Re: [ovs-dev] [PATCH ovn] ovn-controller: Fix the missing flows with monitor-all set to True

On Thu, Jul 16, 2020 at 1:34 AM Dumitru Ceara wrote: > > On 7/15/20 7:56 PM, num...@ovn.org wrote: > > From: Numan Siddique > > > > When the config ovn-monitor-all is set to True, then ovn-controller is not > > programming the flows completely when it binds the logical port of > > a logical

[ovs-dev] [ovs-dev 1/1] datapath-windows: Reset ct_mark/ct_label to support ALG

The ct_mark/ct_label setting on related connection keep the same behavior with Linux datapath. If one CT entry has parent/master entry, its ct_mark and ct_label should inherit from the corresponding part of parent/master entry at initialization. Signed-off-by: Jinjun Gao ---

Re: [ovs-dev] [PATCH ovs-dev v1] netdev-offload: Flush offloaded rules when ports removed

On Thu, Jun 18, 2020 at 7:41 PM Tonghao Zhang wrote: > > On Thu, Jun 18, 2020 at 7:18 PM Simon Horman > wrote: > > > > On Tue, Jun 16, 2020 at 04:25:21PM +0800, Tonghao Zhang wrote: > > > On Mon, Jun 15, 2020 at 7:13 PM Ilya Maximets wrote: > > > > > > > > On 6/14/20 1:40 PM, Tonghao Zhang

Re: [ovs-dev] [PATCH v3 1/4] netdev-afxdp: Enable loading XDP program.

On Mon, Jun 29, 2020 at 8:30 AM Toshiaki Makita wrote: > > From: William Tu > > Now netdev-afxdp always forwards all packets to userspace because > it is using libbpf's default XDP program, see 'xsk_load_xdp_prog'. > There are some cases when users want to keep packets in kernel instead > of

Re: [ovs-dev] [PATCH v3 1/4] netdev-afxdp: Enable loading XDP program.

On Mon, Jun 29, 2020 at 8:30 AM Toshiaki Makita wrote: > > From: William Tu > > Now netdev-afxdp always forwards all packets to userspace because > it is using libbpf's default XDP program, see 'xsk_load_xdp_prog'. > There are some cases when users want to keep packets in kernel instead > of

Re: [ovs-dev] [PATCH 00/25] netdev datapath vxlan offload

Hi Eli, I'm having another question, thanks in advance! On Mon, Jan 20, 2020 at 7:09 AM Eli Britstein wrote: > > In the netdev datapath, packets arriving over a tunnel are processed by > more than one flow. For example a packet that should be decapsulated and > forwarded to another port is

[ovs-dev] Post pandemia - Guía integral para la reactivación

Buenos día Quise aprovechar la oportunidad de hacerte una invitación para tomar nuestro curso: Nombre: Guía integral para la reactivación post-pandemia Horario: de 10:00 am a 17:00 pm. ¿Cuándo?: 06 de Agosto Formato: En línea con interacción en vivo. Lugar: En Vivo desde su computadora

Re: [ovs-dev] [PATCH v6] Bareudp Tunnel Support

On 7/10/20 7:20 PM, Ilya Maximets wrote: > On 6/29/20 3:31 PM, Martin Varghese wrote: >> From: Martin Varghese >> >> There are various L3 encapsulation standards using UDP being discussed to >> leverage the UDP based load balancing capability of different networks. >> MPLSoUDP (__

Re: [ovs-dev] [PATCH v2] dpif-netdev: Avoid deadlock with offloading during PMD thread deletion.

On 7/15/20 8:30 PM, Stokes, Ian wrote: >> >> On 15/07/2020 17:00, Ilya Maximets wrote: >>> Main thread will try to pause/stop all revalidators during datapath >>> reconfiguration via datapath purge callback (dp_purge_cb) while >>> holding 'dp->port_mutex'. And deadlock happens in case any of >>>

Re: [ovs-dev] [PATCH] rpm-fedora: Add missing dist library

On 7/15/20 5:55 PM, Gregory Rose wrote: > > On 7/15/2020 8:21 AM, Van Haaren, Harry wrote: >>> -Original Message- >>> From: Ilya Maximets >>> Sent: Wednesday, July 15, 2020 2:28 PM >>> To: Van Haaren, Harry ; Ilya Maximets >>> ; William Tu ; Gregory Rose >>> ; Stokes, Ian >>> Cc:

Re: [ovs-dev] [PATCH ovn] ovn-controller: Fix the missing flows with monitor-all set to True

On 7/15/20 7:56 PM, num...@ovn.org wrote: > From: Numan Siddique > > When the config ovn-monitor-all is set to True, then ovn-controller is not > programming the flows completely when it binds the logical port of > a logical switch until a full recompute is triggered. > > This issue was

Re: [ovs-dev] [PATCH v2] dpif-netdev: Avoid deadlock with offloading during PMD thread deletion.

> > On 15/07/2020 17:00, Ilya Maximets wrote: > > Main thread will try to pause/stop all revalidators during datapath > > reconfiguration via datapath purge callback (dp_purge_cb) while > > holding 'dp->port_mutex'. And deadlock happens in case any of > > revalidator threads is already waiting

[ovs-dev] [PATCH ovn] ovn-controller: Fix the missing flows with monitor-all set to True

From: Numan Siddique When the config ovn-monitor-all is set to True, then ovn-controller is not programming the flows completely when it binds the logical port of a logical switch until a full recompute is triggered. This issue was introduced by the commit in Fixes - which added incremental

Re: [ovs-dev] [PATCH v6 0/8] netdev datapath: Partial action offload

Following my comment about possibility to double encap (https://patchwork.ozlabs.org/project/openvswitch/patch/20200629095020.8491-4-sriharsha.basavapa...@broadcom.com/), with this patch-set there are also scenarios in which double push-vlan or double pop-vlan will occur. For example there is

[ovs-dev] [PATCH ovn RFC v2 12/12] Move egress to a separate rule and finalize

From: Anton Ivanov 1. Move egress to a separate rule function 2. Clean-up build_lrouter_flows from variables which are not used Signed-off-by: Anton Ivanov --- northd/ovn-northd.c | 107 1 file changed, 58 insertions(+), 49 deletions(-) diff --git

[ovs-dev] [PATCH ovn RFC v2 10/12] Move GW redirect rules to a new function

From: Anton Ivanov Signed-off-by: Anton Ivanov --- northd/ovn-northd.c | 80 ++--- 1 file changed, 46 insertions(+), 34 deletions(-) diff --git a/northd/ovn-northd.c b/northd/ovn-northd.c index 0b4926a75..81a6407a9 100644 --- a/northd/ovn-northd.c +++

[ovs-dev] [PATCH ovn RFC v2 11/12] Move ARP request to a separate rule function

From: Anton Ivanov Signed-off-by: Anton Ivanov --- northd/ovn-northd.c | 139 1 file changed, 75 insertions(+), 64 deletions(-) diff --git a/northd/ovn-northd.c b/northd/ovn-northd.c index 81a6407a9..6b7a57c41 100644 --- a/northd/ovn-northd.c +++

Re: [ovs-dev] [PATCH v2] dpif-netdev: Avoid deadlock with offloading during PMD thread deletion.

Bleep bloop. Greetings Ilya Maximets, I am a robot and I have tried out your patch. Thanks for your contribution. I encountered some error that I wasn't expecting. See the details below. checkpatch: WARNING: Comment with 'xxx' marker #172 FILE: lib/dpif-netdev.c:3365: * XXX: Main thread

Re: [ovs-dev] RFC DEMO of using parallel processing in OVN

On 14/07/2020 17:06, Anton Ivanov wrote: On 14/07/2020 16:27, Dumitru Ceara wrote: On 7/10/20 10:39 AM, anton.iva...@cambridgegreys.com wrote: Hi all, This is a series of patches to demo the use in OVN of the parallel processing of hashes patchset which I submitted to OVS. The OVS patch

[ovs-dev] [PATCH ovn RFC v2 09/12] Move packet size rules to a separate function

From: Anton Ivanov Signed-off-by: Anton Ivanov --- northd/ovn-northd.c | 752 +++- 1 file changed, 397 insertions(+), 355 deletions(-) diff --git a/northd/ovn-northd.c b/northd/ovn-northd.c index 7932ab9e3..0b4926a75 100644 --- a/northd/ovn-northd.c +++

[ovs-dev] [PATCH ovn RFC v2 07/12] Move out multicast lookup into a function

From: Anton Ivanov Signed-off-by: Anton Ivanov --- northd/ovn-northd.c | 128 1 file changed, 69 insertions(+), 59 deletions(-) diff --git a/northd/ovn-northd.c b/northd/ovn-northd.c index c6e58576a..61f9b4ab8 100644 --- a/northd/ovn-northd.c +++

[ovs-dev] Path Series to prepare ovn-northd for parallelization

Hi All, This is the build_lrouter_flows portion of the patchset to prepare ovn-northd for parallelisation. I will split the patches into smaller ones this time and 3 series: 1. build_lrouter_flows() reorg 2. build_lswitch_flows() reorg 3. actual parallelization I have tried to preserve

[ovs-dev] [PATCH ovn RFC v2 08/12] Move Logical router ingress policy to a function

From: Anton Ivanov Signed-off-by: Anton Ivanov --- northd/ovn-northd.c | 52 ++--- 1 file changed, 30 insertions(+), 22 deletions(-) diff --git a/northd/ovn-northd.c b/northd/ovn-northd.c index 61f9b4ab8..7932ab9e3 100644 --- a/northd/ovn-northd.c +++

[ovs-dev] [PATCH ovn RFC v2 01/12] Move out Table 0 operations to functions

From: Anton Ivanov Signed-off-by: Anton Ivanov --- northd/ovn-northd.c | 71 - 1 file changed, 38 insertions(+), 33 deletions(-) diff --git a/northd/ovn-northd.c b/northd/ovn-northd.c index 192198272..4374b88a6 100644 --- a/northd/ovn-northd.c +++

[ovs-dev] [PATCH ovn RFC v2 05/12] Move Logical router ingress table IP_ROUTING to a function

From: Anton Ivanov Signed-off-by: Anton Ivanov --- northd/ovn-northd.c | 97 ++--- 1 file changed, 56 insertions(+), 41 deletions(-) diff --git a/northd/ovn-northd.c b/northd/ovn-northd.c index 95c17af98..807bf23ec 100644 --- a/northd/ovn-northd.c +++

[ovs-dev] [PATCH ovn RFC v2 04/12] Move NAT and Load Balancing to a separate function

From: Anton Ivanov Signed-off-by: Anton Ivanov --- northd/ovn-northd.c | 144 ++-- 1 file changed, 87 insertions(+), 57 deletions(-) diff --git a/northd/ovn-northd.c b/northd/ovn-northd.c index 8a0e28040..95c17af98 100644 --- a/northd/ovn-northd.c +++

[ovs-dev] [PATCH ovn RFC v2 06/12] Move static route datapath mapping to a function

From: Anton Ivanov Signed-off-by: Anton Ivanov --- northd/ovn-northd.c | 96 - 1 file changed, 51 insertions(+), 45 deletions(-) diff --git a/northd/ovn-northd.c b/northd/ovn-northd.c index 807bf23ec..c6e58576a 100644 --- a/northd/ovn-northd.c +++

[ovs-dev] [PATCH ovn RFC v2 03/12] Move out Table 3 operations to functions

From: Anton Ivanov Signed-off-by: Anton Ivanov --- northd/ovn-northd.c | 143 ++-- 1 file changed, 84 insertions(+), 59 deletions(-) diff --git a/northd/ovn-northd.c b/northd/ovn-northd.c index 1ec530854..8a0e28040 100644 --- a/northd/ovn-northd.c +++

[ovs-dev] [PATCH ovn RFC v2 02/12] Move out Table 1 operations to functions

From: Anton Ivanov The functions are numbered in 10 step intervals to minimize renumbering if new pipeline stages are added in the future. Signed-off-by: Anton Ivanov --- northd/ovn-northd.c | 75 +++-- 1 file changed, 46 insertions(+), 29 deletions(-)

Re: [ovs-dev] [PATCH v2] dpif-netdev: Avoid deadlock with offloading during PMD thread deletion.

On 15/07/2020 17:00, Ilya Maximets wrote: > Main thread will try to pause/stop all revalidators during datapath > reconfiguration via datapath purge callback (dp_purge_cb) while > holding 'dp->port_mutex'. And deadlock happens in case any of > revalidator threads is already waiting on

[ovs-dev] [PATCH v2] dpif-netdev: Avoid deadlock with offloading during PMD thread deletion.

Main thread will try to pause/stop all revalidators during datapath reconfiguration via datapath purge callback (dp_purge_cb) while holding 'dp->port_mutex'. And deadlock happens in case any of revalidator threads is already waiting on 'dp->port_mutex' while dumping offloaded flows:

Re: [ovs-dev] [PATCH] rpm-fedora: Add missing dist library

On 7/15/2020 8:21 AM, Van Haaren, Harry wrote: -Original Message- From: Ilya Maximets Sent: Wednesday, July 15, 2020 2:28 PM To: Van Haaren, Harry ; Ilya Maximets ; William Tu ; Gregory Rose ; Stokes, Ian Cc: d...@openvswitch.org Subject: Re: [PATCH] rpm-fedora: Add missing dist

Re: [ovs-dev] [PATCH v3 0/4] XDP offload using flow API provider

Hi Toshiaki, Thanks for the interesting patch series! I haven't finished reviewing, but I went through a couple of steps to make sure it works in my environment. Comments below: On Tue, Jun 30, 2020 at 12:30:29AM +0900, Toshiaki Makita wrote: > This patch adds an XDP-based flow cache using the

Re: [ovs-dev] [PATCH] rpm-fedora: Add missing dist library

> -Original Message- > From: Ilya Maximets > Sent: Wednesday, July 15, 2020 2:28 PM > To: Van Haaren, Harry ; Ilya Maximets > ; William Tu ; Gregory Rose > ; Stokes, Ian > Cc: d...@openvswitch.org > Subject: Re: [PATCH] rpm-fedora: Add missing dist library > > On 7/15/20 2:36 PM, Van

Re: [ovs-dev] [PATCH] rpm-fedora: Add missing dist library

On 7/14/2020 4:33 PM, William Tu wrote: On Tue, Jul 14, 2020 at 03:40:39PM -0700, Gregory Rose wrote: On 7/14/2020 1:49 PM, Greg Rose wrote: libopenvswitchavx512.a is needed for the fedora rpm spec. Signed-off-by: Greg Rose --- rhel/openvswitch-fedora.spec.in | 1 + 1 file changed, 1

Re: [ovs-dev] [PATCH v3] ovsdb-tool: Add a db consistency check to the ovsdb-tool check-cluster command

On 7/15/20 4:22 PM, Federico Paolinelli wrote: > > > On Tue, Jul 14, 2020 at 10:22 AM Dumitru Ceara > wrote: > > On 7/10/20 10:18 AM, Federico Paolinelli wrote: > > There are some occurrences where the database ends up in an > inconsistent > > state.

Re: [ovs-dev] [PATCH net-next] net: openvswitch: reorder masks array based on usage

On Wed, 2020-07-15 at 14:09 +0200, Eelco Chaudron wrote: > This patch reorders the masks array every 4 seconds based on their > usage count. This greatly reduces the masks per packet hit, and > hence the overall performance. Especially in the OVS/OVN case for > OpenShift. > > Here are some

Re: [ovs-dev] [PATCH v3] ovsdb-tool: Add a db consistency check to the ovsdb-tool check-cluster command

On Tue, Jul 14, 2020 at 10:22 AM Dumitru Ceara wrote: > On 7/10/20 10:18 AM, Federico Paolinelli wrote: > > There are some occurrences where the database ends up in an inconsistent > > state. This happened in ovn-k8s and is described in [0]. > > Here we are adding a supported way to check that a

Re: [ovs-dev] [PATCH] rpm-fedora: Add missing dist library

On 7/15/20 2:36 PM, Van Haaren, Harry wrote: >> -Original Message- >> From: Ilya Maximets >> Sent: Wednesday, July 15, 2020 1:22 PM >> To: Van Haaren, Harry ; Ilya Maximets >> ; William Tu ; Gregory Rose >> ; Stokes, Ian >> Cc: d...@openvswitch.org >> Subject: Re: [PATCH] rpm-fedora: Add

Re: [ovs-dev] [PATCH] rhel: openvswitch-fedora.spec.in: Fix installed but not packaged

> -Original Message- > From: Roi Dayan > Sent: Wednesday, July 15, 2020 1:41 PM > To: d...@openvswitch.org > Cc: Van Haaren, Harry ; Ilya Maximets > ; Ben Pfaff ; Simon Horman > ; Roi Dayan > Subject: [PATCH] rhel: openvswitch-fedora.spec.in: Fix installed but not > packaged > > With

[ovs-dev] [PATCH] rhel: openvswitch-fedora.spec.in: Fix installed but not packaged

With the cited commit, we get an error from rpmbuild about installed but not packaged /usr/lib64/libopenvswitchavx512.a. Fix it by treating it as the other la files. Fixes: 352b6c7116cd ("dpif-lookup: add avx512 gather implementation.") Signed-off-by: Roi Dayan ---

Re: [ovs-dev] [PATCH] rpm-fedora: Add missing dist library

> -Original Message- > From: Ilya Maximets > Sent: Wednesday, July 15, 2020 1:22 PM > To: Van Haaren, Harry ; Ilya Maximets > ; William Tu ; Gregory Rose > ; Stokes, Ian > Cc: d...@openvswitch.org > Subject: Re: [PATCH] rpm-fedora: Add missing dist library > > On 7/15/20 1:45 PM, Van

Re: [ovs-dev] [PATCH] rpm-fedora: Add missing dist library

On 7/15/20 1:45 PM, Van Haaren, Harry wrote: >> -Original Message- >> From: Ilya Maximets >> Sent: Wednesday, July 15, 2020 10:16 AM >> To: William Tu ; Gregory Rose ; >> Van Haaren, Harry ; Stokes, Ian >> >> Cc: d...@openvswitch.org; Ilya Maximets ; >> i.maxim...@ovn.org >> Subject: Re:

[ovs-dev] [PATCH net-next] net: openvswitch: reorder masks array based on usage

This patch reorders the masks array every 4 seconds based on their usage count. This greatly reduces the masks per packet hit, and hence the overall performance. Especially in the OVS/OVN case for OpenShift. Here are some results from the OVS/OVN OpenShift test, which use 8 pods, each pod having

Re: [ovs-dev] [PATCH] rpm-fedora: Add missing dist library

> -Original Message- > From: Ilya Maximets > Sent: Wednesday, July 15, 2020 10:16 AM > To: William Tu ; Gregory Rose ; > Van Haaren, Harry ; Stokes, Ian > > Cc: d...@openvswitch.org; Ilya Maximets ; > i.maxim...@ovn.org > Subject: Re: [PATCH] rpm-fedora: Add missing dist library > > On

Re: [ovs-dev] [PATCH] dpif-netdev: Avoid deadlock with offloading during PMD thread deletion.

On 15/07/2020 11:02, Ilya Maximets wrote: > On 7/15/20 11:44 AM, Kevin Traynor wrote: >> On 14/07/2020 19:24, Ilya Maximets wrote: >>> On 7/14/20 7:52 PM, Ilya Maximets wrote: On 7/14/20 5:33 PM, Kevin Traynor wrote: > On 09/07/2020 16:19, Ilya Maximets wrote: >> @@ -3175,11 +3219,31

Re: [ovs-dev] RFC DEMO of using parallel processing in OVN

On 7/15/20 11:54 AM, Anton Ivanov wrote: > > On 15/07/2020 09:57, Dumitru Ceara wrote: >> On 7/14/20 6:06 PM, Anton Ivanov wrote: >>> On 14/07/2020 16:27, Dumitru Ceara wrote: On 7/10/20 10:39 AM, anton.iva...@cambridgegreys.com wrote: > Hi all, > > This is a series of patches to

Re: [ovs-dev] [PATCH] dpif-netdev: Avoid deadlock with offloading during PMD thread deletion.

On 7/15/20 11:44 AM, Kevin Traynor wrote: > On 14/07/2020 19:24, Ilya Maximets wrote: >> On 7/14/20 7:52 PM, Ilya Maximets wrote: >>> On 7/14/20 5:33 PM, Kevin Traynor wrote: On 09/07/2020 16:19, Ilya Maximets wrote: > @@ -3175,11 +3219,31 @@ dpif_netdev_get_flow_offload_status(const

Re: [ovs-dev] RFC DEMO of using parallel processing in OVN

On 15/07/2020 09:57, Dumitru Ceara wrote: On 7/14/20 6:06 PM, Anton Ivanov wrote: On 14/07/2020 16:27, Dumitru Ceara wrote: On 7/10/20 10:39 AM, anton.iva...@cambridgegreys.com wrote: Hi all, This is a series of patches to demo the use in OVN of the parallel processing of hashes patchset

Re: [ovs-dev] [PATCH] dpif-netdev: Avoid deadlock with offloading during PMD thread deletion.

On 14/07/2020 19:24, Ilya Maximets wrote: > On 7/14/20 7:52 PM, Ilya Maximets wrote: >> On 7/14/20 5:33 PM, Kevin Traynor wrote: >>> On 09/07/2020 16:19, Ilya Maximets wrote: Main thread will try to pause/stop all revalidators during datapath reconfiguration via datapath purge callback

Re: [ovs-dev] [PATCH 0/5] Stream Record/Replay.

On 7/14/20 4:56 PM, Ilya Maximets wrote: > On 7/14/20 3:34 PM, Dumitru Ceara wrote: >> On 6/30/20 3:24 AM, Ilya Maximets wrote: >>> This patch set adds new stream provider and other functionality in >>> order to record all the incoming data on all the steams (ssl, tcp, >>> unixctl) of openvswitch

Re: [ovs-dev] [PATCH] rpm-fedora: Add missing dist library

On 7/15/20 1:33 AM, William Tu wrote: > On Tue, Jul 14, 2020 at 03:40:39PM -0700, Gregory Rose wrote: >> >> >> On 7/14/2020 1:49 PM, Greg Rose wrote: >>> libopenvswitchavx512.a is needed for the fedora rpm spec. >>> >>> Signed-off-by: Greg Rose >>> --- >>> rhel/openvswitch-fedora.spec.in | 1 +

Re: [ovs-dev] RFC DEMO of using parallel processing in OVN

On 7/14/20 6:06 PM, Anton Ivanov wrote: > On 14/07/2020 16:27, Dumitru Ceara wrote: >> On 7/10/20 10:39 AM, anton.iva...@cambridgegreys.com wrote: >>> Hi all, >>> >>> This is a series of patches to demo the use in OVN of the parallel >>> processing of hashes patchset which I submitted to OVS. >>>