Re: [ovs-dev] [PATCH ovn v2 2/3] northd.c: Lock to protect against possible od->group corruption.

On Mon, Oct 4, 2021 at 5:29 AM Anton Ivanov wrote: > > > On 03/10/2021 23:45, Han Zhou wrote: > > When parallel build is used, od->group can be updated by threads outside > > of the function do_ovn_lflow_add_pd (for lb related flow building). So > > use the function

Re: [ovs-dev] [PATCH ovn v2 1/3] northd.c: Remove redundant condition check in ovn_dp_group_add_with_reference().

On Mon, Oct 4, 2021 at 5:29 AM Anton Ivanov wrote: > > > On 03/10/2021 23:45, Han Zhou wrote: > > Signed-off-by: Han Zhou > > --- > > northd/northd.c | 2 +- > > 1 file changed, 1 insertion(+), 1 deletion(-) > > > > diff --git a/northd/northd.c b/northd/northd.c > > index cf2467fe1..027c5b170

Re: [ovs-dev] [PATCH ovn v5 10/12] lib: Add infrastructure for plug providers.

On Tue, Sep 28, 2021 at 8:58 AM Frode Nordahl wrote: > > New lib/plug-provider module contains the infrastructure for > registering plug provider classes which may be hosted inside or > outside the core OVN repository. > > New controller/plug module adds internal interface for interacting > with

Re: [ovs-dev] [PATCH ovn v6 09/12] tests: Use built objects for unit test deps

On Tue, Sep 28, 2021 at 11:54 AM Frode Nordahl wrote: > > Currently the tests/automake.mk adds source files from the project > as needed and rebuilds these for the unit test programs. > > Use the already built objects instead. > > Signed-off-by: Frode Nordahl Acked-by: Numan Siddique Numan >

Re: [ovs-dev] [PATCH ovn v6 08/12] ovn-controller: Move shared functions to ovn-util

On Tue, Sep 28, 2021 at 11:54 AM Frode Nordahl wrote: > > The `get_tunnel_type` (and associated enum) and `get_bridge` > functions are used in modules other than the `ovn-controller` > module. > > Since the `ovn-controller` module also has the `main` function > this organization makes it hard or

Re: [ovs-dev] [PATCH ovn v6 07/12] binding: Make local_binding data structure public

On Tue, Sep 28, 2021 at 11:54 AM Frode Nordahl wrote: > > The binding module maintains a shash of lports bound to the local > chassis. Other modules have interest in the same data for lookup, > and it would be wasteful to reimplement the same data structure > elsewhere. > > The incremental

Re: [ovs-dev] [PATCH ovn v6 05/12] patch: Consume ovsport functions.

On Tue, Sep 28, 2021 at 11:54 AM Frode Nordahl wrote: > > Make use of the common functions for maintaining OVS ports. > > Signed-off-by: Frode Nordahl Acked-by: Numan Siddique Numan > --- > controller/patch.c | 39 ++- > 1 file changed, 6 insertions(+),

Re: [ovs-dev] [PATCH ovn v6 04/12] controller: Move OVS port functions to new module.

On Tue, Sep 28, 2021 at 11:54 AM Frode Nordahl wrote: > > Up until now the controller patch module has been the only > consumer of functions to maintain OVS ports and interfaces. > > With the introduction of infrastructure for plugging providers > these functions will also be consumed by the

Re: [ovs-dev] [PATCH ovn v6 01/12] ovn-sb: Add requested_chassis column to Port_Binding.

On Tue, Sep 28, 2021 at 11:54 AM Frode Nordahl wrote: > > To allow for ovn-controller to efficiently monitor all > Port_Binding records destined to it, we add a new > requested_chassis column with weakRef to the Chassis table. > > The ovn-controller can monitor this column and only process >

Re: [ovs-dev] [PATCH ovn v6 02/12] test: Fix options:requested-chassis with hostname

On Tue, Sep 28, 2021 at 11:54 AM Frode Nordahl wrote: > > This test currently passes, but is broken in two ways. > > 1) The `fetch_column` helper should be used to retrieve the value >of hostname, not `fetch` wich results in a "fetch: command not >found" error which is currently not

Re: [ovs-dev] [PATCH ovn v6 03/12] controller: Make use of Port_Binding:requested_chassis

On Tue, Sep 28, 2021 at 11:54 AM Frode Nordahl wrote: > > Improve the efficiency of the requested-chassis feature by using > the new Southbound Port_Binding:requested_chassis column instead > of each chassis performing string comparison for every > Port_Binding record processed. > >

Re: [ovs-dev] [PATCH ovn] spec: add tcpdump to BuildRequires

On Sat, Sep 18, 2021 at 9:54 AM Vladislav Odintsov wrote: > > tcpdump is used in tests. When RPM package is built it > appropriate test fails if tcpdump package is not installed. > > This commit fixes this issue by adding a new conditional > BuildRequires: tcpdump to specfile. > > Signed-off-by:

Re: [ovs-dev] [PATCH v4 ovn] Fix basic multicast flows for vxlan (non-vtep) tunnels

On Mon, Oct 4, 2021 at 12:37 PM Ihar Hrachyshka wrote: > > The 15-bit port key range used for multicast groups can't be covered > by 12-bit key space available for port keys in VXLAN. To make > multicast keys work, we have to transform 16-bit multicast port keys > to 12-bit keys before fanning

Re: [ovs-dev] Issue with one of your old patches, commit c645550b

This code in odp-util is basically unmaintainable. I've never been able to figure out how to properly test or reason about it. It's a huge reason that I want to get rid of the kernel module. On Mon, Oct 04, 2021 at 11:49:41AM +0200, Eelco Chaudron wrote: > Hi Ben, > > Any chance you can take a

Re: [ovs-dev] [PATCH ovn] northd: Change the default value of ignore_lsp_down to true.

On Mon, Oct 4, 2021 at 4:56 PM Mark Michelson wrote: > > On 10/4/21 4:10 PM, Han Zhou wrote: > > > > > > > > On Mon, Oct 4, 2021 at 11:30 AM Numan Siddique > > wrote: > > > > > > On Mon, Oct 4, 2021 at 1:37 PM Han Zhou > > wrote: > > > > > > > >

Re: [ovs-dev] [PATCH v5 3/3] conntrack: limit port clash resolution attempts

not a full review, but see some questions/remarks below. we...@ucloud.cn writes: > From: wenxu > > In case almost or all available ports are taken, clash resolution can > take a very long time, resulting in pmd hang in conntrack. > > This can happen when many to-be-natted hosts connect to same

Re: [ovs-dev] [PATCH v5 2/3] conntrack: select correct sport range for well-known origin sport

we...@ucloud.cn writes: > From: wenxu > > Like the kernel datapath. The sport nat range for well-konwn origin > sport should limit in the well-known ports. > > Signed-off-by: wenxu > --- LGTM, Acked-by: Paolo Valerio ___ dev mailing list

[ovs-dev] [ovn] bug: ovn-northd was blocked for changes handling

Hi, I’ve faced with a next issue using latest OVN (master branch) with OVN-IC enabled: ovn-northd CPU utilisation was at ~70-80% and ovsdb-server serving OVN_Southbound DB was also under heavy load. In ovn-northd.log file there were warnings: 2021-10-01T11:14:43.548Z|18845|northd|INFO|deleting

Re: [ovs-dev] [PATCH ovn] northd: Change the default value of ignore_lsp_down to true.

On 10/4/21 4:10 PM, Han Zhou wrote: On Mon, Oct 4, 2021 at 11:30 AM Numan Siddique > wrote: > > On Mon, Oct 4, 2021 at 1:37 PM Han Zhou > wrote: > > > > On Mon, Oct 4, 2021 at 3:10 AM Mark Gray > wrote: > > >

Re: [ovs-dev] [PATCH ovn] northd: Change the default value of ignore_lsp_down to true.

On Mon, Oct 4, 2021 at 11:30 AM Numan Siddique wrote: > > On Mon, Oct 4, 2021 at 1:37 PM Han Zhou wrote: > > > > On Mon, Oct 4, 2021 at 3:10 AM Mark Gray wrote: > > > > > > On 02/10/2021 08:11, Han Zhou wrote: > > > > The current default behavior is that ARP responder flows for VIFs are > > > >

Re: [ovs-dev] [PATCH v3 ovn] Enforce datapath and port key constraints in vxlan mode

On Thu, Sep 30, 2021 at 6:05 PM Ihar Hrachyshka wrote: > > With vxlan enabled for in-cluster communication, the ranges available > for port and datapath keys are limited to 12 bits (including multigroup > keys). (The default range is 16 bit long.) > > This means that OVN should avoid allocating,

Re: [ovs-dev] [PATCH ovn] northd: Change the default value of ignore_lsp_down to true.

On Mon, Oct 4, 2021 at 1:37 PM Han Zhou wrote: > > On Mon, Oct 4, 2021 at 3:10 AM Mark Gray wrote: > > > > On 02/10/2021 08:11, Han Zhou wrote: > > > The current default behavior is that ARP responder flows for VIFs are > > > added by northd after the port-binding state is UP, which creates more

Re: [ovs-dev] [PATCH ovn] controller: configure only matching encaps between chassis

On Mon, Sep 20, 2021 at 6:16 AM Dumitru Ceara wrote: > > On 9/16/21 7:30 PM, Vladislav Odintsov wrote: > > Previously tunnels encap on a chassis was based on the remote chassis > > "best" encap type. Suppose we have 2 chassis: one with STT, another with > > GENEVE. In this case on chassis 1 there

Re: [ovs-dev] [PATCH branch-21.09 1/2] Set release date for 21.09.0.

On 10/4/21 12:53 PM, Numan Siddique wrote: On Fri, Oct 1, 2021 at 3:11 PM Mark Michelson wrote: Signed-off-by: Mark Michelson For both the patches : Acked-by: Numan Siddique Numan Thank you Numan. I pushed the first commit, created the 21.09.0 tag, then pushed the second commit.

Re: [ovs-dev] [PATCH ovn] northd: Change the default value of ignore_lsp_down to true.

On Mon, Oct 4, 2021 at 3:10 AM Mark Gray wrote: > > On 02/10/2021 08:11, Han Zhou wrote: > > The current default behavior is that ARP responder flows for VIFs are > > added by northd after the port-binding state is UP, which creates more > > trouble than benefit in most use cases. To make the

Re: [ovs-dev] [PATCH ovn v2 3/3] northd.c: Optimize parallel build performance with hash based locks.

On Mon, Oct 4, 2021 at 9:57 AM Anton Ivanov wrote: > > > On 04/10/2021 16:58, Han Zhou wrote: > > > > On Mon, Oct 4, 2021 at 2:31 AM Anton Ivanov < anton.iva...@cambridgegreys.com> wrote: > > > > > > On 03/10/2021 23:45, Han Zhou wrote: > > > The current implementation of parallel build in northd

Re: [ovs-dev] [PATCH ovn v2 3/3] northd.c: Optimize parallel build performance with hash based locks.

On 04/10/2021 16:58, Han Zhou wrote: On Mon, Oct 4, 2021 at 2:31 AM Anton Ivanov mailto:anton.iva...@cambridgegreys.com>> wrote: > > > On 03/10/2021 23:45, Han Zhou wrote: > > The current implementation of parallel build in northd with dp-groups > > enabled results in bad performance when the

Re: [ovs-dev] [PATCH branch-21.09 1/2] Set release date for 21.09.0.

On Fri, Oct 1, 2021 at 3:11 PM Mark Michelson wrote: > > Signed-off-by: Mark Michelson For both the patches : Acked-by: Numan Siddique Numan > --- > NEWS | 2 +- > debian/changelog | 2 +- > 2 files changed, 2 insertions(+), 2 deletions(-) > > diff --git a/NEWS b/NEWS > index

[ovs-dev] [PATCH v4 ovn] Fix basic multicast flows for vxlan (non-vtep) tunnels

The 15-bit port key range used for multicast groups can't be covered by 12-bit key space available for port keys in VXLAN. To make multicast keys work, we have to transform 16-bit multicast port keys to 12-bit keys before fanning out packets through VXLAN tunnels. Otherwise significant bits are

Re: [ovs-dev] [PATCH ovn v4 3/4] ic: add support for routing tables in adv/learn routes

On Sun, Sep 19, 2021 at 5:24 PM Vladislav Odintsov wrote: > > Previously support for multiple routing tables was added > to northd code. > This commit expands support for multiple routing tables > by adding support of advertising and learning routes with > their routing table information. > > To

Re: [ovs-dev] [PATCH ovn v4 2/4] northd, utils: support for RouteTables in LRs

On Sun, Sep 19, 2021 at 5:24 PM Vladislav Odintsov wrote: > > Signed-off-by: Vladislav Odintsov > --- > northd/northd.c | 159 --- > northd/ovn-northd.8.xml | 63 -- > ovn-nb.ovsschema| 5 +- > ovn-nb.xml | 30 +++ > tests/ovn-ic.at | 4

Re: [ovs-dev] [PATCH ovn v4 1/4] ic: process only local port_bindings

On Sun, Sep 19, 2021 at 5:24 PM Vladislav Odintsov wrote: > > This commit adds a small optimization by utilizing ovsdb_index > to iterate over port_bindings. > Prior to this change each iteration checked availability_zone > and continued processing only if port_binding belons to local AZ. > > Now

Re: [ovs-dev] [PATCH ovn 2/2] ovn-parallel-hmap: Remove the unused mutex in worker_control.

On Mon, Oct 4, 2021 at 2:35 AM Anton Ivanov wrote: > > > On 04/10/2021 02:37, Han Zhou wrote: > > It is not used, and seems not going to be needed, so remove it. > > > > Signed-off-by: Han Zhou > > --- > > lib/ovn-parallel-hmap.c | 1 - > > lib/ovn-parallel-hmap.h | 1 - > > 2 files changed,

Re: [ovs-dev] [PATCH ovn v2 3/3] northd.c: Optimize parallel build performance with hash based locks.

On Mon, Oct 4, 2021 at 2:31 AM Anton Ivanov wrote: > > > On 03/10/2021 23:45, Han Zhou wrote: > > The current implementation of parallel build in northd with dp-groups > > enabled results in bad performance when the below assumption is not > > true: > > > > * 3. Most RW ops are not flow adds,

[ovs-dev] [ovn] 21.09.0 release not ready yet?

Hi, As far I know OVN 21.09.0 release previously was scheduled for 01.10.2021, but wasn't released yet. I’m wonder if my patch series is still scheduled to be in this release: https://patchwork.ozlabs.org/project/ovn/list/?series=263029 Also, it’d be good if this patch also can be in this

Re: [ovs-dev] [PATCH v2] checkpatch: check if some tags are wrongly written

On Fri, Sep 10, 2021 at 11:08 AM Timothy Redaelli wrote: > > Currently, there are some patches with the tags wrongly written (with > space instead of dash ) and this may prevent some automatic system or CI > to detect them correctly. > > This commit adds a check in checkpatch to be sure the tag

Re: [ovs-dev] Encap & decap actions for MPLS Packet type

On Thu, Sep 30, 2021 at 02:18:42PM -0400, Aaron Conole wrote: > Martin Varghese writes: > > > acon...@redhat.com,echau...@redhat.com > > Cc: d...@openvswitch.org echau...@redhat.com > > Bcc: > > Subject: Re: [ovs-dev] [PATCH v6] Encap & Decap actions for MPLS packet > > type. > > Reply-To: >

Re: [ovs-dev] [PATCH ovn] northd: Change the default value of ignore_lsp_down to true.

On 02/10/2021 08:11, Han Zhou wrote: > The current default behavior is that ARP responder flows for VIFs are > added by northd after the port-binding state is UP, which creates more > trouble than benefit in most use cases. To make the default behavior > desirable for majority of the use cases,

Re: [ovs-dev] Issue with one of your old patches, commit c645550b

Hi Ben, Any chance you can take a peek at the email below? I hope you can remember this from 2018 :) Thanks, Eelco On 15 Sep 2021, at 15:30, Eelco Chaudron wrote: > Hi Ben, > > The mentioned fix below is causing a regression, actually two related ones: > > 1) If IGMP snooping is disabled,

Re: [ovs-dev] [PATCH ovn 2/2] ovn-parallel-hmap: Remove the unused mutex in worker_control.

On 04/10/2021 02:37, Han Zhou wrote: It is not used, and seems not going to be needed, so remove it. Signed-off-by: Han Zhou --- lib/ovn-parallel-hmap.c | 1 - lib/ovn-parallel-hmap.h | 1 - 2 files changed, 2 deletions(-) diff --git a/lib/ovn-parallel-hmap.c b/lib/ovn-parallel-hmap.c

Re: [ovs-dev] [PATCH ovn 1/2] ovn-parallel-hmap.h: Minor fixes for hashrow_lock.

On 04/10/2021 02:37, Han Zhou wrote: Although not used currently, it is better to fix: 1. The type of the mask field should be the same as hmap->mask: size_t. 2. Calculating the index is better to use & instead of %. Signed-off-by: Han Zhou --- lib/ovn-parallel-hmap.h | 6 +++--- 1 file

Re: [ovs-dev] [PATCH ovn v2 3/3] northd.c: Optimize parallel build performance with hash based locks.

On 03/10/2021 23:45, Han Zhou wrote: The current implementation of parallel build in northd with dp-groups enabled results in bad performance when the below assumption is not true: * 3. Most RW ops are not flow adds, but changes to the * od groups. In fact most (if not all) of the ovn

Re: [ovs-dev] [PATCH ovn v2 2/3] northd.c: Lock to protect against possible od->group corruption.

On 03/10/2021 23:45, Han Zhou wrote: When parallel build is used, od->group can be updated by threads outside of the function do_ovn_lflow_add_pd (for lb related flow building). So use the function ovn_dp_group_add_with_reference() to update it in function do_ovn_lflow_add() when it is not a

Re: [ovs-dev] [PATCH ovn v2 1/3] northd.c: Remove redundant condition check in ovn_dp_group_add_with_reference().

On 03/10/2021 23:45, Han Zhou wrote: Signed-off-by: Han Zhou --- northd/northd.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/northd/northd.c b/northd/northd.c index cf2467fe1..027c5b170 100644 --- a/northd/northd.c +++ b/northd/northd.c @@ -4475,7 +4475,7 @@