Re: [PATCH 6/24] make atomic_read() behave consistently on frv

[Michael Buesch]

On Wednesday 15 August 2007 15:29:43 Arnd Bergmann wrote:
> On Wednesday 15 August 2007, Paul E. McKenney wrote:
> 
> > ACCESS_ONCE() is indeed intended to be used when actually loading or
> > storing the variable.  That said, I must admit that it is not clear to me
> > why you would want to add an extra order() rather than ACCESS_ONCE()ing
> > one or both of the adjacent accesses to that same variable.
> > 
> > So, what am I missing?
> 
> You're probably right, the only case I can construct is something like
> 
> 	if (ACCESS_ONCE(x)) {
> 		...
> 		ACCESS_ONCE(x)++;
> 	}
> 
> which would be slightly less efficient than
> 
> 	if (x)
> 		x++;
> 	order(x);
> 
> because in the first case, you need to do two ordered accesses
> but only one in the second case. However, I can't think of a case
> where this actually makes a noticable difference in real life.

How can this example actually get used in a sane and race-free
way? This would need locking around the whole if
statement. But locking is a barrier.

-- 
Greetings Michael.
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/