I take that back, it may still be necessary because you need to compute the 
checksum beforehand.

-----Original Message-----
From: Daniel Ng, Mr 
Sent: Tuesday, October 27, 2009 11:56 AM
To: Xi Chen
Cc: gini@cs.mcgill.ca
Subject: RE: UDP programming

Ok, it seems that this is not even necessary, as this step is performed in 
IPOutgoingPacket.  

-----Original Message-----
From: gini-boun...@cs.mcgill.ca [mailto:gini-boun...@cs.mcgill.ca] On Behalf Of 
Daniel Ng, Mr
Sent: October 27, 2009 11:40 AM
To: Xi Chen
Cc: gini@cs.mcgill.ca
Subject: Re: [gini] UDP programming

I imagine most groups will ask this question at some point, so here it is.  Use 
findRouteEntry from routetable.c first to find your interface number, then use 
that number in findInterfaceIP from mtu.c to get the source address.  As for 
the port I believe it is your choice.  This is not necessarily the only 
solution, so if you have it working another way (other than hardcoding), there 
is no need to change it.

Daniel

-----Original Message-----
From: Xi Chen 
Sent: October 27, 2009 10:38 AM
To: Daniel Ng, Mr
Subject: RE: UDP programming

Hi Daniel,

I understand when we process the UDP packet, we can get the source IP and port 
from the IP header, but what about when we send the UDP packet? Do you know 
what method to call to get those information?

Thanks,

Xi

_______________________________________________
gini mailing list
gini@cs.mcgill.ca
http://mailman.cs.mcgill.ca/mailman/listinfo/gini
_______________________________________________
gini mailing list
gini@cs.mcgill.ca
http://mailman.cs.mcgill.ca/mailman/listinfo/gini

Reply via email to