Re: [ovs-dev] [PATCH v17 3/8] dpif-offload-provider: Introduce dpif-offload-provider layer

Hi Eelco, Thanks for reviewing this patchset again. I will address them next week. Thanks, Chris On 11/5/2021 8:18 PM, Eelco Chaudron wrote: On 21 Oct 2021, at 10:00, Chris Mi wrote: Some offload actions require functionality that is not netdev based, but dpif. For example, sFlow action req

[ovs-dev] [PATCH RFC v2] netdev-dpdk: Add Windows support.

The patch adds OVS-DPDK supports on Windows. Motivation -- Currently OVS supports multiple datapath implementations: Linux kernel, Windows kernel, userspace with OVS-DPDK Linux, and HW offload. Adding any new feature to OVS datapath requires OS-specific expertise and usually ends up with f

Re: [ovs-dev] [PATCH] ovsdb: Don't let transaction history grow larger than the database.

On 10/23/21 08:40, Han Zhou wrote: > > > On Thu, Oct 14, 2021 at 7:58 AM Ilya Maximets > wrote: >> >> If user frequently changes a lot of rows in a database, transaction >> history could grow way larger than the database itself.  This wastes >> a lot of memory and also

Re: [ovs-dev] [PATCH ovn v10 0/4] Introduce infrastructure for VIF plug providers.

fre. 5. nov. 2021, 20:43 skrev Han Zhou : > > > On Fri, Nov 5, 2021 at 7:00 AM Frode Nordahl > wrote: > > > > Introduce infrastructure for VIF plug providers and add feature to > > ovn-controller to add and remove ports on the integration bridge as > > directed by CMS through Logical_Switch_Port

Re: [ovs-dev] [PATCH ovn v10 0/4] Introduce infrastructure for VIF plug providers.

On Fri, Nov 5, 2021 at 7:00 AM Frode Nordahl wrote: > > Introduce infrastructure for VIF plug providers and add feature to > ovn-controller to add and remove ports on the integration bridge as > directed by CMS through Logical_Switch_Port options. > > Traditionally it has been the CMSs responsibil

[ovs-dev] [PATCH v4] ovs-tcpdump: Improve performance with dummy interface

Currently the ovs-tcpdump utility creates a virtual tunnel to send packets to. This method functions perfectly fine, however, it can greatly impact performance of the monitored port. It has been reported to reduce packet throughput significantly. I was able to reproduce a reduction in throughput o

Re: [ovs-dev] [PATCH v2] Backup and remove existing DB when joining cluster

On 11/5/21 7:24 PM, Terry Wilson wrote: > ovsdb-tool join-cluster requires a remote addr, so the existing > code that tried to join a cluster without it when there was an > existing $DB_FILE would fail. > > Instead, if we are trying to specifically join a cluster and there > is an existing $DB_FIL

[ovs-dev] [PATCH v2] Backup and remove existing DB when joining cluster

ovsdb-tool join-cluster requires a remote addr, so the existing code that tried to join a cluster without it when there was an existing $DB_FILE would fail. Instead, if we are trying to specifically join a cluster and there is an existing $DB_FILE, back it up and remove the original before continu

[ovs-dev] [PATCH v2] utilities: Add another GDB macro for ovs-vswitchd

This commit adds a basic packet metadata macro to the already existing macros in ovs_gdb.py, ovs_dump_packets will print out information about one or more packets. It feeds packets into tcpdump, and the user can pass in tcpdump options to modify how packets are parsed or even write out packets to a

[ovs-dev] [PATCH ovn 2/2] northd: Use ct_(snat/dnat)_in_czone action for distributed routers.

From: Numan Siddique Make of use of these new actions for the distributed routers for NAT. These new actions ensure that both sNAT and dNAT happens in the same zone. This approach solves a couple of problems: - The datapath flows generated for external traffic which requires dNAT (N -> S)

[ovs-dev] [PATCH ovn 1/2] actions: Add new actions - ct_dnat_in_czone and ct_snat_in_czone.

From: Numan Siddique These actions are very similar to ct_dnat and ct_snat respectively with one difference. These new actions use the same zone id for natting. Upcoming patch will make use of these actions. Signed-off-by: Numan Siddique --- include/ovn/actions.h| 2 ++ include/ovn/

Re: [ovs-dev] [PATCH] tests: Really fix requested-chassis localport test

Bleep bloop. Greetings Frode Nordahl, 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. git-am: error: sha1 information is lacking or useless (tests/ovn.at). error: could not build fake ancestor

Re: [ovs-dev] [PATCH ovn v1 5/5] ovn_detrace.py: add helper to create CookieHandlers

On 10/21/21 11:10 AM, Adrian Moreno wrote: > That way external programs can use such handlers to create all the > supported handlers. > > Signed-off-by: Adrian Moreno > --- Looks good to me, thanks! Acked-by: Dumitru Ceara ___ dev mailing list d...

Re: [ovs-dev] [PATCH ovn v1 4/5] ovn-detrace: expose version as global variable

On 10/21/21 11:10 AM, Adrian Moreno wrote: > That way importers can check it to determine whether a particular > feature exists or not. > > Signed-off-by: Adrian Moreno > --- Looks good to me, thanks! Acked-by: Dumitru Ceara ___ dev mailing list d.

Re: [ovs-dev] [PATCH ovn v1 3/5] ovn-detrace: rename ovn-detrace to ovn_detrace.py

On 10/21/21 11:10 AM, Adrian Moreno wrote: > For external python programs to be able to load ovn-detrace as a python > module easily, it should end in .py and should have underscores. > > Move ovn-detrace to ovn_detrace.py and create a symlink with the old name > for backwards compatibility. > >

Re: [ovs-dev] [PATCH ovn v1 2/5] ovn-detrace: use configurable printer object

On 10/21/21 11:10 AM, Adrian Moreno wrote: > Intead of declaring global printing functions, use a configurable > printer object that has to declare print_p and print_h. > > That way external programs that may want to use ovn-detrace can obtain > the information provided by ovn-detrace by just re-i

Re: [ovs-dev] [PATCH ovn v1 1/5] manpages.mk: fix dependencies path

On 10/21/21 11:10 AM, Adrian Moreno wrote: > Currently, if "make" is run after the project is built, the root > manpage (ovn-detrace.1) is rebuilt unnecessarily. > > The reason is that its dependencies are wrong: files such as > lib/common.man or lib/ovs.tmac do not exist in the project's root > p

Re: [ovs-dev] [PATCH] tests: Really fix requested-chassis localport test

On Fri, Nov 5, 2021 at 5:06 PM Frode Nordahl wrote: > > Use OVS_WAIT_UNTIL instead of AT_CHECK, which should avoid race > conditions where the check is ran before the tables are populated. > > Fixes: ad77db239d9a ("controller: Make use of > Port_Binding:requested_chassis.") > Signed-off-by: Frode

[ovs-dev] [PATCH] tests: Really fix requested-chassis localport test

Use OVS_WAIT_UNTIL instead of AT_CHECK, which should avoid race conditions where the check is ran before the tables are populated. Fixes: ad77db239d9a ("controller: Make use of Port_Binding:requested_chassis.") Signed-off-by: Frode Nordahl --- tests/ovn.at | 30 -- 1

Re: [ovs-dev] [PATCH v2] python: idl: Avoid pre-allocating column defaults

On 11/5/21 3:13 PM, Terry Wilson wrote: > Many python implementations pre-allocate space for multiple > objects in empty dicts and lists. Using a custom dict-like object > that only generates these objects when they are accessed can save > memory. > > On a fairly pathological case where the DB has

[ovs-dev] [PATCH v2] python: idl: Avoid pre-allocating column defaults

Many python implementations pre-allocate space for multiple objects in empty dicts and lists. Using a custom dict-like object that only generates these objects when they are accessed can save memory. On a fairly pathological case where the DB has 1000 networks each with 100 ports, with only 'name'

Re: [ovs-dev] [PATCH v7] Encap & Decap actions for MPLS packet type.

On 10/12/21 06:23, Martin Varghese wrote: > From: Martin Varghese Hi, Martin. Thanks for working on this! I didn't review most of the logic of the action itself, but see some comments inline. Bets regards, Ilya Maximets. > > The encap & decap actions are extended to support MPLS packet type.

Re: [ovs-dev] [PATCH] [python] Avoid pre-allocating column defaults

Thanks for the review! On Fri, Nov 5, 2021 at 5:43 AM Dumitru Ceara wrote: > > Hi Terry, > > Nit: I'd prefix the commit subject line with "python: idl: ". > > On 10/12/21 9:53 PM, Terry Wilson wrote: > > Many python implementations pre-allocate space for multiple > > objects in empty dicts and li

[ovs-dev] [PATCH ovn v10 0/4] Introduce infrastructure for VIF plug providers.

Introduce infrastructure for VIF plug providers and add feature to ovn-controller to add and remove ports on the integration bridge as directed by CMS through Logical_Switch_Port options. Traditionally it has been the CMSs responsibility to create Virtual Interfaces (VIFs) as part of instance (Con

[ovs-dev] [PATCH ovn v10 1/4] lib: Add infrastructure for VIF plug providers.

New lib/vif-plug-provider module contains the infrastructure for registering VIF plug provider classes which may be hosted inside or outside the core OVN repository. New controller/vif-plug module adds internal interface for interacting with the VIF plug providers. Extend build system to allow bu

[ovs-dev] [PATCH ovn v10 2/4] ovn-controller: Prepare VIF plug provider infrastructure.

Add `enable-dummy-plug` option - To enable testing of the VIF plug provider infrastructure without building OVN with an external VIF plug provider we include a dummy implementation which can be enabled using this command line option. Signed-off-by: Frode Nordahl --- controller/ovn-controller.c |

[ovs-dev] [PATCH ovn v10 4/4] NEWS: Add note on infrastructure for VIF plug providers.

Signed-off-by: Frode Nordahl --- NEWS | 6 ++ 1 file changed, 6 insertions(+) diff --git a/NEWS b/NEWS index 5f448e67d..97a833a1b 100644 --- a/NEWS +++ b/NEWS @@ -6,6 +6,12 @@ Post v21.09.0 if not desired. - Added Load_Balancer_Group support, which simplifies large scale configu

[ovs-dev] [PATCH ovn v10 3/4] controller: Consider plugging VIF on CMS request.

When OVN is linked with an appropriate VIF plug provider, CMS can request an OVN controller to plug individual lports into the Open vSwitch instance it manages. The port and interface record will be maintained throughout the lifetime of the lport and it will be removed on release of lport. Add po

Re: [ovs-dev] [PATCH ovn v9 2/4] ovn-controller: Prepare VIF plug provider infrastructure.

On Fri, Nov 5, 2021 at 2:41 PM 0-day Robot wrote: > > Bleep bloop. Greetings Frode Nordahl, 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. > > > build: > gcc -std=gnu99 -DHAVE_CONFIG_H

Re: [ovs-dev] [PATCH v2] Backup and remove existing DB when joining cluster

On 11/5/21 2:29 PM, Terry Wilson wrote: > On Fri, Nov 5, 2021 at 6:07 AM Dumitru Ceara wrote: >> >> On 10/20/21 10:56 PM, Terry Wilson wrote: >>> ovsdb-tool join-cluster requires a remote addr, so the existing >>> code that tried to join a cluster without it when there was an >>> existing $DB_FILE

Re: [ovs-dev] [PATCH ovn v9 2/4] ovn-controller: Prepare VIF plug provider infrastructure.

Bleep bloop. Greetings Frode Nordahl, 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. build: gcc -std=gnu99 -DHAVE_CONFIG_H -I. -I ./include -I ./include -I ./ovn -I ./include -I ./lib -I

Re: [ovs-dev] [PATCH v5 ovn] controller: add memory accounting for local_datapath

On 10/21/21 11:47 AM, Lorenzo Bianconi wrote: > Similar to if-status-mgr, track memory allocated for local_datapath. > > Signed-off-by: Lorenzo Bianconi > --- Looks good to me, thanks! Acked-by: Dumitru Ceara ___ dev mailing list d...@openvswitch.o

[ovs-dev] [PATCH ovn v9 1/4] lib: Add infrastructure for VIF plug providers.

New lib/vif-plug-provider module contains the infrastructure for registering VIF plug provider classes which may be hosted inside or outside the core OVN repository. New controller/vif-plug module adds internal interface for interacting with the VIF plug providers. Extend build system to allow bu

[ovs-dev] [PATCH ovn v9 0/4] Introduce infrastructure for VIF plug providers.

Introduce infrastructure for VIF plug providers and add feature to ovn-controller to add and remove ports on the integration bridge as directed by CMS through Logical_Switch_Port options. Traditionally it has been the CMSs responsibility to create Virtual Interfaces (VIFs) as part of instance (Con

[ovs-dev] [PATCH ovn v9 3/4] controller: Consider plugging VIF on CMS request.

When OVN is linked with an appropriate VIF plug provider, CMS can request an OVN controller to plug individual lports into the Open vSwitch instance it manages. The port and interface record will be maintained throughout the lifetime of the lport and it will be removed on release of lport. Signed

[ovs-dev] [PATCH ovn v9 2/4] ovn-controller: Prepare VIF plug provider infrastructure.

Add port by interfaces index - To be able to effectively remove ports previously plugged by us we need to look up ports by interface records. Add Port_Binding by requested_chassis index - To be able to effectively iterate over ports destined to our chassis we need to look up Port_Binding records b

[ovs-dev] [PATCH ovn v9 4/4] NEWS: Add note on infrastructure for VIF plug providers.

Signed-off-by: Frode Nordahl --- NEWS | 6 ++ 1 file changed, 6 insertions(+) diff --git a/NEWS b/NEWS index 5f448e67d..97a833a1b 100644 --- a/NEWS +++ b/NEWS @@ -6,6 +6,12 @@ Post v21.09.0 if not desired. - Added Load_Balancer_Group support, which simplifies large scale configu

Re: [ovs-dev] [PATCH v2] Backup and remove existing DB when joining cluster

On Fri, Nov 5, 2021 at 6:07 AM Dumitru Ceara wrote: > > On 10/20/21 10:56 PM, Terry Wilson wrote: > > ovsdb-tool join-cluster requires a remote addr, so the existing > > code that tried to join a cluster without it when there was an > > existing $DB_FILE would fail. > > > > Instead, if we are tryi

Re: [ovs-dev] [PATCH v17 5/8] dpif-offload-netlink: Implement dpif-offload-provider API

On 21 Oct 2021, at 10:01, Chris Mi wrote: > Implement dpif-offload API for netlink datapath. > > Signed-off-by: Chris Mi > Reviewed-by: Eli Britstein > Acked-by: Eelco Chaudron Chris, if you make changes to this patch after I have ACKed it, you need to remove the acked-by. See some comments i

Re: [ovs-dev] [PATCH v17 7/8] netdev-offload-tc: Add offload support for sFlow

Hi Chris, Only one thing you forgot to fix from the previous review. If you make no other changes than the below, you can add my ack. //Eelco On 21 Oct 2021, at 10:01, Chris Mi wrote: > Create a unique group ID to map the sFlow info when offloading sFlow > action to TC. When showing the offloa

Re: [ovs-dev] [PATCH v17 3/8] dpif-offload-provider: Introduce dpif-offload-provider layer

On 21 Oct 2021, at 10:00, Chris Mi wrote: > Some offload actions require functionality that is not netdev > based, but dpif. For example, sFlow action requires to create > a psample netlink socket to receive the sampled packets from > TC or kernel driver. > > Create dpif-offload-provider layer t

Re: [ovs-dev] [PATCH ovn v8 0/4] Introduce infrastructure for VIF plug providers

On Fri, Nov 5, 2021 at 8:50 AM Han Zhou wrote: > > > > On Tue, Nov 2, 2021 at 4:29 AM Frode Nordahl > wrote: > > > > Introduce infrastructure for VIF plug providers and add feature to > > ovn-controller to add and remove ports on the integration bridge as > > directed by CMS through Logical_Swit

Re: [ovs-dev] [PATCH ovn v8 4/4] NEWS: Add note on infrastructure for VIF plug providers.

On Fri, Nov 5, 2021 at 8:44 AM Han Zhou wrote: > > > > On Tue, Nov 2, 2021 at 4:29 AM Frode Nordahl > wrote: > > > > Signed-off-by: Frode Nordahl > > --- > > NEWS | 7 +++ > > 1 file changed, 7 insertions(+) > > > > diff --git a/NEWS b/NEWS > > index 5f448e67d..65bf7d30f 100644 > > --- a/N

Re: [ovs-dev] [PATCH ovn v8 2/4] ovn-controller: Prepare VIF plug provider infrastructure.

On Fri, Nov 5, 2021 at 8:41 AM Han Zhou wrote: > > > > On Tue, Nov 2, 2021 at 4:29 AM Frode Nordahl > wrote: > > > > Add port by interfaces index - To be able to effectively remove > > ports previously plugged by us we need to look up ports by > > interface records. > > > > Add Port_Binding by r

Re: [ovs-dev] [PATCH ovn v8 1/4] lib: Add infrastructure for VIF plug providers.

On Fri, Nov 5, 2021 at 8:40 AM Han Zhou wrote: > > > > On Tue, Nov 2, 2021 at 4:29 AM Frode Nordahl > wrote: > > > > New lib/vif-plug-provider module contains the infrastructure for > > registering VIF plug provider classes which may be hosted inside > > or outside the core OVN repository. > > >

Re: [ovs-dev] [PATCH] utilities: Add another GDB macro for ovs-vswitchd

Thanks for adding a new command to the GDB macros. See some comments below… //Eelco On 4 Nov 2021, at 19:03, Mike Pattrick wrote: > This commit adds a basic packet metadata macro to the already existing > macros in ovs_gdb.py, ovs_dump_packets will print out information about > one or more pack

Re: [ovs-dev] [PATCH v2] Backup and remove existing DB when joining cluster

On 10/20/21 10:56 PM, Terry Wilson wrote: > ovsdb-tool join-cluster requires a remote addr, so the existing > code that tried to join a cluster without it when there was an > existing $DB_FILE would fail. > > Instead, if we are trying to specifically join a cluster and there > is an existing $DB_F

Re: [ovs-dev] [PATCH] [python] Avoid pre-allocating column defaults

Hi Terry, Nit: I'd prefix the commit subject line with "python: idl: ". On 10/12/21 9:53 PM, Terry Wilson wrote: > Many python implementations pre-allocate space for multiple > objects in empty dicts and lists. Using a custom dict-like object > that only generates these objects when they are acce

[ovs-dev] [PATCH] dpif-netlink: Improve feature negotiation for older kernels.

OVS_DP_F_UNALIGNED is already set, no need to set again. If restarting ovs, dp is already created. So dpif_netlink_dp_transact() will return EEXIST. No need to probe again. Signed-off-by: Chris Mi --- lib/dpif-netlink.c | 3 +-- 1 file changed, 1 insertion(+), 2 deletions(-) diff --git a/lib/dp

Re: [ovs-dev] [PATCH] [python] Allow custom transaction ops to be added

On 10/14/21 12:40 AM, Terry Wilson wrote: > It can be useful to be able to send raw transaction operations > through the Idl's connection. For example, to clean up MAC_Binding > entries for floating IPs without having to monitor the MAC_Binding > table which can be quite large. > > Signed-off-by:

Re: [ovs-dev] [PATCH v6 2/2] dpctl: dpif: allow viewing and configuring dp cache sizes

On 4 Nov 2021, at 16:21, Ilya Maximets wrote: > Hi, Eelco. Thanks for the patch! > > I'd like to test it a bit more on my setup before applying, but the > code looks mostly OK to me. > > See some nits inline. > > Bets regards, Ilya Maximets. > > On 9/6/21 10:53, Eelco Chaudron wrote: >> This p

[ovs-dev] [PATCH ovn v2] nb: Add support for gateway_mtu_bypass.

There are various costs (e.g., not being able to perform hardware offload in some cases) when using check_pkt_larger() so the CMS can now limit the impact by bypassing the packet length checks for specific types of traffic (e.g., TCP). Reported-at: https://bugzilla.redhat.com/show_bug.cgi?id=20117

Re: [ovs-dev] [PATCH v6 1/2] dpctl: dpif: add kernel datapath cache hit output

On 4 Nov 2021, at 15:51, Ilya Maximets wrote: > Hi, Eelco. Sorry for this taking so long. > > The patch looks good to me, just one small nit below. > > On 9/6/21 10:53, Eelco Chaudron wrote: >> This patch adds cache usage statistics to the output: >> >> $ ovs-dpctl show >> system@ovs-system: >

[ovs-dev] [PATCH ovn] ovs: Bump submodule to include ovsdb/IDL optimizations.

There are two relevant OVS commits picked with this submodule update. The first one significantly improves performance of OVN database clients and the second one adds memory reporting for the OVN IDL clients. 1bdda7b6d ovsdb-idl: Use functions to apply diff in place. 066741d9c ovsdb-idl: Add m

Re: [ovs-dev] [PATCH] ovsdb-idl: Use functions to apply diff in place.

On 11/5/21 12:34 AM, Ilya Maximets wrote: > On 10/21/21 11:16, Dumitru Ceara wrote: >> On large scale deployments with records that contain large sets, this >> significantly improves client side performance as it avoids comparing >> full contents of the old and new rows. >> >> Signed-off-by: Dumitr

Re: [ovs-dev] [PATCH ovn v8 0/4] Introduce infrastructure for VIF plug providers

On Tue, Nov 2, 2021 at 4:29 AM Frode Nordahl wrote: > > Introduce infrastructure for VIF plug providers and add feature to > ovn-controller to add and remove ports on the integration bridge as > directed by CMS through Logical_Switch_Port options. > > Traditionally it has been the CMSs responsibil

Re: [ovs-dev] [PATCH ovn v8 4/4] NEWS: Add note on infrastructure for VIF plug providers.

On Tue, Nov 2, 2021 at 4:29 AM Frode Nordahl wrote: > > Signed-off-by: Frode Nordahl > --- > NEWS | 7 +++ > 1 file changed, 7 insertions(+) > > diff --git a/NEWS b/NEWS > index 5f448e67d..65bf7d30f 100644 > --- a/NEWS > +++ b/NEWS > @@ -6,6 +6,13 @@ Post v21.09.0 > if not desired. >

Re: [ovs-dev] [PATCH ovn v8 3/4] controller: Consider plugging VIF on CMS request.

On Tue, Nov 2, 2021 at 4:29 AM Frode Nordahl wrote: > > When OVN is linked with an appropriate VIF plug provider, > CMS can request an OVN controller to plug individual lports into > the Open vSwitch instance it manages. > > The port and interface record will be maintained throughout the > lifetim

Re: [ovs-dev] [PATCH ovn v8 2/4] ovn-controller: Prepare VIF plug provider infrastructure.

On Tue, Nov 2, 2021 at 4:29 AM Frode Nordahl wrote: > > Add port by interfaces index - To be able to effectively remove > ports previously plugged by us we need to look up ports by > interface records. > > Add Port_Binding by requested_chassis index - To be able to > effectively iterate over ports

Re: [ovs-dev] [PATCH ovn v8 1/4] lib: Add infrastructure for VIF plug providers.

On Tue, Nov 2, 2021 at 4:29 AM Frode Nordahl wrote: > > New lib/vif-plug-provider module contains the infrastructure for > registering VIF plug provider classes which may be hosted inside > or outside the core OVN repository. > > New controller/vif-plug module adds internal interface for > interac