Re: [PATCH][5/8] Arch agnostic completely out of line locks / ppc64

[Linus Torvalds]

On Thu, 9 Sep 2004, Zwane Mwaikambo wrote:
> 
> I think cpu_relax() (or some other primitive) should actually take a 
> parameter, this will allow for us to use monitor/mwait on i386 too so 
> that in cases where we're spinning waiting on memory modify we could do 
> something akin to the following;
> 
> while (spin_is_locked(lock))
> 	cpu_relax(lock);

You can't do it that way. It needs to be arch-specific. When using 
something like monitor/mwait (or a "futex", which something like UML might 
use), you need to load the value you want to wait on _before_. So the 
interface literally would have to be the monitor/mwait interface:

	for (;;) {
		lockval = monitor(lock);
		if (!is_locked(lockval))
			break;
		mwait(lock, lockval);
	}

and the fact is, this is all much better just done in the arch-specific 
spinlock code. 

		Linus
-
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/