On Sun, Jan 08, 2023 at 10:11:25AM +0100, Mirsad Goran Todorovac wrote:
[root@pc-mtodorov marvin]# tcpdump --immediate-mode -p -v -i veth0 -n dropped privs to tcpdump tcpdump: listening on veth0, link-type EN10MB (Ethernet), capture size 262144 bytes 08:30:22.835825 IP (tos 0x0, ttl 64, id 2490, offset 0, flags [none], proto UDP (17), length 78) 198.18.0.1.35195 > 198.18.0.2.vxlan: VXLAN, flags [I] (0x08), vni 100 ARP, Ethernet (len 6), IPv4 (len 4), Request who-has 198.19.0.2 tell 198.19.0.1, length 28 08:30:22.835926 IP (tos 0x0, ttl 64, id 1388, offset 0, flags [none], proto UDP (17), length 78) 198.18.0.2.35195 > 198.18.0.1.vxlan: VXLAN, flags [I] (0x08), vni 100 ARP, Ethernet (len 6), IPv4 (len 4), Reply 198.19.0.2 is-at a6:45:d5:c4:93:1f, length 28 08:30:22.835976 IP (tos 0xc0, ttl 64, id 29533, offset 0, flags [none], proto ICMP (1), length 106) 198.18.0.1 > 198.18.0.2: ICMP host 198.18.0.1 unreachable - admin prohibited filter, length 86 IP (tos 0x0, ttl 64, id 1388, offset 0, flags [none], proto UDP (17), length 78) 198.18.0.2.35195 > 198.18.0.1.vxlan: VXLAN, flags [I] (0x08), vni 100 ARP, Ethernet (len 6), IPv4 (len 4), Reply 198.19.0.2 is-at a6:45:d5:c4:93:1f, length 28
For some reasons, your host doesn't accept the VXLAN packets received over veth0. I guess there are some firewalling rules incompatible with this tests script.
-------- >8 --------
Isolate testing environment and ensure everything is cleaned up on exit.
diff --git a/tools/testing/selftests/net/l2_tos_ttl_inherit.sh b/tools/testing/selftests/net/l2_tos_ttl_inherit.sh
Wow, Guillaueme, this patch actually made things unstuck :)
Great! The patch isolates the testing environment, making it less dependent from the host that runs it. So the routing and firewalling configurations don't interfere anymore.
The entire tools/tests/selftests/net section now had a PASS w "OK", save for a couple of tests here:
not ok 1 selftests: nci: nci_dev # exit=1 not ok 12 selftests: net: nat6to4.o not ok 13 selftests: net: run_netsocktests # exit=1 not ok 29 selftests: net: udpgro_bench.sh # exit=255 not ok 30 selftests: net: udpgro.sh # exit=255 not ok 37 selftests: net: fcnal-test.sh # TIMEOUT 1500 seconds not ok 38 selftests: net: l2tp.sh # exit=2 not ok 46 selftests: net: icmp_redirect.sh # exit=1 not ok 55 selftests: net: vrf_route_leaking.sh # exit=1 not ok 59 selftests: net: udpgro_fwd.sh # exit=1 not ok 60 selftests: net: udpgro_frglist.sh # exit=255 not ok 61 selftests: net: veth.sh # exit=1 not ok 68 selftests: net: srv6_end_dt46_l3vpn_test.sh # exit=1 not ok 69 selftests: net: srv6_end_dt4_l3vpn_test.sh # exit=1 not ok 75 selftests: net: arp_ndisc_evict_nocarrier.sh # exit=255 not ok 83 selftests: net: test_ingress_egress_chaining.sh # exit=1 not ok 1 selftests: net/hsr: hsr_ping.sh # TIMEOUT 45 seconds not ok 3 selftests: net/mptcp: mptcp_join.sh # exit=1
If you are interested in additional diagnostics, this is a very interesting part of the Linux kernel testing ...
There was apparent hang in selftest/net/fcnal-test.sh as well. I can help you with the diagnostics if you wish? Thanks.
If I could make them all work both on Ubuntu 22.10 kinetic kudu and AlmaLinux 8.7 stone smilodon (CentOS fork), this would be a milestone for me :)
I'm surprised you have so many failures. Feel free to report them individually. Don't forget to Cc the authors of the scripts. Just pay attention not to overwhelm people.
I can probably help with the l2tp.sh failure and maybe with the fcnal-test.sh hang. Please report them in their own mail thread.
Have a nice day!
Regards, Mirsad
-- Mirsad Goran Todorovac Sistem inženjer Grafički fakultet | Akademija likovnih umjetnosti Sveučilište u Zagrebu System engineer Faculty of Graphic Arts | Academy of Fine Arts University of Zagreb, Republic of Croatia The European Union