[ovs-dev] OVN 20.12 Branched

Hi all, OVN 20.12 has been branched. Ideally we will have 2 weeks of testing and bug fixes and release on 18 December. If there is a delay for some reason, it will run into the Christmas/New Year's holidays and likely end up being released in January. Even if it is released in January, we

Re: [ovs-dev] [PATCH ovn v3 0/6] Combine Logical Flows by Logical Datapath.

I pushed the series to master. Thanks Ilya and Dumitru! On 12/4/20 12:54 PM, Ilya Maximets wrote: This series is aimed to reduce size of a Southbound DB. In modern OVN cluster there is a big probability that same logical flows are installed for several logical switches. Currently, OVN will

Re: [ovs-dev] [PATCH net] net: openvswitch: fix TTL decrement exception action execution

On Fri, 4 Dec 2020 07:16:23 -0500 Eelco Chaudron wrote: > Currently, the exception actions are not processed correctly as the wrong > dataset is passed. This change fixes this, including the misleading > comment. > > In addition, a check was added to make sure we work on an IPv4 packet, > and

Re: [ovs-dev] [PATCH net] openvswitch: fix error return code in validate_and_copy_dec_ttl()

On Fri, 04 Dec 2020 13:07:48 +0100 Eelco Chaudron wrote: > > Fix to return a negative error code from the error handling > > case instead of 0, as done elsewhere in this function. > > > > Changing 'return start' to 'return action_start' can fix this bug. > > > > Fixes: 69929d4c49e1 ("net:

[ovs-dev] Taller práctico de valuación de puestos

Webinar en Vivo: Taller práctico de valuación de puestos Jueves 14 de Enero - Horario de 10:00 a 17:00 Hrs. ¿Por qué él gana más que yo si tiene menos tiempo en la empresa? ¿Cuánto le vamos a pagar al nuevo empleado? ¿Con qué sueldo publico la vacante del nuevo puesto?, Todas estas preguntas

Re: [ovs-dev] [PATCH 1/4] dpif-netdev: Flush offload rules upon port deletion

On 12/2/20 8:23 AM, Eli Britstein wrote: > When a port is deleted, flow deletion requests are posted, and the netdev > is removed from offload netdevs map. Following flow deletion handling may > be done after the netdev has already been removed from the offload > netdevs map, so the HW rule is not

Re: [ovs-dev] [PATCH] ovsdb-cluster.at: Fix infinite loop in torture tests.

On 11/26/20 11:17 AM, Ilya Maximets wrote: > For some reason, while running cluster torture tests in GitHub Actions > workflow, failure of 'echo' command doesn't fail the loop and subshell > never exits, but keeps infinitely printing errors after breaking from > the loop on the right side of the

Re: [ovs-dev] [PATCH] ovsdb-idl: Fix expected condition seqno when changes are pending.

On 12/4/20 3:54 PM, Dumitru Ceara wrote: > Commit 17f22fe46142 tried to address this but only covered some of the > cases. > > The correct way to report the expected seqno is to take into account if > there already is a condition change that was requested to the server but > not acked yet. In

Re: [ovs-dev] [PATCH ovn] northd: Fix iteration over vip backends.

On 12/4/20 7:50 PM, Ilya Maximets wrote: > During refactoring, direct access such as 'lb->vips[i].backends[j].op' > was simplified by the presence of the 'backend_nb' pointer. But instead > of the inner one, the index of the outer loop was used. This forces > northd to only use one backend per

Re: [ovs-dev] [PATCH ovn v3 3/6] controller: Add support for Logical Datapath Groups.

On 12/4/20 6:54 PM, Ilya Maximets wrote: > ovn-controller will receive updates from Logical_DP_Group table > and process logical flows accordingly. Feature is fully backward > compatible since old 'logical_datapath' column kept as is. > It will also be used by nothd to not create datapath groups

Re: [ovs-dev] [PATCH] netdev-afxdp: updated documentation

On 12/2/20 7:21 PM, Toms Atteka wrote: > If system is missing ethtool, test fails silently. > > Signed-off-by: Toms Atteka > --- > Documentation/intro/install/afxdp.rst | 2 +- > 1 file changed, 1 insertion(+), 1 deletion(-) > > diff --git a/Documentation/intro/install/afxdp.rst >

[ovs-dev] [PATCH ovn] northd: Fix iteration over vip backends.

During refactoring, direct access such as 'lb->vips[i].backends[j].op' was simplified by the presence of the 'backend_nb' pointer. But instead of the inner one, the index of the outer loop was used. This forces northd to only use one backend per vip, and can also lead to invalid memory accesses.

[ovs-dev] [PATCH ovn v3 1/6] tests: Sort flow and database dumps.

The order in which flows appears in the dump is not strictly determined. Need to sort the output to make tests independent from the system and implementation details. Signed-off-by: Ilya Maximets Acked-by: Dumitru Ceara Acked-by: Mark Michelson --- tests/ovn-northd.at | 10 +-

[ovs-dev] [PATCH ovn v3 3/6] controller: Add support for Logical Datapath Groups.

ovn-controller will receive updates from Logical_DP_Group table and process logical flows accordingly. Feature is fully backward compatible since old 'logical_datapath' column kept as is. It will also be used by nothd to not create datapath groups if there is ony one datapath in it.

[ovs-dev] [PATCH ovn v3 5/6] northd: Add support for Logical Datapath Groups.

northd will create only one logical flow with a single logical datapath group instead of several lflows that differs only by a datapath column. We're avoiding creation of datapath groups with only one datapath to save space. Old 'logical_datapath' column used instead. Feature is disabled.

[ovs-dev] [PATCH ovn v3 6/6] northd: Config option to enable use of Logical Datapath Groups.

New configuration knob: ovn-nbctl set NB_Global . options:use_logical_dp_groups=true Feature disabled by default, because it could cause performance issues on ovn-controller side, specifically, if you have a lot of logical switches or routers local to the same node, i.e. handled by the same

[ovs-dev] [PATCH ovn v3 2/6] ovn-sb.ovsschema: Add Logical Datapath Groups.

New table 'Logical_Datapath_Group'. Each row of this table will store a set of logical datapaths. Each logical flow now has a reference to a logical datapath group, so the same logical flow could be applied to many different logical datapaths in a group. Signed-off-by: Ilya Maximets Acked-by:

[ovs-dev] [PATCH ovn v3 4/6] utilities: Add support for Logical Datapath Groups.

Signed-off-by: Ilya Maximets Acked-by: Dumitru Ceara Acked-by: Mark Michelson --- utilities/ovn-detrace.in | 8 ++- utilities/ovn-sbctl.c| 108 --- utilities/ovn-trace.c| 40 +++ 3 files changed, 115 insertions(+), 41 deletions(-)

[ovs-dev] [PATCH ovn v3 0/6] Combine Logical Flows by Logical Datapath.

This series is aimed to reduce size of a Southbound DB. In modern OVN cluster there is a big probability that same logical flows are installed for several logical switches. Currently, OVN will create a separate logical flow for each logical switch (datapath) filled with exact same match, actions

Re: [ovs-dev] [PATCH ovn v2 3/6] controller: Add support for Logical Datapath Groups.

On 12/4/20 5:36 PM, Dumitru Ceara wrote: > On 12/3/20 3:03 PM, Ilya Maximets wrote: >> ovn-controller will receive updates from Logical_DP_Group table >> and process logical flows accordingly. Feature is fully backward >> compatible since old 'logical_datapath' column kept as is. >> It will also

Re: [ovs-dev] [PATCH ovn v2 6/6] northd: Config option to enable use of Logical Datapath Groups.

On 12/3/20 3:03 PM, Ilya Maximets wrote: > New configuration knob: > ovn-nbctl set NB_Global . options:use_logical_dp_groups=true > > Feature disabled by default, because it could cause performance > issues on ovn-controller side, specifically, if you have a lot > of logical switches or routers

Re: [ovs-dev] [PATCH ovn v2 4/6] utilities: Add support for Logical Datapath Groups.

On 12/3/20 3:03 PM, Ilya Maximets wrote: > Signed-off-by: Ilya Maximets > --- > utilities/ovn-detrace.in | 8 ++- > utilities/ovn-sbctl.c| 108 --- > utilities/ovn-trace.c| 40 +++ > 3 files changed, 115 insertions(+), 41 deletions(-) >

Re: [ovs-dev] [PATCH ovn v2 2/6] ovn-sb.ovsschema: Add Logical Datapath Groups.

On 12/3/20 3:03 PM, Ilya Maximets wrote: > New table 'Logical_Datapath_Group'. Each row of this table will store > a set of logical datapaths. Each logical flow now has a reference to > a logical datapath group, so the same logical flow could be applied to > many different logical datapaths in a

Re: [ovs-dev] [PATCH ovn v2 1/6] tests: Sort flow and database dumps.

On 12/3/20 3:03 PM, Ilya Maximets wrote: > The order in which flows appears in the dump is not strictly > determined. Need to sort the output to make tests independent > from the system and implementation details. > > Signed-off-by: Ilya Maximets > --- Acked-by: Dumitru Ceara Thanks!

Re: [ovs-dev] [PATCH ovn 0/4] introduce BFD support in ovn-controller

On Fri, Dec 4, 2020 at 8:04 PM Lorenzo Bianconi wrote: > > > > > On Wed, Nov 25, 2020 at 3:30 PM Lorenzo Bianconi > > wrote: > > > > > > Introduce BFD protocol in ovn-controller according to RFC5880 [0] > > > We added BFD implementation in ovn since layered protocols usually > > > request to >

Re: [ovs-dev] [PATCH ovn v2 3/6] controller: Add support for Logical Datapath Groups.

On 12/3/20 3:03 PM, Ilya Maximets wrote: > ovn-controller will receive updates from Logical_DP_Group table > and process logical flows accordingly. Feature is fully backward > compatible since old 'logical_datapath' column kept as is. > It will also be used by nothd to not create datapath groups

Re: [ovs-dev] [PATCH ovn v2 0/6] Combine Logical Flows by Logical Datapath.

For the series, Acked-by: Mark Michelson On 12/3/20 9:03 AM, Ilya Maximets wrote: This series is aimed to reduce size of a Southbound DB. In modern OVN cluster there is a big probability that same logical flows are installed for several logical switches. Currently, OVN will create a separate

Re: [ovs-dev] [PATCH ovn] ovn-controller: Relax get_nb_cfg() condition seqno check.

On 12/4/20 9:13 AM, Han Zhou wrote: > > > On Thu, Dec 3, 2020 at 6:49 AM Dumitru Ceara > wrote: >> >> The IDL might decide to resend pending monitor condition changes >> implicitly (one such case is at reconnect) in which case the client >> (ovn-controller) might end

[ovs-dev] [PATCH] ovsdb-idl: Fix expected condition seqno when changes are pending.

Commit 17f22fe46142 tried to address this but only covered some of the cases. The correct way to report the expected seqno is to take into account if there already is a condition change that was requested to the server but not acked yet. In that case, the new condition change request will be

Re: [ovs-dev] [PATCH ovn 0/4] introduce BFD support in ovn-controller

> > On Wed, Nov 25, 2020 at 3:30 PM Lorenzo Bianconi > wrote: > > > > Introduce BFD protocol in ovn-controller according to RFC5880 [0] > > We added BFD implementation in ovn since layered protocols usually request > > to > > enable it on ovn entities (e.g. logical router ports) while ovs > >

Re: [ovs-dev] [PATCH ovn 0/4] introduce BFD support in ovn-controller

On Wed, Nov 25, 2020 at 3:30 PM Lorenzo Bianconi wrote: > > Introduce BFD protocol in ovn-controller according to RFC5880 [0] > We added BFD implementation in ovn since layered protocols usually request to > enable it on ovn entities (e.g. logical router ports) while ovs implementation > relies

[ovs-dev] [PATCH net] net: openvswitch: fix TTL decrement exception action execution

Currently, the exception actions are not processed correctly as the wrong dataset is passed. This change fixes this, including the misleading comment. In addition, a check was added to make sure we work on an IPv4 packet, and not just assume if it's not IPv6 it's IPv4. Small cleanup which

Re: [ovs-dev] [PATCH net] openvswitch: fix error return code in validate_and_copy_dec_ttl()

On 4 Dec 2020, at 12:43, Wang Hai wrote: Fix to return a negative error code from the error handling case instead of 0, as done elsewhere in this function. Changing 'return start' to 'return action_start' can fix this bug. Fixes: 69929d4c49e1 ("net: openvswitch: fix TTL decrement action

[ovs-dev] [PATCH net] openvswitch: fix error return code in validate_and_copy_dec_ttl()

Fix to return a negative error code from the error handling case instead of 0, as done elsewhere in this function. Changing 'return start' to 'return action_start' can fix this bug. Fixes: 69929d4c49e1 ("net: openvswitch: fix TTL decrement action netlink message format") Reported-by: Hulk Robot

Re: [ovs-dev] [PATCH v7 11/11] netdev-offload-tc: Add offload support for sFlow

On 4 Dec 2020, at 6:34, Chris Mi wrote: On 12/2/2020 11:08 PM, Eelco Chaudron wrote: On 30 Nov 2020, at 5:48, Chris Mi wrote: Create a unique group ID to map the sFlow info when offloading sFlow action to TC. When showing the offloaded datapath flows, translate the group ID from TC

Re: [ovs-dev] [PATCH ovn] ovn-controller: Relax get_nb_cfg() condition seqno check.

On Thu, Dec 3, 2020 at 6:49 AM Dumitru Ceara wrote: > > The IDL might decide to resend pending monitor condition changes > implicitly (one such case is at reconnect) in which case the client > (ovn-controller) might end waiting for a condition seqno that has > already been satisfied. > > In order