ipfw/ipf IP filtering thoughts

[Richard Coleman]

Antti Louko wrote:

> Generally, I like the (Free)BSD way of doing things.  But the IP
> filtering modules available for FreeBSD lack one feature when compared
> to Linux way (ipchains and iptables).
> 
> In ipchains and iptables you have a sequential list of rules, very
> much like in ipfw and ipf, but you can have several different lists
> which have symbolic names and you can make calls from lists to other
> lists based on normal packet criteria.  If the list is exchausted, the
> scan returns to the previous list.  This makes it possible to make
> filtering decisions much more efficient in complex situation.  You can
> for example scan a certain list only for eg. packets going to for
> example port 25 and so on.  In FreeBSD, you don't have this
> "subroutine call" feature at all and you are limited to only one
> sequential list with a "goto".

In ipf, you can use head/group tags to create a tree-like structure for 
the rules processing.  This works roughly like a subroutine call.  But 
admittedly this starts to look very confusing if you have many levels of 
these.

Pf handles this much cleaner in that the tree structure is derived 
automatically, as long as the rules are ordered correctly.  There is a 
port for pf in the ports tree.  But I haven't tried it, since ipf is 
working fine for me.

Richard Coleman
richardcoleman at mindspring.com