Tcpdumping on lagg interface with em driver

[Paul]

I am setting up two new FREEBSD_7 RELEASE routers.
 
So far so good on the setup, everything looks good then I go to do some 
testing.
I have polling enabled with high settings 500 burst 8000 max, etc.   I 
can't check exactly because I managed to nuke the networking in the 
kernel somehow by doing this.....
Testing with random ip packet generator sending 200kpps  directly to the 
ip address of lagg1 interface. say 1.1.1.1 for example, so we have 200k 
packets/second going to 1.1.1.1 on lagg1
I do tcpdump -n -i lagg1 and that's it.. networking code dead.  All 
interfaces are sending weird packets (corrupted,etc)
I did manage one ssh login back to the box for a short period and there 
is no information in dmesg..
All the em interfaces on the machine (there's 8) were not working 
properly.  The way we could ssh in was through the BCE interface but it 
was very shaky and eventually died off after trying to do ifconfig em6 
down em7 down etc.. and back up.

tcpdump -n -i lagg1 works fine without the 200kpps going to it.. Is this 
a known issue or something new?  How can I test this?

On a side note,  what exactly does the lost polls and suspect polls and 
stalled mean?   I see some incrementing slowly and would like some help 
tuning the polling values to rid myself of new increments of the 
lost/suspsect/stalled. But first this tcpdump problem should be 
addressed.   

Thank you!!

Paul