Guide to DECthreads

• Naming conventions
  • exceptions
• Normal mutexes

• One-time initialization
  • of threads
• OpenVMS operating system
  • 64-bit addressing
  • compiling DECthreads-based images
  • condition values used by DECthreads
  • DCL command operation with DECthreads
  • debugging threads
  • interactions with DECthreads
  • linker options for kernel threads
  • linking DECthreads-based images
  • system services
    • blocking
  • using DECthreads with

• Pipelining functional model
• POSIX 1003.1c-1995 standard
  • See POSIX.1c standard
• POSIX 1003.4a/Draft 4 (d4) interface
• POSIX.1c standard #1
• POSIX.1c standard #2
• POSIX.1c style (pthread) interface
• POSIX for OpenVMS layered product
  • interoperability with DECthreads
• Priority
  • inversion
    • avoiding
    • definition of
• Priority inversion
  • of threads
• Process contention scope
  • of threads
• Processes
  • creating
  • states
• Processors
  • releasing thread's control of
• pthread.h header file #1
• pthread.h header file #2
• pthread.h header file #3
• pthread interface #1
• pthread interface #2
  • optionally implemented routines
  • summary of routines
• pthread_attr_stacksize() routine
• pthread_exc_get_status_np() routine
• pthread_exc_matches_np() routine
• pthread_exc_report_np() routine
• pthread_exc_set_status_np() routine
• pthread_once() routine
• pthread_once_t data structure #1
• pthread_once_t data structure #2

• Race conditions #1
• Race conditions #2
  • among threads
  • avoiding
• RAISE exception
• Readers/writer locks
  • creating
  • destroying
  • in DECthreads tis interface
  • initializing
  • locking for read access #1
  • locking for read access #2
  • locking for write access #1
  • locking for write access #2
  • tis interface routines for
  • unlocking for read access
  • unlocking for write access
  • using in thread-safe library code
• Recursive mutexes
• Reentrant code
  • required for multithreaded programming
  • required for thread-safe code
  • See also Thread-reentrant code
• RERAISE macro #1
• RERAISE macro #2
• RERAISE macro #3
• RR (round-robin) scheduling policy
  • of threads

• Sample programs
  • multithreaded prime number search
• Scheduling parameters
  • of threads
    • getting
    • setting
• Scheduling parameters attribute
  • of threads
    • getting
    • setting
• Scheduling policy attribute
  • of threads
    • getting
    • setting
• Scheduling priorities
  • of threads
    • calculating
    • effects of scheduling policy
• Scheduling priority attribute
  • of threads
• Sequence numbers
  • of threads
    • getting
• Signal masks
  • of threads
    • getting
    • setting
• Signals
  • reported as exceptions
  • using under Digital UNIX
• sigwait service
• Stack guard regions
  • of threads
• Stack memory
  • using from threads
• Stacks
  • of threads
    • handling overflow
    • managing
    • size of
    • using a stack guard region
• Stacksize attribute
  • of threads
    • getting
    • setting
• Static memory
  • using before release of mutex
  • using from threads
• Status exceptions
• Synchronization objects
  • condition variables
  • mutexes
• Synchronous cancelation
  • of threads
    • cleanup from
• SYSGEN
  • MULTITHREAD parameter
• System contention scope
• System services
  • cancelability from DECthreads #1
  • cancelability from DECthreads #2

• THIS_CATCH exception object
• Thread attributes objects
  • creating
  • destroying
  • initializing
• Thread-independent services
  • See also tis interface
• Thread-reentrant code
• Thread-safe code
  • in libraries
  • lack of
  • requires reentrant compilation
  • using condition variables
  • using mutexes
  • using readers/writer locks
  • using thread-specific data
• Thread-specific data
  • in DECthreads tis interface
  • keys
    • creating #1
    • creating #2
    • destroying #1
    • destroying #2
    • getting #1
    • getting #2
    • setting #1
    • setting #2
  • tis
  • using in thread-safe library code
• THREADCP tool
• Threads
  • advantages of
  • avoiding nonreentrant routines
  • cancelability state
  • cancelability state attribute
    • setting #1
    • setting #2
  • cancelability type
  • cancelability type attribute
    • setting
  • canceling #1
  • canceling #2
    • asynchronously
    • synchronously
  • cleanup
    • from asynchronous cancelation
    • from synchronous cancelation
  • cleanup handlers #1
  • cleanup handlers #2
  • cleanup handlers #3
  • context-switching
    • in kernel
    • in user mode
  • creating #1
  • creating #2
  • deadlocks
  • delaying execution of
  • destroying #1
  • destroying #2
  • detaching #1
  • detaching #2
  • detachstate attribute #1
  • detachstate attribute #2
  • example of cancelation code
  • execution of
  • guardsize attribute #1
  • guardsize attribute #2
  • identifiers
    • comparing
    • getting #1
    • getting #2
  • inherit scheduling attribute #1
  • inherit scheduling attribute #2
  • joining on
  • joining on another thread
  • locking mutexes #1
  • locking mutexes #2
  • on multiprocessor systems
  • one-time initialization of #1
  • one-time initialization of #2
  • one-time initialization of #3
  • prioritizing #1
  • prioritizing #2
    • effects of scheduling policy
  • priority inversion
  • process contention scope
  • race conditions among
  • real-time scheduling
    • under Digital UNIX
  • reentrant code necessary for
  • requesting delivery of cancelation request #1
  • requesting delivery of cancelation request #2
  • scheduling
    • inherit scheduling attribute
    • scheduling policy attribute
    • scheduling priority attribute
  • scheduling parameters
    • getting
    • setting
  • scheduling parameters attribute #1
  • scheduling parameters attribute #2
  • scheduling policy attribute #1
  • scheduling policy attribute #2
  • See also Multithreaded programming
  • sending signals to
  • sequence numbers
    • getting
  • setting attributes
  • signal masks
    • getting
    • setting
  • stack guard region
  • stack overflow
  • stack size
  • stacks
    • handling overflow
    • managing
    • size of
    • stack guard region
    • tracing
  • stacksize attribute #1
  • stacksize attribute #2
  • starting
  • states
  • synchronizing access to memory
  • system contention scope
  • terminating #1
  • terminating #2
    • causative events
    • due to error
    • normally
    • series of actions #1
    • series of actions #2
    • via pthread_exit() routine
  • thread-specific data
  • timeslicing
  • unlocking mutexes #1
  • unlocking mutexes #2
  • unlocking the global mutex #1
  • unlocking the global mutex #2
  • using shared memory
  • waiting for another thread to terminate #1
  • waiting for another thread to terminate #2
  • waiting on mutexes #1
  • waiting on mutexes #2
  • wakeups for
    • broadcasting #1
    • broadcasting #2
    • signaling #1
    • signaling #2
    • signaling #3
  • yielding processor to another thread
• Throughput scheduling policy
  • of threads
• Time
  • expiration
    • obtaining
• Timeslicing
  • of threads
• tis interface #1
• tis interface #2
  • features of
  • run-time linkages to routines
  • summary of routines
• TRY macro
  • restrictions
• Two-level scheduling
  • under Digital UNIX
  • under OpenVMS Alpha

• Upcalls
  • due to $HIBER and $WAKE system services

• Virtual processors
  • for kernel threads

  6493PI01.HTM
  OSSG Documentation
  22-NOV-1996 13:20:26.80

Copyright © Digital Equipment Corporation 1996. All Rights Reserved.

Legal