Open Source
POSIX Threads for Win32



Q 1     What is it?

Q 2     Which of the several dll versions do I use?
        What are all these pthread*.dll and pthread*.lib files?

Q 3     What is the library naming convention?

Q 4     Cleanup code default style or: it used to work when I built
        the library myself, but now it doesn't - why?

Q 5     Why is the default library version now less exception-friendly?

Q 6     Should I use Cygwin or Mingw32 as a development environment?

Q 7     Now that pthreads-win32 builds under Mingw32, why do I get
        memory access violations (segfaults)?

Q 8     How do I use pthread.dll for Win32 (Visual C++ 5.0)

Q 9     Cancelation doesn't work for me, why?

Q 10    How do I generate pthreadGCE.dll and libpthreadw32.a for use
        with Mingw32?

Q 11    Why isn't pthread_t defined as a scalar (e.g. pointer or int)
        like it is for other POSIX threads implementations?


Q 1     What is it?

Pthreads-win32 is 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. The library also supports
some of the functionality of the Open Group's Single Unix
specification, version 2, namely mutex types.

See the file "ANNOUNCE" for more information including standards
conformance details and list of supported routines.


Q 2     Which of the several dll versions do I use?
---     or,
        What are all these pthread*.dll and pthread*.lib files?

Simply, you only use one of them, but you need to choose carefully.

The most important choice you need to make is whether to use a
version that uses exceptions internally, or not (there are versions
of the library that use exceptions as part of the thread
cancelation and cleanup implementation, and one that uses
setjmp/longjmp instead).

There is some contension amongst POSIX threads experts as
to how POSIX threads cancelation and exit should work
with languages that include exceptions and handlers, e.g.
C++ and even C (Microsoft's Structured Exceptions).

The issue is: should cancelation of a thread in, say,
a C++ application cause object destructors and C++ exception
handlers to be invoked as the stack unwinds during thread
exit, or not?

There seems to be more opinion in favour of using the
standard C version of the library (no EH) with C++ applications
since this appears to be the assumption commercial pthreads
implementations make. Therefore, if you use an EH version
of pthreads-win32 then you may be under the illusion that
your application will be portable, when in fact it is likely to
behave very differently linked with other pthreads libraries.

Now you may be asking: why have you kept the EH versions of
the library?

There are a couple of reasons:
- there is division amongst the experts and so the code may
  be needed in the future. (Yes, it's in the repository and we
  can get it out anytime in the future, but ...)
- pthreads-win32 is one of the few implementations, and possibly
  the only freely available one, that has EH versions. It may be
  useful to people who want to play with or study application
  behaviour under these conditions.


Q 3     What is the library naming convention?

Because the library is being built using various exception
handling schemes and compilers - and because the library
may not work reliably if these are mixed in an application,
each different version of the library has it's own name.

Note 1: the incompatibility is really between EH implementations
of the different compilers. It should be possible to use the
standard C version from either compiler with C++ applications
built with a different compiler. If you use an EH version of
the library, then you must use the same compiler for the
application. This is another complication and dependency that
can be avoided by using only the standard C library version.

Note 2: if you use a standard C pthread*.dll with a C++
application, then any functions that you define that are
intended to be called via pthread_cleanup_push() must be

Note 3: the intention is to also name either the VC or GC
version (it should be arbitrary) as pthread.dll, including
pthread.lib and libpthread.a as appropriate.

In general:

        [VG] indicates the compiler
        V       - MS VC
        G       - GNU C

        {SE,CE,C} indicates the exception handling scheme
        SE      - Structured EH
        CE      - C++ EH
        C       - no exceptions - uses setjmp/longjmp

For example:
        pthreadVSE.dll  (MSVC/SEH)
        pthreadGCE.dll  (GNUC/C++ EH)
        pthreadGC.dll   (GNUC/not dependent on exceptions)

The GNU library archive file names have changed to:



Q 4     Cleanup code default style or: it used to work when I built
---     the library myself, but now it doesn't - why?

Up to and including snapshot 2001-07-12, if not defined, the cleanup
style was determined automatically from the compiler used, 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().

After snapshot 2001-07-12, unless your 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/all
commercial Unix POSIX threads implementations.

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

THE POINT OF ALL THIS IS: if you have not been defining one of these
explicitly, then the defaults have been set according to the compiler
and language you are using, as described at the top of this

THIS NOW CHANGES, as has been explained above. For example:

If you were building your application with MSVC++ i.e. using C++
exceptions (rather than SEH) and not explicitly defining one of
__CLEANUP_*, then __CLEANUP_C++ was 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
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, nor when 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 local
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.


Q 5     Why is the default library version now less exception-friendly?

Because most commercial Unix POSIX threads implementations don't allow you to
choose to have stack unwinding. (Compaq's TRU64 Unix is possibly an exception.)

Therefore, providing it in pthread-win32 as a default could be dangerous
and non-portable. We still provide the choice but you must now consciously
make it.

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.)
- 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.


Q 6     Should I use Cygwin or Mingw32 as a development environment?

Important: see Q7 also.

Use Mingw32 with the MSVCRT library to build applications that use
the pthreads DLL.

Cygwin's own internal support for POSIX threads is growing.
Consult that project's documentation for more information.


Q 7     Now that pthreads-win32 builds under Mingw32, why do I get
---     memory access violations (segfaults)?

The latest Mingw32 package has thread-safe exception handling (see Q10).
Also, see Q6 above.


Q 8     How do I use pthread.dll for Win32 (Visual C++ 5.0)

> I'm a "rookie" when it comes to your pthread implementation.  I'm currently
> desperately trying to install the prebuilt .dll file into my MSVC compiler.
> Could you please provide me with explicit instructions on how to do this (or
> direct me to a resource(s) where I can acquire such information)?
> Thank you,

You should have a .dll, .lib, .def, and three .h files. It is recommended
that you use pthreadVC.dll, rather than pthreadVCE.dll or pthreadVSE.dll
(see Q2 above).

The .dll can go in any directory listed in your PATH environment
variable, so putting it into C:\WINDOWS should work.

The .lib file can go in any directory listed in your LIB environment

The .h files can go in any directory listed in your INCLUDE
environment variable.

Or you might prefer to put the .lib and .h files into a new directory
and add its path to LIB and INCLUDE. You can probably do this easiest
by editing the file:-

C:\Program Files\DevStudio\vc\bin\vcvars32.bat

The .def file isn't used by anything in the pre-compiled version but 
is included for information.



Q 9     Cancelation doesn't work for me, why?

> I'm investigating a problem regarding thread cancelation. The thread I want
> to cancel has PTHREAD_CANCEL_ASYNCHRONOUS, however, this piece of code
> blocks on the join():
>               if ((retv = Pthread_cancel( recvThread )) == 0)
>               {
>                       retv = Pthread_join( recvThread, 0 );
>               }
> Pthread_* are just macro's; they call pthread_*.
> The thread recvThread seems to block on a select() call. It doesn't get
> cancelled.
> Two questions:
> 1) is this normal behaviour?
> 2) if not, how does the cancel mechanism work? I'm not very familliar to
> win32 programming, so I don't really understand how the *Event() family of
> calls work.

The answer to your first question is, normal POSIX behaviour would  
be to asynchronously cancel the thread. However, even that doesn't
guarantee cancelation as the standard only says it should be
cancelled as soon as possible.

Snapshot 99-11-02 or earlier only partially supports asynchronous cancellation.
Snapshots since then simulate async cancelation by poking the address of
a cancelation routine into the PC of the threads context. This requires
the thread to be resumed in some way for the cancelation to actually
proceed. This is not true async cancelation, but it is as close as we've
been able to get to it.

If the thread you're trying to cancel is blocked (for instance, it could be
waiting for data from the network), it will only get cancelled when it unblocks
(when the data arrives). For true pre-emptive cancelation in these cases,
pthreads-win32 from snapshot 2004-05-16 can automatically recognise and use the
QueueUserAPCEx package by Panagiotis E. Hadjidoukas. This package is available
from the pthreads-win32 ftp site and is included in the pthreads-win32
self-unpacking zip from 2004-05-16 onwards.

Using deferred cancelation would normally be the way to go, however,
even though the POSIX threads standard lists a number of C library
functions that are defined as deferred cancelation points, there is
no hookup between those which are provided by Windows and the
pthreads-win32 library.

Incidently, it's worth noting for code portability that the older POSIX
threads standards cancelation point lists didn't include "select" because
(as I read in Butenhof) it wasn't part of POSIX. However, it does appear in
the SUSV3.

Effectively, the only mandatory cancelation points that pthreads-win32
recognises are those the library implements itself, ie.

The following routines from the non-mandatory list in SUSV3 are
cancelation points in pthreads-win32:


The following routines from the non-mandatory list in SUSV3 are not
cancelation points in pthreads-win32:


Pthreads-win32 also provides two functions that allow you to create
cancelation points within your application, but only for cases where
a thread is going to block on a Win32 handle. These are:

        pthreadCancelableWait(HANDLE waitHandle) /* Infinite wait */
        pthreadCancelableTimedWait(HANDLE waitHandle, DWORD timeout)


Q 10    How do I create thread-safe applications using
----    pthreadGCE.dll, libpthreadw32.a and Mingw32?

This should not be a problem with recent versions of MinGW32.

For early versions, see Thomas Pfaff's email at:

Q 11    Why isn't pthread_t defined as a scalar (e.g. pointer or int)
        like it is for other POSIX threads implementations?

Originally pthread_t was defined as a pointer (to the opaque pthread_t_
struct) and later it was changed to a struct containing the original
pointer plus a sequence counter. This is allowed under both the original
POSIX Threads Standard and the current Single Unix Specification.

When pthread_t is a simple pointer to a struct some very difficult to
debug problems arise from the process of freeing and later allocing
thread structs because new pthread_t handles can acquire the identity of
previously detached threads. The change to a struct was made, along with
some changes to their internal managment, in order to guarantee (for
practical applications) that the pthread_t handle will be unique over the
life of the running process.

Where application code attempts to compare one pthread_t against another
directly, a compiler error will be emitted because structs can't be
compared at that level. This should signal a potentially serious problem
in the code design, which would go undetected if pthread_t was a scalar.

The POSIX Threading API provides a function named pthread_equal() to
compare pthread_t thread handles.

Other pthreads implementations, such as Sun's, use an int as the handle
but do guarantee uniqueness within the process scope. Win32 scalar typed
thread handles also guarantee uniqueness in system scope. It wasn't clear
how well the internal management of these handles would scale as the
number of threads and the fragmentation of the sequence numbering
increased for applications where thousands or millions of threads are
created and detached over time. The current management of threads within
pthreads-win32 using structs for pthread_t, and reusing without ever
freeing them, reduces the management time overheads to a constant, which
could be important given that pthreads-win32 threads are built on top of
Win32 threads and will therefore include that management overhead on top
of their own. The cost is that the memory resources used for thread
handles will remain at the peak level until the process exits.

While it may be inconvenient for developers to be forced away from making
assumptions about the internals of pthread_t, the advantage for the
future development of pthread-win32, as well as those applications that
use it and other pthread implementations, is that the library is free to
change pthread_t internals and management as better methods arise.

Maintained by Ross Johnson