From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from mails.dpdk.org (mails.dpdk.org [217.70.189.124]) by inbox.dpdk.org (Postfix) with ESMTP id 8DCEE45770; Thu, 8 Aug 2024 23:48:39 +0200 (CEST) Received: from mails.dpdk.org (localhost [127.0.0.1]) by mails.dpdk.org (Postfix) with ESMTP id 8560042DE3; Thu, 8 Aug 2024 23:48:39 +0200 (CEST) Received: from mail-oi1-f175.google.com (mail-oi1-f175.google.com [209.85.167.175]) by mails.dpdk.org (Postfix) with ESMTP id 8BF2B42D89 for ; Thu, 8 Aug 2024 23:48:37 +0200 (CEST) Received: by mail-oi1-f175.google.com with SMTP id 5614622812f47-3db1e21b0e4so1040279b6e.2 for ; Thu, 08 Aug 2024 14:48:37 -0700 (PDT) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=iol.unh.edu; s=unh-iol; t=1723153716; x=1723758516; darn=dpdk.org; h=content-transfer-encoding:cc:to:subject:message-id:date:from :mime-version:from:to:cc:subject:date:message-id:reply-to; bh=ycaKMMG1s5wKB5++TkpxWhxxk8pKgyjZkn2QYgxq79c=; b=OIFgsnKoiQmJQxg8sbj1ES+RN+8Ii2P1zlvn3iNkD6EqHFtFmGX9AVu0J7DJH1SOsF CARuq8UHCcl4DOwkKGUTTQGSdFdM9+4Fy2YqhDHQFvHtSPjxb3qsIWZQeN55TcXtpQb+ nP5ycrkiG9qNN3E+kN9NIXbHa0q5GcDv2tmXw= X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20230601; t=1723153716; x=1723758516; h=content-transfer-encoding:cc:to:subject:message-id:date:from :mime-version:x-gm-message-state:from:to:cc:subject:date:message-id :reply-to; bh=ycaKMMG1s5wKB5++TkpxWhxxk8pKgyjZkn2QYgxq79c=; b=Z+SVsWAG5cPxnjyMeoTF0Gf5jgE01vzZlmoPZD9mYEqC+MvHOnlbGnLmPfl+Dt9KtX 00uMn7Pxg0bXPtlsyWHR5sp2a/X/pxZnHDymTl2MrYmkalAntdqv/AtQ4aByg/tJbQre qYxv5jgJ8KjbL16bDx3qXxRR539TQNPGHFNUNk9u0EQ7C7zoRTV/Valqzzh7DwlHJ2Gn +3S17v8whkJPRkQ1/InYNs2LScmJxKUwKpYUktzat0vWBYrP+6Z5iXROcsQLprOWyl1U PjYdOEgDoCHroDvgCzpo12wPNVPUU33lQMLZvm4lbfIVKxuZ6SD708aO0iopQYFHxbkL D7zQ== X-Gm-Message-State: AOJu0YzOlLMuyvqY7nuqWQrn72/a2TMORVQdmqlVyLqBucCJBBlj3FBc BFgzYLNCxCjEOycVV/AayYq8MCm1dc3XehXRsCDPODO+nnzzo9+nGhIsJQCROmDqhi+T2tSm9en WONG28kTF8VU6g23f+5J9EMsm+I436fo17G1vEp08M3exHnhESZ8= X-Google-Smtp-Source: AGHT+IGKJIrUCxW0v8HE/TVPjGyOanDdu9icgEmfcZssZKwyfGl7tUR/m52DzwLypqTcvq5IVfsJqckLj9HnlzB9ydk= X-Received: by 2002:a05:6808:23c7:b0:3da:a032:24bc with SMTP id 5614622812f47-3dc3b44d817mr4022163b6e.33.1723153716394; Thu, 08 Aug 2024 14:48:36 -0700 (PDT) MIME-Version: 1.0 From: Patrick Robb Date: Thu, 8 Aug 2024 17:48:22 -0400 Message-ID: Subject: Community CI Meeting Minutes - August 8, 2024 To: ci@dpdk.org Cc: dev , dts@dpdk.org Content-Type: text/plain; charset="UTF-8" Content-Transfer-Encoding: quoted-printable X-BeenThere: ci@dpdk.org X-Mailman-Version: 2.1.29 Precedence: list List-Id: DPDK CI discussions List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Errors-To: ci-bounces@dpdk.org ##################################################################### August 8, 2024 Attendees 1. Patrick Robb 2. Juraj Linke=C5=A1 3. Aaron Conole 4. Alex Chapman 5. Dean Marx 6. Jeremy Spewock 7. Luca Vizzarro 8. Nicholas Pratte 9. Paul Szczepanek 10. Manit Mahajan ##################################################################### Minutes =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D General Announcements * Testing presentations at DPDK Summit: * UNH Lab Update and CI tooling updates talk was approved. Patrick and Adam Hassick will give this talk * Luca=E2=80=99s proposal for a DTS talk was approved, but the remote presentation option was rejected. So, Patrick and (Nick or Dean) will give this presentation. * AWS CI Testing initiative status? * What testing tooling/frameworks can AWS people use, for CI testing and reporting to the mailing list? * 1. They mentioned they have some internal DPDK test tools * 2. dpdk-test/meson test from the DPDK repo * 3. DTS (would be most appropriate in a couple months once more testsuites are merged to DPDK and we can figure out running the framework in a cloud environment) * What utilities do they need in order to setup CI? * Each lab has a different automation framework setup (UNH uses Jenkins, for instance) but the dpdk-ci repo (https://git.dpdk.org/tools/dpdk-ci/) contains bash/python scripts which can be used to stand up the CI process regardless of the automation tool in use: * downloading patches from DPDK patchwork (download-patch.sh) * UNH also wrote a python equivalent for downloading patchseries, which we can publish to dpdk-ci if there is interest. * Making a guess of what the correct DPDK branch to apply on is, based on files modified (https://git.dpdk.org/tools/dpdk-ci/tree/tools/pw_maintainers_cli.py) * (if needed) Apply patches to DPDK with git-pw and create a tarball of the resulting DPDK artifact to test on (https://git.dpdk.org/tools/dpdk-ci/tree/tools/create_series_artifact.py) * Send test report to the DPDK test-report mailing list (https://git.dpdk.org/tools/dpdk-ci/tree/tools/send-patch-report.sh) * And more utilities, which may be needed depending on the specific infra/process you decide to stand up * There are some people wanting to visit the UNH lab. Need to resurrect the mailing list conversation. * Most people will probably try to leave the morning of the 26th and drive to UNH (5 hours), and can arrive early afternoon at UNH =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D CI Status --------------------------------------------------------------------- UNH-IOL Community Lab * Windows Server testbeds: Tyler from Microsoft let us know that we can discontinue usage of the 2019 server. So, we will only run from server 2022 now, we just need to verify we are running all 3 compile toolchains on server 2022. As a reminder, they are: * LLVM 14.0.0 (or later) and Microsoft MSVC linker. * Microsoft Visual Studio 2022 (any edition). * The MinGW-w64 10.0 (or later) toolchain (native). * One more note about Windows compile jobs is that the prebuilt mingw64 binaries are not longer available at Sourceforge. It is necessary to download them at https://winlibs.com/ * Depends-on support on Patchwork: submitted a v2 for the series * https://patchwork.ozlabs.org/project/patchwork/list/?submitter=3D88738 * Pending reporting: * This is working, but we have gotten complaints about spamming the mailing list with too many reports. The reason is that we send a new report every time a testing job finishes for a particular environment (so we add a new row to the results table for a patchwork context). The community has requested that we change to a model of =E2=80=9Csend a report when testing for a patchwork context begins, then as tests finish, only send a new report if the overall test status for the context has changed, then send a report when all testing for the patchwork context is complete.=E2=80=9D We are making these updates now. * Testbed maintenance: Updating baremetal systems, firmware, VMs, containers definitions via dockerfiles, distro upgrades. We do this at the start of every release cycle. * DPDK Dashboard: Have made some improvements to how we query for test results, and started caching some requests we make to our results REST API, which have improved performance quite a bit. Still not lightning fast, but will no longer get timeout errors trying to load the results detail view for a patchseries etc. --------------------------------------------------------------------- Intel Lab * Juraj=E2=80=99s DTS API doc generation patch failed Intel=E2=80=99s CI, d= ue to =E2=80=9C-- exclude=3Ddoc/* *=E2=80=9D build parameter being used. They have stated thi= s is done to provide CI stability, and suggested in this case maintainers make an exception and merge the patch without a pass from Intel lab. Conversation is ongoing on the mailing list here, but Juraj/Patrick trying to figure out why this option is needed: https://inbox.dpdk.org/ci/CAJvnSUB_Ff_vw5p_VRoVstSrkXts8O9bWEuOSXvmTnMyBj27= sQ@mail.gmail.com/T/#t * This issue is no longer relevant (in terms of Juraj=E2=80=99s patch) a= s he has moved .rst sources to a different location and the series is now patching * We should still get a resolution regarding their build process, how changes can be authorized etc. * They appear to have an apply process in which they skip a patch if the patch only contains changes within the /doc directory. If a patch contains change in /doc, and also elsewhere, it is applied. --------------------------------------------------------------------- Github Actions * Aaron does intend to adjust his depends-on code once that support is added to Patchwork and DPDK patchwork upgrades to the latest version. --------------------------------------------------------------------- Loongarch Lab * None, Patrick Robb still needs to work with Min Zhou on retests =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D DTS Improvements & Test Development * DTS 24.11 roadmap: https://docs.google.com/document/d/1Rcp1-gZWzGGCCSkbEsigrd0-NoQmknv6ZS7V2CP= dgFo/edit * First 4 items will be taken by Juraj And Tomas * Filtering noise packets within testpmd verbose output * Need to identify packets via verbose output based on some unique ident= ifier * Option 1: * Could possibly use the SRC mac address * Can also check the packet layers, to see whether they match, adding a 2nd check beyond the SRC mac address check * Option 2: * Can use checksums * Will DPDK always report packets with a bad checksum? * NOTE: We have noticed that there is a bug for mlx5, which does not differentiate between a bad checksum and no checksum. Bugzilla ticket is here: https://bugs.dpdk.org/show_bug.cgi?id=3D1514 * Option 3: * Add support to testpmd code to show the port number * This appears to be the best candidate which will give us a high confidence in its stability, and allow for the greatest test coverage going forward * Test cases for hardware reported values, like Speed capabilities, RSS Key Update, etc. * Requires a config file for the testsuite registering the expected output to be returned by testpmd * Release notes: * The release notes updates should come in with the patchseries which add a particular feature (/doc/release_notes/24.11_release_notes.rst or something similar) * Would be good to add a list of bulleted testsuites which are added once per cycle * API Docs Generation: * Juraj has majorly refactored the suite for v16. So, he is needing new reviews. * Thomas has provided some comments, which Juraj has now addressed * There may be minor changes ahead, but it=E2=80=99s essentially done (s= o, ready for reviews again) * We no longer require that DTS dependencies are satisfied in order to generate documentation. There is a sphinx option which allows you to flag dependencies as not required to build docs. * This is so there is less hassle to building documentation * Where do we put the source .rst files? It was previously in the dts folder. Now moved to where DPDK API docs files are located (/doc/api). Now all the docs files are located in one place. * Ethertype testsuite RFC was posted, but the logic behind the testcases needs to be validated (original testplan is not valid), so we should provide an explanation of any new testcases on the mailing list =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D Any other business * Next Meeting Aug 22, 2024