Thanks for your responce, in my case I have a wired scenario and I use the default (static) routing. As I say in the previous post, I do all the work in the connector class. I don't really care from which layer the packet is dropped, as long as I have access to the packet to read the information from it.
So, if I got this correct, your suggestion is to move the code out from the connector class and put it in the routing agent for example? In my case that I use static routing, which is the file that I should edit? Sorry for the amount of question but it is the first time I work with the c++ code of ns2. So far I never had to change anything to perform my simulations. Thanks again, Giorgos On 3 August 2012 16:07, Ali Rostami <ali.rostami...@gmail.com> wrote: > first of all you should determine which layer is responsible for dropping > the packet? > I suppose it's routing agent, for example. then, put this line wherever > you want to drop the packet p: > drop(p, DROP_RTR_YOUR_REASON); > put this at line 97 in file %NS_DIRECTORY%/trace/cmu-trace.h > #define DROP_RTR_YOUR_REASON "XXX" > then run the following command: > touch %NS_DIRECTORY%/trace/cmu-trace.cc > now you can make the ns code. when the ns drops a packet under your > condition, the trace will be like this: > D 2.828954441 1 RTR XXX 2 cbr 512 [0 0 0 0] ------- [1:0 20:0 32 0] [2] 0 > 3 0 > > Good luck! > > > > On Fri, Aug 3, 2012 at 4:49 PM, Giorgos Karanikas > <computera...@gmail.com>wrote: > >> >> >> To provide some more info: >> >> Suppose I have the following topology: >> >> http://old.nabble.com/file/p34250683/topology.jpeg >> >> What I want is node D to drop packets from some specific nodes. What I've >> done is the following: >> I added one variables in the node.cc && node.h to define if the packets >> from >> that node are to be dropped and I also added one variable to the packet.cc >> (in the header of the packet) to define after how many hops the packet is >> to >> be dropped. >> >> In the connector.cc I have changed the recv(Packet* p, Handler* h) >> function >> in order to check if the packet is to be dropped. If not then I update the >> packet's hop variable I created (increase it to show the number of hops >> that >> packet has been over). >> If the packet is to be dropped though I call the drop(p) function. >> >> As I said that does not work properly since (probably) the Packet::free(p) >> function is used as I don't see the packet drop events in the traces. >> >> Anyone knows what I'm doing wrong? Should I define a drop-target for D and >> connect it to the agents from the nodes that I want packets to be dropped? >> And exactly how am I supposed to do that (since I tried but I failed)? >> >> Thanks in advance and I hope that these information give a clearer view of >> my situation. >> -- >> View this message in context: >> http://old.nabble.com/Properly-drop-packet-tp34250208p34250683.html >> Sent from the ns-users mailing list archive at Nabble.com. >> >> >
