Re: [ovs-dev] [PATCH v6 ovn 4/9] controller: configure qos through ovs qos table and do not run tc directly

On Wed, May 10, 2023 at 10:22 AM Lorenzo Bianconi wrote: > > Rework OVN QoS implementation in order to configure it through OVS QoS > table instead of running tc command directly bypassing OVS. > > Acked-By: Ihar Hrachyshka > Reviewed-by: Simon Horman > Tested-by: Rodolfo Alonso >

Re: [ovs-dev] [PATCH v2] ofproto-dpif-xlate: Fix use-after-free when xlate_actions().

On 5/6/23 12:00, Yunjian Wang wrote: > Currently, bundle->cvlans and xbundle->cvlans are pointing to the > same memory location. This can cause issues if the main thread > modifies bundle->cvlans and frees it while the revalidator thread > is still accessing xbundle->cvlans. This leads to

Re: [ovs-dev] [PATCH v2] netdev-offload: Fix deadlock/recursive use of the netdev_hmap_rwlock rwlock.

On 5/9/23 16:29, Eelco Chaudron wrote: > When doing performance testing with OVS v3.1 we ran into a deadlock > situation with the netdev_hmap_rwlock read/write lock. After some > debugging, it was discovered that the netdev_hmap_rwlock read lock > was taken recursively. And well in the folowing

Re: [ovs-dev] [PATCH v3 2/2] userspace: Add new option srv6_flowlabel in SRv6 tunnel.

On 5/9/23 11:38, Nobuhiro MIKI wrote: > It supports flowlabel based load balancing by controlling the flowlabel > of outer IPv6 header, which is already implemented in Linux kernel as > seg6_flowlabel sysctl [1]. > > [1]: https://docs.kernel.org/networking/seg6-sysctl.html > > Signed-off-by:

Re: [ovs-dev] [PATCH v3 1/2] netdev-native-tnl: Add ipv6_label param in netdev_tnl_ip_build_header.

On 5/9/23 11:37, Nobuhiro MIKI wrote: > For tunnels such as SRv6, some popular vendor appliances support > IPv6 flowlabel based load balancing. In preparation for OVS to > support it, this patch modifies the encapsulation to allow IPv6 > flowlabel to be configured. > > Signed-off-by: Nobuhiro

Re: [ovs-dev] [PATCH v2 2/2] test-stream: Add ssl tests for stream open block

On 5/9/23 10:25, Stefan Hoffmann wrote: > This tests stream.c and stream.py with ssl connection at > CHECK_STREAM_OPEN_BLOCK. > For the tests, ovsdb needs to be build with libssl. > > Signed-off-by: Stefan Hoffmann > --- > tests/ovsdb-idl.at | 32 >

Re: [ovs-dev] [PATCH v3 ovn] mirror: fix ovn mirror support with IPv6

On 5/10/23 21:15, Han Zhou wrote: > > > On Wed, May 10, 2023 at 10:36 AM Dumitru Ceara > wrote: >> >> On 5/10/23 18:01, Han Zhou wrote: >> > On Fri, May 5, 2023 at 4:18 AM Lorenzo Bianconi > mailto:lorenzo.bianc...@redhat.com>> >> > wrote: >> >> >> >> Fix IPv6 support

Re: [ovs-dev] [PATCH v3 ovn] mirror: fix ovn mirror support with IPv6

On Wed, May 10, 2023 at 10:36 AM Dumitru Ceara wrote: > > On 5/10/23 18:01, Han Zhou wrote: > > On Fri, May 5, 2023 at 4:18 AM Lorenzo Bianconi < lorenzo.bianc...@redhat.com> > > wrote: > >> > >> Fix IPv6 support for OVN mirroring configuring properly interface type > >> in OVS interface table >

Re: [ovs-dev] [PATCH v2 1/2] tests-ovsdb: switch OVSDB_START_IDLTEST to macro

On 5/9/23 12:05, Stefan Hoffmann wrote: > The failed pipeline is bfd-decay again, which seems to fail out of > reason sometimes. Test works on my machine. > > Is there a way to rerun the pipeline? > > https://github.com/ovsrobot/ovs/actions/runs/4924091327 Unfortunately, there is no way to

Re: [ovs-dev] [PATCH v2 1/2] tests-ovsdb: switch OVSDB_START_IDLTEST to macro

On 5/9/23 10:20, Stefan Hoffmann wrote: > Define bash function as macro now. Later we can extend this macro for other > usecases. > > Signed-off-by: Stefan Hoffmann > --- > tests/ovsdb-idl.at | 71 +- > 1 file changed, 38 insertions(+), 33

Re: [ovs-dev] [PATCH v3 2/2] netdev-dpdk: Add support for userspace port-based ingress packet-per-second policing.

On 5/1/23 13:33, mit...@outlook.com wrote: > From: Lin Huang > > OvS has supported packet-per-second policer which can be set at ingress > and egress side in kernel datapath. But the userspace datapath dosen't > support for ingress and egress packet-per-second policing now. > > So, this patch

Re: [ovs-dev] [PATCH v3 ovn] mirror: fix ovn mirror support with IPv6

On 5/10/23 18:01, Han Zhou wrote: > On Fri, May 5, 2023 at 4:18 AM Lorenzo Bianconi > wrote: >> >> Fix IPv6 support for OVN mirroring configuring properly interface type >> in OVS interface table >> >> Acked-by: Mark Michelson >> Reported-at: https://bugzilla.redhat.com/show_bug.cgi?id=2168119

Re: [ovs-dev] [PATCH] tc: fix crash on EAGAIN return from recvmsg on netlink socket.

Hi Frode, Thanks for fixing this, just had one nit below since you're pushing out a new PS. On 5/10/23 11:12 AM, Frode Nordahl wrote: The tc module combines the use of the `tc_transact` helper function for communication with the in-kernel tc infrastructure with assertions on the reply data

Re: [ovs-dev] [PATCH ovn 1/2] mirror: Enhance port-mirroring to support local OVS port as target.

On Mon, May 8, 2023 at 11:06 AM Mark Michelson wrote: > > Hi Han, > > For both patches in the series, > > Acked-by: Mark Michelson > Thanks Mark. I applied to main. > I want to put a reply to patch 2 directly, but unfortunately I can't > find it in my email. Sometimes patches get sent to spam,

Re: [ovs-dev] [PATCH] tc: fix crash on EAGAIN return from recvmsg on netlink socket.

Hello again, While we have confirmed this patch fixes a real issue in production and it passes the regular functional tests, I was apparently a bit quick out the door with this patch. It appears to break the system tests due to some unexpected WRN level logging: ../../tests/system-traffic.at:23:

Re: [ovs-dev] [PATCH v3 ovn] mirror: fix ovn mirror support with IPv6

On Fri, May 5, 2023 at 4:18 AM Lorenzo Bianconi wrote: > > Fix IPv6 support for OVN mirroring configuring properly interface type > in OVS interface table > > Acked-by: Mark Michelson > Reported-at: https://bugzilla.redhat.com/show_bug.cgi?id=2168119 > Fixes: ba8aa26e44cb ("OVN Remote Port

Re: [ovs-dev] [PATCH v3 1/2] netdev-dpdk: Add support for userspace port-based egress packet-per-second policing.

Hi simon, Thanks for your reviewing. On 5/10/2023 10:34 PM, Simon Horman wrote: On Mon, May 01, 2023 at 07:33:38PM +0800, mit...@outlook.com wrote: From: Lin Huang OvS has supported packet-per-second policer which can be set at ingress and egress side in kernel datapath. But the userspace

Re: [ovs-dev] [PATCH v2 7/7] tc: Add vxlan encap action with gbp option offload

On Tue, Apr 25, 2023 at 03:41:22PM +0300, Roi Dayan via dev wrote: > From: Gavin Li > > Add TC offload support for vxlan encap with gbp option > > Signed-off-by: Gavin Li > Reviewed-by: Gavi Teitz > Reviewed-by: Roi Dayan No objections. But I guess this will also change wrt netlink

Re: [ovs-dev] [PATCH v2 6/7] tc: Pass encap entirely to nl_msg_put_act_tunnel_key_set

On Tue, Apr 25, 2023 at 03:41:21PM +0300, Roi Dayan via dev wrote: > From: Gavin Li > > Most of the data members of struct tc_action{ } are defined as anonymous > struct in place. Instead of passing all members of an anonymous struct, > which is not flexible to new members being added, expose

Re: [ovs-dev] [PATCH v2 5/7] tc: Add vxlan gbp option flower match offload

On Tue, Apr 25, 2023 at 03:41:20PM +0300, Roi Dayan via dev wrote: > From: Gavin Li > > Add TC offload support for filtering vxlan tunnels with gbp option > > Signed-off-by: Gavin Li > Reviewed-by: Gavi Teitz ... > @@ -3439,11 +3479,12 @@ nl_msg_put_flower_tunnel_opts(struct ofpbuf

Re: [ovs-dev] [PATCH v3 1/2] netdev-dpdk: Remove requested descriptors from get_config.

On Thu, May 4, 2023 at 7:36 PM Kevin Traynor wrote: > > There is no need to display 'requested_rx/tx_descriptors' and > 'configured_rx/tx_descriptors' as they will be the same. > > It simpler to just have a single 'n_rxq/txq_desc' value. It is* > > Suggested-by: Ilya Maximets > Signed-off-by:

Re: [ovs-dev] [PATCH v2 2/7] odp-util: Extract vxlan gbp option decoding to a function

On Tue, Apr 25, 2023 at 03:41:17PM +0300, Roi Dayan via dev wrote: > From: Gavin Li > > Extract vxlan gbp option decoding to odp_decode_gbp_raw to be used in > following commits. > > Signed-off-by: Gavin Li > Reviewed-by: Roi Dayan Reviewed-by: Simon Horman

Re: [ovs-dev] [PATCH v3 2/2] netdev-dpdk: Check rx/tx descriptor sizes for device.

On Thu, May 4, 2023 at 7:36 PM Kevin Traynor wrote: > > By default OVS configures 2048 descriptors for tx and rx queues > on DPDK devices. It also allows the user to configure those values. > > If the values used are not acceptable to the device then queue setup > would fail. > > The device

Re: [ovs-dev] [PATCH v2 3/7] odp-util: Extract vxlan gbp option encoding to a function

On Tue, Apr 25, 2023 at 03:41:18PM +0300, Roi Dayan via dev wrote: > From: Gavin Li > > Extract vxlan gbp option encoding to odp_encode_gbp_raw to be used in > following commits. > > Signed-off-by: Gavin Li > Reviewed-by: Roi Dayan Reviewed-by: Simon Horman

Re: [ovs-dev] [PATCH v2 1/7] tc: Pass tunnel entirely to tunnel option parse and put functions

On Tue, Apr 25, 2023 at 03:41:16PM +0300, Roi Dayan via dev wrote: > From: Gavin Li > > Tc flower tunnel key options were encoded in nl_msg_put_flower_tunnel_opts > and decoded in nl_parse_flower_tunnel_opts. Only geneve was supported. > > To avoid adding more arguments to the function to

[ovs-dev] [PATCH] tc: fix crash on EAGAIN return from recvmsg on netlink socket.

The tc module combines the use of the `tc_transact` helper function for communication with the in-kernel tc infrastructure with assertions on the reply data by `ofpbuf_at_assert` on the received data prior to further processing. `tc_transact` in turn calls `nl_transact`, which via

Re: [ovs-dev] [PATCH v26 8/8] system-offloads-traffic.at: Add sFlow offload test cases

On 27 Apr 2023, at 3:27, Chris Mi wrote: >>> recirc_id(0),in_port(1),eth(src=06:9c:de:63:c9:40,dst=01:00:5e:00:00:fb),eth_type(0x0800),ipv4(tos=0/0x3,frag=no), >>> packets:13, bytes:1883, used:5.700s, >>>

Re: [ovs-dev] [PATCH v26 8/8] system-offloads-traffic.at: Add sFlow offload test cases

On 26 Apr 2023, at 4:47, Chris Mi wrote: >>> + >>> +P0_IFINDEX=$(cat /sys/class/net/ovs-p0/ifindex) >>> +m4_define([DUMP_SFLOW], [sed -e >>> "s/used:[[0-9]].[[0-9]]*s/used:0.001s/;s/eth(src=[[a-z0-9:]]*,dst=[[a-z0-9:]]*)/eth(macs)/;s/pid=[[0-9]]*/pid=1/;s/output=$P0_IFINDEX/output=1/"]) >>>

Re: [ovs-dev] [PATCH v2 4/7] tc: Add NLA_F_NESTED to nested netlink messages

On Wed, May 10, 2023 at 04:28:06PM +0200, Ilya Maximets wrote: > On 5/10/23 16:10, Simon Horman wrote: > > On Wed, Apr 26, 2023 at 06:39:22PM +0200, Ilya Maximets wrote: > >> On 4/25/23 14:41, Roi Dayan wrote: > >>> From: Gavin Li > >>> > >>> Linux kernel netlink module added NLA_F_NESTED flag

Re: [ovs-dev] [PATCH v3 1/2] netdev-dpdk: Add support for userspace port-based egress packet-per-second policing.

On Mon, May 01, 2023 at 07:33:38PM +0800, mit...@outlook.com wrote: > From: Lin Huang > > OvS has supported packet-per-second policer which can be set at ingress > and egress side in kernel datapath. But the userspace datapath dosen't s/dosen't/doesn't/ > support for ingress and egress

Re: [ovs-dev] [PATCH v2 4/7] tc: Add NLA_F_NESTED to nested netlink messages

On 5/10/23 16:10, Simon Horman wrote: > On Wed, Apr 26, 2023 at 06:39:22PM +0200, Ilya Maximets wrote: >> On 4/25/23 14:41, Roi Dayan wrote: >>> From: Gavin Li >>> >>> Linux kernel netlink module added NLA_F_NESTED flag checking for nested >>> netlink messages in 5.2. A nested message without the

[ovs-dev] [PATCH v6 ovn 6/9] controller: use unsigned long long int for qos_max_rate/qos_min_rate/qos_burst

This patch allow to configure max/min rate greater than 4Gbps Acked-By: Ihar Hrachyshka Tested-by: Rodolfo Alonso Signed-off-by: Lorenzo Bianconi --- controller/binding.c | 43 ++- 1 file changed, 26 insertions(+), 17 deletions(-) diff --git

[ovs-dev] [PATCH v6 ovn 7/9] northd: apply QoS rules on the localnet port related to LSP ports

This patch allows to apply QoS rules on the localnet port related to logical switch ports running on the same datapath. Considering the following netowrk configuration: LSP{0,1} -- LogicalSwitch -- Localnet0 It is possible to apply the following QoS rules on Localnet0 on egress traffic entering

[ovs-dev] [PATCH v6 ovn 8/9] controller: get rid of egress_ifaces sset

egress_ifaces sset is no longer used by ovn-controller, so get rid of it Signed-off-by: Lorenzo Bianconi --- controller/binding.c| 40 - controller/binding.h| 1 - controller/ovn-controller.c | 10 ++ 3 files changed, 2 insertions(+),

[ovs-dev] [PATCH v6 ovn 9/9] update NEWS with new QoS info

Reported-at: https://bugzilla.redhat.com/show_bug.cgi?id=2129742 Reported-at: https://bugzilla.redhat.com/show_bug.cgi?id=2139100 Acked-By: Ihar Hrachyshka Signed-off-by: Lorenzo Bianconi --- NEWS | 6 ++ 1 file changed, 6 insertions(+) diff --git a/NEWS b/NEWS index 1f1ac8c5a..f467d7bbe

[ovs-dev] [PATCH v6 ovn 5/9] controller: improve ovs port lookup by name and qos

Introduce ovsport_lookup_by_name and ovsport_lookup_by_qos routines in order to speed-up ovs port lookup based on port name or qos. Acked-By: Ihar Hrachyshka Tested-by: Rodolfo Alonso Signed-off-by: Lorenzo Bianconi --- controller/binding.c| 25 +++--

[ovs-dev] [PATCH v6 ovn 4/9] controller: configure qos through ovs qos table and do not run tc directly

Rework OVN QoS implementation in order to configure it through OVS QoS table instead of running tc command directly bypassing OVS. Acked-By: Ihar Hrachyshka Reviewed-by: Simon Horman Tested-by: Rodolfo Alonso Signed-off-by: Lorenzo Bianconi --- controller/binding.c| 352

[ovs-dev] [PATCH v6 ovn 3/9] northd: introduce qos_physical_network in port_binding other_config column

This is a preliminary patch to rework OVN QoS implementation in order to configure it through OVS QoS table instead of running tc command directly bypassing OVS. Signed-off-by: Lorenzo Bianconi --- controller/binding.c | 7 ++- northd/northd.c | 8 ovn-sb.xml | 5

[ovs-dev] [PATCH v6 ovn 2/9] controller: add incremental processing for ovn-controller qos_map

This is a preliminary patch to rework OVN QoS implementation in order to configure it through OVS QoS table instead of running tc command directly bypassing OVS. Signed-off-by: Lorenzo Bianconi --- controller/binding.c| 254 +--- controller/binding.h

[ovs-dev] [PATCH v6 ovn 1/9] controller: remove tunnel interfaces from egress_ifaces sset

Remove tunnel interfaces from egress list in order to not shape them. Acked-By: Ihar Hrachyshka Tested-by: Rodolfo Alonso Signed-off-by: Lorenzo Bianconi --- controller/binding.c | 9 - 1 file changed, 9 deletions(-) diff --git a/controller/binding.c b/controller/binding.c index

[ovs-dev] [PATCH v6 ovn 0/9] Configure OVN QoS thorugh OvS db

Rework OVN QoS implementation in order to configure it through OVS QoS table instead of running tc command directly bypassing OVS. This series allows to apply QoS rules on the localnet port related to logical switch ports running on the same datapath. Considering the following netowrk

Re: [ovs-dev] [PATCH v2 4/7] tc: Add NLA_F_NESTED to nested netlink messages

On Wed, Apr 26, 2023 at 06:39:22PM +0200, Ilya Maximets wrote: > On 4/25/23 14:41, Roi Dayan wrote: > > From: Gavin Li > > > > Linux kernel netlink module added NLA_F_NESTED flag checking for nested > > netlink messages in 5.2. A nested message without the flag set will be > > treated as

Re: [ovs-dev] [PATCH 2/2] conntrack: Release nat_conn in case both keys have the same hash.

On 5/4/23 19:21, Paolo Valerio wrote: > Ilya Maximets writes: > >> On 4/19/23 20:40, Paolo Valerio wrote: >>> During the creation of a new connection, there's a chance both key and >>> rev_key end up having the same hash. This is more common in the case >>> of all-zero snat with no collisions.

Re: [ovs-dev] [PATCH v26 7/8] netdev-offload-tc: Add offload support for sFlow

On 26 Apr 2023, at 4:44, Chris Mi wrote: >>> + >>> +static int >>> +offload_sample_init(struct offload_sample *sample, >>> +const struct nlattr *next_actions, >>> +size_t next_actions_len, bool tunnel, >>> +const struct flow_tnl

Re: [ovs-dev] [PATCH v2 1/2] tests-ovsdb: switch OVSDB_START_IDLTEST to macro

On 5/10/23 13:48, Stefan Hoffmann wrote: > On Wed, 2023-05-10 at 13:15 +0200, Ilya Maximets wrote: >> On 5/10/23 11:43, Stefan Hoffmann wrote: >>> Define bash function as macro now. Later we can extend this macro for other >>> usecases. >>> >>> Signed-off-by: Stefan Hoffmann >> >> Hi, Stefan. Is

Re: [ovs-dev] [PATCH] dpif-netdev: Add config option for vhost tx_steering default mode

On Fri, May 5, 2023 at 9:41 PM Ilya Maximets wrote: > > On 5/3/23 00:20, Christophe Fontaine wrote: > > Current tx_steering mode has to be configured on each Interface. > > The global config option "vhost-tx-steering-default" allows > > the user to select a default mode (thread or hash) for > >

Re: [ovs-dev] [PATCH v2 1/2] tests-ovsdb: switch OVSDB_START_IDLTEST to macro

On Wed, 2023-05-10 at 13:15 +0200, Ilya Maximets wrote: > On 5/10/23 11:43, Stefan Hoffmann wrote: > > Define bash function as macro now. Later we can extend this macro for other > > usecases. > > > > Signed-off-by: Stefan Hoffmann > > Hi, Stefan. Is there any difference between this patch and

Re: [ovs-dev] [PATCH v5 ovn 0/7] Configure OVN QoS thorugh OvS db

On 4/25/23 20:36, Lorenzo Bianconi wrote: > Rework OVN QoS implementation in order to configure it through OVS QoS > table instead of running tc command directly bypassing OVS. Hi, Lorenzo. One question: Does this patch set impact the behavior around custom qdiscs applied externally? For

Re: [ovs-dev] [PATCH v26 5/8] netdev-offload: Add netdev offload recv and recv_wait APIs

On 26 Apr 2023, at 4:38, Chris Mi wrote: > On 4/12/2023 10:07 PM, Eelco Chaudron wrote: >> On 29 Mar 2023, at 13:42, Chris Mi wrote: >> >>> Iterate each registered offload API. It's not a problem for today >>> since we only have one implementation. >> See comment inline. >> >> //Eelco >> >>>

Re: [ovs-dev] [PATCH v26 4/8] netdev-offload-tc: Add sample offload API for TC

On 26 Apr 2023, at 4:42, Chris Mi wrote: >>> /* DPIF_UC_ACTION only. */ >>> struct nlattr *userdata;/* Argument to OVS_ACTION_ATTR_USERSPACE. >>> */ >>> -struct nlattr *out_tun_key;/* Output tunnel key. */ >>> -struct nlattr *actions;/* Argument to

Re: [ovs-dev] [PATCH ovn] tests: Fix flakiness of policy based routing on slower systems

On Wed, May 10, 2023 at 1:11 PM Xavier Simonart wrote: > Hi Ales > Hi Xavier, thank you for the review. > > Thanks for the patch. > I have small questions/comments. > Otherwise, it looks good to me. > Thanks > Xavier > > On Wed, May 10, 2023 at 10:04 AM Ales Musil wrote: > >> The test

[ovs-dev] [PATCH ovn v2] tests: Fix flakiness of policy based routing on slower systems

The test expected that the packet statistics will be immediately reflected after packet inject, however that might not be true on slower systems. Use OVS_WAIT_UNTIL instead to ensure that the packet really went through. Also align the IPv4 and IPv6 variant and fix some checks that were suggesting

Re: [ovs-dev] [PATCH v2 1/2] tests-ovsdb: switch OVSDB_START_IDLTEST to macro

On 5/10/23 11:43, Stefan Hoffmann wrote: > Define bash function as macro now. Later we can extend this macro for other > usecases. > > Signed-off-by: Stefan Hoffmann Hi, Stefan. Is there any difference between this patch and the previous (also v2):

Re: [ovs-dev] [PATCH ovn] tests: Fix flakiness of policy based routing on slower systems

Hi Ales Thanks for the patch. I have small questions/comments. Otherwise, it looks good to me. Thanks Xavier On Wed, May 10, 2023 at 10:04 AM Ales Musil wrote: > The test expected that the packet statistics will be > immediately reflected after packet inject, however that > might not be true

[ovs-dev] OVS patches public meeting

Hi All, There is a public call once a month to discuss OVS patches. It is not required for anyone to attend, but sometimes it can be useful to have a quick sync on status etc. It is on today and every fourth Wednesday from today at @ 1600 UTC for 30 mins. Everyone is free to join. If anyone

Re: [ovs-dev] [PATCH v26 3/8] netdev-offload-tc: Introduce group ID management API

On 26 Apr 2023, at 4:36, Chris Mi wrote: > On 4/12/2023 10:06 PM, Eelco Chaudron wrote: >>> +{ >>> +new->type = old->type; >>> +new->action = xmemdup(old->action, old->action->nla_len); >>> +new->userspace_actions = old->userspace_actions >>> + ?

[ovs-dev] [PATCH v2 1/2] tests-ovsdb: switch OVSDB_START_IDLTEST to macro

Define bash function as macro now. Later we can extend this macro for other usecases. Signed-off-by: Stefan Hoffmann --- tests/ovsdb-idl.at | 71 +- 1 file changed, 38 insertions(+), 33 deletions(-) diff --git a/tests/ovsdb-idl.at

[ovs-dev] [PATCH v2] dpctl: Fix dereferencing null pointer in parse_ct_limit_zones().

Command with empty string following "dpctl/ct-get-limits zone=" such as "ovs-appctl dpctl/ct-get-limits zone=" will cause parse_ct_limit_zones() dereferencing null. Signed-off-by: Zhiqi Chen --- lib/dpctl.c| 5 +++-- tests/dpctl.at | 16 2 files changed, 19 insertions(+),

Re: [ovs-dev] [PATCH ovn v3] Pass localnet traffic through CT when a LB is configured.

On Mon, May 8, 2023 at 5:19 PM Mark Michelson wrote: > Current code always skips conntrack for traffic that ingresses or > egresses on a localnet port. However, this makes it impossible for > traffic to be load-balanced when it arrives on a localnet port. > > This patch allows for traffic to be

[ovs-dev] [PATCH ovn] tests: Fix flakiness of policy based routing on slower systems

The test expected that the packet statistics will be immediately reflected after packet inject, however that might not be true on slower systems. Use OVS_WAIT_UNTIL instead to ensure that the packet really went through. Also align the IPv4 and IPv6 variant and fix some checks that were suggesting