Re: [vpp-dev] tcpdump like feature

Any input on this? -=-=-=-=-=-=-=-=-=-=-=- Links: You receive all messages sent to this group. View/Reply Online (#15482): https://lists.fd.io/g/vpp-dev/message/15482 Mute This Topic: https://lists.fd.io/mt/71384285/21656 Group Owner: vpp-dev+ow...@lists.fd.io Unsubscribe: https://lists.fd.io/g/vp

Re: [vpp-dev] tcpdump like feature

I am using vpp 19.04. -=-=-=-=-=-=-=-=-=-=-=- Links: You receive all messages sent to this group. View/Reply Online (#15458): https://lists.fd.io/g/vpp-dev/message/15458 Mute This Topic: https://lists.fd.io/mt/71384285/21656 Group Owner: vpp-dev+ow...@lists.fd.io Unsubscribe: https://lists.fd.io/g

[vpp-dev] tcpdump like feature

Hello all, Do we have anything similar to "tcpdump -ni " in vpp? I can see two options. But one of them is not working for multi-core. And another is not allowing me to capture packets based on a tcpdump filter. Option # 1 vpp# pcap tx trace on intfc GigabitEthernet0/11/0 file gige_tx.pcap I

Re: [vpp-dev] vlib_node_add_next usage #vpp

[Edited Message Follows] Thanks Christian and Florin. I am trying without a foreach loop. Things are looking good so far. -=-=-=-=-=-=-=-=-=-=-=- Links: You receive all messages sent to this group. View/Reply Online (#14110): https://lists.fd.io/g/vpp-dev/message/14110 Mute This Topic: https://l

Re: [vpp-dev] vlib_node_add_next usage #vpp

Thanks Christian and Florin. I am trying without a foreach loop. things are looking good so far. -=-=-=-=-=-=-=-=-=-=-=- Links: You receive all messages sent to this group. View/Reply Online (#14110): https://lists.fd.io/g/vpp-dev/message/14110 Mute This Topic: https://lists.fd.io/mt/34376308/216

Re: [vpp-dev] vlib_node_add_next usage #vpp

The session_register_transport() has the foreach code. 1381 /* *INDENT-OFF* */ 1382 if (output_node != ~0) 1383 { 1384 foreach_vlib_main ( https://docs.fd.io/vpp/19.08/db/d96/threads_8h.html#a26faa6aef056b1bbb12dc83ce874923c ) (({ 1385 next_index = vlib_node_add_next ( https://docs.fd.io/vpp/19

[vpp-dev] vlib_node_add_next usage #vpp

Hello, I am seeing vlib_node_add_next is getting used in two different ways. vlib_node_add_next(vm, node_index, next_node_index); And foreach_vlib_main (({ next_index = vlib_node_add_next (this_vlib_main, node_index, next_node_index); })); The first one (without foreach_vlib_main) calls vlib_n

[vpp-dev] VPP/RSS questions

Hi,  I am trying to play with the RSS feature by enabling multiple rx queues for the dpdk interface. However, when I do "show hardware-interfaces detail", I can see the lines below for RSS.      rss active:        ipv4-tcp ipv6-tcp-ex ipv6-tcp ipv6-udp ipv6-ex     rss supported:     none It sa

Re: [vpp-dev] NAT handoff node graph question

I think I found it.  vlib_frame_queue_main_init (sm->in2out_output_node_index, 0); So, it will be passing to the in2out node of the next core.  Thanks Damjan. -=-=-=-=-=-=-=-=-=-=-=- Links: You receive all messages sent to this group. View/Reply Online (#12490): https://lists.fd.io/g/vpp-dev/mess

[vpp-dev] VPP unit test with multiple workers

I can see that we can create pg- interface for unit testing various vpp features.  Is it possible to test multiple workers? How can I make sure the packets will be sent to different workers on a pg- interface? -=-=-=-=-=-=-=-=-=-=-=- Links: You receive all messages sent to this group. View/Reply

[vpp-dev] VPP unite test with multiple worker threads

Hello,  I can see that we can create pg- interface for unit testing various vpp features.  Is it possible to test multiple workers? How can I make sure the packets will be sent to different workers on a pg- interface? -=-=-=-=-=-=-=-=-=-=-=- Links: You receive all messages sent to this group. V

Re: [vpp-dev] NAT handoff node graph question

Hi Damjan,  Thanks for the reply. I am not very familiar with VPP yet.  Are you suggesting that the pkt will start its journey all the way from dpdk-input on the new core?  Can't we just send the pkt directly to nat44-in2out node on the new core?  In the code, I see that, we are calling vlib_get

[vpp-dev] NAT handoff node graph question

Hello,  Once handoff node decides to handoff a pkt to another core, which node on the other core picks the pkt up?  I am using 1807, and it seems like nst44-handoff-classify will send the pkt to either nat44-in2out-worker-handoff or nat44-out2in-worker-handoff node.  nat44-in2out-worker-handoff

[vpp-dev] multi-queue tap interface #vpp

Does vpp support multi-queue tap interface? If yes, how do I create a tap interface with multi-queue support? -=-=-=-=-=-=-=-=-=-=-=- Links: You receive all messages sent to this group. View/Reply Online (#12030): https://lists.fd.io/g/vpp-dev/message/12030 Mute This Topic: https://lists.fd.io/mt

Re: [vpp-dev] CPU usage on vpp_main thread is high when using tap cli between vpp and another userspace application #vpp

Hi Damjan, Thanks for your quick response. Will try tapv2. Are there any recommendations / best_practices for vpp to send/receive packets to/from a socket based application running side by side? Will a veth pair be any better than tapv2 in this case? -R > On Jan 24, 2019, at 11:43 PM, Damja

[vpp-dev] CPU usage on vpp_main thread is high when using tap cli between vpp and another userspace application #vpp

Hello,  I am trying to use vpp with 4 worker threads.  Client is sending packet to server through vpp. This packet comes to vpp, and gets picked up from a VF. VPP sends this packet to the another user-space application over a tapcli interface.  The user-space application processes the packet, an

[vpp-dev] Sending information from vpp to Linux user space daemon periodically#vpp

Hello,  I have a use-case where vpp should be able to send data to an user-space daemon running on the same Linux box.  The data should be sent periodically. Please let me know if there is any recommended way to achieve this.       So far, I have come across the options below.  Events type vpp/