Re: [vpp-dev] _mm_testz_si128 Woes

2018-02-19 Thread Florin Coras
Okay, I take that back. Google tells me these are SSE4 intrinsics, not SSE2 … Florin > On Feb 19, 2018, at 5:50 PM, Ed Kern wrote: > > yes…it was considered a trivial rebase…. > > IF that was the case then ill take the wrap for that one wrt it getting a +1… > > in a nutshell.. patch 2 faile

Re: [vpp-dev] _mm_testz_si128 Woes

2018-02-19 Thread Ed Kern
yes…it was considered a trivial rebase…. IF that was the case then ill take the wrap for that one wrt it getting a +1… in a nutshell.. patch 2 failed…got its -1 ok fine. someone came and did a trivial rebase…IF those last three jobs docs, csit-virl and whatever were also not set to run on tri

Re: [vpp-dev] _mm_testz_si128 Woes

2018-02-19 Thread Florin Coras
That’s .. interesting. Patch set 3 didn’t run all the jobs. Any idea why that would happen? Instead of reverting the patch, here’s a hammer-based attempt at solving the problem [1]. Damjan, feel free to propose a prettier solution :) Florin [1] https://gerrit.fd.io/r/#/c/10660/

Re: [vpp-dev] _mm_testz_si128 Woes

2018-02-19 Thread Ed Kern
On Feb 19, 2018, at 2:47 PM, Jon Loeliger mailto:j...@netgate.com>> wrote: Hey hey hey, I tried to build a variant of vppsb and failed miserably for lack of the __mm_testz_si128 symbol. Wont make you feel better but my centos builds have been rolling and croaking over this for a few hours a

[vpp-dev] _mm_testz_si128 Woes

2018-02-19 Thread Jon Loeliger
Hey hey hey, I tried to build a variant of vppsb and failed miserably for lack of the __mm_testz_si128 symbol. >From whence does that arrive on my doorstep? I'm trying to figure out if vppsb needs to #define something new, or #include something first, or if my system is suddenly missing somethin

Re: [vpp-dev] TCP Throughput testing

2018-02-19 Thread Florin Coras
Cool! :-) Cheers, Florin > On Feb 19, 2018, at 9:03 AM, Ray Kinsella wrote: > > Insufficient hugepages in the VM - thanks for everyone's help! > > > Ray K > > On 19/02/2018 15:38, Ray Kinsella wrote: >> Part of the mystery is solved in any case, >> Looks to be related to running inside the

Re: [vpp-dev] FD.io CSIT rls1801 report published

2018-02-19 Thread Dave Wallace
WooHoo!  Congrats to the CSIT community. -daw- On 02/19/2018 12:06 PM, Maciek Konstantynowicz (mkonstan) wrote: Dear FD.io community, *The CSIT rls1801 report is now published on FD.io docs site:* *html: *https://docs.fd.io/csit/rls1801/report/ *pdf:* https://

[vpp-dev] FD.io CSIT rls1801 report published

2018-02-19 Thread Maciek Konstantynowicz (mkonstan)
Dear FD.io community, The CSIT rls1801 report is now published on FD.io docs site: html: https://docs.fd.io/csit/rls1801/report/ pdf: https://docs.fd.io/csit/rls1801/report/_static/archive/csit_rls1801.pdf Many Thanks to All CSIT Contributors and Committers th

Re: [vpp-dev] TCP Throughput testing

2018-02-19 Thread Ray Kinsella
Insufficient hugepages in the VM - thanks for everyone's help! Ray K On 19/02/2018 15:38, Ray Kinsella wrote: Part of the mystery is solved in any case, Looks to be related to running inside the VM. DBGvpp# test tcp clients nclients 4 mbytes 256 test-timeout 100 uri tcp://192.168.1.1/9000

[vpp-dev] maximum number of interfaces ? #vpp

2018-02-19 Thread georgeroman
Hi All,  I am very new to VPP. I basically just compiled it, assigned some interfaces and was ready to start testing performance improvements over kernel with an ixia box.  All in all i want to use 8 x 10G interfaces in my setup and pass through them a total of ~36Gbps of traffic but i got stuc

Re: [vpp-dev] TCP Throughput testing

2018-02-19 Thread Ray Kinsella
Part of the mystery is solved in any case, Looks to be related to running inside the VM. DBGvpp# test tcp clients nclients 4 mbytes 256 test-timeout 100 uri tcp://192.168.1.1/9000 4 three-way handshakes in 4.10 seconds .98/s Test started at 10.538212 Test finished at 26.065026 1073741824 bytes

Re: [vpp-dev] vpp rspan

2018-02-19 Thread Neale Ranns
Support for ERSPAN: https://tools.ietf.org/html/draft-foschiano-erspan-03 was recently added: https://gerrit.fd.io/r/#/c/10524/ /neale From: on behalf of Gulakh Date: Monday, 19 February 2018 at 12:08 To: "vpp-dev@lists.fd.io" Subject: [vpp-dev] vpp rspan Hi I'm wondering if VPP support

Re: [vpp-dev] Unable to add interface to VPP on ARM64.

2018-02-19 Thread Christophe Fontaine
Hi, I never tried to do a virt-install on arm platform: I always use cloud images. (Either Centos or Fedora based). One more thing: maybe you also need to update other dependencies, along qemu. An apt-get update and/or apt-get upgrade may fix the issues you have. Christophe From: adarsh

Re: [vpp-dev] VPP Performance on preempt_rt Linux Kernel

2018-02-19 Thread Keith Burns
Something something corporate network, something switch to plain old internet something: https://wiki.fd.io/view/File:FD.io_mini-summit_916_Vhost_Performance_and_Optimization.pptx On Mon, Feb 19, 2018 at 6:00 AM Keith Burns wrote: > I thought Pierre Pfister (cc) had done some work with RT linu

Re: [vpp-dev] VPP Performance on preempt_rt Linux Kernel

2018-02-19 Thread Keith Burns
I thought Pierre Pfister (cc) had done some work with RT linux or at least with standard kernel and RT settings when doing vhost-user enhancements. I wanted to check his presentation at an fd.io mini-summit on our wiki, but it appears that wiki.fd.io is down (as is wiki.opendaylight.org at least f

Re: [vpp-dev] Unable to add interface to VPP on ARM64.

2018-02-19 Thread via Lists.Fd.Io
Hi, Thanks for the steps we were now able to install to 2.10 version but when we try to start VM we are getting below error, pls check and let us know if you faced the same or need any other change . ubuntu@vasily:~$ sudo virt-install --name dut3 --ram 4096 --disk path=dut3.img,size=30 --vcpus 2

Re: [vpp-dev] Unable to add interface to VPP on ARM64.

2018-02-19 Thread Christophe Fontaine
Hi, Which version of qemu do you use ? I had this kind of issue when using an old version of qemu: upgrading qemu to 2.10 resolved my issue. In qemu 2.5 (the default version with Ubuntu 16.04 in my case), the IO region is not mapped, which ends to this kind of error (the ressource0 is the IO r

Re: [vpp-dev] Unable to add interface to VPP on ARM64.

2018-02-19 Thread via Lists.Fd.Io
Hi, I am trying to  upgrade from 2.5.0 and it is showing this as the latest version We are using Ubuntu 16-04 version, can you pls let us know how you updated the version. ubuntu@ubuntu:~$ ubuntu@ubuntu:~$ qemu-system-aarch64 --version QEMU emulator version 2.5.0 (Debian 1:2.5+dfsg-5ubuntu10.21)

Re: [vpp-dev] Unable to add interface to VPP on ARM64.

2018-02-19 Thread Christophe Fontaine
Hi, As it is a dev platform, I upgraded qemu the ugly way, by upgrading the apt sources file. root@nxp2088:~# apt show qemu Package: qemu Version: 1:2.10+dfsg-0ubuntu3.4 Priority: optional Section: universe/metapackages APT-Sources: http://us.ports.ubuntu.com/ubuntu-ports artful-updates/univer

Re: [vpp-dev] VPP vs XDP-eBPF performance numbers - for linux containers

2018-02-19 Thread Ray Kinsella
Hi Avi, The only thing I would add a priori. Is that the cost of scalar packet processing in the kernel may mean that the cost of the single packet copy will be comparatively in the noise, and that userspace vSwitch will give you additional features/flexibility versus what you can realize in

Re: [vpp-dev] VPP Performance on preempt_rt Linux Kernel

2018-02-19 Thread Ray Kinsella
Not sure of any testing with the RT Linux Kernel. Have you seen the latency graphs? https://docs.fd.io/csit/rls1710/report/vpp_performance_tests/packet_latency_graphs/index.html Ray K On 16/02/2018 19:24, David Bainbridge wrote: Has there been any performance testing of VPP using a RT linux ke

[vpp-dev] vpp rspan

2018-02-19 Thread Gulakh
Hi I'm wondering if VPP supports RSPAN. I've found the documentation about SPAN (port mirroring) but nothing about RemoteSPAN. I would appreciate if you could help. thnks

Re: [vpp-dev] VPP vs XDP-eBPF performance numbers - for linux containers

2018-02-19 Thread Jerome Tollet
Hi Avi, As I already told you in a previous message. With tapv2, there's no longer copy from user to kernel (there's still kernel to user copy). Jerome Le 19/02/2018 09:35, « vpp-dev@lists.fd.io au nom de Avi Cohen (A) » a écrit : Thank you Ray dpdk is not running in the container -

Re: [vpp-dev] VPP vs XDP-eBPF performance numbers - for linux containers

2018-02-19 Thread Avi Cohen (A)
Thank you Ray dpdk is not running in the container - this is customer container and I cannot force the customer to use dpdk. Regarding the "not compatible" - I mean pkts received at vpp user-space, destined to C1 (container) , should be copied to the kernel and then to C1 IP-stack (e.g. AF_P

Re: [vpp-dev] Mellanox Connectx-4 - /nasm-2.13.01: No such file or directory

2018-02-19 Thread Avi Cohen (A)
Thank you Venkat and John I'll check that and update Avi > -Original Message- > From: vpp-dev@lists.fd.io [mailto:vpp-dev@lists.fd.io] On Behalf Of murali > venkat > Sent: Wednesday, 14 February, 2018 8:16 PM > To: vpp-dev@lists.fd.io > Subject: Re: [vpp-dev] Mellanox Connectx-4 - /nasm-2