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 CFFEA43E82; Thu, 25 Apr 2024 05:46:16 +0200 (CEST) Received: from mails.dpdk.org (localhost [127.0.0.1]) by mails.dpdk.org (Postfix) with ESMTP id CFA2E435FE; Thu, 25 Apr 2024 05:46:15 +0200 (CEST) Received: from mail-oo1-f45.google.com (mail-oo1-f45.google.com [209.85.161.45]) by mails.dpdk.org (Postfix) with ESMTP id B4AB9435F4 for ; Thu, 25 Apr 2024 05:46:13 +0200 (CEST) Received: by mail-oo1-f45.google.com with SMTP id 006d021491bc7-5ad2da2196aso373032eaf.2 for ; Wed, 24 Apr 2024 20:46:13 -0700 (PDT) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=iol.unh.edu; s=unh-iol; t=1714016773; x=1714621573; darn=dpdk.org; h=cc:to:subject:message-id:date:from:mime-version:from:to:cc:subject :date:message-id:reply-to; bh=7YD2dPt0biLBVgI606Iaemxv3GdJuMfmFcJD2lf0VB4=; b=SHjJNNsFVUPjjApHLI2Aic2T4RJ3OXgXCDHPF/IzAJ1WStR/CYDJwzTdlqS4TMpsLQ HLWMA6TI7gvOQeXJAbQS0TMbI2NBSAldhN5L3JcRD0331tv9RsF+fbg7PiirLHY3MIQF eVQZI/r4YXh1obrVVtx5auHYQTWOxp7iYyW6Y= X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20230601; t=1714016773; x=1714621573; h=cc:to:subject:message-id:date:from:mime-version:x-gm-message-state :from:to:cc:subject:date:message-id:reply-to; bh=7YD2dPt0biLBVgI606Iaemxv3GdJuMfmFcJD2lf0VB4=; b=VvUT84rH7u/jdytDFT+CiboEz6TN9Pa2LXd4nWvt4e1byU/Amrt58sq9Xmk6qKzeI/ 3AYNJhuyTwUgYRw8fYEZdEkyUdKsB4J0rwRliujkrvE7JKno/5bqpfK+EVp58Ky6G/kv bn9daJr/hkohiOF1MlBW0rNNSRhuOXB10KiGIkJMhpbuKvtMOL+btW5f6w7S99e3jvSu NT48TOBk/lcC4nqYzHS7AfGjvcB1KQ6ciS+BJ5mC4bx9DVLGITXdyVRpsqLKPLBjtEZY eHj1EckvjF3nSnA9SD8N+8u3jQfZEHI0JpX9VvigRO4suWoYQENKlXphQpDC4E1sSSNp fe9Q== X-Gm-Message-State: AOJu0YyWuI2ZHnidrNp9d23gF7fv/v1HgNnhGGp3d3UECRzMM9Msai7q n+bV7ggIlryRW/V1BHSHmBoKEXYeYAC457thw7swqR2x+dcTh3DsLuApbMBFhO7mvYDgLBHNXCy 4PtU8TYdEJxH6Oqyf/wrvxXF74jS6DBeDvx5aQnW4zghJurDSeqg= X-Google-Smtp-Source: AGHT+IGHuCb8Q7inWW6DQkcgOMk303087qDeW3zE5G18JnbPK8mlhmHfWwr/nYS/2vS6NT/wvA9qd0EbA7znrj8qeJM= X-Received: by 2002:a4a:5103:0:b0:5ac:5c89:2d08 with SMTP id s3-20020a4a5103000000b005ac5c892d08mr4562471ooa.1.1714016772621; Wed, 24 Apr 2024 20:46:12 -0700 (PDT) MIME-Version: 1.0 From: Patrick Robb Date: Wed, 24 Apr 2024 23:46:01 -0400 Message-ID: Subject: DTS WG Meeting Minutes - April 24, 2024 To: dev Cc: ci@dpdk.org, dts@dpdk.org Content-Type: multipart/alternative; boundary="000000000000df24650616e3a0ae" 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 --000000000000df24650616e3a0ae Content-Type: text/plain; charset="UTF-8" Content-Transfer-Encoding: quoted-printable ##################################################################### Attendees * Patrick Robb * Juraj Linke=C5=A1 * Paul Szczepanek * Luca Vizzarro * Nicholas Pratte ##################################################################### 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 * What does the path to project wide adoption look like for DTS? * Tech board laid out at the end of 2023 what the process should look like for building DTS adoption in DPDK * 1. Current DTS developers start writing testsuites for dpdk library X, demonstrating the framework (easily and effectively) supports testsuites for the library. If any gaps in the framework are discovered, submit the needed DTS framework patches for writing testsuites for dpdk library X. * 1.1 As this process gets under way, it=E2=80=99s important to st= art working with library maintainers to validate the new testsuites, and start to build the loose framework for what broad adoption will look like. * 1.2 To this end, UNH developers are supposed to write ethdev testsuites in 2024. The current timeline is that techboard will assess ethdev testsuite support in DTS at the end of 2024. * Paul notes that Luca too will be writing some ethdev suites in the future * Group should determine the =E2=80=9Cshortlist=E2=80=9D of eth= dev testsuite * Patrick Robbto the full list down for the next bi-weekly dts meeting * Always make a bugzilla ticket per testsuite before doing any work * please make a ticket for jumboframes: https://bugs.dpdk.org/show_bug.cgi?id=3D1421 * 2. When DTS developers indicate DTS support for library X is ready, tech board will take it up as a discussion point and give that an up/down vote. And =E2=80=9Cup=E2=80=9D vote will set the new policy that when DPDK = developers submit a DPDK patch to library X, they are now required to also submit an associated testsuite/testcase addition which provides some test coverage for their new feature. * 2.1 DPDK CI labs will =E2=80=9Cpick up=E2=80=9D the new testsuit= e additions alongside the new patches and run across their hardware. * Jeremy has been on another project for a few weeks, but that=E2=80=99s wr= apped up now and he=E2=80=99s back to focusing on DTS =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D Patch discussions * Hugepages patch: * V4 is submitted with some small improvements recommended by Juraj: https://patchwork.dpdk.org/project/dpdk/patch/20240418161026.2839-1-npratte= @iol.unh.edu/ * Juraj will send a review tag * Jumboframes: Nick is writing this now, V1 should be coming soon * Bugzilla: https://bugs.dpdk.org/show_bug.cgi?id=3D1421 * Testpmd show port info/show port stats: https://patchwork.dpdk.org/project/dpdk/list/?series=3D31729 * Testpmd statefullness / params class * V2 will be forthcoming here * Jeremy can give a review on the V2 now that=E2=80=99s he=E2=80=99s bac= k to DTS world=E2=80=A6 * Jeremy should assess for his patch whether show port info should be used as a tool to check capabilities * Skip test cases based on capabilities * Jeremy still needs to provide feedback from his new scatter testsuite rebased off of this patch * Replace XML-RPC server with scapy shell: * Coming soon to a mailing list near you. * API-Docs generation patch * Review are needed for the latest version * Jeremy Spewockplease review * Nicholas Pratteplease build the docs again and provide another tested-by * Renaming =E2=80=9CExecution=E2=80=9D * Idea is to rename to testrun * As it is a tiny patch it can be applied soon * Juraj=E2=80=99s patch to remove OS-UDP suite * We want to remove the code which was added only to support the os-udp suite. This includes the code which allows for network configuration in the os. But, we may need some of this for scapy on the TG. * Bugzilla: https://bugs.dpdk.org/show_bug.cgi?id=3D1414 * Juraj submitted some small patches with cleanups like changing implicit object inheritance=E2=80=A6 etc. * Interactive shells: * By default, the interactive shells are not run with admin privileges. For shells handling any DPDK apps, it may make sense to run with admin privilege. =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D Bugzilla discussions * Had some discussion on 1360 relating to how users should define node ports for an execution: https://bugs.dpdk.org/show_bug.cgi?id=3D1360 * It=E2=80=99s not really execution specific, as the execution has thing= s which we are configuring for a specific testrun. There will never be a different value for a specific port on a node, so this information should not go into the execution config block. * Individual nodes do not need to have the full port links stored as attributes. The port links can be created by the testsuite alone. * So, going forward the SUTNode and TGNode should contain ONLY local pci addresses, and these nodes are not aware of the full links * Store a list of =E2=80=9CNICs=E2=80=9D in the node config block, then = specify the specific TG NIC and SUT NIC to use in the execution config for that given execution * New framework assumption: If I use NIC A and NIC B in a testsuite, the first port defined for NIC A is assumed to be linked to the first port defined in NIC B, 2nd to 2nd, 3rd to 3rd, etc. * Patrick Robbwrite example configuration on this bugzilla ticket to represent the above idea =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=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 * How to document test plans in new DTS? Can this always go in the testsuite class docstring? What if the writeup is very long? --000000000000df24650616e3a0ae Content-Type: text/html; charset="UTF-8" Content-Transfer-Encoding: quoted-printable
##########################################################= ###########
Attendees
* Patrick Robb
* Juraj Linke=C5=A1
* Paul= Szczepanek
* Luca Vizzarro
* Nicholas Pratte

################= #####################################################
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 Announcem= ents
* What does the path to project wide adoption look like for DTS? =C2=A0 =C2=A0* Tech board laid out at the end of 2023 what the process sh= ould look like for building DTS adoption in DPDK
=C2=A0 =C2=A0 =C2=A0 * = 1. Current DTS developers start writing testsuites for dpdk library X, demo= nstrating the framework (easily and effectively) supports testsuites for th= e library. If any gaps in the framework are discovered, submit the needed D= TS framework patches for writing testsuites for dpdk library X.
=C2=A0 = =C2=A0 =C2=A0 =C2=A0 =C2=A0* 1.1 As this process gets under way, it=E2=80= =99s important to start working with library maintainers to validate the ne= w testsuites, and start to build the loose framework for what broad adoptio= n will look like.
=C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0* 1.2 To this end, U= NH developers are supposed to write ethdev testsuites in 2024. The current = timeline is that techboard will assess ethdev testsuite support in DTS at t= he end of 2024.
=C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 * Paul notes t= hat Luca too will be writing some ethdev suites in the future
=C2=A0 =C2= =A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 * Group should determine the =E2=80=9Cshort= list=E2=80=9D of ethdev testsuite
=C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2= =A0 =C2=A0 =C2=A0* Patrick Robbto the full list down for the next bi-weekly= dts meeting
=C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 * Always make a b= ugzilla ticket per testsuite before doing any work
=C2=A0 =C2=A0 =C2=A0 = =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0* please make a ticket for jumboframes: <= a href=3D"https://bugs.dpdk.org/show_bug.cgi?id=3D1421">https://bugs.dpdk.o= rg/show_bug.cgi?id=3D1421
=C2=A0 =C2=A0 =C2=A0 * 2. When DTS develop= ers indicate DTS support for library X is ready, tech board will take it up= as a discussion point and give that an up/down vote. And =E2=80=9Cup=E2=80= =9D vote will set the new policy that when DPDK developers submit a DPDK pa= tch to library X, they are now required to also submit an associated testsu= ite/testcase addition which provides some test coverage for their new featu= re.
=C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0* 2.1 DPDK CI labs will =E2=80=9C= pick up=E2=80=9D the new testsuite additions alongside the new patches and = run across their hardware.
* Jeremy has been on another project for a fe= w weeks, but that=E2=80=99s wrapped up now and he=E2=80=99s back to focusin= g on DTS

=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
Patch discussions
* Hugepages patch:
=C2=A0 =C2=A0* V4 is submitt= ed with some small improvements recommended by Juraj: https://patchwork.dpdk.org/project/dpdk/patch/20240418161026.2839-1-np= ratte@iol.unh.edu/
=C2=A0 =C2=A0* Juraj will send a review tag
* = Jumboframes: Nick is writing this now, V1 should be coming soon
=C2=A0 = =C2=A0* Bugzilla: = https://bugs.dpdk.org/show_bug.cgi?id=3D1421
* Testpmd show port inf= o/show port stats: https://patchwork.dpdk.org/project/dpdk/list/?series=3D317= 29
* Testpmd statefullness / params class =C2=A0 =C2=A0 =C2=A0 =C2= =A0
=C2=A0 =C2=A0* V2 will be forthcoming here
=C2=A0 =C2=A0* Jeremy = can give a review on the V2 now that=E2=80=99s he=E2=80=99s back to DTS wor= ld=E2=80=A6
=C2=A0 =C2=A0* Jeremy should assess for his patch whether sh= ow port info should be used as a tool to check capabilities
* Skip test = cases based on capabilities
=C2=A0 =C2=A0* Jeremy still needs to provide= feedback from his new scatter testsuite rebased off of this patch
* Rep= lace XML-RPC server with scapy shell:
=C2=A0 =C2=A0* Coming soon to a ma= iling list near you.
* API-Docs generation patch
=C2=A0 =C2=A0* Revie= w are needed for the latest version
=C2=A0 =C2=A0* Jeremy Spewockplease = review
=C2=A0 =C2=A0* Nicholas Pratteplease build the docs again and pro= vide another tested-by
* Renaming =E2=80=9CExecution=E2=80=9D
=C2=A0= =C2=A0* Idea is to rename to testrun
=C2=A0 =C2=A0* As it is a tiny pat= ch it can be applied soon
* Juraj=E2=80=99s patch to remove OS-UDP suite=
=C2=A0 =C2=A0* We want to remove the code which was added only to suppo= rt the os-udp suite. This includes the code which allows for network config= uration in the os. But, we may need some of this for scapy on the TG.
= =C2=A0 =C2=A0* Bugzilla: https://bugs.dpdk.org/show_bug.cgi?id=3D1414
* Juraj submit= ted some small patches with cleanups like changing implicit object inherita= nce=E2=80=A6 etc.
* Interactive shells:
=C2=A0 =C2=A0* By default, th= e interactive shells are not run with admin privileges. For shells handling= any DPDK apps, it may make sense to run with admin privilege.
=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
Bu= gzilla discussions
* Had some discussion on 1360 relating to how users s= hould define node ports for an execution: https://bugs.dpdk.org/show_bug.cgi?id=3D1360=C2=A0 =C2=A0* It=E2=80=99s not really execution specific, as the executio= n has things which we are configuring for a specific testrun. There will ne= ver be a different value for a specific port on a node, so this information= should not go into the execution config block.
=C2=A0 =C2=A0* Individua= l nodes do not need to have the full port links stored as attributes. The p= ort links can be created by the testsuite alone.
=C2=A0 =C2=A0 =C2=A0 * = So, going forward the SUTNode and TGNode should contain ONLY local pci addr= esses, and these nodes are not aware of the full links
=C2=A0 =C2=A0* St= ore a list of =E2=80=9CNICs=E2=80=9D in the node config block, then specify= the specific TG NIC and SUT NIC to use in the execution config for that gi= ven execution
=C2=A0 =C2=A0 =C2=A0 * New framework assumption: If I use = NIC A and NIC B in a testsuite, the first port defined for NIC A is assumed= to be linked to the first port defined in NIC B, 2nd to 2nd, 3rd to 3rd, e= tc.
=C2=A0 =C2=A0 =C2=A0 * Patrick Robbwrite example configuration on th= is bugzilla ticket to represent the above idea

=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=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
* How to docu= ment test plans in new DTS? Can this always go in the testsuite class docst= ring? What if the writeup is very long?
--000000000000df24650616e3a0ae--