Platform-specific spinlocks disabled by default unless BOOST_LWM_USE_SPINLOCK is defined.

[SVN r14213]
This commit is contained in:
Peter Dimov
2002-06-20 15:16:03 +00:00
parent 951c2b7e83
commit a322dc54dc

View File

@ -15,10 +15,10 @@
// This software is provided "as is" without express or implied
// warranty, and with no claim as to its suitability for any purpose.
//
// typedef <implementation-defined> boost::detail::lightweight_mutex;
// typedef <unspecified> boost::detail::lightweight_mutex;
//
// boost::detail::lightweight_mutex meets the Mutex concept requirements
// See http://www.boost.org/libs/thread/doc/mutex_concept.html#Mutex
// boost::detail::lightweight_mutex meets a subset of the Mutex concept
// requirements: http://www.boost.org/libs/thread/doc/mutex_concept.html#Mutex
//
// * Used by the smart pointer library
// * Performance oriented
@ -30,6 +30,23 @@
// lightweight_mutex does not guarantee fairness.
// * Never keep a lightweight_mutex locked for long periods.
//
// The current implementation can use a pthread_mutex, a CRITICAL_SECTION,
// or a platform-specific spinlock.
//
// You can force a particular implementation by defining BOOST_LWM_USE_PTHREADS,
// BOOST_LWM_USE_CRITICAL_SECTION, or BOOST_LWM_USE_SPINLOCK.
//
// If neither macro has been defined, the default is to use a spinlock on Win32,
// and a pthread_mutex otherwise.
//
// Note that a spinlock is not a general synchronization primitive. In particular,
// it is not guaranteed to be a memory barrier, and it is possible to "livelock"
// if a lower-priority thread has acquired the spinlock but a higher-priority
// thread is spinning trying to acquire the same lock.
//
// For these reasons, spinlocks have been disabled by default except on Windows,
// where a spinlock can be several orders of magnitude faster than a CRITICAL_SECTION.
// Note: lwm_linux.hpp has been disabled by default; see the comments
// inside for more info.
@ -37,8 +54,7 @@
#include <boost/config.hpp>
//
// Note to implementors: if you write a platform-specific lightweight_mutex
// Note to implementors: if you write a platform-specific spinlock
// for a platform that supports pthreads, be sure to test its performance
// against the pthreads-based version using shared_ptr_timing_test.cpp and
// shared_ptr_mt_test.cpp. Custom versions are usually not worth the trouble
@ -47,11 +63,10 @@
// Be sure to compare against a "real" pthreads library;
// shared_ptr_timing_test.cpp will compile succesfully with a stub do-nothing
// pthreads library, since it doesn't create any threads.
//
#ifndef BOOST_HAS_THREADS
# include <boost/detail/lwm_nop.hpp>
#elif defined(BOOST_USE_ASM_ATOMIC_H)
#elif defined(BOOST_LWM_USE_SPINLOCK) && defined(BOOST_USE_ASM_ATOMIC_H)
# include <boost/detail/lwm_linux.hpp>
#elif defined(BOOST_LWM_USE_CRITICAL_SECTION)
# include <boost/detail/lwm_win32_cs.hpp>
@ -59,9 +74,9 @@
# include <boost/detail/lwm_pthreads.hpp>
#elif defined(WIN32) || defined(_WIN32) || defined(__WIN32__)
# include <boost/detail/lwm_win32.hpp>
#elif defined(__sgi)
#elif defined(BOOST_LWM_USE_SPINLOCK) && defined(__sgi)
# include <boost/detail/lwm_irix.hpp>
#elif defined(__GLIBCPP__)
#elif defined(BOOST_LWM_USE_SPINLOCK) && defined(__GLIBCPP__)
# include <boost/detail/lwm_gcc.hpp>
#elif defined(BOOST_HAS_PTHREADS)
# define BOOST_LWM_USE_PTHREADS