RFC: untimeout

[Till Straumann]

I wanted to add an 'untimeout' implementation (when porting
drivers, timeout/untimeout are often required) to
libnetworking/rtems/rtems_glue.c and came across
the following glitch:

If the networking daemon is soundly asleep (RTEMS_NO_TIMEOUT)
and a timeout is scheduled by another task (e.g., driver task)
then the timeout might never expire. I believe the 'timeout'
routine should wake up the network daemon if the calltodo
list is empty prior to enqueueing the new entry.

Should we use an extra event number for this purpose?

-- Till