Currently, when the client is IPv4 and the destination is IPv4, transparent proxying works.
When the client is IPv6 and the destination is IPv6, it does not, because the wrong type of socket is created.
When the client is IPv4 and the destination is IPv6, no mapping is done and an address family error is thrown.
The address family error is thrown because when using ::ffff:192.0.2.1 as source address, the socket will create an IPv4 packet, not an IPv6 packet.
This patch fixes the socket type for IPv6 to IPv6, and also changes the source address from ::ffff:192.0.2.1 to ::192.0.2.1 for IPv4 to IPv6.
In the server logs on the destinations the IPv4-address is as human readable as possible, and replies go to ::0.0.0.0/96.
These replies can be captured using:
ip6tables -t mangle -A PREROUTING -m socket --transparent -j MARK --set-mark 1
ip -6 rule add fwmark 0x1 lookup 100
ip -6 route add local ::/96 dev lo table 100
With nftables in the table ip6 mangle chain PREROUTING: socket transparent 1 mark set 0x1 accept
Hi,
Currently, when the client is IPv4 and the destination is IPv4, transparent proxying works. When the client is IPv6 and the destination is IPv6, it does not, because the wrong type of socket is created. When the client is IPv4 and the destination is IPv6, no mapping is done and an address family error is thrown.
The address family error is thrown because when using ::ffff:192.0.2.1 as source address, the socket will create an IPv4 packet, not an IPv6 packet.
This patch fixes the socket type for IPv6 to IPv6, and also changes the source address from ::ffff:192.0.2.1 to ::192.0.2.1 for IPv4 to IPv6.
In the server logs on the destinations the IPv4-address is as human readable as possible, and replies go to ::0.0.0.0/96.
These replies can be captured using: ip6tables -t mangle -A PREROUTING -m socket --transparent -j MARK --set-mark 1 ip -6 rule add fwmark 0x1 lookup 100 ip -6 route add local ::/96 dev lo table 100
With nftables in the table ip6 mangle chain PREROUTING: socket transparent 1 mark set 0x1 accept