comments on approach: adding new ip packet header from userspace

[Amer Aijaz]

I wanted advice on something I want to do. 

Do you foresee problems, if using QUEUE target I send the ipv4 packets from 
the output chain to a user space program, which would change the IPv4 
packet in a way that it adds a new header and a value in it .. updates header 
length and checksums and reinjects this altered packet back in output chain 
.. I am sure that if I am able to strip off this additional header on the 
computer on the other end of the network cable, using the same QUEUE and 
libipq, this change would be invisible for the upper layers .. is that right?

Is there any problem that you foresee in my approach?

I would treasure some expert comments on this.

Thanks

Amer