My UDP socket is bind()ing to port 53 (DNS). Does UDP have a TIME_WAIT state or is using SO_REUSEADDR pointless on UDP sockets?
UDP doesn't have connections, so there's nothing analogous to TIME_WAIT. You don't need to use SO_REUSEADDR.
If you're listening on a broadcast or multicast address, you might need to use SO_REUSEPORT, so that if there are multiple listeners on the same machine they don't conflict. However, from what I can tell, this doesn't exist on Linux.
Related
I would like to write a daemon that is available on a large number of AF_INET-SOCK_STREAM and AF_INET-SOCK_DGRAM sockets for network debugging purposes.
To avoid excessive resource usage I want to avoid opening a large number of ports on the application layer but try to multiplex the connections per socket type on lower layers.
Knowledge of the original incoming port on the application layer is a requirement.
I have successfully implemented a daemon that listens on an AF_INET SOCK_STREAM socket that is multiplexed by an iptables REDIRECT rule. The original incoming port of the connection can be retrieved by calling getsockopt with SO_ORIGINAL_DST. As I understand this does not work with AF_INET SOCK_DGRAM.
I have also successfully implemented a daemon that listens on an AF_INET SOCK_DGRAM socket that is multiplexed by an iptables TPROXY rule. The original incoming port of the connection can be retrieved by using recvmsg() and consuming the available ancillary message containing information about the connection before multiplexing. As I understand this does not work with AF_INET SOCK_STREAM.
Is there a transport-layer-agnostic way of multiplexing such socket connections and retrieving information about the original incoming port? Possibly even suitable for protocols like SCTP or DCCP?
With TCP you have a connection. The target port for this connection is the same for all packets inside this connection. In this case each connected socket (result from accept) equals to a single connection and the incoming port is a property of this socket. It does not matter in this case if the listening socket will accept connections on multiple ports, all what matters is the connected socket.
With UDP you don't have a connection. Instead the same socket is used to receive packages from multiple clients and in your case to multiple incoming ports. Source and destination IP and port are thus a property of each packet and not of the socket.
That's why you need different interfaces to retrieve the original incoming port: a socket based for TCP and a packet based for UDP.
What is the best way to tell if traffic is reaching a socket?
netstat -i is convenient for checking counters in/out of an interface, but how to tell if that traffic is reaching a socket?
Thanks
You can use netstat to check the count of ESTABLISHED connection to the port number that the socket is using to listen. Alternatively tcpdump can show you packets flowing in and out of that port.
situation: I need to broadcast from client using UDP from some free port, then to accept tcp connection on client from server on the port with the same number, but TCP. That's why I need to listen (and bind) to this port BEFORE broadcasting. The port cannot be const, because I can have multiple clients running on one machine.
So here some questions, which can help me to make this situation clearer:
If I made sendto from unbinded UDP socket, is it binded to any free port and all next sendto messages will go from this port, or each time the port will be chosen for a new message?
Can I ask system to reserve some free port for me? (I need to reserve two ports with the same numbers for UDP and TCP connections)
I'm sure there is a known way to handle these situations, what is it?
1)If I made sendto from unbinded UDP socket, is it binded to any free port and all next sendto messages will go from this port
Yes.
or each time the port will be chosen for a new message?
No.
2) Can I ask system to reserve some free port for me? (I need to reserve two ports with the same numbers for UDP and TCP connections)
That's what happens when the auto-bind occurs. You can do explicitly by binding to port number zero, but it isn't necessary. It also does not guarantee that you can bind both UDP and TCP to the same port number.
3) I'm sure there is a known way to handle these situations, what is it?
You've found it. Let the auto-bind happen.
I've found some answers on stackoverflow
You can bind to 0 port, that is specified it struct semaddr_in. That will allow you get an unused port for your type of connection. It is not defined to be free for other types of connection.
Look at #remy's answer at Bind to any port available
Yes, they can because headers are specified for UDP or TCP protocols. So the machine can tell one from another.
See Can TCP and UDP sockets use the same port?
If you bind to 0 port you'll get ability to listen to it before calling sendto
I need to create a virtual IP network over TCP connection. The hosting system is Linux, with the TUN/TAP kernel driver, it's quite easy to receive & re-inject IP packets of the virtual network.
The difficult part is to transmit the received IP packets to another host. For some non-technical reasons, I can only transmit the packets over TCP protocol but not UDP. Transmit IP packets over UDP is easy, but with TCP it becomes tricky, here's the reason:
UDP protocol doesn't support retransmission/reordering, just like IP. So, if one UDP packet is sent for every received virtual IP packet, the kernel TCP/IP protocol stack would still see virtual IP packet loss/duplication/reordering(those are required for TCP/IP to work well, if those "features" are missing, the TCP connection speed on the virtual network would suffer). If IP packets are transmitted over TCP all required "features" will be missing, unless they are simulated some how.
It seems I have to fake some kind of packet duplication/loss/reordering on TCP connection, or patch the kernel TCP/IP protocol stack. Both options aren't easy.
Is there any other simpler solution to my problem ? or did I just go into a completely wrong direction ? I'm all ears.
==== UPDATE ====
I'm thinking about using raw IP socket (which could get rid of all the TCP retransmission/reordering stuff on the physical network easily while still using TCP packets) to transmit the received virtual network IP packets. But on the receiving host, how can I only receive the packets I'm interested in and return all other IP packets to the kernel TCP/IP stack ?
First of all, you do not want to make a VPN over TCP because you would end up with tcp-over-tcp eventually. The main issue is that the timers of your inner TCP and outer TCP might differ significantly which negatively impacts your TCP session reliability. You can find a bit longer explanation here.
UDP protocol doesn't support retransmission/reordering, just like IP. So, if one UDP packet is sent for every received virtual IP packet, the kernel TCP/IP protocol stack would still see virtual IP packet loss/duplication/reordering(those are required for TCP/IP to work well, if those "features" are missing, the TCP connection speed on the virtual network would suffer). If IP packets are transmitted over TCP all required "features" will be missing, unless they are simulated some how.
This does not make sense, if your outer layer uses TCP as a transport mechanism, nothing stops your inner layer to still use the full ip/tcp stack, including those features. They can conflict badly like I said, but it's not that this functionality disappears or breaks completely.
It seems like you actually want to use TCP just to have the headers and ignore the actual protocol, this would indeed avoid the issues with tcp over tcp. However, once again this is a very bad idea. Flow processing for firewalls, NAT, DPI, tcp boosters, becomes more and more common, if you fake TCP packets you might up stressing those boxes, possibly detoriating your own connection once again.
So you should ask yourself why you can't use UDP, and if no alternative protocol (header) is okay, like GRE or L2TP.
How to open a raw socket for sending from specific TCP port? I want to have all my connections always go from a range of ports below ephemerals.
If you are using raw sockets, then just fill in the correct TCP source port in the packet header.
If, instead, you are using the TCP socket interface (socket(), connect() and friends), then you can set the source port by calling the bind() system call for the client socket - exactly as you would to set the listening port for the server socket.
Making a tcp connection using raw sockets is somewhere between difficult and impossible; you'd need to implement the entire tcp protocol in your program AND also stop the kernel from sending its own replies to the packets (if the kernel has IP bound on that address on that interface).
This is probably not what you want. However, if you did want it, it is trivial to send tcp frames with any source port you want, as you get to specify it in the tcp header, which of course, if you're implementing your own TCP layer, you'll need to understand.