summaryrefslogtreecommitdiff
path: root/mutex.c
AgeCommit message (Collapse)Author
2011-03-24First pass of robust mutexesrpj
2007-01-06See ChangeLogrpj
2005-01-01Copyright year updatedrpj
2004-11-22Copyright updaterpj
2004-05-17re-indentation, bug fixes, hooks for pre-emptive async cancelationrpj
2003-08-19Updated Copyright notice (date and contact address).rpj
2002-02-18Major reorganisation of source code; new routine and tests added.rpj
2002-02-08Copyright notice updates; module splitting; POSIX versionrpj
conditional compilation; refinements; bug fixes. See the ChangeLog.
2002-02-02 * cancel.c: Rearranged some code and introduced checksrpj
to disable cancelation at the start of a thread's cancelation run to prevent double cancelation. The main problem arises if a thread is canceling and then receives a subsequent async cancel request. * private.c: Likewise. * condvar.c: Place pragmas around cleanup_push/pop to turn off inline optimisation (/Obn where n>0 - MSVC only). Various optimisation switches in MSVC turn this on, which interferes with the way that cleanup handlers are run in C++ EH and SEH code. Application code compiled with inline optimisation must also wrap cleanup_push/pop blocks with the pragmas, e.g. #pragma inline_depth(0) pthread_cleanup_push(...) ... pthread_cleanup_pop(...) #pragma inline_depth(8) * rwlock.c: Likewise. * mutex.c: Remove attempts to inline some functions. * signal.c: Modify misleading comment. tests/ * mutex8: New test. * mutex8n: New test. * mutex8e: New test. * mutex8r: New test. * cancel6a: New test. * cancel6d: New test. * cleanup0.c: Add pragmas for inline optimisation control. * cleanup1.c: Add pragmas for inline optimisation control. * cleanup2.c: Add pragmas for inline optimisation control. * cleanup3.c: Add pragmas for inline optimisation control. * condvar7.c: Add pragmas for inline optimisation control. * condvar8.c: Add pragmas for inline optimisation control. * condvar9.c: Add pragmas for inline optimisation control.
2002-02-01 * semaphore.c (sem_trywait): Fix missing errno returnrpj
for systems that define NEED_SEM (e.g. early WinCE). * mutex.c (pthread_mutex_timedlock): Return ENOTSUP for systems that define NEED_SEM since they don't have sem_trywait().
2002-02-01pthread_mutex_timedlock: Add missing InterlockedDecrement(lock_idx)rpj
for the case where abstime has passed already. Make routines INLINEable since some are called by other routines in this module.
2002-01-31Fix critical section init.rpj
2002-01-312002-01-27 Ross Johnson <rpj@special.ise.canberra.edu.au>rpj
* mutex.c (pthread_mutex_timedlock): New function suggested by Alexander Terekhov. The logic required to implement this properly came from Alexander, with some collaboration with Thomas Pfaff. (pthread_mutex_unlock): Wrap the waiters check and sema post in a critical section to prevent a race with pthread_mutex_timedlock. (ptw32_timed_semwait): New function; returns a special result if the absolute timeout parameter represents a time already passed when called; used by pthread_mutex_timedwait(). Have deliberately not reused the name "ptw32_sem_timedwait" because they are not the same routine. * condvar.c (ptw32_cond_timedwait): Use the new sem_timedwait() instead of ptw32_sem_timedwait(), which now has a different function. See previous. * implement.h: Remove prototype for ptw32_sem_timedwait. See next. (pthread_mutex_t_): Add critical section element for access to lock_idx during mutex post-timeout processing. * semaphore.h (sem_timedwait): See next. * semaphore.c (sem_timedwait): See next. * private.c (ptw32_sem_timedwait): Move to semaphore.c and rename as sem_timedwait(). 2002-01-18 Ross Johnson <rpj@special.ise.canberra.edu.au> * sync.c (pthread_join): Was getting the exit code from the calling thread rather than the joined thread if defined(__MINGW32__) && !defined(__MSVCRT__). 2002-01-15 Ross Johnson <rpj@special.ise.canberra.edu.au> * pthread.h: Unless the build explicitly defines __CLEANUP_SEH, __CLEANUP_CXX, or __CLEANUP_C, then the build defaults to __CLEANUP_C style cleanup. This style uses setjmp/longjmp in the cancelation and thread exit implementations and therefore won't do stack unwinding if linked to applications that have it (e.g. C++ apps). This is currently consistent with most/all commercial Unix POSIX threads implementations. * spin.c (pthread_spin_init): Edit renamed function call. * nonportable.c (pthread_num_processors_np): New. (pthread_getprocessors_np): Renamed to ptw32_getprocessors and moved to private.c. * private.c (pthread_getprocessors): Moved here from nonportable.c. * pthread.def (pthread_getprocessors_np): Removed from export list. * rwlock.c (pthread_rwlockattr_init): New. (pthread_rwlockattr_destroy): New. (pthread_rwlockattr_getpshared): New. (pthread_rwlockattr_setpshared): New.
2002-01-08 * mutex.c (pthread_mutex_trylock): userpj
ptw32_interlocked_compare_exchange function pointer rather than ptw32_InterlockedCompareExchange() directly to retain portability to non-iX86 processors, e.g. WinCE etc. The pointer will point to the native OS version of InterlockedCompareExchange() if the OS supports it (see ChangeLog entry of 2001-10-17).
2002-01-07 * mutex.c (pthread_mutex_init): Remove criticalrpj
section calls. (pthread_mutex_destroy): Likewise. (pthread_mutex_unlock): Likewise. (pthread_mutex_trylock): Likewise; recursive mutexes now increment the lock count rather than return EBUSY; errorcheck mutexes return EDEADLCK rather than EBUSY. This behaviour is consistent with the Solaris pthreads implementation. * implement.h (pthread_mutex_t_): Remove critical section element - no longer needed.
2001-12-20 * mutex.c (pthread_mutex_lock): Decrementing lock_idx wasrpj
not thread-safe. (pthread_mutex_trylock): Likewise.
2001-12-18Include an interim complete Copyright notice in each source file;rpj
Remove unused autoconf files; Add the file COPYING, which describes the pthread-win32 license relative to the FSF LGPL.
2001-10-15 * spin.c (pthread_spin_lock): PTHREAD_SPINLOCK_INITIALIZERrpj
was causing a program fault. (pthread_spin_init): Could have alloced memory without freeing under some error conditions. * mutex.c (pthread_mutex_init): Move memory allocation of mutex struct after checking for PROCESS_SHARED.
2001-07-18Fix potential NULL pointer dereferences.rpj
Contributed by - "Scott McCaskill" <scott@magruder.org> * mutex.c (pthread_mutexattr_init): Return ENOMEM immediately and don't dereference the NULL pointer if calloc fails. (pthread_mutexattr_getpshared): Don't dereference a pointer that is possibly NULL. * barrier.c (pthread_barrierattr_init): Likewise (pthread_barrierattr_getpshared): Don't dereference a pointer that is possibly NULL. * condvar.c (pthread_condattr_getpshared): Don't dereference a pointer that is possibly NULL.
2001-07-18 * mutex.c (pthread_mutexattr_init): Return ENOMEMrpj
immediately and don't dereference the NULL pointer if calloc fails. - "Scott McCaskill" <scott@magruder.org>
2001-07-06Spinlocks and barriers fixed and working. Beta level.rpj
* spin.c: Revamped and working; included static initialiser. * barrier.c: Likewise. * condvar.c: Macro constant change; inline auto init routine. * mutex.c: Likewise. * rwlock.c: Likewise. * private.c: Add support for spinlock initialiser. * global.c: Likewise. * implement.h: Likewise. * pthread.h (PTHREAD_SPINLOCK_INITIALIZER): Fix typo. tests/ChangeLog: * spin3.c: Changed test and fixed. * spin4.c: Fixed. * barrier3.c: Fixed. * barrier4.c: Fixed.
2001-07-05Added new routines from POSIX 1003.1j. This is alpha level code.rpj
* spin.c: New module implementing spin locks. * barrier.c: New module implementing barriers. * pthread.h (_POSIX_SPIN_LOCKS): defined. (_POSIX_BARRIERS): Defined. (pthread_spin_*): Defined. (pthread_barrier*): Defined. (PTHREAD_BARRIER_SERIAL_THREAD): Defined. * implement.h (pthread_spinlock_t_): Defined. (pthread_barrier_t_): Defined. (pthread_barrierattr_t_): Defined. * mutex.c (pthread_mutex_lock): Return with the error if an auto-initialiser initialisation fails. * nonportable.c (pthread_getprocessors_np): New; gets the number of available processors for the current process.
2001-07-012001-06-25 Ross Johnson <rpj@setup1.ise.canberra.edu.au>rpj
* create.c (pthread_create): Add priority inheritance attributes. * mutex.c (pthread_mutex_lock): Remove some overhead for PTHREAD_MUTEX_NORMAL mutex types. Specifically, avoid calling pthread_self() and pthread_equal() to check/set the mutex owner. Introduce a new pseudo owner for this type. Test results suggest increases in speed of up to 90% for non-blocking locks. This is the default type of mutex used internally by other synchronising objects, ie. condition variables and read-write locks. The test rwlock7.c shows about a 30-35% speed increase over snapshot 2001-06-06. The price of this is that the application developer must ensure correct behaviour, or explicitly set the mutex to a safer type such as PTHREAD_MUTEX_ERRORCHECK. For example, PTHREAD_MUTEX_NORMAL (or PTHREAD_MUTEX_DEFAULT) type mutexes will not return an error if a thread which is not the owner calls pthread_mutex_unlock. The call will succeed in unlocking the mutex if it is currently locked, but a subsequent unlock by the true owner will then fail with EPERM. This is however consistent with some other implementations. (pthread_mutex_unlock): Likewise. (pthread_mutex_trylock): Likewise. (pthread_mutex_destroy): Likewise. * attr.c (pthread_attr_init): PTHREAD_EXPLICIT_SCHED is the default inheritance attribute; THREAD_PRIORITY_NORMAL is the default priority for new threads. * sched.c (pthread_attr_setschedpolicy): Added routine. (pthread_attr_getschedpolicy): Added routine. (pthread_attr_setinheritsched): Added routine. (pthread_attr_getinheritsched): Added routine. * pthread.h (sched_rr_set_interval): Added as a macro; returns -1 with errno set to ENOSYS. 2001-06-23 Ross Johnson <rpj@setup1.ise.canberra.edu.au> *sched.c (pthread_attr_setschedparam): Add priority range check. (sched_setscheduler): New function; checks for a valid pid and policy; checks for permission to set information in the target process; expects pid to be a Win32 process ID, not a process handle; the only scheduler policy allowed is SCHED_OTHER. (sched_getscheduler): Likewise, but checks for permission to query. * pthread.h (SCHED_*): Moved to sched.h as defined in the POSIX standard. * sched.h (SCHED_*): Moved from pthread.h. (pid_t): Defined if necessary. (sched_setscheduler): Defined. (sched_getscheduler): Defined. * pthread.def (sched_setscheduler): Exported. (sched_getscheduler): Likewise. 2001-06-23 Ross Johnson <rpj@setup1.ise.canberra.edu.au> Contributed by - Ralf Brese <Ralf.Brese@pdb4.siemens.de> * create.c (pthread_create): Set thread priority from thread attributes.
2001-06-05 * mutex.c (pthread_mutexattr_init): Removesnap-2001-06-06rpj
ptw32_mutex_default_kind.
2001-06-05 * nonportable.c (pthread_mutex_setdefaultkind_np):rpj
Remove - should not have been included in the first place. (pthread_mutex_getdefaultkind_np): Likewise. * global.c (ptw32_mutex_default_kind): Likewise. * mutex.c (pthread_mutex_init): Remove use of ptw32_mutex_default_kind. * pthread.h (pthread_mutex_setdefaultkind_np): Likewise. (pthread_mutex_getdefaultkind_np): Likewise. * pthread.def (pthread_mutexattr_setkind_np): Added. (pthread_mutexattr_getkind_np): Likewise. * README: Many changes that should have gone in before the last snapshot. * README.NONPORTABLE: New - referred to by ANNOUNCE but never created; documents the non-portable routines included in the library - moved from README with new routines added. * ANNOUNCE (pthread_mutexattr_setkind_np): Added to compliance list. (pthread_mutexattr_getkind_np): Likewise.
2001-06-05Update some comments.rpj
2001-05-312001-05-30 Ross Johnson <rpj@setup1.ise.canberra.edu.au>rpj
* pthread.h (rand_r): Fake using _seed argument to quell compiler warning (compiler should optimise this away later). * GNUmakefile (OPT): Leave symbolic information out of the library and increase optimisation level - for smaller faster prebuilt dlls. 2001-05-29 Ross Johnson <rpj@setup1.ise.canberra.edu.au> Contributed by - Milan Gardian <Milan.Gardian@LEIBINGER.com> * Makefile: fix typo. * pthreads.h: Fix problems with stdcall/cdecl conventions, in particular remove the need for PT_STDCALL everywhere; remove warning supression. * (errno): Fix the longstanding "inconsistent dll linkage" problem with errno; now also works with /MD debugging libs - warnings emerged when compiling pthreads library with /MD (or /MDd) compiler switch, instead of /MT (or /MTd) (i.e. when compiling pthreads using Multithreaded DLL CRT instead of Multithreaded statically linked CRT). * create.c (pthread_create): Likewise; fix typo. * private.c (ptw32_threadStart): Eliminate use of terminate() which doesn't throw exceptions. * Remove unnecessary #includes from a number of modules - [I had to #include malloc.h in implement.h for gcc - rpj]. 2001-05-29 Ross Johnson <rpj@setup1.ise.canberra.edu.au> Contributed by - Thomas Pfaff <tpfaff@gmx.net> * pthread.h (PTHREAD_MUTEX_DEFAULT): New; equivalent to PTHREAD_MUTEX_DEFAULT_NP. * (PTHREAD_MUTEX_NORMAL): Similarly. * (PTHREAD_MUTEX_ERRORCHECK): Similarly. * (PTHREAD_MUTEX_RECURSIVE): Similarly. * (pthread_mutex_setdefaultkind_np): New; Linux compatibility stub for pthread_mutexattr_settype. * (pthread_mutexattr_getkind_np): New; Linux compatibility stub for pthread_mutexattr_gettype. * mutex.c (pthread_mutexattr_settype): New; allow the following types of mutex: PTHREAD_MUTEX_DEFAULT_NP PTHREAD_MUTEX_NORMAL_NP PTHREAD_MUTEX_ERRORCHECK_NP PTHREAD_MUTEX_RECURSIVE_NP * Note that PTHREAD_MUTEX_DEFAULT is equivalent to PTHREAD_MUTEX_NORMAL - ie. mutexes should no longer be recursive by default, and a thread will deadlock if it tries to relock a mutex it already owns. This is inline with other pthreads implementations. * (pthread_mutex_lock): Process the lock request according to the mutex type. * (pthread_mutex_init): Eliminate use of Win32 mutexes as the basis of POSIX mutexes - instead, a combination of one critical section and one semaphore are used in conjunction with Win32 Interlocked* routines. * (pthread_mutex_destroy): Likewise. * (pthread_mutex_lock): Likewise. * (pthread_mutex_trylock): Likewise. * (pthread_mutex_unlock): Likewise. * Use longjmp/setjmp to implement cancelation when building the library using a C compiler which doesn't support exceptions, e.g. gcc -x c (note that gcc -x c++ uses exceptions). * Also fixed some of the same typos and eliminated PT_STDCALL as Milan Gardian's patches above. 2001-02-07 Ross Johnson <rpj@special.ise.canberra.edu.au> Contributed by - Alexander Terekhov <TEREKHOV@de.ibm.com> * rwlock.c: Revamped. * implement.h (pthread_rwlock_t_): Redefined. This implementation does not have reader/writer starvation problem. Rwlock attempts to behave more like a normal mutex with races and scheduling policy determining who is more important; It also supports recursive locking, has less synchronization overhead (no broadcasts at all, readers are not blocked on any condition variable) and seem to be faster than the current implementation [W98 appears to be approximately 15 percent faster at least - on top of speed increase from Thomas Pfaff's changes to mutex.c - rpj].
2000-12-29./ChangeLog:rpj
2000-12-29 Ross Johnson <rpj@special.ise.canberra.edu.au> * Makefile: Back-out "for" loops which don't work. * GNUmakefile: Remove the fake.a target; add the "realclean" target; don't remove built libs under the "clean" target. * config.h: Add a guard against multiple inclusion. * semaphore.h: Add some defines from config.h to make semaphore.h independent of config.h when building apps. * pthread.h (_errno): Back-out previous fix until we know how to fix it properly. * implement.h (lockCount): Add missing element to pthread_mutex_t_. * sync.c (pthread_join): Spelling fix in comment. * private.c (ptw32_threadStart): Reset original termination function (C++). (ptw32_threadStart): Cleanup detached threads early in case the library is statically linked. (ptw32_callUserDestroyRoutines): Remove [SEH] __try block from destructor call so that unhandled exceptions will be passed through to the system; call terminate() from [C++] try block for the same reason. * tsd.c (pthread_getspecific): Add comment. * mutex.c (pthread_mutex_init): Initialise new elements in pthread_mutex_t. (pthread_mutex_unlock): Invert "pthread_equal()" test. 2000-12-28 Ross Johnson <rpj@special.ise.canberra.edu.au> * semaphore.c (mode_t): Use ifndef HAVE_MODE_T to include definition. * config.h.in (HAVE_MODE_T): Added. (_UWIN): Start adding defines for the UWIN package. ./tests/ChangeLog: 2000-12-29 Ross Johnson <rpj@special.ise.canberra.edu.au> * GNUmakefile: Add mutex4 test; ensure libpthreadw32.a is removed for "clean" target. * Makefile: Add mutex4 test. * exception3.c: Remove SEH code; automatically pass the test under SEH (which is an N/A environment). * mutex4.c: New test. * eyal1.c (do_work_unit): Add a dummy "if" to force the optimiser to retain code; reduce thread work loads. * condvar8.c (main): Add an additional "assert" for debugging; increase pthread_cond_signal timeout.
2000-12-282000-10-10 Ross Johnson <rpj@setup1.ise.canberra.edu.au>rpj
* misc.c (pthread_self): Restore Win32 "last error" cleared by TlsGetValue() call in pthread_getspecific() - "Steven Reddie" <smr@essemer.com.au> 2000-09-20 Ross Johnson <rpj@setup1.ise.canberra.edu.au> * mutex.c (pthread_mutex_lock): Record the owner of the mutex. This requires also keeping count of recursive locks ourselves rather than leaving it to Win32 since we need to know when to NULL the thread owner when the mutex is unlocked. (pthread_mutex_trylock): Likewise. (pthread_mutex_unlock): Check that the calling thread owns the mutex, decrement the recursive lock count, and NULL the owner if zero. Return EPERM if the mutex is owned by another thread. * implement.h (pthread_mutex_t_): Add ownerThread and lockCount members. - reported by Arthur Kantor <akantor@bexusa.com>
2000-09-132000-09-13 Ross Johnson <rpj@setup1.ise.canberra.edu.au>rpj
* mutex.c (pthread_mutex_init): Call TryEnterCriticalSection through the pointer rather than directly so that the dll can load on Windows versions that can't resolve the function, eg. Windows 95 - "Jef Gearhart" <jgearhart@tpssys.com>
2000-09-022000-09-02 Ross Johnson <rpj@special.ise.canberra.edu.au>rpj
* rwlock.c (_rwlock_check*): Renamed to ptw32_rwlock_check*. * mutex.c (_mutex_check*): Renamed to ptw32_mutex_check*. * condvar.c (cond_timed*): Renamed to ptw32_cond_timed*. (_cond_check*): Renamed to ptw32_cond_check*.
2000-08-182000-08-18 Ross Johnson <rpj@special.ise.canberra.edu.au>rpj
* mutex.c (pthread_mutex_destroy): Check that the mutex isn't held; invalidate the mutex as early as possible to avoid contention; not perfect - FIXME! * rwlock.c (pthread_rwlock_init): Remove redundant assignment to "rw". (pthread_rwlock_destroy): Invalidate the rwlock before freeing up any of it's resources - to avoid contention. * private.c (ptw32_tkAssocCreate): Change assoc->lock to use a dynamically initialised mutex - only consumes a W32 mutex or critical section when first used, not before. * mutex.c (pthread_mutex_init): Remove redundant assignment to "mx". (pthread_mutexattr_destroy): Set attribute to NULL before freeing it's memory - to avoid contention.
2000-08-182000-08-18 Ross Johnson <rpj@setup1.ise.canberra.edu.au>rpj
* Several: Fix typos from scripted edit session yesterday. * nonportable.c (pthread_mutexattr_setforcecs_np): Moved this function from mutex.c. (pthread_getw32threadhandle_np): New function to return the win32 thread handle that the POSIX thread is using. * mutex.c (pthread_mutexattr_setforcecs_np): Moved to new file "nonportable.c". * pthread.h (PTW32_BUILD): Only redefine __except and catch compiler keywords if we aren't building the library (ie. PTW32_BUILD is not defined) - this is safer than defining and then undefining if not building the library. * implement.h: Remove __except and catch undefines. * Makefile (CFLAGS): Define PTW32_BUILD. * GNUmakefile (CFLAGS): Define PTW32_BUILD. * All appropriate: Change Pthread_exception* to ptw32_exception* to be consistent with internal identifier naming. * private.c (ptw32_throw): New function to provide a generic exception throw for all internal exceptions and EH schemes. (ptw32_threadStart): pthread_exit() value is now returned via the thread structure exitStatus element. * exit.c (pthread_exit): pthread_exit() value is now returned via the thread structure exitStatus element. * cancel.c (ptw32_cancel_self): Now uses ptw32_throw. (pthread_setcancelstate): Ditto. (pthread_setcanceltype): Ditto. (pthread_testcancel): Ditto. (pthread_cancel): Ditto. * misc.c (CancelableWait): Ditto. * exit.c (pthread_exit): Ditto. * All applicable: Change PTW32_ prefix to PTW32_ prefix to remove leading underscores from private library identifiers.
2000-08-172000-08-17 Ross Johnson <rpj@special.ise.canberra.edu.au>rpj
* All applicable: Change _pthread_ prefix to ptw32_ prefix to remove leading underscores from private library identifiers (single and double leading underscores are reserved in the ANSI C standard for compiler implementations).
2000-07-252000-07-25 Ross Johnson <rpj@special.ise.canberra.edu.au>rpj
* sched.c (sched_get_priority_max): Handle different WinCE and Win32 priority values together. (sched_get_priority_min): Ditto. - Tristan Savatier <tristan@mpegtv.com> * create.c (pthread_create): Force new threads to wait until pthread_create has the new thread's handle; we also retain a local copy of the handle for internal use until pthread_create returns. * private.c (_pthread_threadStart): Initialise ei[]. (_pthread_threadStart): When beginthread is used to start the thread, force waiting until the creator thread had the thread handle. * cancel.c (_pthread_cancel_thread): Include context switch code for defined(_X86_) environments in addition to _M_IX86. * rwlock.c (pthread_rwlock_destroy): Assignment changed to avoid compiler warning. * private.c (_pthread_get_exception_services_code): Cast NULL return value to avoid compiler warning. * cleanup.c (pthread_pop_cleanup): Initialise "cleanup" variable to avoid compiler warnings. * misc.c (_pthread_new): Change "new" variable to "t" to avoid confusion with the C++ keyword of the same name. * condvar.c (cond_wait_cleanup): Initialise lastWaiter variable. (cond_timedwait): Remove unused local variables. to avoid compiler warnings. * dll.c (dllMain): Remove 2000-07-21 change - problem appears to be in pthread_create(). 2000-07-22 Ross Johnson <rpj@special.ise.canberra.edu.au> * tsd.c (pthread_key_create): If a destructor was given and the pthread_mutex_init failed, then would try to reference a NULL pointer (*key); eliminate this section of code by using a dynamically initialised mutex (PTHREAD_MUTEX_INITIALIZER). * tsd.c (pthread_setspecific): Return an error if unable to set the value; simplify cryptic conditional. * tsd.c (pthread_key_delete): Locking threadsLock relied on mutex_lock returning an error if the key has no destructor. ThreadsLock is only initialised if the key has a destructor. Making this mutex a static could reduce the number of mutexes used by an application since it is actually created only at first use and it's often destroyed soon after. 2000-07-22 Ross Johnson <rpj@special.ise.canberra.edu.au> * FAQ: Added Q5 and Q6. tests/ChangeLog: 2000-07-25 Ross Johnson <rpj@special.ise.canberra.edu.au> * runtest.bat: modified to work under W98. * runall.bat: Add new tests; modified to work under W98. It was ok under NT. * Makefile: Add new tests. * exception1.c: New; Test passing exceptions back to the application and retaining library internal exceptions. * join0.c: New; Test a single join.
1999-11-041999-11-05 Ross Johnson <rpj@ixobrychus.canberra.edu.au>rpj
* general: Patched for portability to WinCE. The details are described in the file WinCE-PORT. Follow the instructions in README.WinCE to make the appropriate changes in config.h. - Tristan Savatier <tristan@mpegtv.com>
1999-09-16Thu Sep 16 1999 Ross Johnson <rpj@swan.canberra.edu.au>rpj
* rwlock.c (pthread_rwlock_destroy): Add serialisation. (_rwlock_check_need_init): Check for detroyed rwlock. * rwlock.c: Check return codes from _rwlock_check_need_init(); modify comments; serialise access to rwlock objects during operations; rename rw_mutex to rw_lock. * implement.h: Rename rw_mutex to rw_lock. * mutex.c (pthread_mutex_destroy): Add serialisation. (_mutex_check_need_init): Check for detroyed mutex. * condvar.c (pthread_cond_destroy): Add serialisation. (_cond_check_need_init): Check for detroyed condvar. * mutex.c: Modify comments. * condvar.c: Modify comments.
1999-09-11Sat Sep 10 12:56:13 1999 Ross Johnson <rpj@swan.canberra.edu.au>rpj
* mutex.c (pthread_mutex_destroy): Don't free mutex memory if mutex is PTHREAD_MUTEX_INITIALIZER and has not been initialised yet.
1999-09-10Wed Sep 8 12:56:13 1999 Ross Johnson <rpj@swan.canberra.edu.au>rpj
* mutex.c (pthread_mutex_destroy): Free mutex memory. - Milan Gardian <mg@tatramed.sk>
1999-03-16Tue Mar 16 1999 Ross Johnson <rpj@ixobrychus.canberra.edu.au>rpj
* all: Add GNU LGPL and Copyright and Warranty.
1999-03-11Thu Mar 11 09:01:48 1999 Ross Johnson <rpj@ixobrychus.canberra.edu.au>rpj
* pthread.h (pthread_mutex_t): revert to (pthread_mutex_t *); define a value to serve as PTHREAD_MUTEX_INITIALIZER. (pthread_mutex_t_): remove staticinit and valid elements. (pthread_cond_t): revert to (pthread_cond_t_ *); define a value to serve as PTHREAD_COND_INITIALIZER. (pthread_cond_t_): remove staticinit and valid elements. * mutex.c (pthread_mutex_t args): adjust indirection of references. (all functions): check for PTHREAD_MUTEX_INITIALIZER value; check for NULL (invalid). * condvar.c (pthread_cond_t args): adjust indirection of references. (all functions): check for PTHREAD_COND_INITIALIZER value; check for NULL (invalid). Wed Mar 10 17:18:12 1999 Ross Johnson <rpj@ixobrychus.canberra.edu.au> * misc.c (CancelableWait): Undo changes from Mar 8 and 7. tests/ChangeLog Fri Mar 12 08:34:15 1999 Ross Johnson <rpj@ixobrychus.canberra.edu.au> * eyal1.c (main): Fix trylock loop; was not waiting for thread to lock the "started" mutex.
1999-03-07Mon Mar 8 11:18:59 1999 Ross Johnson <rpj@ixobrychus.canberra.edu.au>rpj
* misc.c (CancelableWait): Ensure cancelEvent handle is the lowest indexed element in the handles array. Enhance test for abandoned objects. * pthread.h (PTHREAD_MUTEX_INITIALIZER): Trailing elements not initialised are set to zero by the compiler. This avoids the problem of initialising the opaque critical section element in it. (PTHREAD_COND_INITIALIZER): Ditto. * semaphore.c (_pthread_sem_timedwait): Check sem == NULL earlier. Sun Mar 7 12:31:14 1999 Ross Johnson <rpj@ixobrychus.canberra.edu.au> * condvar.c (pthread_cond_init): set semaphore initial value to 0, not 1. cond_timedwait was returning signaled immediately. * misc.c (CancelableWait): Place the cancel event handle first in the handle table for WaitForMultipleObjects. This ensures that the cancel event is recognised and acted apon if both objects happen to be signaled together. * private.c (_pthread_cond_test_init_lock): Initialise and destroy. * implement.h (_pthread_cond_test_init_lock): Add extern. * global.c (_pthread_cond_test_init_lock): Add declaration. * condvar.c (pthread_cond_destroy): check for valid initialised CV; flag destroyed CVs as invalid. (pthread_cond_init): pthread_cond_t is no longer just a pointer. This is because PTHREAD_COND_INITIALIZER needs state info to reside in pthread_cond_t so that it can initialise on first use. Will work on making pthread_cond_t (and other objects like it) opaque again, if possible, later. (cond_timedwait): add check for statically initialisation of CV; initialise on first use. (pthread_cond_signal): check for valid CV. (pthread_cond_broadcast): check for valid CV. (_cond_check_need_init): Add. * pthread.h (PTHREAD_COND_INITIALIZER): Fix. (pthread_cond_t): no longer a pointer to pthread_cond_t_. (pthread_cond_t_): add 'staticinit' and 'valid' elements. tests/ChangeLog Sun Mar 7 10:41:52 1999 Ross Johnson <rpj@ixobrychus.canberra.edu.au> * Makefile (condvar3, condvar4): Add tests. * condvar4.c (General): Reduce to simple test case; prerequisite is condvar3.c; add description. * condvar3.c (General): Reduce to simple test case; prerequisite is condvar2.c; add description. * condvar2.c (General): Reduce to simple test case; prerequisite is condvar1.c; add description. * condvar1.c (General): Reduce to simple test case; add description. * Template.c (Comments): Add generic test detail.
1999-02-20Sat Feb 20 16:03:30 1999 Ross Johnson <rpj@ixobrychus.canberra.edu.au>rpj
* dll.c (DLLMain): Expand TryEnterCriticalSection support test. * mutex.c (pthread_mutex_trylock): The check for _pthread_try_enter_critical_section == NULL should have been removed long ago.
1999-02-19Fri Feb 19 16:2s1703:30 1999 Ross Johnson <rpj@ixobrychus.canberra.edu.au>rpj
* sync.c (pthread_join): Fix pthread_equal() test. * mutex.c (pthread_mutex_trylock): Check mutex != NULL before using it.
1999-02-02Mods to compile under Mingw32 egcs.rpj
Tue Feb 2 18:07:43 1999 Ross Johnson <rpj@swan.canberra.edu.au> * implement.h: Add #include <pthread.h>. Change sem_t to _pthread_sem_t. Various patches by Kevin Ruland <Kevin.Ruland@anheuser-busch.com> * signal.c (pthread_sigmask): Add and modify casts. Reverse LHS/RHS bitwise assignments. * pthread.h: Remove #include <semaphore.h>. (_PTHREAD_ATTR_VALID): Add cast. (struct pthread_t_): Add sigmask element. * dll.c: Add "extern C" for DLLMain. (DllMain): Add cast. * create.c (pthread_create): Set sigmask in thread. * condvar.c: Remove #include. Change sem_* to _pthread_sem_*. * attr.c: Changed #include. * Makefile.in: Additional targets and changes to build the library as a DLL. Fri Jan 29 11:56:28 1999 Ross Johnson <rpj@swan.canberra.edu.au> * Makefile.in (OBJS): Add semaphore.o to list. * semaphore.c (_pthread_sem_timedwait): Move from private.c. Rename sem_* to _pthread_sem_*. * pthread.h (pthread_cond_t): Change type of sem_t. _POSIX_SEMAPHORES no longer defined. * semaphore.h: Contents moved to implement.h. Removed from source tree. * implement.h: Add semaphore function prototypes and ename all functions to prepend '_pthread_'. They are now private to the pthreads-win32 implementation. * private.c: Change #warning. Move _pthread_sem_timedwait() to semaphore.c. * cleanup.c: Change #warning. * misc.c: Remove #include <errno.h> * pthread.def: Cleanup CVS merge conflicts. * global.c: Ditto. * ChangeLog: Ditto. * cleanup.c: Ditto.
1999-01-23Fixing problems building with GNU compilers:rpj
Sun Jan 24 01:34:52 1999 Ross Johnson <rpj@ixobrychus.canberra.edu.au> * semaphore.c (sem_wait): Remove second arg to pthreadCancelableWait() call. Fri Jan 22 14:31:59 1999 Ross Johnson <rpj@swan.canberra.edu.au> * Makefile.in (CFLAGS): Remove -fhandle-exceptions. Not needed with egcs. Add -g for debugging. * create.c (pthread_create): Replace __stdcall with PT_STDCALL macro. This is a hack and must be fixed. * misc.c (CancelableWait): Remove redundant statement. * mutex.c (pthread_mutexattr_init): Cast calloc return value. * misc.c (CancelableWait): Add cast. (pthread_self): Add cast. * exit.c (pthread_exit): Add cast. * condvar.c (pthread_condattr_init): Cast calloc return value. * cleanup.c: Reorganise conditional compilation. * attr.c (pthread_attr_init): Remove unused 'result'. Cast malloc return value. * private.c (_pthread_callUserDestroyRoutines): Redo conditional compilation. * misc.c (CancelableWait): C++ version uses 'throw'. * cancel.c (pthread_testcancel): Ditto. * implement.h (class pthread_exception): Define for C++. * pthread.h: Fix C, C++, and Win32 SEH condition compilation mayhem around pthread_cleanup_* defines. C++ version now uses John Bossom's cleanup handlers. (pthread_attr_t): Make 'valid' unsigned. Define '_timeb' as 'timeb' for Ming32. Define PT_STDCALL as nothing for Mingw32. May be temporary. * cancel.c (pthread_testcancel): Cast return value.
1999-01-20Wed Jan 20 09:31:28 1999 Ross Johnson <rpj@ixobrychus.canberra.edu.au>rpj
* pthread.h (pthread_mutexattr_t): Changed to a pointer. * mutex.c (pthread_mutex_init): Conditionally create Win32 mutex - from John Bossom's implementation. (pthread_mutex_destroy): Conditionally close Win32 mutex - from John Bossom's implementation. (pthread_mutexattr_init): Replaced by John Bossom's version. (pthread_mutexattr_destroy): Ditto. (pthread_mutexattr_getpshared): New function from John Bossom's implementation. (pthread_mutexattr_setpshared): New function from John Bossom's implementation. Tue Jan 19 18:27:42 1999 Ross Johnson <rpj@swan.canberra.edu.au> * pthread.h (pthreadCancelableTimedWait): New prototype. (pthreadCancelableWait): Remove second argument. * misc.c (CancelableWait): New static function is pthreadCancelableWait() renamed. (pthreadCancelableWait): Now just calls CancelableWait() with INFINITE timeout. (pthreadCancelableTimedWait): Just calls CancelableWait() with passed in timeout. * private.c (_pthread_sem_timedwait): 'abstime' arg really is absolute time. Calculate relative time to wait from current time before passing timeout to new routine pthreadCancelableTimedWait(). - Scott Lightner <scott@curriculum.com> Tue Jan 19 10:27:39 1999 Ross Johnson <rpj@ixobrychus.canberra.edu.au> * pthread.h (pthread_mutexattr_setforcecs_np): New prototype. * mutex.c (pthread_mutexattr_init): Init 'pshared' and 'forcecs' attributes to 0. (pthread_mutexattr_setforcecs_np): New function (not portable). * pthread.h (pthread_mutex_t): Add 'mutex' element. Set to NULL in PTHREAD_MUTEX_INITIALIZER. The pthread_mutex_*() routines will try to optimise performance by choosing either mutexes or critical sections as the basis for pthread mutexes for each indevidual mutex. (pthread_mutexattr_t_): Add 'forcecs' element. Some applications may choose to force use of critical sections if they know that:- the mutex is PROCESS_PRIVATE and, either the OS supports TryEnterCriticalSection() or pthread_mutex_trylock() will never be called on the mutex. This attribute will be setable via a non-portable routine. Note: We don't yet support PROCESS_SHARED mutexes, so the implementation as it stands will default to Win32 mutexes only if the OS doesn't support TryEnterCriticalSection. On Win9x, and early versions of NT 'forcecs' will need to be set in order to get critical section based mutexes. Sun Jan 17 12:01:26 1999 Ross Johnson <rpj@ixobrychus.canberra.edu.au> * pthread.h (PTHREAD_MUTEX_INITIALIZER): Init new 'staticinit' value to '1' and existing 'valid' value to '1'. * global.c (_pthread_mutex_test_init_lock): Add. * implement.h (_pthread_mutex_test_init_lock.): Add extern. * private.c (_pthread_processInitialize): Init critical section for global lock used by _mutex_check_need_init(). (_pthread_processTerminate): Ditto (:s/Init/Destroy/). * dll.c (dllMain): Move call to FreeLibrary() so that it is only called once when the process detaches. * mutex.c (_mutex_check_need_init): New static function to test and init PTHREAD_MUTEX_INITIALIZER mutexes. Provides serialised access to the internal state of the uninitialised static mutex. Called from pthread_mutex_trylock() and pthread_mutex_lock() which do a quick unguarded test to check if _mutex_check_need_init() needs to be called. This is safe as the test is conservative and is repeated inside the guarded section of _mutex_check_need_init(). Thus in all calls except the first calls to lock static mutexes, the additional overhead to lock any mutex is a single memory fetch and test for zero. * pthread.h (pthread_mutex_t_): Add 'staticinit' member. Mutexes initialised by PTHREAD_MUTEX_INITIALIZER aren't really initialised until the first attempt to lock it. Using the 'valid' flag (which flags the mutex as destroyed or not) to record this information would be messy. It is possible for a statically initialised mutex such as this to be destroyed before ever being used. * mutex.c (pthread_mutex_trylock): Call _mutex_check_need_init() to test/init PTHREAD_MUTEX_INITIALIZER mutexes. (pthread_mutex_lock): Ditto. (pthread_mutex_unlock): Add check to ensure we don't try to unlock an unitialised static mutex. (pthread_mutex_destroy): Add check to ensure we don't try to delete a critical section that we never created. Allows us to destroy a static mutex that has never been locked (and hence initialised). (pthread_mutex_init): Set 'staticinit' flag to 0 for the new mutex.
1998-12-06As well as patches from Anders, this commit includes some older changesrpj
that hadn't been checked in. Those are to finish off the original TSD management scheme which will now start to be replaced by John Bossom's implementation. Anders' description of his changes:- Fixes the problem with `TryEnterCriticalSection' on w95 systems. Instead of directly calling `TryEnterCriticalSection' it tries to get a pointer at DLL startup and only calls the function if it exists. See ChangeLog for things changed. 1998-12-05 Anders Norlander <anorland@hem2.passagen.se> * implement.h (_pthread_try_enter_critical_section): New extern * dll.c (_pthread_try_enter_critical_section): New pointer to TryEnterCriticalSection if it exists; otherwise NULL. * dll.c (PthreadsEntryPoint): Initialize _pthread_try_enter_critical_section at startup and release kernel32 handle when DLL is being unloaded. * mutex.c (pthread_mutex_trylock): Replaced check for NT with a check if _pthread_try_enter_critical_section is valid pointer to a function. Call _pthread_try_enter_critical_section instead of TryEnterCriticalSection to avoid errors on Win95. Sun Nov 15 21:24:06 1998 Ross Johnson <rpj@ixobrychus.canberra.edu.au> * cleanup.c (_pthread_destructor_run_all): Declare missing void * arg. Fixup CVS merge conflicts. Fri Oct 30 15:15:50 1998 Ross Johnson <rpj@swan.canberra.edu.au> * cleanup.c (_pthread_handler_push): Fixed bug appending new handler to list reported by Peter Slacik <Peter.Slacik@leibinger.freinet.de>. (new_thread): Rename poorly named local variable to "new_handler". Sat Oct 24 18:34:59 1998 Ross Johnson <rpj@ixobrychus.canberra.edu.au> * global.c: Add TSD key management array and index declarations. * implement.h: Ditto for externs. Fri Oct 23 00:08:09 1998 Ross Johnson <rpj@ixobrychus.canberra.edu.au> * implement.h (_PTHREAD_TSD_KEY_REUSE): Add enum. * private.c (_pthread_delete_thread): Add call to _pthread_destructor_run_all() to clean up the threads keys. * cleanup.c (_pthread_destructor_run_all): Check for no more dirty keys to run destructors on. Assume that the destructor call always succeeds and set the key value to NULL.
1998-10-041998-10-05 Ben Elliston <bje@cygnus.com>bje
* global.c (PTHREAD_MUTEX_INITIALIZER): Move to pthread.h. * pthread.h (PTHREAD_MUTEX_INITIALIZER): Define. (pthread_mutex_t): Reimplement as a struct containing a valid flag. If the flag is ever down upon entry to a mutex operation, we call pthread_mutex_create() to initialise the object. This fixes the problem of how to handle statically initialised objects that can't call InitializeCriticalSection() due to their context. * mutex.c (pthread_mutex_init): Set valid flag. (pthread_mutex_destroy): Clear valid flag. (pthread_mutex_lock): Check and handle the valid flag. (pthread_mutex_unlock): Likewise. (pthread_mutex_trylock): Likewise. * tests/mutex3.c: New file; test for the static initialisation macro. Passes.
1998-10-031998-10-03 Ben Elliston <bje@cygnus.com>bje
* mutex.c (pthread_mutex_init): Correct function signature.