PThreads Programming. A POSIX Standard for Better Multiprocessing

Spis treści książki

• Pthreads Programming
• A Note Regarding Supplemental Files
• Examples
• Preface
  • Organization
  • Example Programs
    • FTP
  • Typographical Conventions
  • Acknowledgments
• 1. Why Threads?
  • What Are Pthreads?
  • Potential Parallelism
  • Specifying Potential Parallelism in a Concurrent Programming Environment
    • UNIX Concurrent Programming: Multiple Processes
      • Creating a new process: fork
    • Pthreads Concurrent Programming: Multiple Threads
      • Creating a new thread: pthread_create
      • Threads are peers
  • Parallel vs. Concurrent Programming
  • Synchronization
    • Sharing Process Resources
    • Communication
    • Scheduling
  • Who Am I? Who Are You?
  • Terminating Thread Execution
    • Exit Status and Return Values
    • Pthreads Library Calls and Errors
  • Why Use Threads Over Processes?
  • A Structured Programming Environment
  • Choosing Which Applications to Thread
• 2. Designing Threaded Programs
  • Suitable Tasks for Threading
  • Models
    • Boss/Worker Model
    • Peer Model
    • Pipeline Model
  • Buffering Data Between Threads
  • Some Common Problems
  • Performance
  • Example: An ATM Server
    • The Serial ATM Server
      • Handling asynchronous events: blocking with select
      • Handling file I/O: blocking with read/write
    • The Multithreaded ATM Server
      • Model: boss/worker model
      • The boss thread
      • Dynamically detaching a thread
      • A worker thread
      • Synchronization: whats needed
      • Future enhancements
  • Example: A Matrix Multiplication Program
    • The Serial Matrix-Multiply Program
    • The Multithreaded Matrix-Multiply Program
      • Passing data to a new thread
      • Synchronization in the matrix-multiply program
• 3. Synchronizing Pthreads
  • Selecting the Right Synchronization Tool
  • Mutex Variables
    • Using Mutexes
    • Error Detection and Return Values
    • Using pthread_mutex_trylock
    • When Other Tools Are Better
    • Some Shortcomings of Mutexes
    • Contention for a Mutex
    • Example: Using Mutexes in a Linked List
    • Complex Data Structures and Lock Granularity
    • Requirements and Goals for Synchronization
    • Access Patterns and Granularity
    • Locking Hierarchies
    • Sharing a Mutex Among Processes
  • Condition Variables
    • Using a Mutex with a Condition Variable
    • When Many Threads Are Waiting
    • Checking the Condition on Wake Up: Spurious Wake Ups
    • Condition Variable Attributes
    • Condition Variables and UNIX Signals
    • Condition Variables and Cancellation
  • Reader/Writer Locks
  • Synchronization in the ATM Server
    • Synchronizing Access to Account Data
    • Limiting the Number of Worker Threads
    • Synchronizing a Server Shutdown
  • Thread Pools
    • An ATM Server Example That Uses a Thread Pool
      • Initializing a thread pool
      • Checking for work
      • Adding work
      • Deleting a thread pool
      • Adapting the atm_server_init and main routines
• 4. Managing Pthreads
  • Setting Thread Attributes
    • Setting a Threads Stack Size
    • Specifying the Location of a Threads Stack
    • Setting a Threads Detached State
    • Setting Multiple Attributes
    • Destroying a Thread Attribute Object
  • The pthread_once Mechanism
    • Example: The ATM Servers Communication Module
      • Using a statically initialized mutex
      • Using the pthread_once mechanism
  • Keys: Using Thread-Specific Data
    • Initializing a Key: pthread_key_create
    • Associating Data with a Key
    • Retrieving Data from a Key
    • Destructors
  • Cancellation
    • The Complication with Cancellation
    • Cancelability Types and States
    • Cancellation Points: More on Deferred Cancellation
    • A Simple Cancellation Example
      • The bullet_proof thread: no effect
      • The ask_for_it thread: deferred cancellation
      • The sitting_duck thread: asynchronous cancellation
    • Cleanup Stacks
    • Cancellation in the ATM Server
      • Aborting a deposit
  • Scheduling Pthreads
    • Scheduling Priority and Policy
    • Scheduling Scope and Allocation Domains
    • Runnable and Blocked Threads
    • Scheduling Priority
    • Scheduling Policy
    • Using Priorities and Policies
    • Setting Scheduling Policy and Priority
    • Inheritance
    • Scheduling in the ATM Server
  • Mutex Scheduling Attributes
    • Priority Ceiling
    • Priority Inheritance
    • The ATM Example and Priority Inversion
• 5. Pthreads and UNIX
  • Threads and Signals
    • Traditional Signal Processing
      • Sending signals and waiting for signals
      • Using a signal mask to block signals
    • Signal Processing in a Multithreaded World
      • Synchronously generated signals
      • Asynchronously generated signals
      • Per-thread signal masks
      • Per-process signal actions
      • Putting it all together
    • Threads in Signal Handlers
    • A Simple Example
    • Some Signal Issues
    • Handling Signals in the ATM Example
  • Threadsafe Library Functions and System Calls
    • Threadsafe and Reentrant Functions
    • Example of Thread-Unsafe and Threadsafe Versions of the Same Function
    • Functions That Return Pointers to Static Data
    • Library Use of errno
    • The Pthreads Standard Specifies Which Functions Must Be Threadsafe
      • Alternative interfaces for functions that return static data
      • Additional routines for performance considerations
      • File-locking functions for threads
      • Where are the threadsafe functions?
    • Using Thread-Unsafe Functions in a Multithreaded Program
  • Cancellation-Safe Library Functions and System Calls
    • Asynchronous Cancellation-Safe Functions
    • Cancellation Points in System and Library Calls
  • Thread-Blocking Library Functions and System Calls
  • Threads and Process Management
    • Calling fork from a Thread
      • Fork-handling stacks
    • Calling exec from a Thread
    • Process Exit and Threads
  • Multiprocessor Memory Synchronization
• 6. Practical Considerations
  • Understanding Pthreads Implementation
    • Two Worlds
    • Two Kinds of Threads
    • Whos Providing the Thread?
      • User-space Pthreads implementations
      • Kernel threadbased Pthreads implementations
      • Two-level scheduler Pthreads implementations: the best of both worlds
  • Debugging
    • Deadlock
    • Race Conditions
    • Event Ordering
    • Less Is Better
    • Trace Statements
    • Debugger Support for Threads
    • Example: Debugging the ATM Server
      • Debugging a deadlock caused by a missing unlock
      • Debugging a race condition caused by a missing lock
  • Performance
    • The Costs of Sharing Too MuchLocking
    • Thread Overhead
      • Thread context switches
    • Synchronization Overhead
    • How Do Your Threads Spend Their Time?
    • Performance in the ATM Server Example
      • Performance depends on input workload: increasing clients and contention
      • Performance depends on a good locking strategy
      • Performance depends on the type of work threads do
      • Key performance issues between using threads and using processes
  • Conclusion
• A. Pthreads and DCE
  • The Structure of a DCE Server
  • What Does the DCE Programmer Have to Do?
  • Example: The ATM as a DCE Server
• B. Pthreads Draft 4 vs. the Final Standard
  • Detaching a Thread
  • Mutex Variables
  • Condition Variables
  • Thread Attributes
  • The pthread_once Function
  • Keys
  • Cancellation
  • Scheduling
  • Signals
  • Threadsafe System Interfaces
  • Error Reporting
  • System Interfaces and Cancellation-Safety
  • Process-Blocking Calls
  • Process Management
• C. Pthreads Quick Reference
• D. About the Authors
• Index
• About the Authors
• Colophon
• Copyright

pokaż cały spis treści

ukryj recenzje