please see PR 9978 about TCP behavior change due to RST rate limitation.
	There are multiple ways to address this problem:
	- disable RST rate limitation for local packets, like we did for
	  IPv4 ICMP error rate lmitation.
	- do not use ratelimit(9), use ppsratelimit(9).  ratelimit does not
	  work right for smaller-than-1/hz limitation.
	- do not do rate limitation at all.
	- use more clever algorithm for rate limitation
	the last one should need longer research, should be post 1.5 item.

	I believe the right way (quick solution for 1.5) is the second one.
	we cannot really trust local traffic (local traffic does not fill
	network bandwidth, however, it still chews CPU time) so the first 
	one will be ruled out.  Is it okay if I make the following changes:
	- tcp_input to use ppsratelimit(9), instead of ratelimit(9)
	and also,
	- remove "loopback is okay" code from IPv4 ICMP error rate limitation.

itojun