Memory Multiplexing (3+4)

Question 1

two main approaches to multiplexing in an OS

Answer 1

multiplex CPU operations - schedule to run concurrently on single processor
memory mult - use v mem to enable many processes to share phys mem

Question 2

reason for multiplexing

Answer 2

limited hardware resources

Question 3

Explain the concept of protection in memory MUX

Answer 3

2 or more processes cannot accidently overwrite each others data

Question 4

Example of memory protection

Answer 4

kernel data being protected from user program data

a specific range of physical memory addresses is reserved for kernel data only and not user data

Question 5

Explain the notion of controlled overlap

Answer 5

The overlapping of concurrently running process needs to be controlled for protection to be implemented properly - mechanisms needed

Question 6

Explain the concept of translation in memory MUX

Answer 6

mapping from virtual addresses to physical address

RAM uses physical addresses and CPU uses virtual addresses giving it access to all programs that need to be run

Question 7

What are the different types of memory found in the memory hierarchy?

Answer 7

CPU registers, cache, Ram, disk

Question 8

What is cache memory in the OS context

Answer 8

computer memory with very short acces time used for storage of frequently used data or instructions

Question 9

Issues that arise with memory translation

Answer 9

many virtual addresses, few physical addresses - OS needs to map to a valid physical address that is free

Question 10

Issues that arise with memory assignment

Answer 10

Assigned phys mem space must be free and large enough for the chunk but not too large

Question 11

Issues that arise with memory protection

Answer 11

Processes must not access same address at same time - must not overwrite each other - process 1 must execute before being swapped out

Question 12

Issues that arise with memory sharing

Answer 12

n processes share same address in phys mem
Problem - n processes must be divided properly so that each process is assigned phys mem and that equal processes share phys address - must be protected from each other

Question 13

Explain the notion that a process doesnt need occupy continuous mem spaces

Answer 13

Process in Vmem is swapped into main memory - can be in pieces.
Process referenced by virtual memory can be swapped into main mem and executed bit by bit

Question 14

Explain the role of the CPU in generating physical addresses

Answer 14

Addresses generated by cpu point directly to bytes is phys mem - when CPU assigns an executing process to physical memory, the process is assigned a chunk of memory that comprises of data words

Question 15

Define words in the OS context

Answer 15

fixed size data chunks

Question 16

What does it mean for a phys memory space to be valid

Answer 16

large enough and free

Question 17

What is the MMU used for

Answer 17

address translation

Question 18

Briefly explain what happens in a VA to PA translation

Answer 18

MMU checks cache, maps VA to PA so that processes referenced by VA can be loaded into the phys mem at the PA

Question 19

Issue that arises with use of translating between VA and PA

Answer 19

physical memory access conflicts - 2 or more VA map to same PA

Question 20

First part of solution to physical memory access conflicts

Answer 20

OS needs to make sure that different VA space is used to address the memory needed to run every process that has been scheduled to run - each process has different VA?

Question 21

Pages?

Answer 21

Memory chuncks that processes are broken into

Question 22

program vs process

Answer 22

program - code, libraries, data needed to run

process - program executable running

Question 23

What is exe.

Answer 23

processes that has been scheduled to run by the OS and is currently suspended or running
Machine code instructions that incorporates everything needed to run the process

Question 24

Processes stack

Answer 24

executables scheduled to run by OS

stores in VMem and loaded into main mem bit by bit

Question 25

VAS

Answer 25

Virtual address space where process stack is stored - many pages of memory in secondary storage

Question 26

Describe the process of compiling a program into an exe

Answer 26

Build - removes comments, adds header file (if macros used, they will be replaced with code). Compiler will take the code and generate assembly code. Assembler converts assembly into object code (machine code). Linker links machine code with library function code. Merges other code into one exe file. Load exe into main mem

Question 27

Explain dynamic linking

Answer 27

Delays the resolving of some undefined symbols to runtime. Executable code will still have some undefined symbols and list of libraries that will provide definition for said symbols. Load the program will load these libraries

Question 28

Explain the use of stub code in dynamic linking

Answer 28

locates memory-resident library at run time - where the stub (executable code) replaces itself with the library address and executes the library at the address

Question 29

Explain static linking

Answer 29

Linking of libraries done at compile time

Question 30

Linker or link editor

Answer 30

programming language utility program that takes one or more object files generated by a compiler and combines them into a single executable file, library file or object file

Question 31

Paging

Answer 31

deciding which pages to swap into and out of main memory

Question 32

page table

Answer 32

holds mapping function

Question 33

page replacement algorithm

Answer 33

decides which pages to swap into and out of main memory

Question 34

What is loading into memory for execution

Answer 34

process with own kernel and process virtual memory parts. Kernel part contains: OS specific libraries and toher data that must be execited
Virtual memory parts: memory chuncks each storing a specific part of the process where chunks are stacked on top of each other

Question 35

What is uni-programming

Answer 35

No translation or protection - only one process runs a time and a process always runs at the same main mem space but can access any main mem address

Question 36

What is multi-programming

Answer 36

original concept of multiprogamming
each process allocated a specific space in main mem but no translation or protection - linker and loader dynamically adjusted program addresses when they were loaded to prevent overlap

Question 37

Issue with original concept of multiprogamming

Answer 37

no protection - bugs in one program could cause another process to crash

Question 38

What is multi-programming with protection

Answer 38

Two special CPU registers to enable process protection without translation
BaseAddr and LimitAddr prevented users from straying out of allocated memory addresses - if a user program tried to access illegal address - cause error and process would not execure