Hi Ilya, Thanks for the clarification. Good to know nc in netdev CI is the nmap version, that makes this much simpler.
For v3 I'll: 1. Replace ncat with nc (same --sctp flag, same output format). Add a which nc + --sctp probe so the test skips cleanly on systems where nc doesn't support SCTP. 2. Add wait $server_pid after each explicit kill to properly reap the background daemon processes (addressing sashiko's comment about orphaned background processes). Will send v3 shortly. Best regards, Minxi Ilya Maximets <[email protected]> 于2026年7月14日周二 08:17写道： > On 7/14/26 2:01 PM, 侯敏熙 wrote: > > Hi Ilya, > > > > Thanks for pointing this out. I looked into this and found that > > nc (netcat-openbsd) on Debian/Ubuntu does not have SCTP support at > > all. The upstream OpenBSD nc only handles TCP, UDP, and > > Unix-domain sockets, and the Debian patches don't add SCTP either. > > AFAICT, 'nc' in netdev CI is the nmap version of netcat, it's not > the openbsd one. Does nmap version support sctp? Google seems to > think so at least. The test should be skipped though if sctp is > not supported. > > Note: sashiko had some comments on the daemon pid handling in v2. > > Best regards, Ilya Maximets. > > > > > I see two cleaner alternatives that already have precedent in the > > kernel selftests: > > > > 1. socat: net/netfilter/nft_queue.sh already uses > > "socat -u SCTP-LISTEN:PORT STDOUT" for SCTP data-plane testing. > > Many drivers/net/ selftests already gate on require_cmd("socat"). > > > > 2. A small C helper: similar to sctp_hello.c used by sctp_vrf.sh > > and conntrack_sctp_collision.sh. The shell test would call > > "./sctp_test server/client IP PORT" instead of ncat. > > > > Which approach would you prefer for v3? > > > > Best regards, > > Minxi > > > > 侯敏熙 <[email protected]> 于2026年7月13日周一 11:02写道： > > > >> Ah so many thanks for your suggestions, Please forgive me, > >> I've always used ncat to write test cases. > >> > >> I will re-spin this patch. > >> > >> Ilya Maximets <[email protected]> 于2026年7月13日周一 10:39写道： > >> > >>> On 7/7/26 5:47 AM, Minxi Hou wrote: > >>>> Register OVS_KEY_ATTR_SCTP in the flow key parser so that sctp() > >>>> can be used in flow specifications. The ovs_key_sctp class already > >>>> exists (with src/dst fields matching the TCP/UDP siblings) but was > >>>> not wired into the parser, so the token was silently dropped and the > >>>> kernel rejected the flow. > >>>> > >>>> Add test_sctp_connect_v4 exercising the SCTP flow key with > >>>> port-specific matching: sctp(dst=4443) for client-to-server and > >>>> sctp(src=4443) for server-to-client. > >>>> > >>>> Wait for ncat readiness with ovs_wait instead of a fixed sleep so > >>>> the test does not race against ncat startup. Use grep -c on the > >>>> listening message to distinguish between successive ncat instances > >>>> that share the same stderr log. Kill the previous server before > >>>> respawning to avoid EADDRINUSE on the SCTP port. > >>>> > >>>> Signed-off-by: Minxi Hou <[email protected]> > >>>> --- > >>>> v1 -> v2: replace sleep with ovs_wait on ncat listening output, > >>>> kill previous ncat server before respawning to avoid > >>>> port conflict (Aaron review feedback) > >>>> > >>>> .../selftests/net/openvswitch/openvswitch.sh | 102 > ++++++++++++++++++ > >>>> .../selftests/net/openvswitch/ovs-dpctl.py | 5 + > >>>> 2 files changed, 107 insertions(+) > >>>> > >>>> diff --git a/tools/testing/selftests/net/openvswitch/openvswitch.sh > >>> b/tools/testing/selftests/net/openvswitch/openvswitch.sh > >>>> index 2954245129a2..9c364eeb2ec2 100755 > >>>> --- a/tools/testing/selftests/net/openvswitch/openvswitch.sh > >>>> +++ b/tools/testing/selftests/net/openvswitch/openvswitch.sh > >>>> @@ -32,6 +32,7 @@ tests=" > >>>> dec_ttl ttl: dec_ttl decrements > >>> IP TTL > >>>> flow_set flow-set: Flow modify > >>>> action_set set: SET action rewrites > >>> fields > >>>> + sctp_connect_v4 sctp: SCTP flow key > >>> matching > >>>> psample psample: Sampling > packets > >>> with psample" > >>>> > >>>> info() { > >>>> @@ -443,6 +444,107 @@ test_action_set() { > >>>> return 0 > >>>> } > >>>> > >>>> +# sctp_connect_v4 test > >>>> +# - sctp(dst=4443) matches client-to-server INIT > >>>> +# - sctp(src=4443) matches server-to-client INIT-ACK > >>>> +# - remove flows and verify connection fails, reinstall and recover > >>>> +test_sctp_connect_v4() { > >>>> + local t="test_sctp_connect_v4" > >>>> + > >>>> + which ncat >/dev/null 2>&1 || return $ksft_skip > >>> > >>> Not a full review, but can we avoid ncat? We already use nc in other > >>> tests, using different implementations of the same thing in different > >>> tests doesn't sound like a good idea. On mnay systems nc and ncat > >>> will be different implementations with different options and behavior. > >>> > >>> Best regards, Ilya Maximets. > >>> > >> > > _______________________________________________ > > dev mailing list > > [email protected] > > https://mail.openvswitch.org/mailman/listinfo/ovs-dev > > > > _______________________________________________ dev mailing list [email protected] https://mail.openvswitch.org/mailman/listinfo/ovs-dev
