Daniel Hartmeier <[EMAIL PROTECTED]> writes:
[...]

>
> Sounds like the source/destination address/port pair is reused before
> the state is removed. For instance, if the first connections uses
> 10.1.2.3:789 <-> 10.2.3.4:514, the client is not allowed to reuse source
> port 789 before 2MSL (about 120 seconds) for another connection to
> 10.2.3.4:514. If it does, it violates TCP, and confuses pf.

Yes that is the problem I get BAD State with the misc debug. How come
this violates TCP?

-----
Dec 17 09:45:02 fw /bsd: pf: BAD state: TCP 10.32.254.13:514 10.32.254.13:514 
10.32.10.3:1023 [lo=4091004600 high=4091200080 win=63712 modulator=0 wscale=0] 
[lo=4088272098 high=4088335810 win=48870 modulator=0 wscale=2] 9:9 S seq=4111446242 
ack=4088272098 len=0 ackskew=0 pkts=4313:2382 dir=in,fwd
Dec 17 09:45:02 fw /bsd: pf: State failure on: 1       | 5  
-----
[...]

Is there any solution in order to keep state and manage to forward
packet for that crappy rsh implementation (linux is not ok, sgi is ok).

>From my point of view i don't see any apart from disabling the
stateful inspection.

Thanks for you help

-- 
Lo�c

Reply via email to