Question

Does the stack reserved/commited for a thread get freed when

• the thread terminates
• the thread object is destroyed
  (i.e. the thread is terminated and all handles to the thread are closed)

?

More broadly, are there significant resources associated with a thread that has terminated, but still exists since there are valid handles to it?

---

Reason: I need to modify a kind of “scoped singleton”, so it doesn’t return a single object, but a per-thread object. I cannot rely on thread creation/termination notices, much less on process-wide ones.

At the moment, I store the objects in a map<ThreadID, Object>, with a cache cleanup policy that’s suitable for my application. To protect myself from the OS “recycling” thread ID’s, I keep an handle to the thread open. (Rec

A side effect would be holding open handles to long-terminated threads in some corner cases.

Answer 1

According to “Windows VIA C/C++” by Richter and Nasarre (A must-have book for any C++ Windwos programmer) p.154:

Terminating a Thread

A thread can be terminated in four
ways:

• The thread function returns. (This is highly recommended.)
• The thread kills itself by calling the ExitThread function. (Avoid this
  method.)
• A thread in the same process or in another one calls the TerminateThread
  function. (Avoid this method.)
• The process containing the thread terminates. (Avoid this method.)

The Thread Function Returns

You should always design your thread
functions so that they return when you
want the thread to terminate. This is
the only way to guarantee that all
your thread’s resources are cleaned up
properly.

Having your thread function return
ensures the following:

• All C++ objects created in your thread function will be destroyed
  properly via their destructors.
• The operating system will properly free the memory used by the thread’s
  stack.
• The system will set the thread’s exit code (maintained in the thread’s
  kernel object) to your thread
  function’s return value.
• The system will decrement the usage count of the thread’s kernel object.

The ExitThread Function

You can force your thread to terminate by having it call ExitThread:

VOID ExitThread(DWORD dwExitCode);

This function terminates the thread
and causes the operating system to
clean up all the operating system
resources that were used by the
thread. However, your C/C++ resources
(such as C++ class objects) will not
be destroyed. For this reason, it is
much better to simply return from your
thread function instead of calling
ExitThread yourself.

Of course, you use ExitThread’s
dwExitCode parameter to tell the
system what to set the thread’s exit
code to. The ExitThread function does
not return a value because the thread
has terminated and cannot execute any
more code.

Note The recommended way to have a
thread terminate is by having its
thread function simply return (as
described in the previous section).
However, if you use the method
described in this section, be aware
that the ExitThread function is the
Windows function that kills a thread.
If you are writing C/C++ code, you
should never call ExitThread. Instead,
you should use the C++ run-time
library function _endthreadex. If you
do not use Microsoft’s C++ compiler,
your compiler vendor will have its own
alternative to ExitThread. Whatever
this alternative is, you must use it.

The TerminateThread Function

A call to
TerminateThread also kills a thread:

BOOL TerminateThread( HANDLE
hThread, DWORD dwExitCode);

Unlike ExitThread, which always kills
the calling thread, TerminateThread
can kill any thread. The hThread
parameter identifies the handle of the
thread to be terminated. When the
thread terminates, its exit code
becomes the value you passed as the
dwExitCode parameter. Also, the
thread’s kernel object has its usage
count decremented.

Note The TerminateThread function is
asynchronous. That is, it tells the
system that you want the thread to
terminate but the thread is not
guaranteed to be killed by the time
the function returns. If you need to
know for sure that the thread has
terminated, you might want to call
WaitForSingleObject or a similar function,
passing the handle of the thread.

A well-designed application never uses
this function because the thread being
terminated receives no notification
that it is dying. The thread cannot
clean up properly, and it cannot
prevent itself from being killed.

Note When a thread dies by returning
or calling ExitThread, the stack for
the thread is destroyed. However, if
TerminateThread is used, the system
does not destroy the thread’s stack
until the process that owned the
thread terminates. Microsoft purposely
implemented TerminateThread in this
way. If other still-executing threads
were to reference values on the
forcibly killed thread’s stack, these
other threads would raise access
violations. By leaving the killed
thread’s stack in memory, other
threads can continue to execute just
fine.

In addition, dynamic-link libraries
(DLLs) usually receive notifications
when a thread is terminating. If a
thread is forcibly killed with
TerminateThread, however, the DLLs do
not receive this notification, which
can prevent proper cleanup.

When a Thread Terminates

The following actions occur when a
thread terminates:

All User object handles owned by the
thread are freed. In Windows, most
objects are owned by the process
containing the thread that creates the
objects. However, a thread owns two
User objects: windows and hooks. When
a thread dies, the system
automatically destroys any windows and
uninstalls any hooks that were created
or installed by the thread. Other
objects are destroyed only when the
owning process terminates.

The thread’s exit code changes from
STILL_ACTIVE to the code passed to
ExitThread or TerminateThread.

The state of the thread kernel object
becomes signaled.

If the thread is the last active
thread in the process, the system
considers the process terminated as
well.

The thread kernel object’s usage count
is decremented by 1.

When a thread terminates, its
associated thread kernel object
doesn’t automatically become freed
until all the outstanding references
to the object are closed.

Once a thread is no longer running,
there isn’t much any other thread in
the system can do with the thread’s
handle. However, these other threads
can call GetExitCodeThread to check
whether the thread identified by
hThread has terminated and, if it has,
determine its exit code:

BOOL GetExitCodeThread( HANDLE
hThread, PDWORD pdwExitCode);

The exit code value is returned in the
DWORD pointed to by pdwExitCode. If
the thread hasn’t terminated when
GetExitCodeThread is called, the
function fills the DWORD with the
STILL_ACTIVE identifier (defined as
0x103). If the function is successful,
TRUE is returned.