problem using ng_patch

[Nikolay Denev]

On Aug 11, 2012, at 11:07 AM, h bagade <bagadeh at gmail.com> wrote:

> Hi all,
> 
> I want to use the node ng_patch, to set the ToS field of special class of
> packets. I try to test the function by a simple test scenario and
> encountered problem using it. I have no idea why the problem occurs.
> 
> Here I explain the test scenario I've used.
> 
> I have a topology like this:
> 
> |A:192.168.8.8|<---->|192.168.8.26--(B)--192.168.7.26|<---->|C:192.168.7.20|
> --------------------------------
> A, C: two end stations
> B: a router
> --------------------------------
> netgraph settings:
> kldload ng_ipfw
> ngctl mkpeer ipfw: patch 300 in
> ngctl name ipfw:300 tos
> ngctl msg tos: setconfig {count=1 csum_flags=1 ops=[ {mode=1 value=0x05
> length=1 offset=1}]}
> --------------------------------
> ipfw rule:
> ipfw add 20 netgraph 300 icmp from any to 192.168.7.20
> 
> This configuration works well and when A pings C or C pings A, the packets
> destined to 192.168.7.20(station C) gets the ToS: 0x05.
> The problem occurs when I change the ipfw rule to the following;
> 
> ipfw add 20 netgraph 300 icmp from 192.168.7.20 to any
> 
> By this rule, neither A can ping C nor C can ping A! the packets sent to
> ng_patch node never comes back to the next ipfw rule!
> 
> I don't know what's the difference between these two scenarios (only the
> checking from destination address is changed to source address), but it's
> what I saw in my tests. I really hope to understand what's happening.
> 
> Any hints or comments would help
> _______________________________________________
> freebsd-net at freebsd.org mailing list
> http://lists.freebsd.org/mailman/listinfo/freebsd-net
> To unsubscribe, send any mail to "freebsd-net-unsubscribe at freebsd.org"

Hi,

Do you have "sysctl net.inet.ip.fw.one_pass=0" set?

Regards,