Re: [vpp-dev] Can "use mq eventfd" solve epoll wait h igh cpu usage problem?

Hi Hanlin, Stephen and Ping have made a lot of progress with Envoy and VCL, but I’ll let them comment on that. Regards, Florin > On Oct 31, 2019, at 9:44 PM, wanghanlin wrote: > > OK，I got it. Thanks a lot. > By the way, can VCL adapt to envoy or any progress about this？ > > Regards, >

Re: [vpp-dev] Can "use mq eventfd" solve epoll wait h =?utf-8?Q?_igh_cpu_usage_problem??=

OK，I got it. Thanks a lot.By the way, can VCL adapt to envoy or any progress about this？ Regards,Hanlin

Re: [EXT] Re: [vpp-dev] cross-compilation support

Hi, >> cmake /path/to/vpp/src [your favorite xcompile args] cmake —build . Instead of having a shell script to pass cmake commands, I really liked the idea of passing vpp_cmake_args from build-data/platforms/*.mk (https://gerrit.fd.io/r/#/c/vpp/+/21035/23/build-data/platforms/native.mk@42).

Re: [vpp-dev] Can "use mq eventfd" solve epoll wait h igh cpu usage problem?

Hi Hanlin, If a worker’s mq uses eventfds for notifications, we could nest it in libc_epfd, i.e.,the epoll fd we create for the linux fds. So, if an app's worker calls epoll_wait, in ldp we can epoll_wait on libc_epfd and if we get an event on the mq’s eventfd, we can call vls_epoll_wait with

Re: [vpp-dev] nat44 bug - created nat sessions aren't automatically cleaned up

Hi Filip, Thank you! I can do some tests if you want. On Thu, Oct 31, 2019 at 7:14 AM Filip Varga -X (fivarga - PANTHEON TECH SRO at Cisco) wrote: > Hi, > > > > We have registred this issue int NAT plugin and i am already working on > solving the issue. If you need to post any additional

Re: [vpp-dev] Can "use mq eventfd" solve epoll wait h =?utf-8?Q?igh_cpu_usage_problem??=

Do you mean, if just use eventfds only, then I needn't set timeout  to 0 in ldp_epoll_pwait?If so, then how to process unhandled_evts_vector in vppcom_epoll_wait timely? What I'm saying is,  another thread add event to unhandled_evts_vector during epoll_wait, or 

Re: [vpp-dev] vpp with gtpu

Hi, Neale, Hongjun The environment issue was fixed. Thanks for your help. Thanks, -Ruijing From: Guo, Ruijing Sent: Thursday, October 31, 2019 4:43 PM To: Neale Ranns (nranns) ; Ni, Hongjun ; vpp-dev@lists.fd.io Subject: RE: [vpp-dev] vpp with gtpu Good catch, neale I will check why it use

Re: [vpp-dev] RDMA fix needed in 19.08 also

Elias, Thanks for telling! I have cherry-picked https://gerrit.fd.io/r/#/c/vpp/+/23164/ and will merge it tomorrow. --a > On 31 Oct 2019, at 19:18, Elias Rudberg wrote: > > 386ebb6e2b -=-=-=-=-=-=-=-=-=-=-=- Links: You receive all messages sent to this group. View/Reply Online (#14446):

Re: [EXT] [vpp-dev] cross-compilation support

> On Oct 31, 2019, at 5:39 PM, Damjan Marion via Lists.Fd.Io > wrote: > > > >> On 31 Oct 2019, at 21:41, Christian Hopps wrote: >> >>> >>> On Oct 31, 2019, at 4:03 PM, Damjan Marion via Lists.Fd.Io >>> wrote: >>> >>> >>> Let me clarify a bit more, as i can see that this still may be

Re: [EXT] [vpp-dev] cross-compilation support

> On 31 Oct 2019, at 21:41, Christian Hopps wrote: > >> >> On Oct 31, 2019, at 4:03 PM, Damjan Marion via Lists.Fd.Io >> wrote: >> >> >> Let me clarify a bit more, as i can see that this still may be confusing: >> >> 1) VPP is using cmake to build binaries and packages, all cmake related

Re: [EXT] [vpp-dev] cross-compilation support

> On Oct 31, 2019, at 4:03 PM, Damjan Marion via Lists.Fd.Io > wrote: > > > Let me clarify a bit more, as i can see that this still may be confusing: > > 1) VPP is using cmake to build binaries and packages, all cmake related stuff > is in src/. > > 2) there is build-root/Makefile and

Re: [vpp-dev] event-logger output format

Dave, Couldn’t have explained it better! Thanks much! Given that most CPU’s are 2+ cores, it should be okay but, you say “Note that most debug CLI commands are not marked thread-safe” Does it mean, it might they would still revert to a single-core like operation? Thanks! -- Regards, Balaji.

Re: [EXT] [vpp-dev] cross-compilation support

Let me clarify a bit more, as i can see that this still may be confusing: 1) VPP is using cmake to build binaries and packages, all cmake related stuff is in src/. 2) there is build-root/Makefile and files in build-data/* which are part of old build system called ebuild - ebuild is very

Re: [EXT] [vpp-dev] cross-compilation support

I have similar scheme on my mind, where you can have platform specific mk files loaded…. > On 31 Oct 2019, at 20:23, Nitin Saxena wrote: > > Hi, > >>> cmake /path/to/vpp/src [your favorite xcompile args] cmake —build . > Instead of having a shell script to pass cmake commands, I really liked

Re: [vpp-dev] event-logger output format

Two cases: in single-core operation, debug CLI commands change the NDR / max throughput. A single “format” call might cost as many clock cycles as processing 50 packets. If the one and only core is busy printing something, it’s not processing packets. In multi-core operation, CLI commands

Re: [vpp-dev] cross-compilation support

I slightly modified my patch, so it builds both native packages and cross packages out of same makefile. Native: $ make pkg $ ls -la build/native-release total 99352 drwxrwxr-x 3 damarion damarion 4096 Oct 31 20:00 . drwxrwxr-x 10 damarion damarion 4096 Oct 31 19:55 .. drwxrwxr-x 22

Re: [vpp-dev] cross-compilation support

Nobody requires you to use docker, you are free to pass right arguments straight to the cmake. All the stuff in build-data, build-root, build/ is optional, and it is there to help developers to stage workspace and create packages but VPP can be built as simple as: mkdir build cd build cmake

Re: [vpp-dev] Centos 8 and python scripts

On 31/10/19 13:28, Renato Botelho do Couto wrote: > On 31/10/19 13:17, Paul Vinciguerra wrote: >> Speaking for myself only, I would prefer that we didn't introduce per >> platform changes.  It would seem better to apply a single change that >> makes it through the CI gates. > > Right, I'll work

Re: [vpp-dev] cross-compilation support

I mean to call out the use of docker/qemu. In order to allow for this the user has to have extra privileges. As I mentioned, locally we use docker for building, I like it, it's a nice tool; however, my personal opinion is that it seems wrong to require docker and/or qemu for compiling an open

Re: [vpp-dev] cross-compilation support

That is just copy paste from Ben’s command line. As. i wrote, this is just about presenting the concept... > On 31 Oct 2019, at 18:39, Christian Hopps wrote: > > Requiring root privileges for building software is pretty atypical and > limiting, even for x-compiling. > > Thanks, > Chris. >

[vpp-dev] RDMA fix needed in 19.08 also

It seems like the rdma plugin is currently not working in the stable/1908 branch. It stopped working after commit b4c5f16889. In the master branch, the rdma plugin stopped working in commit 534de8b2a7 but started working again after the fix in commit 386ebb6e2b with commit message "rdma: build:

Re: [vpp-dev] Centos 8 and python scripts

On 31/10/19 13:39, Andrew  Yourtchenko wrote: > Renato, > > For my own curiosity sake - would you have a minute to try the following > patch and instructions within it and tell the result ? There is a list of good/bad shebangs used by brp-mangle-shebangs script. This custom shebang line is not

Re: [vpp-dev] cross-compilation support

Requiring root privileges for building software is pretty atypical and limiting, even for x-compiling. Thanks, Chris. > On Oct 31, 2019, at 12:45 PM, Damjan Marion via Lists.Fd.Io > wrote: > > > So what people think about following approach for cross-compilation: > > Patch:

Re: [vpp-dev] per-node tracing

$ git grep -h -B20 VLIB_NODE_FLAG_TRACE_SUPPORTED | grep VLIB_REGISTER_NODE VLIB_REGISTER_NODE (avf_input_node) = { VLIB_REGISTER_NODE (dpdk_input_node) = { VLIB_REGISTER_NODE (dpdk_crypto_input_node) = VLIB_REGISTER_NODE (ixge_input_node, static) = { VLIB_REGISTER_NODE (mrvl_pp2_input_node) = {

[vpp-dev] cross-compilation support

So what people think about following approach for cross-compilation: Patch: https://gerrit.fd.io/r/c/vpp/+/23153 At the moment this is just to show idea, so it works only for x86-to-x86 and between ubuntu versions only but it can be easily extended to

Re: [vpp-dev] Centos 8 and python scripts

Renato, For my own curiosity sake - would you have a minute to try the following patch and instructions within it and tell the result ? https://gerrit.fd.io/r/#/c/vpp/+/23152/ --a > On 31 Oct 2019, at 17:28, Renato Botelho do Couto via Lists.Fd.Io > wrote: > > On 31/10/19 13:17, Paul

Re: [vpp-dev] Centos 8 and python scripts

On 31/10/19 13:17, Paul Vinciguerra wrote: > Speaking for myself only, I would prefer that we didn't introduce per > platform changes.  It would seem better to apply a single change that > makes it through the CI gates. Right, I'll work on a patch changing all shebang lines to python3 and submit

Re: [vpp-dev] event-logger output format

Do you think it would be prudent to publish a ‘warning’ message on its impact on packet forwarding when this cmd is used? Thanks! -- Regards, Balaji. From: on behalf of Aleksander Djuric Date: Thursday, October 31, 2019 at 8:07 AM To: "vpp-dev@lists.fd.io" Subject: Re: [vpp-dev]

Re: [vpp-dev] Centos 8 and python scripts

Speaking for myself only, I would prefer that we didn't introduce per platform changes. It would seem better to apply a single change that makes it through the CI gates. Is it time for a (non-voting) Centos8 CI job? On Thu, Oct 31, 2019 at 12:06 PM Renato Botelho do Couto wrote: > On 31/10/19

Re: [vpp-dev] Centos 8 and python scripts

On 31/10/19 11:09, Paul Vinciguerra wrote: > I am aware of the issue.  I would suggest changing them all to python3.  > We can work through any issues that arise after the code passes the CI > gates.  Most of those shebangs are in place for developers using make > test-shell, which could be

Re: [vpp-dev] per-node tracing

grep for VLIB_NODE_FLAG_TRACE_SUPPORTED Cf: Damjan’s commit 29dc11bd --a > On 31 Oct 2019, at 16:45, Paul Vinciguerra wrote: > >  > Is there some simple way to identify the nodes that are supported under trace > add? > > > trace add: node 'ethernet-input' doesn't support per-node tracing.

[vpp-dev] per-node tracing

Is there some simple way to identify the nodes that are supported under trace add? trace add: node 'ethernet-input' doesn't support per-node tracing. There may be another way to initiate trace on this node. -=-=-=-=-=-=-=-=-=-=-=- Links: You receive all messages sent to this group. View/Reply

Re: [vpp-dev] Can "use_mq_eventfd" solve epoll_wait high cpu usage problem?

Hi, use_mq_eventfd will help with vcl but as you’ve noticed it won’t help for ldp because there we need to poll both vcl and linux fds. Because mutex-condvar notifications can’t be epolled we have to constantly switch between linux and vcl epolled fds. One option going forward would be to

Re: [vpp-dev] event-logger output format

Sample output: > > vpp# show event-logger 2 > 82 of 131072 events in buffer, logger running > 19.946551561: api-msg: trace_plugin_msg_ids > 19.946557011: api-msg-done(mp-safe): trace_plugin_msg_ids > vpp# show event-logger 2 no-delta > 82 of 131072 events in buffer, logger running >

Re: [vpp-dev] event-logger output format

Hi Dave, Thanks! Now it looks like this: > show event-logger [all] [] [delta|no-delta|date-time] Updated patch: https://gerrit.fd.io/r/c/vpp/+/23146 On Thu, Oct 31, 2019 at 04:47 PM, Dave Barach wrote: > > > > Yes, we see things differently. I never look at elog data in the debug > CLI. I

Re: [vpp-dev] Centos 8 and python scripts

Why not delete them altogether then? --a > On 31 Oct 2019, at 15:09, Paul Vinciguerra wrote: > >  > I am aware of the issue. I would suggest changing them all to python3. We > can work through any issues that arise after the code passes the CI gates. > Most of those shebangs are in place

Re: [vpp-dev] Centos 8 and python scripts

I am aware of the issue. I would suggest changing them all to python3. We can work through any issues that arise after the code passes the CI gates. Most of those shebangs are in place for developers using make test-shell, which could be evoking the script via 'python2 ' On Thu, Oct 31, 2019 at

Re: [vpp-dev] event-logger output format

Yes, we see things differently. I never look at elog data in the debug CLI. I save the data and deal with it offline. See https://fd.io/docs/vpp/master/gettingstarted/developers/eventviewer.html. We often use elog buffer sizes measured in millions of events. The offline tools are happy to deal

[vpp-dev] Centos 8 and python scripts

I've been working to build VPP on CentOS 8 and rpmbuild is complaining about using '#!/usr/bin/env python' or '#!/usr/bin/python' on scripts' shebang line. It says it must to define explicitly if it's using python2 or python3. # git grep 'env python$' | wc -l 117 There are several scripts with

[vpp-dev] Can "use_mq_eventfd" solve epoll_wait high cpu usage problem?

hi ALL， I found app using VCL "epoll_wait" still occupy 70% cpu with "use_mq_eventfd" configuration even if very little traffic.Then I investigate code in ldp_epoll_pwait, vls_epoll_wait is called with timeout equal to 0.Then I have two questions:1. What

Re: [vpp-dev] Change in vpp[master]: ebuild: Cross compilation aarch64 Ubuntu support

> On 31 Oct 2019, at 13:18, Benoit Ganne (bganne) wrote: > >> I was going to remain silent, but since there's now multiple people saying >> this sounds good -- I think this sounds horrible. :) >> To wit, it seems too complex and too much setup/overhead. I'll try and >> look closer at this soon

Re: [vpp-dev] event-logger output format

Hi Dave, Thank you very much for your reply. I think it can be useful to discuss this patch here at first. As I see, we looks at these things little bit differently ) I'm use event-logger mostly for registering such fast events as ACL filter matching. I think that the human-readable format is

Re: [vpp-dev] Change in vpp[master]: ebuild: Cross compilation aarch64 Ubuntu support

> I was going to remain silent, but since there's now multiple people saying > this sounds good -- I think this sounds horrible. :) > To wit, it seems too complex and too much setup/overhead. I'll try and > look closer at this soon to see if I can feed back our local changes that > seem to be

[vpp-dev] mildly off-topic: non-member posts to vpp dev list

Folks, Behind the scenes, the vpp dev mailing list receives a lot of spam from non-list-member accounts. I delete them from the moderation queue periodically, roughly twice a week. At cleanup time, legitimate emails stuck in the moderation queue will appear on the dev list. Anyone who might

Re: [vpp-dev] how to update openssl in the vpp

Thank you very much for your reply: Our system uses centos7, vpp-based ipsec wants to use the encryption algorithm sm4. but the current centos7 openssl version 1.0.2k, but the sm4 algorithm is in the higher openssl version 1.1.1. we try to compile the openssl 1.1.1 version code, replace

Re: [vpp-dev] event-logger output format

Please see https://fd.io/docs/vpp/master/gettingstarted/developers/gitreview.html. Event log entry timestamps have nanosecond granularity. It might work out to do something like the following: 2019/10/31 09:18:36:937:123:895 ## initial full timestamp +0: api-msg: trace_plugin_msg_ids

Re: [vpp-dev] vpp with gtpu

Good catch, neale I will check why it use 192.168.1.1. From: Neale Ranns (nranns) Sent: Thursday, October 31, 2019 4:39 PM To: Guo, Ruijing ; Ni, Hongjun ; vpp-dev@lists.fd.io Subject: Re: [vpp-dev] vpp with gtpu Hi Ruijing, Comments inline. /neale From: mailto:vpp-dev@lists.fd.io>> on

Re: [vpp-dev] vpp with gtpu

Hi Ruijing, Comments inline. /neale From: on behalf of "Guo, Ruijing" Date: Thursday 31 October 2019 at 09:29 To: "Ni, Hongjun" , "vpp-dev@lists.fd.io" Subject: Re: [vpp-dev] vpp with gtpu Hi, Hongjun After adding the configuration, ping still failed as: Packet 14 00:02:10:052842:

Re: [vpp-dev] vpp with gtpu

Hi, Hongjun After adding the configuration, ping still failed as: Packet 14 00:02:10:052842: af-packet-input af_packet: hw_if_index 1 next-index 4 tpacket2_hdr: status 0x81 len 146 snaplen 146 mac 66 net 80 sec 0x5dbb241d nsec 0x2a83d8a6 vlan 0 vlan_tpid 0 00:02:10:052858:

Re: [vpp-dev] vpp with gtpu

Hi Ruijing, It seems that GTPU did not failed on the fib check. Please add below items on your configuration: sudo vppctl create gtpu tunnel src 192.168.1.22 dst 192.168.1.21 teid 13 encap-vrf-id 0 decap-next ip4 Thanks, Hongjun From: Guo, Ruijing Sent: Thursday, October 31, 2019 3:23 PM To:

Re: [vpp-dev] vpp with gtpu

Hi, Hongjun Packets are attached Thanks, -Ruijing From: Ni, Hongjun Sent: Thursday, October 31, 2019 3:12 PM To: Guo, Ruijing ; vpp-dev@lists.fd.io Subject: RE: vpp with gtpu Hi Ruijing, It seems that there may be some issue on the gtpu decapsulation. Could you help to capture the traffic

Re: [vpp-dev] vpp with gtpu

Hi Ruijing, It seems that there may be some issue on the gtpu decapsulation. Could you help to capture the traffic between the vpp1 side and vpp2 side? Thanks, Hongjun From: Guo, Ruijing Sent: Thursday, October 31, 2019 2:57 PM To: vpp-dev@lists.fd.io Cc: Ni, Hongjun Subject: vpp with gtpu

[vpp-dev] vpp with gtpu

Hi, I tried vpp (18.04-release, 19.08.1-release) with gtpu. it cannot work. Could you pls help? Vpp1 setup: sudo vppctl create host-interface name eth1 sudo vppctl set inter stat host-eth1 up sudo vppctl set int mac address host-eth1 52:54:00:43:88:63 sudo vppctl set int ip add host-eth1

[vpp-dev] event-logger output format

Hi all, In my opinion, the event-logger cli command has uncomfortable format of output. I have small suggestion about it. I think that the event-logger output should be printed in following manner: vpp# show event-logger > 72 of 131072 events in buffer, logger running > 2019/10/31 09:18:36:937: