polling probrem?

[Katsuji Ishikawa]

I have rebuid the FreeBSD 5.3-RELEASE kernel with options below:

options         DEVICE_POLLING
options         HZ=1000

on ThinkPad iSeries 1800.

And setting 'sysctl kern.polling.enable=1', then
running dnetc installed from ports(dnetc-2.9008.491_1,1), I detected
considerable amount of packet loss.
Top command shows that dnetc is using 1000% (not 100%, but 1000%!).

ping result
(snip)
64 bytes from 192.168.2.200: icmp_seq=258 ttl=64 time=9597.83 ms
64 bytes from 192.168.2.200: icmp_seq=268 ttl=64 time=9624.56 ms
64 bytes from 192.168.2.200: icmp_seq=278 ttl=64 time=9612.66 ms
64 bytes from 192.168.2.200: icmp_seq=288 ttl=64 time=9639.19 ms
64 bytes from 192.168.2.200: icmp_seq=298 ttl=64 time=19652 ms
64 bytes from 192.168.2.200: icmp_seq=318 ttl=64 time=9579.19 ms
64 bytes from 192.168.2.200: icmp_seq=328 ttl=64 time=9606.89 ms
64 bytes from 192.168.2.200: icmp_seq=329 ttl=64 time=8606.83 ms
64 bytes from 192.168.2.200: icmp_seq=338 ttl=64 time=19564.4 ms
64 bytes from 192.168.2.200: icmp_seq=340 ttl=64 time=17564.2 ms
64 bytes from 192.168.2.200: icmp_seq=358 ttl=64 time=9547.07 ms
64 bytes from 192.168.2.200: icmp_seq=361 ttl=64 time=6546.53 ms
^C
--- 192.168.2.200 ping statistics ---
372 packets transmitted, 201 packets received, 45% packet loss
round-trip min/avg/max = 0.367/1158.48/19652 ms 

top result
 PID USERNAME PRI NICE SIZE RES STATE TIME WCPU CPU COMMAND
434 dnetc 139 20 964K 764K RUN 11:09 1259.46% 1259.42% dnetc 

Is this the problem with polling?
Please let me know.

Thank you.