In some email I received from Manuel Kasper, sie wrote: > On 5.2.2003 17:16, "ryan beasley" <[EMAIL PROTECTED]> wrote: > > > Assuming my understanding of the interface indexes is correct, I > > would've expected something like > > interfaces: in qfe0,qfe1 out qfe1,qfe0 where 192.168.2.12 is on qfe0 > > and 212.238.176.190 on qfe1 > > Me too... I found a bug in ipfilter 3.4.29 that was related to resyncing ipf > with "ipf -y" - each time the first input interface in each state entry got > overwritten with the first output interface because of a small mistake in > ip_state.c (is->is_ifpin is used in ip_statesync() where is->is_ifp[i] > should be used). Darren knows about the bug; I'm not sure if it's fixed in > 3.4.31, though.
There is a bug there in 3.4.31. Patch can be found listed on: http://coombs.anu.edu.au/~avalon/ipfpatches/index.html Darren
