671
Subj : Re: Stack in thread-local virtual memory?
To   : comp.programming.threads
From : Ulrich Eckhardt
Date : Tue Aug 23 2005 08:14 am

David Hopwood wrote:
> Ulrich Eckhardt wrote:
>>    foo a_foo;
>>    thread t = create_thread( &thread_entry, &a_foo);
>>    join(t);
> 
> Can 'join' fail?

This is all pseudocode, the 'foo' was a handle to an event, the join() was
in fact a WaitForSingleObject() on that event without a timeout. I tried
to make sure that I check each and every call's returnvalue for failure,
too. Using debugging output, I was able to see that the program did
nothing unusual or unexpected, except for the already described access
violation on the main thread's stack.

>> Simple thing, fork off a worker thread and wait until it has finished -
>> to be precise I only waited until it had initialised and passed an
>> pointer to a win32 event handle, but for sure I did not return from
>> main() so the reference to a_foo was valid all the time.
> 
> This looks like the kind of code that could be mis-optimized by a
> compiler that is not sufficiently aware of threading. In a
> single-threaded program, depending on the code of 'join', it could 
> be correct to treat a_foo as dead after the create_thread call.

No, I still used the a_foo (an event handle) afterwards in order to wait
for the forked thread to signal it.

> I gather that since you no longer have access to the machine where it
> failed, you can't disassemble main() to see whether it looks correct?

I'm contemplating rebuilding that environment - not knowing what was going
on there is making me nuts. :(

Uli

.

0