Re: test6-lowlatency-D1 results: 50msec + hangs ....

[Philipp Rumpf]

On Sat, Aug 05, 2000 at 06:37:33PM -0700, David S. Miller wrote:
> while (spin_is_locked(&lock)) {
> 	/* Wait for others to release their references. */
> 	barrier();
> }
> 
> This type of code does happen, what is the correct return value so
> that spin_is_locked() handles this kind of case?

0.  It's one of the "spin_is_locked_and_the_holder_might_be_someone_else"
cases.  Maybe you misunderstood, but I'm proposing to introduce a new
function, not change the existing one.  Maybe move to a better name for it.

Or did I miss your point about the code above ?

	Philipp