Can fragments of a packet be refragmented again

icmpipmturouting

In IPv4, fragmentation is done by routers on way to the destination if DF(do not fragment) flag is not set in the IP packet.

Once a packet is fragmented, its fragments may take different paths (due to various reasons like topology changes) to the destination.

If, on some link again in the path to destination, one routers find that the link MTU is smaller than the frame size, then either the packet needs to be fragmented or dropped.

Can fragments of a packet be refragmented again? If yes, what will be the value of MF flag in the new individual fragments created by this?

Best Answer

Yes, they can be refragmented. Every refragment but the last will have the MF flag set. The last fragment will have the same MF flag as the original fragment. So it's just as if they had originally been fragmented that way.

Related Solutions

Implementing IP fragmentation

If the packets are correctly fragmented then reassembled on client side, but the reverse does not work,

A -----> B
         |
A <--x-- B

I would suggest first to try to inverse the roles (since the problem is on the way back) and check if, this time, the fragmented packets/reassembling can be performed by A

A <----- B
|
?

If not, there is a problem in the packets management from B to A, could be a firewall limitation, or any router or switch you may have in between.

Linux – set an interface to not fragment packets

If the DF bit is set then the router, router here being any device that routes a packet from one interface to another, will not fragment. That's the purpose of allowing the host to set the DF bit because you do not want fragmentation.

Regardless, if a router fragments a packet that has the DF bit set then it is not RFC compliant. Rather or not to send an ICMP unreachable message is a different story.

Typically, you want this behavior because it allows for PMTU discovery but a lot of people disable it in fear of DoS attacks, drains routers resources to generate those messages given enough packets.

So now back to your real question ha ha, sorry if that was stuff you already knew. iptables has the ability to drop based on packet size. If you don't want to send those ICMP unreachable messages you can use iptables for that as well...

http://www.linuxtopia.org/Linux_Firewall_iptables/x2682.html

Best Answer

Related Solutions

Implementing IP fragmentation

Linux – set an interface to not fragment packets

Related Topic