Transcript 08 Operating System Support

Chapter 8
Operating System Support
Focus on Architecture
Functions of Operating System
• Managing Resources
Access to System Utilities
Access to files
Access to I/O devices
Managing Interrupts & Bus Control (DMA)
Error detection and response
Accounting
• Scheduling Processes (or tasks)
Program creation
Program execution
I/O Processing
• Managing Memory Utilization
Partitioning,
Paging,
Virtual memory,
Segmentation
Types of Operating Systems
What do we want to Optimize in each?
• Interactive
• Batch
• Uni-tasking
• Multi-tasking
• Real-Time
Batch Operating System Resident Monitor
• Jobs are batched in queue
• Monitor handles scheduling
• Monitor controls sequence of
events to process batch
•When one job is finished,
control returns to Monitor
which loads next job
Desirable Hardware Support for OS
• Memory protection
— To protect the Monitor & Utilities
• Timer
— To prevent a job monopolizing the system
• Privileged instructions
— Only executed by Monitor
— e.g. I/O, files
• Interrupts
— Allows for relinquishing and regaining control
• DMA
— Allows for optimizing bus usage
The Case for Multi-programmed Batch Systems
• I/O devices are very slow
 Waiting is inefficient use of computer
• When one program is waiting for I/O,
another can use the CPU
Multi-Programming with Three Programs
Utilization:
Uni-programmed vs Multi-programmed
Multiprogramming Resource Utilization
Scheduling
• What are the challenges of scheduling ?
—When do we start a new program ?
—When a program is blocked, what program is
executed next?
—When several programs are blocked, which gets
first dibs at the resource needed?
—How long can one process monopolize a resource?
How can we model Processing?
A Five State
Process
Model
Types of Scheduling Needed
Keeping Track of Process Information
If Processes can be interrupted and restarted,
what is necessary to be able to stop and
restart programs?
Process Control Block Layout
Scheduling Time Sequence Example:
Some Key Elements of O/S
Process Scheduling:
What happened to the Medium-term Queue ?
Memory Management
What are the Memory Management Issues?
• How can we keep the optimum “portions” of programs
in memory to optimize the use of the resources ?
• How can we protect one program from corrupting
other programs ?
• What do we do when all programs in memory and
stalled and memory is full ?
Memory Management
• Uni-programming
— Memory split into two
— One for Operating System (monitor)
— One for currently executing program
• Multi-programming
— “User” part is sub-divided and shared among active
processes
• Note: Memory size implications
- 16 bits  64K memory addresses
- 24 bits  16M memory addresses
- 32 bits  4G memory addresses
- 64 bits 
? memory addresses
Swapping
• Problem: I/O is so slow compared with CPU
that even in multi-programming system, CPU
can be idle most of the time
• Solutions:
— Increase amount of main memory
– Expensive
— Swapping
How Does Swapping Work?
• Long term queue of “processes stored on disk”
• Processes moved in as space becomes available
• As a process completes it is moved out of main
memory to make room for other process(es)
• If none of the processes in memory are ready (i.e. all
I/O blocked)
— Swap out a blocked process (intermediate queue)
— Swap in a ready process or a new process
But swapping is an I/O process…
Isn’t I/O slow?
So why does swapping make sense ?
Implementation
of Swapping
Other Structures to help with Memory Management
• Partitioning
• Paging
Partitioning
Partitioning:
Splitting memory into sections to allocate to processes
(including Operating System!)
Fixed Size Partitions:
• Equal-sized partitions
—Potentially a lot of wasted memory
• Variable-sized partitions
—Process is stored into smallest reasonable “hole”
Dynamic partitions
— memory leak
— need periodic compaction
Fixed Partitioning
Effect of Dynamic Partitioning – memory leaks
What about Relocation Challenges?
• Can’t expect that process will load into the same
place in memory as last time, but
• Instructions contain addresses
— For Locations of data
— For Addresses for instructions (branching)
• How about having logical address and physical
addresses ?
— Logical address - relative to beginning of program
— Physical address - actual location in memory (this time)
• Mechanisms:
— Use Base Address
— Automatic (hardware) Conversion
Paging
• Split memory into equal sized, “small” chunks
- Page frames
• Operating System maintains list of free
frames
• Then:
— Allocate the required number page frames to a process
– A process does not require contiguous page frames
— Each process has its own page table
Allocation of Free Frames
Paging Addresses- Logical and Physical
Paging Implementation Issues
• Demand paging
—Do not require all pages of a process in memory
—Bring in pages as required
• Page fault
—Required page is not in memory
—Operating System must swap in required page
—May need to swap out a page to make space
—Perhaps select page to throw out based on recent
history
Paging has a Potential for “Thrashing”
• Too many processes in too little memory
• Operating System spends all its time
swapping
• Little or no real work is done
• Solutions
— Good page replacement algorithms
— Reduce number of processes running
— Add more memory
Virtual Memory
• We do not need all pages of a process in memory for
it to run - We can swap in pages only as required
• So - we can now run processes that are bigger than
total memory available!
Differentiation “memory models”:
• Main memory is called real memory
— physical memory
• User/programmer can see much bigger memory space
— virtual memory
Implications:
• Page Tables can become huge and can’t fit into
memory
— need multiple level tables, or
— inverted tables (Why inverted tables ?)
Alternate Inverted Page Table Structure
Problem with Inverted Page Table
• Every virtual memory reference causes two
physical memory accesses
—Fetch page table entry
—Fetch data
• Use Translation Lookaside Buffer
— special cache for page table(s)
TLB and Cache Operation
(special Cache for tables)
Segmentation
• Segmentation is “partitioning” memory that is
visible to the programmer
- Note: Paging is not visible to the programmer
• Usually different segments are allocated to
program and data
• There may be a number of program and data
segments per process (program)
— e.g. to support protection levels, priority levels,
organization, flexibility, etc.
Advantages of Segmentation
• Simplifies handling of growing data structures
• Allows programs to be altered and recompiled
independently, without re-linking and re-loading
• Lends itself to sharing among processes
• Lends itself to protection
Can paging and segmentation be combined?
Example:
Pentium II Address Translation Mechanism
“Segment” uses 2 bits to provide 4 levels of protection, typically:
0: OS kernel,
1: OS,
2: apps needing special security,
3: general apps
Pentium II
(Uses hardware for segmentation & paging)
• Unsegmented, unpaged
— virtual address = physical address
— Used in Low complexity, High performance systems
• Unsegmented, paged
— Memory viewed as paged linear address space
— Protection and management via paging (Ex: Berkeley UNIX)
• Segmented, unpaged
— Collection of local address spaces
— Protection to single byte level, Translation table needed is on chip when
segment is in memory, provide predictable access times
• Segmented, paged
— Segmentation used to define logical memory partitions subject to access
control
— Paging manages allocation of memory within partitions (Ex: Unix System V)
OS Review
Scheduling:
uni-programming
multi-programming
time-sharing
long-term scheduler (queue of all jobs potentially schedulable)
short-term scheduler (queue of processes that are ready to execute)
medium-term scheduling (queue of jobs that can reside in memory)
blocked monitoring (queue of processes blocked for resources)
new – ready – running – blocked – exit
state machine
Memory management:
partitioning
paging – frames, pages, page fault, page table, logical/physical addr
virtual memory – inverted page table, Translation Lookaside Buffer
segmentation