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 077244645A; Sun, 23 Mar 2025 18:16:34 +0100 (CET) Received: from mails.dpdk.org (localhost [127.0.0.1]) by mails.dpdk.org (Postfix) with ESMTP id DCE3C40277; Sun, 23 Mar 2025 18:16:33 +0100 (CET) Received: from mail-pj1-f45.google.com (mail-pj1-f45.google.com [209.85.216.45]) by mails.dpdk.org (Postfix) with ESMTP id DEA5040150 for ; Sun, 23 Mar 2025 18:16:31 +0100 (CET) Received: by mail-pj1-f45.google.com with SMTP id 98e67ed59e1d1-30332dfc821so1010220a91.3 for ; Sun, 23 Mar 2025 10:16:31 -0700 (PDT) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=iol.unh.edu; s=unh-iol; t=1742750191; x=1743354991; darn=dpdk.org; h=cc:to:subject:message-id:date:from:mime-version:from:to:cc:subject :date:message-id:reply-to; bh=opCZYwGGdgfqshZLjp4uBog8cIN/pECe2B1Xd1FLy4k=; b=Y0PWsnylBsm6wVzN69upUshx0LfldrR770QVNE/7yaHUJGK5MmpKiKPSgwZdDO6rsQ R1lnLT4H/M2DevyKrbie+QjtJB2/oe/7T3nUrfDpcdYQUiKalCEt7YVAvxLUT5K48Xbb gxZf9liXODwuoa24UeyQKOHrhcQTVRqzV2w9I= X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20230601; t=1742750191; x=1743354991; h=cc:to:subject:message-id:date:from:mime-version:x-gm-message-state :from:to:cc:subject:date:message-id:reply-to; bh=opCZYwGGdgfqshZLjp4uBog8cIN/pECe2B1Xd1FLy4k=; b=ktqU4k7BYHwW07+yZgQDjJgmjfDJX4xnLrhHVMlxEGE0uC4ca7v869549YcTFYjlEx mGbnLhqJQeKLHRxbQcdhQNdBeMJlZg4hb4Mm5bu0KrFGLYfjZ6U7eNrtUanJcixzi0kf NddlRBDyeWoyRqM70eXP5dfKFUYAbksKxmWKB15wIOyywF0kT3FWQ5UVqCj6KVXFMxxE InE2Dwy3WnI4RwnoGToFgBg5n0Uj4K/UXF45+90Syiwf5on8kovXOkWoFFuNNmWANwr0 Blr5WXvKMvVYbDxN5KP77g0Ap8smbHuwAhKXdZgx0Iqxej28eqI+dfjwoQg2b1Pr+F2s oQfg== X-Gm-Message-State: AOJu0YzHDMmkB/yK6DHV/0N9kTVeOTlar6M1VuUf3BTQZLCFV24XYEao SlQLA6mnJu/i2qKWk/WTUQ6IR7Y4vvR7EnGea1zqbiHpWcTL02Cx76ZMLJ019Xh5UHFmoKWp9Y2 +KqEnu1NnpPybsuUPkkud9kbsUXyybQj51aZ6mN/xQm28R7sXnEV3vg== X-Gm-Gg: ASbGncs3T1U4ggETs83xnQeaLshIcEjLoN9aDuN50mH4bkAhrCltOmRKyn9NthCFjpV QUhJKH6pf/mZLN5MMlDUqM2ntniVtjq9jzhpdRhhniL9RRRigrTeaUqQT7JbfuXQ92llfkb06vp lNr5h9dZ3RkrLRN/XChbd2ZbT1 X-Google-Smtp-Source: AGHT+IHWmuKTSR63EqNL5VxObqRgQkY9XhUFR7g1gg0zx6yL6cFeeQ/16JpSy8aY/iV6K8BKIHDX2vbdMC+vaWEf948= X-Received: by 2002:a17:90b:3906:b0:2ee:aa28:79aa with SMTP id 98e67ed59e1d1-3030fe8be5dmr13253785a91.6.1742750190107; Sun, 23 Mar 2025 10:16:30 -0700 (PDT) MIME-Version: 1.0 From: Patrick Robb Date: Sun, 23 Mar 2025 13:12:38 -0400 X-Gm-Features: AQ5f1Jo2Fwd1eMwYCtZtr3vbDGaVXtLjRCLF83YCnZEpBguX0CX8Ml7M2z1AqP8 Message-ID: Subject: Community CI Meeting Minutes - March 20, 2025 To: ci@dpdk.org Cc: dev Content-Type: multipart/alternative; boundary="000000000000039ced063105a60c" X-BeenThere: dev@dpdk.org X-Mailman-Version: 2.1.29 Precedence: list List-Id: DPDK patches and discussions List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Errors-To: dev-bounces@dpdk.org --000000000000039ced063105a60c Content-Type: text/plain; charset="UTF-8" Content-Transfer-Encoding: quoted-printable ##################################################################### March 20, 2025 Attendees 1. Patrick Robb 2. Paul Szczepanek 3. Luca Vizzarro 4. Cody Cheng 5. Aaron Conole 6. Matthew McGovern 7. Dean Marx ##################################################################### 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 * RC3 has been released * Next-dts has been pulled to main * DPDK 25.03 will be released in a few days * Pw-ci project: Aaron is syncing his internal repo to the upstream * Adds recheck support, and other reporting scripts =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 * New DTS: * Running a broader set of =E2=80=9Cnew DTS=E2=80=9D tests on Intel XL71= 0 and CX5 NICs now: * Unified Packet * Promiscuous Support * PMD Buffer Scatter * Dynamic Queue Config * Dynamic Config * MAC Filter * Cody has begun work on setting up a VM which will be locked to the minimum kernel version supported by DPDK - this system will do build tests. * According to DPDK docs, that is kernel version 4.19, but it links to a kernel mirror site which indicates a newer minimum supported kernel - Cody is going to reach out on the mailing list for clarification * The patch adding patch series dependency support to DPDK Patchwork has been accepted: https://patchwork.ozlabs.org/project/patchwork/list/?series=3D442332&state= =3D* * We will need to work with Ali after the next patchwork release to get the DPDK Patchwork instance updated to the latest release * Build failure emails: * Build failure emails were not including the Meson and Ninja logs due to a bad HTTP server reconfiguration we had deployed. This issue is now resolved and we see the most recent build failure comes with the relevant logs. --------------------------------------------------------------------- Intel Lab * None --------------------------------------------------------------------- Github Actions * Cirrus CI: No one is working on the FreeBSD support, so unit tests still fail, and this decreases the value added by Cirrus CI * Working on an update which will associate a run number with a recheck request --------------------------------------------------------------------- AWS * They are working on their test-report email templating script * There was some conversation at the joint Governing Board and Tech Board meeting about platforms AWS will run from and their testing goals: * Platforms: includes x86 and ARM Graviton based systems * Goals: Will setup a CI lab which will target build and unit tests first, then target some functional and performance testing with DTS. --------------------------------------------------------------------- Loongarch Lab * None =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 Packet Sniffer update: * Breaks the sniffer out into a new class which is run separately from Scapy TG class * It can start/stop and collect packets whenever directed to do so via the TestSuite API * Async sniffer will run throughout the entire testrun instead of starting and stopping. * Adds a facility to trigger a stop condition for the sniffer (like stop as soon as you sniff packet X) * This is not exposed via the TestSuite API currently * Currently Sniffer is integrated into the scapy class, but there is no particular reason for this - it could reasonably be moved to the TestSuite and exposed * DTS Roadmap 25.07 * Rte_flow * Performance TG and forwarding tests * Improved documentation * TestSuite API * Paul is interested in moving from the TestSuite class as API to a =E2=80=9CDPDK=E2=80=9D API for DTS testsuites. This would involved committi= ng to a stable API. We would be able to change framework internals without breaking testsuites. * The biggest advantage would be discoverability for a new test writer who has not seen DTS before * Cloud environment support for DTS * VF Support * VF support: * No movement on the VF support series * TestSuite Development: * Dean has submitted updates for unified packet testsuite and checksum offload testsuite which switches from matching packets on mac dst address to relying on l4 port number. * Testrun duration: * Perhaps we should profile the testsuites into different buckets and add config options for running certain groups of tests * By keeping the sniffer process on the entire time, we will cut out a lot of time waiting for sniffer to start and stop * One option would be to keep testpmd open through the entire testrun, and just reset it for each testsuite without restarting it and going through EAL bring up (takes a few seconds). On the other hand, the appeal of restarting testpmd completely is we know it is totally reset. * Possible keep 1 testpmd shell per testsuite, and that testsuite is responsible for cleaning up the shell properly in between testcases =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 * Azure: We got more details from Matthew about the DPDK testing goals for Microsoft Azure * Environments: They can run testing on cloud instances based on Intel x86, AMD x86, and ARM systems * Mellanox NICs are available to some cloud instances on Azure * They would like to run functional and performance tests and products reports for those for dpdk.org * Performance tests: will be simple single and multi-queue testpmd MPPS forwarding tests * They cannot commit to a report cadence right now, but discussed the idea of releasing a report per DPDK release, as well as rolling reports on main * Next Meeting Apr 3, 2025 --000000000000039ced063105a60c Content-Type: text/html; charset="UTF-8" Content-Transfer-Encoding: quoted-printable
##########################################################= ###########
March 20, 2025
Attendees
1. Patrick Robb
2. Paul Sz= czepanek
3. Luca Vizzarro
4. Cody Cheng
5. Aaron Conole
6. Matt= hew McGovern
7. Dean Marx

#######################################= ##############################
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
* RC3 has bee= n released
=C2=A0 =C2=A0* Next-dts has been pulled to main
=C2=A0 =C2= =A0* DPDK 25.03 will be released in a few days
* Pw-ci project: Aaron is= syncing his internal repo to the upstream
=C2=A0 =C2=A0* Adds recheck s= upport, and other reporting scripts
=C2=A0 =C2=A0
=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
* New DTS:
=C2=A0 =C2=A0* Running a broader set of =E2=80=9Cnew = DTS=E2=80=9D tests on Intel XL710 and CX5 NICs now:
=C2=A0 =C2=A0 =C2=A0= * Unified Packet
=C2=A0 =C2=A0 =C2=A0 * Promiscuous Support
=C2=A0 = =C2=A0 =C2=A0 * PMD Buffer Scatter
=C2=A0 =C2=A0 =C2=A0 * Dynamic Queue = Config
=C2=A0 =C2=A0 =C2=A0 * Dynamic Config
=C2=A0 =C2=A0 =C2=A0 * M= AC Filter
* Cody has begun work on setting up a VM which will be locked = to the minimum kernel version supported by DPDK - this system will do build= tests.
=C2=A0 =C2=A0* According to DPDK docs, that is kernel version 4.= 19, but it links to a kernel mirror site which indicates a newer minimum su= pported kernel - Cody is going to reach out on the mailing list for clarifi= cation
* The patch adding patch series dependency support to DPDK Patchw= ork has been accepted: https://patchwork.ozlabs.org/pro= ject/patchwork/list/?series=3D442332&state=3D*
=C2=A0 =C2=A0* We= will need to work with Ali after the next patchwork release to get the DPD= K Patchwork instance updated to the latest release
* Build failure email= s:
=C2=A0 =C2=A0* Build failure emails were not including the Meson and = Ninja logs due to a bad HTTP server reconfiguration we had deployed. This i= ssue is now resolved and we see the most recent build failure comes with th= e relevant logs.
=C2=A0 =C2=A0
=C2=A0 =C2=A0-------------------------= --------------------------------------------
Intel Lab
* None

= ---------------------------------------------------------------------
Gi= thub Actions
* Cirrus CI: No one is working on the FreeBSD support, so u= nit tests still fail, and this decreases the value added by Cirrus CI
* = Working on an update which will associate a run number with a recheck reque= st

-----------------------------------------------------------------= ----
AWS
* They are working on their test-report email templating scr= ipt
* There was some conversation at the joint Governing Board and Tech = Board meeting about platforms AWS will run from and their testing goals:=C2=A0 =C2=A0* Platforms: includes x86 and ARM Graviton based systems
= =C2=A0 =C2=A0* Goals: Will setup a CI lab which will target build and unit = tests first, then target some functional and performance testing with DTS.<= br>=C2=A0 =C2=A0
-------------------------------------------------------= --------------
Loongarch Lab
* None


=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 Develop= ment
* DTS Packet Sniffer update:
=C2=A0 =C2=A0* Breaks the sniffer o= ut into a new class which is run separately from Scapy TG class
=C2=A0 = =C2=A0* It can start/stop and collect packets whenever directed to do so vi= a the TestSuite API
=C2=A0 =C2=A0* Async sniffer will run throughout the= entire testrun instead of starting and stopping.
=C2=A0 =C2=A0* Adds a = facility to trigger a stop condition for the sniffer (like stop as soon as = you sniff packet X)
=C2=A0 =C2=A0 =C2=A0 * This is not exposed via the T= estSuite API currently
=C2=A0 =C2=A0* Currently Sniffer is integrated in= to the scapy class, but there is no particular reason for this - it could r= easonably be moved to the TestSuite and exposed
* DTS Roadmap 25.07
= =C2=A0 =C2=A0* Rte_flow
=C2=A0 =C2=A0* Performance TG and forwarding tes= ts
=C2=A0 =C2=A0* Improved documentation
=C2=A0 =C2=A0* TestSuite API=
=C2=A0 =C2=A0 =C2=A0 * Paul is interested in moving from the TestSuite = class as API to a =E2=80=9CDPDK=E2=80=9D API for DTS testsuites. This would= involved committing to a stable API. We would be able to change framework = internals without breaking testsuites.
=C2=A0 =C2=A0 =C2=A0 * The bigges= t advantage would be discoverability for a new test writer who has not seen= DTS before
=C2=A0 =C2=A0* Cloud environment support for DTS
=C2=A0 = =C2=A0* VF Support
* VF support:
=C2=A0 =C2=A0* No movement on the VF= support series
* TestSuite Development:
=C2=A0 =C2=A0* Dean has subm= itted updates for unified packet testsuite and checksum offload testsuite w= hich switches from matching packets on mac dst address to relying on l4 por= t number.
* Testrun duration:
=C2=A0 =C2=A0* Perhaps we should profil= e the testsuites into different buckets and add config options for running = certain groups of tests
=C2=A0 =C2=A0* By keeping the sniffer process on= the entire time, we will cut out a lot of time waiting for sniffer to star= t and stop
=C2=A0 =C2=A0* One option would be to keep testpmd open throu= gh the entire testrun, and just reset it for each testsuite without restart= ing it and going through EAL bring up (takes a few seconds). On the other h= and, the appeal of restarting testpmd completely is we know it is totally r= eset.
=C2=A0 =C2=A0 =C2=A0 * Possible keep 1 testpmd shell per testsuite= , and that testsuite is responsible for cleaning up the shell properly in b= etween testcases

=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
* Azure: We got more details from Matthew a= bout the DPDK testing goals for Microsoft Azure
=C2=A0 =C2=A0* Environme= nts: They can run testing on cloud instances based on Intel x86, AMD x86, a= nd ARM systems
=C2=A0 =C2=A0* Mellanox NICs are available to some cloud = instances on Azure
=C2=A0 =C2=A0* They would like to run functional and = performance tests and products reports for those for dpdk.org
=C2=A0 =C2=A0 =C2=A0 * Performance tests: will be sim= ple single and multi-queue testpmd MPPS forwarding tests
=C2=A0 =C2=A0* = They cannot commit to a report cadence right now, but discussed the idea of= releasing a report per DPDK release, as well as rolling reports on main* Next Meeting Apr 3, 2025
--000000000000039ced063105a60c--