Netgraph ng_patch and ng_input: where to find packets?

[Victor Gamov]

ipfw allow log udp from 192.168.230.9 to 192.168.230.128 dst-port 1234

this rule added to ipfw before ngtee action and I see patched packets at ipfw now -- its marked as received via vlan999 still.  Yes, it's OK.

Also, I make 3 actions at ng_patch now:
set TTL=3
set src_ip=192.168.230.9 (vlan333)
set dst_ip=192.168.230.128 now.

But packets still does not exists on vlan333 as outgoing.

Any suggestions?

Is it possible patched packets silently drops by kernel ?

On 26Nov, 2013, at 13:44, Victor Gamov wrote:

> 
> On 26Nov, 2013, at 03:57, Julian Elischer wrote:
> 
>> On 11/24/13, 5:05 AM, Victor Gamov wrote:
>>> Hi All
>>> 
>>> I want to get 2 or 3 copies of input packet at my system to resend it to new destinations.  So I prepare following configuration:
>>> 
>>> # ipfw add 10000 ngtee 100 udp from any to 239.0.0.19 dst-port 1234 in via vlan999
>>> 
>>> # ngctl mkpeer ipfw: hub 100 hub-in
>>> # ngctl name ipfw:100 hub100
>>> 
>>> # ngctl mkpeer hub100: patch hub100-out1 in
>>> # ngctl name hub100:hub100-out1 patch100
>>> # ngctl msg patch100: setconfig '{ count=1 csum_flags=1 ops=[ { value=0xc0a8e680 offset=16 length=4 mode=1 } ] }'
>>> 
>>> Now when I connect to patch:out as
>>> # nghook -a patch100: out
>>> 
>>> then I see packets with new IP:
>>> 
>>> 0000:  45 00 05 40 00 00 40 00 ff 11 b9 27 c0 a8 0d 12
>>> 0010:  c0 a8 e6 80 04 dc 04 dc 05 2c 00 00 47 4c ef 1a
>>> 
>>> Now I want to put this packets back into IP processing to send it to new destination 192.168.230.128 (0xc0a8e680):
>>> 
>>> # ngctl mkpeer patch100: ip_input out new100_to_dst_1
>>> 
>>> But packets not shown on outgoing interface:
>>> 
>>> # ifconfig vlan333
>>> vlan333: flags=8843<UP,BROADCAST,RUNNING,SIMPLEX,MULTICAST> metric 0 mtu 1500
>>> 	options=103<RXCSUM,TXCSUM,TSO4>
>>> 	ether 00:1b:21:5b:7e:e9
>>> 	inet 192.168.230.9 netmask 0xffffff00 broadcast 192.168.230.255
>>> 
>>> # arp 192.168.230.128
>>> ? (192.168.230.128) at 62:99:4c:3b:22:fc on vlan333 expires in 1190 seconds
>> I would looking at giving the packet back to the firewall as suggested..
>> 
>> netgraph cookie
>>            Divert packet into netgraph with given cookie.  The search termi-
>>            nates.  If packet is later returned from netgraph it is either
>>            accepted or continues with the next rule, depending on
>>            net.inet.ip.fw.one_pass sysctl variable.
>> see ng_ipfw for more details..
> 
> Yes I read this manuals :-)   But I still can't see packets neither at ipfw nor at outgoing interface.
> 
> net.inet.ip.fw.one_pass: 0
> net.inet.ip.forwarding: 1
> 
> Is my original idea is correct?

--
CU,
Victor Gamov