Subj : Re: Memory visibility with Pthreads
To   : comp.programming.threads
From : Alexander Terekhov
Date : Thu Aug 11 2005 01:50 pm


Christoph Bartoschek wrote:
> 
> Alexander Terekhov wrote:
> 
> >
> > The standard doesn't say "fully". Certain functions merely "synchronize
> > memory" and only as long as the program is data race-free. Very relaxed
> > memory models like RCpc are fully compatible with Pthreads.
> 
> How does the  "synchronize memory" instead of "fully" apply to my example in
> the first post? 

: int A = 0;
: int B = 0;
: 
: Thread 1 executes the following code:
: 
: A = 10;
: pthread_mutex_lock(mutex);
: B = 10;
: pthread_mutex_unlock(mutex);
: 
: Thread 2 executes another codesequence:
: 
: pthread_mutex_lock(mutex);
: if (B == 10) {
:   printf("The value of A is %d\n", A);
: } else {
:   printf("B is not 10\n");
: }
: pthread_mutex_unlock(mutex);

> Does it become invalid? 

No. 

> And is the programm data race-free?

Maybe. Provided that thread 1 can execute "A = 10; ..." code only 
once... and also watch out implementation-specific "memory location" 
granularity.

> I guess it is, because there is no possiblity that one thread writes to a
> variable and the other one tries to access it.

I meant:

http://tinyurl.com/77hvz

regards,
alexander.

.