Re: [RFC] Userspace RCU: (ab)using futexes to save cpu cycles and energy

["Chris Friesen" <cfriesen@nortel.com>]

On 09/23/2009 11:48 AM, Mathieu Desnoyers wrote:

> Here are the primitives I've created. I'd like to have feedback on my
> futex use, just to make sure I did not do any incorrect assumptions.

> /*
>  * Wake-up any waiting defer thread. Called from many concurrent threads.
>  */
> static void wake_up_defer(void)
> {
>         if (unlikely(atomic_read(&defer_thread_futex) == -1))
>                 atomic_set(&defer_thread_futex, 0);
>                 futex(&defer_thread_futex, FUTEX_WAKE,
>                       0, NULL, NULL, 0);
> }

Is it a problem if multiple threads all hit the defer_thread_futex==-1
case simultaneously?  If so, maybe this should use an atomic
test-and-set operation so that only one thread actually calls futex().

> /*
>  * Defer thread waiting. Single thread.
>  */
> static void wait_defer(void)
> {
>         atomic_dec(&defer_thread_futex);
>         if (atomic_read(&defer_thread_futex) == -1)
>                 futex(&defer_thread_futex, FUTEX_WAIT, -1,
>                       NULL, NULL, 0);
> }

Is it a problem if the value of defer_thread_futex changes to zero after
the dec but before the test?

Chris