[ovs-dev] [PATCH v4 0/2] Add Performance Measurement Library

This set of commits adds a new library for OVS that allows for measuring the performance of operations in OVS and compiling statistics from these measurements. For developers, this can provide a measurement of something that is either finer or coarser-grained than what is easily measured with a

Re: [ovs-dev] [PATCH] ofproto-dpif-xlate: native tunnel using valid tun_src specified by flow or port options

Hi Ben, Are there some idea for this patch? BR wenxu 在 2018-02-03 08:47:41，"wenxu" 写道： Hi ben, This patch can be a bugfix. The tunnel_src of packet maybe not the IP address set by gre port options:local_ip dpdk-br has two address 10.1.1.7/24 and 10.1.1.254/32

[ovs-dev] [PATCH v4 1/2] Add performance measuring API

This is similar to the existing coverage and perf-counter APIs in OVS. However, rather than keeping counters, this is aimed at timing how long operations take to perform. "Operations" in this case can be anything from a loop iteration, to a function, to something more complex. The library will

[ovs-dev] [PATCH v4 2/2] Measure performance of ovn-controller loop.

This modifies ovn-controller to measure the amount of time it takes to detect a change in the southbound database and generate the resulting flow table. This may require multiple iterations of the ovn-controller loop. The statistics can be queried using: ovs-appctl -t ovn-controller

Re: [ovs-dev] [PATCH 0/8] Linux upstream kernel backports and updates

> On Feb 19, 2018, at 4:30 PM, Pravin Shelar wrote: > > On Mon, Feb 19, 2018 at 3:23 PM, Justin Pettit wrote: >> >>> On Feb 19, 2018, at 9:50 AM, Justin Pettit wrote: >>> >>> On Feb 16, 2018, at 5:32 PM, Justin Pettit

Re: [ovs-dev] [PATCH] datapath-windows: Support to selectively compile targets

I agree with having the target specified during make and not during configure. There is no reason to re-do the whole configure process (as Alin mentioned that it is particularly slow on Windows) only to change the target version. Thanks, Alin Balutoiu > -Original Message- > From:

Re: [ovs-dev] [PATCH 4/4] ovn-northd: Reduce amount of flow hashing.

> On Fri, Feb 16, 2018 at 05:10:29PM +0100, Jakub Sitnicki wrote: >> Ben, >> >> On Wed, Feb 14, 2018 at 09:54 PM GMT, Ben Pfaff wrote: >> > Jakub Sitnicki demonstrated that repeatedly calculating row hashes is >> > expensive, so this should improve ovn-northd performance. >> > >> > Reported-by:

Re: [ovs-dev] [PATCH v5] Configurable Link State Change (LSC) detection mode

Hi Robert, I did not see a reply to my v4 reply, so I ask this question again:     Did you try what happens if the PMD does not support LSC and its enabled? Also, it would be easier to understand what you are going to change if you answer the questions in the review emails. For example,

[ovs-dev] [PATCH] AUTHORS: update e-mail address of Zoltan Balogh

Signed-off-by: Zoltan Balogh CC: Zoltan Balogh --- AUTHORS.rst | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/AUTHORS.rst b/AUTHORS.rst index 5fbb27653..2ef51dcef 100644 --- a/AUTHORS.rst +++ b/AUTHORS.rst @@

Re: [ovs-dev] [RFC 2/2] ingress scheduling: Provide per interface ingress priority

Not a full review. Two general comments inline. > Allow configuration to specify an ingress priority for interfaces. > Modify ovs-netdev datapath to act on this configuration so that packets > on interfaces with a higher priority will tend be processed ahead of > packets on lower priority

Re: [ovs-dev] Please tag v2.9.0 and v2.8.2

On Mon, Feb 19, 2018 at 03:13:32PM -0800, Justin Pettit wrote: > Hi, Ben. Can you tag "v2.9.0" on commit bd916d13db ("Set release dates for > 2.9.0."), and "v2.8.2" on commit ba44001de6 ("Set release date for 2.8.2.")? Done. ___ dev mailing list

Re: [ovs-dev] [RFC 0/2] Ingress Scheduling

> -Original Message- > From: Ilya Maximets [mailto:i.maxim...@samsung.com] > Sent: Tuesday, February 20, 2018 2:55 PM > To: ovs-dev@openvswitch.org; O Mahony, Billy > Subject: Re: Re: [ovs-dev] [RFC 0/2] Ingress Scheduling > > > This patch set implements the

Re: [ovs-dev] [RFC 2/2] ingress scheduling: Provide per interface ingress priority

> -Original Message- > From: Ilya Maximets [mailto:i.maxim...@samsung.com] > Sent: Tuesday, February 20, 2018 3:10 PM > To: ovs-dev@openvswitch.org; O Mahony, Billy > Subject: Re: [ovs-dev] [RFC 2/2] ingress scheduling: Provide per interface > ingress >

Re: [ovs-dev] Hardware Acceleration in OVS-DPDK

Hello All, As discussed in the last meeting, I have created a OVS 2.7 fork with our hardware acceleration implementation as below. https://github.com/sugchand/ovs.git (branch - dpdk-hw-accel-intel) Few points on the implementation. 1)

[ovs-dev] [RFC 0/8] ovn-controller Incremental Processing

ovn-controller currently recomputes everything when there are any changes of input, which leads to high CPU usages and slow in end-to-end flow enforcement in response to changes. It even wastes CPU to recompute flows for unrelated inputs such as pinctrl events. This patch series implements

[ovs-dev] [RFC 2/8] ovn-controller: Track OVSDB changes

Track OVSDB changes for future patches of incremental processing Signed-off-by: Han Zhou --- ovn/controller/bfd.c| 4 ++-- ovn/controller/binding.c| 16 ovn/controller/encaps.c | 12 ++-- ovn/controller/ovn-controller.c | 22

[ovs-dev] [RFC 8/8] ovn-controller: port-binding incremental processing for physical flows

This patch implements change handler for port-binding in flow_output for physical flows computing, so that physical flow computing will be incremental. This patch together with previous incremental processing engine related changes supports incremental processing for lflow changes and

[ovs-dev] [RFC 1/8] ovn-controller: Incremental processing engine

This patch implements the engine which will be used in future patches for ovn-controller incremental processing. Signed-off-by: Han Zhou --- ovn/lib/automake.mk| 4 +- ovn/lib/inc-proc-eng.c | 97 ovn/lib/inc-proc-eng.h | 118

[ovs-dev] [RFC 5/8] ovn-controller: split ovs_idl inputs in incremental engine

Create nodes for ovs_idl inputs and add ovs_idl inputs and SB inputs as dependencies for runtime_data. With this patch there is no recomputed if there is no change in input. For example, pinctrl input will not trigger flow recompute any more. Signed-off-by: Han Zhou ---

[ovs-dev] [RFC 3/8] ovn-controller: Initial use of incremental engine in main

Incremental proccessing engine is used to compute flows. In this patch we only create 2 engine nodes with simple dependency: runtime_data -> flow_output In each iteration everything is still recomputed. Signed-off-by: Han Zhou --- ovn/controller/ofctrl.c | 21 ++-

[ovs-dev] [RFC 4/8] ovn-controller: Split SB inputs as separate incremental engine nodes

This patch expands the incremental processing by spliting SB inputs from runtime_data and add them as input for flow_output. Signed-off-by: Han Zhou --- ovn/controller/ovn-controller.c | 41 + 1 file changed, 41 insertions(+) diff --git

[ovs-dev] [RFC 6/8] ovn-controller: Incremental logical flow processing

Persistents flow-table and implements change handler of flow_output for SB lflow changes. (TODO: test case 2338 failed) Signed-off-by: Han Zhou --- include/ovn/actions.h | 3 + ovn/controller/lflow.c | 104 ovn/controller/lflow.h

[ovs-dev] [RFC 7/8] ovn-controller: runtime_data change handler for SB port-binding

Evaluates change for SB port-binding in runtime_data node. If the port-binding change has no impact for the runtime_data it will not trigger runtime_data change. Signed-off-by: Han Zhou --- ovn/controller/binding.c| 85 +

[ovs-dev] [PATCH v2 1/2] OVN: add icmp4{} action support

icmp4 action is used to replace the IPv4 packet been processed with an ICMPv4 packet initialized based on incoming IPv4 one. Ethernet and IPv4 fields not listed are not changed: - ip.proto = 1 (ICMPv4) - ip.frag = 0 (not a fragment) - ip.ttl = 255 - icmp4.type = 3 (destination unreachable) -

[ovs-dev] [PATCH v2 0/2] add acl reject rule support introducing icmp4 action

Changes since v1: - removed 1/3 of v1 since it has been already merged - added ovn automatic tests for icmp4 action and for ACL reject rules - used tos of incoming IP frames for ICMP dst unreachable packets - updated ovn-sb.xml - used icmp4 instead of icmp as naming convention Changes since RFC:

Re: [ovs-dev] [PATCH] datapath-windows: Support to selectively compile targets

That is not consistent with how we specify the Debug/Release flag. There is no need to do the whole configure if you specify the targets correctly. Running multiple make commands is confusing for the user in terms of the current workflow. Thanks, Shashank

Re: [ovs-dev] [RFC 1/4] dpif-netdev: Refactor datapath flow cache

Hi Yipeng, Thanks for the RFC series. This patch series need to be rebased. I applied this on an older commit to do initial testing. Some comments below. I see that DFC cache is implemented in similar lines of EMC cache except that it holds Million entries and uses more bits of RSS hash to

Re: [ovs-dev] [RFC 2/4] dpif-netdev: Fix EMC key length

This fix is needed and can be included in 1/4 in next revision. - Bhanuprakash. >-Original Message- >From: ovs-dev-boun...@openvswitch.org [mailto:ovs-dev- >boun...@openvswitch.org] On Behalf Of Yipeng Wang >Sent: Thursday, January 18, 2018 6:20 PM >To: d...@openvswitch.org;