PTHREADS-WIN32 SNAPSHOT 2002-??-?? ---------------------------------- Web Site: http://sources.redhat.com/pthreads-win32/ FTP Site: ftp://sources.redhat.com/pub/pthreads-win32 Coordinator: Ross Johnson We are pleased to announce the availability of a new snapshot of Pthreads-win32, an Open Source Software implementation of the Threads component of the POSIX 1003.1c 1995 Standard for Microsoft's Win32 environment. Some functions from POSIX 1003.1b are also supported including semaphores. Other related functions include the set of read-write lock functions. Parts of the implementation also comply with the Open Group's Single Unix specification for compatibility with major Unix implementations and Linux. Pthreads-win32 is free software, distributed under the GNU Lesser General Public License (LGPL). Please see the 'Acknowledgements' section at the end of this announcement for the list of contributors. ------------------------------- Changes since the last snapshot ------------------------------- Cleanup code default style. (IMPORTANT) ---------------------------------------------------------------------- Previously, if not defined, the cleanup style was determined automatically from the compiler/language, and one of the following was defined accordingly: __CLEANUP_SEH MSVC only __CLEANUP_CXX C++, including MSVC++, GNU G++ __CLEANUP_C C, including GNU GCC, not MSVC These defines determine the style of cleanup (see pthread.h) and, most importantly, the way that cancelation and thread exit (via pthread_exit) is performed (see the routine ptw32_throw() in private.c). In short, the exceptions versions of the library throw an exception when a thread is canceled or exits (via pthread_exit()), which is caught by a handler in the thread startup routine, so that the the correct stack unwinding occurs regardless of where the thread is when it's canceled or exits via pthread_exit(). In this and future snapshots, unless the build explicitly defines (e.g. via a compiler option) __CLEANUP_SEH, __CLEANUP_CXX, or __CLEANUP_C, then the build NOW always defaults to __CLEANUP_C style cleanup. This style uses setjmp/longjmp in the cancelation and pthread_exit implementations, and therefore won't do stack unwinding even when linked to applications that have it (e.g. C++ apps). This is for consistency with most current commercial Unix POSIX threads implementations. Compaq's TRU64 may be an exception (no pun intended) and possible future trend. Although it was not clearly documented before, it is still necessary to build your application using the same __CLEANUP_* define as was used for the version of the library that you link with, so that the correct parts of pthread.h are included. That is, the possible defines require the following library versions: __CLEANUP_SEH pthreadVSE.dll __CLEANUP_CXX pthreadVCE.dll or pthreadGCE.dll __CLEANUP_C pthreadVC.dll or pthreadGC.dll E.g. regardless of whether your app is C or C++, if you link with pthreadVC.lib or libpthreadGC.a, then you must define __CLEANUP_C. THE POINT OF ALL THIS IS: if you have not been defining one of these explicitly, then the defaults as described at the top of this section were being used. THIS NOW CHANGES, as has been explained above, but to try to make this clearer here's an example: If you were building your application with MSVC++ i.e. using C++ exceptions and not explicitly defining one of __CLEANUP_*, then __CLEANUP_C++ was automatically defined for you in pthread.h. You should have been linking with pthreadVCE.dll, which does stack unwinding. If you now build your application as you had before, pthread.h will now automatically set __CLEANUP_C as the default style, and you will need to link with pthreadVC.dll. Stack unwinding will now NOT occur when a thread is canceled, or the thread calls pthread_exit(). Your application will now most likely behave differently to previous versions, and in non-obvious ways. Most likely is that locally instantiated objects may not be destroyed or cleaned up after a thread is canceled. If you want the same behaviour as before, then you must now define __CLEANUP_C++ explicitly using a compiler option and link with pthreadVCE.dll as you did before. WHY ARE WE MAKING THE DEFAULT STYLE LESS EXCEPTION-FRIENDLY? Because no commercial Unix POSIX threads implementation allows you to choose to have stack unwinding. Therefore, providing it in pthread-win32 as a default is dangerous. We still provide the choice but unless you consciously choose to do otherwise, your pthreads applications will now run or crash in similar ways irrespective of the threads platform you use. Or at least this is the hope. WHY NOT REMOVE THE EXCEPTIONS VERSIONS OF THE LIBRARY ALTOGETHER? There are a few reasons: - because there are well respected POSIX threads people who believe that POSIX threads implementations should be exceptions aware and do the expected thing in that context. (There are equally respected people who believe it should not be easily accessible, if it's there at all, for unconditional conformity to other implementations.) - because pthreads-win32 is one of the few implementations that has the choice, perhaps the only freely available one, and so offers a laboratory to people who may want to explore the effects; - although the code will always be around somewhere for anyone who wants it, once it's removed from the current version it will not be nearly as visible to people who may have a use for it. Source module splitting ----------------------- In order to enable smaller image sizes to be generated for applications that link statically with the library, most routines have been separated out into individual source code files. This is being done in such a way as to be backward compatible. The old source files are reused to congregate the individual routine files into larger translation units (via a bunch of # includes) so that the compiler can still optimise wherever possible, e.g. through inlining, which can only be done within the same translation unit. It is also possible to build the entire library by compiling the single file named "pthread.c", which just #includes all the secondary congregation source files. The compiler may be able to use this to do more inlining of routines. Although the GNU compiler is able to produce libraries with the necessary separation (the -ffunction-segments switch), AFAIK, the MSVC and other compilers don't have this feature. Finally, since I use makefiles and command-line compilation, I don't know what havoc this reorganisation may wreak amongst IDE project file users. You should be able to continue using your existing project files without modification. New non-portable function ------------------------- pthread_num_processors_np(): Returns the number of processors in the system that are available to the process, as determined from the processor affinity mask. Platform dependence ------------------- As Win95 doesn't provide one, the library now contains it's own InterlockedCompareExchange() routine, which is used whenever Windows doesn't provide it. InterlockedCompareExchange() is used to implement spinlocks and barriers, and also in mutexes. This routine relies on the CMPXCHG machine instruction which is not available on i386 CPUs. This library (from snapshot 20010712 onwards) is therefore no longer supported on i386 processor platforms. New routines ------------ For source code portability only - rwlocks cannot be process shared yet. pthread_rwlockattr_init() pthread_rwlockattr_destroy() pthread_rwlockattr_setpshared() pthread_rwlockattr_getpshared() As defined in the new POSIX standard, and the Single Unix Spec version 3: sem_timedwait() pthread_mutex_timedlock() pthread.h no longer includes windows.h -------------------------------------- [Not yet for G++] This was done to prevent conflicts. HANDLE, DWORD, and NULL are temporarily defined within pthread.h if they are not already. Bug fixes --------- * Fixed potential NULL pointer dereferences in pthread_mutexattr_init, pthread_mutexattr_getpshared, pthread_barrierattr_init, pthread_barrierattr_getpshared, and pthread_condattr_getpshared. - Scott McCaskill * Removed potential race condition in pthread_mutex_trylock and pthread_mutex_lock; - Alexander Terekhov * The behaviour of pthread_mutex_trylock in relation to recursive mutexes was inconsistent with commercial implementations. Trylock would return EBUSY if the lock was owned already by the calling thread regardless of mutex type. Trylock now increments the recursion count and returns 0 for RECURSIVE mutexes, and will return EDEADLK rather than EBUSY for ERRORCHECK mutexes. This is consistent with Solaris. - Thomas Pfaff * Found a fix for the library and workaround for applications for the known bug #2, i.e. where __CLEANUP_CXX or __CLEANUP_SEH is defined. See the "Known Bugs in this snapshot" section below. This could be made transparent to applications by replacing the macros that define the current C++ and SEH versions of pthread_cleanup_push/pop with the C version, but AFAIK cleanup handlers would not then run in the correct sequence with destructors and exception cleanup handlers when an exception occurs. * Cancelation once started in a thread cannot now be inadvertantly double canceled. That is, once a thread begins it's cancelation run, cancelation is disabled and a subsequent cancel request will return an error (ESRCH). --------------------------- Known bugs in this snapshot --------------------------- 1. Under MS VC++ (only tested with version 6.0), a term_func set via the standard C++ set_terminate() function is not called for some reason. The code in private.c ptw32_threadStart() retrieves and calls the user supplied terminate routine, which works as expected under MinGW32 g++, but doesn't run under MS VC++ 6.0, presumably because set_terminate() returns NULL. 2. Cancellation problems in optimised code - Milan Gardian Workaround [rpj - 2 Feb 2002] ----------------------------- The problem disappears when /Ob0 is used, i.e. /O2 /Ob0 works OK, but if you want to use inlining optimisation you can be much more specific about where it's switched off and on by using a pragma. So the inlining optimisation is interfering with the way that cleanup handlers are run. It appears to relate to auto-inlining of class methods since this is the only auto inlining that is performed at /O1 optimisation (functions with the "inline" qualifier are also inlined, but the problem doesn't appear to involve any such functions in the library or testsuite). In order to confirm the inlining culprit, the following use of pragmas eliminate the problem but I don't know how to make it transparent, putting it in, say, pthread.h where pthread_cleanup_push defined as a macro. #pragma inline_depth(0) pthread_cleanup_push(handlerFunc, (void *) &arg); /* ... */ pthread_cleanup_pop(0); #pragma inline_depth() Note the empty () value after the pop macro. This resets depth to the default. Or you can specify a non-zero depth here. The pragma is also needed (and now used) within the library itself wherever cleanup handlers are used (condvar.c and rwlock.c). Use of these pragmas allows compiler optimisations /O1 and /O2 to be used for either or both the library and applications. Experimenting further, I found that wrapping the actual cleanup handler function with #pragma auto_inline(off|on) does NOT work. MSVC6.0 doesn't appear to support the C99 standard's _Pragma directive, however, later versions may. This form is embeddable inside #define macros, which would be ideal because it would mean that it could be added to the push/pop macro definitions in pthread.h and hidden from the application programmer. [/rpj] Original problem description ---------------------------- The cancellation (actually, cleanup-after-cancel) tests fail when using VC (professional) optimisation switches (/O1 or /O2) in pthreads library. I have not investigated which concrete optimisation technique causes this problem (/Og, /Oi, /Ot, /Oy, /Ob1, /Gs, /Gf, /Gy, etc.), but here is a summary of builds and corresponding failures: * pthreads VSE (optimised tests): OK * pthreads VCE (optimised tests): Failed "cleanup1" test (runtime) * pthreads VSE (DLL in CRT, optimised tests): OK * pthreads VCE (DLL in CRT, optimised tests): Failed "cleanup1" test (runtime) Please note that while in VSE version of the pthreads library the optimisation does not really have any impact on the tests (they pass OK), in VCE version addition of optimisation (/O2 in this case) causes the tests to fail uniformly - either in "cleanup0" or "cleanup1" test cases. Please note that all the tests above use default pthreads DLL (no optimisations, linked with either static or DLL CRT, based on test type). Therefore the problem lies not within the pthreads DLL but within the compiled client code (the application using pthreads -> involvement of "pthread.h"). I think the message of this section is that usage of VCE version of pthreads in applications relying on cancellation/cleanup AND using optimisations for creation of production code is highly unreliable for the current version of the pthreads library. Caveats ------- 1. Due to what is believed to be a C++ compliance error in VC++, if your application contains catch(...) blocks in your POSIX threads then you will need to replace the "catch(...)" with the macro "PtW32Catch", eg. #ifdef PtW32Catch PtW32Catch { ... } #else catch(...) { ... } #endif Otherwise neither pthreads cancelation nor pthread_exit() will work reliably. Level of standards conformance ------------------------------ The following POSIX 1003.1c/1b/1j options are defined: _POSIX_THREADS _POSIX_THREAD_SAFE_FUNCTIONS _POSIX_THREAD_ATTR_STACKSIZE _POSIX_THREAD_PRIORITY_SCHEDULING _POSIX_SEMAPHORES _POSIX_READER_WRITER_LOCKS _POSIX_SPIN_LOCKS _POSIX_BARRIERS The following POSIX 1003.1c options are not defined: _POSIX_THREAD_ATTR_STACKADDR _POSIX_THREAD_PRIO_INHERIT _POSIX_THREAD_PRIO_PROTECT _POSIX_THREAD_PROCESS_SHARED The following functions are implemented: --------------------------- PThreads --------------------------- pthread_attr_init pthread_attr_destroy pthread_attr_getdetachstate pthread_attr_getstackaddr pthread_attr_getstacksize pthread_attr_setdetachstate pthread_attr_setstackaddr pthread_attr_setstacksize pthread_create pthread_detach pthread_equal pthread_exit pthread_join pthread_once pthread_self pthread_cancel pthread_cleanup_pop pthread_cleanup_push pthread_setcancelstate pthread_setcanceltype pthread_testcancel --------------------------- Thread Specific Data --------------------------- pthread_key_create pthread_key_delete pthread_setspecific pthread_getspecific --------------------------- Mutexes --------------------------- pthread_mutexattr_init pthread_mutexattr_destroy pthread_mutexattr_getpshared pthread_mutexattr_setpshared pthread_mutexattr_gettype pthread_mutexattr_settype (types: PTHREAD_MUTEX_DEFAULT PTHREAD_MUTEX_NORMAL PTHREAD_MUTEX_ERRORCHECK PTHREAD_MUTEX_RECURSIVE ) pthread_mutex_init pthread_mutex_destroy pthread_mutex_lock pthread_mutex_trylock pthread_mutex_timedlock pthread_mutex_unlock --------------------------- Condition Variables --------------------------- pthread_condattr_init pthread_condattr_destroy pthread_condattr_getpshared pthread_condattr_setpshared pthread_cond_init pthread_cond_destroy pthread_cond_wait pthread_cond_timedwait pthread_cond_signal pthread_cond_broadcast --------------------------- Read/Write Locks - POSIX 1j --------------------------- pthread_rwlock_init pthread_rwlock_destroy pthread_rwlock_tryrdlock pthread_rwlock_trywrlock pthread_rwlock_rdlock pthread_rwlock_rwlock pthread_rwlock_unlock pthread_rwlockattr_init pthread_rwlockattr_destroy pthread_rwlockattr_getpshared pthread_rwlockattr_setpshared --------------------------- Spin Locks - POSIX 1j --------------------------- pthread_spin_init pthread_spin_destroy pthread_spin_lock pthread_spin_unlock pthread_spin_trylock --------------------------- Barriers - POSIX 1j --------------------------- pthread_barrier_init pthread_barrier_destroy pthread_barrier_wait pthread_barrierattr_init pthread_barrierattr_destroy pthread_barrierattr_getpshared pthread_barrierattr_setpshared --------------------------- Semaphores - POSIX 1b --------------------------- sem_init sem_destroy sem_post sem_wait sem_trywait sem_timedwait sem_open (returns an error ENOSYS) sem_close (returns an error ENOSYS) sem_unlink (returns an error ENOSYS) sem_getvalue (returns an error ENOSYS) --------------------------- RealTime Scheduling --------------------------- pthread_attr_getschedparam pthread_attr_setschedparam pthread_attr_getinheritsched pthread_attr_setinheritsched pthread_attr_getschedpolicy (only supports SCHED_OTHER) pthread_attr_setschedpolicy (only supports SCHED_OTHER) pthread_getschedparam pthread_setschedparam pthread_getconcurrency pthread_setconcurrency pthread_attr_getscope pthread_attr_setscope (only supports PTHREAD_SCOPE_SYSTEM) sched_get_priority_max (POSIX 1b) sched_get_priority_min (POSIX 1b) sched_rr_get_interval (POSIX 1b - returns an error ENOTSUP) sched_setscheduler (POSIX 1b - only supports SCHED_OTHER) sched_getscheduler (POSIX 1b - only supports SCHED_OTHER) sched_yield (POSIX 1b) --------------------------- Signals --------------------------- pthread_sigmask --------------------------- Non-portable routines (see the README.NONPORTABLE file for usage) --------------------------- pthread_getw32threadhandle_np pthread_delay_np pthread_mutexattr_getkind_np pthread_mutexattr_setkind_np (types: PTHREAD_MUTEX_FAST_NP, PTHREAD_MUTEX_ERRORCHECK_NP, PTHREAD_MUTEX_RECURSIVE_NP, PTHREAD_MUTEX_ADAPTIVE_NP, PTHREAD_MUTEX_TIMED_NP) pthread_num_processors_np pthread_win32_process_attach_np (Required when statically linking the library) pthread_win32_process_detach_np (Required when statically linking the library) pthread_win32_thread_attach_np (Required when statically linking the library) pthread_win32_thread_detach_np (Required when statically linking the library) --------------------------- Static Initializers --------------------------- PTHREAD_ONCE_INIT PTHREAD_MUTEX_INITIALIZER PTHREAD_COND_INITIALIZER PTHREAD_RWLOCK_INITIALIZER PTHREAD_SPINLOCK_INITIALIZER --------------------------- Thread-Safe C Runtime Library (macros) --------------------------- strtok_r asctime_r ctime_r gmtime_r localtime_r rand_r The following functions are not implemented: --------------------------- RealTime Scheduling --------------------------- pthread_mutex_getprioceiling pthread_mutex_setprioceiling pthread_mutex_attr_getprioceiling pthread_mutex_attr_getprotocol pthread_mutex_attr_setprioceiling pthread_mutex_attr_setprotocol --------------------------- Fork Handlers --------------------------- pthread_atfork --------------------------- Stdio --------------------------- flockfile ftrylockfile funlockfile getc_unlocked getchar_unlocked putc_unlocked putchar_unlocked --------------------------- Thread-Safe C Runtime Library --------------------------- readdir_r getgrgid_r getgrnam_r getpwuid_r getpwnam_r --------------------------- Signals --------------------------- pthread_kill sigtimedwait sigwait sigwaitinfo The library includes two non-API functions for creating cancellation points in applications and libraries: pthreadCancelableWait pthreadCancelableTimedWait Availability ------------ The prebuilt DLL, export libs (for both MSVC and Mingw32), and the header files (pthread.h, semaphore.h, sched.h) are available along with the complete source code. The source code can be found at: ftp://sources.redhat.com/pub/pthreads-win32 and as individual source code files at ftp://sources.redhat.com/pub/pthreads-win32/source The pre-built DLL, export libraries and include files can be found at: ftp://sources.redhat.com/pub/pthreads-win32/dll-latest Mailing List ------------ There is a mailing list for discussing pthreads on Win32. To join, send email to: pthreads-win32-subscribe@sourceware.cygnus.com Application Development Environments ------------------------------------ See the README file for more information. MSVC: MSVC using SEH works. Distribute pthreadVSE.dll with your application. MSVC using C++ EH works. Distribute pthreadVCE.dll with your application. MSVC using C setjmp/longjmp works. Distribute pthreadVC.dll with your application. Mingw32: See FAQ Questions 6 and 10. Mingw using C++ EH works. Distribute pthreadGCE.dll with your application. Mingw using C setjmp/longjmp works. Distribute pthreadGC.dll with your application. Cygwin: (http://sourceware.cygnus.com/cygwin/) Developers using Cygwin will not need pthreads-win32 since it has POSIX threads support. Refer to its documentation for details and extent. UWIN: UWIN is a complete Unix-like environment for Windows from AT&T. Pthreads-win32 doesn't currently support UWIN (and vice versa), but that may change in the future. Generally: For convenience, the following pre-built files are available on the FTP site (see Availability above): pthread.h - for POSIX 1c threads semaphore.h - for POSIX 1b semaphores sched.h - for POSIX 1b scheduling pthreadVCE.dll - built with MSVC++ compiler using C++ EH pthreadVCE.lib pthreadVC.dll - built with MSVC compiler using C setjmp/longjmp pthreadVC.lib pthreadVSE.dll - built with MSVC compiler using SEH pthreadVSE.lib pthreadGCE.dll - built with Mingw32 G++ 2.95.2-1 pthreadGC.dll - built with Mingw32 GCC 2.95.2-1 using setjmp/longjmp libpthreadGCE.a - derived from pthreadGCE.dll libpthreadGC.a - derived from pthreadGC.dll gcc.dll - needed if distributing applications that use pthreadGCE.dll These are the only files you need in order to build POSIX threads applications for Win32 using either MSVC or Mingw32. See the FAQ file in the source tree for additional information. Documentation ------------- Currently, there is no documentation included in the package apart from the copious comments in the source code. For POSIX Thread API programming, several reference books are available: Programming with POSIX Threads David R. Butenhof Addison-Wesley (pub) Pthreads Programming By Bradford Nichols, Dick Buttlar & Jacqueline Proulx Farrell O'Reilly (pub) On the web: see the links at the bottom of the pthreads-win32 site: http://sources.redhat.com/pthreads-win32/ Acknowledgements ---------------- This library is based substantially on a Win32 pthreads implementation contributed by John Bossom . The implementation of Condition Variables uses algorithms developed by Alexander Terekhov and Louis Thomas. The implementation of POSIX mutexes has been improved by Thomas Pfaff. The implementation of read/write locks was contributed by Aurelio Medina and improved by Alexander Terekhov. Many others have contributed significant time and effort to solve critical problems in order to make the library workable, robust and reliable. There is also a separate CONTRIBUTORS file. This file and others are on the web site: http://sources.redhat.com/pthreads-win32 As much as possible, the ChangeLog file acknowledges contributions to the code base in more detail. Enjoy! Ross Johnson