Re: [PATCH] First part of TCP-MD5 inbound verification

Just a note that, as discussion on nanog shows, it's very important to only do the md5 check if the incoming packet is going to be accepted and processed, rather than the intuitive order of checking the sig first. That's because checking first allows an easy DoS, since checking is cpu-intensive. B

[Fwd: NetBSD Security Advisory 2004-006: TCP protocol andimplementation vulnerability]

FYI--- Begin Message --- The additional implementation flaw of BSD based TCP/IP stacks has been fixed in FreeBSD in revision 1.81 of tcp_input.c in 1998 for FreeBSD 2.2 and 3.0 and all releases since about six years ago. -- Andre NetBSD Security-Officer wrote: > > -BEGIN PGP SIGNED MESSAG

Re: simulating an LFN over 1Gb LAN Ethernet?

Luigi Rizzo wrote: On Thu, Apr 22, 2004 at 04:26:26PM +0200, Heinz Knocke wrote: I'd like to simulate an LFN over LAN - my idea is to install testing software on 2 hosts, traffic between them would be routed by the 3rd one - a FreeBSD based router. To simulate long RTT the router would have to del

Re: RFC: if_clone overhaul

On Thu, Apr 22, 2004 at 04:13:40PM +0200, Andre Oppermann wrote: > Brooks Davis wrote: > > > > Please test/review the following patch to the network interface cloneing > > code. This code is a major overhaul of the cloning infrastructure. > > > > The significant include: > > - Split the code ou

Re: simulating an LFN over 1Gb LAN Ethernet?

Heinz Knocke wrote: > > Hi! > > I'd like to simulate an LFN over LAN - my idea is to install testing software > on 2 hosts, traffic between them would be routed by the 3rd one - a FreeBSD > based router. To simulate long RTT the router would have to delay packet > forwarding in at least one direc

Re: simulating an LFN over 1Gb LAN Ethernet?

On Thu, Apr 22, 2004 at 04:26:26PM +0200, Heinz Knocke wrote: > Hi! > > I'd like to simulate an LFN over LAN - my idea is to install testing software on 2 > hosts, traffic between them would be routed by the 3rd one - a FreeBSD based router. > To simulate long RTT the router would have to delay

simulating an LFN over 1Gb LAN Ethernet?

Hi! I'd like to simulate an LFN over LAN - my idea is to install testing software on 2 hosts, traffic between them would be routed by the 3rd one - a FreeBSD based router. To simulate long RTT the router would have to delay packet forwarding in at least one direction - does anyone know how to d

Re: RFC: if_clone overhaul

Brooks Davis wrote: > > Please test/review the following patch to the network interface cloneing > code. This code is a major overhaul of the cloning infrastructure. > > The significant include: > - Split the code out into if_clone.[ch]. > - Locked struct if_clone. Derived from work by Mauryc

[PATCH] First part of TCP-MD5 inbound verification

Hey guys, I'm really pressed for time at the moment and people are demanding a lot of other things from me. So I'd like to float this patch set against HEAD which does inbound TCP-MD5 verification, so far for SYNs only. I took a decision to use sysctls rather than enlarge struct tcpstat to avoid

Re: Assigning a specific IP address and Interface with MPD

Hi, Jay Hall schrieb: Is it possible to do this? Or should I work on trying to add the routes as the interfaces come up based on the IP address they are assigned? 2 possibilities: a) use RADIUS or b) use an iface-up/down script where you add/remove your routes. bye, -- -

Assigning a specific IP address and Interface with MPD

I have setup a VPN for the company I work for in which all of the remote offices connect to the Headquarters office using MPD. And this works great until I have to re-establish the connections. What I am trying to do, and maybe there is a better way, is to control what ng interface a client co