From: Jeff Dike <jdike@ka...> - 2001-10-01 17:17:03
> There are other ways to get this functionality, including booting a
> UML on an mcast network connection and using plain tcpdump, but for
> various reasons I want to use uml_router instead.
Care to elaborate? I'm curious what they are.
> Jeff, once tcpdump.org figures out when the code will be more
> mainstream, I'll redo the patch with #ifdef, etc. so that you can just
> apply it.
OK, it's not really fit to go in the pool right now :-)
From: Michael Richardson <mcr@sa...> - 2001-10-02 01:39:11
-----BEGIN PGP SIGNED MESSAGE-----
>>>>> "Jeff" == Jeff Dike <jdike@...> writes:
Jeff> mcr@... said:
>> There are other ways to get this functionality, including booting a
>> UML on an mcast network connection and using plain tcpdump, but for
>> various reasons I want to use uml_router instead.
Jeff> Care to elaborate? I'm curious what they are.
Solution 1: boot another UML and run tcpdump.
1) overhead. That's a lot just to run tcpdump.
2) promisc. While mcast has promisc, uml_router does not as
far as I can see. I have need to use uml_router later.
3) independant view. UML uses Linux networking code. I would
therefore be using code that I may be suspicious of.
Solution 2: use tcpdump on existing UML.
1) I can't get boot time packet traces.
2) UML network devices see two copies of mcast'ed traffic. One
going out, and the other when it echos back.
3) now I'm really using code under test.
I could also use ethertap/tun, and run tcpdump on host. This requires root
on host, and gets the hosts' network mixed into things.
] ON HUMILITY: to err is human. To moo, bovine. | firewalls [
] Michael Richardson, Sandelman Software Works, Ottawa, ON |net architect[
] mcr@... http://www.sandelman.ottawa.on.ca/ |device driver[
] panic("Just another NetBSD/notebook using, kernel hacking, security guy"); [
-----BEGIN PGP SIGNATURE-----
Comment: Finger me for keys
-----END PGP SIGNATURE-----