At the end of ip6_forward(), is there a good reason why IPSTATS_MIB_OUTFORWDATAGRAMS and IPSTATS_MIB_OUTOCTETS are incremented before the NF_HOOK? If the hook steals or drops the packet, this counts still go up, which seems incorrect.
v4/ip_forward() increments these counters in ip_forward_finish(). It seems that v6 should do it in ip6_forward_finish() ?? Thoughts? Jeff
