You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
{{ message }}
This repository was archived by the owner on May 10, 2020. It is now read-only.
How challenging do you think it would be to add TCP support? I'm diving into the code myself and looking at potential gotchas so I'm wondering what you think as well.
Currently, it filters using icmpid, which is a great trick, but that wouldn't work for TCP. For TCP, it may be possible using a FWMARK, however, those are lost when the routing across network namespaces. Instead, I'm guessing I'd need to possibly:
Create a TCP socket on the python side
Bind the source port before sending any packets
Connect
Track every time iptables modified the source/dest IPs
Alternatively, I guess I could track the TCP sequence numbers rather than the source/dest which might be more reliable.
What do you think?
The text was updated successfully, but these errors were encountered:
Sign up for freeto subscribe to this conversation on GitHub.
Already have an account?
Sign in.
How challenging do you think it would be to add TCP support? I'm diving into the code myself and looking at potential gotchas so I'm wondering what you think as well.
Currently, it filters using icmpid, which is a great trick, but that wouldn't work for TCP. For TCP, it may be possible using a FWMARK, however, those are lost when the routing across network namespaces. Instead, I'm guessing I'd need to possibly:
Alternatively, I guess I could track the TCP sequence numbers rather than the source/dest which might be more reliable.
What do you think?
The text was updated successfully, but these errors were encountered: