Operating System Concepts

1
Operating System Concepts

• Ku-Yaw Chang
• canseco_at_mail.dyu.edu.tw
• Assistant Professor, Department of Computer
  Science and Information Engineering
• Da-Yeh University

2
Chapter 5 Threads

• A process
• With a single thread of control
• With multiple threads of control
• Multithreaded computer systems
• Pthreads
• Solaris 2 threads
• Windows 2000 threads
• Linux threads
• Java threads

3
Chapter 5 Threads

1. Overview
2. Multithreading Models
3. Threading Issues
4. Pthreads
5. Solaris 2 Threads

1. Windows 2000 Threads
2. Linux Threads
3. Java Threads
4. Summary
5. Exercises

4
5.1 Overview

• A thread
• A basic unit of CPU utilization
• A thread ID
• A program counter
• A register set
• A stack
• Called a lightweight process (LWP)
• A traditional process has a single thread of
  control, called a heavyweight process (HWP)

5
Single- and Multithreaded processes
6
5.1.1 Motivation

• A web browser
• One thread display images or text
• Another thread retrieves data from the network
• A word processor
• One thread for displaying graphics
• Another thread for reading keystrokes
• Third thread for performing spelling and grammar
  checking
• Process creation very heavyweight.
• If new process will perform the same tasks as the
  existing process, why incur all that overhead?

7
5.1.2 Benefits

• Responsiveness
• Allow a program to continue even if part of it is
  blocked or is performing a lengthy operation
• Resource sharing
• Memory
• Different threads all within the same address
  space
• Resources
• Economy
• More economical to create and context switch
  threads
• Utilization of multiprocessor architectures
• Increase concurrency

8
5.1.3 User and Kernel Threads

• User threads
• Implemented by a thread library at the user level
• Supported above the kernel
• Advantage
• Fast to create and manage
• Disadvantage
• A thread may cause the entire process to block
• Examples
• POSIX Pthreads
• Mach C-threads
• Solaris 2 UI-threads

9
5.1.3 User and Kernel Threads

• Kernel threads
• The kernel performs thread creation, scheduling,
  and management
• Supported directly by OS
• Advantage
• Kernel can schedule another thread when a thread
  is blocked
• Disadvantage
• Slow to create and manage
• Examples
• Windows NT/2000
• Solaris 2
• BeOS
• Tru64 Unix

10
5.1.3 User and Kernel Threads

• Java threads
• Created and managed by the Java virtual machine
  (JVM)
• Do not fall under the realm of either user or
  kernel threads

11
Chapter 5 Threads

1. Overview
2. Multithreading Models
3. Threading Issues
4. Pthreads
5. Solaris 2 Threads

1. Windows 2000 Threads
2. Linux Threads
3. Java Threads
4. Summary
5. Exercises

12
5.2 Multithreading Models

• Many systems provide support for both user and
  kernel threads
• Different multithreading models
• Many-to-one model
• One-to-one model
• Many-to-many model

13
5.2.1 Many-to-One Model

• Many user-level threads mapped to one kernel
  thread
• Advantage
• Efficient
• Disadvantage
• Entire process may block if a thread makes a
  blocking system call
• Unable to run in parallel on multiprocessors
• Used on systems that do not support kernel
  threads
• Green threads a thread library available for
  Solaris 2

14
5.2.1 Many-to-One Model
15
5.2.2 One-to-One Model

• Each user thread mapped to one kernel thread
• Advantage
• More concurrency
• Disadvantage
• Overhead of creating kernel threads can burden
  the performance of an application
• Restrict the number of threads
• Example
• Windows NT/2000 and OS/2

16
5.2.2 One-to-One Model
17
5.2.3 Many-to-Many Model

• Multiplex many user-level threads to a small or
  equal number of kernel threads
• Advantage
• Developers can create as many as user threads as
  wish
• More concurrency
• Example
• Solaris 2
• Windows NT/2000 with the ThreadFiber package
• IRIX
• HP-UX
• Tru64 Unix

18
5.2.3 Many-to-Many Model
19
Chapter 5 Threads

1. Overview
2. Multithreading Models
3. Threading Issues
4. Pthreads
5. Solaris 2 Threads

1. Windows 2000 Threads
2. Linux Threads
3. Java Threads
4. Summary
5. Exercises

20
5.3 Threading Issues

• The fork and exec System Calls
• Cancellation
• Signal Handling
• Thread Pools
• Thread-Specific Data

21
5.3.1 The fork and execSystem Calls

• One thread in a program calls fork
• New process duplicate all threads
• New process is single threaded
• The exec system call works in the same way as
  described before.
• Some systems have two versions of fork
• If exec is called immediately, then duplicating
  all threads is unnecessary.
• If not, the separate process should duplicate all
  threads.

22
5.3.2 Cancellation

• Thread cancellation
• The task of terminating a thread before it has
  completed
• Target thread
• The thread that is to be cancelled
• Two different scenarios
• Asynchronous cancellation
• One thread immediately terminates the target
  threads.
• Deferred cancellation
• The target thread periodically checks if it is
  should terminate, allowing itself an opportunity
  to terminate itself in an orderly fashion.
• Cancellation points (Pthreads)

23
5.3.3 Signal Handling

• A signal
• Used in Unix systems
• Notify a process that a particular event has
  occurred
• May be received either synchronously or
  asynchronously
• All signals follow the same pattern
• A signal is generated by the occurrence of a
  particular event.
• A generated signal is delivered to a process.
• Once delivered, the signal must be handled.

24
5.3.3 Signal Handling

• Synchronous signal
• An illegal memory access or divided by zero
• Asynchronous signal
• Terminating a process with specific keystrokes (
  ltcontrolgtltCgt )
• Signal may be handled by
• A default signal handler
• Every signal has a default signal handler run by
  the kernel
• A user-defined signal handler
• Override the default signal handler

25
5.3.3 Signal Handling

• Delivering signals is more complicated in
  multithreaded programs
• Deliver the signal to the thread to which the
  signal applies
• Deliver the signal to every thread in the process
• Deliver the signal to certain threads in the
  process
• Assign a specific thread to receive all signals
  for the process

26
5.3.3 Signal Handling

• Windows 2000
• Does not explicitly support signals
• Emulation using Asynchronous Procedure Calls
  (APCs)
• Allow a thread to specify a function that is to
  be called when the thread receives notification
  of a particular event

27
5.3.4 Thread Pools

• Potential problems
• Amount of time required to create the thread
  prior to servicing the request
• Unlimited threads could exhaust the system
  resources
• Thread pool
• To create a number of threads at process startup
  and place them in a pool
• Sit and wait for work
• A server receives a request, a thread is awaken
  from the pool
• Once the thread completes its work, it returns to
  the pool
• If the pools contains no available thread, the
  server waits until one become free

28
5.3.5 Thread-Specific Data

• Threads belonging to a process share the data of
  the process.
• One of the benefits of multithreaded programming
• Each thread also needs its own copy of certain
  data called thread specific data

29
Chapter 5 Threads

1. Overview
2. Multithreading Models
3. Threading Issues
4. Pthreads
5. Solaris 2 Threads

1. Windows 2000 Threads
2. Linux Threads
3. Java Threads
4. Summary
5. Exercises

30
5.4 Pthreads

• a POSIX standard (IEEE 1003.1c) API for thread
  creation and synchronization
• A specification for thread behavior, not an
  implementation
• Common in UNIX-based systems
• Generally not supported in Windows systems

31
Chapter 5 Threads

1. Overview
2. Multithreading Models
3. Threading Issues
4. Pthreads
5. Solaris 2 Threads

1. Windows 2000 Threads
2. Linux Threads
3. Java Threads
4. Summary
5. Exercises

32
5.5 Solaris 2 Threads

• Read it by yourself
• P. 141 to 143

33
Chapter 5 Threads

1. Overview
2. Multithreading Models
3. Threading Issues
4. Pthreads
5. Solaris 2 Threads

1. Windows 2000 Threads
2. Linux Threads
3. Java Threads
4. Summary
5. Exercises

34
5.6 Windows 2000 Threads

• Win32 API
• Primary API for the family of MS OSs
• Provide one-to-one mapping model
• A fiber library provides the many-to-many model
• Primary data structures
• ETHREAD (executive thread block)
• In kernel space
• KTHREAD (kernel thread block)
• In kernel space
• TEB (Thread environment block)
• User-space data structure

35
Chapter 5 Threads

1. Overview
2. Multithreading Models
3. Threading Issues
4. Pthreads
5. Solaris 2 Threads

1. Windows 2000 Threads
2. Linux Threads
3. Java Threads
4. Summary
5. Exercises

36
5.7 Linux Threads

• Read it by yourself
• P. 144 to 145

37
Chapter 5 Threads

1. Overview
2. Multithreading Models
3. Threading Issues
4. Pthreads
5. Solaris 2 Threads

1. Windows 2000 Threads
2. Linux Threads
3. Java Threads
4. Summary
5. Exercises

38
5.8 Java Threads

• Java
• Support threads at the language level
• For the creation and management of threads
• Threads are managed by JVM
• Not by a user-level library or kernel
• A Java program
• Only a main method runs as a single thread in JVM

39
5.8.1 Thread Creation

• Thread creation
• To create a new class that derived from the
  Thread class
• Override the run method of the Thread class
• Calling the start method actually creates the new
  thread

40
Java Thread States
41
5.8.2 The JVM andthe Host Operation System

• JVM does not indicate how Java thread are to be
  mapped to the underlying OS
• Leaving the decision to the particular
  implementation of the JVM
• Windows use one-to-one model

42
Chapter 5 Threads

1. Overview
2. Multithreading Models
3. Threading Issues
4. Pthreads
5. Solaris 2 Threads

1. Windows 2000 Threads
2. Linux Threads
3. Java Threads
4. Summary
5. Exercises

43
Summary

• P.147

44
Chapter 5 Threads

1. Overview
2. Multithreading Models
3. Threading Issues
4. Pthreads
5. Solaris 2 Threads

1. Windows 2000 Threads
2. Linux Threads
3. Java Threads
4. Summary
5. Exercises

45
Exercises

• 5.1
• 5.2
• 5.3
• 5.8

46
The End