Il 08/10/2013 22:16, Hans de Goede ha scritto: > No, it is calling main_loop_wait with nonblocking set to 0, so > normally the lock would get released. But > timerlistgroup_deadline_ns(&main_loop_tlg) is returning 0, > causing timeout_ns to be 0, and this causes the lock to not get > released.
Yes, this was my understanding of the patch as well. Before Alex's series, this would be capped to MIN_REARM_TIMER_NS (250 us). This is why I mentioned 250 us. However, I agree with Alex that it looks a bit fishy and I'd like to know what timer is it that is continuously set to expire in the past. Paolo
