Broadcast storming problem?

[Andre Oppermann]

Colin Percival wrote:
> 
> [CC: secteam, since this relates to a recent advisory]
> 
>    In http://www.freebsd.org/cgi/query-pr.cgi?pr=kern/64053 a
> problem is reported as having been introduced by the recent
> TCP reassembly patch.
>    Could someone look into this please?

Reading the description of the problem I find it highly unlikely
that the TCP reassembly changes would have caused this.  The TCP
reassembly code only comes into play when a packet belongs to an
established TCP session.  And it doesn't generate any kind of
packet by itself, it only drops them if there are too many.

The problem in this PR must come from something else in IP stack.
I don't have a clear idea what it can be and I'm at conference
till tomorrow and only have slow modem access.

-- 
Andre