My Ack still stands, but I wanted to point something out in-line On Mon, Aug 3, 2026 at 2:29 PM Mark Michelson <[email protected]> wrote: > > Thanks Ales, > > Acked-by: Mark Michelson <[email protected]> > > On Mon, Aug 3, 2026 at 3:13 AM Ales Musil via dev > <[email protected]> wrote: > > > > Bump the submodule to include: > > ea7f2165889f ("packets: Add support for unicast ND NS compose.") > > > > at the same time this includes 9b63c2ead11e ("ofproto-dpif: Avoid > > race between recirc id free and the leak check.") which allows > > us to remove the test workaround. > > > > Signed-off-by: Ales Musil <[email protected]> > > --- > > controller/pinctrl.c | 5 +++-- > > ovs | 2 +- > > tests/ovn.at | 12 ++---------- > > 3 files changed, 6 insertions(+), 13 deletions(-) > > > > diff --git a/controller/pinctrl.c b/controller/pinctrl.c > > index ce7b9ac26..e6cdd7ca4 100644 > > --- a/controller/pinctrl.c > > +++ b/controller/pinctrl.c > > @@ -5206,7 +5206,7 @@ send_self_originated_neigh_packet(struct rconn > > *swconn, > > in6_addr_get_mapped_ipv4(local), > > in6_addr_get_mapped_ipv4(target)); > > } else { > > - compose_nd_ns(&packet, eth, local, target); > > + compose_nd_ns(&packet, true, eth, eth_addr_zero, local, target); > > } > > > > /* Inject GARP request. */ > > @@ -6575,7 +6575,8 @@ pinctrl_handle_nd_ns(struct rconn *swconn, const > > struct flow *ip_flow, > > hton128(flow_get_xxreg(&pin->flow_metadata.flow, 0)); > > memcpy(&ipv6_nh, &nexthop_be, sizeof ipv6_nh); > > > > - compose_nd_ns(&packet, ip_flow->dl_src, &ipv6_src, &ipv6_nh); > > + compose_nd_ns(&packet, true, ip_flow->dl_src, eth_addr_zero, &ipv6_src, > > + &ipv6_nh); > > > > /* Reload previous packet metadata and set actions from userdata. */ > > set_actions_and_enqueue_msg(swconn, &packet, > > diff --git a/ovs b/ovs > > index bdb95cc19..45ee6f7d9 160000 > > --- a/ovs > > +++ b/ovs > > @@ -1 +1 @@ > > -Subproject commit bdb95cc1920d4ab66fe062a9470eeb33a51d33e2 > > +Subproject commit 45ee6f7d96dec264df2c044bc9afa61d4ef5af37
The important bit of this patch is the new subproject commit. However, I should note that the old subproject commit is not correct based on what the current subproject commit is in main. Therefore this patch does not apply cleanly currently. When this is merged, the submodule bump will need to be manually recreated. > > diff --git a/tests/ovn.at b/tests/ovn.at > > index 42f1934fb..7395c8bd4 100644 > > --- a/tests/ovn.at > > +++ b/tests/ovn.at > > @@ -10336,11 +10336,7 @@ check ovn-nbctl --wait=hv set Logical_Switch_Port > > ln1 \ > > options:localnet_learn_fdb=true > > test_arp_nd_localnet > > > > -OVN_CLEANUP([hv1 > > -/left allocated/d > > -], [hv2 > > -/left allocated/d > > -]) > > +OVN_CLEANUP([hv1], [hv2]) > > AT_CLEANUP > > ]) > > > > @@ -25516,11 +25512,7 @@ OVS_WAIT_UNTIL( > > 'hv2/br-phys-1_n1-tx.pcap expected_empty' ], > > [$at_diff -F'^---' exp rcv]) > > > > -OVN_CLEANUP([hv1 > > -/left allocated/d > > -], [hv2 > > -/left allocated/d > > -]) > > +OVN_CLEANUP([hv1], [hv2]) > > AT_CLEANUP > > ]) > > > > -- > > 2.55.0 > > > > _______________________________________________ > > 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
