[Ns-bugs] [Bug 748] Cloned TCP socket uses wrong source address

[code@nsnam.ece.gatech.edu]

http://www.nsnam.org/bugzilla/show_bug.cgi?id=748


Josh Pelkey <jpelkey at gatech.edu> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |jpelkey at gatech.edu,
                   |                            |riley at ece.gatech.edu




--- Comment #1 from Josh Pelkey <jpelkey at gatech.edu>  2010-02-02 15:08:54 EDT ---
(In reply to comment #0)
> Created an attachment (id=668)
 --> (http://www.nsnam.org/bugzilla/attachment.cgi?id=668) [details]
> patch proposal
> 
> A listening TCP socket, to which a connection is opened, forks a new socket and
> determines the source address to use from the route lookup, instead of using
> the destination address to which the connection request originally has been
> sent to. This leaves the connection in a state where no communication is
> possible, as the endpoint of the forked socket does not match the address the
> client socket is sending packets to.
> 
> To solve this, it is required to pass the packet's destination address to the
> socket. This also requires changes to the Ipv4Endpoint API. 
> 
> The proposed patch includes changes to the TCP forking process and the
> Ipv4Endpoint API. Other protocols are only modified to match the new
> Ipv4Endpoint API to be able to compile it, but do not make explicit use of it.

Hi Fabian,

I think I follow this mostly.  Shouldn't the source address from the route
lookup give you what you want?  I'm not saying that it is the right way to do
it, but you suggested that it doesn't work.

Finally, the patch attached can no longer successfully be applied to ns-3-dev. 
In any case, I can still follow what you've done from it.

-- 
Configure bugmail: http://www.nsnam.org/bugzilla/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.