On Jul 03, 2013, Blair Zajac wrote:

> On 7/3/13 8:23 PM, Michael Rash wrote:
> > On Jul 03, 2013, Blair Zajac wrote:
> >> On 6/11/13 2:35 PM, Michael Rash wrote:
> >>> On Jun 11, 2013, Blair Zajac wrote:
> >>>> On 6/10/13 8:12 PM, Michael Rash wrote:
> >>> Now, the above assumes that the test suite is working properly to begin
> >>> with on your system, and the previous output you sent makes me suspect
> >>> that you may need to adjust your local ipfw policy to accept traffic
> >>> over the loopback interface since fwknopd didn't appear to receive any
> >>> SPA traffic.
> >>
> >> Yup, appears like that.  What do you suggest to test this correctly?
> >
> > Perhaps adding a rule like "ipfw 00010 allow all from any to any via lo0"
> > would work?  You can test with tcpdump and netcat - the following
> > scenario works, then the fwknop test suite should also work I think:
> 
> This doesn't work:
> 
> # ipfw 00010 allow all from any to any via lo0
> ipfw: bad command `allow'

Bleh, sorry, was quoting the wrong ipfw syntax (list output vs. new
rule).  Although I don't have a system where I can test this, it should
be something like "ipfw add 00010 set 1 pass udp from any to me dst-port 62201"
and then make sure packets to set 1 can be seen via loopback (I think).

> > # tcpdump -i lo0 -l -nn -s 0 -X port 62201
> > # echo "test" | nc -u 127.0.0.1 62201
> 
> This works with no ipfw changes:
> 
> # tcpdump -i lo0 -l -nn -s 0 -X port 62201
> tcpdump: verbose output suppressed, use -v or -vv for full protocol decode
> listening on lo0, link-type NULL (BSD loopback), capture size 65535 bytes
> 20:54:42.886619 IP 127.0.0.1.64601 > 127.0.0.1.62201: UDP, length 5
>       0x0000:  4500 0021 1bb4 0000 4011 0000 7f00 0001  E..!....@.......
>       0x0010:  7f00 0001 fc59 f2f9 000d fe20 7465 7374  .....Y......test
>       0x0020:  0a                                       .
> ^C
> 1 packets captured
> 2 packets received by filter
> 0 packets dropped by kernel

Ok, I wonder if there is a link level issue here in fwknopd related to the
BSD equivalent of the Linux 'cooked' interface or something like that.
I need to get a Mac.  :)

> > If you see the packet in the tcpdump output, then ipfw is not in the
> > way, and hence fwknop traffic should also be seen by fwknopd over
> > loopback.
> 
> BTW, if you want to punt on PPC support, that's fine with me.  I have a 
> patched 2.0.4 that works.  I'm hoping to get the next generation of 
> MacBook Pro, whenever that is, so the x86_86 version should work.

Ok, I'd like to get it working, but it may need to wait until 2.5.1.

Thanks,

--Mike

> Blair

------------------------------------------------------------------------------
This SF.net email is sponsored by Windows:

Build for Windows Store.

http://p.sf.net/sfu/windows-dev2dev
_______________________________________________
Fwknop-discuss mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/fwknop-discuss

Reply via email to