use ml_purge to flush the arp hold lists instead of dequeue/free loops.

this is an interesting change because the loops have to decrement
the global hold count too. they looked like this:

	while ((mh = ml_dequeue(&la->la_ml)) != NULL) {
		la_hold_total--;
		m_freem(mh);
	}

because ml_purge returns how many mbufs were freed, we can do this:

	la_hold_total -= ml_purge(&la->la_ml);

ok mpi@ bluhm@

0 files changed, 0 insertions, 0 deletions