Re: [vpp-dev] Replacing pycodestyle with black

Also an item to ponder - line length. PEP8 says line length should be <= 79 IIRC. Black advocates line length of 88, saying it produces least lines of code by experience and it’s the default. https://black.readthedocs.io/en/stable/the_black_code_style/current_style.html#line-length On 2 May

Re: [vpp-dev] #epoll problems with epoll

Hi, It can be that vpp reset the connection due to some bogus packets or because the connection could not be allocated. Check “show errors”. Given the config, in particular, 30M fifos (those are pretty large) and 32k event queue size, maybe add: segment-size 70 add-segment-size 7

[vpp-dev] #epoll problems with epoll

[Edited Message Follows] Hi, Florin, I'm using epoll with LDP, but sometimes I receive EPOLLRDHUP|EPOLLHUP unexpectedly when the opposite side didn't send the FIN or RST packet. Here is my vcl config, am Iconfiguring something wrong? Thank you very much for your help. * vcl { rx-fifo-size 3

Re: [vpp-dev] Crash in tcp46_input_inline, vpp 21.06

Hi Vijay, Error seems unrelated to the patch. Hard to say from the trace what’s going on. If f5 is to be believed, wrk pointer looks corrupted. Hard to say why that happened given that tcp_get_worker should not fail. Maybe inspect &tcp_main.wrk_ctx[thread_index] or run a debug image. Regards,

[vpp-dev] #epoll problems with epoll

[Edited Message Follows] Hi, Florin, I'm using epoll with LDP, but sometimes I receive EPOLLRDHUP|EPOLLHUP unexpectedly when the opposite side didn't send the FIN packet. Here is my vcl config, am Iconfiguring something wrong? Thank you very much for your help. * vcl { api-socket-name /run/v

[vpp-dev] #epoll problems with epoll

Hi, Florin, I'm using epoll with LDP, but sometimes I receive EPOLLRDHUP|EPOLLHUP unexpectedly when the opposite side didn't send the FIN packet. Here is my vcl config, am Iconfiguring something wrong? Thank you very much for your help. -=-=-=-=-=-=-=-=-=-=-=- Links: You receive all messages sen

[vpp-dev] Crash in tcp46_input_inline, vpp 21.06

Hi experts, I am seeing the below pasted call stack. We are using vpp 21.06. The crash is in tcp_update_time_now() called inside tcp46_input_inline. I had not done any changes to *tcp46_input_inline *graph node. All I did was add this patch mentioned below. This patch was applied recently in the f

[vpp-dev] Connect QEMU VM to Internet through VPP/DPDK #azure #vhost #dpdk

Hello, I'm new to VPP. I have a scenario where I need to run a QEMU VM instance on an Azure VM and connect the QEMU VM to the Internet through VPP. The QEMU VM is connected to the VPP through vhost-user interface and a NIC from the host Azure VM is also attached to the VPP. However, I'm struggl

Re: [vpp-dev] Query on VPP in Aure

Hello Matt, Thank you again for the response. So we were able to compile but now are unable to start VPP. We are seeing the following errors: I rebuilt, and got: May 02 20:20:24 demo001 vpp[2161]: vpp[2161]: perfmon: skipping source 'intel-uncore' - intel_uncore_init: no uncore units found Ma

Re: [vpp-dev] Replacing pycodestyle with black

I'll work on expanding Klement's patch to include all python files in the repo. I'm thinking it makes more sense to have a single target 'make checkstyle-python' (and corresponding 'fixstyle-python') that runs black against all of the '*.py' files in the repo and deprecate the 'test-checkstyl

Re: [vpp-dev] Replacing pycodestyle with black

I’d be happy to have formatting consistency across the project. Cheers Ole > On 2 May 2022, at 19:17, Klement Sekera wrote: > >  +Real Ole. > >>> On 2 May 2022, at 19:15, Klement Sekera via lists.fd.io >>> wrote: >>> >>> Hi Dave, >>> >>> good catch about requirements.txt. Should be fixe

Re: [vpp-dev] Replacing pycodestyle with black

Hoi, As a (small) voice of support, Black is a great formatter. Happy to support it in vpp's Python codebase! It's also the vppcfg formatter of choice :) groet, Pim On Mon, May 2, 2022 at 3:56 AM Klement Sekera wrote: > Hey vpp-dev, > > looking at our current code style and its PEP8 compliance

Re: [vpp-dev] Query on VPP in Aure

Sorry, there was a typo. You should use DPDK_MLX5_COMMON_PMD=y instead of DPDK_MLX_COMMON_PMD=y. -Matt On Mon, May 2, 2022 at 11:28 AM Johnny Martinez wrote: > Hello Matt! > > *Thank you very much for the reply!* > > Our end goal is to be using Alma Linux but we went ahead and tried it in a >

[vpp-dev] Help requested with VPP not working with 25G intel adapter #vpp-dev

Hello, Trying to get an Intel adapter to work with VPP and cannot figure out the error. Any help is welcome and appreciated! Adapter:  Intel(R) Ethernet Network Adapter E810-XXVDA2, driver: ice (version:  5.13.0-40-generic) VPP: v22.02 Linux: kernel 5.13.0-40-generic, Ubuntu focal 20.04.3 LTS W

Re: [vpp-dev] Replacing pycodestyle with black

+Real Ole. On 2 May 2022, at 19:15, Klement Sekera via lists.fd.io mailto:klement=graphiant@lists.fd.io>> wrote: Hi Dave, good catch about requirements.txt. Should be fixed now. I’m not sure how to proceed with rest of code, seems to be mostly Ole’s domain. Thoughts,

Re: [vpp-dev] Replacing pycodestyle with black

Hi Dave, good catch about requirements.txt. Should be fixed now. I’m not sure how to proceed with rest of code, seems to be mostly Ole’s domain. Thoughts, Ole? Thanks, Klement On 2 May 2022, at 18:50, Dave Wallace mailto:dwallac...@gmail.com>> wrote: Hi Klement, I agree that black is a much

Re: [vpp-dev] test performance of nginx using vpp host stack#vpp-hoststack

Hi, That indeed looks like an issue due to vpp not being able to recycle connections fast enough. There are only 64k connections available between vpp and the upstream server, so recycling them as fast as possible, i.e., with 0 timeout as the kernel does after tcp_max_tw_buckets threshold is h

Re: [vpp-dev] Replacing pycodestyle with black

Hi Klement, I agree that black is a much better PEP8 compliance/formatter for python code.  I've been using it on my personal python projects for a while now. I left a minor comment on your gerrit change wrt. installation of pinned version as part of the rest of the python virtualenv. Otherwi

[vpp-dev] FD.io Nexus SHUTDOWN: 2022-05-16 16:00 UTC

*What*: LF will be decommissioning https://nexus.fd.io on 2022-05-16 at 16:00 UTC. We believe all required artifacts have been migrated to PackageCloud (packages) and S3 (logs and docs). If your project has artifacts in Nexus that need to be migrated please migrate required artifacts or open

Re: [vpp-dev] test performance of nginx using vpp host stack#vpp-hoststack

Hi, The short link means that after the client send GET request, the client send tcp FIN packet. Instead, the long link means that after the client send GET request,  the client send next http GET request by using the same link and don't need to send syn packet. We found that when vpp and the up

Re: [vpp-dev] test performance of nginx using vpp host stack#vpp-hoststack

Hi, As per this [1], after tcp_max_tw_buckets threshold is hit timewait time is 0 and this [2] explains what will go wrong. Assuming you’re hitting the threshold, 1s timewait-time in vpp will probably not be enough to match performance. Not sure what you mean by “short link”. If you can’t us

[vpp-dev] Replacing pycodestyle with black

Hey vpp-dev, looking at our current code style and its PEP8 compliance I see a couple of issues 1.) it’s not really PEP8 compliant 2.) we have no test-fixstyle I decided to make it more PEP8 compliant and found out that to just fix W503 to W504 (per https://peps.python.org/pep-0008/#should-a-

[vpp-dev] RCT - alternative router configuration tool

Hi, I'd like to share the simple configuration utility for VPP project. It is an alternative router configuration tool based on JSON configuration files. It has a limited feature set (ip addres, vxlan, bridge, dot1q,...). Most changes in configuration are reflected in the VPP process using the GoV