Hi Ed, When creating an adjacency for the first time we send out a probe packet (ARP/ND) to see if the peer will resolve it.
/neale From: Ed Warnicke <hagb...@gmail.com> Date: Friday, 20 August 2021 at 16:54 To: Neale Ranns <ne...@graphiant.com> Cc: zsta...@gmail.com <zsta...@gmail.com>, Artem Glazychev <artem.glazyc...@xored.com>, vpp-dev <vpp-dev@lists.fd.io> Subject: Re: [vpp-dev] Adding new route via ipv6 link-local Neale, Question... looking at the trace it looks like we are having an issue with fib_sas6_get ... could you help me understand why we are calling fib_sas6_get in this situation? Ed On Fri, Aug 20, 2021 at 9:49 AM Neale Ranns <ne...@graphiant.com<mailto:ne...@graphiant.com>> wrote: Hi Ed, It does. But in this case the command is adding a route through an interface, which is not the same as configuring a prefix on an interface (nor ip6 enabling it). /neale From: Ed Warnicke <hagb...@gmail.com<mailto:hagb...@gmail.com>> Date: Friday, 20 August 2021 at 16:40 To: Neale Ranns <ne...@graphiant.com<mailto:ne...@graphiant.com>> Cc: zsta...@gmail.com<mailto:zsta...@gmail.com> <zsta...@gmail.com<mailto:zsta...@gmail.com>>, Artem Glazychev <artem.glazyc...@xored.com<mailto:artem.glazyc...@xored.com>>, vpp-dev <vpp-dev@lists.fd.io<mailto:vpp-dev@lists.fd.io>> Subject: Re: [vpp-dev] Adding new route via ipv6 link-local Neale, I'm a bit confused... I had previously thought that *adding* an IPv6 address to a link *did* IPv6 enable it... Ed On Fri, Aug 20, 2021 at 9:20 AM Neale Ranns <ne...@graphiant.com<mailto:ne...@graphiant.com>> wrote: Hi Artem, Stanislav, I agree graceful handling is needed, return false if there is no link local. If you enabled ip6 on the link first, this wouldn’t be a problem. The link will need to be ip6 enabled if you want something to reply. /neale From: vpp-dev@lists.fd.io<mailto:vpp-dev@lists.fd.io> <vpp-dev@lists.fd.io<mailto:vpp-dev@lists.fd.io>> on behalf of Stanislav Zaikin via lists.fd.io<http://lists.fd.io> <zstaseg=gmail....@lists.fd.io<mailto:gmail....@lists.fd.io>> Date: Friday, 20 August 2021 at 16:10 To: Artem Glazychev <artem.glazyc...@xored.com<mailto:artem.glazyc...@xored.com>> Cc: vpp-dev <vpp-dev@lists.fd.io<mailto:vpp-dev@lists.fd.io>> Subject: Re: [vpp-dev] Adding new route via ipv6 link-local Hello Artem, Looks like we need (at least) to add a check for nullptr. DBGvpp# create tap tap0 DBGvpp# set interface state tap0 up DBGvpp# ip route add fa::1/120 via fe80::1 tap0 Thread 1 "vpp_main" received signal SIGSEGV, Segmentation fault. 0x00007ffff774ba90 in ip6_address_copy (dst=0x7fffaf487aa0, src=0x0) at /home/zstas/vpp/src/vnet/ip/ip6_packet.h:129 129 dst->as_u64[0] = src->as_u64[0]; (gdb) bt #0 0x00007ffff774ba90 in ip6_address_copy (dst=0x7fffaf487aa0, src=0x0) at /home/zstas/vpp/src/vnet/ip/ip6_packet.h:129 #1 0x00007ffff774b892 in fib_sas6_get (sw_if_index=1, dst=0x7fffb91b4e50, src=0x7fffaf487aa0) at /home/zstas/vpp/src/vnet/fib/fib_sas.c:105 On Fri, 20 Aug 2021 at 15:20, Artem Glazychev <artem.glazyc...@xored.com<mailto:artem.glazyc...@xored.com>> wrote: Hello, Found a crash: if we add the route via fe80:: after the interface upping - vpp crashes. For example: DBGvpp# create tap tap0 DBGvpp# set int state tap0 up DBGvpp# ip route add fa::1/120 via fe80::1 tap0 make: *** [Makefile:537: run] Aborted If we swap set ... up and ip route ... - no crashes. Could you please explain me if such a situation is possible at all? In any case, I believe that vpp should not crash. Thanks. -- Best regards Stanislav Zaikin
-=-=-=-=-=-=-=-=-=-=-=- Links: You receive all messages sent to this group. View/Reply Online (#20002): https://lists.fd.io/g/vpp-dev/message/20002 Mute This Topic: https://lists.fd.io/mt/85020362/21656 Group Owner: vpp-dev+ow...@lists.fd.io Unsubscribe: https://lists.fd.io/g/vpp-dev/unsub [arch...@mail-archive.com] -=-=-=-=-=-=-=-=-=-=-=-
