2001-05-30 Ross Johnson <rpj@setup1.ise.canberra.edu.au>

	* 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].

1 files changed, 123 insertions, 54 deletions

diff --git a/ANNOUNCE b/ANNOUNCE
index b997141..809dc7d 100644
--- a/ANNOUNCE
+++ b/ANNOUNCE
@@ -1,5 +1,5 @@
 

-                 PTHREADS-WIN32 SNAPSHOT 2000-12-29

+                 PTHREADS-WIN32 SNAPSHOT ????-??-??

                  ----------------------------------

        Web Site: http://sources.redhat.com/pthreads-win32/

       FTP Site: ftp://sources.redhat.com/pub/pthreads-win32

@@ -13,6 +13,10 @@ 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

+Unix 98 specification for compatibility with major Unix

+implementations and Linux.

+

 Pthreads-win32 is free software, distributed under the GNU Library

 General Public License (LGPL).

 

@@ -26,57 +30,114 @@ Change Summary (since the last snapshot)
 (See the ChangeLog file for details.)

 

 New:

--	New non-portable functions for use when statically linking the library

-	(see the README file):

-		pthread_delay_np

-		pthread_win32_process_attach_np

-		pthread_win32_process_detach_np

-		pthread_win32_thread_attach_np

-		pthread_win32_thread_detach_np

-	Also, the start of support within the library for statically

-	linked applications, but this has not been tested by statically

-	linking yet.

--	A fully working MinGW32 g++ compiled version of the library

-	is now supplied. To use with g++ compiled applications

-	please see the FAQ "How do I generate pthreadGCE.dll and

-	libpthreadw32.a for use with Mingw32", which you need to do

-	before compiling and linking applications as well.

--	Unhandled exceptions, whether SEH or C++, are no longer

-	silently caught and ignored but are passed out of threads for

-	the system to deal with. C++ applications can override

-	the default termination handler via set_terminate().

-	set_terminate() works fine for MinGW32 g++ but not yet with MS VC++

-	(see known bugs (2)) even though it's the same code. See

-	tests/exception3.c for an example of using set_terminate().

+-	Async cancellation should now work for Windows running

+	on the following processors: IX86, MIPS, ALPHA, PPC.

+	- contributors name misplaced (please contact me if it's you)

+ 

+-       New functions (no-ops) for source code compatibility:

+                pthread_getconcurrency()

+                        Returns the value previously set by

+			pthread_setconcurrency() or 0 (zero)

+			if no value was previously set.

+			The implementation does not currently use

+			this value.

+                pthread_setconcurrency()

+                        Accepts any value >= 0 but does not

+                        have any effect; returns '0' or

+                        EINVAL if the value is < 0

+                pthread_attr_getscope()

+                pthread_attr_setscope()

+                        Currently only return ENOSYS

+

+-       The following mutex types and related functions are now

+        supported:

+

+                pthread_mutexattr_gettype()

+                pthread_mutexattr_settype()

+                PTHREAD_MUTEX_DEFAULT

+                PTHREAD_MUTEX_NORMAL or PTHREAD_MUTEX_FAST_NP

+                PTHREAD_MUTEX_ERRORCHECK or PTHREAD_MUTEX_ERRORCHECK_NP

+                PTHREAD_MUTEX_RECURSIVE or PTHREAD_MUTEX_RECURSIVE_NP

+

+	The *_NP versions are for Linux portability.

+

+        In this implementation PTHREAD_MUTEX_DEFAULT is

+        mapped to PTHREAD_MUTEX_NORMAL for compatibility

+	with major Unix vendors. This is also the default

+	if no type is specified.

+

+	Please note that the default behaviour in previous

+	versions of the library was equivalent to

+	PTHREAD_MUTEX_RECURSIVE. This will be a problem

+	for some applications unless they are updated to

+	explicitly set the type of any mutexes that are, or

+	could be, locked recursively.

+

+        PTHREAD_MUTEX_NORMAL will cause thread deadlock

+        if the owner of a mutex tries to relock it without

+        first unlocking it. It has slightly less overhead

+	that other types.

+

+-       Pthreads-win32 mutexes are now based on Win32

+        critical sections for all Windows versions. The

+	implementation no longer depends on

+	TryEnterCriticalSection.

+        - Thomas Pfaff <tpfaff@gmx.net>

 

 Bugs fixed:

--	Pthread_mutex_unlock() now returns EPERM if the unlocking

-	thread does not currently hold the thread.

--	Args in the ctime_r macro in pthread.h are fixed.

--	Made semaphore.h independent of config.h when building

-	applications. Don't know how this went unnoticed for so long

-	- the default defines must have been correct for my test

-	environment.

-

-Some new tests have been added.

+-       Pthread_mutex_trylock() now properly returns EBUSY

+        even when the current thread owns the mutex.

+        Consequently, pthread_mutex_destroy() will no longer

+        destroy a locked mutex (it will return EBUSY).

+        - Thomas Pfaff <tpfaff@gmx.net>

 

+-	Errno warnings and building with /MD;

+	calling conventions (stdcall v cdecl);

+	thread termination problems;

+	- Milan Gardian <Milan.Gardian@LEIBINGER.com>

 

 Known bugs in this snapshot

 ---------------------------

 

-1. Asynchronous cancelation only works on Intel X86 machines.

-

-2. Under MS VC++ (only tested with version 6.0), a term_func

+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() -

    makes an explicit call to terminate() which works as expected

    under MinGW32 g++ doesn't appear to run the term_func under

    MC VC++ 6.0.

 

-3. This is an interim snapshot as there are still some additional

-   patches to go in, eg. to fix problems with errno support under

-   some circumstances. Some people are seeing compile warnings

-   to do with _errno.

+2. Cancellation problems in optimised code

+   - Milan Gardian

+

+   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

 -------

@@ -166,7 +227,11 @@ The following functions are implemented:
       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

@@ -219,6 +284,10 @@ The following functions are implemented:
       pthread_attr_setschedparam  

       pthread_getschedparam

       pthread_setschedparam

+      pthread_getconcurrency (always returns '0' indicating system default)

+      pthread_setconcurrency (accepts any value >= 0 but is ignored)

+      pthread_attr_getscope  (returns an error ENOSYS)

+      pthread_attr_setscope  (returns an error ENOSYS)

       sched_get_priority_max (POSIX 1b)

       sched_get_priority_min (POSIX 1b)

       sched_yield            (POSIX 1b)

@@ -229,15 +298,15 @@ The following functions are implemented:
       pthread_sigmask

 

       ---------------------------

-      Non-portable routines (see the README file for usage)

+      Non-portable routines (see the README.NONPORTABLE file for usage)

       ---------------------------

-      pthread_mutexattr_setforcecs_np

       pthread_getw32threadhandle_np

-	pthread_delay_np

-	pthread_win32_process_attach_np

-	pthread_win32_process_detach_np

-	pthread_win32_thread_attach_np

-	pthread_win32_thread_detach_np

+      pthread_delay_np

+      pthread_mutex_setdefaulttype_np

+      pthread_win32_process_attach_np

+      pthread_win32_process_detach_np

+      pthread_win32_thread_attach_np

+      pthread_win32_thread_detach_np

 

       ---------------------------

       Static Initializers (macros)

@@ -265,10 +334,8 @@ The following functions are not implemented:
       ---------------------------

       pthread_attr_getinheritsched

       pthread_attr_getschedpolicy

-      pthread_attr_getscope

       pthread_attr_setinheritsched

       pthread_attr_setschedpolicy

-      pthread_attr_setscope

       pthread_mutex_getprioceiling

       pthread_mutex_setprioceiling

       pthread_mutex_attr_getprioceiling

@@ -344,20 +411,22 @@ Mailing List
 There is a mailing list for discussing pthreads on Win32. To join,

 send email to:

 

-        pthreads-win32-subscribe@sources.redhat.com

+        pthreads-win32-subscribe@sourceware.cygnus.com

       

 

 Application Development Environments

 ------------------------------------

       

 MSVC:

-MSVC using SEH works.

-MSVC using C++ EH works.

+MSVC using SEH works. Distribute pthreadVSE.dll with your application.

+MSVC using C++ EH works. Distribute pthreadVCE.dll with your application.

 

 Mingw32:

 You need gcc-2.95.2-1 modified as per pthreads-win32 FAQ answer (6), with

 binutils-19990818-1 and msvcrt runtime-2000-03-27. Mingw32 must use

-the thread-safe MSVCRT library (see the FAQ).

+the thread-safe MSVCRT library (see the FAQ). You need to distribute

+the gcc.dll DLL from Mingw32 with your application (as well as

+pthreadGCE.dll of course).

 

 Cygwin: (http://sourceware.cygnus.com/cygwin/)

 Cygwin aims to provide a complete POSIX environment on top of Win32, including

@@ -382,7 +451,7 @@ For convenience, the following pre-built files are available on the FTP site
 	pthreadVCE.lib

 	pthreadVSE.dll	- built with MSVC compiler using SEH

 	pthreadVSE.lib

-	pthreadGCE.dll	- built with Mingw32 G++

+	pthreadGCE.dll	- built with Mingw32 G++ 2.95.2-1 

 	libpthreadw32.a	- derived from pthreadGCE.dll

 

 These are the only files you need in order to build POSIX threads