Re: [boost] Re: read/write mutexes and boost

From: E. Gladyshev (egladysh_at_[hidden])
Date: 2003-10-20 15:53:36

> Daniel Spangenberg <dsp_at_[hidden]> writes:
> Question is: Can I use the current thread library to realize a
> read/write mutex (accepting
> multiple readers from different threads but prevents writer from
> access)
> on my own?
> If so, how should I do this? If this would help to give a more
> precise answer: My
> target platform is Win32, although a generic solution would be
> wunderful, of course ;-))
> (More worse: It must be a mutex accessible/protectable from threads
> of different
> processes...).

[...]

If you are working on an industrial mission critical control application
adn thinking about using boost::mutex,
I'd suggest to take a look on how boost::mutex is actually implemented
on Win32 platform.
As far as I remember, it doesn't handle the WAIT_ABANDONED condition
correctly if at all. Here is how it was in boost v1.30

void try_mutex::do_lock()
{
    int res = 0;
    res = WaitForSingleObject(reinterpret_cast<HANDLE>(m_mutex), INFINITE);
    assert(res == WAIT_OBJECT_0);
}

bool try_mutex::do_trylock()
{
    unsigned int res = 0;
    res = WaitForSingleObject(reinterpret_cast<HANDLE>(m_mutex), 0);
    assert(res != WAIT_FAILED && res != WAIT_ABANDONED);
    return res == WAIT_OBJECT_0;
}

If in try_mutex::do_lock(), the 'assert' statement is a debug
version only macro, the try_mutex::do_lock() method is basically 
ignoring the WaitForSingleObject() return code completely.
I don't think that this implementation is acceptable for
a serious multi-thread application. 

Also, in most critical (especially cross-process) applications WAIT_ABANDONED
is a "normal" condition that must be dealt with.

If you are interested, we can discuss this condition
in more details. It is not as straight forward as it might seem.

Eugene

__________________________________
Do you Yahoo!?
The New Yahoo! Shopping - with improved product search
http://shopping.yahoo.com