ath: truncated-ip

[Sam Leffler]

Jiri Mikulas wrote:
> Hello
> Maybe there is problem with ath
> I have point-to-point uplink with ath 5213 based cards
> with
> 
>   6.0-CURRENT #2: Thu Jun 30 02:39:34 CEST 2005
>   hw.ath.hal.version: 0.9.14.9
>   ifconfig_ath0="10.27.64.78/30 ssid HA2BH media OFDM/24Mbps mode 11a up"
>    bhillz-F# ifconfig ath0 list ap
>    SSID            BSSID              CHAN RATE  S:N   INT CAPS
>    HA2BH           00:90:4b:ca:33:52  149   54M 32:0   100 E
> 
> and on the other side:
>    5.4-STABLE #6: Sat May 28 11:01:07 CEST 2005
>    hw.ath.hal.version: 0.9.6.3
>    ifconfig_ath0="10.27.64.77/30 ssid HA2BH channel 149 media autoselect
> mode 11a mediaopt hostap up"
>    habr# wicontrol ath0 -l | grep signal
>         Quality/Signal/Noise [signal]:  [ 0 / 34 / 0 ]
> 
> 
> The problem is, that "client" side seems to be sendig corrupted data
> (tcpdump above), I tried to switch client<->hostap cards (in config) but
> the same thing again.
> I also tried CURRENT <-> CURRENT  machines, it do the same thing
> I have more interfaces (fxp, wi, vlan) on these boxes and OSPF runs
> there without any problem, so I think it could be ath specific problem ?
> 
> Does anybody have any idea where could be the problem ?
> If will anybody be interested in, I can provide more info.

This is likely tcpdump not knowing that there is some padding in the 
packets for the atheros h/w; can't recall which cases this occurs and/or 
whether it's exposed at the 802.11 layer (it definitely can show up if 
you tap with -y IEEE802_11_RADIO).  I would sniff from another station 
to be sure.

Can you be sure packets are being lost and/or discarded?  statistics at 
the ath and/or 802.11 level will tell you.

	Sam

> 
> ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
> habr# tcpdump -i ath0 proto ospf
> tcpdump: verbose output suppressed, use -v or -vv for full protocol decode
> listening on ath0, link-type EN10MB (Ethernet), capture size 96 bytes
> 18:01:48.994769 IP 10.27.64.78 > OSPF-ALL.MCAST.NET: OSPFv2, Hello (1),
> length: 44
> 18:01:49.013819 IP truncated-ip - 11200 bytes missing! 10.27.64.78 >
> OSPF-ALL.MCAST.NET: OSPFv2, Hello (1), length: 11244
> 18:01:49.398200 IP 10.27.64.77 > OSPF-ALL.MCAST.NET: OSPFv2, Hello (1),
> length: 48
> 18:01:58.995047 IP 10.27.64.78 > OSPF-ALL.MCAST.NET: OSPFv2, Hello (1),
> length: 44
> 18:01:59.048889 IP truncated-ip - 11200 bytes missing! 10.27.64.78 >
> OSPF-ALL.MCAST.NET: OSPFv2, Hello (1), length: 11244
> 18:01:59.408431 IP 10.27.64.77 > OSPF-ALL.MCAST.NET: OSPFv2, Hello (1),
> length: 48
> 18:02:08.995264 IP 10.27.64.78 > OSPF-ALL.MCAST.NET: OSPFv2, Hello (1),
> length: 44
> 18:02:09.038394 IP truncated-ip - 11200 bytes missing! 10.27.64.78 >
> OSPF-ALL.MCAST.NET: OSPFv2, Hello (1), length: 11244
> 18:02:09.418535 IP 10.27.64.77 > OSPF-ALL.MCAST.NET: OSPFv2, Hello (1),
> length: 48
> 18:02:18.995734 IP 10.27.64.78 > OSPF-ALL.MCAST.NET: OSPFv2, Hello (1),
> length: 44
> 18:02:19.016646 IP truncated-ip - 11200 bytes missing! 10.27.64.78 >
> OSPF-ALL.MCAST.NET: OSPFv2, Hello (1), length: 11244
> 18:02:19.428675 IP 10.27.64.77 > OSPF-ALL.MCAST.NET: OSPFv2, Hello (1),
> length: 48
> 18:02:28.996688 IP 10.27.64.78 > OSPF-ALL.MCAST.NET: OSPFv2, Hello (1),
> length: 44
> 18:02:29.051713 IP truncated-ip - 11200 bytes missing! 10.27.64.78 >
> OSPF-ALL.MCAST.NET: OSPFv2, Hello (1), length: 11244
> 18:02:29.438828 IP 10.27.64.77 > OSPF-ALL.MCAST.NET: OSPFv2, Hello (1),
> length: 48
> 
> ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
> bhillz-F# tcpdump -i ath0 proto ospf
> tcpdump: verbose output suppressed, use -v or -vv for full protocol decode
> listening on ath0, link-type EN10MB (Ethernet), capture size 96 bytes
> 18:01:49.389444 IP 10.27.64.78 > OSPF-ALL.MCAST.NET: OSPFv2, Hello,
> length: 44
> 18:01:59.390520 IP 10.27.64.78 > OSPF-ALL.MCAST.NET: OSPFv2, Hello,
> length: 44
> 18:02:09.391534 IP 10.27.64.78 > OSPF-ALL.MCAST.NET: OSPFv2, Hello,
> length: 44
> 18:02:19.392793 IP 10.27.64.78 > OSPF-ALL.MCAST.NET: OSPFv2, Hello,
> length: 44
> 18:02:29.394546 IP 10.27.64.78 > OSPF-ALL.MCAST.NET: OSPFv2, Hello,
> length: 44
> 
> 
> _______________________________________________
> freebsd-current at freebsd.org mailing list
> http://lists.freebsd.org/mailman/listinfo/freebsd-current
> To unsubscribe, send any mail to "freebsd-current-unsubscribe at freebsd.org"
>