[ovs-dev] [PATCH ovn] system-tests: Try to load modules only if they weren't loaded before

When the system is busy with multiple test running concurrently the modprobe might fail with "Device or resource busy", which will fail the whole test. Instead of always trying to load the module check if it's already loaded. Also replace the AT_CHECK with OVS_WAIT_UNTIL which should reduce the cha

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

On Fri, Apr 28, 2023 at 7:20 AM Lorenzo Bianconi < lorenzo.bianc...@redhat.com> wrote: > > > On Tue, Apr 25, 2023 at 3:56 PM 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] seq: Make read of the current value atomic

On 3/27/23 13:25, Eelco Chaudron wrote: > Make the read of the current seq->value atomic, i.e., not needing to > acquire the global mutex when reading it. On 64-bit systems, this > incurs no overhead, and it will avoid the mutex and potentially > a system call. > > For incrementing the value follo

Re: [ovs-dev] [PATCH ovn 5/6] Implement MTU Path Discovery for multichassis ports

On 5/2/23 21:12, Ihar Hrachyshka wrote: When a multichassis port belongs to a switch with a localnet port, packets originating or directed to the multichassis port are NOT sent thorugh the localnet port. Instead, tunneling is enforced in-cluster to guarantee delivery of all packets to all chassis

Re: [ovs-dev] [PATCH ovn 3/6] if-status: track interfaces for additional chassis

On 5/2/23 21:12, Ihar Hrachyshka wrote: This will allow all chassis hosting a port to extract interface MTU from if-status-mgr. This will be used in a later patch to calculate the effective path MTU for each port. In addition, it's the right thing to do to claim and mark an interface on all chas

[ovs-dev] [PATCH ovn] docs: document that vxlan is supported for encap type

Most of the documentation was updated when the feature was implemented, but these two places were omitted by mistake. Fixes: b07f1bc3d068 ("Add VXLAN support for non-VTEP datapath bindings") Signed-off-by: Ihar Hrachyshka --- controller/ovn-controller.8.xml | 10 -- ovn-sb.xml

[ovs-dev] [PATCH ovn v2 4/6] Add new egress tables to accommodate for too-big packets handling

The new tables will be used in a later patch as follows: table=37, OFTABLE_OUTPUT_INIT: becomes an initial entry point into the egress pipeline that serves a semantic goal. (Not doing any actual processing at the moment.) table=38, OFTABLE_OUTPUT_LARGE_PKT_DETECT: detect "too-big" IP packets and

[ovs-dev] [PATCH ovn v2 5/6] Implement MTU Path Discovery for multichassis ports

When a multichassis port belongs to a switch with a localnet port, packets originating or directed to the multichassis port are NOT sent thorugh the localnet port. Instead, tunneling is enforced in-cluster to guarantee delivery of all packets to all chassis of the port. This behavior has an unfort

[ovs-dev] [PATCH ovn v2 6/6] Update multichassis physical flows on interface MTU update

Make ICMP Path MTU Discovery flows in table=38 react to underlying interface MTU update. NOTE: ideally, OVN would support Logical_Port MTU, in which case we wouldn't have to track OVSDB for interfaces, and we would also be able to react to MTU changes regardless of interface location. This patch i

[ovs-dev] [PATCH ovn v2 2/6] Track interface MTU in if-status-mgr

This will be used in a later patch to calculate the effective interface MTU after considering tunneling overhead. Signed-off-by: Ihar Hrachyshka --- controller/binding.c | 4 ++-- controller/if-status.c | 31 +++ controller/if-status.h | 3 +++ 3 files changed, 32

[ovs-dev] [PATCH ovn v2 3/6] if-status: track interfaces for additional chassis

This will allow all chassis hosting a port to extract interface MTU from if-status-mgr. This will be used in a later patch to calculate the effective path MTU for each port. In addition, it's the right thing to do to claim and mark an interface on all chassis as ovn-installed, even if the chassis

[ovs-dev] [PATCH ovn v2 0/6] Implement MTU Path Discovery for multichassis ports

This series fixes a non-optimal behavior with some multichassis ports. Specifically, - when a multichassis port belongs to a switch that also has a localnet port, - because ingress and egress traffic for the port is funnelled through tunnels to guarantee delivery of packets to all chassis tha

[ovs-dev] [PATCH ovn v2 1/6] Track ip version of tunnel in chassis_tunnel struct

This will be used in a later patch to calculate tunneling overhead for effective path MTU. Signed-off-by: Ihar Hrachyshka --- controller/local_data.c | 2 ++ controller/local_data.h | 1 + 2 files changed, 3 insertions(+) diff --git a/controller/local_data.c b/controller/local_data.c index acaf

Re: [ovs-dev] [PATCH v3] ci: Separate DPDK from OVS build.

On 4/28/23 14:54, David Marchand wrote: > Let's separate DPDK compilation from the rest of OVS build: > - this avoids multiple jobs building DPDK in parallel, which especially > affects builds in the dpdk-latest branch, > - we separate concerns about DPDK build requirements from OVS build > req

[ovs-dev] [PATCH ovn] northd: Update obsolete datapath groups instead of re-creating.

Every time we add a new switch or router it may result in addition or removal of a few datapath in datapath groups. Every time that happens, northd will remove obsolete groups, create new groups that differ by a few datapaths, and update all affected Sb LBs and logical flows with references to the

Re: [ovs-dev] [PATCH ovn v3 1/2] CI: Add a couple of periodic jobs using ovn-fake-multinode.

On Thu, Mar 30, 2023 at 8:54 AM Dumitru Ceara wrote: > > On 3/29/23 16:01, num...@ovn.org wrote: > > From: Numan Siddique > > > > This patch adds a couple of jobs using ovn-fake-multinode. > > It first builds 2 ovn-fake-multinode container images > > - one with OVN 22.03 > > - one with presen

Re: [ovs-dev] [PATCH ovn v2] utilities: increase OVSDB inactivity probe interval for ovn-*ctl

On 5/3/23 16:12, Vladislav Odintsov wrote: > Hi Dumitru and Mark, > > thanks for the review! > >> On 3 May 2023, at 16:56, Dumitru Ceara wrote: >> >> On 5/3/23 15:47, Mark Michelson wrote: >>> On 5/3/23 09:03, Dumitru Ceara wrote: Hi Vladislav, On 5/3/23 02:55, Vladislav Odintsov

Re: [ovs-dev] [PATCH ovn v2] utilities: increase OVSDB inactivity probe interval for ovn-*ctl

Hi Dumitru and Mark, thanks for the review! > On 3 May 2023, at 16:56, Dumitru Ceara wrote: > > On 5/3/23 15:47, Mark Michelson wrote: >> On 5/3/23 09:03, Dumitru Ceara wrote: >>> Hi Vladislav, >>> >>> On 5/3/23 02:55, Vladislav Odintsov wrote: For large OVN_Southbound (or other) database

Re: [ovs-dev] OVN LTS/non-LTS bug fix backport strategy

On 5/2/23 15:00, Ilya Maximets wrote: > On 5/2/23 06:50, Han Zhou wrote: >> >> >> On Fri, Apr 28, 2023 at 11:38 AM Mark Michelson > > wrote: >>> >>> Sorry for the top-post, but trying to write my responses inline at this >>> point is basically impossible for me if I want

Re: [ovs-dev] [PATCH 4/7] netdev-linux: use 64-bit rates in htb tc classes

On 4/25/23 14:51, Simon Horman wrote: On Fri, Apr 21, 2023 at 05:16:48PM +0200, Adrian Moreno wrote: Currently, htb rates are capped at ~34Gbps because they are internally expressed as 32-bit fields. Move min and max rates to 64-bit fields and use TCA_HTB_RATE64 and TCA_HTB_CEIL64 to configur

Re: [ovs-dev] [PATCH dpdk-latest v4 1/5] netdev-dpdk: use flow transfer proxy

Hello Simon, This patch has me intrigued. By the looks of it, it bears uncanny resemblance to patch [1] by another author. Is your patch based on patch [1]? If yes, could you please comment on the following: 1) Your patch does not seem to reference the original author. Why is it so? Is there

Re: [ovs-dev] [PATCH ovn v2] utilities: increase OVSDB inactivity probe interval for ovn-*ctl

On 5/3/23 15:47, Mark Michelson wrote: > On 5/3/23 09:03, Dumitru Ceara wrote: >> Hi Vladislav, >> >> On 5/3/23 02:55, Vladislav Odintsov wrote: >>> For large OVN_Southbound (or other) databases the default interval of >>> 5000 ms >>> could be not sufficient to run.  This patch adds configuration o

Re: [ovs-dev] [PATCH 1/7] netdev: add netdev_get_speed() to nedev API

On 4/25/23 14:47, Simon Horman wrote: On Fri, Apr 21, 2023 at 05:16:45PM +0200, Adrian Moreno wrote: Currently, the netdev's speed is being calculated by taking the link's feature bits (using netdev_get_features()) and transforming them into bps. This mechanism can be both inaccurate and diff

Re: [ovs-dev] [PATCH 6/7] netdev-linux: refactor nl_msg_put_act_police

On 4/26/23 15:39, Simon Horman wrote: On Fri, Apr 21, 2023 at 05:16:50PM +0200, Adrian Moreno wrote: In preparation for supporting 64-bit rates in tc policies, move the allocation and initialization of struct tc_police object inside nl_msg_put_act_police(). That way, the function is now called

Re: [ovs-dev] [PATCH ovn v2] utilities: increase OVSDB inactivity probe interval for ovn-*ctl

On 5/3/23 09:03, Dumitru Ceara wrote: Hi Vladislav, On 5/3/23 02:55, Vladislav Odintsov wrote: For large OVN_Southbound (or other) databases the default interval of 5000 ms could be not sufficient to run. This patch adds configuration of OVSDB inactivity probes for ovn-*ctl utilities. Initial

Re: [ovs-dev] [PATCH v2 ovn] mirror: fix ovn mirror support with IPv6

Thanks Lorenzo I have one minor comment down below, but it's very small and can be fixed by whoever merges the patch: Acked-by: Mark Michelson On 4/28/23 11:06, Lorenzo Bianconi wrote: Fix IPv6 support for OVN mirroring configuring properly interface type in OVS interface table Reported-at

Re: [ovs-dev] [PATCH ovn v2] utilities: increase OVSDB inactivity probe interval for ovn-*ctl

Hi Vladislav, On 5/3/23 02:55, Vladislav Odintsov wrote: > For large OVN_Southbound (or other) databases the default interval of 5000 ms > could be not sufficient to run. This patch adds configuration of OVSDB > inactivity probes for ovn-*ctl utilities. > > Initially, on the utility start the ha

Re: [ovs-dev] [PATCH v2] netdev-dpdk: Check rx/tx descriptor sizes for device.

On 03/05/2023 12:37, Ilya Maximets wrote: On 5/3/23 11:49, Kevin Traynor wrote: On 28/04/2023 19:16, Ilya Maximets wrote: On 4/14/23 17:44, 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.

Re: [ovs-dev] [PATCH v2] netdev-dpdk: Check rx/tx descriptor sizes for device.

On 5/3/23 11:49, Kevin Traynor wrote: > On 28/04/2023 19:16, Ilya Maximets wrote: >> On 4/14/23 17:44, 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

Re: [ovs-dev] [PATCH v2] netdev-dpdk: Check rx/tx descriptor sizes for device.

On 28/04/2023 19:16, Ilya Maximets wrote: On 4/14/23 17:44, 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.

Re: [ovs-dev] [PATCH ovn 1/2] tests: define fmt_pkt function to construct packets with scapy

On Tue, May 02, 2023 at 02:29:51PM +0200, Dumitru Ceara wrote: > On 4/20/23 16:28, Ihar Hrachyshka wrote: > > I don't see how this is any better. Instead of calling to python that > > imports GPL libraries, we now call to scapy that calls to python that > > imports GPL libraries. In both cases we c