Lenovo X60 em

[Jon Otterholm]

Jack Vogel wrote:
> On 1/16/07, Jon Otterholm <jon.otterholm at ide.resurscentrum.se> wrote:
>> Hi.
>>
>> I have trouble with high latency on my new X60 with em-interface. Anyone
>> else with the same problem? I´m running 6.2-RELEASE.
>
> Would you please give a bit more detail.
>
> Jack
Here comes some info:

uname -a
FreeBSD onob2.domain.local 6.2-RELEASE FreeBSD 6.2-RELEASE #0: Fri Jan
12 11:05:30 UTC 2007
root at dessler.cse.buffalo.edu:/usr/obj/usr/src/sys/SMP  i386

$ ifconfig em0
em0: flags=8843<UP,BROADCAST,RUNNING,SIMPLEX,MULTICAST> mtu 1500
        options=b<RXCSUM,TXCSUM,VLAN_MTU>
        inet 192.168.98.98 netmask 0xffffff00 broadcast 192.168.98.255
        ether 00:16:d3:35:0a:a9
        media: Ethernet autoselect (100baseTX <full-duplex>)
        status: active

dmesg
em0: <Intel(R) PRO/1000 Network Connection Version - 6.2.9> port
0x2000-0x201f mem 0xee000000-0xee01ffff irq 16 at device 0.0 on pci2
em0: Ethernet address: 00:16:d3:35:0a:a9

pciconf -l
em0 at pci2:0:0:   class=0x020000 card=0x207e17aa chip=0x109a8086 rev=0x00
hdr=0x00

I think I found a pattern to work with. If I do a echo-reply wit the
"-D" (no fragment) and increase the packet size (-s) to 1472 I get
normal response times:

onob2# ping -D -s 1470 www.sunet.se
PING www.sunet.se (192.36.125.18): 1470 data bytes
1478 bytes from 192.36.125.18: icmp_seq=0 ttl=57 time=47.770 ms
1478 bytes from 192.36.125.18: icmp_seq=1 ttl=57 time=46.128 ms
1478 bytes from 192.36.125.18: icmp_seq=2 ttl=57 time=47.038 ms
1478 bytes from 192.36.125.18: icmp_seq=3 ttl=57 time=46.457 ms
1478 bytes from 192.36.125.18: icmp_seq=4 ttl=57 time=46.375 ms
1478 bytes from 192.36.125.18: icmp_seq=5 ttl=57 time=46.291 ms
1478 bytes from 192.36.125.18: icmp_seq=6 ttl=57 time=46.707 ms
1478 bytes from 192.36.125.18: icmp_seq=7 ttl=57 time=46.623 ms
1478 bytes from 192.36.125.18: icmp_seq=8 ttl=57 time=47.541 ms
1478 bytes from 192.36.125.18: icmp_seq=9 ttl=57 time=46.458 ms

But when decreasing packet size:

onob2# ping -D -s 64 www.sunet.se
PING www.sunet.se (192.36.125.18): 64 data bytes
72 bytes from 192.36.125.18: icmp_seq=0 ttl=57 time=102.082 ms
72 bytes from 192.36.125.18: icmp_seq=1 ttl=57 time=897.502 ms
72 bytes from 192.36.125.18: icmp_seq=2 ttl=57 time=896.420 ms
72 bytes from 192.36.125.18: icmp_seq=3 ttl=57 time=27.858 ms
72 bytes from 192.36.125.18: icmp_seq=4 ttl=57 time=894.255 ms
72 bytes from 192.36.125.18: icmp_seq=5 ttl=57 time=893.245 ms
72 bytes from 192.36.125.18: icmp_seq=6 ttl=57 time=892.244 ms
72 bytes from 192.36.125.18: icmp_seq=7 ttl=57 time=891.502 ms
72 bytes from 192.36.125.18: icmp_seq=8 ttl=57 time=890.421 ms
72 bytes from 192.36.125.18: icmp_seq=9 ttl=57 time=785.403 ms

After some time (~1h) all ICMP seems to work, but when using tcp it
doesn't work. For example I tried a "pkg_add -r bash2" and it timed out.
I haven't had time to make any dumps on traffic, I would appreciate some
hints on how to approach this to get some useful output.

BTW: "netstat -i" shows no i/o-error.

//Jon