On Mon, Nov 14, 2011 at 11:12 AM, Paul Querna <[email protected]> wrote:

>
> The problem became that in trunk, we had to told the lock for the
> timeout queues while we were doing the pollset operation.   The
> pollset already had its own internal mutex too, for its own rings.  So
> we were double locking a piece of fairly high used code.
>


> In regards to reducing latency, I'm thinking about using the
> apr_pollset_wakeup API, which uses an internal pipe under the covers.
>

So it's a tradeoff: the cost of double locking vs. the cost of the extra
syscalls and extra context switch.  I will take a look at the double
locking in trunk.  It sure would be nice to have our cake and eat it too.

Greg

Reply via email to