Subj : Re: Challenge: Multithreading & Synchronization
To   : comp.programming.threads
From : Giancarlo Niccolai
Date : Mon May 23 2005 03:54 pm

doug wrote:

> 
> Actually, I have a (hopefully) quick question.  Is the above algorithm
> safe
> in a system where you dynamically create the recursive mutex?  E.g.  one
> thread allocates a control block (which includes creating a recursive
> mutex), and a second thread comes along later and attempts to acquire the
> sema4.  Is it not possible that this second thread might misread
> uninitialised 'garbage' from the owner field, and may think it already
> owns
> the sema4?  (Pretty low odds, I know, but still finite.)

Theoretically yes, but then the problem is NOT in the mutex, but in your
app, trying to access a non initialized mutex when it is still not
available. Doing this kind of things results in anything between a correct
operation and a core dump, inclusive, depending by the odds, whichever kind
of data is a stake: a recursive mutex, a normal mutex, or any object at
all... So, I suppose the problem is not quite relevant.

Mutexes are ANYHOW to be initialized in a safe moment in the first place...

Gian.

.