The linux kernel IP stack won't deliver multicast packets unless you open the appropriate listener socket. Obviously, without looking at your application code I have no idea if that's your problem, but from the sounds of it that is the most likely issue. Nothing you've said sounds like a tcpreplay bug to me. -- Aaron Turner https://synfin.net/ Twitter: @synfinatic Those who would give up essential Liberty, to purchase a little temporary Safety, deserve neither Liberty nor Safety. -- Benjamin Franklin
On Fri, Jul 8, 2016 at 1:53 PM, James deCocq <dotmar...@hotmail.com> wrote: > Hi all, > > NICETIES > tcpreplay version: 4.1.0; CentOS 6.7; x86_64 > Multiple Solarflare SFC9120 > > BACKGROUND: > Our application can be built to use one of: > * kernel sockets > * OpenOnload (bespoke network stack, kernel bypass) > * ef_vi (direct to/from VNICs, no stack) > > PROBLEM: > * We test using tcpreplay of captured live multicast packets. > * Replayed packets go out eth1 to the switch and back to eth2. > * When application is built to use onload or ef_vi, it works as expected. > * However when built for kernel stack, replayed packets arrive at eth2 > as expected, but are subsequently dropped by kernel. > > DETAIL: > * Multicast drops appear to originate from 2 spots in the linux stack: > most from ip_rcv(), and some in tpacket_rcv() (the AF_PACKET stuff) > * Captured live data is massaged for playback thusly, addresses redacted: > tcprewrite -C --srcipmap=nnn.nnn.nnn.nnn/26:nnn.nnn.nnn.nnn > --dstipmap=nnn.nnn.nnn.nnn/24:nnn.nnn.nnn.nnn/24 > --enet-smac=00:0F:53:xx:yy:zz --enet-vlan=del -i in.pcap -o out.pcap > * All mods and IP checksums look good in Wireshark, and again, > the application receives expected flow if the kernel stack is not involved. > > QUESTION: > * Of course I noted the caveat about replay on the same box; however > I was not certain that the caveat applied to our use case, > tcpreplay -> eth1 -> switch -> eth2 ->kernel; > however drops from kernel's tpacket_rcv() would seem to suggest yes? > * So is tcpreplay running concurrently with our application likely to be > the cause of the kernel dropping multicast packets in this situation? > > Thanks very much for any assistance you can provide. > > james > > ------------------------------------------------------------------------------ > Attend Shape: An AT&T Tech Expo July 15-16. Meet us at AT&T Park in San > Francisco, CA to explore cutting-edge tech and listen to tech luminaries > present their vision of the future. This family event has something for > everyone, including kids. Get more information and register today. > http://sdm.link/attshape > _______________________________________________ > Tcpreplay-users mailing list > Tcpreplay-users@lists.sourceforge.net > https://lists.sourceforge.net/lists/listinfo/tcpreplay-users > Support Information: http://tcpreplay.synfin.net/trac/wiki/Support ------------------------------------------------------------------------------ Attend Shape: An AT&T Tech Expo July 15-16. Meet us at AT&T Park in San Francisco, CA to explore cutting-edge tech and listen to tech luminaries present their vision of the future. This family event has something for everyone, including kids. Get more information and register today. http://sdm.link/attshape _______________________________________________ Tcpreplay-users mailing list Tcpreplay-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/tcpreplay-users Support Information: http://tcpreplay.synfin.net/trac/wiki/Support
