Review for Midterm

1
Review for Midterm

• Fred Kuhns
• (fredk_at_arl.wustl.edu, http//www.arl.wustl.edu/fr
  edk)
• Department of Computer Science and Engineering
• Washington University in St. Louis

2
Computer Systems

• Consists of hardware and software that combine to
  solve some identified problem. That is, it
  implements an application.
• Software solves some problem entertainment,
  information management, scientific problem
  solving etc.
• Hardware provides the provides basic computing
  resources (Processor(s), Memory, System Bus and
  I/O modules).
• Software is divided into two categories
• Application software specialized for the
  problem domain.
• System software general programming
  environment. Simplifies application programming
  and can be used for different application
  domains.
• Users include people, computers or applications.

3
System Software

• Primary goals
• User environment Execute user programs and make
  solving user problems easier.
• Resource Management Make the computer system
  convenient to use
• Examples of system software
• run-time system provides language specific
  libraries , such as I/O, graphics and math.
• database management systems
• windowing system
• Operating system interface to hardware,
  resource abstractions and sharing.

4
Operating Systems

• Requirements
• Resource abstractions (data and special
  devices)
• Process abstraction (i.e. instance of a program)
• Sharing and isolation
• Management dimensions
• Time Space Synchronization and deadlock
  Accounting and status information
• Functions
• Device, Process, Resource, Memory and File
  management
• Implementation issues
• Protection (modes) Trusted control program
  (kernel) Requesting services (system calls)
• Issues
• Performance protection and security
  correctness maintainability commercial
  standards and open systems.

5
Resources

• Resource
• abstraction that represents any hardware or
  software component
• models components and their operation
• Resource Sharing sharing in a multiprogramming
  environment
• Space-multiplexed vs. Time-multiplexed
• isolate resource access to implement allocation
  policy
• authorized versus non-authorized sharing
• OS implements
• fundamental abstraction of hardware components
• core mechanisms to manage resource sharing

6
Model and Core Abstractions

• Basic model access, transform and store
  information.
• Abstractions facilitate processing and
  conceptualization.
• Common abstractions in modern systems
• Process basic unit of computation
• Program in execution.
• encapsulates notion of computation and resource
  ownership
• File basic unit of information storage.
• typically persistent storage is used for files
• also non-persistent such as memory based files
  and file systems
• A resource must satisfy
• a process must request it from the OS
• a process must suspend operation until resource
  is allocated

7
Processes and Resources
An application consists of one or more processes,
a set of resources and state
Resources
Processes
...
...
memory Ri
...
CPU
display Rj
Operating System
8
Multiprogramming and Process Model
traces for P1, P2 and P3
P1
P2
P3
Shared Code (Program)
9
OS Organization
10
OS Techniques

• Controlling access to hardware
• hardware support for privilege levels typically
  two user and system
• Use trusted program (Kernel) for privileged
  operations, resource management, protection
  enforcement (isolation), and sharing.
• Users request OS services using a well defined
  interface that validates user.
• Common schemes System calls or Message passing

11
Typical System Call Approach
execution environment
application
trap
libraries
user
System call interface
kernel
System Services
Kernel interface
handlers
File Subsystem
Memory Subsystem
hardware
12
Typical Message Passing Approach
Client
microkernel
13
Microkernels

• Small operating system core containing essential
  OS functions.
• Most services implemented as external subsystems
  device drivers file systems virtual memory
  manager windowing system and security
• Design
• Primitive memory management mapping each virtual
  page to a physical page frame grant, map and
  flush.
• Inter-process communication
• I/O and interrupt management
• Benefits
• Portability isolate port specific code to
  microkernel
• Reliability modular design, small microkernel,
  simpler validation
• Uniform interface all services are provided by
  means of message passing
• Extensibility allows the addition of new
  services
• Flexibility existing features can be subtracted
• Distributed system support message are sent
  without knowing what the target machine is or
  where it is located
• Object-oriented operating system components are
  objects with clearly defined interfaces that can
  be interconnected to form software

14
(No Transcript)
15
Basic Concepts - UNIX

• Two privilege levels user and system mode
• Process has "protected" address space -
  protection
• Process share same kernel space - efficiency
• Per process state in kernel - efficiency
• Per process kernel stack, practical
  considerations
• system versus process context virtual property
• Kernel is trusted program that runs directly on
  HW
• Loaded at boot time and initializes system,
  creates initial system processes. , remains in
  memory and manages the system
• Resource manager/mediator
• Well defined entry points syscalls, exceptions
  or interrupts.
• Performs privileged operations.
• Entry into the Kernel
• Synchronous System call interface Hardware
  exceptions
• Asynchronous Hardware interrupts
• Scheduled System processes - swapper and
  pagedaemon.

16
Mode, Space and Context
Privileged
user
kernel
mode
context
process
Application (user code)
System calls Exceptions
kernel
X not allowed
Interrupts, System tasks
system space
17
Von Neumann Architecture
Central Processing Unit (CPU)
18
Computer Architecture Processor
CPU
Memory
0
PC
1
MAR
IR
Instruction
Reg N
MBR
Reg 1
Reg 0
Instruction
I/O AR
Instruction
execution unit
I/O BR
Data
Devices
Data
Data
Data
.
.
Buffers
N
MBR - Memory Buffer Register I/O AR -
Input/Output Address Register I/O BE -
Input/Output Buffer Register
PC - Program Counter IR - Instruction
Register MAR - Memory Address Register
19
Instruction Cycle
Interrupt Cycle
Execute Cycle
Fetch Cycle
Interrupts Disabled
Fetch Next Instruction
Execute Instruction
Check for Process Int
START
Interrupts Enabled
HALT
20
Interrupts
Processor
Device table
dispatcher (interrupt handler)
X
Bus
command
status
rt-counter
Timer
Example clock interrupt is generated every 10ms,
used by OS for accounting and resource (CPU)
sharing.
21
I/O and Devices
Processor
Device table
dispatcher (interrupt handler)
X
Bus
command
status
data 0
data 1
Device Controller (firmware and logic)
...
data N-1
Device X
22
I/O Techniques

• Direct I/O with polling, aka Programmed I/O
• Processor does all the work. Poll for results.
• Interrupt Driven I/O
• Device notifies CPU when I/O operation complete
• Memory Mapped I/O
• rather than reading/writing to controller
  registers the device is mapped into the OS memory
  space
• increased efficiency
• Direct memory access (DMA)
• DMA controller read and write directly to memory,
  freeing the CPU to do other things.
• CPU notified when DMA complete

23
The Process Model

• Process Sequential execution, characterized by
  trace
• notion of execution and resource ownership
• Resource allocations
• Explicit Memory, Files, Others
• Implicit CPU, interleave execution fairness,
  deadlines, throughput
• State kept in Process Control Block
• Process state, Program counter, CPU registers,
  CPU scheduling information, Memory-management
  information, Accounting information, I/O status
  information
• Process scheduling queues may migrate between
  queues
• Job queue Ready queue Device queues.
• Context Switch
• system saves state of old process and load the
  saved state of new process.
• Context-switch time is overhead the system does
  no useful work while switching.
• Time dependent on hardware support.

24
Resource Schedulers

• Long-term scheduler job scheduler
• selects which processes should be brought into
  the ready queue.
• invoked infrequently (seconds, minutes)
• controls the degree of multiprogramming
• Medium-term scheduler
• allocates memory for process.
• invoked periodically or as needed.
• Short-term scheduler CPU scheduler
• selects which process should be executed next and
  allocates CPU.
• invoked frequently (ms)

25
Example Process Representation
Process P2 State
Hardware State (registers)
init P0
Program counter
Kernel Process Table

P0 HW state resources
Process P3
Memory base

Process State (logical)
P2 HW state resources
Process P1

PN HW state resources
Process P2
26
Typical Process States
suspend
dispatch
admit
time-out
suspend
event
release
admit
activate
wait
suspend
event
activate
27
Cooperating Processes

• Independent process cannot affect or be affected
  by the execution of another process.
• Cooperating process can affect or be affected by
  the execution of another process
• Advantages of process cooperation
• Information sharing Computation speed-up
  Modularity Convenience
• Dangers of process cooperation
• Data corruption, deadlocks, increased complexity
• Requires processes to synchronize their
  processing
• Cooperating processes require special mechanisms
  for synchronization, deadlock control and
  communication

28
IPC Mechanisms

• Mechanisms used for communication and
  synchronization
• Message Passing Shared Memory Synchronization
  Debugging Event Notification
• Message passing
• no shared variables two operations for fixed or
  variable sized message send(message),
  receive(message)
• Processes exchange messages over a communication
  link.
• Implementation of communication link (logical
  properties)
• Direct or Indirect (Naming)
• Synchronization blocking versus non-blocking
• Buffering Zero, bounded or unbounded. Automatic
  or Explicit buffering
• Send-by-copy or send-by-reference
• fixed or variable sized messages

29
Direct versus Indirect

• Direct
• Link Properties established automatically
  between pairs explicit addressing one link per
  pair of communicating processes symmetric or
  asymmetric addressing
• Disadvantage explicit use of ID
• Indirect
• Link properties use common mailbox link may be
  shared by multiple processes Supports multiple
  links between process pairs
• Ownership
• Processes only owner may receive, others can
  write
• Kernel creator may read. Read permission may be
  transferred to other processes.
• Disadvantages more complex management

30
Multiprocessor Systems

• Motivations Performance, Fault tolerance
• Architectures SISD, SIMD, MISD, MIMD
• Classifications Tightly or Loosely coupled
• Memory Access UMA, NUMA, NORMA
• OS Structure Separate Supervisor, Master/Slave,
  SMP

31
Multiprocessors
MultiProcessor
SIMD
MIMD
Shared Memory (tightly coupled)
Distributed Memory (loosely coupled)
Symmetric (SMP)
Clusters
Master/Slave
32
Typical SMP System
CPU
CPU
CPU
CPU
500MHz
cache
MMU
cache
MMU
cache
MMU
cache
MMU
System/Memory Bus

• Issues
• Memory contention
• Limited bus BW
• I/O contention
• Cache coherence

I/O subsystem
50ns
Bridge
INT
ether
System Functions (timer, BIOS, reset)
scsi

• Typical I/O Bus
• 33MHz/32bit (132MB/s)
• 66MHz/64bit (528MB/s)

video
33
SMP Operating Systems

• Design Issues
• Threads - effectiveness of parallelism depends on
  performance of primitives used to express and
  control concurrency.
• Process Synchronization more than disabling
  interrupts
• Process Scheduling Global versus Local (per
  CPU) Task affinity for a particular CPU
  resource accounting inter-thread dependencies
• Memory Management cache coherence memory access
  synchronization balancing overhead with
  increased concurrency
• Reliability and fault Tolerance
• Defintions
• Parallelism degree to which a multiprocessor
  application achieves parallel execution
• Concurrency Maximum parallelism an application
  can achieve with unlimited processors
• System Concurrency supports multiple threads of
  control
• User Concurrency User space threads (coroutines)
  provide a natural programming model for
  concurrent applications.

34
Threads
Multithreaded Process Model
Single-Threaded Process Model
Thread
Thread
Thread
Thread Control Block
Thread Control Block
Thread Control Block
Process Control Block
User Stack
Process Control Block
User Stack
User Stack
User Stack
Kernel Stack
User Address Space
User Address Space
Kernel Stack
Kernel Stack
Kernel Stack
35
Thread Concepts

• Process Model Resource ownership and dispatching
• Thread model separates the notion of execution
  from the Process abstraction
• Represents an execution path and computational
  state
• All threads of a process share a common address
  space.
• One or more threads per process, each having
• Execution state Saved thread context when not
  running Execution stack Per-thread static
  storage for local variables Shared access to
  process resources
• Effectiveness of parallel computing depends on
  the performance of the primitives used to express
  and control parallelism
• Thread model is useful for expressing the
  intrinsic concurrency of a program regardless of
  resulting performance

36
Thread Support

• User space threads implemented in user space
  system library
• Examples POSIX Pthreads, Mach C-threads, Solaris
  threads
• Benefits no modifications required to kernel
  flexible and low cost
• Drawbacks can not block without blocking entire
  process no parallelism (not recognized by
  kernel)
• Kernel level threads kernel support thread is
  the basic scheduling entity
• Examples Windows 95/98/NT/2000, Solaris, Tru64
  UNIX, Linux
• Benefits coordination between scheduling and
  synchronization less overhead than a process
  suitable for parallel application
• Drawbacks more expensive than user-level
  threads generality leads to greater overhead
• Combined Approach support both models
• Benefits Less time to do the following
• create a new thread terminate a thread switch
  between two threads within the same process
  communicate.

37
Threading Models

• User threads Many-to-One
• kernel threads One-to-One
• Mixed user and kernel Many-to-Many

38
Relationships
39
Threading Issues

• fork and exec duplicate one, some or all
  threads?
• Cancellation cancel the target thread and free
  resources
• asynchronous cancellation VS deferred
  cancellation
• Signals generation, posting and delivery
• synchronous signals should go to thread causing
  the signal
• what about asynchronous signals?
• Bounding the number of threads created in a
  dynamic environment
• use thread pools
• Al threads share some address space
• use of thread specific data

40
Solaris Threading Model
Process 2
Process 1
user
Int kthr
kernel
hardware
41
Concurrency

• Concurrent programs may exhibit a dependency on
  process/thread execution sequence or processor
  speed
• Race condition final value of the shared data
  may depend on the order in which it was accessed
• Prevent race conditions, concurrent processes
  synchronize their access to the shared memory
  locations known as the critical section problem
• There are two basic issues resulting from the
  need to support concurrency
• Mutual exclusion combines a sequence of actions
  into a critical section which appear to execute
  atomically
• Conditional synchronization. Processes/threads
  must be able to wait for the data to arrive or
  constraints (assertions) to be satisfied.
• Avoiding race conditions
• Disjoint variables Weakened assertions Global
  invariants Synchronization
• At-Most-Once property only one variable may be
  in the read and write set of two or more
  processes

42
Critical Section Problem

• Entry/exit protocol satisfies
• Mutual Exclusion At most one in CS
• Absence of deadlock 2 or more then at least one
  enters CS
• Absence of Unnecessary delay if attempting to
  enter CS and others not in CS (or exited) then
  process must enter
• Eventual entry cant wait forever no
  starvation.

Task A while (True) entry protocol
critical section exit protocol
non-critical section
43
Concurrency What toKnow

• Be able to asses mutual exclusion algorithm using
  the 4 properties on previous slide
• Busy waiting versus conditional synchronization
• be able to describe lost wakeup problem
• Be ble to describe and understand the example
  solutions for
• consumer-producer bounded-buffer readers and
  writers dining philosophers
• Understand the following
• interrupt blocking, Petersons solution hardware
  instruction support (test_and_set instruction)

44
Concurrency

• Understand and be able to show an example
  implementation of semaphores and monitors using
  primitive spin locks and condition variables.

45
Deadlock

• Be able to define the term deadlock
• Know the necessary conditions for a deadlock to
  exist
• Know and be able to use/construct the resource
  allocation graph
• Know the approaches to deadlock handling and be
  able to work out a detailed example
• prevention ensure one of the necessary
  conditions does not hold
• avoidance know in advance resource requirements
  then asses each request
• detection and recovery

46
Scheduling

• Understand the concept of a burst cycle and the
  difference between I/O and CPU bound processes
• Be able to construct a gantt chart illustrating
  the behavior of any of the scheduling algorithms
  we have covered.
• Policy versus mechanism
• all material covered today