Current-Users archive

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index][Old Index]

wm(4) and IPv6



Hi folks,

I am experiencing strange problems with Ipv6 and the wm(4) driver: when, after a delay of inactivity let's say, about one minute or so, I try to ping my router, the first three packets are lost, the forth returns with an exceedingly long round trip time, and the following gets normal:

> ping6 freeboxcimax
PING6(56=40+8+8 bytes) 2a01:5d8:5243:b634::2 --> 2a01:5d8:5243:b634::1
16 bytes from 2a01:5d8:5243:b634::1, icmp_seq=4 hlim=64 time=1001.221 ms
16 bytes from 2a01:5d8:5243:b634::1, icmp_seq=5 hlim=64 time=0.865 ms
16 bytes from 2a01:5d8:5243:b634::1, icmp_seq=6 hlim=64 time=0.648 ms
16 bytes from 2a01:5d8:5243:b634::1, icmp_seq=7 hlim=64 time=0.639 ms
16 bytes from 2a01:5d8:5243:b634::1, icmp_seq=8 hlim=64 time=0.637 ms
16 bytes from 2a01:5d8:5243:b634::1, icmp_seq=9 hlim=64 time=0.788 ms
16 bytes from 2a01:5d8:5243:b634::1, icmp_seq=10 hlim=64 time=2.844 ms
16 bytes from 2a01:5d8:5243:b634::1, icmp_seq=11 hlim=64 time=0.778 ms
16 bytes from 2a01:5d8:5243:b634::1, icmp_seq=12 hlim=64 time=0.641 ms
16 bytes from 2a01:5d8:5243:b634::1, icmp_seq=13 hlim=64 time=0.793 ms
16 bytes from 2a01:5d8:5243:b634::1, icmp_seq=14 hlim=64 time=0.637 ms
^C
--- freebox ping6 statistics ---
15 packets transmitted, 11 packets received, 26.7% packet loss
round-trip min/avg/max/std-dev = 0.637/91.863/1001.221/301.601 ms

When contacted from outside, the ping6 fails. Traceroute6 gives this:

> traceroute6 -v -I cimax
traceroute6 to cimax (2a01:5d8:5243:b634::2) from 2a01:5d8:5139:15db::2, 64 hops max, 16 byte packets 1 freebox 64 bytes of data to 2a01:5d8:5139:15db::2 3.326 ms 1.658 ms 1.121 ms 2 freeboxcimax 64 bytes of data to 2a01:5d8:5139:15db::2 62.155 ms 64.912 ms 56.792 ms
 3  freeboxcimax 64 bytes of data to 2a01:5d8:5139:15db::2  3061.78 ms !A
24 bytes from 2a01:5d8:5139:15db::1 to 2a01:5d8:5139:15db::2: icmp type 136 (Neighbor Advertisement) code 0
0000: 2a0105d8 513915db 00000000 00000001
0010: 00000000 00000002

32 bytes from fe80::207:cbff:fe4b:c921%ath0 to 2a01:5d8:5139:15db::2: icmp type 135 (Neighbor Solicitation) code 0
0000: 2a0105d8 513915db 00000000 00000002
0010: 01010007 cb4bc921 2a0105d8 5243b634
  3056.9 ms !A  3058.85 ms !A

And vice-versa:

traceroute6 -v maison
traceroute6 to maison (2a01:5d8:5139:15db::2) from 2a01:5d8:5243:b634::2, 64 hops max, 12 byte packets
 1
24 bytes from 2a01:5d8:5243:b634::1 to 2a01:5d8:5243:b634::2: icmp type 136 (Neighbor Advertisement) code 0
0000: 2a0105d8 5243b634 00000000 00000001
0010: 00000000 00000000

32 bytes from fe80::207:cbff:feb1:c40d%wm0 to 2a01:5d8:5243:b634::2: icmp type 135 (Neighbor Solicitation) code 0
0000: 2a0105d8 5243b634 00000000 00000002
0010: 01010007 cbb1c40d 00000000 00000000

24 bytes from fe80::207:cbff:feb1:c40d%wm0 to fe80::21d:9ff:fe80:9414: icmp type 136 (Neighbor Advertisement) code 0
0000: fe800000 00000000 0207cbff feb1c40d
0010: 01010007 cbb1c40d

32 bytes from fe80::207:cbff:feb1:c40d%wm0 to fe80::21d:9ff:fe80:9414: icmp type 135 (Neighbor Solicitation) code 0
0000: fe800000 00000000 021d09ff fe809414
0010: 01010007 cbb1c40d 00000000 00000000
* freeboxcimax 68 bytes of data to 2a01:5d8:5243:b634::2 1.607 ms 0.768 ms 2 freebox 68 bytes of data to 2a01:5d8:5243:b634::2 60.004 ms 57.851 ms 58.099 ms 3 maison 68 bytes of data to 2a01:5d8:5243:b634::2 63.043 ms 59.09 ms 58.341 ms

Once the level 3 link is up again, everything works smoothly, until it is paused and the same phenomenon shows up again.

This seems not to be a router default, since I have a machine at the same address (but ...::3%sip0), with -current also, that works like a charm. Also, nothing like this happens when I use IPv4 instead.

Any idea or same behavior?

Vincent

Oh, by the way, I have this config on the faulty machine:

wm0 at pci0 dev 25 function 0: 82801I LAN Controller, rev. 2
wm0: interrupting at ioapic0 pin 20 (irq 10)
wm0: PCI-Express bus
wm0: FLASH
wm0: Ethernet address 00:1d:09:80:94:14
inphy0 at wm0 phy 1: i82562EM 10/100 media interface, rev. 0

Note that also Gigabit Ethernet capable, I have only a 100 Megabit phy. Thanks Dell. Could it pose problems?



Home | Main Index | Thread Index | Old Index