Evan
-- | Evan Harris - Consultant, Harris Enterprises - eharris@puremagic.com | | Custom Solutions for your Software, Networking, and Telephony NeedsOn Sat, 24 Apr 1999, Evan Harris wrote:
> I upgraded to 2.2.6 from 2.2.3 and found that ipip tunneling is partially > broken. Tried 2.2.5 and it's broken there too. Couldn't try 2.2.4 > because it has compile problems. > > It manifests itself as communication originating from the remote tunnel > endpoint being ignored. Traffic from hosts other than the remote endpoint > that is routed over the tunnel IS seen and acted on. > > In this log, 10.1.1.1 is the remote endpoint, 10.1.1.2 is tunneled, and > 10.1.1.3 is another machine on the remote endpoints network. These are > tcpdumps of traffic as seen by the 10.1.1.2 system. > > With kernel 2.2.5 (same results with 2.2.6) > 20:00:39.041879 10.1.1.3 > 10.1.1.2: icmp: echo request (ipip) > 20:00:39.041979 10.1.1.2 > 10.1.1.3: icmp: echo reply (ipip) > 20:02:36.712234 10.1.1.1 > 10.1.1.2: icmp: echo request (ipip) > 20:02:37.712373 10.1.1.1 > 10.1.1.2: icmp: echo request (ipip) > > With kernel 2.2.3 > 20:20:43.066396 10.1.1.3 > 10.1.1.2: icmp: echo request (ipip) > 20:20:43.066493 10.1.1.2 > 10.1.1.3: icmp: echo reply (ipip) > 20:21:14.144984 10.1.1.1 > 10.1.1.2: icmp: echo request (ipip) > 20:21:14.145081 10.1.1.2 > 10.1.1.1: icmp: echo reply > > Of course, in the log from 2.2.3, the fact that the icmp reply is not > (ipip) shows that there is an asymmetrical route for that traffic, but > that doesn't seem to cause problems with 2.2.3. Maybe the problem is > stricter checking of interfaces traffic is recieved/sent on in the later > kernels, and that is the cause of the problem? If so, is there a > workaround? > > I would be happy to provide any further information that is needed to > track down the problem. > > Evan
- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.rutgers.edu Please read the FAQ at http://www.tux.org/lkml/