from:"Kinsella, Ray"

Re: [vpp-dev] VPP performance tuning guide

2018-01-25 Thread Kinsella, Ray


Hi Charlie,


Please see this preso and associated paper from Kubecon - it is a pretty 
comprehensive guide on where to start.


https://wiki.fd.io/images/3/31/Benchmarking-sw-data-planes-Dec5_2017.pdf

Also consider replicating the CSIT test environment locally.
It is very well documented and you can then inspect the best way to 
benchmark VPP etc.


Ray K


On 24/01/2018 21:23, Li, Charlie wrote:

Hi All,

I am looking for some guidance on how to tune the VPP performance, specifically 
how to tune the IP Forwarding performance for small packets.

Appreciate if someone can point me to some documents or online resources on 
this topic.

My setup is simple, just an XL710-QDA2 NIC card with two 40G ports and VPP is 
configured to forward IP traffic between the two ports.

Basically I am using the default /etc/vpp/startup.conf (with PCI device address 
and core numbers modified). The throughput for bi-directional traffic with 
small packets is far below line rate. Throwing more cores does not seem to 
improve.

Regards,
Charlie Li

___
vpp-dev mailing list
vpp-dev@lists.fd.io
https://lists.fd.io/mailman/listinfo/vpp-dev


___
vpp-dev mailing list
vpp-dev@lists.fd.io
https://lists.fd.io/mailman/listinfo/vpp-dev

Re: [vpp-dev] VPP 18.01 Release artifacts are now available on nexus.fd.io

2018-01-25 Thread Kinsella, Ray


W00t!

On 25/01/2018 05:23, Dave Wallace wrote:

Folks,

The VPP 18.01 Release artifacts are now available on nexus.fd.io

The ubuntu.xenial and centos packages can be installed following the 
recipe on the wiki: 
https://wiki.fd.io/view/VPP/Installing_VPP_binaries_from_packages


Thank you to all of the VPP community who have contributed to the 18.01 
VPP Release.



Elvis has left the building!
-daw-



___
vpp-dev mailing list
vpp-dev@lists.fd.io
https://lists.fd.io/mailman/listinfo/vpp-dev


___
vpp-dev mailing list
vpp-dev@lists.fd.io
https://lists.fd.io/mailman/listinfo/vpp-dev

Re: [vpp-dev] [tsc] [nsh_sfc-dev] Stepping down as NSH_SFC PTL

2018-01-17 Thread Kinsella, Ray


Hongjun is elected as NSH_SFC PLT.
Trishan – can you make the update from the LF side?

Thanks and congrats Hongjun,

Ray K

From: tsc-boun...@lists.fd.io [mailto:tsc-boun...@lists.fd.io] On Behalf Of 
Keith Burns
Sent: Wednesday 10 January 2018 00:27
To: Zhou, Danny 
Cc: t...@lists.fd.io; nsh_sfc-...@lists.fd.io; vpp-dev@lists.fd.io
Subject: Re: [tsc] [vpp-dev] [nsh_sfc-dev] Stepping down as NSH_SFC PTL

+1 Hongjun!

On Jan 4, 2018 7:15 PM, "Zhou, Danny" 
mailto:danny.z...@intel.com>> wrote:
+1 for Hongjun

From: a...@yunify.org 
[mailto:a...@yunify.org]
Sent: Friday, January 5, 2018 9:06 AM
To: Zhou, Danny mailto:danny.z...@intel.com>>; 
t...@lists.fd.io; 
nsh_sfc-...@lists.fd.io
Cc: vpp-dev@lists.fd.io
Subject: Re: [nsh_sfc-dev] Stepping down as NSH_SFC PTL

+1

Thanks for all your hard work!

Ash

Ash Young
M: 408-628-7223

On Wed, Jan 3, 2018 at 8:32 PM -0600, "Zhou, Danny" 
mailto:danny.z...@intel.com>> wrote:
Hi TSC,
I've been taking the NSH_SFC PTL role 1 year ago. It has been a pleasure, but 
also a burden as there are some personal reasons make me cannot focus on 
NSH_SFC anymore, so I am stepping down as NSH_SFC PTL with immediate effect. 
Additionally, I would like to nominate Hongjun Ni, who is the most active 
NSH_SFC maintainer in 2017, as new NSH_SFC PTL for his excellent contribution 
as below:

•  Upstreamed 42 patches totally cover Ingress Classifier, SFF, Egress 
Classifier, NSH-Proxy, NSH-aware NAT, MD-Type2, NSH over Ethernet and NSH TTL.

•  Published four NSH_SFC Releases: 17.01, 17.04, 17.07 and 17.10.

NSH_SFC maintainers,

Pls vote for Hongjun Ni by directly replying this email.

-Danny

___
vpp-dev mailing list
vpp-dev@lists.fd.io
https://lists.fd.io/mailman/listinfo/vpp-dev
___
vpp-dev mailing list
vpp-dev@lists.fd.io
https://lists.fd.io/mailman/listinfo/vpp-dev

Re: [vpp-dev] VPP Design Documents

2017-12-21 Thread Kinsella, Ray


In fact we do.

VPP Design and Implementation Preso.
https://wiki.fd.io/view/File:Fd.io_vpp_overview_29.03.17.pptx

See Course 8 in Intel Network Builders (Free) - Overview of VPP.
https://builders.intel.com/university/networkbuilders/coursescategory/basic-training

I think you get the slides also with the basic training.
Shooting the video was fun (not).

Ray K

On 20/12/2017 11:53, Damjan Marion wrote:

Sorry, we don't have such documents



On 19 Dec 2017, at 14:29, Holoo Gulakh  wrote:

Hi,
I want to analyze and understand VPP code, but as you know getting insight into 
how things are implemented from the source code is difficult.

Can I have VPP design documents (software engineering) to understand:
+how project is structured and divided in multiple parts??
+how different parts communicate each other??
+how to make change in some part of code??
and so on

I need more details than what provided in wiki and readme files.

thanks in advance
___
vpp-dev mailing list
vpp-dev@lists.fd.io
https://lists.fd.io/mailman/listinfo/vpp-dev

___
vpp-dev mailing list
vpp-dev@lists.fd.io
https://lists.fd.io/mailman/listinfo/vpp-dev



___
vpp-dev mailing list
vpp-dev@lists.fd.io
https://lists.fd.io/mailman/listinfo/vpp-dev

Re: [vpp-dev] Measure plugin performances

2017-12-11 Thread Kinsella, Ray


Justin,

Suggest adding vpphost and iperfvpp to different network namespaces.
All that is happening at the moment is the kernel is recognizing it owns 
both IPs and bypassing everything.


You also need to add vppiperf and vppout to a bridge domain in VPP.

Take a look at VPP bootstrap in vppsb, it does 98% of what you are 
trying to do here.


https://git.fd.io/vppsb/tree/vpp-bootstrap

Ray K


On 11/12/2017 12:18, Justin Iurman wrote:

Hi everyone,

Just wanted to know if there is already something inside VPP to measure some metrics when 
testing performances of our own plugin ? I know there are some useful info with 
"vppctl show ..." but is it enough/all ?

Actually, based on this link ( 
https://wiki.fd.io/view/VPP/Progressive_VPP_Tutorial#Exercise:_Connecting_two_vpp_instances
 ), I was trying to reproduce the almost-the-same topology combined with iperf. 
Here is how it looks like (hope format doesn't get scratched here):
   _   __
  | |host-vppiperf   iperfvpp |  |
  |   VPP   |--- 10.10.2.1/24 --- 10.10.2.2/24 ---|  SERVER  |
  |_| |__|
   |
   host-vppout
  10.10.1.2/24
   |
   |
   |
vpphost (CLIENT)
  10.10.1.1/24

And here is the VPP-topology-building procedure:

## BEGIN ##
## create vpp and connect it to host
sudo vpp unix { log /tmp/vpp.log cli-listen /run/vpp/cli-vpp.sock } api-segment 
{ prefix vpp } plugins { plugin dpdk_plugin.so { disable } }

# veth interface between host and vpp
sudo ip link add name vppout type veth peer name vpphost
sudo ip link set dev vppout up
sudo ip link set dev vpphost up
sudo ip addr add 10.10.1.1/24 dev vpphost

# host interface between host and vpp
sudo vppctl -s /run/vpp/cli-vpp.sock create host-interface name vppout
sudo vppctl -s /run/vpp/cli-vpp.sock set int state host-vppout up
sudo vppctl -s /run/vpp/cli-vpp.sock set int ip address host-vppout 10.10.1.2/24

## create iperf server and connect it to vpp
sudo ip link add name vppiperf type veth peer name iperfvpp
sudo ip link set dev vppiperf up
sudo ip link set dev iperfvpp up
sudo ip addr add 10.10.2.2/24 dev iperfvpp

sudo vppctl -s /run/vpp/cli-vpp.sock create host-interface name vppiperf
sudo vppctl -s /run/vpp/cli-vpp.sock set int state host-vppiperf up
sudo vppctl -s /run/vpp/cli-vpp.sock set int ip address host-vppiperf 
10.10.2.1/24

#disabled
#sudo ip route add 10.10.2.0/24 via 10.10.1.2
#sudo ip route add 10.10.1.0/24 via 10.10.2.1
## END ##

When running iperf:

# server bound to iperfvpp (10.10.2.2)
iperf3 -s -B 10.10.2.2 -D
# client
iperf3 -c 10.10.2.2 -B 10.10.1.1 -u -4 -k 10

It works but there is no trace of those packets inside VPP. Note: pings work 
and I'm able to see them inside VPP. What am I missing ? I think I can guess 
what the problem is but still don't know how to solve it. Any tips/help ? 
Thanks !

Justin
___
vpp-dev mailing list
vpp-dev@lists.fd.io
https://lists.fd.io/mailman/listinfo/vpp-dev



___
vpp-dev mailing list
vpp-dev@lists.fd.io
https://lists.fd.io/mailman/listinfo/vpp-dev

Re: [vpp-dev] How can i use VPP as MPLS PE/P device

2017-12-11 Thread Kinsella, Ray


At the moment - there is no direct/easy way to do this AFAIK.

Router plugin is the best example of this, the other option is to use HC 
instead of FRR for BGP.


Ray K


On 08/12/2017 21:23, Michael Borokhovich wrote:
So, for the control plane, we can use e.g., FRR that will populate 
Linux's FIB and MPLS table. Then, we need to sync this info to VPP's 
FIB and VPP's MPLS table.


While the "router plugin" provides support for FIB synchronization, 
there is no support for MPLS sync. Does anyone know if there are plans 
to add this MPLS support to router plugin? Otherwise, what would be 
the alternative best way of synchronizing Linux MPLS table with VPP?


Thanks!
Michael.


On Wed, Dec 6, 2017 at 2:15 PM, Luke, Chris > wrote:


But to make sure we’re clear, while VPP can provide the dataplane
of a P/PE, but something else has to provide the control plane
(eg, LDP, BGP, SDN controller, etc)

Chris.

*From: *mailto:vpp-dev-boun...@lists.fd.io>> on behalf of "Neale Ranns
(nranns)" mailto:nra...@cisco.com>>
*Date: *Wednesday, December 6, 2017 at 09:33
*To: *"wangchuan...@163.com "
mailto:wangchuan...@163.com>>, vpp-dev
mailto:vpp-dev@lists.fd.io>>
*Subject: *Re: [vpp-dev] How can i use VPP as MPLS PE/P device

Another hastily assembled, on-demand guide:

https://wiki.fd.io/view/VPP/MPLS_FIB

/neale

*From: *mailto:vpp-dev-boun...@lists.fd.io>> on behalf of
"wangchuan...@163.com "
mailto:wangchuan...@163.com>>
*Date: *Wednesday, 6 December 2017 at 09:11
*To: *vpp-dev mailto:vpp-dev@lists.fd.io>>
*Subject: *[vpp-dev] How can i use VPP as MPLS PE/P device

hi all，

    I want to configure my testing MPLS network.

how can I configure VPP to act as PE or P using CLI cmd?

who can help?




best regards!



simon wang


___
vpp-dev mailing list
vpp-dev@lists.fd.io 
https://lists.fd.io/mailman/listinfo/vpp-dev





___
vpp-dev mailing list
vpp-dev@lists.fd.io
https://lists.fd.io/mailman/listinfo/vpp-dev


___
vpp-dev mailing list
vpp-dev@lists.fd.io
https://lists.fd.io/mailman/listinfo/vpp-dev

Re: [vpp-dev] i40e in a sorry state?

2017-12-07 Thread Kinsella, Ray


perfect, happy to have helped.


On 06/12/2017 13:46, Jon Loeliger wrote:

that something else was happening.  I dug a little deeper and found an


___
vpp-dev mailing list
vpp-dev@lists.fd.io
https://lists.fd.io/mailman/listinfo/vpp-dev

Re: [vpp-dev] Building and running sample plugin

2017-12-07 Thread Kinsella, Ray



You can find previous guidance on how to build it here.

https://docs.fd.io/vpp/17.10/sample_plugin_doc.html

Ray K

On 06/12/2017 19:44, Pradeep Patel (pradpate) wrote:


I am trying to build and run sample plugin using make option. I see 
sample plugin so gets created but failed due to undefined symbol: 
sample_main. Any pointers with be helpful.


Regards,
Pradeep

>make build SAMPLE_PLUGIN=yes

> make run SAMPLE_PLUGIN=yes

vagrant@localhost:/vpp$ make run SAMPLE_PLUGIN=yes

WARNING: STARTUP_CONF not defined or file doesn't exist.

Running with minimal startup config:  unix { interactive cli-listen 
/run/vpp/cli.sock gid 1000 }


vlib_plugin_early_init:356: plugin path 
/vpp/build-root/install-vpp_debug-native/sample-plugin/lib64/vpp_plugins:/vpp/build-root/install-vpp_debug-native/vpp/lib64/vpp_plugins


load_one_plugin:184: Loaded plugin: acl_plugin.so (Access Control Lists)

load_one_plugin:184: Loaded plugin: dpdk_plugin.so (Data Plane 
Development Kit (DPDK))


load_one_plugin:184: Loaded plugin: flowprobe_plugin.so (Flow per Packet)

load_one_plugin:184: Loaded plugin: gtpu_plugin.so (GTPv1-U)

load_one_plugin:184: Loaded plugin: ila_plugin.so (Identifier-locator 
addressing for IPv6)


load_one_plugin:184: Loaded plugin: ioam_plugin.so (Inbound OAM)

load_one_plugin:114: Plugin disabled (default): ixge_plugin.so

load_one_plugin:184: Loaded plugin: lb_plugin.so (Load Balancer)

load_one_plugin:184: Loaded plugin: libsixrd_plugin.so (IPv6 Rapid 
Deployment on IPv4 Infrastructure (RFC5969))


load_one_plugin:184: Loaded plugin: memif_plugin.so (Packet Memory 
Interface (experimetal))


load_one_plugin:184: Loaded plugin: nat_plugin.so (Network Address 
Translation)


load_one_plugin:184: Loaded plugin: pppoe_plugin.so (PPPoE)

load_one_plugin:142: 
/vpp/build-root/install-vpp_debug-native/sample-plugin/lib64/vpp_plugins/sample_plugin.so: 
undefined symbol: sample_main


load_one_plugin:143: Failed to load plugin 'sample_plugin.so'

Aborted

Makefile:434: recipe for target 'run' failed

make: *** [run] Error 134

vagrant@localhost:/vpp$ vim ^C



___
vpp-dev mailing list
vpp-dev@lists.fd.io
https://lists.fd.io/mailman/listinfo/vpp-dev


___
vpp-dev mailing list
vpp-dev@lists.fd.io
https://lists.fd.io/mailman/listinfo/vpp-dev

Re: [vpp-dev] i40e in a sorry state?

2017-12-06 Thread Kinsella, Ray



We may need the contents of PFHMC_ERRORINFO and PFHMC_ERRORDATA 
registers to figure this out.

I suspect that this may be something to do with interrupts being enabled.
However from reading the code, looks like interrupts should be disabled 
by default unless explictly enabled.


Ray K

https://www.intel.com/content/dam/www/public/us/en/documents/datasheets/xl710-10-40-controller-datasheet.pdf

Section 7.9.5

HMC related errors are reported through the PFHMC_ERRORINFO (see Section 
11.1.2.8.56) and PFHMC_ERRORDATA (see Section 11.1.2.8.57) registers. 
The HMC_ERR interrupt status bit in the PFINT_ICR0 register may also 
deliver an interrupt for HMC errors if the interrupt is enabled in the 
PFINT_ICR0_ENA register. When the HMC detects an error, it sets the 
PFHMC_ERRORINFO.ERROR_DETECTED bit along with the relevant information 
in the other fields of the PFHMC_ERRORINFO and PFHMC_ERRORDATA 
registers. No further notification of subsequent HMC errors associated 
with any given PF will be issued until the current error is acknowledged 
by writing a 0 to the PFHMC_ERRORINFO.ERROR_DETECTED bit. Table 7-187 
describes the errors detected for each HMC object and the behavior 
associated with each error.



On 05/12/2017 19:37, Jon Loeliger wrote:

Guys,

I've updated VPP to vpp.x86_64 0:18.01-rc0~405_g7f0d1d3 and when
I enable a interface, I get this love note:

    i40e_dev_interrupt_handler(): ICR0: HMC error
    Aborted

This used to work, of course.  Some more details below.

Any notions?

Thanks,
jdl


# cat /etc/vpp/startup.conf

unix {
    nodaemon
    log /tmp/vpp.log
    full-coredump
    cli-listen /run/vpp/cli.sock
    gid vpp
}

dpdk {
    uio-driver igb_uio
}

api-trace {
    on
}

api-segment {
    gid vpp
}



# /usr/bin/vpp -c /etc/vpp/startup.conf
vlib_plugin_early_init:356: plugin path /usr/lib/vpp_plugins
load_one_plugin:184: Loaded plugin: acl_plugin.so (Access Control Lists)
load_one_plugin:184: Loaded plugin: dpdk_plugin.so (Data Plane 
Development Kit (DPDK))

load_one_plugin:184: Loaded plugin: flowprobe_plugin.so (Flow per Packet)
load_one_plugin:184: Loaded plugin: gtpu_plugin.so (GTPv1-U)
load_one_plugin:184: Loaded plugin: ila_plugin.so (Identifier-locator 
addressing for IPv6)

load_one_plugin:184: Loaded plugin: ioam_plugin.so (Inbound OAM)
load_one_plugin:114: Plugin disabled (default): ixge_plugin.so
load_one_plugin:184: Loaded plugin: lb_plugin.so (Load Balancer)
load_one_plugin:184: Loaded plugin: libsixrd_plugin.so (IPv6 Rapid 
Deployment on IPv4 Infrastructure (RFC5969))
load_one_plugin:184: Loaded plugin: memif_plugin.so (Packet Memory 
Interface (experimetal))
load_one_plugin:184: Loaded plugin: nat_plugin.so (Network Address 
Translation)

load_one_plugin:184: Loaded plugin: pppoe_plugin.so (PPPoE)
load_one_plugin:184: Loaded plugin: router.so (router)
load_one_plugin:184: Loaded plugin: stn_plugin.so (VPP Steals the NIC 
for Container integration)
/usr/bin/vpp[4604]: tap_inject_interface_add_del:474: 
tap_inject_interface_add_del: Adding interface with hw_if_index 0
/usr/bin/vpp[4604]: tap_inject_is_config_enabled:122: 
tap_inject_is_config_enabled: Value of im->flags is 0
/usr/bin/vpp[4604]: tap_inject_interface_add_del:477: 
tap_inject_interface_add_del: tap_inject is disabled in config
/usr/bin/vpp[4604]: load_one_plugin:63: Loaded plugin: 
/usr/lib/vpp_api_test_plugins/acl_test_plugin.so
/usr/bin/vpp[4604]: load_one_plugin:63: Loaded plugin: 
/usr/lib/vpp_api_test_plugins/dpdk_test_plugin.so
/usr/bin/vpp[4604]: load_one_plugin:63: Loaded plugin: 
/usr/lib/vpp_api_test_plugins/flowprobe_test_plugin.so
/usr/bin/vpp[4604]: load_one_plugin:63: Loaded plugin: 
/usr/lib/vpp_api_test_plugins/gtpu_test_plugin.so
/usr/bin/vpp[4604]: load_one_plugin:63: Loaded plugin: 
/usr/lib/vpp_api_test_plugins/ioam_export_test_plugin.so
/usr/bin/vpp[4604]: load_one_plugin:63: Loaded plugin: 
/usr/lib/vpp_api_test_plugins/ioam_pot_test_plugin.so
/usr/bin/vpp[4604]: load_one_plugin:63: Loaded plugin: 
/usr/lib/vpp_api_test_plugins/ioam_trace_test_plugin.so
/usr/bin/vpp[4604]: load_one_plugin:63: Loaded plugin: 
/usr/lib/vpp_api_test_plugins/ioam_vxlan_gpe_test_plugin.so
/usr/bin/vpp[4604]: load_one_plugin:63: Loaded plugin: 
/usr/lib/vpp_api_test_plugins/lb_test_plugin.so
/usr/bin/vpp[4604]: load_one_plugin:63: Loaded plugin: 
/usr/lib/vpp_api_test_plugins/memif_test_plugin.so
/usr/bin/vpp[4604]: load_one_plugin:63: Loaded plugin: 
/usr/lib/vpp_api_test_plugins/nat_test_plugin.so
/usr/bin/vpp[4604]: load_one_plugin:63: Loaded plugin: 
/usr/lib/vpp_api_test_plugins/pppoe_test_plugin.so
/usr/bin/vpp[4604]: load_one_plugin:63: Loaded plugin: 
/usr/lib/vpp_api_test_plugins/udp_ping_test_plugin.so
/usr/bin/vpp[4604]: load_one_plugin:63: Loaded plugin: 
/usr/lib/vpp_api_test_plugins/vxlan_gpe_ioam_export_test_plugin.so
/usr/bin/vpp[4604]: tap_inject_is_config_enabled:122: 
tap_inject_is_config_enabled: Value of im->flags is 0
/usr/bin/vpp[4604]: vlib_pci_bind_to_uio: Skipping PCI device 
:0

Re: [vpp-dev] Bug?

2017-12-05 Thread Kinsella, Ray

I imagine the bridge ports are being flooded looking for the final 
destination.



Ray K


On 05/12/2017 03:55, Yuliang Li wrote:

Hi,

I am using vpp v18.01-rc0~160-ge13c799 (show version in vppctl). I 
feel I find a bug.


I have two physical interfaces TenGigabitEthernet5/0/0 (int1) 
and TenGigabitEthernet5/0/1 (int2).


Normally, if I send traffic (from another server) to int1, I expect 
int2 should NOT receive the traffic. This is true when int2 is in l3 mode.


However, if I set int2 to l2 mode (e.g., set interface l2 bridge 
TenGigabitEthernet5/0/1 2), int2 will receive the same traffic as int1 
does. This problem happens no matter int1 is in l2 or l3 mode. I use 
the trace to confirm that int2 receives the same traffic as int1 does.


Does anyone know this problem?

Thanks,
--
Yuliang Li
PhD student
Department of Computer Science
Yale University


___
vpp-dev mailing list
vpp-dev@lists.fd.io
https://lists.fd.io/mailman/listinfo/vpp-dev


___
vpp-dev mailing list
vpp-dev@lists.fd.io
https://lists.fd.io/mailman/listinfo/vpp-dev

Re: [vpp-dev] FW: how to redirect specific UDP port data to a new node?

2017-12-04 Thread Kinsella, Ray


Another good example is vxlan.
Not sure what will happen if multiple graph nodes are fighting for the 
same port though.


Ray K

https://git.fd.io/vpp/tree/src/vnet/vxlan/vxlan.c

|clib_error_t *vxlan_init (vlib_main_t *vm) { vxlan_main_t * vxm = 
&vxlan_main; vxm->vnet_main = vnet_get_main(); vxm->vlib_main = vm; /* 
initialize the ip6 hash */ vxm->vxlan6_tunnel_by_key = 
hash_create_mem(0, sizeof(vxlan6_tunnel_key_t), sizeof(uword)); 
vxm->vtep6 = hash_create_mem(0, sizeof(ip6_address_t), sizeof(uword)); 
vxm->mcast_shared = hash_create_mem(0, sizeof(ip46_address_t), 
sizeof(mcast_shared_t)); udp_register_dst_port (vm, UDP_DST_PORT_vxlan, 
vxlan4_input_node.index, /* is_ip4 */ 1); udp_register_dst_port (vm, 
UDP_DST_PORT_vxlan6, vxlan6_input_node.index, /* is_ip4 */ 0); 
fib_node_register_type(FIB_NODE_TYPE_VXLAN_TUNNEL, &vxlan_vft); return 
0; } |




On 04/12/2017 08:19, Klement Sekera -X (ksekera - PANTHEON TECHNOLOGIES 
at Cisco) wrote:

Are you implementing a new graph node, which consumes UDP traffic? If so, then 
the BFD example is a way to redirect interesting traffic to your new node...


-Original Message-
From: lin huang [mailto:mit...@outlook.com]
Sent: Monday, December 4, 2017 9:16 AM
To: Klement Sekera -X (ksekera - PANTHEON TECHNOLOGIES at Cisco)
; vpp-dev@lists.fd.io
Subject: RE: [vpp-dev] FW: how to redirect specific UDP port data to a new
node?

Hi Klement,

I want to redirect data to a new udp port. Do I need to modify VPP
code?
Adding a new UDP_DST_PORT_ enum types??

Thanks a lot!!!


-Original Message-
From: Klement Sekera -X (ksekera - PANTHEON TECHNOLOGIES at Cisco)
[mailto:ksek...@cisco.com]
Sent: Monday, December 04, 2017 4:12 PM
To: lin huang ; vpp-dev@lists.fd.io
Subject: RE: [vpp-dev] FW: how to redirect specific UDP port data to a
new node?

Hi Lin Huang,

You can take a look at bfd_udp_init() in bfd_udp.c to see how the BFD
feature registers to listen on BFD ports using udp_register_dst_port().

Thanks,
Klement


-Original Message-
From: vpp-dev-boun...@lists.fd.io
[mailto:vpp-dev-boun...@lists.fd.io]
On Behalf Of lin huang
Sent: Monday, December 4, 2017 9:03 AM
To: vpp-dev@lists.fd.io
Subject: [vpp-dev] FW: how to redirect specific UDP port data to a new

node?





From: lin huang
Sent: Monday, December 04, 2017 2:53 PM
To: vpp-dev-boun...@lists.fd.io
Subject: how to redirect specific UDP port data to a new node?



Hi, all.

  I want to know how to redirect specific UDP port data to a
new

node?

Dose udp_register_dst_port make effecet? Or using the arc with feature?



Thanks!!!

___
vpp-dev mailing list
vpp-dev@lists.fd.io
https://lists.fd.io/mailman/listinfo/vpp-dev



___
vpp-dev mailing list
vpp-dev@lists.fd.io
https://lists.fd.io/mailman/listinfo/vpp-dev

Re: [vpp-dev] VPP high level working flowchart

2017-12-04 Thread Kinsella, Ray


Hi Holoo,

Can I recommend that you play at bit with VPP's packet tracing facility.

The following wiki page explains setting up the tracer a bit.

https://wiki.fd.io/view/VPP/How_To_Connect_A_PCI_Interface_To_VPP#Show_trace

You will see from tracing a that FD.io VPP exists independently of the 
Linux Networking Stack.

With it's own L2 + L3 implementations.

Ray K


On 04/12/2017 07:28, Holoo Gulakh wrote:

Hi,
According to my researches around VPP and its workflow, I have 
prepared a flowchart showing VPP's workflow.


This flowchart says that all ingress packet to interfaces handled by 
VPP are sent to VPP without Linux network stack intervention. 
afterward based on the packet type/destination one of the flows goes on.


Is this flowchart describing VPP's workflow correctly?? (it might be 
helpful if you add more details to it if it's correct OR otherwise 
correct it if it's wrong)


thanks in advace


___
vpp-dev mailing list
vpp-dev@lists.fd.io
https://lists.fd.io/mailman/listinfo/vpp-dev


___
vpp-dev mailing list
vpp-dev@lists.fd.io
https://lists.fd.io/mailman/listinfo/vpp-dev

Re: [vpp-dev] Router plugin capability question (VRFs and MPLS)

2017-11-24 Thread Kinsella, Ray



If you mean the router plugin from the sandbox,
the short answer is yes, it doesn't support any of these.

Better approach is to use Honeycomb, with or without ODL.

Ray K


On 22/11/2017 18:53, Michael Borokhovich wrote:

Hi,

Does router plugin support the following features?

1) Multiple VRFs
2) MPLS

From our initial experiments, the above features are not supported.
Multiple VRFs do not work (I tried with namespaces).
The MPLS information is not passed from Linux to VPP's FIBs.

Please let me know what you think.

Thanks,
Michael.



___
vpp-dev mailing list
vpp-dev@lists.fd.io
https://lists.fd.io/mailman/listinfo/vpp-dev


___
vpp-dev mailing list
vpp-dev@lists.fd.io
https://lists.fd.io/mailman/listinfo/vpp-dev

Re: [vpp-dev] VM instantiation failed when I got a tutorial through

2017-11-24 Thread Kinsella, Ray


Try sending the VM's console to stdout and see what you learn.

||-append "|console=ttyS0" -serial stdio|


Ray K

On 24/11/2017 06:48, 박민철 wrote:

Hi,
I'm new to vpp and trying to bench mark this.
I followed the tutorial, 
https://wiki.fd.io/view/VPP/Use_VPP_to_connect_VMs_Using_Vhost-User_Interface.
When I try to instantiate the VM with below command, it seems like to 
be hanged and there's no reponse from terminal.


qemu-system-x86_64 \
 -enable-kvm -m 1024 \
 -bios OVMF.fd \
 -smp 4 -cpu host \
 -vga none -nographic \
 -drive file="1-clear-14200-kvm.img",if=virtio,aio=threads \
 -chardev socket,id=char1,path=/tmp/sock1.sock \
 -netdev type=vhost-user,id=mynet1,chardev=char1,vhostforce \
 -device virtio-net-pci,mac=00:00:00:00:00:01,netdev=mynet1 \
 -object 
memory-backend-file,id=mem,size=1024M,mem-path=/dev/hugepages,share=on \
 -numa node,memdev=mem -mem-prealloc \
 -debugcon file:debug.log -global isa-debugcon.iobase=0x402

Is there any one who faced with same problem as me?
My environment is like this,
 - Dell R730
 - Intel X720 NIC
 - 16 Core, 2 Socket
 - 64GB Ram
 - Ubuntu 16.04

Regards,




___
vpp-dev mailing list
vpp-dev@lists.fd.io
https://lists.fd.io/mailman/listinfo/vpp-dev


___
vpp-dev mailing list
vpp-dev@lists.fd.io
https://lists.fd.io/mailman/listinfo/vpp-dev

[vpp-dev] FOSDEM 2018

2017-11-16 Thread Kinsella, Ray


Folks,

The Call for Content for FOSDEM 2018 is closing today!
The "SDN and NFV room" at FOSDEM is great way to get the FD.io message 
out there.


Last minute submissions would be most welcome!
To submit see,

https://blogs.gnome.org/bolsh/2017/11/01/fosdem-2018-sdnnfv-devroom-call-for-content/

Thanks,

Ray K

___
vpp-dev mailing list
vpp-dev@lists.fd.io
https://lists.fd.io/mailman/listinfo/vpp-dev

Re: [vpp-dev] FD.io CSIT rls1710 report published

2017-11-03 Thread Kinsella, Ray


w00t!


On 02/11/2017 17:32, Maciek Konstantynowicz (mkonstan) wrote:

Dear FD.io  community,

The CSIT rls1710 report is now published on FD.io  docs 
site:
https://docs.fd.io/csit/rls1710/report/ 



Many Thanks to All CSIT Contributors and Committers to make CSIT rls1710
happen !

Many Thanks to all folks that collaborated with CSIT including VPP
project, HoneyComb, goVPP, TRex and other sub-projects.

Linked report includes results of i) VPP performance and functional
tests, ii) Testpmd performance tests, iii) Honeycomb functional tests,
and iv) reference to VPP unit tests.

Points of note in the report:

    - CSIT infrastructure changes and added tests including VPP, HoneyComb

https://docs.fd.io/csit/rls1710/report/vpp_performance_tests/csit_release_notes.html#changes-in-csit-release
https://docs.fd.io/csit/rls1710/report/honeycomb_functional_tests/csit_release_notes.html#changes-in-csit-release

    - Measured VPP performance improvements and changes

https://docs.fd.io/csit/rls1710/report/vpp_performance_tests/csit_release_notes.html#performance-improvements
https://docs.fd.io/csit/rls1710/report/vpp_performance_tests/csit_release_notes.html#other-performance-changes

    - Performance graphs - throughput and latency for VPP and DPDK-Testpmd

https://docs.fd.io/csit/rls1710/report/vpp_performance_tests/packet_throughput_graphs/index.html
https://docs.fd.io/csit/rls1710/report/vpp_performance_tests/packet_latency_graphs/

https://docs.fd.io/csit/rls1710/report/dpdk_performance_tests/packet_throughput_graphs/index.html
https://docs.fd.io/csit/rls1710/report/dpdk_performance_tests/packet_latency_graphs/index.html

   - VPP configs used per test case

https://docs.fd.io/csit/rls1710/report/test_configuration/vpp_performance_configuration/index.html
https://docs.fd.io/csit/rls1710/report/test_configuration/vpp_functional_configuration/index.html

   - VPP operational data - "show runtime” outputs at NDR rate with 
CPU core

     clock cycles spent per worker node per packet

https://docs.fd.io/csit/rls1710/report/test_operational_data/vpp_performance_operational_data/index.html

    - CSIT Framework Documentation - including the new Presentation 
and Analytics Layer


https://docs.fd.io/csit/rls1710/report/csit_framework_documentation/index.html
https://docs.fd.io/csit/rls1710/report/csit_framework_documentation/pal_lld.html

Please send your feedback and comments by email to 
csit-...@lists.fd.io .


-Maciek
On behalf of FD.io  CSIT project.


___
vpp-dev mailing list
vpp-dev@lists.fd.io
https://lists.fd.io/mailman/listinfo/vpp-dev


___
vpp-dev mailing list
vpp-dev@lists.fd.io
https://lists.fd.io/mailman/listinfo/vpp-dev

Re: [vpp-dev] VPP 17.10 Released

2017-10-27 Thread Kinsella, Ray


Much kudos


On 26/10/2017 22:36, Florin Coras wrote:

Dear VPP community,

The VPP 17.10 release is complete. The release artefacts are now available on 
the nexus server and the repos have been updated.

I’d like to take this opportunity to thank you all for your continued support 
for VPP.

Regards,
Florin
___
vpp-dev mailing list
vpp-dev@lists.fd.io
https://lists.fd.io/mailman/listinfo/vpp-dev


___
vpp-dev mailing list
vpp-dev@lists.fd.io
https://lists.fd.io/mailman/listinfo/vpp-dev

Re: [vpp-dev] What can be traced?

2017-10-25 Thread Kinsella, Ray


10 is how many frame to capture in the trace.

Ray K


On 25/10/2017 23:09, John Wei wrote:


*vppctl  trace add* af-packet-input 10
was mentioned in this write up:
https://wiki.fd.io/view/VPP/Progressive_VPP_Tutorial#Action:_Add_trace

Is there write-up on
- list of things that we can turn on tracing?
- what is the number "10" means?

John






___
vpp-dev mailing list
vpp-dev@lists.fd.io
https://lists.fd.io/mailman/listinfo/vpp-dev


___
vpp-dev mailing list
vpp-dev@lists.fd.io
https://lists.fd.io/mailman/listinfo/vpp-dev

Re: [vpp-dev] IS THERE ANY SETUP FOR BUILDING AND TESTING ALL THE FD.IO MODULES TOGETHER

2017-10-25 Thread Kinsella, Ray


Hi Sharath,

Not at the moment, you need to grab the ones you and build individually.

Which modules are you interested in?

Ray K


On 25/10/2017 13:28, Sharath Kumar wrote:

Hello all,

I am a newbie to fd.io .

I have started with setting up dev environment by following this link 
https://wiki.fd.io/view/VPP/Setting_Up_Your_Dev_Environment .


This setup is building only VPP libs. (so I was considering it as a 
VPP setup and not a FD.IO  setup)


IS THERE ANY SETUP FOR BUILDING AND TESTING ALL THE FD.IO 
 MODULES TOGETHER ??


I have searched for FD.IO  setup but I could find only 
the individual module setups.

can anyone please help me ?

With Thanks & Regards,
Sharath.




___
vpp-dev mailing list
vpp-dev@lists.fd.io
https://lists.fd.io/mailman/listinfo/vpp-dev


___
vpp-dev mailing list
vpp-dev@lists.fd.io
https://lists.fd.io/mailman/listinfo/vpp-dev

Re: [vpp-dev] vppctl -s option?

2017-10-25 Thread Kinsella, Ray


If you do a 'file' on vppctl now - you will see its a binary.
Strange must have been an installer issue.

Ray K

On 25/10/2017 02:06, John Wei wrote:

Hi Ray,
I upgraded my vpp version to 17.10-rc2. And vppctl does accept -s option.
Maybe we need to document that -s in only supported started 17.10?

John


On Tue, Oct 24, 2017 at 10:18 AM, John Wei <mailto:johnt...@gmail.com>> wrote:


Add file output:

file `type -p vppctl`
/bin/vppctl: Python script, ASCII text executable


On Tue, Oct 24, 2017 at 10:17 AM, John Wei mailto:johnt...@gmail.com>> wrote:

See the vppctl output below.

Thanks s for the help.
John

# vppctl  show int
              Name               Idx    State         
Counter          Count
local0                            0     down

# vppctl -s /run/vpp/cli-vpp.sock show int
Usage: vppctl [options]

vppctl: error: no such option: -s
# type -p vppctl
/bin/vppctl
# vppctl show ver
vpp v17.07.01-release built by jenkins on
centos7-basebuild-2c-8g-2985 at Fri Sep 15 10:32:55 UTC 2017

# vppctl -s /run/vpp/cli-vpp.sock show ver
Usage: vppctl [options]

vppctl: error: no such option: -s
# ls -l /run/vpp
total 0
srwxrwxr-x. 1 root root 0 Oct 23 19:29 cli-vpp1.sock
srwxrwxr-x. 1 root root 0 Oct 23 19:20 cli-vpp.sock


On Tue, Oct 24, 2017 at 2:17 AM, Kinsella, Ray mailto:m...@ashroe.eu>> wrote:

Hi John,

That is strange.
Can you do a

file 'type -p vppctl'

And report back ?

Thanks,

Ray K



On 24/10/2017 03:47, John Wei wrote:

I was reading the VPP tutorial below:

https://wiki.fd.io/view/VPP/Progressive_VPP_Tutorial#Exercise:_vpp_basics

<https://wiki.fd.io/view/VPP/Progressive_VPP_Tutorial#Exercise:_vpp_basics>

In the example, "vppctl -s /run/vpp/cli-vpp1.sock" was used.
But, my vppctl does not support "-s" option.

I have version v17.07.01.
Do I need to get newer version to get support of "-s"?

John





___
vpp-dev mailing list
vpp-dev@lists.fd.io <mailto:vpp-dev@lists.fd.io>
https://lists.fd.io/mailman/listinfo/vpp-dev
<https://lists.fd.io/mailman/listinfo/vpp-dev>



___
vpp-dev mailing list
vpp-dev@lists.fd.io <mailto:vpp-dev@lists.fd.io>
https://lists.fd.io/mailman/listinfo/vpp-dev
<https://lists.fd.io/mailman/listinfo/vpp-dev>






___
vpp-dev mailing list
vpp-dev@lists.fd.io
https://lists.fd.io/mailman/listinfo/vpp-dev

Re: [vpp-dev] vppctl -s option?

2017-10-24 Thread Kinsella, Ray


Hi John,

That is strange.
Can you do a

file 'type -p vppctl'

And report back ?

Thanks,

Ray K


On 24/10/2017 03:47, John Wei wrote:

I was reading the VPP tutorial below:
https://wiki.fd.io/view/VPP/Progressive_VPP_Tutorial#Exercise:_vpp_basics

In the example, "vppctl -s /run/vpp/cli-vpp1.sock" was used.
But, my vppctl does not support "-s" option.

I have version v17.07.01.
Do I need to get newer version to get support of "-s"?

John





___
vpp-dev mailing list
vpp-dev@lists.fd.io
https://lists.fd.io/mailman/listinfo/vpp-dev


___
vpp-dev mailing list
vpp-dev@lists.fd.io
https://lists.fd.io/mailman/listinfo/vpp-dev

[vpp-dev] Fwd: Call for Content for the SDN & NFV DevRoom 2018

2017-10-20 Thread Kinsella, Ray

For our Europe based folks, FOSDEM is always fun!

Ray K

-Original Message-
From: network-devroom-boun...@lists.fosdem.org 
[mailto:network-devroom-boun...@lists.fosdem.org] On Behalf Of Dave Neary
Sent: Monday 16 October 2017 21:53
To: network-devr...@lists.fosdem.org; FOSDEM Devroom Managers 

Subject: Call for Content for the SDN & NFV DevRoom 2018

Hi everyone,

We are pleased to announce the Call for Participation in the FOSDEM
2018 Software Defined Networking and Network Functions Virtualization DevRoom!

Important dates:
* Nov 16: Deadline for submissions
* Dec  1: Speakers notified of acceptance
* Dec  5: Schedule published

This year, as it has for the past two years, the DevRoom topics will cover two 
distinct fields:
* Software Defined Networking (SDN), covering virtual switching, open source 
SDN controllers, virtual routing
* Network Functions Virtualization (NFV), covering open source network 
functions, NFV management and orchestration tools, and topics related to the 
creation of an open source NFV platform

We are now inviting proposals for talks about Free/Libre/Open Source Software 
on the topics of SDN and NFV. This is an exciting and growing field, and FOSDEM 
gives an opportunity to reach a unique audience of very knowledgeable and 
highly technical free and open source software activists.

This year, the DevRoom will focus on the emergence of cloud native Virtual 
Network Functions, and the management and performance requirements of those 
applications, in addition to our traditional focus on high performance packet 
processing.

A representative, but not exhaustive, list of the projects and topics we would 
like to see on the schedule are:

* Low-level networking and switching: IOvisor, eBPF, XDP, fd.io, Open vSwitch, 
OpenDataplane, Free Range Routing, ...
* SDN controllers and overlay networking: OpenStack Neutron, Calico, 
OpenDaylight, ONOS, Plumgrid, OVN, OpenContrail, Midonet, ...
* NFV Management and Orchestration: ONAP, ManageIQ, Juju, OpenBaton, Tacker, 
OSM, network management, PNDA.io, ...
* NFV related features: Service Assurance, enforcement of Quality of Service, 
Service Function Chaining, fault management, dataplane acceleration, security, 
...

Talks should be aimed at a technical audience, but should not assume that 
attendees are already familiar with your project or how it solves a general 
problem. Talk proposals can be very specific solutions to a problem, or can be 
higher level project overviews for lesser known projects.

Please include the following information when submitting a proposal:

Your name
The title of your talk (please be descriptive, as titles will be listed 
with around 250 from other projects)
Short abstract of one or two paragraphs
Short bio (with photo)

The deadline for submissions is November 16th 2017. FOSDEM will be held on the 
weekend of February 3-4, 2018 and the SDN/NFV DevRoom will take place on 
Saturday, February 3, 2017. Please use the following website to submit your 
proposals:

  https://penta.fosdem.org/submission/FOSDEM18

(you do not need to create a new Pentabarf account if you already have one from 
past years).

You can also join the devroom’s mailing list, which is the official 
communication channel for the DevRoom: network-devr...@lists.fosdem.org 
(subscription page: https://lists.fosdem.org/listinfo/network-devroom)

– The Networking DevRoom 2018 Organization Team

--
Dave Neary - NFV/SDN Community Strategy
Open Source and Standards, Red Hat - http://community.redhat.com
Ph: +1-978-399-2182 / Cell: +1-978-799-3338 
___
network-devroom mailing list
network-devr...@lists.fosdem.org
https://lists.fosdem.org/listinfo/network-devroom

___
vpp-dev mailing list
vpp-dev@lists.fd.io
https://lists.fd.io/mailman/listinfo/vpp-dev

Re: [vpp-dev] vpp performnace measurement in sriov configuration.

2017-08-10 Thread Kinsella, Ray


Hi Rahul,

So there a few additional hoops you have to jump through to get good 
performance. The biggest of which is making sure that


1. The host is setup to make hugepages available.
2. QEMU is pointed at the hugepages.
3. The guest is setup to make hugepages available.

The following documented from DPDK covers it in some detail.

http://dpdk.readthedocs.io/en/v16.04/sample_app_ug/vhost.html

Ray K

On 09/08/2017 23:33, Rahul Negi wrote:

Hi All,
 I am working on an use case where i want to measure the Maximum PPS
handled by vpp in sriov configuration.I have created a virtual machine
on a host having specifications as follows:-
1.RHEL 7.3 installed
2.Intel X540 10 gig NIC attached

I have created a virtual function from one of the interface of 10 gig
NIC(for e.g ens3f0) and attached it to my Virtual Machine.I have
installed ubuntu 16.04 on my Vm. My vpp version running on vm is 17.10.

vpp# show version verbose
Version:  v17.10-rc0~86-g7d4a22c
Compiled by:  root
Compile host: ubuntu
Compile date: Wed Jul 26 18:56:51 EDT 2017
Compile location: /root/vpp/vpp
Compiler: GCC 5.4.0 20160609
Current PID:  5006

Currently My Vm is with 4vcpu and vpp  cpu model is as follow:-
1.one main thread
2.Three worker threads

I am not able to get more than 4Mpps with this configuration of vpp.When
i am pumping more than 4Mpps traffic to my VM. I can see this
counter rx_no_dma_resources: 5628104 getting incremented at the host.As
there is no eth stats available in sriov configuration for virtual
functions attached to VM.

Guest vcpus are pin to host physical cpus.

So 4Mpps is the expected number that we can get in this configuration?

Thanks,
Rahul


___
vpp-dev mailing list
vpp-dev@lists.fd.io
https://lists.fd.io/mailman/listinfo/vpp-dev


___
vpp-dev mailing list
vpp-dev@lists.fd.io
https://lists.fd.io/mailman/listinfo/vpp-dev

Re: [vpp-dev] 17.07 Release

2017-08-06 Thread Kinsella, Ray


Thanks !

On 31/07/2017 13:51, Neale Ranns (nranns) wrote:

Hi Chris,

Thanks for fixing it!
Release notes now available at:
  https://docs.fd.io/vpp/17.07/release_notes_1707.html

regards,
neale

-Original Message-
From: "Luke, Chris" 
Date: Monday, 31 July 2017 at 18:01
To: "Neale Ranns (nranns)" , "Kinsella, Ray" , 
"vpp-dev@lists.fd.io" 
Subject: Re: [vpp-dev] 17.07 Release

The next merge job to run on each branch will trigger it; or at least I 
hope it does.

I’ll do a ‘remerge’ on the current HEAD commit to see if that will do the 
needful

Chris.


On 7/31/17, 10:34, "vpp-dev-boun...@lists.fd.io on behalf of Neale Ranns 
(nranns)"  wrote:

Hi Ray,

The release notes will appear here eventually:
 https://docs.fd.io/vpp/17.07/release_notes.html

there was a breakage in the generation of the docs, which Chris fixed, 
and was recently merged:
  https://gerrit.fd.io/r/#/c/7818/

hopefully we will get the docs updated when the next patched is merged.

Regards,
neale

-Original Message-
From:  on behalf of "Kinsella, Ray" 

Date: Monday, 31 July 2017 at 15:42
To: "vpp-dev@lists.fd.io" 
Subject: Re: [vpp-dev] 17.07 Release

Hi Neale,

Thanks for this - great work.
Are there release notes archived anywhere?

Ray K

On 20/07/2017 16:56, Neale Ranns (nranns) wrote:
>
> Dear VPP community,
>
> The VPP 17.07 release is complete. The release artefacts are now 
available on the nexus server.
>
> I’d like to take this opportunity to thank you all for your 
continued support for VPP.
>
> Best regards,
> Neale
>
>
> ___
> vpp-dev mailing list
> vpp-dev@lists.fd.io
> https://lists.fd.io/mailman/listinfo/vpp-dev
>
___
vpp-dev mailing list
vpp-dev@lists.fd.io
https://lists.fd.io/mailman/listinfo/vpp-dev

___
vpp-dev mailing list
vpp-dev@lists.fd.io
https://lists.fd.io/mailman/listinfo/vpp-dev



___
vpp-dev mailing list
vpp-dev@lists.fd.io
https://lists.fd.io/mailman/listinfo/vpp-dev

Re: [vpp-dev] 17.07 Release

2017-07-31 Thread Kinsella, Ray


Hi Neale,

Thanks for this - great work.
Are there release notes archived anywhere?

Ray K

On 20/07/2017 16:56, Neale Ranns (nranns) wrote:


Dear VPP community,

The VPP 17.07 release is complete. The release artefacts are now available on 
the nexus server.

I’d like to take this opportunity to thank you all for your continued support 
for VPP.

Best regards,
Neale


___
vpp-dev mailing list
vpp-dev@lists.fd.io
https://lists.fd.io/mailman/listinfo/vpp-dev


___
vpp-dev mailing list
vpp-dev@lists.fd.io
https://lists.fd.io/mailman/listinfo/vpp-dev

Re: [vpp-dev] VPP/How To Build The Sample Plugin

2017-06-01 Thread Kinsella, Ray

Its conservative - makes fewer changes and preserves the separate 
packaging for the sample plugin, so its a good approach.


My only piece of feedback on it, is that we are starting to end up with 
build 'config' divided across 'build-data' and environment variables, 
which is something we may need to make consistent over time.


I don't know if the sample plugin is important enough to pollute the 
'make help' with detail of how to build it. I thought the 'build-data' 
config and markdown document describing the plugin and how to build it 
was a better approach.


In anycase (env var or build-data config), I am happy to excise from my 
original patch, the markdown doc, update it with the new build details 
and doxygen comments etc and submit those seperately. Once this patch 
goes through.


Ray K

On 31/05/2017 18:28, Damjan Marion (damarion) wrote:


Something like:

https://gerrit.fd.io/r/#/c/6962/



On 31 May 2017, at 18:50, Damjan Marion (damarion)  wrote:

What about:

export VPP_WITH_SAMPLE_PLUGIN=yes
make build
make run

Does this work for you?


On 31 May 2017, at 18:45, Kinsella, Ray  wrote:

I think the idea that a user needs to go all the effort to install VPP and then 
Sample plugin, in order to run trivial sample code as way too much work.

This is something we expect, they are going to copy and base a new plugin off. 
So why we would bother making them go to all the effort to install it on their 
system just to play with it.

Either way, I don't see this being reconciled, so I consigned the patch to 
/dev/null.

Ray K

On 31/05/2017 17:31, Damjan Marion (damarion) wrote:


On 31 May 2017, at 18:18, Kinsella, Ray  wrote:


Ok - but that doesn't get us any closer to helping newbies use the sample 
plugin with 'make build' and 'make run', right? They still need to install vpp, 
then the sample-plugin - lots of hoops.


make run is not built for running out-of-tree plugins but this should work:

make bootstrap
make pkg-deb
dpkg -i build-root/*.deb
cd src/examples/sample-plugin
autoreconf -fis
./configure
make
sudo make install



I disagree with a documentation heavy approach in principle, the wiki suggests, 
that it similarly goes 'out of sync' quiet quickly.

BTW - I wasn't advocating PLUGIN_DISABLED, I provided build-data configs in the 
same way we do enabling/disabling dpdk features.


ok



The updated patch provides the separation between example/sample plugins and 
plugins that was asked for. It re-uses all the same autotools configs as 
src/plugins, so shouldn't go out of sync.


i still disagree, sample-plugin should be stand-alone autotools project, you 
are removing configure.ac so for me it is no-go.



Ray K


On 31/05/2017 17:05, Damjan Marion (damarion) wrote:


I do not agree with that proposal, I think we need to have one sample of 
out-of-tree plugin as it is today.

Still, I agree that we need to help newbies and my proposal is that we just 
document build process for out-of-tree plugins with simple README.md inside 
src/examples/sample-plugin.

btw I consider use of PLUGIN_DISABLED (as default choice) as evil, as it mens 
that plugin will go out of sync sooner or later.



On 31 May 2017, at 17:37, Kinsella, Ray  wrote:


Ok, typically example/sample code is intended to be used by the newest of the 
new, newbies. So the sample plugin should work with 'make build' and 'make run' 
with the minimum of hoops to enable. Asking these users to install and 
configure VPP, then do the same for the sample plugin is too much. I think that 
this thread exists, is testament that the UX could be better - too many hoops.

So here I what I suggest to fix.

We create src/examples/plugins, put the sample plugin in here.

The examples plugins (src/examples/plugins) are in-tree plugins and build in 
exactly the same way as src/plugins from a build PoV (PLUGIN_ENABLED etc), with 
the exception that the examples plugins are disabled by default. They also live 
in the sample directory with no symlinks etc to src/plugin. We then provide a 
way to explicitly enable them with a build-data config.

I reworked the patch along these lines, does it make sense?

Ray K

On 31/05/2017 10:15, Damjan Marion (damarion) wrote:


The idea of sample plugin is to show people how to build out-of-tree plugin. As 
that plugin was broken several times due to changes we made I created special 
ebuild package which builds sample plugin as part of verify job to ensure that 
plugin will not be broken again due to changes in vpp.

Saying that, I strongly disagree that we move sample plugin into src/plugins, 
as that is place for in-tree plugins which actually do something useful.
If people want to create additional in-tree plugin, there is many samples 
already in src/plugins so I don't see an need for additional one.

So to continue discussion on this particular change, what do you think that it 
i

Re: [vpp-dev] VPP/How To Build The Sample Plugin

2017-05-31 Thread Kinsella, Ray

I think the idea that a user needs to go all the effort to install VPP 
and then Sample plugin, in order to run trivial sample code as way too 
much work.


This is something we expect, they are going to copy and base a new 
plugin off. So why we would bother making them go to all the effort to 
install it on their system just to play with it.


Either way, I don't see this being reconciled, so I consigned the patch 
to /dev/null.


Ray K

On 31/05/2017 17:31, Damjan Marion (damarion) wrote:


On 31 May 2017, at 18:18, Kinsella, Ray  wrote:


Ok - but that doesn't get us any closer to helping newbies use the sample 
plugin with 'make build' and 'make run', right? They still need to install vpp, 
then the sample-plugin - lots of hoops.


make run is not built for running out-of-tree plugins but this should work:

make bootstrap
make pkg-deb
dpkg -i build-root/*.deb
cd src/examples/sample-plugin
autoreconf -fis
./configure
make
sudo make install



I disagree with a documentation heavy approach in principle, the wiki suggests, 
that it similarly goes 'out of sync' quiet quickly.

BTW - I wasn't advocating PLUGIN_DISABLED, I provided build-data configs in the 
same way we do enabling/disabling dpdk features.


ok



The updated patch provides the separation between example/sample plugins and 
plugins that was asked for. It re-uses all the same autotools configs as 
src/plugins, so shouldn't go out of sync.


i still disagree, sample-plugin should be stand-alone autotools project, you 
are removing configure.ac so for me it is no-go.



Ray K


On 31/05/2017 17:05, Damjan Marion (damarion) wrote:


I do not agree with that proposal, I think we need to have one sample of 
out-of-tree plugin as it is today.

Still, I agree that we need to help newbies and my proposal is that we just 
document build process for out-of-tree plugins with simple README.md inside 
src/examples/sample-plugin.

btw I consider use of PLUGIN_DISABLED (as default choice) as evil, as it mens 
that plugin will go out of sync sooner or later.



On 31 May 2017, at 17:37, Kinsella, Ray  wrote:


Ok, typically example/sample code is intended to be used by the newest of the 
new, newbies. So the sample plugin should work with 'make build' and 'make run' 
with the minimum of hoops to enable. Asking these users to install and 
configure VPP, then do the same for the sample plugin is too much. I think that 
this thread exists, is testament that the UX could be better - too many hoops.

So here I what I suggest to fix.

We create src/examples/plugins, put the sample plugin in here.

The examples plugins (src/examples/plugins) are in-tree plugins and build in 
exactly the same way as src/plugins from a build PoV (PLUGIN_ENABLED etc), with 
the exception that the examples plugins are disabled by default. They also live 
in the sample directory with no symlinks etc to src/plugin. We then provide a 
way to explicitly enable them with a build-data config.

I reworked the patch along these lines, does it make sense?

Ray K

On 31/05/2017 10:15, Damjan Marion (damarion) wrote:


The idea of sample plugin is to show people how to build out-of-tree plugin. As 
that plugin was broken several times due to changes we made I created special 
ebuild package which builds sample plugin as part of verify job to ensure that 
plugin will not be broken again due to changes in vpp.

Saying that, I strongly disagree that we move sample plugin into src/plugins, 
as that is place for in-tree plugins which actually do something useful.
If people want to create additional in-tree plugin, there is many samples 
already in src/plugins so I don't see an need for additional one.

So to continue discussion on this particular change, what do you think that it 
is broken?

For me sequence:

autoreconf -fis
./configure
make
make install

Works perfectly fine. Off-course you need to have install vpp-dev package on 
your system...



On 30 May 2017, at 13:30, Kinsella, Ray  wrote:

The UX for the sample plugin is broken. Especially when you consider that the 
people most likely to try it and use it, are those least familiar with VPP.

I tried the use it a few months ago in training and found the UX similar then. 
So I put together a number of changes to integrate the plugin into the VPP 
build system, provide a build var to enable and disable it, and then added 
documentation for anyone new to VPP.

https://gerrit.fd.io/r/#/c/6920/

Thanks,

Ray K

On 27/05/2017 18:27, Luke, Chris wrote:

Wishes often come true when you turn them into patches. :)



Chris.



*From:* vpp-dev-boun...@lists.fd.io [mailto:vpp-dev-boun...@lists.fd.io]
*On Behalf Of *Burt Silverman
*Sent:* Saturday, May 27, 2017 10:33
*To:* Kinsella, Ray 
*Cc:* Alessio Silvestro ; vpp-dev

*Subject:* Re: [vpp-dev] VPP/How To Build The Sample Plugin



Thanks, Ray, this is exactly what I needed, by coincidence. I wish your
item 2. was placed co

Re: [vpp-dev] VPP/How To Build The Sample Plugin

2017-05-31 Thread Kinsella, Ray

Ok - but that doesn't get us any closer to helping newbies use the
sample plugin with 'make build' and 'make run', right? They still need
to install vpp, then the sample-plugin - lots of hoops.

I disagree with a documentation heavy approach in principle, the wiki
suggests, that it similarly goes 'out of sync' quiet quickly.

BTW - I wasn't advocating PLUGIN_DISABLED, I provided build-data configs
in the same way we do enabling/disabling dpdk features.

The updated patch provides the separation between example/sample plugins
and plugins that was asked for. It re-uses all the same autotools
configs as src/plugins, so shouldn't go out of sync.

Ray K

On 31/05/2017 17:05, Damjan Marion (damarion) wrote:

I do not agree with that proposal, I think we need to have one sample of
out-of-tree plugin as it is today.

Still, I agree that we need to help newbies and my proposal is that we just
document build process for out-of-tree plugins with simple README.md inside
src/examples/sample-plugin.

btw I consider use of PLUGIN_DISABLED (as default choice) as evil, as it mens
that plugin will go out of sync sooner or later.

On 31 May 2017, at 17:37, Kinsella, Ray wrote:

Ok, typically example/sample code is intended to be used by the newest of the
new, newbies. So the sample plugin should work with 'make build' and 'make run'
with the minimum of hoops to enable. Asking these users to install and
configure VPP, then do the same for the sample plugin is too much. I think that
this thread exists, is testament that the UX could be better - too many hoops.

So here I what I suggest to fix.

We create src/examples/plugins, put the sample plugin in here.

The examples plugins (src/examples/plugins) are in-tree plugins and build in
exactly the same way as src/plugins from a build PoV (PLUGIN_ENABLED etc), with
the exception that the examples plugins are disabled by default. They also live
in the sample directory with no symlinks etc to src/plugin. We then provide a
way to explicitly enable them with a build-data config.

I reworked the patch along these lines, does it make sense?

Ray K

On 31/05/2017 10:15, Damjan Marion (damarion) wrote:

The idea of sample plugin is to show people how to build out-of-tree plugin. As
that plugin was broken several times due to changes we made I created special
ebuild package which builds sample plugin as part of verify job to ensure that
plugin will not be broken again due to changes in vpp.

Saying that, I strongly disagree that we move sample plugin into src/plugins,
as that is place for in-tree plugins which actually do something useful.
If people want to create additional in-tree plugin, there is many samples
already in src/plugins so I don't see an need for additional one.

So to continue discussion on this particular change, what do you think that it
is broken?

For me sequence:

autoreconf -fis
./configure
make
make install

Works perfectly fine. Off-course you need to have install vpp-dev package on
your system...

On 30 May 2017, at 13:30, Kinsella, Ray wrote:

The UX for the sample plugin is broken. Especially when you consider that the
people most likely to try it and use it, are those least familiar with VPP.

I tried the use it a few months ago in training and found the UX similar then.
So I put together a number of changes to integrate the plugin into the VPP
build system, provide a build var to enable and disable it, and then added
documentation for anyone new to VPP.

https://gerrit.fd.io/r/#/c/6920/

Thanks,

Ray K

On 27/05/2017 18:27, Luke, Chris wrote:

Wishes often come true when you turn them into patches. :)

Chris.

*From:* vpp-dev-boun...@lists.fd.io [mailto:vpp-dev-boun...@lists.fd.io]
*On Behalf Of *Burt Silverman
*Sent:* Saturday, May 27, 2017 10:33
*To:* Kinsella, Ray
*Cc:* Alessio Silvestro ; vpp-dev

*Subject:* Re: [vpp-dev] VPP/How To Build The Sample Plugin

Thanks, Ray, this is exactly what I needed, by coincidence. I wish your
item 2. was placed commented out and almost word for word into the
standard $TOPDIR/src/vpp/conf/startup.conf -- that would make things
self documenting.

Burt

On Sat, May 27, 2017 at 8:30 AM, Kinsella, Ray mailto:ray.kinse...@intel.com>> wrote:

So there is an easier way

1. make -C build-root PLATFORM=vpp TAG=vpp sample-plugin-install

2. adjusting the plugin path depending on where the VPP src is, add
the following to your startup.conf

plugins
{
path
/root/src/vpp/build-root/install-vpp-native/sample-plugin/lib64/vpp_plugins/:/root/src/vpp/build-root/install-vpp_debug-native/vpp/lib64/vpp_plugins
}

Ray K

On 26/05/2017 12:53, Alessio Silvestro wrote:

Hi all,

I am trying to build the sample vpp-engine plug-in as explained here
(https://wiki.fd.io/view/VPP/How_To_Build_The_Sample_Plugin
<https://wiki.fd.noclick_io/view/VPP/How_To_Build_The_Sample_Plugin>).

I alr

Re: [vpp-dev] VPP/How To Build The Sample Plugin

2017-05-31 Thread Kinsella, Ray

Ok, typically example/sample code is intended to be used by the newest
of the new, newbies. So the sample plugin should work with 'make build'
and 'make run' with the minimum of hoops to enable. Asking these users
to install and configure VPP, then do the same for the sample plugin is
too much. I think that this thread exists, is testament that the UX
could be better - too many hoops.

So here I what I suggest to fix.

We create src/examples/plugins, put the sample plugin in here.

The examples plugins (src/examples/plugins) are in-tree plugins and
build in exactly the same way as src/plugins from a build PoV
(PLUGIN_ENABLED etc), with the exception that the examples plugins are
disabled by default. They also live in the sample directory with no
symlinks etc to src/plugin. We then provide a way to explicitly enable
them with a build-data config.

I reworked the patch along these lines, does it make sense?

Ray K

On 31/05/2017 10:15, Damjan Marion (damarion) wrote:

So to continue discussion on this particular change, what do you think that it
is broken?

For me sequence:

autoreconf -fis
./configure
make
make install

Works perfectly fine. Off-course you need to have install vpp-dev package on
your system...

On 30 May 2017, at 13:30, Kinsella, Ray wrote:

The UX for the sample plugin is broken. Especially when you consider that the
people most likely to try it and use it, are those least familiar with VPP.

https://gerrit.fd.io/r/#/c/6920/

Thanks,

Ray K

On 27/05/2017 18:27, Luke, Chris wrote:

Wishes often come true when you turn them into patches. :)

Chris.

*From:* vpp-dev-boun...@lists.fd.io [mailto:vpp-dev-boun...@lists.fd.io]
*On Behalf Of *Burt Silverman
*Sent:* Saturday, May 27, 2017 10:33
*To:* Kinsella, Ray
*Cc:* Alessio Silvestro ; vpp-dev

*Subject:* Re: [vpp-dev] VPP/How To Build The Sample Plugin

Burt

On Sat, May 27, 2017 at 8:30 AM, Kinsella, Ray mailto:ray.kinse...@intel.com>> wrote:

So there is an easier way

1. make -C build-root PLATFORM=vpp TAG=vpp sample-plugin-install

2. adjusting the plugin path depending on where the VPP src is, add
the following to your startup.conf

plugins
{
path

/root/src/vpp/build-root/install-vpp-native/sample-plugin/lib64/vpp_plugins/:/root/src/vpp/build-root/install-vpp_debug-native/vpp/lib64/vpp_plugins
}

Ray K

On 26/05/2017 12:53, Alessio Silvestro wrote:

Hi all,

I am trying to build the sample vpp-engine plug-in as explained here
(https://wiki.fd.io/view/VPP/How_To_Build_The_Sample_Plugin
<https://wiki.fd.noclick_io/view/VPP/How_To_Build_The_Sample_Plugin>).

I already tested my vpp installation, for instance it works when I
created a Source NAT.

I downloaded the most updated version of the sample-plugin and
run the
following commands:

sudo sh
cd /usr/share/doc/vpp/examples
cd /tmp/sample-plugin
libtoolize
aclocal
autoconf
autoheader

ERROR 1:autoheader: error: AC_CONFIG_HEADERS not found in
configure.ac <http://configure.noclick_ac>
<http://configure.ac <http://configure.noclick_ac>>

automake --add-missing
chmod +x configure
vpp_plugin_configure

ERROR 2:vpp_plugin_configure: command not found

So, first I have an error from the command autoheader.

Second, I do not have the command vpp_plugin_configure.

Any hints?

Best regards,
Alessio

___
vpp-dev mailing list
vpp-dev@lists.fd.io <mailto:vpp-dev@lists.fd.io>
https://lists.fd.io/mailman/listinfo/vpp-dev
<https://lists.fd.noclick_io/mailman/listinfo/vpp-dev>

___
vpp-dev mailing list
vpp-dev@lists.fd.io <mailto:vpp-

Re: [vpp-dev] VPP/How To Build The Sample Plugin

2017-05-30 Thread Kinsella, Ray

The UX for the sample plugin is broken. Especially when you consider
that the people most likely to try it and use it, are those least
familiar with VPP.

I tried the use it a few months ago in training and found the UX similar
then. So I put together a number of changes to integrate the plugin into
the VPP build system, provide a build var to enable and disable it, and
then added documentation for anyone new to VPP.

https://gerrit.fd.io/r/#/c/6920/

Thanks,

Ray K

On 27/05/2017 18:27, Luke, Chris wrote:

Wishes often come true when you turn them into patches. :)

Chris.

*From:* vpp-dev-boun...@lists.fd.io [mailto:vpp-dev-boun...@lists.fd.io]
*On Behalf Of *Burt Silverman
*Sent:* Saturday, May 27, 2017 10:33
*To:* Kinsella, Ray
*Cc:* Alessio Silvestro ; vpp-dev

*Subject:* Re: [vpp-dev] VPP/How To Build The Sample Plugin

Burt

On Sat, May 27, 2017 at 8:30 AM, Kinsella, Ray mailto:ray.kinse...@intel.com>> wrote:

So there is an easier way

1. make -C build-root PLATFORM=vpp TAG=vpp sample-plugin-install

2. adjusting the plugin path depending on where the VPP src is, add
the following to your startup.conf

plugins
{
path

/root/src/vpp/build-root/install-vpp-native/sample-plugin/lib64/vpp_plugins/:/root/src/vpp/build-root/install-vpp_debug-native/vpp/lib64/vpp_plugins
}

Ray K

On 26/05/2017 12:53, Alessio Silvestro wrote:

Hi all,

I am trying to build the sample vpp-engine plug-in as explained here
(https://wiki.fd.io/view/VPP/How_To_Build_The_Sample_Plugin
<https://wiki.fd.noclick_io/view/VPP/How_To_Build_The_Sample_Plugin>).

I already tested my vpp installation, for instance it works when I
created a Source NAT.

I downloaded the most updated version of the sample-plugin and
run the
following commands:

sudo sh
cd /usr/share/doc/vpp/examples
cd /tmp/sample-plugin
libtoolize
aclocal
autoconf
autoheader

ERROR 1:autoheader: error: AC_CONFIG_HEADERS not found in
configure.ac <http://configure.noclick_ac>
<http://configure.ac <http://configure.noclick_ac>>

automake --add-missing
chmod +x configure
vpp_plugin_configure

ERROR 2:vpp_plugin_configure: command not found

So, first I have an error from the command autoheader.

Second, I do not have the command vpp_plugin_configure.

Any hints?

Best regards,
Alessio

___
vpp-dev mailing list
vpp-dev@lists.fd.io <mailto:vpp-dev@lists.fd.io>
https://lists.fd.io/mailman/listinfo/vpp-dev
<https://lists.fd.noclick_io/mailman/listinfo/vpp-dev>

___
vpp-dev mailing list
vpp-dev@lists.fd.io
https://lists.fd.io/mailman/listinfo/vpp-dev

Re: [vpp-dev] VPP/How To Build The Sample Plugin

2017-05-27 Thread Kinsella, Ray


So there is an easier way

1. make -C build-root PLATFORM=vpp TAG=vpp sample-plugin-install

2. adjusting the plugin path depending on where the VPP src is, add the 
following to your startup.conf


plugins
{
path 
/root/src/vpp/build-root/install-vpp-native/sample-plugin/lib64/vpp_plugins/:/root/src/vpp/build-root/install-vpp_debug-native/vpp/lib64/vpp_plugins

}

Ray K


On 26/05/2017 12:53, Alessio Silvestro wrote:

Hi all,

I am trying to build the sample vpp-engine plug-in as explained here
(https://wiki.fd.io/view/VPP/How_To_Build_The_Sample_Plugin).

I already tested my vpp installation, for instance it works when I
created a Source NAT.

I downloaded the most updated version of the sample-plugin and run the
following commands:

   sudo sh
   cd /usr/share/doc/vpp/examples
   cd /tmp/sample-plugin
   libtoolize
   aclocal
   autoconf
   autoheader

ERROR 1:autoheader: error: AC_CONFIG_HEADERS not found in configure.ac


   automake --add-missing
   chmod +x configure
   vpp_plugin_configure

ERROR 2:vpp_plugin_configure: command not found


So, first I have an error from the command autoheader.

Second, I do not have the command vpp_plugin_configure.

Any hints?

Best regards,
Alessio


___
vpp-dev mailing list
vpp-dev@lists.fd.io
https://lists.fd.io/mailman/listinfo/vpp-dev


___
vpp-dev mailing list
vpp-dev@lists.fd.io
https://lists.fd.io/mailman/listinfo/vpp-dev

Re: [vpp-dev] VPP has no interfaces after update from 1704 to 1707 master

2017-05-23 Thread Kinsella, Ray


Is the dpdk plugin enabled?

Ray K

On 23/05/2017 13:01, Michal Cmarada -X (mcmarada - PANTHEON TECHNOLOGIES 
at Cisco) wrote:

Hi VPP devs,



I tried to update VPP to master version from stable 1704 on Centos. In
1704 everything was working. Now in 1707 master VPP does not pick the
physical interfaces. I tried to change the drivers and I whitelisted the
interfaces in startup.conf for vpp. All the interfaces are down so VPP
should pick them. But nothing works. I tried to update dpdk to the
version that was provided on vpp repo but it didn`t help. Can you help
me with this issue?



Here are some details that may help:

OS: centos-release-7-3.1611.el7.centos.x86_64



Installed Packages

Name: vpp

Arch: x86_64

Version : 17.07

Release : rc0~274_gbb292f4~b2376

Size: 1.7 M

Repo: installed

From repo   : fdio-master

Summary : Vector Packet Processing

License : MIT

Description : This package provides VPP executables: vpp, vpp_api_test,
vpp_json_test

: vpp - the vector packet engine

: vpp_api_test - vector packet engine API test tool

: vpp_json_test - vector packet engine JSON test tool



Installed Packages

Name: dpdk

Arch: x86_64

Version : 16.11

Release : 5.el7.centos

Size: 3.5 M

Repo: installed

From repo   : fdio-master

Summary : Set of libraries and drivers for fast packet processing

URL : http://dpdk.org

License : BSD and LGPLv2 and GPLv2

Description : The Data Plane Development Kit is a set of libraries and
drivers for

: fast packet processing in the user space.



vpp# sh pci

Address  Sock VID:PID Link Speed   Driver  Product
NameVital Product Data

:06:00.0   0  1137:0043   8.0 GT/s x16 enic

:07:00.0   0  1137:0043   8.0 GT/s x16 enic

:08:00.0   0  1137:0043   8.0 GT/s x16 enic

:09:00.0   0  1137:0043   8.0 GT/s x16 enic

:10:00.0   0  1137:0043   8.0 GT/s x16 enic

:11:00.0   0  1137:0043   8.0 GT/s x16 enic

:12:00.0   0  1137:0043   8.0 GT/s x16 enic

:0a:00.0   0  1137:0043   8.0 GT/s x16 enic

:0f:00.0   0  1137:0043   8.0 GT/s x16 enic



Thanks

Michal



___
vpp-dev mailing list
vpp-dev@lists.fd.io
https://lists.fd.io/mailman/listinfo/vpp-dev


___
vpp-dev mailing list
vpp-dev@lists.fd.io
https://lists.fd.io/mailman/listinfo/vpp-dev

[vpp-dev] CI Tests Failing

2017-05-22 Thread Kinsella, Ray


Hi folks,

Not sure if it's just me, but some CI tests have suddenly start failing 
for me. Is it just me or a wider problem?


Ray K


CENTOS
https://jenkins.fd.io/job/vpp-verify-master-centos7/5568/

19:52:10 IP Multicast Signabash: line 1: 21723 Segmentation fault 
(core dumped) python run_tests.py -d 
/w/workspace/vpp-verify-master-centos7/test

19:54:03 make[2]: *** [test] Error 139
19:54:03 make[2]: Leaving directory 
`/w/workspace/vpp-verify-master-centos7/test'

19:54:03 make[1]: *** [test] Error 2
19:54:03 make[1]: Leaving directory `/w/workspace/vpp-verify-master-centos7'
19:54:03 make: *** [verify] Error 2
19:54:04 Build step 'Execute shell' marked build as failure


UBUNTU
https://jenkins.fd.io/job/vpp-verify-master-ubuntu1604/5573/console

19:37:57 IP NULL route 
 OK
19:37:57 
FATAL: 
command execution failed

19:48:45 java.io.EOFException
19:48:45 	at 
java.io.ObjectInputStream$PeekInputStream.readFully(ObjectInputStream.java:2638)
19:48:45 	at 
java.io.ObjectInputStream$BlockDataInputStream.readShort(ObjectInputStream.java:3113)
19:48:45 	at 
java.io.ObjectInputStream.readStreamHeader(ObjectInputStream.java:853)

19:48:45at java.io.ObjectInputStream.(ObjectInputStream.java:349)
19:48:45 	at 
hudson.remoting.ObjectInputStreamEx.(ObjectInputStreamEx.java:48)
19:48:45 	at 
hudson.remoting.AbstractSynchronousByteArrayCommandTransport.read(AbstractSynchronousByteArrayCommandTransport.java:34)
19:48:45 	at 
hudson.remoting.SynchronousCommandTransport$ReaderThread.run(SynchronousCommandTransport.java:59)

19:48:45 Caused: java.io.IOException: Unexpected termination of the channel
19:48:45 	at 
hudson.remoting.SynchronousCommandTransport$ReaderThread.run(SynchronousCommandTransport.java:73)
19:48:45 Caused: java.io.IOException: Backing channel 
'ubuntu1604-basebuild-4c-4g-5113' is disconnected.
19:48:45 	at 
hudson.remoting.RemoteInvocationHandler.channelOrFail(RemoteInvocationHandler.java:192)
19:48:45 	at 
hudson.remoting.RemoteInvocationHandler.invoke(RemoteInvocationHandler.java:257)

19:48:45at com.sun.proxy.$Proxy87.isAlive(Unknown Source)
19:48:45 	at 
hudson.Launcher$RemoteLauncher$ProcImpl.isAlive(Launcher.java:1043)
19:48:45 	at 
hudson.Launcher$RemoteLauncher$ProcImpl.join(Launcher.java:1035)
19:48:45 	at 
hudson.tasks.CommandInterpreter.join(CommandInterpreter.java:155)
19:48:45 	at 
hudson.tasks.CommandInterpreter.perform(CommandInterpreter.java:109)
19:48:45 	at 
hudson.tasks.CommandInterpreter.perform(CommandInterpreter.java:66)
19:48:45 	at 
hudson.tasks.BuildStepMonitor$1.perform(BuildStepMonitor.java:20)
19:48:45 	at 
hudson.model.AbstractBuild$AbstractBuildExecution.perform(AbstractBuild.java:779)

19:48:45at hudson.model.Build$BuildExecution.build(Build.java:206)
19:48:45at hudson.model.Build$BuildExecution.doRun(Build.java:163)
19:48:45 	at 
hudson.model.AbstractBuild$AbstractBuildExecution.run(AbstractBuild.java:534)

19:48:45at hudson.model.Run.execute(Run.java:1728)
19:48:45at hudson.model.FreeStyleBuild.run(FreeStyleBuild.java:43)
19:48:45 	at 
hudson.model.ResourceController.execute(ResourceController.java:98)

19:48:45at hudson.model.Executor.run(Executor.java:405)
19:48:45 Build step 'Execute shell' marked build as failure
19:48:45 FATAL: channel is already closed
19:48:45 java.io.EOFException
19:48:45 	at 
java.io.ObjectInputStream$PeekInputStream.readFully(ObjectInputStream.java:2638)
19:48:45 	at 
java.io.ObjectInputStream$BlockDataInputStream.readShort(ObjectInputStream.java:3113)
19:48:45 	at 
java.io.ObjectInputStream.readStreamHeader(ObjectInputStream.java:853)

19:48:45at java.io.ObjectInputStream.(ObjectInputStream.java:349)
19:48:45 	at 
hudson.remoting.ObjectInputStreamEx.(ObjectInputStreamEx.java:48)
19:48:45 	at 
hudson.remoting.AbstractSynchronousByteArrayCommandTransport.read(AbstractSynchronousByteArrayCommandTransport.java:34)
19:48:45 	at 
hudson.remoting.SynchronousCommandTransport$ReaderThread.run(SynchronousCommandTransport.java:59)

19:48:45 Caused: java.io.IOException: Unexpected termination of the channel
19:48:45 	at 
hudson.remoting.SynchronousCommandTransport$ReaderThread.run(SynchronousCommandTransport.java:73)
19:48:45 Caused: hudson.remoting.ChannelClosedException: channel is 
already closed

19:48:45at hudson.remoting.Channel.send(Channel.java:605)
19:48:45at hudson.remoting.Request.call(Request.java:130)
19:48:45at hudson.remoting.Channel.call(Channel.java:829)
19:48:45at hudson.Launcher$RemoteLauncher.launch(Launcher.java:959)
19:48:45at hudson.Launcher$ProcStarter.start(Launcher.java:384)
19:48:45 	at 
com.cloudbees.jenkins.plugins.sshagent.exec.ExecRemoteAgent.stop(ExecRemoteAgent.java:126)
19:48:45 	at 
com.cloudbees.jenkins.plugins.sshagent.SSHAgentBuildWrapper$SSHAgentEnvironment.tearDown(SSHAgentBuildWrapper.java:417)

19:48:45at hudson.model.Build$BuildEx

Re: [vpp-dev] Failed to ping between two VPP

2017-05-18 Thread Kinsella, Ray


Hi Xyxue,

Please try this patch to see if it helps.

https://gerrit.fd.io/r/#/c/6758/1

Ray K

On 18/05/2017 09:28, 薛欣颖 wrote:

Hi neale,

'failed' means this  :
DBGvpp# ping 192.168.153.130

Statistics: 300 sent, 0 received, 100% packet loss

what I know is in'run_ping_ip46_address', L598 got the wrong event_type

I'm not running with multiple worker threads.

My two vpp is in two 14.04
ubuntu,The two host interfaces are in the two VPP. The two interfaces
connect directly.

Thanks,
xyxue


*From:* Neale Ranns (nranns) 
*Date:* 2017-05-17 16:27
*To:* 薛欣颖 ; Damjan Marion

*CC:* vpp-dev ; Dave Barach (dbarach)

*Subject:* Re: [vpp-dev] Failed to ping between two VPP



Hi Xyxue,



I don’t understand what has ‘failed’ here.

The trace shows an echo-reply packet was received, matched against
and outstanding request and punted.



If you are running with multiple worker threads, you’ll need to include;

https://gerrit.fd.io/r/#/c/6736/



the difference in the FIB output from 17.04 to 17.01 is not an
error, it’s just that the rewrite string on the adjacency is no
longer pretty printed/decoded.



Regards,

neale



*From: *薛欣颖
*Date: *Wednesday, 17 May 2017 at 02:07
*To: *Damjan Marion 
*Cc: *vpp-dev , "Dave Barach (dbarach)"
, "Neale Ranns (nranns)" 
*Subject: *Re: Re: [vpp-dev] Failed to ping between two VPP







I sorry to copy the eth2's packet,that is the eth1's packet in vpp1:

00:02:32:132870: af-packet-input
  af_packet: hw_if_index 1 next-index 4
tpacket2_hdr:
  status 0x2001 len 60 snaplen 60 mac 66 net 80
  sec 0x591b9b2d nsec 0x70a3271 vlan 0
00:02:32:134189: ethernet-input
  IP4: 00:0c:29:8f:ad:61 -> 00:0c:29:31:29:ae
00:02:32:134247: ip4-input
  ICMP: 192.168.153.131 -> 192.168.153.130
tos 0x00, ttl 64, length 25, checksum 0x5386
fragment id 0x7307
  ICMP echo_reply checksum 0x8aff
00:02:32:134290: ip4-lookup
  fib 0 dpo-idx 5 flow hash: 0x
  ICMP: 192.168.153.131 -> 192.168.153.130
tos 0x00, ttl 64, length 25, checksum 0x5386
fragment id 0x7307
  ICMP echo_reply checksum 0x8aff
00:02:32:134349: ip4-local
ICMP: 192.168.153.131 -> 192.168.153.130
  tos 0x00, ttl 64, length 25, checksum 0x5386
  fragment id 0x7307
ICMP echo_reply checksum 0x8aff
00:02:32:134413: ip4-icmp-input
  ICMP: 192.168.153.131 -> 192.168.153.130
tos 0x00, ttl 64, length 25, checksum 0x5386
fragment id 0x7307
  ICMP echo_reply checksum 0x8aff
00:02:32:134433: ip4-icmp-echo-reply
  ICMP echo id 29952 seq 0 (unknown)
00:02:32:135147: error-punt
  ip4-icmp-input: unknown type

I know the only way to check the mac on interface "show hardware",how could I 
check the " underlying device"?

Fib entry display 17.04 not exactly correct compared to fib entry 17.01 
:
17.04:
192.168.153.131/32
  unicast-ip4-chain
  [@0]: dpo-load-balance: [index:10 buckets:1 uRPF:9 to:[600:22840] 
via:[205:9430]]
[0] [@5]: ipv4 via 192.168.153.131 host-eth1: 
000c298fad61000c293129ae0800

17.01(not the same device):
192.168.190.138/32
  unicast-ip4-chain
  [@0]: dpo-load-balance: [index:16 buckets:1 uRPF:15 to:[9:756]]
[0] [@5]: ipv4 via 192.168.190.138 host-eth8: IP4: 00:0c:29:b7:19:20 
-> 00:0c:29:57:8f:fa

Will this change affect my test results?



Thanks,

xyxue



*From:* Damjan Marion 

*Date:* 2017-05-16 22:27

*To:* 薛欣颖 

*CC:* vpp-dev@lists.fd.io ; Dave
Barach 

*Subject:* Re: [vpp-dev] Failed to ping between two VPP



On 16 May 2017, at 13:32, Dave Barach (dbarach)
mailto:dbar...@cisco.com>> wrote:



Copying vpp-dev [and Neale Ranns AKA Dr. Fib...]



Thanks… Dave



*From:* 薛欣颖 [mailto:xy...@fiberhome.com]
*Sent:* Tuesday, May 16, 2017 2:16 AM
*To:* Dave Barach (dbarach) mailto:dbar...@cisco.com>>
*Subject:* Failed to ping between two VPP



Hi Dave,



When I use stable 17.04,there is something wrong when I
ping vpp2 from vpp1,that is my configuration:



vpp1:

creat host-interface name eth1
set interface state host-eth1 up

Re: [vpp-dev] Failed to ping between two VPP

2017-05-16 Thread Kinsella, Ray



Configuring the mac address on an af_packet device .
So I am guess that when 'set interface mac' is called on the af_packet 
device, it isn't configuring the underlying device. :-)


Ray K


On 16/05/2017 15:27, Damjan Marion wrote:

Are you sure that this trace belongs to your pings?
Both MAC and IP address is different than one you configured….

___
vpp-dev mailing list
vpp-dev@lists.fd.io
https://lists.fd.io/mailman/listinfo/vpp-dev

Re: [vpp-dev] VPP Nodes and Plugins

2017-05-16 Thread Kinsella, Ray


Hi Tomas,

> But how does af-packet-input decide which of these nodes to run? Is 
it simply in descending order so since sample is at the bottom, it will 
be run before ethernet-input etc?


If the device produce's ptype & ethtype information, dpdk-input will use 
this to identify the next node for any given packet. If not it will 
identify the packet's ptype in software.


As I remember

.runs_before = VNET_FEATURES ("ethernet-input"),

Causes the sample plugin to hook itself in front of the ethernet-input 
graph node. The tracing facility is revealing for this kind of 
understanding.


Ray K


On 16/05/2017 09:55, Tomas Brännström wrote:

Hi
I'm struggling a bit to understand how VPP processes the node graph. If
we take the sample plugin for example, it registers itself to the
"device-input" arc. Then it is run after "af-packet-input" among others.
But looking at the af-packet-input-node, it has the following "next"
nodes in the graph:

af-packet-inputip4-input-no-checksum [0]
 ip4-input [1]
 ip6-input [2]
mpls-input [3]
  ethernet-input [4]
error-drop [5]
  sample [6]

But how does af-packet-input decide which of these nodes to run? Is it
simply in descending order so since sample is at the bottom, it will be
run before ethernet-input etc?

Also, is there a document somewhere that explains exactly how a node is
input into the graph? I think I understand how it works on some level,
but I'm not 100% clear what for example VPP actually does with
".arc_name", ".runs_before" and so on.

/Tomas



___
vpp-dev mailing list
vpp-dev@lists.fd.io
https://lists.fd.io/mailman/listinfo/vpp-dev


___
vpp-dev mailing list
vpp-dev@lists.fd.io
https://lists.fd.io/mailman/listinfo/vpp-dev

Re: [vpp-dev] Failed to ping between two VPP

2017-05-16 Thread Kinsella, Ray


Hi xyxue,

Not sure that `set interface mac addr` will work with af_packet.
Can you provide a little bit more information ...

Are eth1 device's real NICs, or are they Tap devices connected to a 
bridge. In the case of a bridge, I presume its a Linux Kernel bridge?


Ray K


On 16/05/2017 09:16, 薛欣颖 wrote:


Hi guys,

When I use stable 17.04,there is something wrong when I ping vpp2 from vpp1,
I
use" dpkg --purge vpp vpp-dbg vpp-dev vpp-dpdk-dev vpp-dpdk-dkms vpp-lib 
vpp-plugins vpp-python-api
 "to compile vpp.
That is my configuration:

vpp1:
creat host-interface name eth1
set interface state host-eth1 up
set interface ip addr host-eth1 192.168.153.130/24
set interface mac addr host-eth1 00:0c:29:31:29:ae
vpp2:
creat host-interface name eth1
set interface state host-eth1 up
set interface ip addr host-eth1 192.168.153.131/24
set interface mac addr host-eth1 00:0c:29:8f:ad:61

then I " ping 192.168.153.131 "from vpp1,and "ping 192.168.153.130" from
VPP2

the packet trace:
01:20:19:315340: af-packet-input
  af_packet: hw_if_index 2 next-index 4
tpacket2_hdr:
  status 0x2001 len 60 snaplen 60 mac 66 net 80
  sec 0x59196685 nsec 0x18dd5a37 vlan 0
01:20:19:315391: ethernet-input
  IP4: 00:0c:29:31:29:b8 -> 00:0c:29:8f:ad:6b
01:20:19:315411: ip4-input
  ICMP: 192.168.190.134 -> 192.168.190.136
tos 0x00, ttl 64, length 25, checksum 0x5a2b
fragment id 0x2259
  ICMP echo_reply checksum 0x71ff
01:20:19:315430: ip4-lookup
  fib 0 dpo-idx 6 flow hash: 0x
  ICMP: 192.168.190.134 -> 192.168.190.136
tos 0x00, ttl 64, length 25, checksum 0x5a2b
fragment id 0x2259
  ICMP echo_reply checksum 0x71ff
01:20:19:315453: ip4-local
ICMP: 192.168.190.134 -> 192.168.190.136
  tos 0x00, ttl 64, length 25, checksum 0x5a2b
  fragment id 0x2259
ICMP echo_reply checksum 0x71ff
01:20:19:315476: ip4-icmp-input
  ICMP: 192.168.190.134 -> 192.168.190.136
tos 0x00, ttl 64, length 25, checksum 0x5a2b
fragment id 0x2259
  ICMP echo_reply checksum 0x71ff
01:20:19:315486: ip4-icmp-echo-reply
  ICMP echo id 36352 seq 0 (unknown)
01:20:19:315549: error-punt
  ip4-icmp-input: unknown type

the ip fib has something wrong:
192.168.153.130/32
  unicast-ip4-chain
  [@0]: dpo-load-balance: [index:10 buckets:1 uRPF:9 to:[900:34260] 
via:[188:8648]]

[0] [@5]: ipv4 via 192.168.153.130 host-eth1: 000c293129ae000c298fad610800

Is there something I forgot to config ?Did I conifg wrong command?



Thanks,
xyxue


___
vpp-dev mailing list
vpp-dev@lists.fd.io
https://lists.fd.io/mailman/listinfo/vpp-dev


___
vpp-dev mailing list
vpp-dev@lists.fd.io
https://lists.fd.io/mailman/listinfo/vpp-dev

Re: [vpp-dev] CSIT borked on master

2017-05-15 Thread Kinsella, Ray



Does it explain why I am getting a build failure on CentOS ?

https://logs.fd.io/production/vex-yul-rot-jenkins-1/vpp-verify-master-centos7/5402/console.log.gz

+ /usr/lib/rpm/brp-python-bytecompile /usr/bin/python 1
+ /usr/lib/rpm/redhat/brp-python-hardlink
+ /usr/lib/rpm/redhat/brp-java-repack-jars
Processing files: vpp-dpdk-devel-17.05-vpp1.x86_64
Provides: vpp-dpdk-devel = 17.05-vpp1 vpp-dpdk-devel(x86-64) = 17.05-vpp1
Requires(rpmlib): rpmlib(CompressedFileNames) <= 3.0.4-1 
rpmlib(FileDigests) <= 4.6.0-1 rpmlib(PartialHardlinkSets) <= 4.0.4-1 
rpmlib(PayloadFilesHavePrefix) <= 4.0-1
Requires: /bin/bash /bin/sh /usr/bin/env ld-linux-x86-64.so.2()(64bit) 
ld-linux-x86-64.so.2(GLIBC_2.3)(64bit) libc.so.6()(64bit) 
libc.so.6(GLIBC_2.14)(64bit) libc.so.6(GLIBC_2.2.5)(64bit) 
libc.so.6(GLIBC_2.3)(64bit) libc.so.6(GLIBC_2.3.2)(64bit) 
libc.so.6(GLIBC_2.3.4)(64bit) libc.so.6(GLIBC_2.4)(64bit) 
libc.so.6(GLIBC_2.7)(64bit) libc.so.6(GLIBC_2.8)(64bit) 
libcrypto.so.10()(64bit) libcrypto.so.10(libcrypto.so.10)(64bit) 
libdl.so.2()(64bit) libdl.so.2(GLIBC_2.2.5)(64bit) libm.so.6()(64bit) 
libm.so.6(GLIBC_2.2.5)(64bit) libpthread.so.0()(64bit) 
libpthread.so.0(GLIBC_2.12)(64bit) libpthread.so.0(GLIBC_2.2.5)(64bit) 
libpthread.so.0(GLIBC_2.3.2)(64bit) libpthread.so.0(GLIBC_2.3.4)(64bit) 
librt.so.1()(64bit) librt.so.1(GLIBC_2.2.5)(64bit) rtld(GNU_HASH)
Checking for unpackaged file(s): /usr/lib/rpm/check-files 
/w/workspace/vpp-verify-master-centos7/dpdk/rpm/BUILDROOT/vpp-dpdk-17.05-vpp1.x86_64
Wrote: 
/w/workspace/vpp-verify-master-centos7/dpdk/rpm/RPMS/x86_64/vpp-dpdk-devel-17.05-vpp1.x86_64.rpm

Executing(%clean): /bin/sh -e /var/tmp/rpm-tmp.bH2a6n
+ umask 022
+ cd /w/workspace/vpp-verify-master-centos7/dpdk/rpm/BUILD
+ /usr/bin/rm -rf 
/w/workspace/vpp-verify-master-centos7/dpdk/rpm/BUILDROOT/vpp-dpdk-17.05-vpp1.x86_64

+ exit 0
mv rpm/RPMS/x86_64/*.rpm .
git clean -fdx rpm
Removing rpm/BUILD/
Removing rpm/BUILDROOT/
Removing rpm/RPMS/
Removing rpm/SOURCES/
Removing rpm/SPECS/
Removing rpm/SRPMS/
Removing rpm/tmp/
make[2]: Leaving directory `/w/workspace/vpp-verify-master-centos7/dpdk'
sudo rpm -Uih vpp-dpdk-devel-17.05-vpp1.x86_64.rpm

	package vpp-dpdk-devel-17.05-vpp2.x86_64 (which is newer than 
vpp-dpdk-devel-17.05-vpp1.x86_64) is already installed

make[1]: *** [install-rpm] Error 2
make[1]: Leaving directory `/w/workspace/vpp-verify-master-centos7/dpdk'
make: *** [dpdk-install-dev] Error 2
Build step 'Execute shell' marked build as failure


Ray K

On 15/05/2017 11:54, Damjan Marion (damarion) wrote:


This issue is caused by bug in DPDK 17.05 caused by following commit:

http://dpdk.org/browse/dpdk/commit/?id=ee1843b

It happens only with old QEMU emulation (I repro it with “pc-1.0”) which
VIRL uses.

Fix (revert) is in gerrit:

https://gerrit.fd.io/r/#/c/6690/

Regards,

Damjan



On 13 May 2017, at 20:34, Neale Ranns (nranns) mailto:nra...@cisco.com>> wrote:


Hi Chris,

Yes, every CSIT job on master is borked.
I think I’ve narrowed this down to all VAT sw_interface_dump returning
bogus/garbage MAC addresses. No Idea why, can’t repro yet. I’ve a
speculative DPDK 17.05 bump backout job in the queue, for purposes of
elimination.

Regards,
/neale




*From: *"Luke, Chris" mailto:chris_l...@comcast.com>>
*Date: *Saturday, 13 May 2017 at 19:04
*To: *"Neale Ranns (nranns)" mailto:nra...@cisco.com>>, "yug...@telincn.com
" mailto:yug...@telincn.com>>, vpp-dev mailto:vpp-dev@lists.fd.io>>
*Subject: *RE: [vpp-dev] Segmentation fault in recursivly lookuping
fib entry.

CSIT seems to be barfing on every job at the moment :(

*From:* vpp-dev-boun...@lists.fd.io
 [mailto:vpp-dev-boun...@lists.fd.io] *On
Behalf Of *Neale Ranns (nranns)
*Sent:* Saturday, May 13, 2017 11:20
*To:* yug...@telincn.com ; vpp-dev
mailto:vpp-dev@lists.fd.io>>
*Subject:* Re: [vpp-dev] Segmentation fault in recursivly lookuping
fib entry.


https://gerrit.fd.io/r/#/c/6674/


/neale


*From: *"yug...@telincn.com "
mailto:yug...@telincn.com>>
*Date: *Saturday, 13 May 2017 at 14:24
*To: *"Neale Ranns (nranns)" mailto:nra...@cisco.com>>, vpp-dev mailto:vpp-dev@lists.fd.io>>
*Subject: *Re: Re: [vpp-dev] Segmentation fault in recursivly
lookuping fib entry.

Hi neale,
Could you leave me a msg then?

Thanks,
Ewan


yug...@telincn.com 


*From:* Neale Ranns (nranns) 
*Date:* 2017-05-13 20:33
*To:* yug...@telincn.com ; vpp-dev

*Subject:* Re: [vpp-dev] Segmentation fault in recursivly lookuping
fib entry.
Hi Ewan,

That’s a bug. I’ll fix it ASAP.

Thanks,
neale


*From: *mailto:vpp-dev-boun...@lists.fd.io>> on behalf of
"yug...@telincn.com "
mailto:yug.

Re: [vpp-dev] any "cloud" facilities for testing VPPs and VNFs at high load ?

2017-05-15 Thread Kinsella, Ray



Would you consider developing and contributing vSwitch scaling test 
cases to CSIT as one method?


Ray K

On 14/05/2017 15:43, madhu sharma wrote:


I am looking for a way to create a test setup for my VPP-based VNF,
running at 10-100gbps throughput.

Any recommendations ? Short of putting together their own network lab
with in-house hardware, how do folks test their virtual switches, VPPs,
and VNFs ?

At a minimum, I want to run my VNF on say 2-8 cores, and generate
stateful traffic using Trex on a connected VM - via virtual or real NICs.

I came across one such facility. Advantech, also in collaboration with
TwinPrime offers it here:
http://www2.advantech.com/nc/newsletter/NCG/RES/environment.html
I also see references to an Intel/Cisco NFV Quick Start Lab.

Anyone tried AWS or Azure ?

thanks,
- Madhu



___
vpp-dev mailing list
vpp-dev@lists.fd.io
https://lists.fd.io/mailman/listinfo/vpp-dev


___
vpp-dev mailing list
vpp-dev@lists.fd.io
https://lists.fd.io/mailman/listinfo/vpp-dev

Re: [vpp-dev] Ip fragment reassembly

2017-05-15 Thread Kinsella, Ray


Hi Xyxue,

Please try

https://gerrit.fd.io/r/#/c/6695/

And see if it resolves your issue.

Ray K

On 10/05/2017 13:35, 薛欣颖 wrote:

Hi guys,

Whether vpp support ip fragment reassembly on host-interface?
Can I use only" set interface mtu 2000 host-eth2 "to set mtu.

Here's my test result:
vpp+dpdk:
mtu 1500  when the packet size 1500+  it  dropped;
mtu 2000  when the packet size 2000+  it  dropped;
mtu 300  when the packet size < 1500 it passed;
no fragmentation in my test;
vpp without dpdk:
Configuring MTU is not effective.

Thanks,
xyxue


___
vpp-dev mailing list
vpp-dev@lists.fd.io
https://lists.fd.io/mailman/listinfo/vpp-dev


___
vpp-dev mailing list
vpp-dev@lists.fd.io
https://lists.fd.io/mailman/listinfo/vpp-dev

Re: [vpp-dev] Ip fragment reassembly

2017-05-10 Thread Kinsella, Ray


Hi Xyxue,

Likelyhood is there the MTU is not being set on the underlying 
interface, the host interface does great rx/tx but not much else.


Can you check with IP tools (ip link)?

Ray K


On 10/05/2017 13:35, 薛欣颖 wrote:

Hi guys,

Whether vpp support ip fragment reassembly on host-interface?
Can I use only" set interface mtu 2000 host-eth2 "to set mtu.

Here's my test result:
vpp+dpdk:
mtu 1500  when the packet size 1500+  it  dropped;
mtu 2000  when the packet size 2000+  it  dropped;
mtu 300  when the packet size < 1500 it passed;
no fragmentation in my test;
vpp without dpdk:
Configuring MTU is not effective.

Thanks,
xyxue


___
vpp-dev mailing list
vpp-dev@lists.fd.io
https://lists.fd.io/mailman/listinfo/vpp-dev


___
vpp-dev mailing list
vpp-dev@lists.fd.io
https://lists.fd.io/mailman/listinfo/vpp-dev

Re: [vpp-dev] Questions regarding running VPP in KVM VM and the C API

2017-05-05 Thread Kinsella, Ray





1) I'm trying to get VPP up and running in a Ubuntu KVM VM in an
OpenStack environment. However I have trouble even starting it. I've
installed the packages from binary (latest master) but the service fails
to start.



ok ... after you installed the packages, you might need a reboot. VPP 
installs /etc/sysctl.d/80-vpp.conf to automagically reserve huge pages 
at startup. But this is only activated at boot.



*"Main heap allocation failure!"* is printed in

/var/log/upstart/vpp.log so I was thinking it might have something do to
with the HugePages settings, but lowering the amount of requested pages
in the config does not help. Also it works without issues in VirtualBox
which had even fewer pages.


From memory you have to make sure to put the heapsize parameter in the 
right place, I found appending to the end the safest.


echo -e "heapsize 64M" | sudo tee -a /etc/vpp/startup.conf



Is there something specific that must be done to get VPP running in a
KVM VM?



I use libvirt as a vagrant provider all the time, so no ... it should be 
fine.


Ray K
___
vpp-dev mailing list
vpp-dev@lists.fd.io
https://lists.fd.io/mailman/listinfo/vpp-dev

Re: [vpp-dev] SNAT 64bit performance issue

2017-05-03 Thread Kinsella, Ray


Does 'show run' offer any clues?

Ray K

On 03/05/2017 07:35, Juan Salmon wrote:

Thanks Matus for your reply.

But I tested SNAT in 32 bit mode with the same number of users and it
worked 1000 times faster.

Best Regards,
Juan Salmon.


___
vpp-dev mailing list
vpp-dev@lists.fd.io
https://lists.fd.io/mailman/listinfo/vpp-dev


___
vpp-dev mailing list
vpp-dev@lists.fd.io
https://lists.fd.io/mailman/listinfo/vpp-dev

Re: [vpp-dev] Running VPP on windows (with vagrant)

2017-04-07 Thread Kinsella, Ray


As do I, works fine.

Ray K

On 05/04/2017 16:59, Ed Warnicke wrote:

Mathieu,

I routinely build in AWS using  fairly stock Ubuntu/Centos images.

Ed

On Tue, Apr 4, 2017 at 7:38 AM, Mathieu Duperre (mduperre)
mailto:mdupe...@cisco.com>> wrote:

Hi, 

__ __

(my apology if that’s not the right forum, not sure where else to
ask). Was anyone able to build vpp either:

__ __

-in a windows machine running vagrant (ubuntu or centos)? I tried
both, could not build it properly (Vagrant image not installing few
libs, make not making, etc.)

-in AWS?  I was able to build, with a few errors here and there, but
i got a segmentation fault when trying to start vpp.

__ __

I just want to know if that was done before. thanks.

__ __

__ __


http://wwwin.cisco.com/c/dam/cec/organizations/gmcc/services-tools/signaturetool/images/logo/logo_Cisco_Blue.png

 



*Mathieu Duperre*

SOFTWARE SYSTEMS ENGINEER.SALES

mdupe...@cisco.com 

Tel: *+1 919 927 1296 *



*Cisco Systems, Inc.*




Canada
cisco.com 



__ __

http://www.cisco.com/assets/swa/img/thinkbeforeyouprint.gifThink
before you print.

This email may contain confidential and privileged material for the
sole use of the intended recipient. Any review, use, distribution or
disclosure by others is strictly prohibited. If you are not the
intended recipient (or authorized to receive for the recipient),
please contact the sender by reply email and delete all copies of
this message.

Please click here

for Company Registration Information.

__ __


___
vpp-dev mailing list
vpp-dev@lists.fd.io 
https://lists.fd.io/mailman/listinfo/vpp-dev





___
vpp-dev mailing list
vpp-dev@lists.fd.io
https://lists.fd.io/mailman/listinfo/vpp-dev


___
vpp-dev mailing list
vpp-dev@lists.fd.io
https://lists.fd.io/mailman/listinfo/vpp-dev

Re: [vpp-dev] Out of Tree Build Failure on vnet/rewrite.h

2017-03-31 Thread Kinsella, Ray

Would this TC cover the actual problem though, as I though in this case 
the header was never "installed".


Ray K

On 31/03/2017 14:46, Burt Silverman wrote:

Exported is the wrong word. I meant "installed."



___
vpp-dev mailing list
vpp-dev@lists.fd.io
https://lists.fd.io/mailman/listinfo/vpp-dev


___
vpp-dev mailing list
vpp-dev@lists.fd.io
https://lists.fd.io/mailman/listinfo/vpp-dev

Re: [vpp-dev] VPP Overview - design and Implementation

2017-03-31 Thread Kinsella, Ray


New version uploaded with updates to

* Segment Routing
* Hierarchical FIB

Ray K


On 29/03/2017 14:58, Kinsella, Ray wrote:

Hi folks,

As discussed on yesterday's call, I have been pulling together material
to support FD.io VPP conference presentations. This is designed to be
material people can pull from and customize based on their target
audience, presentation length and own areas of interest.

That said my feeling is that

 * Slides 4, 5, 6 - VPP intro
 * Slide 35 - VPP summary

belong is all presentations.

You can find the draft preso here,

https://wiki.fd.io/view/File:Fd.io_vpp_overview_29.03.17.pptx

All comments and feedback are welcome.

Ray K
___
vpp-dev mailing list
vpp-dev@lists.fd.io
https://lists.fd.io/mailman/listinfo/vpp-dev

___
vpp-dev mailing list
vpp-dev@lists.fd.io
https://lists.fd.io/mailman/listinfo/vpp-dev

[vpp-dev] VPP Overview - design and Implementation

2017-03-29 Thread Kinsella, Ray


Hi folks,

As discussed on yesterday's call, I have been pulling together material 
to support FD.io VPP conference presentations. This is designed to be 
material people can pull from and customize based on their target 
audience, presentation length and own areas of interest.


That said my feeling is that

 * Slides 4, 5, 6 - VPP intro
 * Slide 35 - VPP summary

belong is all presentations.

You can find the draft preso here,

https://wiki.fd.io/view/File:Fd.io_vpp_overview_29.03.17.pptx

All comments and feedback are welcome.

Ray K
___
vpp-dev mailing list
vpp-dev@lists.fd.io
https://lists.fd.io/mailman/listinfo/vpp-dev

Re: [vpp-dev] Tap IF Names

2017-03-28 Thread Kinsella, Ray


+1 to Jon's comments.


On 24/03/2017 14:07, Pierre Pfister (ppfister) wrote:

Hello Jon,

No strong opinion on my side, but I'd just like to notice that there might be 
cases where multiple interfaces, in linux, have the same name, if they are in 
different network namespaces.
VPP could literally control the back-end of thousands of containers' 
interfaces, all called eth0.


Well your backend and your frontend device names are typically named 
different. You are correct in that frontend device in the container is 
always eth0, the backend device for each container's eth0 is uniquely 
named in the default network namespace.


Ray K
___
vpp-dev mailing list
vpp-dev@lists.fd.io
https://lists.fd.io/mailman/listinfo/vpp-dev

Re: [vpp-dev] vhost user and upstream dpdk

2017-03-21 Thread Kinsella, Ray


+1 to Thomas's comments, completely agree.


Ray K

On 20/03/2017 13:20, Thomas F Herbert wrote:

Damjan,

Thanks for your response. My answer is inline below.

--TFH


On 03/16/2017 02:25 PM, Damjan Marion wrote:



On 16 Mar 2017, at 14:24, Thomas F Herbert
<therb...@redhat.com> wrote:


Do we still have a plan convergence toward upstream DPDK for vhost-user?

This has come up in a discussion in this bugz,
https://bugzilla.redhat.com/show_bug.cgi?id=1422534

and this patch to vhost_user https://gerrit.fd.io/r/#/c/5700/

--TFH



No, we don't have such plans, i don't see a reason why we would invest
significant time in adopting 3rd party library if we have own working
code.

what is your particular concern?

DPDK has had a vhost user implementation for quite some time and it is
used by other projects built on DPDK. It is generally condidered better
to not duplicate similar functionality found in upstream projects. Also,
it is far easier for downstream organizations supporting both DPDK/OVS
and VPP to have a single code base for an identical functaionality.





--
*Thomas F Herbert*
SDN Group
Office of Technology
*Red Hat*


___
vpp-dev mailing list
vpp-dev@lists.fd.io
https://lists.fd.io/mailman/listinfo/vpp-dev


___
vpp-dev mailing list
vpp-dev@lists.fd.io
https://lists.fd.io/mailman/listinfo/vpp-dev

Re: [vpp-dev] HP performance issues

2017-03-06 Thread Kinsella, Ray


Hi Sean,

Since I guess all this is running a VM (right?), it would be good to 
understand what KVM stat dump looks like. For the HP+Dell compared to 
UCS/SuperMicro.


sudo ./perf stat -e 'kvm:*' -a sleep 1h

Are we using nested virtualization?

Ray K

On 03/03/2017 15:14, Sean Chandler (sechandl) wrote:

Hi Folks,

For several months we have been fighting performance issues on both HP and Dell 
systems.  The scenario is l2bd and one vhost so a traffic generator and a 
single DUT running vpp.

Various CPU jitter mitigation methods have been tried and integrated into CSIT 
but the performance for one core on thread one queue is half of what automated 
testing has shown on UCS or super micro.

Who would be the best person or persons to help investigate the situation?  
What info can I provide to help debug.  I've been working with the CSIT folks 
but we have run dry of ideas.

Thanks in advance!
-s
___
vpp-dev mailing list
vpp-dev@lists.fd.io
https://lists.fd.io/mailman/listinfo/vpp-dev


___
vpp-dev mailing list
vpp-dev@lists.fd.io
https://lists.fd.io/mailman/listinfo/vpp-dev

Re: [vpp-dev] vpp w/ dpdk-as-plugin

2017-03-02 Thread Kinsella, Ray


Good job - is it being benchmarked in CSIT yet.

RAy K

On 02/03/2017 00:07, Luke, Chris wrote:

Ooh, nice





*From:* vpp-dev-boun...@lists.fd.io [mailto:vpp-dev-boun...@lists.fd.io]
*On Behalf Of *Dave Barach (dbarach)
*Sent:* Wednesday, March 1, 2017 18:09
*To:* Damjan Marion (damarion) 
*Cc:* vpp-dev@lists.fd.io
*Subject:* [vpp-dev] vpp w/ dpdk-as-plugin



“If you’re reading this message, vpp w/ dpdk-as-a-plugin is now running
my home gateway.”



No issues noted... Nice job, Damjan!



dbarach@vppgate:~$ telnet 0 5002

Trying 0.0.0.0...

Connected to 0.

Escape character is '^]'.

_____   _  ___

 __/ __/ _ \  (_)__| | / / _ \/ _ \

_/ _// // / / / _ \   | |/ / ___/ ___/

/_/ /(_)_/\___/   |___/_/  /_/



vpp# sh plug

Plugin path is: /usr/lib/vpp_plugins



 Plugin   Version

  1. ioam_plugin.so   17.04-rc0~329-gc3a814b

  2. ila_plugin.so17.04-rc0~329-gc3a814b

  3. dpdk_plugin.so   17.04-rc0~329-gc3a814b

  4. acl_plugin.so17.04-rc0~329-gc3a814b

  5. flowperpkt_plugin.so 17.04-rc0~329-gc3a814b

  6. snat_plugin.so   17.04-rc0~329-gc3a814b

  7. lb_plugin.so 17.04-rc0~329-gc3a814b

  8. libsixrd_plugin.so   17.04-rc0~329-gc3a814b

vpp#



Thanks… Dave





___
vpp-dev mailing list
vpp-dev@lists.fd.io
https://lists.fd.io/mailman/listinfo/vpp-dev


___
vpp-dev mailing list
vpp-dev@lists.fd.io
https://lists.fd.io/mailman/listinfo/vpp-dev

Re: [vpp-dev] schedule vpp graph nodes on seperate cores

2017-02-24 Thread Kinsella, Ray


inline

On 24/02/2017 16:56, yusuf khan wrote:

Hi,

On my test setup, which is server machine with Intel(R) Xeon(R) CPU
E5-2680 v3 @ 2.50GHz and 16GB RAM. I am comparing dpdk aesni_mb
l2fwd-crypto decryption results with VPP dpdk-esp-decrypt node. I see 1
Mpps difference. l2wd-crypto can process 3 Mpps(size 168) but vpp dpdk
crypto can only process around 2 Mpps.


Right but there is no comparison in terms of L2 feature support, the 
DPDK code is sample code only, VPP is a full l2 stack.



There is 1 Mpps difference. May be due to the fact that dpdk-input and
crypto-input are process on the same core?


Well VPP does a double pass to route the encrypted packet, that the DPDK 
IPSEC sample code does not do, but again I would argue that is behaviour 
you want.




Is it possible to run dpdk-esp-decrypt node on another core just for
encryption decryption. This way we can avoid re-ordering of the packets
and at the same time use multicore for single ipsec flow.


Each core run's an entire copy of the VPP stack, you can't break a 
pipeline up over multiple core's easily AFAIK.


Ray K
___
vpp-dev mailing list
vpp-dev@lists.fd.io
https://lists.fd.io/mailman/listinfo/vpp-dev

Re: [vpp-dev] Vpp in a container

2017-02-13 Thread Kinsella, Ray


Hi Ragav,

So there a couple of things you can look at.

A bunch of work has been done on the orchestration side by Keith Burns. 
This involved wiring up the Kubernetes, Calico to VPP.  Keith can point 
you at the latest patches - but I believe the most of these are upstream.


You can see Keith's demo video of this here.
https://www.youtube.com/watch?v=Wzmr3_kQLIE

On the setup/installation side, I am happy to report, due to binary 
packaging VPP works no problem in a Container. I have used to myself on 
vpp_bootstrap, which is a container based quick-start development 
environment. If you take a look at the vagrant setup scripts here they 
will show you how to install etc.


You can find vpp-bootstrap here.
https://git.fd.io/vppsb/tree/vpp-bootstrap

Finally on the performance question, Container setup's tend to use 
veth's and memory mapped raw sockets to talk to the Container (af_packet 
+ PACKET_TX_RING). So VPP dataplane performance in this setup tends to 
be constrained by pushing and pulling packets to and from the host kernel.


Ray K

On 10/02/2017 01:40, Raghav Kaushik (rakaushi) wrote:

Hi Folks,



I’m trying to find some data about pros and cons of running VPP in a
container.



Has this been tried before? Are there some performance number
comparisons available ?



Any pointers will be much appreciated.



Thanks,

Raghav



___
vpp-dev mailing list
vpp-dev@lists.fd.io
https://lists.fd.io/mailman/listinfo/vpp-dev


___
vpp-dev mailing list
vpp-dev@lists.fd.io
https://lists.fd.io/mailman/listinfo/vpp-dev

Re: [vpp-dev] Error reading from file descriptor 9: Input/output error

2017-01-25 Thread Kinsella, Ray


Hi Damjan,

I don't get how the the following message is being blurbed to the Syslog

"PCI INTX mask not supported\n"

Is causing the VPP log to run out of space and resulting in ... ?

"file descriptor 9: Input/output error".

On the patch ...

http://dpdk.org/dev/patchwork/patch/11622/

Steve Hemminger's advice looks like the right path to resolve.

Ray K

On 25/01/2017 14:08, Damjan Marion wrote:


This is long standing issue in DPDK e1000 driver likely caused by bad
e1000 emulation.
Still, bad emulation is still not good excuse for lack of error message
throttling.

There is even patch submitted to fix this issue:

http://dpdk.org/dev/patchwork/patch/11622/

But it never went upstream…

As Dave suggested, switch to vmxnet3 will solve this issue…

Thanks,

Damjan


On 25 Jan 2017, at 14:54, Andrew Li (zhaoxili) mailto:zhaox...@cisco.com>> wrote:

Thank you, Dave, and Damjan, that’s clear enough for this
question…We’ll plan to switch to Vmxnet3 then.

- Andrew
*From: *"Dave Barach (dbarach)" mailto:dbar...@cisco.com>>
*Date: *Wednesday, 25 January 2017 at 9:37 PM
*To: *"Andrew Li (zhaoxili)" mailto:zhaox...@cisco.com>>, Damjan Marion mailto:dmarion.li...@gmail.com>>
*Cc: *"vpp-dev@lists.fd.io "
mailto:vpp-dev@lists.fd.io>>
*Subject: *RE: [vpp-dev] Error reading from file descriptor 9:
Input/output error

Vmxnet3 interfaces are strongly preferred: for performance, as well as
to make this specific issue disappear.

We have a patch floating around which turns off the disk-filling
syslog message, but if you simply switch to vmxnet3 interfaces the
pain will go away without rebuilding images, etc. etc.

HTH... D

*From:* vpp-dev-boun...@lists.fd.io
 [mailto:vpp-dev-boun...@lists.fd.io] *On
Behalf Of *Andrew Li (zhaoxili)
*Sent:* Wednesday, January 25, 2017 6:51 AM
*To:* Damjan Marion mailto:dmarion.li...@gmail.com>>
*Cc:* vpp-dev@lists.fd.io 
*Subject:* Re: [vpp-dev] Error reading from file descriptor 9:
Input/output error

Hi Damjan,

Yes…and will it hurt anything other than printing this error message?
If so we could move to VMXNET3.

Thanks,
Andrew
*From: *Damjan Marion mailto:dmarion.li...@gmail.com>>
*Date: *Wednesday, 25 January 2017 at 6:27 PM
*To: *"Andrew Li (zhaoxili)" mailto:zhaox...@cisco.com>>
*Cc: *"vpp-dev@lists.fd.io "
mailto:vpp-dev@lists.fd.io>>
*Subject: *Re: [vpp-dev] Error reading from file descriptor 9:
Input/output error



On 25 Jan 2017, at 05:43, Andrew Li (zhaoxili) mailto:zhaox...@cisco.com>> wrote:

Hi vpp-dev,

I’m encountering this strange issue: VPP keeps generating this
error message into /var/log/syslog:

Jan 24 23:28:19 localhost vpp[4749]: EAL: Error reading from file
descriptor 9: Input/output error
Jan 24 23:28:19 localhost vpp[4749]: /usr/bin/vpp[4749]: EAL:
Error reading from file descriptor 9: Input/output error

And my disk space is being eaten quickly…Haven’t encountered this
before.

VPP version(latest master):
stack@devstack-vpp2:~/src/vpp$ git log
commit f69ecfe09db52c672ccbe47e714bc9c9a70d5539
Author: Andrew Yourtchenko mailto:ayour...@gmail.com>>
Date:   Tue Jan 24 15:47:27 2017 +0100

Does anyone know what this error means? Or how to solve this?
Looks like a dpdk issue.


Are you using ESXi VM with e1000 interfaces?







___
vpp-dev mailing list
vpp-dev@lists.fd.io
https://lists.fd.io/mailman/listinfo/vpp-dev


___
vpp-dev mailing list
vpp-dev@lists.fd.io
https://lists.fd.io/mailman/listinfo/vpp-dev

Re: [vpp-dev] [csit-dev] VPP 17.01 Released

2017-01-23 Thread Kinsella, Ray


+1 Great Work!

Ray K

On 20/01/2017 23:50, Florin Coras wrote:

Awesome!

Thank you for the hard work Damjan and congratulations to the entire community!

Florin


On Jan 20, 2017, at 2:51 PM, Damjan Marion (damarion)  
wrote:


The VPP 17.01 release i up. Many thanks to all contributors and testers
and specially Ed for helping to roll this release over the finish line.

.rpm and .deb packages are uploaded to the nexus server.

New features in the VPP 17.01:

- Integrated November 2016 DPDK release
- Complete rework of Forwarding Information Base (FIB)
- Performance Improvements
 - Improvements in DPDK input and output nodes
 - Improvements in L2 path
 - Improvmeents in IPv4 lookup node
- Feature Arcs Improvements
 - Consolidation of the code
 - New feature arcs
   - device-input
   - interface-output
- DPDK Cryptodev Support
 - Software and Hardware Crypto Support
- DPDK HQoS support
- Simple Port Analyzer (SPAN)
- Bidirectional Forwarding Detection
 - Basic implementation
- IPFIX Improvements
- L2 GRE over IPSec tunnels
- Link Layer Discovery Protocol (LLDP)
- Vhost-user Improvements
 - Performance Improvements
 - Multi-queue
 - Reconnect
- LISP Enhancements
 - Source/Dest control plane support
 - L2 over LISP and GRE
 - Map-Register/Map-Notify/RLOC-probing support
 - L2 API improvements, overall code hardening
- Plugins:
 - New: ACL
 - New: Flow per Packet
 - Improved: SNAT
   - Multi-threading
   - Flow export
- Doxygen Enhancements
- Luajit API bindings
- API Refactoring
 - file split
 - message signatures
- Python and Scapy based unit testing infrastructure
 - Infrastructure
 - Various tests
- Packet Generator improvements
- TUN/TAP jumbo frames support
- Other various bug fixes and improvements

Thanks,

Damjan
___
csit-dev mailing list
csit-...@lists.fd.io
https://lists.fd.io/mailman/listinfo/csit-dev


___
csit-dev mailing list
csit-...@lists.fd.io
https://lists.fd.io/mailman/listinfo/csit-dev


___
vpp-dev mailing list
vpp-dev@lists.fd.io
https://lists.fd.io/mailman/listinfo/vpp-dev

Re: [vpp-dev] Avoid constant rebuilding of dpdk during the build process

2017-01-10 Thread Kinsella, Ray



Seems like a welcome, logical, incremental step towards using DPDK 
binary packaging.


Would it make sense to always first try to pull and install the binary 
package from the nexus server.


Only building locally when either the download fails or when forced
to do so.

Then for those people who are developing using a stable release,
they save the unnecessary time of building DPDK?

Ray K

On 10/01/2017 00:04, Damjan Marion (damarion) wrote:


Currently we are maintaining own set of DPDK patches and building
DPDK from source, this is something which will unlikely change short term.
Still it is quite time consuming and annoying that we are re-building dpdk 
over-and-over without real need to do so.

I just submitted RFC patch to gerrit[1] which changes this a bit and would
like to hear feedback from other people. Changes are:

- dpdk is not built anymore during the vpp build process, installed dpdk 
libs/headers are used instead

- I added packaging for dpdk header/libs which generates .deb files. It works 
in the following way:

$ make -C dpdk install-deb

Creates and installs (dpkg -i):

$ ls dpdk/*.deb
dpdk/vpp-dpdk-dbg_16.11-vpp1_amd64.deb
dpdk/vpp-dpdk-dev_16.11-vpp1_amd64.deb

- this installation process can be part of bootstrap, or people can fetch those 
debs from nexus server

- When we add new patch we simply bump PKG_PREFIX inside dpdk/Makefile to be 
“vpp2” instead of “vpp1” and so on.
  Immediately after that (during the next build) people should be warned that 
they have outdated dpdk package installed.
  Something like:

$ make -C dpdk check-deb
make: Entering directory '/home/damarion/src/vpp/dpdk'
==
 Outdated DPDK package detected:
  Installed: vpp-dpdk-dev 16.11-vpp0
  Current:   vpp-dpdk-dev 16.11-vpp1

 Please upgrade by invoking 'make -C dpdk install-deb'
==

- We should support option for people to use instead official DPDK packages, 
but they should be warned
  that they might not get optimal performance in all cases.

- All written above should also apply for RPMs

[1] https://gerrit.fd.io/r/#/c/4613

Thanks,

Damjan
___
vpp-dev mailing list
vpp-dev@lists.fd.io
https://lists.fd.io/mailman/listinfo/vpp-dev


___
vpp-dev mailing list
vpp-dev@lists.fd.io
https://lists.fd.io/mailman/listinfo/vpp-dev

Re: [vpp-dev] Committer / Maintainer model.

2016-12-21 Thread Kinsella, Ray




If somebody submits plugin change + 3 lines of new
code in critical ip4 path, I will greatly benefit from +1 received from
plugin maintainer and i will focus on this 3 lines.
If I don’t have +1 from plugin maintainer, i will just left whole diff
in the gerrit, simply because i don't have a clue about that plugin.


But really, the committer will have nothing to add about the plugin in 
this case.


A more efficient process would be to let an IPv4 committer review and 
commit the IPv4 change, with the Plugin committer separately doing 
likewise.


Many open source project require that patches be made separately for 
component.


Ray K
___
vpp-dev mailing list
vpp-dev@lists.fd.io
https://lists.fd.io/mailman/listinfo/vpp-dev

Re: [vpp-dev] Committer / Maintainer model.

2016-12-21 Thread Kinsella, Ray


Hi Joel,

Thanks for asking - I think I would be fine with that model. I think it 
would clarify who owns what and who is to go to for what.


A list of committers and a second list of maintainers (reviewers), feels 
like bureaucracy.


* I can see people on the maintainer list getting frustrated over time 
without an obvious path to become a committer.
* If a maintainer requires a committer's final  approval in any case, it 
may not solve the scaling problem we have at the moment.
* In practice, I am not sure what sure what a committers approval will 
add on top of the maintainers approval.


Thanks,

Ray K


On 21/12/2016 15:47, Joel Halpern wrote:

Ray, I think you are suggesting a more basic change.  That instead of having 
committers for a project, we have committers on a per-feature basis (where a 
single committer may be listed for several features in a project.)  Is that the 
model you are looking for?

Yours,
Joel

-Original Message-
From: vpp-dev-boun...@lists.fd.io [mailto:vpp-dev-boun...@lists.fd.io] On 
Behalf Of Kinsella, Ray
Sent: Wednesday, December 21, 2016 10:44 AM
To: Damjan Marion 
Cc: vpp-dev 
Subject: Re: [vpp-dev] Committer / Maintainer model.


If the bottleneck here is committers are not familiar enough with a feature to 
review or approve - and these are stacking up, should they really have have 
commit rights for that feature?

If we are asking maintainers to be domain experts for a feature, to own it, why 
have the extra bureaucracy of the additional committer approval.
Will 'committer' approval on top of 'maintainer' approval ultimately anything 
but overhead?

Ray K


On 21/12/2016 15:36, Damjan Marion wrote:



On 21 Dec 2016, at 15:58, Kinsella, Ray  wrote:

My 2c is that my experience of this model is that maintainers typically get 
frustrated and disillusioned over time and become inactive, as they feel they 
are minions, doing the tough and often invisble review work but with no real 
authority over a feature. You end up with maintainer churn etc.


My view here is completelly different, maintainer should have authority of his 
feature.
I.e. if somebody owns vagrant stuff, he should be asked to +1 on any 
non-trivial change.
If maintainer is busy with something else and not taking much care
about it anymore, then we either need to find another one or we should remove 
it completelly from the repo.

What we have today is broken, there is many unmaintained stuff in the
repo which went out of sync with time and if somebody submits patch to
the gerrit, I (and I guess other committers also) don't have a clue what to do 
with it, as i really never started vagrant in my life.

Same story with vppctl, rpm packaging, different plugins...





___
vpp-dev mailing list
vpp-dev@lists.fd.io
https://lists.fd.io/mailman/listinfo/vpp-dev


___
vpp-dev mailing list
vpp-dev@lists.fd.io
https://lists.fd.io/mailman/listinfo/vpp-dev

Re: [vpp-dev] Committer / Maintainer model.

2016-12-21 Thread Kinsella, Ray



If the bottleneck here is committers are not familiar enough with a 
feature to review or approve - and these are stacking up, should they 
really have have commit rights for that feature?


If we are asking maintainers to be domain experts for a feature, to own 
it, why have the extra bureaucracy of the additional committer approval. 
Will 'committer' approval on top of 'maintainer' approval ultimately 
anything but overhead?


Ray K


On 21/12/2016 15:36, Damjan Marion wrote:



On 21 Dec 2016, at 15:58, Kinsella, Ray  wrote:

My 2c is that my experience of this model is that maintainers typically get 
frustrated and disillusioned over time and become inactive, as they feel they 
are minions, doing the tough and often invisble review work but with no real 
authority over a feature. You end up with maintainer churn etc.


My view here is completelly different, maintainer should have authority of his 
feature.
I.e. if somebody owns vagrant stuff, he should be asked to +1 on any 
non-trivial change.
If maintainer is busy with something else and not taking much care about it 
anymore, then we
either need to find another one or we should remove it completelly from the 
repo.

What we have today is broken, there is many unmaintained stuff in the repo 
which went out of sync with time
and if somebody submits patch to the gerrit, I (and I guess other committers 
also) don't have a clue what to
do with it, as i really never started vagrant in my life.

Same story with vppctl, rpm packaging, different plugins...





___
vpp-dev mailing list
vpp-dev@lists.fd.io
https://lists.fd.io/mailman/listinfo/vpp-dev

Re: [vpp-dev] Committer / Maintainer model.

2016-12-21 Thread Kinsella, Ray




My 2c is that my experience of this model is that maintainers typically get 
frustrated and disillusioned over time and become inactive, as they feel they 
are minions, doing the tough and often invisble review work but with no real 
authority over a feature. You end up with maintainer churn etc.

That is why I push for a strong correlation between maintainers and committers. 
Committers should be a maintainer of something and the number of maintainers 
who are not committers should be small. A maintainer with no commit rights, 
should be a committer in waiting with a clear path to graduating to be a 
committer.


I see that. Interesting to see how it develops. Not suggesting we should cast 
this in stone. .


Understand why we want to make a change, want to ensure we move to a 
sustainable model.



I imagined the opposite. That a maintainer had super-powers for his feature, 
but perhaps wasn't so interested in the rest of the system. Then the role of 
the committer is more one of architectural oversight and someone who ensures 
that the right hoops were jumped.


In that case the committer, needs to review everything that a maintainer 
has already approved, to ensure it is 'architecturally agreeable' and 
'boxes have been ticked'. Kinda sounds like we are doubling the work?




I kind of think of a maintainer as a committer for a particular 
feature/area/plugin.


Well if they are a 'committer' then dispense with the extra bureaucracy 
and give them commit rights?




Try and then adjust as we go along?

Cheers,
Ole





Ray K


On 21/12/2016 12:18, Ole Troan wrote:
Guys,

The discussion we had on the Tuesday meeting about the maintainers file got me 
thinking.

I think there are two problems with the current model.
- It puts a high burden on the committers, since at least one among the 
committer set has to know every piece of the code.
 Aka. it doesn't scale very well.
- It doesn't allow the code authors aka maintainers of a given component any 
control of changes to 'their' component.

Proposal:
1) Each component has a set of owners (or named list).
2) When a change is submitted. A gerrit hook processes the list of changed 
files, and automatically adds the component owner lists or individuals to the 
reviews list.
   (e.g. Ole Troan (vnet/map), Neale Ranns (vnet/ip))
3) A committer will then see when maintainers have +1 for each affected 
component. And then either submit or chase component owners for review.

A maintainer can then raise a "discuss" by a -1.

If there is support for this we need to:
- create maintainer file (and keep this up to date)
- write gerrit hook

Btw, Damjan pointed me to the Xen project's definition of maintainers vs 
committers:

Maintainers

Maintainers own one or several components in the Xen tree. A maintainer reviews 
and approves changes that affect their components. It is a maintainer's prime 
responsibility to review, comment on, co-ordinate and accept patches from other 
community member's and to maintain the design cohesion of their components. 
Maintainers are listed in a MAINTAINERS file in the root of the source tree.

Committers

Committers are Maintainers that are allowed to commit changes into the source 
code repository. The committer acts on the wishes of the maintainers and 
applies changes that have been approved by the respective maintainer(s) to the 
source tree. Due to their status in the community, committers can also act as 
referees should disagreements amongst maintainers arise. Committers are listed 
on the sub-project's team portal (e.g. Hypervisor team portal).


Views?

Best regards,
Ole
___
vpp-dev mailing list
vpp-dev@lists.fd.io
https://lists.fd.io/mailman/listinfo/vpp-dev


___
vpp-dev mailing list
vpp-dev@lists.fd.io
https://lists.fd.io/mailman/listinfo/vpp-dev

Re: [vpp-dev] Committer / Maintainer model.

2016-12-21 Thread Kinsella, Ray


Thanks Ole,

I suspect the Maintainers versus Committers description was included for 
my benefit :-)


My 2c is that my experience of this model is that maintainers typically 
get frustrated and disillusioned over time and become inactive, as they 
feel they are minions, doing the tough and often invisble review work 
but with no real authority over a feature. You end up with maintainer 
churn etc.


That is why I push for a strong correlation between maintainers and 
committers. Committers should be a maintainer of something and the 
number of maintainers who are not committers should be small. A 
maintainer with no commit rights, should be a committer in waiting with 
a clear path to graduating to be a committer.


Ray K

On 21/12/2016 12:18, Ole Troan wrote:

Guys,

The discussion we had on the Tuesday meeting about the maintainers file got me 
thinking.

I think there are two problems with the current model.
- It puts a high burden on the committers, since at least one among the 
committer set has to know every piece of the code.
  Aka. it doesn't scale very well.
- It doesn't allow the code authors aka maintainers of a given component any 
control of changes to 'their' component.

Proposal:
1) Each component has a set of owners (or named list).
2) When a change is submitted. A gerrit hook processes the list of changed 
files, and automatically adds the component owner lists or individuals to the 
reviews list.
(e.g. Ole Troan (vnet/map), Neale Ranns (vnet/ip))
3) A committer will then see when maintainers have +1 for each affected 
component. And then either submit or chase component owners for review.

A maintainer can then raise a "discuss" by a -1.

If there is support for this we need to:
 - create maintainer file (and keep this up to date)
 - write gerrit hook

Btw, Damjan pointed me to the Xen project's definition of maintainers vs 
committers:

Maintainers

Maintainers own one or several components in the Xen tree. A maintainer reviews 
and approves changes that affect their components. It is a maintainer's prime 
responsibility to review, comment on, co-ordinate and accept patches from other 
community member's and to maintain the design cohesion of their components. 
Maintainers are listed in a MAINTAINERS file in the root of the source tree.

Committers

Committers are Maintainers that are allowed to commit changes into the source 
code repository. The committer acts on the wishes of the maintainers and 
applies changes that have been approved by the respective maintainer(s) to the 
source tree. Due to their status in the community, committers can also act as 
referees should disagreements amongst maintainers arise. Committers are listed 
on the sub-project's team portal (e.g. Hypervisor team portal).


Views?

Best regards,
Ole
___
vpp-dev mailing list
vpp-dev@lists.fd.io
https://lists.fd.io/mailman/listinfo/vpp-dev


___
vpp-dev mailing list
vpp-dev@lists.fd.io
https://lists.fd.io/mailman/listinfo/vpp-dev

Re: [vpp-dev] Committer / Maintainer model.

2016-12-21 Thread Kinsella, Ray

The the addition of a Maintainers file would be a positive step. Can I 
also recommend that we include and distribute get_maintainer.pl. A 
script which automates the process of identifying the correct maintainer 
for a patch.


See: http://lxr.free-electrons.com/source/scripts/get_maintainer.pl

Ray K

On 21/12/2016 14:31, Damjan Marion wrote:



On 21 Dec 2016, at 14:35, Ole Troan  wrote:


That makes sense to me. We have a few problems to address:

1. Patches not being reviewed in a timely fashion
2. Committers spending too much time reviewing patches
3. Committers reviewing patches outside their area of expertise
4. Inability to filter code review request emails from general chit-chat, 
contributing to (1).

How about this?

O Based on the set of files / directories involved in a patch, gerrit 
automatically adds reviewers
O Email patch review requests w/ subject lines of the form [CODE-REVIEW] or 
some such
O Send email "to" the area owner(s), "cc" committers [who should know when 
folks are unavailable, etc.]
O Area owners score patches.
  O If they're also committers, feel free to +2.
  O Otherwise, email committers when satisfied to ensure timely merges

We know that gerrit can automatically add reviewers (see also "fd.io JJB").

This scheme depends on identifying folks that we trust to enforce a certain level of 
"truth, justice, and the right way," but it should help a lot in terms of our 
current committer scaling problem.


Wouldn't this be a self-selecting bunch, assuming it is the people who authored 
the code for the area in the first place?

Thanks for fleshing this out!

I think the added emails would help a lot in identifying who's responsible for 
what action. Currently we're suffering a bit from the bystander effect.

I'd be happy to have a first go at a maintainers file. I suggest we inherit the 
linux kernel format?

https://www.kernel.org/doc/linux/MAINTAINERS


I really support idea of putting MAINTAINERS file and using this standard 
format,
unless somebody can point us to law or other legal act which denies use of word 
“maintainer” in the fd.io.




___
vpp-dev mailing list
vpp-dev@lists.fd.io
https://lists.fd.io/mailman/listinfo/vpp-dev


___
vpp-dev mailing list
vpp-dev@lists.fd.io
https://lists.fd.io/mailman/listinfo/vpp-dev

Re: [vpp-dev] QQs on VPP clone.sh

2016-12-21 Thread Kinsella, Ray


As promised, the rsync changes are here.

https://gerrit.fd.io/r/4449

I also came across another issues with Vagrant ubuntu setup, which I 
resolved here.


https://gerrit.fd.io/r/4450

Ray K


On 20/12/2016 22:34, Ed Warnicke wrote:

I think that has the potential to be a brilliant solution.  Could you
push a gerrit so its easy for folks to grab it to try and see how it goes?

Ed

On Tue, Dec 20, 2016 at 11:15 AM, Kinsella, Ray mailto:ray.kinse...@intel.com>> wrote:

Hi Ed,

For those of us that have to suffer a corporate firewall with a
socks proxy this is problematic. It stalls on the git clone as
vagrant-proxy doesn't take care of setting up the socks proxy.

And to be honest, I wouldn't feel comfortable for vagrant-proxy to
start copying my ssh_config and keys to the vagrant box. In addition
cloning vpp a second time, inside the vagrant box feels alot like
duplication of network bits.

Can I propose a simpler work-around, using rsync instead of nfs to
copy (instead of share) the vpp folder to the vagrant box.

  config.vm.synced_folder "../../", "/vpp", type: "rsync",
  rsync__auto: false,
  rsync__exclude: [
"build-root/build*/",
"build-root/install*/",
"build-root/images*/",
"build-root/*.deb",
"build-root/*.rpm",
"build-root/*.changes",
"build-root/python",
"build-root/deb/debian/*.dkms",
"build-root/deb/debian/*.install",
"build-root/deb/debian/changes",
"build-root/tools"]

Seems to work fine, and means we can remove the 'clone.sh' script.
What do you think?

Regards,

Ray K

On 13/12/2016 18:20, Ed Warnicke wrote:

Ray,

The reason is because vagrant produces a bunch of 'issues' when
trying
to build from a 'mounted' share.

When running vagrant, /vpp gets mounted with where ever the
source is
*outside* the vagrant VM

Once inside, if you try to build from /vpp, a variety of things
can and
    do break sadly :(

So we clone vpp again into the local-to-the-VM storage.

Does that answer your question?

Ed

On Tue, Dec 13, 2016 at 10:47 AM, Kinsella, Ray
mailto:ray.kinse...@intel.com>
<mailto:ray.kinse...@intel.com <mailto:ray.kinse...@intel.com>>>
wrote:

Folks,

Two QQ on clone.sh
(https://git.fd.io/vpp/tree/build-root/vagrant/clone.sh
<https://git.fd.io/vpp/tree/build-root/vagrant/clone.sh>
<https://git.fd.io/vpp/tree/build-root/vagrant/clone.sh
<https://git.fd.io/vpp/tree/build-root/vagrant/clone.sh>>)


1. Why are we re-cloning source already available in /vpp?
2. Why do we ssh to remote server on the third last line ?
(this presumably hangs?).

Ray K


# Setup the vpp code
cd ~vagrant/
sudo -u vagrant mkdir git
cd git/
echo "SSH_AUTH_SOCK  $SSH_AUTH_SOCK x"
chmod 777 $SSH_AUTH_SOCK

CLONE_URL=`cd /vpp;git remote -v | grep origin |grep fetch |awk
'{print $2}'`
echo "CLONE_URL $CLONE_URL"
echo $CLONE_URL | grep -q "^ssh:"
if [ $? == 0 ]; then
SSH_HOST=`echo $CLONE_URL| awk -F/ '{print $3}'`
SSH_PORT=`echo $SSH_HOST| awk -F: '{print $2}'`
if [ -n $SSH_PORT ]; then
SSH_PORT="-p $SSH_PORT"
fi
SSH_HOST=`echo $SSH_HOST| awk -F: '{print $1}'`
echo "SSH_HOST $SSH_HOST"
echo "SSH_PORT $SSH_PORT"
sudo -HE -u vagrant ssh -oStrictHostKeyChecking=no -v
$SSH_PORT
$SSH_HOST
fi
sudo -HE -u vagrant git clone $CLONE_URL
___
vpp-dev mailing list
vpp-dev@lists.fd.io <mailto:vpp-dev@lists.fd.io>
<mailto:vpp-dev@lists.fd.io <mailto:vpp-dev@lists.fd.io>>
https://lists.fd.io/mailman/listinfo/vpp-dev
<https://lists.fd.io/mailman/listinfo/vpp-dev>
<https://lists.fd.io/mailman/listinfo/vpp-dev
<https://lists.fd.io/mailman/listinfo/vpp-dev>>




___
vpp-dev mailing list
vpp-dev@lists.fd.io
https://lists.fd.io/mailman/listinfo/vpp-dev

Re: [vpp-dev] QQs on VPP clone.sh

2016-12-20 Thread Kinsella, Ray


Hi Ed,

For those of us that have to suffer a corporate firewall with a socks 
proxy this is problematic. It stalls on the git clone as vagrant-proxy 
doesn't take care of setting up the socks proxy.


And to be honest, I wouldn't feel comfortable for vagrant-proxy to start 
copying my ssh_config and keys to the vagrant box. In addition cloning 
vpp a second time, inside the vagrant box feels alot like duplication of 
network bits.


Can I propose a simpler work-around, using rsync instead of nfs to copy 
(instead of share) the vpp folder to the vagrant box.


  config.vm.synced_folder "../../", "/vpp", type: "rsync",
  rsync__auto: false,
  rsync__exclude: [
"build-root/build*/",
"build-root/install*/",
"build-root/images*/",
"build-root/*.deb",
"build-root/*.rpm",
"build-root/*.changes",
"build-root/python",
"build-root/deb/debian/*.dkms",
"build-root/deb/debian/*.install",
"build-root/deb/debian/changes",
"build-root/tools"]

Seems to work fine, and means we can remove the 'clone.sh' script.
What do you think?

Regards,

Ray K

On 13/12/2016 18:20, Ed Warnicke wrote:

Ray,

The reason is because vagrant produces a bunch of 'issues' when trying
to build from a 'mounted' share.

When running vagrant, /vpp gets mounted with where ever the source is
*outside* the vagrant VM

Once inside, if you try to build from /vpp, a variety of things can and
do break sadly :(

So we clone vpp again into the local-to-the-VM storage.

Does that answer your question?

Ed

On Tue, Dec 13, 2016 at 10:47 AM, Kinsella, Ray mailto:ray.kinse...@intel.com>> wrote:

Folks,

Two QQ on clone.sh
(https://git.fd.io/vpp/tree/build-root/vagrant/clone.sh
<https://git.fd.io/vpp/tree/build-root/vagrant/clone.sh>)

1. Why are we re-cloning source already available in /vpp?
2. Why do we ssh to remote server on the third last line ?
(this presumably hangs?).

Ray K


# Setup the vpp code
cd ~vagrant/
sudo -u vagrant mkdir git
cd git/
echo "SSH_AUTH_SOCK  $SSH_AUTH_SOCK x"
chmod 777 $SSH_AUTH_SOCK

CLONE_URL=`cd /vpp;git remote -v | grep origin |grep fetch |awk
'{print $2}'`
echo "CLONE_URL $CLONE_URL"
echo $CLONE_URL | grep -q "^ssh:"
if [ $? == 0 ]; then
SSH_HOST=`echo $CLONE_URL| awk -F/ '{print $3}'`
SSH_PORT=`echo $SSH_HOST| awk -F: '{print $2}'`
if [ -n $SSH_PORT ]; then
SSH_PORT="-p $SSH_PORT"
fi
SSH_HOST=`echo $SSH_HOST| awk -F: '{print $1}'`
echo "SSH_HOST $SSH_HOST"
echo "SSH_PORT $SSH_PORT"
sudo -HE -u vagrant ssh -oStrictHostKeyChecking=no -v $SSH_PORT
$SSH_HOST
fi
sudo -HE -u vagrant git clone $CLONE_URL
___
vpp-dev mailing list
vpp-dev@lists.fd.io <mailto:vpp-dev@lists.fd.io>
https://lists.fd.io/mailman/listinfo/vpp-dev
<https://lists.fd.io/mailman/listinfo/vpp-dev>



___
vpp-dev mailing list
vpp-dev@lists.fd.io
https://lists.fd.io/mailman/listinfo/vpp-dev

Re: [vpp-dev] VPP and DPDK

2016-12-14 Thread Kinsella, Ray



The 'fix' in the Niantic vPMD was rejected upstream in DPDK for solid 
reasons - it causes a performance degradation in the Niantic vPMD .


I completely agree that this 'fix', fixing Niantic PType support is 
beneficial for the VPP workload, there is no argument with this 
upstream. The point was that this would be at the expense of every other 
workload that was not using PType support with the Niantic vPMD.


So this is a place where the requirements of VPP and other DPDK 
consumers may diverge. I think its a good example of where packaging 
projects are required to impartially service the needs of all consumers. 
I think every patch to DPDK (or indeed any open source project) needs to 
be reviewed on this basis.


That all said, we are looking at fixing the Niantic vPMD, in the same 
way we fixed the Fortville vPMD so hopefully we can have the best of 
both worlds.


Ray K


On 08/12/2016 16:53, Ed Warnicke wrote:

Looping in the deb_dpdk and rpm_dpdk projects to see if we can work
together on getting vpp out of the business of building DPDK (and DPDK
packages) while still retaining the ability to quickly incorporate fixes
like the one the Niantic driver.

Ed

On Thu, Dec 8, 2016 at 5:34 AM, Damjan Marion (damarion)
mailto:damar...@cisco.com>> wrote:


I think it is right time to start discussing post 17.01 release
developments
and here I would like to propose some changes related to DPDK
integration.
I see a need for changes in 2 areas:

1) building DPDK as part of VPP build and patch management
2) moving DPDK related code (input node, tx, buffer, eal) to
separate shared library


## Building DPDK as part of VPP build and patch management

Unfortunately with DPDK 16.11 we are back in situation that we need
to patch DPDK. Niantic/ixgbe
vector PMD is not setting IP checksum offload flags so we need to
rollback this change.
Still, we don’t have any functional reasons not to compile/link VPP
with unpatched DPDK.

Going forward my proposal is:
1. removing dpdk build out of VPP git tree and depending on DPDK
rpm/deb packages instead
2. Providing our optimized versions of those packages (ie. which
include change to Niantic driver)
3. Users/developers can chose between using standard packages and
optimized ones


##  moving DPDK related code (input node, tx, buffer, eal) to
separate shared library

In 17.01 release I already did 1st step in that direction by
removing almost all code which
touches rte_mbuf and make handling of rte_mbuf duty of dpdk-input
and dpdk-tx nodes.

Reason for that is very simple, we want to have unique behavior
independently of what interface
types are used. So packets going in and out on vhost_user,
af_packet, netmap or
packet-generator interface should really don’t care about rte_mbuf.
If packet arrives on vhost_user interface and goes out on dpdk, then
dpdk-tx function will simply
fix rte_mbuf.

Next step which I would like to propose is moving DPDK related code
to separate shared library (e.g. libvpp-dpdk).
There are few reasons I have on my mind:
  * avoid building 2 different versions of VPP (vpp and vpp_lite),
quite a lot of people are using vpp_lite in cases when they develop
features and they simply don't have or don’t need access to physical
interfaces
  * “make test” infrastructure is using vpp_lite
  * supporting ongoing efforts to use VPP with different libraries
like ODP or Mellanox OFED.
  * avoid using weak references to rte_* functions which we have in
the code today

This change should be purely mechanical, it should not introduce any
limitations in how we use DPDK today.

Thanks,

Damjan

___
vpp-dev mailing list
vpp-dev@lists.fd.io 
https://lists.fd.io/mailman/listinfo/vpp-dev





___
vpp-dev mailing list
vpp-dev@lists.fd.io
https://lists.fd.io/mailman/listinfo/vpp-dev


___
vpp-dev mailing list
vpp-dev@lists.fd.io
https://lists.fd.io/mailman/listinfo/vpp-dev

[vpp-dev] QQs on VPP clone.sh

2016-12-13 Thread Kinsella, Ray


Folks,

Two QQ on clone.sh
(https://git.fd.io/vpp/tree/build-root/vagrant/clone.sh)

1. Why are we re-cloning source already available in /vpp?
2. Why do we ssh to remote server on the third last line ?
(this presumably hangs?).

Ray K


# Setup the vpp code
cd ~vagrant/
sudo -u vagrant mkdir git
cd git/
echo "SSH_AUTH_SOCK  $SSH_AUTH_SOCK x"
chmod 777 $SSH_AUTH_SOCK

CLONE_URL=`cd /vpp;git remote -v | grep origin |grep fetch |awk '{print 
$2}'`

echo "CLONE_URL $CLONE_URL"
echo $CLONE_URL | grep -q "^ssh:"
if [ $? == 0 ]; then
SSH_HOST=`echo $CLONE_URL| awk -F/ '{print $3}'`
SSH_PORT=`echo $SSH_HOST| awk -F: '{print $2}'`
if [ -n $SSH_PORT ]; then
SSH_PORT="-p $SSH_PORT"
fi
SSH_HOST=`echo $SSH_HOST| awk -F: '{print $1}'`
echo "SSH_HOST $SSH_HOST"
echo "SSH_PORT $SSH_PORT"
sudo -HE -u vagrant ssh -oStrictHostKeyChecking=no -v $SSH_PORT 
$SSH_HOST

fi
sudo -HE -u vagrant git clone $CLONE_URL
___
vpp-dev mailing list
vpp-dev@lists.fd.io
https://lists.fd.io/mailman/listinfo/vpp-dev

Re: [vpp-dev] vxlan setup guidance

2016-12-13 Thread Kinsella, Ray



I had manually set flags=0x8 to get scapy + vpp to work correctly also.
Might be something to fix in Scapy and upstream?

Ray K

On 09/12/2016 19:34, Matej Klotton -X (mklotton - PANTHEON TECHNOLOGIES 
at Cisco) wrote:

My understanding is, if is VXLAN then UDP destination port is 4789 and
must have set I bit to 1. If is frame encapsulated into VXLAN GPE then
UDP destination port is 4790 and must have set I and P bits (probably
some others too). The RFCs also defines that if is received wrong bit
set it should be ignored.


___
vpp-dev mailing list
vpp-dev@lists.fd.io
https://lists.fd.io/mailman/listinfo/vpp-dev

Re: [vpp-dev] vxlan setup guidance

2016-12-02 Thread Kinsella, Ray


Discovered the problem ...

For reason you have the explicitly set the valid vni flag in scapy.

x=Ether()/IP(dst="172.16.1.3")/UDP(sport=1337,dport=4789)/VXLAN(vni=42,flags=8)/Ether()/IP()/UDP()

Ray K

On 02/12/2016 14:43, Kinsella, Ray wrote:


VPP is complaining that a tunnel for vni 42 does not exist, when clearly it 
does.
I would have thought my simple set up below, would allow packets to find their 
way through the tunnel to 192.168.1.3.

I am definitely doing something wrong.

Ray K

Packet 2

00:00:19:676348: af-packet-input
  af_packet: hw_if_index 1 next-index 3
tpacket2_hdr:
  status 0x2001 len 84 snaplen 84 mac 66 net 80
  sec 0x58418737 nsec 0x37c99699 vlan 0
00:00:19:679815: ethernet-input
  IP4: 00:17:3e:37:19:89 -> 02:fe:16:35:df:49
00:00:19:679851: ip4-input
  UDP: 172.16.1.2 -> 172.16.1.3
tos 0x00, ttl 64, length 70, checksum 0x2081
fragment id 0x0001
  UDP: 1337 -> 4789
length 50, checksum 0xd6ca
00:00:19:679886: ip4-lookup
  fib 0 dpo-idx 5 flow hash: 0x
  UDP: 172.16.1.2 -> 172.16.1.3
tos 0x00, ttl 64, length 70, checksum 0x2081
fragment id 0x0001
  UDP: 1337 -> 4789
length 50, checksum 0xd6ca
00:00:19:679938: ip4-local
UDP: 172.16.1.2 -> 172.16.1.3
  tos 0x00, ttl 64, length 70, checksum 0x2081
  fragment id 0x0001
UDP: 1337 -> 4789
  length 50, checksum 0xd6ca
00:00:19:679975: ip4-udp-lookup
  UDP: src-port 1337 dst-port 4789
00:00:19:679996: vxlan4-input
  VXLAN decap error - tunnel for vni 42 does not exist
00:00:19:680008: error-drop
  vxlan4-input: packets with bad flags field in vxlan header

DBGvpp# show vxlan tunnel
[0] src 172.16.1.3 dst 172.16.1.2 vni 42 encap_fib_index 1 sw_if_index 3 
fib_entry_index 20

- setup -
create host-interface name veth_link1
set interface ip address host-veth_link1 172.16.1.3/24
set interface state host-veth_link1 up

create loopback interface
set interface ip table loop0 7
set interface ip address loop0 192.168.1.3/24

create vxlan tunnel src 172.16.1.3 dst 172.16.1.2 vni 42 encap-vrf-id 7 
decap-next l2

set interface state vxlan_tunnel0 up
set interface state loop0 up

trace add af-packet-input 50
---



-Original Message-
From: Neale Ranns (nranns) [mailto:nra...@cisco.com]
Sent: Thursday, December 1, 2016 7:11 PM
To: Kinsella, Ray ; John Lo (loj) ; 
Keith
Burns (krb) ; Dave Barach (dbarach) 
Cc: vpp-dev@lists.fd.io
Subject: Re: [vpp-dev] vxlan setup guidance


Hi Ray,

An observation inline...



On 01/12/2016, 18:23, "vpp-dev-boun...@lists.fd.io on behalf of Kinsella, Ray"
 wrote:


Guys,

I am struggling with VXLan setup with VPP.
On Linux it's no problem ...

Container 1
sudo ip addr add 172.16.1.2/24 dev veth_link1
sudo ip link set dev veth_link1 up
sudo ip link add vxlan0 type vxlan id 42 remote 172.16.1.3 dev 
veth_link1
sudo ip addr add 192.168.1.2 dev vxlan0
sudo ip link set dev vxlan0 up

Container 2
sudo ip addr add 172.16.1.3/24 dev veth_link1
sudo ip link set dev veth_link1 up
sudo ip link add vxlan0 type vxlan id 42 remote 172.16.1.2 dev 
veth_link1
sudo ip addr add 192.168.1.3 dev vxlan0
sudo ip link set dev vxlan0 up

Container 1 and 2 are connected on the host with a VPP l2 bridge.
All works fine I can ping from between 172.16.1.2 and 172.16.1.3, 192.168.1.2

and 192.168.1.3.


However when I reconfigure Container 2 to use VPP, as follows.
(delete vxlan0, ip assignments to Linux before I start).

create host-interface name veth_link1
set interface ip address host-veth_link1 172.16.1.3/24
set interface state host-veth_link1 up

At this point, I can ping between 172.16.1.2 and 172.16.1.3.
Still all works fine ... I then create the vxlan tunnel and then bridge

create vxlan tunnel src 172.16.1.3 dst 172.16.1.2 vni 42
set interface l2 bridge host-veth_link1 100


Here you added the underlay link into the VxLAN overlay bridge-domain, so the
underlay IP traffic is bridged. Remove this and I expect it will work.


/neale



set interface l2 bridge vxlan_tunnel0 100

However when after I create the vxlan tunnel and add it to the a bridge.
Pings in 172.16.1.0/24  stop and I get the following

vagrant@ctwo:~$ sudo tcpdump -i veth_link1 -n
tcpdump: WARNING: veth_link1: no IPv4 address assigned
tcpdump: verbose output suppressed, use -v or -vv for full protocol
decode listening on veth_link1, link-type EN10MB (Ethernet), capture
size 65535 bytes
18:14:31.942718 ARP, Request who-has 172.16.1.3 tell 172.16.1.2, length
28
18:14:31.943169 IP 172.16.1.3.28173 > 172.16.1.2.4789: VXLAN, flags [I]
(0x08), vni 42 ARP, Request who-has 172.16.1.3 tell 172.16.1.2, length
28
18:14:32.940710 ARP, Request who-has 172.16.1.3 tell 172.16.1.2, length
28
18:14:32.941011 IP 172.16.1.3.28173 > 172.16.1.2.4789:

Re: [vpp-dev] vxlan setup guidance

2016-12-02 Thread Kinsella, Ray


VPP is complaining that a tunnel for vni 42 does not exist, when clearly it 
does.
I would have thought my simple set up below, would allow packets to find their 
way through the tunnel to 192.168.1.3.

I am definitely doing something wrong.

Ray K

Packet 2

00:00:19:676348: af-packet-input
  af_packet: hw_if_index 1 next-index 3
tpacket2_hdr:
  status 0x2001 len 84 snaplen 84 mac 66 net 80
  sec 0x58418737 nsec 0x37c99699 vlan 0
00:00:19:679815: ethernet-input
  IP4: 00:17:3e:37:19:89 -> 02:fe:16:35:df:49
00:00:19:679851: ip4-input
  UDP: 172.16.1.2 -> 172.16.1.3
tos 0x00, ttl 64, length 70, checksum 0x2081
fragment id 0x0001
  UDP: 1337 -> 4789
length 50, checksum 0xd6ca
00:00:19:679886: ip4-lookup
  fib 0 dpo-idx 5 flow hash: 0x
  UDP: 172.16.1.2 -> 172.16.1.3
tos 0x00, ttl 64, length 70, checksum 0x2081
fragment id 0x0001
  UDP: 1337 -> 4789
length 50, checksum 0xd6ca
00:00:19:679938: ip4-local
UDP: 172.16.1.2 -> 172.16.1.3
  tos 0x00, ttl 64, length 70, checksum 0x2081
  fragment id 0x0001
UDP: 1337 -> 4789
  length 50, checksum 0xd6ca
00:00:19:679975: ip4-udp-lookup
  UDP: src-port 1337 dst-port 4789
00:00:19:679996: vxlan4-input
  VXLAN decap error - tunnel for vni 42 does not exist
00:00:19:680008: error-drop
  vxlan4-input: packets with bad flags field in vxlan header

DBGvpp# show vxlan tunnel
[0] src 172.16.1.3 dst 172.16.1.2 vni 42 encap_fib_index 1 sw_if_index 3 
fib_entry_index 20

- setup -
create host-interface name veth_link1
set interface ip address host-veth_link1 172.16.1.3/24
set interface state host-veth_link1 up

create loopback interface
set interface ip table loop0 7
set interface ip address loop0 192.168.1.3/24

create vxlan tunnel src 172.16.1.3 dst 172.16.1.2 vni 42 encap-vrf-id 7 
decap-next l2

set interface state vxlan_tunnel0 up
set interface state loop0 up

trace add af-packet-input 50
---


>-Original Message-
>From: Neale Ranns (nranns) [mailto:nra...@cisco.com]
>Sent: Thursday, December 1, 2016 7:11 PM
>To: Kinsella, Ray ; John Lo (loj) ; 
>Keith
>Burns (krb) ; Dave Barach (dbarach) 
>Cc: vpp-dev@lists.fd.io
>Subject: Re: [vpp-dev] vxlan setup guidance
>
>
>Hi Ray,
>
>An observation inline...
>
>
>
>On 01/12/2016, 18:23, "vpp-dev-boun...@lists.fd.io on behalf of Kinsella, Ray"
> wrote:
>
>>Guys,
>>
>>I am struggling with VXLan setup with VPP.
>>On Linux it's no problem ...
>>
>>Container 1
>>  sudo ip addr add 172.16.1.2/24 dev veth_link1
>>  sudo ip link set dev veth_link1 up
>>  sudo ip link add vxlan0 type vxlan id 42 remote 172.16.1.3 dev 
>> veth_link1
>>  sudo ip addr add 192.168.1.2 dev vxlan0
>>  sudo ip link set dev vxlan0 up
>>
>>Container 2
>>  sudo ip addr add 172.16.1.3/24 dev veth_link1
>>  sudo ip link set dev veth_link1 up
>>  sudo ip link add vxlan0 type vxlan id 42 remote 172.16.1.2 dev 
>> veth_link1
>>  sudo ip addr add 192.168.1.3 dev vxlan0
>>  sudo ip link set dev vxlan0 up
>>
>>Container 1 and 2 are connected on the host with a VPP l2 bridge.
>>All works fine I can ping from between 172.16.1.2 and 172.16.1.3, 192.168.1.2
>and 192.168.1.3.
>>
>>However when I reconfigure Container 2 to use VPP, as follows.
>>(delete vxlan0, ip assignments to Linux before I start).
>>
>>  create host-interface name veth_link1
>>  set interface ip address host-veth_link1 172.16.1.3/24
>>  set interface state host-veth_link1 up
>>
>>At this point, I can ping between 172.16.1.2 and 172.16.1.3.
>>Still all works fine ... I then create the vxlan tunnel and then bridge
>>
>>  create vxlan tunnel src 172.16.1.3 dst 172.16.1.2 vni 42
>>  set interface l2 bridge host-veth_link1 100
>
>Here you added the underlay link into the VxLAN overlay bridge-domain, so the
>underlay IP traffic is bridged. Remove this and I expect it will work.
>
>
>/neale
>
>
>>  set interface l2 bridge vxlan_tunnel0 100
>>
>>However when after I create the vxlan tunnel and add it to the a bridge.
>>Pings in 172.16.1.0/24  stop and I get the following
>>
>>vagrant@ctwo:~$ sudo tcpdump -i veth_link1 -n
>>tcpdump: WARNING: veth_link1: no IPv4 address assigned
>>tcpdump: verbose output suppressed, use -v or -vv for full protocol
>>decode listening on veth_link1, link-type EN10MB (Ethernet), capture
>>size 65535 bytes
>>18:14:31.942718 ARP, Request who-has 172.16.1.3 tell 172.16.1.2, length
>>28
>>18:14:31.943169 IP 172.16.1.3.28173 > 172.16.1.2.4789: VXLAN, flags [I]
>>

[vpp-dev] vxlan setup guidance

2016-12-01 Thread Kinsella, Ray

Guys,

I am struggling with VXLan setup with VPP.
On Linux it's no problem ... 

Container 1
sudo ip addr add 172.16.1.2/24 dev veth_link1
sudo ip link set dev veth_link1 up
sudo ip link add vxlan0 type vxlan id 42 remote 172.16.1.3 dev 
veth_link1
sudo ip addr add 192.168.1.2 dev vxlan0
sudo ip link set dev vxlan0 up

Container 2
sudo ip addr add 172.16.1.3/24 dev veth_link1
sudo ip link set dev veth_link1 up
sudo ip link add vxlan0 type vxlan id 42 remote 172.16.1.2 dev 
veth_link1
sudo ip addr add 192.168.1.3 dev vxlan0
sudo ip link set dev vxlan0 up

Container 1 and 2 are connected on the host with a VPP l2 bridge. 
All works fine I can ping from between 172.16.1.2 and 172.16.1.3, 192.168.1.2 
and 192.168.1.3.

However when I reconfigure Container 2 to use VPP, as follows.
(delete vxlan0, ip assignments to Linux before I start).

create host-interface name veth_link1
set interface ip address host-veth_link1 172.16.1.3/24
set interface state host-veth_link1 up

At this point, I can ping between 172.16.1.2 and 172.16.1.3.
Still all works fine ... I then create the vxlan tunnel and then bridge

create vxlan tunnel src 172.16.1.3 dst 172.16.1.2 vni 42
set interface l2 bridge host-veth_link1 100
set interface l2 bridge vxlan_tunnel0 100

However when after I create the vxlan tunnel and add it to the a bridge.
Pings in 172.16.1.0/24  stop and I get the following 

vagrant@ctwo:~$ sudo tcpdump -i veth_link1 -n
tcpdump: WARNING: veth_link1: no IPv4 address assigned
tcpdump: verbose output suppressed, use -v or -vv for full protocol decode
listening on veth_link1, link-type EN10MB (Ethernet), capture size 65535 bytes
18:14:31.942718 ARP, Request who-has 172.16.1.3 tell 172.16.1.2, length 28
18:14:31.943169 IP 172.16.1.3.28173 > 172.16.1.2.4789: VXLAN, flags [I] (0x08), 
vni 42
ARP, Request who-has 172.16.1.3 tell 172.16.1.2, length 28
18:14:32.940710 ARP, Request who-has 172.16.1.3 tell 172.16.1.2, length 28
18:14:32.941011 IP 172.16.1.3.28173 > 172.16.1.2.4789: VXLAN, flags [I] (0x08), 
vni 42
ARP, Request who-has 172.16.1.3 tell 172.16.1.2, length 28

Container 2's ICMP requests for who has 172.16.1.2, are getting encapsulated as 
VXLAN.
So Container 1 never responds as owning 172.16.1.2.

172.16.1.2 and 172.16.1.3  should be my VTEPs.
I should be able to ping back and forth between them without encapsulation.

Suspect I am making a simple mistake in my setup, can someone provide guidance. 

Ray K


___
vpp-dev mailing list
vpp-dev@lists.fd.io
https://lists.fd.io/mailman/listinfo/vpp-dev

[vpp-dev] Rebuilding with disabled plugin

2016-11-28 Thread Kinsella, Ray


Hi folks,

The only reliable way I found to rebuild after enabling a plugin was to 
do a full distclean and rebuild. As follows ...


cd vpp
(cd build-root; make distclean; rm .bootstrap.ok)
sed -i "s/PLUGIN_DISABLED[(]sample[)]/PLUGIN_ENABLED\(sample\)/" 
plugins/configure.ac

make PLATFORM=vpp_lite build
make PLATFORM=vpp_lite build-vpp-api
make PLATFORM=vpp_lite plugins

A full rebuild is clearly a PITA, in theory I should be able to do a

configure --enable-sample-plugin

in the plugins directory but I don't see an obvious way to do this.
Also if I do this by hacking  I end up with missing directories under 
build-root.


Was wondering if there was a more obvious approach I was missing?

Ray K
___
vpp-dev mailing list
vpp-dev@lists.fd.io
https://lists.fd.io/mailman/listinfo/vpp-dev

Re: [vpp-dev] [tldk-dev] TLDK meeting Calendar entry updated

2016-11-23 Thread Kinsella, Ray


Thanks Keith.

On 22/11/2016 16:40, Keith Burns wrote:

Please note I updated https://wiki.fd.io/view/Meetings to reflect that
TLDK meeting is:

NOW:
- 7am Friday Pacific

WAS:
- 8am Wed Pacfic



___
tldk-dev mailing list
tldk-...@lists.fd.io
https://lists.fd.io/mailman/listinfo/tldk-dev


___
vpp-dev mailing list
vpp-dev@lists.fd.io
https://lists.fd.io/mailman/listinfo/vpp-dev

Re: [vpp-dev] ioam-plugin build fails

2016-11-22 Thread Kinsella, Ray


Hi Mohsin,

That seems to work - does that mean that build-vpp-api should be a 
sub-target of plugins?


Ray K

On 21/11/2016 22:45, Mohsin Kazmi (sykazmi) wrote:

make build-vpp-api

___
vpp-dev mailing list
vpp-dev@lists.fd.io
https://lists.fd.io/mailman/listinfo/vpp-dev

[vpp-dev] ioam-plugin build fails

2016-11-21 Thread Kinsella, Ray


Hi folks,

Building of the ioam-plugin fails, complaining of missing symbols.

==> default: io/fd/vpp/jvpp/ioamtrace/JVppIoamtraceImpl.java:13: error: 
package io.fd.vpp.jvpp.callback does not exist

==> default: import io.fd.vpp.jvpp.callback.JVppCallback;
==> default:   ^
==> default: io/fd/vpp/jvpp/ioamtrace/JVppIoamtraceImpl.java:14: error: 
cannot find symbol

==> default: import io.fd.vpp.jvpp.VppConnection;


Full error log is on pastebin (http://pastebin.com/Fk6MhNCh)
This is with vpp_lite on VPP 16.09.

Any clues where I am going wrong?

Ray K
___
vpp-dev mailing list
vpp-dev@lists.fd.io
https://lists.fd.io/mailman/listinfo/vpp-dev

Re: [vpp-dev] https://gerrit.fd.io/r/#/c/2428/ isn't included in 16.09

2016-11-08 Thread Kinsella, Ray



Hi Yi,

Good you have a workaround but I agree with Keith.

We still need understand the scope of the issue, perhaps raise a JIRA 
bug. I can help out and raise the bug if required.


Does this mean that SFC w/VPP + ODL is completely broken - what is missing?

Ray K


On 07/11/2016 01:40, Yang, Yi Y wrote:

Hi, Keith



I have built a VPP version 16.09+this fix patch and upload it to my
personal repo, our sfc104 demo will use this to fix it. Once 17.01 is
ready for use, we will switch to 17.01, thank you so much for taking it
seriously.



*From:*Keith Burns [mailto:alaga...@gmail.com]
*Sent:* Saturday, November 5, 2016 1:15 AM
*To:* Yang, Yi Y ; Dave Barach (dbarach)

*Cc:* vpp-dev@lists.fd.io
*Subject:* RE: [vpp-dev] https://gerrit.fd.io/r/#/c/2428/ isn't included
in 16.09



Yi,



Putting the list back on.



I don’t see significant justification from you to do a respin.



i.e. “NSH doesn’t work” (sic ODL SFC) isn’t enough justification. Is ODL
SFC looking to do a respin?



How come this wasn’t caught sooner, as in, we all got feedback that ODL
SFC -> HC -> VPP + NSH_SFC was working for SFF functionality.



Can this wait for 17.01 ?



What is the scope of the issue? i.e. how many other projects are
impacted and can you explain the scope of the issue?



Is there a JIRA bug for instance?






From: Yang, Yi Y  
Reply: Yang, Yi Y  
Date: November 3, 2016 at 6:37:47 PM
To: Keith Burns  , Dave
Barach (dbarach)  
Subject:  RE: [vpp-dev] https://gerrit.fd.io/r/#/c/2428/ isn't included
in 16.09



Thanks, Keith, Vpp Weekly Meeting  time is too late for me, so I
can’t make it.



Dave, you’re vpp PTL, can you help build vpp 16.09.01 with this fix?



*From:*Keith Burns [mailto:alaga...@gmail.com]
*Sent:* Tuesday, November 1, 2016 10:11 PM
*To:* Dave Barach (dbarach) mailto:dbar...@cisco.com>>; Yang, Yi Y mailto:yi.y.y...@intel.com>>; vpp-dev@lists.fd.io

*Subject:* Re: [vpp-dev] https://gerrit.fd.io/r/#/c/2428/ isn't
included in 16.09



Yi,



Looks like this was submitted the day of RC0 but didn't get to
review until a bit after.



1. This is why I sent multiple emails along the lines of RC0 is coming

2. Post that, you needed a bug ID and a cherry-pick.



As for a respin this mechanically should be easier now we are on
Nexus Pro BUT I'm unsure of the implications to other projects who
may be using 16.09 artifacts and I believe we would want to respin
this as 16.09.01 or some such. I'll defer to others with more
experience



Do please come to vpp meeting

On Tue, Nov 1, 2016, 6:08 AM Dave Barach (dbarach)
mailto:dbar...@cisco.com>> wrote:

Please attend the vpp project meeting today at 11:00 AM to
discuss this request.



Thanks… Dave



*From:* vpp-dev-boun...@lists.fd.io

[mailto:vpp-dev-boun...@lists.fd.io
] *On Behalf Of *Yang, Yi Y
*Sent:* Tuesday, November 1, 2016 9:01 AM
*To:* vpp-dev@lists.fd.io 
*Subject:* [vpp-dev] https://gerrit.fd.io/r/#/c/2428/ isn't
included in 16.09



Hi, guys



Although vpp 16.09 is released on Sep 16, but
https://gerrit.fd.io/r/#/c/2428/ isn’t included in it, so ODL
sfc can’t work correctly with 16.09 debian repo, anybody can
help cherry-pick https://gerrit.fd.io/r/#/c/2428/ to 16.09
branch and rebuild 16.09 binary packages for Ubuntu 14.04?

___
vpp-dev mailing list
vpp-dev@lists.fd.io 
https://lists.fd.io/mailman/listinfo/vpp-dev



___
vpp-dev mailing list
vpp-dev@lists.fd.io
https://lists.fd.io/mailman/listinfo/vpp-dev


___
vpp-dev mailing list
vpp-dev@lists.fd.io
https://lists.fd.io/mailman/listinfo/vpp-dev

Re: [vpp-dev] jVpp as independent fd.io project

2016-10-26 Thread Kinsella, Ray


Hi Marek,

I am not sure that jVpp is useful outside of VPP,
So outside of helping with 1), I am not sure what the benefit would be.

I would have to imagine 1) is fixable in some other way.

Ray K

On 26/10/2016 08:06, Marek Gradzki -X (mgradzki - PANTHEON TECHNOLOGIES 
at Cisco) wrote:

Hi,



this is follow-up to the discussion started on the last VPP weekly meeting.



In my opinion, the biggest advantages are:



1) it would make VPP build faster and more stable

2) it would make possible to use build system that supports dependency
management

   and hence is more suitable for java build (e.g. Maven or Gradle)

Having 2) we could for example translate vpe.api to yang (messages as
RPCs), generate Java bindings (requires yangtools from ODL)

and have NETCONF support in HC for all VPP APIs.



The jVpp project could have two parts:

-Java API generator that could be referenced from outside (e.g.
in form of a maven plugin)

-actual API generation for all the plugins that live in VPP
codebase,

other fd.io plugins (e.g. NSH) could be also included or use API generator.



Regards,

Marek



___
vpp-dev mailing list
vpp-dev@lists.fd.io
https://lists.fd.io/mailman/listinfo/vpp-dev


___
vpp-dev mailing list
vpp-dev@lists.fd.io
https://lists.fd.io/mailman/listinfo/vpp-dev

Re: [vpp-dev] Problems with sub-interfaces

2016-10-21 Thread Kinsella, Ray


added a question in comments.

On 21/10/2016 08:42, Tibor Frank -X (tifrank - PANTHEON TECHNOLOGIES at 
Cisco) wrote:

Updated in https://jira.fd.io/browse/VPP-507



Tibor



*From:*John Lo (loj)
*Sent:* 20. októbra 2016 16:49
*To:* Dave Barach (dbarach); Tibor Frank -X (tifrank - PANTHEON
TECHNOLOGIES at Cisco); vpp-dev@lists.fd.io
*Subject:* RE: Problems with sub-interfaces



Since MAC on a sub-interface is the same as that of the main interface,
does it make sense to set MAC on a sub-interface? If so, wouldn’t MAC
also be changed for the main interface and other sub-interfaces on the
same port? I doubt NICs can use separate MACs for sub-interfaces as it
is transparent to them. I think it make sense to allow setting MAC on
the main interface only and the crash as shown in the log while setting
MAC on a sub-interface should be fixed.   -John



*From:*vpp-dev-boun...@lists.fd.io 
[mailto:vpp-dev-boun...@lists.fd.io] *On Behalf Of *Dave Barach (dbarach)
*Sent:* Thursday, October 20, 2016 10:13 AM
*To:* Tibor Frank -X (tifrank - PANTHEON TECHNOLOGIES at Cisco);
vpp-dev@lists.fd.io 
*Subject:* Re: [vpp-dev] Problems with sub-interfaces



Both Jira tickets now in “request information” state, please update...



Thanks… Dave



*From:*vpp-dev-boun...@lists.fd.io 
[mailto:vpp-dev-boun...@lists.fd.io] *On Behalf Of *Tibor Frank -X
(tifrank - PANTHEON TECHNOLOGIES at Cisco)
*Sent:* Thursday, October 20, 2016 9:59 AM
*To:* vpp-dev@lists.fd.io 
*Subject:* [vpp-dev] Problems with sub-interfaces



Hi

I found two issues:

1.  It is not possible to create a sub-interface using VAT -
https://jira.fd.io/browse/VPP-506

2.  It is not possible to set MAC address for a sub-interface -
https://jira.fd.io/browse/VPP-507



Can someone look at it, please.



Thanks

Tibor



___
vpp-dev mailing list
vpp-dev@lists.fd.io
https://lists.fd.io/mailman/listinfo/vpp-dev


___
vpp-dev mailing list
vpp-dev@lists.fd.io
https://lists.fd.io/mailman/listinfo/vpp-dev

Re: [vpp-dev] flowtable proposal

2016-10-19 Thread Kinsella, Ray

Gabriel,

Where do I find the scripts?

Delighted to see contributions, but am a bit fuzzy on where I might use
this.

Ray K

On 19/10/2016 11:11, gannega wrote:

Hi,

Following Christophe Fontaine's work during ietf hackathon in Berlin
(https://github.com/christophefontaine/flowtable-plugin), I wrote a
flowtable as a vpp plugin.

Code is available here:
https://github.com/GabrielGanne/vpp-flowtable/commits/master

It contains the flowtable plugin, a port-mirroring node I used to test
it, and also some quick scripts I used to test it.

Currently :

- Default behavior is to connect transparently to given interface.
- Can reroute packets to given node
- Can receive additional information, and offload sessions

I'm still working on this, do not support multithreading, and haven't
gone through performance tests yet.
Also, I'm planning to maybe add some features such as getting flow stats
at timeout.

If you can, I'd be happy to get feedback on this.

Best regards,

--
Gabriel Ganne

This message and any attachments (the "message") are confidential, intended
solely for the addressees. If you are not the intended recipient, please notify the
sender immediately by e-mail and delete this message from your system. In this case, you
are not authorized to use, copy this message and/or disclose the content to any other
person. E-mails are susceptible to alteration. Neither Qosmos nor any of its subsidiaries
or affiliates shall be liable for the message if altered, changed or falsified.

Ce message et toutes ses pièces jointes (ci-après le "message")sont
confidentiels et établis à l'intention exclusive de ses destinataires. Si vous avez reçu
ce message par erreur, merci d’en informer immédiatement son émetteur par courrier
électronique et d’effacer ce message de votre système. Dans cette hypothèse, vous n’êtes
pas autorisé à utiliser, copier ce message et/ou en divulguer le contenu à un tiers. Tout
message électronique est susceptible d'altération. Qosmos et ses filiales déclinent toute
responsabilité au titre de ce message s'il a été altéré, déformé ou falsifié.
___
vpp-dev mailing list
vpp-dev@lists.fd.io
https://lists.fd.io/mailman/listinfo/vpp-dev

___
vpp-dev mailing list
vpp-dev@lists.fd.io
https://lists.fd.io/mailman/listinfo/vpp-dev

Re: [vpp-dev] Is there a way to change the MAC address in VPP 1609?

2016-10-19 Thread Kinsella, Ray


Hi Chengwei,

Are you using DPDK? as many pmds do support this.
Not sure it's wired into VPP though?

typedef void (*eth_mac_addr_remove_t)(struct rte_eth_dev *dev, uint32_t 
index);

typedef void (*eth_mac_addr_add_t)(struct rte_eth_dev *dev,
  struct ether_addr *mac_addr,
  uint32_t index,
  uint32_t vmdq);
typedef void (*eth_mac_addr_set_t)(struct rte_eth_dev *dev,
  struct ether_addr *mac_addr);

Ray K


On 18/10/2016 21:53, WANG, CHENGWEI  (CHENGWEI) wrote:

Hi VPP developers,



As far as I know, the VPP 1609 version does not have



“Set interface mac address” command line.



I know the latest VPP has this command line but If I use VPP 1609, is
there any other way to change the mac address of an interface?



Thanks

Chengwei



___
vpp-dev mailing list
vpp-dev@lists.fd.io
https://lists.fd.io/mailman/listinfo/vpp-dev


___
vpp-dev mailing list
vpp-dev@lists.fd.io
https://lists.fd.io/mailman/listinfo/vpp-dev

Re: [vpp-dev] Vpp_main running at 100% CPU

2016-10-18 Thread Kinsella, Ray

  0   0  
 3  1.14e40.00
vpe-oam-process any wait 0   0  
   111  1.09e40.00
vpe-route-resolver-process  any wait 0   0  
 3  1.19e40.00

Thank you very much!!
Best Regards,
Miguel Ángel.

-Original Message-
From: Kinsella, Ray [mailto:ray.kinse...@intel.com]
Sent: martes, 18 de octubre de 2016 15:29
To: Miguel Angel Muñoz Gonzalez ; 
vpp-dev@lists.fd.io
Subject: Re: [vpp-dev] Vpp_main running at 100% CPU

Hi Miguel,

If you are using DPDK - it would be my expectation yes.
Can you do a 'show run' in vppctl to see where time is being spent.

Ray K

On 18/10/2016 09:02, Miguel Angel Muñoz Gonzalez wrote:

Hi,

In SFC project from ODL we are trying to use VPP to run a simple
scenario with basic service chain setup.  At the moment I'm running
vpp
1609 with honeycomb in Centos7, as a VirtualBox VM.

Once VPP is started, 'top' command shows vpp_main process taking 100%
CPU in one of the cores and never goes down. Is this normal?

I have VT-x/AMD-V/PAE/NX extensions activated.



Thank you,

Best Regards,

Miguel Ángel.





[root@localhost vagrant]# top -b -n1

top - 07:52:02 up  2:16,  4 users,  load average: 1.16, 1.25, 1.31

Tasks: 197 total,   3 running, 194 sleeping,   0 stopped,   0 zombie

%Cpu(s): 21.7 us, 17.5 sy,  0.0 ni, 60.6 id,  0.1 wa,  0.0 hi,  0.1
si,
0.0 st

KiB Mem : 10163584 total,  2786344 free,  4864612 used,  2512628
buff/cache

KiB Swap:  1572860 total,  1572860 free,0 used.  4967176 avail Mem



  PID USER  PR  NIVIRTRESSHR S  %CPU %MEM TIME+ COMMAND

8033 root  20   0 1933980  30820   4952 R 100.0  0.3  89:25.57 vpp_main

1 root  20   0  126600   7424   2628 S   0.0  0.1   0:01.75 systemd

2 root  20   0   0  0  0 S   0.0  0.0   0:00.00 kthreadd

3 root  20   0   0  0  0 S   0.0  0.0   0:00.18
ksoftirqd/0

5 root   0 -20   0  0  0 S   0.0  0.0   0:00.00
kworker/0:0H

6 root  20   0   0  0  0 S   0.0  0.0   0:00.68
kworker/u4:0

7 root  rt   0   0  0  0 S   0.0  0.0   0:00.07
migration/0







___
vpp-dev mailing list
vpp-dev@lists.fd.io
https://lists.fd.io/mailman/listinfo/vpp-dev


___
vpp-dev mailing list
vpp-dev@lists.fd.io
https://lists.fd.io/mailman/listinfo/vpp-dev

Re: [vpp-dev] Vpp_main running at 100% CPU

2016-10-18 Thread Kinsella, Ray


Hi Miguel,

If you are using DPDK - it would be my expectation yes.
Can you do a 'show run' in vppctl to see where time is being spent.

Ray K

On 18/10/2016 09:02, Miguel Angel Muñoz Gonzalez wrote:

Hi,

In SFC project from ODL we are trying to use VPP to run a simple
scenario with basic service chain setup.  At the moment I’m running vpp
1609 with honeycomb in Centos7, as a VirtualBox VM.

Once VPP is started, ‘top’ command shows vpp_main process taking 100%
CPU in one of the cores and never goes down. Is this normal?

I have VT-x/AMD-V/PAE/NX extensions activated.



Thank you,

Best Regards,

Miguel Ángel.





[root@localhost vagrant]# top -b -n1

top - 07:52:02 up  2:16,  4 users,  load average: 1.16, 1.25, 1.31

Tasks: 197 total,   3 running, 194 sleeping,   0 stopped,   0 zombie

%Cpu(s): 21.7 us, 17.5 sy,  0.0 ni, 60.6 id,  0.1 wa,  0.0 hi,  0.1 si,
0.0 st

KiB Mem : 10163584 total,  2786344 free,  4864612 used,  2512628 buff/cache

KiB Swap:  1572860 total,  1572860 free,0 used.  4967176 avail Mem



  PID USER  PR  NIVIRTRESSHR S  %CPU %MEM TIME+ COMMAND

8033 root  20   0 1933980  30820   4952 R 100.0  0.3  89:25.57 vpp_main

1 root  20   0  126600   7424   2628 S   0.0  0.1   0:01.75 systemd

2 root  20   0   0  0  0 S   0.0  0.0   0:00.00 kthreadd

3 root  20   0   0  0  0 S   0.0  0.0   0:00.18
ksoftirqd/0

5 root   0 -20   0  0  0 S   0.0  0.0   0:00.00
kworker/0:0H

6 root  20   0   0  0  0 S   0.0  0.0   0:00.68
kworker/u4:0

7 root  rt   0   0  0  0 S   0.0  0.0   0:00.07
migration/0







___
vpp-dev mailing list
vpp-dev@lists.fd.io
https://lists.fd.io/mailman/listinfo/vpp-dev


___
vpp-dev mailing list
vpp-dev@lists.fd.io
https://lists.fd.io/mailman/listinfo/vpp-dev

81 matches

Mail list logo