Re: [ovs-dev] [PATCH] dpif-netlink: Allow offloading of flows with dl_type 0x1234.

On 2019-07-30 6:12 PM, Ilya Maximets wrote: > 'dpif_probe_feature()' always has DPIF_FP_PROBE flag set. Other probing > code uses dpif_execute() with DPIF_OP_EXECUTE, hence never calls > parse_flow_put(). > Thus, this 'if' statement is wrong and should be removed as it only > forbids offloading

Re: [ovs-dev] [PATCH v1 0/4] OVN: Vlan backed DVR, enable N-S packet flow

Hi, Sent out the same series (with same version) on OVN repo. Thanks Regards, Ankur -Original Message- From: Ankur Sharma Sent: Tuesday, July 30, 2019 7:23 PM To: ovs-dev@openvswitch.org Subject: RE: [PATCH v1 0/4] OVN: Vlan backed DVR, enable N-S packet flow Hi, I will be sending th

Re: [ovs-dev] OVS/OVN split complete in ovn-org/ovn repository

Thanks a lot for reply Numan. Regards, Ankur From: Numan Siddique Sent: Tuesday, July 30, 2019 10:49 PM To: Ankur Sharma Cc: ovs-dev@openvswitch.org; ovs-discuss Subject: Re: [ovs-dev] OVS/OVN split complete in ovn-org/ovn repository On Wed, Jul 31, 2019 at 7:13 AM Ankur Sharma mailto:anku

[ovs-dev] [PATCH v1 4/4] OVN: Vlan backed DVR N-S, redirect packet via localnet port

Background: With c0974331b7a19a87ab8f1f2cec8fbe366af92fa2, we have added support for E-W workflow for vlan backed DVRs. This series enables N-S workflow for vlan backed DVRs. Key difference between E-W and N-S traffic flow is that N-S flow requires a gateway chassis. A gateway chassis will be res

[ovs-dev] [PATCH v1 3/4] OVN: Vlan backed DVR N-S, avoid get_arp on non redirect chassis.

Background: With c0974331b7a19a87ab8f1f2cec8fbe366af92fa2, we have added support for E-W workflow for vlan backed DVRs. This series enables N-S workflow for vlan backed DVRs. Key difference between E-W and N-S traffic flow is that N-S flow requires a gateway chassis. A gateway chassis will be res

[ovs-dev] [PATCH v1 2/4] OVN: Vlan backed DVR N-S, redirect-type option

Background: With c0974331b7a19a87ab8f1f2cec8fbe366af92fa2, we have added support for E-W workflow for vlan backed DVRs. This series enables N-S workflow for vlan backed DVRs. Key difference between E-W and N-S traffic flow is that N-S flow requires a gateway chassis. A gateway chassis will be res

[ovs-dev] [PATCH v1 1/4] OVN: Do not replace router port mac on gateway chassis.

With 795d7f24ce0e2ed5454e193a059451d237289542 we have added support for E-W routing on vlan backed networks by replacing router port macs with chassis macs. This replacement of router port mac need NOT be done on gateway chassis for following reasons: a. For N-S traffic, gateway chassis will resp

[ovs-dev] [PATCH v1 0/4] OVN: Vlan backed DVR, enable N-S packet flow

Problem Description: Redirection to chassisredirect ports happens only via tunnel encapsulation. As a result, same cannot be leveraged upon for vlan backed environments. This series addresses the issue by allowing redirection to happen without encapsulation. Design: === a. High level design

Re: [ovs-dev] OVS/OVN split complete in ovn-org/ovn repository

On Wed, Jul 31, 2019 at 7:13 AM Ankur Sharma wrote: > Hi Numan, > > I have some basic queries: > a. Will OVN has its own versioning and release now? > My assumption is that we will continue with 2.13 and above and the same 6 month release cycle. We can discuss about this further. b. Will OVS RP

Re: [ovs-dev] [PATCH] ofproto-dpif: Fix for recirc issue with mpls traffic with dp_hash

Hi All, Please consider this as gentle reminder. Thanks & Regards, Surya. -Original Message- From: Surya Rudra Sent: 19 July 2019 11:05 AM To: ovs-dev@openvswitch.org Cc: Surya Rudra Subject: [PATCH] ofproto-dpif: Fix for recirc issue with mpls traffic with dp_hash Fix infinite reci

Re: [ovs-dev] [PATCH v1 0/4] OVN: Vlan backed DVR, enable N-S packet flow

Hi, I will be sending the same series in OVN repo. Thanks Regards, Ankur -Original Message- From: Ankur Sharma Sent: Tuesday, July 30, 2019 3:20 PM To: ovs-dev@openvswitch.org Cc: Ankur Sharma Subject: [PATCH v1 0/4] OVN: Vlan backed DVR, enable N-S packet flow Problem Description: Re

Re: [ovs-dev] OVS/OVN split complete in ovn-org/ovn repository

Hi Numan, I have some basic queries: a. Will OVN has its own versioning and release now? b. Will OVS RPM continue to provide OVN binaries? Sorry, I know these things have been discussed already. I thought I will check on mailing list, rather than digging through the emails 😊. Regards, Ankur --

[ovs-dev] 答复: ovn-controller consuming lots of CPU

Thanks for your reply and guidance, Numan It would not settle down after a while, this issue still exists and lasts for more than 12 hours, I don’t have much logical flows and openflows ovn-sbctl lfow-list | wc –l 392 ovs-ofctl dump-flows br-int | wc –l 1571 I enabled debug log, g

[ovs-dev] 答复: How can we improve veth and tap performance in OVS DPDK?

Thanks Ilya, it works after disable tx offload, the performance is indeed very poor, about one tenth of ovs kernel. This is a very very strong warning for us, I strongly suggest ovs document should tell ovs DPDK users the truth in bold word. For ovn, last year, the information I got is ovn can'

[ovs-dev] [PATCH v1 4/4] OVN: Vlan backed DVR N-S, redirect packet via localnet port

Background: With c0974331b7a19a87ab8f1f2cec8fbe366af92fa2, we have added support for E-W workflow for vlan backed DVRs. This series enables N-S workflow for vlan backed DVRs. Key difference between E-W and N-S traffic flow is that N-S flow requires a gateway chassis. A gateway chassis will be res

[ovs-dev] [PATCH v1 2/4] OVN: Vlan backed DVR N-S, redirect-type option

Background: With c0974331b7a19a87ab8f1f2cec8fbe366af92fa2, we have added support for E-W workflow for vlan backed DVRs. This series enables N-S workflow for vlan backed DVRs. Key difference between E-W and N-S traffic flow is that N-S flow requires a gateway chassis. A gateway chassis will be res

[ovs-dev] [PATCH v1 3/4] OVN: Vlan backed DVR N-S, avoid get_arp on non redirect chassis.

Background: With c0974331b7a19a87ab8f1f2cec8fbe366af92fa2, we have added support for E-W workflow for vlan backed DVRs. This series enables N-S workflow for vlan backed DVRs. Key difference between E-W and N-S traffic flow is that N-S flow requires a gateway chassis. A gateway chassis will be res

[ovs-dev] [PATCH v1 1/4] OVN: Do not replace router port mac on gateway chassis.

With 795d7f24ce0e2ed5454e193a059451d237289542 we have added support for E-W routing on vlan backed networks by replacing router port macs with chassis macs. This replacement of router port mac need NOT be done on gateway chassis for following reasons: a. For N-S traffic, gateway chassis will resp

[ovs-dev] [PATCH v1 0/4] OVN: Vlan backed DVR, enable N-S packet flow

Problem Description: Redirection to chassisredirect ports happens only via tunnel encapsulation. As a result, same cannot be leveraged upon for vlan backed environments. This series addresses the issue by allowing redirection to happen without encapsulation. Design: === a. High level design

[ovs-dev] Видеокурс "Мастер делового администрирования" - ультрасовременные знания, методики и решения. 13_05_2019 02_03 198707

ВИДЕОКУРС МАСТЕР ДЕЛОВОГО АДМИНИСТРИРОВАНИЯ Данный курс разработан практикующими специалистами рынка, ведущими тренерами и преподавателями с огромным опытом в бизнес-образовании. За 2 года подготовки учебной программы в её создании приняли участие более 300 специалистов-практиков, 100 ведущих р

Re: [ovs-dev] OVN Mailing list and patch submission

Mark Michelson writes: > On 7/30/19 2:59 PM, Aaron Conole wrote: >> Mark Michelson writes: >> >>> Hi, >>> >>> Now that the OVN repo is created and new OVN changes are being >>> targeted for it, it seems like a good idea to bring up the topic of >>> creating a mailing list for OVN again. Would it

Re: [ovs-dev] OVN Mailing list and patch submission

On 7/30/19 2:59 PM, Aaron Conole wrote: Mark Michelson writes: Hi, Now that the OVN repo is created and new OVN changes are being targeted for it, it seems like a good idea to bring up the topic of creating a mailing list for OVN again. Would it be possible to get a new OVN mailing list? A s

Re: [ovs-dev] OVN Mailing list and patch submission

Mark Michelson writes: > Hi, > > Now that the OVN repo is created and new OVN changes are being > targeted for it, it seems like a good idea to bring up the topic of > creating a mailing list for OVN again. Would it be possible to get a > new OVN mailing list? > > A subject that goes hand-in-hand

Re: [ovs-dev] [PATCH] vswitchd: Make packet-in controller queue size configurable

Hi Dumitru, The C code looks good to me, but I had some findings in the schema update and documentation below. Is this something that could have a test added? On 7/30/19 6:32 AM, Dumitru Ceara wrote: The ofconn packet-in queue for packets that can't be immediately sent on the rconn connectio

Re: [ovs-dev] [PATCH] OVN: fix DNAT/SNAT system-ovn unit tests

Thanks Lorenzo Looks good I did not retest this patch, but it is the same code you sent before that I had tested On Mon, Jul 29, 2019 at 4:41 AM Lorenzo Bianconi < lorenzo.bianc...@redhat.com> wrote: > Fix conntrack checks in the following tests in tests/system-ovn.at: > - ovn -- DNAT and SNAT o

Re: [ovs-dev] [PATCH] test: do not require python2 for CHECK_CONNTRACK macro

On Tue, Jul 30, 2019 at 10:41 AM Mark Michelson wrote: > Why do these macros require python at all? > alg test tools > > On 7/29/19 7:48 AM, Lorenzo Bianconi wrote: > > Do not strictly require python2 for CHECK_CONNTRACK macro definitions in > > system-{kmod,userspace}-macros.at > > > > Signe

[ovs-dev] OVN Mailing list and patch submission

Hi, Now that the OVN repo is created and new OVN changes are being targeted for it, it seems like a good idea to bring up the topic of creating a mailing list for OVN again. Would it be possible to get a new OVN mailing list? A subject that goes hand-in-hand with a new mailing list is patch

Re: [ovs-dev] [PATCH] test: do not require python2 for CHECK_CONNTRACK macro

Thanks Lorenzo Looks good On Mon, Jul 29, 2019 at 4:48 AM Lorenzo Bianconi < lorenzo.bianc...@redhat.com> wrote: > Do not strictly require python2 for CHECK_CONNTRACK macro definitions in > system-{kmod,userspace}-macros.at > > Signed-off-by: Lorenzo Bianconi > --- > tests/system-kmod-macros

Re: [ovs-dev] [PATCH] test: do not require python2 for CHECK_CONNTRACK macro

Why do these macros require python at all? On 7/29/19 7:48 AM, Lorenzo Bianconi wrote: Do not strictly require python2 for CHECK_CONNTRACK macro definitions in system-{kmod,userspace}-macros.at Signed-off-by: Lorenzo Bianconi --- tests/system-kmod-macros.at | 2 +- tests/system-userspa

Re: [ovs-dev] [PATCH] dpif-netlink: Remove wrong restriction for offloading 0x1234 eth type

On 2019-07-30 6:43 PM, Roi Dayan wrote: > From: Eli Britstein > > 'dpif_probe_feature' always has DPIF_FP_PROBE flag set. Other probing > code uses dpif_execute() which uses DPIF_OP_EXECUTE, hence never calls > parse_flow_put(). So, this 'if' statement is wrong and should be deleted > as it on

Re: [ovs-dev] [PATCH RFC v2 3/8] tc: Introduce tc_id to specify a tc filter

On 7/26/2019 12:47 PM, Simon Horman wrote: Hi Paul, On Thu, Jul 04, 2019 at 05:28:22PM +0300, Paul Blakey wrote: Move all that is needed to identify a tc filter to a new structure, tc_id. This removes a lot of duplication in accessing/creating tc filters. Signed-off-by: Paul Blakey

[ovs-dev] [PATCH] dpif-netlink: Remove wrong restriction for offloading 0x1234 eth type

From: Eli Britstein 'dpif_probe_feature' always has DPIF_FP_PROBE flag set. Other probing code uses dpif_execute() which uses DPIF_OP_EXECUTE, hence never calls parse_flow_put(). So, this 'if' statement is wrong and should be deleted as it only forbids offloading of the real legitimate flows with

[ovs-dev] [PATCH] dpif-netlink: Allow offloading of flows with dl_type 0x1234.

'dpif_probe_feature()' always has DPIF_FP_PROBE flag set. Other probing code uses dpif_execute() with DPIF_OP_EXECUTE, hence never calls parse_flow_put(). Thus, this 'if' statement is wrong and should be removed as it only forbids offloading of the real legitimate flows with dl_type 0x1234. Dummy f

Re: [ovs-dev] [PATCH V2 1/1] dpif-netlink: Log eth type 0x1234 not offloadable

On 30.07.2019 16:50, Paul Blakey wrote: > > On 7/30/2019 3:19 PM, Ilya Maximets wrote: >> On 30.07.2019 13:30, Ilya Maximets wrote: >>> On 30.07.2019 13:23, Eli Britstein wrote: On 7/30/2019 1:10 PM, Ilya Maximets wrote: > On 03.07.2019 7:58, Eli Britstein wrote: >> Ethernet type 0x12

Re: [ovs-dev] [RFC 0/2] OVN: add rate limiting support for controller_events

> On Tue, Jul 30, 2019 at 7:14 PM Lorenzo Bianconi < > lorenzo.bianc...@redhat.com> wrote: > > > Introduce the capability to associate a meter to each controller event > > type in order to not overload the pinctrl thread under heavy load. > > For each event type we need to define a naming conventi

Re: [ovs-dev] [RFC 0/2] OVN: add rate limiting support for controller_events

On Tue, Jul 30, 2019 at 7:14 PM Lorenzo Bianconi < lorenzo.bianc...@redhat.com> wrote: > Introduce the capability to associate a meter to each controller event > type in order to not overload the pinctrl thread under heavy load. > For each event type we need to define a naming convention for the r

Re: [ovs-dev] [PATCH RFC v2 0/8] Introduce connection tracking tc offload

On Mon, Jul 29, 2019 at 01:05:50PM +, Paul Blakey wrote: > Hi, > > I reproduced both traces. > > The first rcu trace is because we deference the qdisc block without rtnl > locked (flower allows it unlocked) > > and the second trace is because res->goto_tp can be null if the last > prio of

Re: [ovs-dev] [PATCH V2 1/1] dpif-netlink: Log eth type 0x1234 not offloadable

On 7/30/2019 3:19 PM, Ilya Maximets wrote: > On 30.07.2019 13:30, Ilya Maximets wrote: >> On 30.07.2019 13:23, Eli Britstein wrote: >>> On 7/30/2019 1:10 PM, Ilya Maximets wrote: On 03.07.2019 7:58, Eli Britstein wrote: > Ethernet type 0x1234 is used for testing and not being offloadable

Re: [ovs-dev] [PATCH RFC v2 0/8] Introduce connection tracking tc offload

On 7/29/2019 4:16 PM, Ilya Maximets wrote: > Hi. > > This patch-set has some checkpatch issues. It'll be good if you > could fix them before submitting the next version. > > Best regards, Ilya Maximets. Sure thanks. ___ dev mailing list d...@openvswitch.

[ovs-dev] [RFC 2/2] OVN: northd: add rate limiting support for SB controller events

Introduce the capability to associate a meter to each controller event type in order to not overload the pinctrl thread under heavy load. Each event type relies on a meter with a defined name: - empty_lb_backends: event-elb Signed-off-by: Lorenzo Bianconi --- ovn/northd/ovn-northd.c | 44 +++

[ovs-dev] [RFC 1/2] OVN: add meter support to trigger_event action

Introduce meter support to trigger_event action in order to not overload pinctrl thread under heavy load Signed-off-by: Lorenzo Bianconi --- include/ovn/actions.h | 1 + ovn/lib/actions.c | 42 -- 2 files changed, 37 insertions(+), 6 deletions(-) dif

[ovs-dev] [RFC 0/2] OVN: add rate limiting support for controller_events

Introduce the capability to associate a meter to each controller event type in order to not overload the pinctrl thread under heavy load. For each event type we need to define a naming convention for the related meter: - empty_lb_backends: event-elb Lorenzo Bianconi (2): OVN: add meter support t

Re: [ovs-dev] ovn-controller consuming lots of CPU

On Tue, Jul 30, 2019 at 2:25 PM Frank Wang(王培辉) wrote: > Hi All, > > > > I’m using the latest ovn 2.11.1 version, I found ovn-controller consuming > a lot of CPU, I’d like to know if this is a bug, This seems to have > happened before, but not in the same place, > > This time the lib/stream-fd.c

Re: [ovs-dev] [PATCH V2 1/1] dpif-netlink: Log eth type 0x1234 not offloadable

On 30.07.2019 13:30, Ilya Maximets wrote: > On 30.07.2019 13:23, Eli Britstein wrote: >> >> On 7/30/2019 1:10 PM, Ilya Maximets wrote: >>> On 03.07.2019 7:58, Eli Britstein wrote: Ethernet type 0x1234 is used for testing and not being offloadable. For testing offloadable features, log abo

[ovs-dev] [PATCH OvS 4/4] ovs-tc: offload MPLS set actions to TC datapath

Recent modifications to TC allows the modifying of fields within the outermost MPLS header of a packet. OvS datapath rules impliment an MPLS set action by supplying a new MPLS header that should overwrite the current one. Convert the OvS datapath MPLS set action to a TC modify action and allow suc

[ovs-dev] [PATCH OvS 2/4] ovs-tc: offload MPLS pop actions to TC datapath

TC now supports an action to pop the outer MPLS header from a packet. The next protocol after the header is required alongside this. Currently, OvS datapath rules also supply this information. Offload OvS MPLS pop actions to TC along with the next protocol. Signed-off-by: John Hurley Reviewed-by

[ovs-dev] [PATCH OvS 3/4] ovs-tc: offload MPLS push actions to TC datapath

TC can now be used to push an MPLS header onto a packet. The MPLS label is the only information that needs to be passed here with the rest reverting to default values if none are supplied. OvS, however, gives the entire MPLS header to be pushed along with the MPLS protocol to use. TC can optionally

[ovs-dev] [PATCH OvS 1/4] compat: add compatibility headers for tc mpls action

OvS includes compat code for several TC actions including vlan, mirred and tunnel key. MPLS actions have recently been added to TC in the kernel. In preparation for adding TC offload code for MPLS, add the MPLS compat code. Signed-off-by: John Hurley Reviewed-by: Simon Horman --- acinclude.m4

[ovs-dev] [PATCH OvS 0/4] Offload MPLS actions to TC datapath

Recent additions to TC in the Linux kernel includes an action module that allows MPLS packet manipulation. Update the OvS TC API to allow OvS datapath MPLS actions to be converted and offloaded to a TC datapath. Actions converted include set, push and pop. TC also offers a decrement TTL MPLS acti

[ovs-dev] [PATCH] vswitchd: Make packet-in controller queue size configurable

The ofconn packet-in queue for packets that can't be immediately sent on the rconn connection was limited to 100 packets (hardcoded value). While increasing this limit is usually not recommended as it might create buffer bloat and increase latency, in scaled scenarios it is useful if the administra

Re: [ovs-dev] [PATCH V2 1/1] dpif-netlink: Log eth type 0x1234 not offloadable

On 30.07.2019 13:23, Eli Britstein wrote: > > On 7/30/2019 1:10 PM, Ilya Maximets wrote: >> On 03.07.2019 7:58, Eli Britstein wrote: >>> Ethernet type 0x1234 is used for testing and not being offloadable. For >>> testing offloadable features, log about using this value. >>> >>> Signed-off-by: Eli

Re: [ovs-dev] [PATCH V2 1/1] dpif-netlink: Log eth type 0x1234 not offloadable

On 7/30/2019 1:10 PM, Ilya Maximets wrote: > On 03.07.2019 7:58, Eli Britstein wrote: >> Ethernet type 0x1234 is used for testing and not being offloadable. For >> testing offloadable features, log about using this value. >> >> Signed-off-by: Eli Britstein >> Acked-by: Roi Dayan >> Signed-off-b

Re: [ovs-dev] [PATCH V2 1/1] dpif-netlink: Log eth type 0x1234 not offloadable

On 03.07.2019 7:58, Eli Britstein wrote: > Ethernet type 0x1234 is used for testing and not being offloadable. For > testing offloadable features, log about using this value. > > Signed-off-by: Eli Britstein > Acked-by: Roi Dayan > Signed-off-by: Eli Britstein > Acked-by: Ben Pfaff > --- > li

[ovs-dev] ovn-controller consuming lots of CPU

Hi All, I’m using the latest ovn 2.11.1 version, I found ovn-controller consuming a lot of CPU, I’d like to know if this is a bug, This seems to have happened before, but not in the same place, This time the lib/stream-fd.c seems has some problem. See the details below 2019-07-30T07:08:09

Re: [ovs-dev] [PATCH] ovn-controller: Fix flow installation latency caused by recompute.

Thanks Mark and Numan for the review. Ben, now this is merged to OVN repo. Could you help backporting it to 2.12? Thanks, Han On Thu, Jul 25, 2019 at 1:23 PM Mark Michelson wrote: > > Forgot the e-mail address after my name :) > > Acked-by: Mark Michelson > > On 7/25/19 4:16 PM, Mark Michelson

Re: [ovs-dev] [PATCH ovn 2/2] ovn-controller: Fix flow installation latency caused by recompute.

On Mon, Jul 29, 2019 at 9:11 PM Numan Siddique wrote: > > > > On Tue, Jul 30, 2019 at 9:39 AM Numan Siddique wrote: >> >> >> >> On Tue, Jul 30, 2019 at 7:10 AM Han Zhou wrote: >>> >>> From: Han Zhou >>> >>> When there are in-flight flow-installations pending to ovs-vswitchd, >>> current increme