Good question, version 4 changed how transparent_mode works and I'm also confused after reading these bullet points. So I second this question.
On Wed, Mar 3, 2010 at 4:33 AM, Mark H <[email protected]> wrote: > Hi, > > I am working with pf_ring for the first time. I currently have the ring > installed and am using transparent_mode=1. Running code similar to the > pktgen tests in a loopback between eth1 and eth2 assigned to the 2 ports of > the Intel 82576. > > Repeated tests show on average ~920 Mbit/sec in both pfcount and cat > /proc/net/pktgen/eth1 dump. > > I presently do not have TNAPI but am considering using it if it gives > better throughput. > > - 'insmod pf_ring.ko transparent_mode=1': PF_RING-aware driver copies > the packets into PF_RING, while the same packet is still passed to TNAPI > - 'insmod pf_ring.ko transparent_mode=2': PF_RING-aware driver copies > the packets into PF_RING, no the packet still is passed to TNAPI > > > I do not understand what the bullet points above from the NTop PF_RING page > are saying. I understood that TNAPI is the driver that passes to the ring. > So if I don't have TNAPI and am in mode 1 what is being passed to where? > And if I do have TNAPI in mode 2 and TNAPI is not getting passed the packet > then what is it doing? > > Can anyone explain simply what the difference is between the 2 modes? > > Also, is there an updated API doc? The version that came with the 4.1.3 > drop is for 3.1 and it has missing argument definitions for the > initialization call per the pfcount example code. > > Thanks > > Mark > > _______________________________________________ > 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
