´ 2 n Solution n allocate physical memory in terms of fixed size chunks of memory, n Question: how do we solve the huge page-table size or pages. problem? n Simpler because it allows use of a bitmap: 00111110000001100 n each bit represents one page of physical memory n 1 means allocated, 0 means unallocated Paging (contd.) Segmentation with paging n Use a page table to translate Virtual address page table size Virtual address n Context switch: similar to VPage # offset error the segmentation scheme Vseg # VPage # offset > n Question: What should be Page table the page size? Page table PPage# ... seg size PPage# ... ... ... n What are the pros/cons of . PPage# ... this scheme? PPage# ... Each segment has PPage # offset > its own page table PPage # offset Physical address error Physical address Paging example Paged Page Tables Physical memory Virtual memory n So far, page tables have to be allocated linearly in memory 0 n Can we page them? a n That is, can we replace page table pointers with virtual addresses b Page Table 4 i n Implication: they can be swapped c j 4 k d l 3 n 8 Put page tables in a special segment that is translated but not e 1 accessible to user programs (part of program’s virtual address space) e f 12 g f g h n Page table for this segment alone is in physical memory h i page size: 4 bytes 16 a j b n Segment table contains page table pointers that are virtual for some c k segments, but physical for some others (used in MIPS and HPs) d l 3 Assignment 2: Overview Assignment 2: user C program halt.c n Objectives n understand how system call really works #include "syscall.h" n understand how to support multiple address spaces n Problems int n implement a set of system calls main() n Exec, join on processes n Create, open, read, write, close on files { n Fork, yield for threads (optional – extra credit) Halt(); n implement multiprogramming /* not reached */ n use bitmap to find unused main memory } n setup the page table (translation is no longer identity) n data copying between user and kernel n n support argument passing for “exec” Note: we don’t use any standard C libraries (because they n support exec of “prog arg1 arg2” instead of exec prog wouldn’t work with the Nachos kernel) n should be easy Traditional OS Structure Assignment 2: user C program shell.c Application User #include "syscall.h” mode …… Application library while( 1 ) int main() { Write(prompt, 2, output); { SpaceId newProc; i = 0; Portable OS layer OpenFileId input = ConsoleInput; Kernel do { OpenFileId output = ConsoleOutput; mode Read(&buffer[i], 1, input); char prompt[2], ch, buffer[60]; Machine dependent layer } while( buffer[i++] != '\n' ); int i; buffer[--i] = '\0'; n How does Nachos’s structure fit into this model? prompt[0] = '-'; n Nachos is the portable OS layer – it simulates the hardware and machine- prompt[1] = '-'; if( i > 0 ) { dependent layer, and it simulates the execution of user programs running …… newProc = Exec(buffer); on top Join(newProc); n Can still use debugger, printf, etc. } n Can run normal UNIX programs concurrently with Nachos } } n Could run Nachos on real hardware by writing a machine-dependent layer Assignment 2: Overview (cont’d) The assembly stub file: start.s n Nachos execution overview: /* ------- System call stub for Halt ----- */ n user program (written in C): halt.c #include "syscall.h" .globl Halt n gcc cross compiler compiling halt.c into MIPS binary code .text .ent Halt .align 2 decstation-ultrix/bin/gcc halt.c start.s -o halt.coff Halt: coff2noff halt.coff halt /* ---------- a stub to main() ------- */ addiu $2,$0,SC_Halt .globl __start syscall Here, halt.coff is like the standard “a.out” file; .ent __start j $31 .end Halt “halt” is a simplified version of “halt.coff” designed for Nachos __start: /* must start at address 0 */ jal main /* ------- System call stub for Exit ------ */ move $4,$0 .globl Exit n nachos loads and runs the user code ( exec or progtest.cc ) jal Exit .ent Exit n initializing an address space /* if we return from main, exit(0) */ Exit: .end __start n set up the page table (mapping address space to physical memory) addiu $2,$0,SC_Exit n zero-ing all memory cells syscall j $31 n copy all segments in “noff” file (e.g., halt) into main memory .end Exit n call the MIPS simulator to run the user code gcc halt.c start.s -o halt.coff ................................ 4.
Load more

Recommended publications
  • Security and Performance Analysis of Custom Memory Allocators Tiffany
    Security and Performance Analysis of Custom Memory Allocators by Tiffany Tang Submitted to the Department of Electrical Engineering and Computer Science in partial fulfillment of the requirements for the degree of Master of Engineering in Electrical Engineering and Computer Science at the MASSACHUSETTS INSTITUTE OF TECHNOLOGY September 2019 c Massachusetts Institute of Technology 2019. All rights reserved. Author.............................................................. Department of Electrical Engineering and Computer Science June 10, 2019 Certified by. Howard E. Shrobe Principal Research Scientist and Associate Director of Security, CSAIL Thesis Supervisor Certified by. Hamed Okhravi Senior Technical Staff, MIT Lincoln Laboratory Thesis Supervisor Accepted by . Katrina LaCurts, Chair, Masters of Engineering Thesis Committee 2 DISTRIBUTION STATEMENT A. Approved for public release. Distribution is unlimited. This material is based upon work supported by the Assistant Secretary of Defense for Research and Engineering under Air Force Contract No. FA8702-15-D-0001. Any opinions, findings, conclusions or recommendations expressed in this material are those of the author(s) and do not necessarily reflect the views of the Assistant Secre- tary of Defense for Research and Engineering. 3 Security and Performance Analysis of Custom Memory Allocators by Tiffany Tang Submitted to the Department of Electrical Engineering and Computer Science on June 10, 2019, in partial fulfillment of the requirements for the degree of Master of Engineering in Electrical Engineering and Computer Science Abstract Computer programmers use custom memory allocators as an alternative to built- in or general-purpose memory allocators with the intent to improve performance and minimize human error. However, it is difficult to achieve both memory safety and performance gains on custom memory allocators.
    [Show full text]
  • CSE421 Midterm Solutions —SOLUTION SET— 09 Mar 2012
    CSE421 Midterm Solutions —SOLUTION SET— 09 Mar 2012 This midterm exam consists of three types of questions: 1. 10 multiple choice questions worth 1 point each. These are drawn directly from lecture slides and intended to be easy. 2. 6 short answer questions worth 5 points each. You can answer as many as you want, but we will give you credit for your best four answers for a total of up to 20 points. You should be able to answer the short answer questions in four or five sentences. 3. 2 long answer questions worth 20 points each. Please answer only one long answer question. If you answer both, we will only grade one. Your answer to the long answer should span a page or two. Please answer each question as clearly and succinctly as possible. Feel free to draw pic- tures or diagrams if they help you to do so. No aids of any kind are permitted. The point value assigned to each question is intended to suggest how to allocate your time. So you should work on a 5 point question for roughly 5 minutes. CSE421 Midterm Solutions 09 Mar 2012 Multiple Choice 1. (10 points) Answer all ten of the following questions. Each is worth one point. (a) In the story that GWA (Geoff) began class with on Monday, March 4th, why was the Harvard student concerned about his grade? p He never attended class. He never arrived at class on time. He usually fell asleep in class. He was using drugs. (b) All of the following are inter-process (IPC) communication mechanisms except p shared files.
    [Show full text]
  • Paging: Smaller Tables
    20 Paging: Smaller Tables We now tackle the second problem that paging introduces: page tables are too big and thus consume too much memory. Let’s start out with a linear page table. As you might recall1, linear page tables get pretty 32 12 big. Assume again a 32-bit address space (2 bytes), with 4KB (2 byte) pages and a 4-byte page-table entry. An address space thus has roughly 232 one million virtual pages in it ( 212 ); multiply by the page-table entry size and you see that our page table is 4MB in size. Recall also: we usually have one page table for every process in the system! With a hundred active processes (not uncommon on a modern system), we will be allocating hundreds of megabytes of memory just for page tables! As a result, we are in search of some techniques to reduce this heavy burden. There are a lot of them, so let’s get going. But not before our crux: CRUX: HOW TO MAKE PAGE TABLES SMALLER? Simple array-based page tables (usually called linear page tables) are too big, taking up far too much memory on typical systems. How can we make page tables smaller? What are the key ideas? What inefficiencies arise as a result of these new data structures? 20.1 Simple Solution: Bigger Pages We could reduce the size of the page table in one simple way: use bigger pages. Take our 32-bit address space again, but this time assume 16KB pages. We would thus have an 18-bit VPN plus a 14-bit offset.
    [Show full text]
  • Virtual Memory Basics
    Operating Systems Virtual Memory Basics Peter Lipp, Daniel Gruss 2021-03-04 Table of contents 1. Address Translation First Idea: Base and Bound Segmentation Simple Paging Multi-level Paging 2. Address Translation on x86 processors Address Translation pointers point to objects etc. transparent: it is not necessary to know how memory reference is converted to data enables number of advanced features programmers perspective: Address Translation OS in control of address translation pointers point to objects etc. transparent: it is not necessary to know how memory reference is converted to data programmers perspective: Address Translation OS in control of address translation enables number of advanced features pointers point to objects etc. transparent: it is not necessary to know how memory reference is converted to data Address Translation OS in control of address translation enables number of advanced features programmers perspective: transparent: it is not necessary to know how memory reference is converted to data Address Translation OS in control of address translation enables number of advanced features programmers perspective: pointers point to objects etc. Address Translation OS in control of address translation enables number of advanced features programmers perspective: pointers point to objects etc. transparent: it is not necessary to know how memory reference is converted to data Address Translation - Idea / Overview Shared libraries, interprocess communication Multiple regions for dynamic allocation
    [Show full text]
  • The RISC-V Instruction Set Manual, Volume II: Privileged Architecture, Version 1.10”, Editors Andrew Waterman and Krste Asanovi´C,RISC-V Foundation, May 2017
    The RISC-V Instruction Set Manual Volume II: Privileged Architecture Privileged Architecture Version 1.10 Document Version 1.10 Warning! This draft specification may change before being accepted as standard by the RISC-V Foundation. While the editors intend future changes to this specification to be forward compatible, it remains possible that implementations made to this draft specification will not conform to the future standard. Editors: Andrew Waterman1, Krste Asanovi´c1;2 1SiFive Inc., 2CS Division, EECS Department, University of California, Berkeley [email protected], [email protected] May 7, 2017 Contributors to all versions of the spec in alphabetical order (please contact editors to suggest corrections): Krste Asanovi´c,Rimas Aviˇzienis,Jacob Bachmeyer, Allen J. Baum, Paolo Bonzini, Ruslan Bukin, Christopher Celio, David Chisnall, Anthony Coulter, Palmer Dabbelt, Monte Dal- rymple, Dennis Ferguson, Mike Frysinger, John Hauser, David Horner, Olof Johansson, Yunsup Lee, Andrew Lutomirski, Jonathan Neusch¨afer,Rishiyur Nikhil, Stefan O'Rear, Albert Ou, John Ousterhout, David Patterson, Colin Schmidt, Wesley Terpstra, Matt Thomas, Tommy Thorn, Ray VanDeWalker, Megan Wachs, Andrew Waterman, and Reinoud Zandijk. This document is released under a Creative Commons Attribution 4.0 International License. This document is a derivative of the RISC-V privileged specification version 1.9.1 released under following license: c 2010{2017 Andrew Waterman, Yunsup Lee, Rimas Aviˇzienis,David Patterson, Krste Asanovi´c.Creative Commons Attribution 4.0 International License. Please cite as: \The RISC-V Instruction Set Manual, Volume II: Privileged Architecture, Version 1.10", Editors Andrew Waterman and Krste Asanovi´c,RISC-V Foundation, May 2017. Preface This is version 1.10 of the RISC-V privileged architecture proposal.
    [Show full text]
  • • Programs • Processes • Context Switching • Protected Mode Execution • Inter-Process Communication • Threads
    • Programs • Processes • Context Switching • Protected Mode Execution • Inter-process Communication • Threads 1 Running Dynamic Code • One basic function of an OS is to execute and manage code dynamically, e.g.: – A command issued at a command line terminal – An icon double clicked from the desktop – Jobs/tasks run as part of a batch system (MapReduce) • A process is the basic unit of a program in execution 2 How to Run a Program? • When you double-click on an .exe, how does the OS turn the file on disk into a process? • What information must the .exe file contain in order to run as a program? 3 Program Formats • Programs obey specific file formats – CP/M and DOS: COM executables (*.com) – DOS: MZ executables (*.exe) • Named after Mark Zbikowski, a DOS developer – Windows Portable Executable (PE, PE32+) (*.exe) • Modified version of Unix COFF executable format • PE files start with an MZ header. Why? – Unix/Linux: Executable and Linkable Format (ELF) – Mac OSX: Mach object file format (Mach-O) 4 test.c #include <stdio.h> int big_big_array[10 * 1024 * 1024]; char *a_string = "Hello, World!"; int a_var_with_value = 100; int main(void) { big_big_array[0] = 100; printf("%s\n", a_string); a_var_with_value += 20; printf("main is : %p\n", &main); return 0; } 5 ELF File Format • ELF Header – Contains compatibility info – Entry point of the executable code • Program header table – Lists all the segments in the file – Used to load and execute the program • Section header table – Used by the linker 6 ELF Header Format• Entry point of typedef struct
    [Show full text]
  • EECS 482 Introduction to Operating Systems
    EECS 482 Introduction to Operating Systems Fall 2019 Baris Kasikci Slides by: Harsha V. Madhyastha Base and bounds ● Load each process into contiguous region of physical memory • Prevent process from accessing data outside its region • Base register: starting physical address physical • Bound register: size of region memory base + bound virtual bound memory base 0 0 February 20, 2019 EECS 482 – Lecture 12 2 Base and bounds ● Pros? • Fast • Simple hardware support ● Cons? • Virtual address space limited by physical memory • No controlled sharing • External fragmentation February 20, 2019 EECS 482 – Lecture 12 3 Base and bounds ● Can’t share part of an address space between processes physical memory data (P2) virtual virtual address address space 1 data (P1) space 2 data code data code code February 20, 2019 EECS 482 – Lecture 12 4 External fragmentation ● Processes come and go, leaving a mishmash of available memory regions • Wasted memory between allocated regions P1 P2 P5 P3 P4 February 20, 2019 EECS 482 – Lecture 12 5 Growing address space physical memory base + bound virtual bound memory base Stack 0 Heap 0 How can stack and heap grow independently? February 20, 2019 EECS 482 – Lecture 12 6 Segmentation ● Divide address space into segments ● Segment: region of memory contiguous in both physical memory and in virtual address space • Multiple segments of memory with different base and bounds. February 20, 2019 EECS 482 – Lecture 12 7 Segmentation virtual physical memory memory segment 3 fff 46ff code stack 4000 0 virtual memory 2fff
    [Show full text]
  • CS5460: Operating Systems
    CS5460: Operating Systems Lecture 13: Memory Management (Chapter 8) CS 5460: Operating Systems Where are we? Basic OS structure,• HW/SW interface, interrupts, scheduling • •Concurrency • Memory management Storage management Other topics CS 5460: Operating Systems Example Virtual Address Space 0xFFFFFFFF Typical address space has 4 parts User stack segment SP – Code: binary image of program – Data: static variables (globals) HP – Heap : explicitly allocated data (malloc) User heap – Stack: implicitly allocated data User Kernel mapped into all processes User data segment MMU hardware: PC User code segment – Remaps virtual addresses to physical 0x80000000 – Supports read-only, supervisor-only Kernel heap – Detects accesses to unmapped regions How can we load two processes Kernel data segment into memory at same time? – Assume each has similar layout Kernel Kernel code segment 0x00000000 CS 5460: Operating Systems Mapping Addresses How can process (virtual) addresses be mapped to physical addresses? – Compile time: Compiler generates physical addresses directly » Advantages: No MMU hardware, no runtime translation overhead » Disadvantages: Inflexible, hard to multiprogram, inefficient use of DRAM – Load time: OS loader “fixes” addresses when it loads program » Advantages: Can support static multiprogramming » Disadvantages: MMU hardware, inflexible, hard to share data, … – Dynamic: Compiler generates address, but OS/HW reinterpret » Advantages: Very flexible, can use memory efficiently » Disadvantages: MMU hardware req’d, runtime
    [Show full text]
  • Automated Verification of RISC-V Kernel Code
    Automated Verification of RISC-V Kernel Code Antoine Kaufmann Big Picture ● Micro/exokernels can be viewed as event-driven ○ Initialize, enter application, get interrupt/syscall, repeat ● Interrupt/syscall handlers are bounded ● Most of the code fiddles with low-level details in some way ○ -> messy, high level language does not help Idea: ● Use SMT solver on instructions and spec ○ (And see how far we get) Overview 1. RISC-V Z3 model 2. Kernel + Proof RISC-V ● Free modular ISA from Berkeley ● Clean slate, compact, and no legacy features ○ 100 pages of spec for user instructions (including extensions) ○ 60 pages for kernel features ● 62 Core RV64-I instructions ○ Basic register operations, branches, linear arithmetic, bit ops ● Supports full blown virtual memory, or base+bounds RISC-V SMT Model Status ● Only 8 RV64-I instructions still missing: ○ fence(i), rdcycle(h), rdtime(h), rdinstrret(h) ● Supported kernel features: ○ Transfers between protection levels: syscall/traps ○ Base and bounds virtual memory ● Missing kernel features: ○ Modelling interrupt causes ○ More than 2 privilege levels ○ Full page table based virtual memory ● Runs (and passes) provided riscv-tests for instructions ○ -> Demo Model: The first attempt ● Pure Z3 expressions for fetch and exec of instructions ○ fetch_and_exec :: machine state -> machine state ● Having pure Z3 expressions everywhere is convenient ● Use simplify after every step to keep compact ● But: Non-trivial Conditional branches cause blow-up ○ Simplify won’t be able to cut down much in next step ○ Expressions
    [Show full text]
  • Operating Systems Principles and Practice, Volume 3: Memory
    Operating Systems Principles & Practice Volume III: Memory Management Second Edition Thomas Anderson University of Washington Mike Dahlin University of Texas and Google Recursive Books recursivebooks.com Operating Systems: Principles and Practice (Second Edition) Volume III: Memory Management by Thomas Anderson and Michael Dahlin Copyright ©Thomas Anderson and Michael Dahlin, 2011-2015. ISBN 978-0-9856735-5-0 Publisher: Recursive Books, Ltd., http://recursivebooks.com/ Cover: Reflection Lake, Mt. Rainier Cover design: Cameron Neat Illustrations: Cameron Neat Copy editors: Sandy Kaplan, Whitney Schmidt Ebook design: Robin Briggs Web design: Adam Anderson SUGGESTIONS, COMMENTS, and ERRORS. We welcome suggestions, comments and error reports, by email to [email protected] Notice of rights. All rights reserved. No part of this book may be reproduced, stored in a retrieval system, or transmitted in any form by any means — electronic, mechanical, photocopying, recording, or otherwise — without the prior written permission of the publisher. For information on getting permissions for reprints and excerpts, contact [email protected] Notice of liability. The information in this book is distributed on an “As Is" basis, without warranty. Neither the authors nor Recursive Books shall have any liability to any person or entity with respect to any loss or damage caused or alleged to be caused directly or indirectly by the information or instructions contained in this book or by the computer software and hardware products described in it. Trademarks: Throughout this book trademarked names are used. Rather than put a trademark symbol in every occurrence of a trademarked name, we state we are using the names only in an editorial fashion and to the benefit of the trademark owner with no intention of infringement of the trademark.
    [Show full text]
  • Virtual Memory
    Fall 2017 :: CSE 306 Introduction to Virtual Memory Nima Honarmand (Based on slides by Prof. Andrea Arpaci-Dusseau) Fall 2017 :: CSE 306 Motivating Virtual Memory • (Very) old days: Uniprogramming — only one process existed at a time • “OS” was little more than a library occupying the beginning of the memory 0 OS Code Physical Memory Heap User Process Stack 2n-1 • Advantage: • Simplicity — No virtualization needed • Disadvantages: • Only one process runs at a time • Process can destroy OS Fall 2017 :: CSE 306 Goals for Multiprogramming • Transparency • Processes are not aware that memory is shared • Works regardless of number and/or location of processes • Protection • Cannot corrupt OS or other processes • Privacy: Cannot read data of other processes • Efficiency • Low run-time overhead • Do not waste memory resources • Sharing • Cooperating processes should be able to share portions of address space Fall 2017 :: CSE 306 Abstraction: Address Space • Address space: Each process’ view of its own memory range • Set of addresses that map to bytes 0 Code • Problem: how can OS provide Heap illusion of private address space to each process? Stack • Address space has static and 2n-1 dynamic components • Static: Code and some global variables • Dynamic: Stack and Heap Fall 2017 :: CSE 306 How to Virtualize Memory? • Problem: How to run multiple processes concurrently? • Addresses are “hard-coded” into program binaries • How to avoid collisions? Fall 2017 :: CSE 306 How to Virtualize Memory? • Possible Solutions for Mechanisms: 1) Time Sharing
    [Show full text]
  • The Abstraction: Address Space
    Memory Management CMPU 334 – Operating Systems Jason Waterman Memory Virtualization • What is memory virtualization? • OS virtualizes its physical memory • OS provides an illusion of memory space per each process • It is as if each process can access the entire memory space • Benefits • Ease of use in programming • Memory efficiency in terms of time and space • The guarantee of isolation for processes as well as OS • Protection from errant accesses of other processes 9/16/21 CMPU 335 -- Operating Systems 2 Early OSes 0KB • Load only one process in memory Operating System (code, data, etc.) • Poor utilization and efficiency 64KB Current Program (code, data, etc.) max Physical Memory 9/16/21 CMPU 335 -- Operating Systems 3 Multiprogramming and Time Sharing 0KB • Load multiple processes in memory Operating System (code, data, etc.) • Execute one for a short while 64KB Free • Switch processes between them in memory 128KB Process C • Increase utilization and efficiency (code, data, etc.) 192KB Process B (code, data, etc.) 256KB • Causes an important protection issue Free • 320KB Errant memory accesses from other processes Process A (code, data, etc.) 384KB Free 448KB Free 512KB Physical Memory 9/16/21 CMPU 335 -- Operating Systems 4 Address Space • The OS creates an abstraction of physical 0KB Program Code memory 1KB • The address space contains all the information Heap about a running process 2KB • program code, heap, stack and etc. (free) 15KB Stack 16KB Address Space 9/16/21 CMPU 335 -- Operating Systems 5 Address Space(Cont.) • Code 0KB
    [Show full text]