Re: [PATCH v1 01/11] selftests: forwarding: custom_multipath_hash.sh: add cleanup for SIGTERM sent by timeout
From: Mirsad Todorovac
Date: Sun Jul 30 2023 - 12:48:24 EST
On 7/30/23 09:53, Ido Schimmel wrote:
On Thu, Jul 27, 2023 at 09:26:03PM +0200, Mirsad Todorovac wrote:
marvin@defiant:~/linux/kernel/linux_torvalds$ grep "not ok" ../kselftest-6.5-rc3-net-forwarding-16.log
not ok 3 selftests: net/forwarding: bridge_mdb.sh # exit=1
Other than one test case (see below), I believe this should be fixed by
the patches I just pushed to the existing branch. My earlier fix was
incomplete which is why it didn't solve the problem.
not ok 5 selftests: net/forwarding: bridge_mdb_max.sh # exit=1
Should be fixed with the patches.
Congratulations! Indeed, it looks a lot better:
marvin@defiant:~/linux/kernel/linux_torvalds$ grep "not ok" ../kselftest-6.5-rc3-net-forwarding-18.log
not ok 3 selftests: net/forwarding: bridge_mdb.sh # exit=1
not ok 11 selftests: net/forwarding: bridge_vlan_mcast.sh # exit=1
not ok 26 selftests: net/forwarding: ip6_forward_instats_vrf.sh # exit=1
not ok 49 selftests: net/forwarding: mirror_gre_changes.sh # exit=1
marvin@defiant:~/linux/kernel/linux_torvalds$ grep -v '^# +' ../kselftest-6.5-rc3-net-forwarding-18.log | grep -A1 -e '\[FAIL\]' | grep -v -e -- | grep -v OK
# TEST: IPv4 (S, G) port group entries configuration tests [FAIL]
# Entry has an unpending group timer after replace
# TEST: IPv6 (S, G) port group entries configuration tests [FAIL]
# Entry has an unpending group timer after replace
# TEST: Vlan mcast_startup_query_interval global option default value [FAIL]
# Wrong default mcast_startup_query_interval global vlan option value
# TEST: Ip6InHdrErrors [FAIL]
# TEST: mirror to gretap: TTL change (skip_hw) [FAIL]
# Expected to capture 10 packets, got 15.
# TEST: mirror to ip6gretap: TTL change (skip_hw) [FAIL]
# Expected to capture 10 packets, got 13.
marvin@defiant:~/linux/kernel/linux_torvalds$
not ok 11 selftests: net/forwarding: bridge_vlan_mcast.sh # exit=1
Nik, the relevant failure is this one:
# TEST: Vlan mcast_startup_query_interval global option default value [FAIL]
# Wrong default mcast_startup_query_interval global vlan option value
Any idea why the kernel will report "mcast_startup_query_interval" as
3124 instead of 3125?
# + jq -e '.[].vlans[] | select(.vlan == 10 and .mcast_startup_query_interval == 3125) '
# + echo -n '[{"ifname":"br0","vlans":[{"vlan":1,"mcast_snooping":1,"mcast_querier":0,"mcast_igmp_version":2,"mcast_mld_version":1,"mcast_last_member_count":2,"mcast_last_member_interval":100,"mcast_startup_query_count":2,"mcast_startup_query_interval":3124,"mcast_membership_interval":26000,"mcast_querier_interval":25500,"mcast_query_interval":12500,"mcast_query_response_interval":1000},{"vlan":10,"vlanEnd":11,"mcast_snooping":1,"mcast_querier":0,"mcast_igmp_version":2,"mcast_mld_version":1,"mcast_last_member_count":2,"mcast_last_member_interval":100,"mcast_startup_query_count":2,"mcast_startup_query_interval":3124,"mcast_membership_interval":26000,"mcast_querier_interval":25500,"mcast_query_interval":12500,"mcast_query_response_interval":1000}]}]'
# + check_err 4 'Wrong default mcast_startup_query_interval global vlan option value'
not ok 26 selftests: net/forwarding: ip6_forward_instats_vrf.sh # exit=1
Please run this one with +x so that we will get more info.
In fact, I have turned it on on all the remaining failing tests.
In case you want to investigate further, please find the debug output log
at the usual place:
https://domac.alu.unizg.hr/~mtodorov/linux/selftests/net-forwarding/kselftest-6.5-rc3-net-forwarding-18.log.xz
https://domac.alu.unizg.hr/~mtodorov/linux/selftests/net-forwarding/bridge_mdb.sh.out.xz
https://domac.alu.unizg.hr/~mtodorov/linux/selftests/net-forwarding/bridge_vlan_mcast.sh.out.xz
https://domac.alu.unizg.hr/~mtodorov/linux/selftests/net-forwarding/ip6_forward_instats_vrf.sh.out.xz
https://domac.alu.unizg.hr/~mtodorov/linux/selftests/net-forwarding/mirror_gre_changes.sh.out.xz
I hope this helps, because you drastically reduced the number of [FAIL] results.
If it matters being heard from me, I think it's a great job!
Kind regards,
Mirsad
not ok 49 selftests: net/forwarding: mirror_gre_changes.sh # exit=1
Petr, please take a look. Probably need to make the filters more
specific. The failure is:
# TEST: mirror to gretap: TTL change (skip_hw) [FAIL]
# Expected to capture 10 packets, got 14.
not ok 84 selftests: net/forwarding: tc_flower_l2_miss.sh # exit=1
Should be fixed with the patches.
marvin@defiant:~/linux/kernel/linux_torvalds$ grep -v "^# +" ../kselftest-6.5-rc3-net-forwarding-16.log | grep -A1 FAIL | grep -v -e -- | grep -v OK
# TEST: IPv6 (S, G) port group entries configuration tests [FAIL]
# "temp" entry has an unpending group timer
Not sure about this one. What is the output with the following diff?
diff --git a/tools/testing/selftests/net/forwarding/bridge_mdb.sh b/tools/testing/selftests/net/forwarding/bridge_mdb.sh
index 8493c3dfc01e..2b2a3b150861 100755
--- a/tools/testing/selftests/net/forwarding/bridge_mdb.sh
+++ b/tools/testing/selftests/net/forwarding/bridge_mdb.sh
@@ -628,6 +628,7 @@ __cfg_test_port_ip_sg()
bridge -d -s mdb show dev br0 vid 10 | grep "$grp_key" | \
grep -q "0.00"
check_fail $? "\"temp\" entry has an unpending group timer"
+ bridge -d -s mdb show dev br0 vid 10 | grep "$grp_key"
bridge mdb del dev br0 port $swp1 $grp_key vid 10
# Check error cases.
# TEST: IPv4 host entries forwarding tests [FAIL]
# Packet not locally received after adding a host entry
# TEST: IPv4 port group "exclude" entries forwarding tests [FAIL]
# Packet from valid source not received on H2 after adding entry
# TEST: IPv4 port group "include" entries forwarding tests [FAIL]
# Packet from valid source not received on H2 after adding entry
# TEST: IGMPv3 MODE_IS_INCLUDE tests [FAIL]
# Source not add to source list
# TEST: ctl4: port: ngroups reporting [FAIL]
# Couldn't add MDB entries
# TEST: ctl4: port maxgroups: reporting and treatment of 0 [FAIL]
# Adding 5 MDB entries failed but should have passed
# TEST: ctl4: port maxgroups: configure below ngroups [FAIL]
# dev veth1: Couldn't add MDB entries
# TEST: ctl4: port: ngroups reporting [FAIL]
# Couldn't add MDB entries
# TEST: ctl4: port maxgroups: reporting and treatment of 0 [FAIL]
# Adding 5 MDB entries failed but should have passed
# TEST: ctl4: port maxgroups: configure below ngroups [FAIL]
# dev veth1 vid 10: Couldn't add MDB entries
# TEST: ctl4: port_vlan: ngroups reporting [FAIL]
# Couldn't add MDB entries
# TEST: ctl4: port_vlan: isolation of port and per-VLAN ngroups [FAIL]
# Couldn't add MDB entries to VLAN 10
# TEST: ctl4: port_vlan maxgroups: reporting and treatment of 0 [FAIL]
# Adding 5 MDB entries failed but should have passed
# TEST: ctl4: port_vlan maxgroups: configure below ngroups [FAIL]
# dev veth1 vid 10: Couldn't add MDB entries
# TEST: ctl4: port_vlan maxgroups: isolation of port and per-VLAN ngroups [FAIL]
# Couldn't add 5 entries
# TEST: Vlan mcast_startup_query_interval global option default value [FAIL]
# Wrong default mcast_startup_query_interval global vlan option value
# TEST: Ip6InHdrErrors [FAIL]
# TEST: mirror to gretap: TTL change (skip_hw) [FAIL]
# Expected to capture 10 packets, got 14.
# TEST: L2 miss - Multicast (IPv4) [FAIL]
# Unregistered multicast filter was not hit before adding MDB entry
marvin@defiant:~/linux/kernel/linux_torvalds$
In case you want to pursue these failures, there is the complete test output log
here:
https://domac.alu.unizg.hr/~mtodorov/linux/selftests/net-forwarding/kselftest-6.5-rc3-net-forwarding-16.log.xz
Thanks again, great work!
Kind regards,
Mirsad