nftables: use after-srcnat postrouting hook. enable ipfrag notrack

This commit is contained in:
bol-van 2022-03-22 13:58:02 +03:00
parent f3508030b1
commit 38883a67a9
3 changed files with 22 additions and 1 deletions

View File

@ -63,6 +63,10 @@ nft_del_all_chains_from_table()
nft_create_chains() nft_create_chains()
{ {
# NOTE : postrouting hook has priority 101 to hook packets already processed by nat
# NOTE : this is the only way to implement ipfrag for forwarded packets if nat is present
# NOTE : to avoid retracking and sport changing use notrack for nfqws generated packets
# NOTE : this also prevents defragmentation of ipv6 fragmented frames in kernels 4.16+
cat << EOF | nft -f - cat << EOF | nft -f -
add chain inet $ZAPRET_NFT_TABLE dnat_output { type nat hook output priority -101; } add chain inet $ZAPRET_NFT_TABLE dnat_output { type nat hook output priority -101; }
flush chain inet $ZAPRET_NFT_TABLE dnat_output flush chain inet $ZAPRET_NFT_TABLE dnat_output
@ -79,7 +83,9 @@ cat << EOF | nft -f -
add rule inet $ZAPRET_NFT_TABLE localnet_protect ip daddr $TPWS_LOCALHOST4 return comment "route_localnet allow access to tpws" add rule inet $ZAPRET_NFT_TABLE localnet_protect ip daddr $TPWS_LOCALHOST4 return comment "route_localnet allow access to tpws"
add rule inet $ZAPRET_NFT_TABLE localnet_protect ip daddr 127.0.0.0/8 drop comment "route_localnet remote access protection" add rule inet $ZAPRET_NFT_TABLE localnet_protect ip daddr 127.0.0.0/8 drop comment "route_localnet remote access protection"
add rule inet $ZAPRET_NFT_TABLE input iif != lo jump localnet_protect add rule inet $ZAPRET_NFT_TABLE input iif != lo jump localnet_protect
add chain inet $ZAPRET_NFT_TABLE postrouting { type filter hook postrouting priority -151; } add chain inet $ZAPRET_NFT_TABLE postrouting { type filter hook postrouting priority 101; }
add chain inet $ZAPRET_NFT_TABLE predefrag { type filter hook output priority -401; }
add rule inet $ZAPRET_NFT_TABLE predefrag mark and $DESYNC_MARK !=0 notrack comment "do not track nfqws generated packets to avoid nat tampering and defragmentation"
flush chain inet $ZAPRET_NFT_TABLE postrouting flush chain inet $ZAPRET_NFT_TABLE postrouting
add set inet $ZAPRET_NFT_TABLE lanif { type ifname; } add set inet $ZAPRET_NFT_TABLE lanif { type ifname; }
add set inet $ZAPRET_NFT_TABLE wanif { type ifname; } add set inet $ZAPRET_NFT_TABLE wanif { type ifname; }
@ -98,6 +104,7 @@ cat << EOF | nft -f - 2>/dev/null
delete chain inet $ZAPRET_NFT_TABLE forward delete chain inet $ZAPRET_NFT_TABLE forward
delete chain inet $ZAPRET_NFT_TABLE input delete chain inet $ZAPRET_NFT_TABLE input
delete chain inet $ZAPRET_NFT_TABLE postrouting delete chain inet $ZAPRET_NFT_TABLE postrouting
delete chain inet $ZAPRET_NFT_TABLE predefrag
delete chain inet $ZAPRET_NFT_TABLE flow_offload delete chain inet $ZAPRET_NFT_TABLE flow_offload
delete chain inet $ZAPRET_NFT_TABLE localnet_protect delete chain inet $ZAPRET_NFT_TABLE localnet_protect
EOF EOF

View File

@ -478,6 +478,13 @@ It must be done manually, `blockcheck.sh` cannot auto fix this for you.
Or just move to `nftables`. You can create hooks with any priority there. Or just move to `nftables`. You can create hooks with any priority there.
Looks like there's no way to do ipfrag using iptables for forwarded traffic if NAT is present.
`MASQUERADE` is terminating target, after it `NFQUEUE` does not work.
nfqws sees packets with internal network source address. If fragmented NAT does not process them.
This results in attempt to send packets to internet with internal IP address.
You need to use nftables instead with hook priority 101 or higher.
## tpws ## tpws
tpws is transparent proxy. tpws is transparent proxy.

View File

@ -507,6 +507,13 @@ options ip6table_raw raw_before_defrag=1
Либо можно раз и навсегда избавиться от этой проблемы, используя nftables. Там можно создать netfilter hook Либо можно раз и навсегда избавиться от этой проблемы, используя nftables. Там можно создать netfilter hook
с любым приоритетом. Используйте приоритет -401 и ниже. с любым приоритетом. Используйте приоритет -401 и ниже.
При использовании iptables и NAT, похоже, что нет способа прицепить обработчик очереди после NAT.
MASQUERADE является финальным таргетом, после него NFQUEUE не срабатывает.
Пакет попадает в nfqws с source адресом внутренней сети, затем фрагментируется и уже не обрабатывается NAT.
Так и уходит во внешюю сеть с src ip 192.168.x.x. Следовательно, метод не срабатывает.
Видимо единственный рабочий метод - отказаться от iptables и использовать nftables.
Хук должен быть с приоритетом 101 или выше.
tpws tpws
----- -----