Hey Daniel,

Thanks for your reply.  I did compile your example

bpfc foo > bar

Which did return
cat bar
{ 0x20, 0, 0, 0xfffff034 },
{ 0x16, 0, 0, 0x00000000 },

I then tried to use it, 

netsniff-ng --dev eth0 -f bar
Running! Hang up with ^C!

           0  packets incoming (0 unread on exit)
           0  packets passed filter
           0  packets failed filter (out of space)
           5  sec, 805074 usec in total

Nothing seems to be passing that filer.  You say, I need a recent kernel, I 
am on Fedora, kernel = 3.5.7+.  I assume I am on an OK kernel.  Perhaps 
there is something silly I am missing.

Thanks,

Dan

On Wednesday, September 4, 2013 1:34:22 PM UTC-4, Daniel Borkmann wrote:
>
> On 09/03/2013 10:23 PM, Daniel Martin wrote: 
> > Hello again, 
> > 
> > I am working on 10Gbps systems.  I am conducting tests that push line 
> rate 
> > speeds and I am simply trying to come up with a way to basically prove 
> the 
> > data is being received at line rate speeds.  From previous posts I 
> realize 
> > I am fairly sure my sniffer machines will not be able to handle 
> capturing 
> > all data "I have configured the systems just about every way I know how 
> > within reason" and I am still getting about 88% packet loss.  I will 
> need 
> > to purchase additional parts to make this work.  However, I am now 
> thinking 
> > since all packets are of uniform size about 24,000 Bytes UDP.  Would it 
> be 
> > practical to try and write a filter that captures the packet, but only 
> > writes the IP Header information to a file and not all the data inside 
> the 
> > packet?  I am thinking that since most of the data being written is the 
> > packet payload maybe I might have better luck simply writing the IP 
> Header 
> > Information, or is this not practical?  Thanks for any advice you can 
> give 
> > me in advance. 
>
> If only headers are sufficient for you, you could try out the following 
> bpfc 
> filter: 
>
> # cat foo 
> ld poff 
> ret a 
> # bpfc foo > bar 
> # netsniff-ng ... -f bar ... 
>
> That should already improve capturing. Note that you need a recent kernel 
> to 
> use this BPF extension. 
>

-- 
You received this message because you are subscribed to the Google Groups 
"netsniff-ng" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to [email protected].
For more options, visit https://groups.google.com/groups/opt_out.

Reply via email to