nat vs. state

[Michael Sierchio]

Earl A. Killian wrote:
> Is it safe to assume packets diverted to NAT are "safe" and don't need
> further checking?  In particular, can the use of dynamic/stateful
> rules be skipped for NAT packets?  It seems so, because NAT is already
> stateful.

Safe?  Define "safe." ;-)

For *dynamic* nat, probably so.  For static nat (port/addr redirect)
you'll probably want to have robust rules after diverting to natd.