Hi Luca,

OK, now it's clear.

---
2 Days ago i sent you mail directly to your address, probably it landed
in spam...
can you check it please?

i asked you for copy of recent ixgbe TNAPI driver for tests we want to
start today at university.

Best Regards,
Alexander


On Thu, 2010-02-18 at 09:13 +0100, Luca Deri wrote:
> Alex
> I mean the userland poll(). If you app has not much to do it has to
> call it (unless you want to active wait) hence the performance gets
> bad.
> 
> if you want to create two rings you need to do
> 
> tcpdump -i e...@0
> and 
> tcpdump -i e...@1
> 
> Luca
> On 02/17/2010 07:11 PM, Alexander Didebulidze wrote: 
> > Hi Luca,
> > 
> > On Wed, 2010-02-17 at 18:10 +0100, Luca Deri wrote:
> >   
> > > Hi Alexander
> > > 
> > > On Feb 17, 2010, at 3:03 PM, Alexander Didebulidze wrote:
> > > 
> > >     
> > > > Hi Luca,
> > > > 
> > > > if I load igb TNAPI driver with parameter RSS=2,2 i get 2 Kernel tnapi
> > > > Threads...
> > > > 
> > > > If i capture from eth0 (without @0 and @1) pf_ring uses only one
> > > > RING-Buffer...
> > > > 
> > > > in /proc/net/pf_ring/info i see: 
> > > > PF_RING Version     : 4.1.0 ($Revision: 4012 $)
> > > > Ring slots          : 4096
> > > > Slot version        : 10
> > > > Capture TX          : Yes [RX+TX]
> > > > IP Defragment       : No
> > > > Transparent mode    : Yes
> > > > Total rings         : 1  <---- ???
> > > > Total plugins       : 0
> > > > 
> > > > would same number of rings as RX-queues improve capturing performance?
> > > >       
> > > DEpends if you have traffic to handle. poll() is very costly so better to 
> > > have something to do that rest/work/rest.... forever
> > >     
> > Which poll() do you mean? do you mean tnapi thread inside driver? I
> > thought that this thread is not polling the NIC. as i can see it sleeps
> > till new packet arrives.
> > (wait_event_interruptible(adapter->tnapi.packet_waitqueue[r_idx]....)
> > 
> > I think I'm also not using ring-polling in user-space apps.(no -a flag
> > etc)
> > 
> >   
> > > > is it possible to use 2 or more rings from unmodified PCAP-based
> > > > capturing applications when using pcap+pfring?
> > > >       
> > > Yes. You can bind a ring per queue
> > >     
> > I don't want to start 2 tcpdumps with -i e...@0 and e...@1
> > I want to use 2 rings with SINGLE tcpdump which is linked against
> > pf_ring capable PCAP. is this also possible?
> > 
> > "tcpdump -i eth0" creates only one ring and i don't see the way to
> > create 2 rings. 
> > 
> > 
> > I also found a bug in PF_RING Revision: 4012 in combination with TNAPI.
> > if i load pf_ring.ko with parameter enable_tx_capture=0 i get no
> > packets:
> > 
> > Tot Packets    : 0
> > Tot Pkt Lost   : 0
> > Tot Insert     : 0
> > Tot Read       : 0
> > Tot Fwd Ok     : 0
> > Tot Fwd Errors : 0
> > 
> > with enable_tx_capture=1 i can receive incoming packets again.
> > since i don't have any outgoing traffic on this interface this parameter
> > is not important for me.
> > (i  don't know if svn trunk also has this problem)
> > 
> > Best Regards,
> > Alexander
> >  
> > P.S. did you get my mail which i sent to you Yesterday? Subject was
> > "TNAPI drivers".  I hope that we can get new driver for tests on 10GE
> > which we want to start tomorrow morning.
> > 
> > 
> > _______________________________________________
> > Ntop-misc mailing list
> > [email protected]
> > http://listgateway.unipi.it/mailman/listinfo/ntop-misc
> >   
> 
> _______________________________________________
> Ntop-misc mailing list
> [email protected]
> http://listgateway.unipi.it/mailman/listinfo/ntop-misc


_______________________________________________
Ntop-misc mailing list
[email protected]
http://listgateway.unipi.it/mailman/listinfo/ntop-misc

Reply via email to