My initial thought in this regard, was to insert the Agent into Node N1's
entry, and then set the Agent's target_ to the Clasifier in Node N1.
Here's what I did in my Agent's command function:
--------------------------------------
int MyAgent::command(int argc, const char*const* argv) {
Tcl& tcl = Tcl::instance(); // tcl instance
if (argc == 3) {
if (strcmp(argv[1], "set-target") == 0) {
target_ = (NsObject*)(TclObject::lookup(argv[2]));
printf("Setting MyAgent target_ to %s\n", argv[2]);
if (target_ == NULL) {
tcl.add_errorf("MyAgent target_ was NULL: %s", argv[2]);
return TCL_ERROR;
}
return(TCL_OK);
}
}
return (Agent::command(argc, argv));
}
--------------------------------------
Then, in tcl I do something like this:
--------------------------------------
set n1_old_entry [$n1 entry] ;# points to classifier_
$n1 insert-entry [new RtModule] $my_agent
$my_agent set-target $n1_old_entry
--------------------------------------
However, when I attept to send a packet across N1, I get a Bus Error.
Any ideas? Am I going about this all wrong?
Thanks in advance!
Brad
On 7/13/06 1:45 AM, "Brad Montgomery" <[EMAIL PROTECTED]> wrote:
>
> All,
>
> I would like to write an Agent that inspects all packets that pass through
> the node to which it is connected. For example:
>
> Given the topology,
>
> N0 - N1 - N2
>
> An Agent attached to N1 would execute it's recv() function even if N2 was
> sending a packet to N0. Is this possible? If so, any pointers to examples
> or descriptions would be greatly appreciated!
>
> Also, as a new NS user, I'm still a little unsure of the exact "path through
> the ns code" that a packet takes on it's way from a sender to a destination.
> Is there a more succinct description of this process than what is in the ns
> documentation?
>
> Thanks!
>
> Brad
>
