Re: [ovs-dev] ovn-controller CPU use and incremental processing dependencies

On Fri, Jun 8, 2018 at 5:21 PM, Ben Pfaff wrote: > > On Thu, Jun 07, 2018 at 04:51:34PM -0700, Ben Pfaff wrote: > > On Thu, Jun 07, 2018 at 04:32:10PM -0700, Han Zhou wrote: > > > On Thu, Jun 7, 2018 at 11:15 AM, Ben Pfaff wrote: > > > > > > > > To review the problem we're trying to solve,

Re: [ovs-dev] ovn-controller CPU use and incremental processing dependencies

On Thu, Jun 07, 2018 at 04:51:34PM -0700, Ben Pfaff wrote: > On Thu, Jun 07, 2018 at 04:32:10PM -0700, Han Zhou wrote: > > On Thu, Jun 7, 2018 at 11:15 AM, Ben Pfaff wrote: > > > > > > To review the problem we're trying to solve, ovn-controller uses too > > > much CPU in many circumstances

[ovs-dev] [PATCH v1] ovn-ctl and ovndb-servers set connection table

only for master node with remote option when using load balancer to manage ovndb clusters via pacemaker. This is will allow setting inactivity probe on the masters. For pacemaker to manage ovndb resources via LB, we skipped creating connection table and hence the inactivity probe was getting set

[ovs-dev] [PATCH 3/5] ovsdb-idl: Redesign use of indexes.

The design of the compound index feature in the C OVSDB IDL was unusual. Indexes were generally referenced only by name rather than by pointer, and could be obtained only from the top-level ovsdb_idl object. To iterate or otherwise search an index required explicitly creating a special

[ovs-dev] [PATCH 4/5] chassis-index: Use OVSDB index mechanism.

It seems like a good idea to use the built-in indexing instead of doing it by hand. Signed-off-by: Ben Pfaff --- ovn/controller/bfd.c| 23 +++--- ovn/controller/bfd.h| 6 ++-- ovn/controller/binding.c| 13 ovn/controller/binding.h| 3 +-

[ovs-dev] [PATCH 5/5] ovn-controller: Use chassis_lookup_by_name() instead of get_chassis().

This was duplicate functionality. Signed-off-by: Ben Pfaff --- ovn/controller/chassis.c| 5 +++-- ovn/controller/chassis.h| 6 -- ovn/controller/ovn-controller.c | 24 ovn/controller/ovn-controller.h | 3 --- 4 files changed, 11 insertions(+), 27

[ovs-dev] [PATCH 2/5] ovn-controller: Pass around pointers to individual tables.

We're working to make ovn-controller compute more incrementally, to reduce CPU usage. To make it easier to keep track of dependencies, it makes sense to pass around pointers to fine-grained resources instead of an entire database at a time. This commit introduces a way to pass individual tables

[ovs-dev] [PATCH 1/5] ovn-controller: Style fixes.

The OVS coding style says that input parameters should come first, followed by output parameters. This changes a few functions in ovn-controller to fit this style. It also marks a number of input parameters 'const', for clarity. Signed-off-by: Ben Pfaff --- ovn/controller/bfd.c|

Re: [ovs-dev] [PATCH v2] datapath-windows: Optimize conntrack performance

Hi Shashank, Thanks for the review. I will address your comments and send out a V3. Regards, Anand Kumar From: Shashank Ram Date: Friday, June 8, 2018 at 10:11 AM To: Anand Kumar Cc: "d...@openvswitch.org" Subject: Re: [ovs-dev] [PATCH v2] datapath-windows: Optimize conntrack performance

[ovs-dev] OVS DPDK: dpdk_merge pull request for branch-2.8

Hi Ben, The following changes since commit 23a1eb14ce99f42c926441330772bce7c00dc446: lib: fix typo in fragment handling error message (2018-06-07 13:54:02 +0200) are available in the git repository at: https://github.com/istokes/ovs dpdk_merge_2_8 for you to fetch changes up to

[ovs-dev] OVS DPDK: dpdk_merge pull request for branch-2.9

Hi Ben, The following changes since commit fa13e226a6a90fc365a40555df7a70043576c022: lib: fix typo in fragment handling error message (2018-06-07 13:54:59 +0200) are available in the git repository at: https://github.com/istokes/ovs dpdk_merge_2_9 for you to fetch changes up to

[ovs-dev] OVS DPDK: dpdk_merge pull request for master

Hi Ben, The following changes since commit cc266ef0c9795e8ffb688a1f67db3b3cd7797c3c: lib: fix typo in fragment handling error message (2018-06-08 14:05:00 +0200) are available in the git repository at: https://github.com/istokes/ovs dpdk_merge for you to fetch changes up to

Re: [ovs-dev] [PATCH v2] datapath-windows: Optimize conntrack performance

On Thu, Jun 7, 2018, 11:52 AM Anand Kumar wrote: > - Move conntrack lock out of NAT module > - Use spinlock instead of read/write lock for conntrack entry. > - Update 'ctTotalRelatedEntries' using interlocked functions > - Refactor conntrack code to make it more readable. > > Testing: >

Re: [ovs-dev] Q on vhost-user and virtio

On Fri, Jun 08, 2018 at 08:29:10AM -0700, Ravi Kerur wrote: > I understand vhost-user and virtio protocol. When you say "The vhost-user > side in OVS is supported and tested. ", what was your testbed configuration > and how did you test it? It is used by OpenStack. NFV is a big use case. --

Re: [ovs-dev] [PATCH] lib: fix typo in fragment handling error message

On Wed, Jun 06, 2018 at 07:41:08AM -0700, Ben Pfaff wrote: > On Wed, Jun 06, 2018 at 11:47:13AM +0200, Simon Horman wrote: > > On Tue, May 29, 2018 at 08:51:15PM +0200, Louis Peens wrote: > > > The error message states that "not_first" is a valid selection > > > for the ip_frag field, but looking

[ovs-dev] Ubuntu supported IPSEC daemon?

Hi James. I have an intern for the summer, Qiuyu, who is working to add IPSEC support to OVN, so that tunnels between hypervisors are protected against compromise of the physical network. An important question for this purpose is what IPSEC daemon does Ubuntu support and prefer? It sounds like

Re: [ovs-dev] [PATCH v3] OVS-DPDK: Change "dpdk-socket-mem" default value.

> > -Original Message- > > From: ovs-dev-boun...@openvswitch.org [mailto:ovs-dev- > > boun...@openvswitch.org] On Behalf Of O Mahony, Billy > > Sent: Wednesday, May 9, 2018 11:14 AM > > To: d...@openvswitch.org > > Cc: Rybka, MarcinX > > Subject: [ovs-dev] [PATCH v3] OVS-DPDK: Change

Re: [ovs-dev] Q on vhost-user and virtio

On Fri, Jun 08, 2018 at 07:10:26AM -0700, Ravi Kerur wrote: > I am referring to upstream code. I want to use OVS vhost-user on a host and > dpdk-app inside a container using virtio-user connect to OVS_vhost_user. Is > this use-case supported/tested? The vhost-user is a master-slave protocol. The

Re: [ovs-dev] Q on vhost-user and virtio

On Fri, Jun 8, 2018 at 6:06 AM, Flavio Leitner wrote: > On Thu, Jun 07, 2018 at 05:09:46PM -0700, Ravi Kerur wrote: > > Hi, > > > > I have a use-case and I want to > > > > (1) create vhost-user ports on OVS-DPDK running on a host > > (2) in a container a dpdk-based application using virtio-user

Re: [ovs-dev] Q on vhost-user and virtio

On Thu, Jun 07, 2018 at 05:09:46PM -0700, Ravi Kerur wrote: > Hi, > > I have a use-case and I want to > > (1) create vhost-user ports on OVS-DPDK running on a host > (2) in a container a dpdk-based application using virtio-user connecting to > vhost-user ports via sockets. > > I looked at

Re: [ovs-dev] [PATCH] Improved Packet Drop Statistics in OVS.

I'm very sorry for double posting. I believe this patch would represent a major feature in term of OVS's usability. Has anyone got any time for reviewing it? Thanks, Federico On 6 June 2018 at 17:37, Jan Scheurich wrote: > The user-space part for packet drop stats should be generic and work

Re: [ovs-dev] [PATCH v10 4/7] netdev-dpdk: implement flow offload with rte flow

Hi Shahaf, a couple of nits below On 05/18/2018 12:14 PM, Shahaf Shuler wrote: From: Finn Christensen The basic yet the major part of this patch is to translate the "match" to rte flow patterns. And then, we create a rte flow with MARK + RSS actions. Afterwards, all

Re: [ovs-dev] [RFC v7 11/13] netdev-dpdk: copy large packet to multi-seg. mbufs

On 05/06/2018 13:03, Ilya Maximets wrote: > On 23.05.2018 19:47, Tiago Lam wrote: >> From: Mark Kavanagh >> >> Currently, packets are only copied to a single segment in >> the function dpdk_do_tx_copy(). This could be an issue in >> the case of jumbo frames, particularly when multi-segment >>

Re: [ovs-dev] [PATCH v4] Upcall/Slowpath rate-limiter for OVS

> Have you considered making this token bucket per-port instead of > per-pmd? As I read it, a greedy port can exhaust all the tokens from a > particular PMD, possibly leading to an unfair performance for that PMD > thread. Am I just being overly paranoid? > [manu] Yes, this is

Re: [ovs-dev] [PATCH v4] Upcall/Slowpath rate-limiter for OVS

Aaron, Thanx for your comments/suggestions. Please see inline. Thanx Manu On 07/06/18, 7:06 PM, "Aaron Conole" wrote: Manohar Krishnappa Chidambaraswamy writes: > Jan, > > Have addressed your comments/suggestions. Please take a look and let me > know if you have