James, all
while I was testing mpls linux I found the following "strange" behaviour
while pinging or generating UDP-TCP/IP traffic. My testbed is in this
case was very simple:
two linux box:
linux box 1
eth1 Link encap:Ethernet HWaddr FE:FD:C0:A8:00:01
inet addr:192.168.0.1 Bcast:192.168.0.255 Mask:255.255.255.0
[root@mandrake82 root]# cat /proc/net/mpls_*
0x40032400 0/0/0 gen 201 0 1 POP PEEK
eth0 0 7
0x00000002 120/8950/0 5 PUSH(gen 200) SET(eth1,192.168.0.2)
[root@mandrake82 root]# /mnt/host/usr/src/iproute2/ip/ip route
192.168.0.2 via 192.168.0.1 dev eth1 lsp 0x2
10.0.0.0/8 dev eth0 proto kernel scope link src 10.192.141.250
[root@mandrake82 root]#
linux box 2
[root@mandrake82 root]# ifconfig
eth0 Link encap:Ethernet HWaddr FE:FD:C0:A8:00:02
inet addr:192.168.0.2 Bcast:192.168.0.255 Mask:255.255.255.0
[root@mandrake82 root]# cat /proc/net/mpls_*
0x40032000 84/6262/0 gen 200 0 1 POP PEEK
eth0 0 6
[root@mandrake82 root]#
[root@mandrake82 root]# /mnt/host/usr/src/iproute2/ip/ip route
192.168.0.1 via 192.168.0.2 dev eth0 lsp 0x4 <---------- or different
lsp, doent matter
I dont remember if this problem was already raised by Oliver. With this
configuration (in the second box there is no PUSH opcode, while in the
first one there is corresponding POP) if I ping 198.168.0.2, or I
generate TCP/UDP traffic things works fine and no packets are discarded
as I could expect. I took a tracedump of eth1 and I could see that
outgoing packet have MPLS shim, while ingoing one no (in the situation
ping 192.168.0.2). As attachment I put the result.
James, is anyone working on this issue?
Thanks
Luca
|
