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 CE35843BB1; Fri, 23 Feb 2024 22:02:08 +0100 (CET) Received: from mails.dpdk.org (localhost [127.0.0.1]) by mails.dpdk.org (Postfix) with ESMTP id C7FB14029A; Fri, 23 Feb 2024 22:02:08 +0100 (CET) Received: from mail-oo1-f41.google.com (mail-oo1-f41.google.com [209.85.161.41]) by mails.dpdk.org (Postfix) with ESMTP id 17C1C4027C for ; Fri, 23 Feb 2024 22:02:07 +0100 (CET) Received: by mail-oo1-f41.google.com with SMTP id 006d021491bc7-59fc2d22cfaso566588eaf.3 for ; Fri, 23 Feb 2024 13:02:07 -0800 (PST) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=iol.unh.edu; s=unh-iol; t=1708722126; x=1709326926; darn=dpdk.org; h=cc:to:subject:message-id:date:from:mime-version:from:to:cc:subject :date:message-id:reply-to; bh=ha3sCB6z+Zv4KtZkCMEwHYltqkk1U3IxSVGP3Tq+lxI=; b=I69HiOhrB5Pu1SxUuDdRE6Kb4WB7ajtckIKoW+ApN++bMcxgB1Smi3tbeq6GCKNKyb xcpjBu2DFA4/XwgXAwAxuGZ1R1Htlzcz8Z/QpF1WL6sRNu9j892O1legOPy1zHmWenjs z71xBAQicoJZ/O3Sh9BLDW9fHnziut1TzQsfw= X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20230601; t=1708722126; x=1709326926; h=cc:to:subject:message-id:date:from:mime-version:x-gm-message-state :from:to:cc:subject:date:message-id:reply-to; bh=ha3sCB6z+Zv4KtZkCMEwHYltqkk1U3IxSVGP3Tq+lxI=; b=VVPu7lge7mPm4lF0Fey6RFOSMP3XIO7JSBCC41Y2lNuVv+zkAY1Rrcko1VCcaeLP6c 6CqPGz7mZmaPS8yqnZrdvNkOuQAF/+tVlt+m3gwjkpLV8OjpOzOQ5+e6g2TOUKt/aRdG mK/wmrw20HzNcQRIVTxkPsl60IuVWZAOk93sG0GGW0+dsZUhwqbYcTJLuzB7uJ8KVDKp 7e9nV16qvUD1oWBPhXzul1PyS/PNR+u0MCsfzVaaQ9asIsQ/AweNSiTWe57FnEhzqybe peLM1NErMmU9Y5Qokq/RAQRtVYWnMsgdSctOb4afIltSQ2dOzDG4OhwH7DxZrO5xZfuf Lv2w== X-Gm-Message-State: AOJu0Yxz5YxA6g0+6nVEh3Lgst1K19iw/OPB0zif7VLv4MaYjcg1xb9C CwRxpPjG0W8cs4G4Nm8+5fGxVLoPcuznjCyCfBkASExi/PyjoS/4QsToGrshuQ8cIStZrDwmti+ 5OJzjhq0M2uywo+yTXFyOD3eVyrNSvAMoYjWpWLScbgU6yVaq+yk7Wg== X-Google-Smtp-Source: AGHT+IEGLh6FGowcEbpOD1/eb/obgQ0/TzJahl4k0VUy/TSaZ2Ef9HDFsCuttKNBcAv2WnKMrY9GiLnVWW45YqiT7cU= X-Received: by 2002:a4a:6f52:0:b0:5a0:525a:1389 with SMTP id i18-20020a4a6f52000000b005a0525a1389mr991934oof.3.1708722126044; Fri, 23 Feb 2024 13:02:06 -0800 (PST) MIME-Version: 1.0 From: Patrick Robb Date: Fri, 23 Feb 2024 16:01:55 -0500 Message-ID: Subject: Community CI Meeting Minutes - February 22, 2024 To: ci@dpdk.org Cc: dev@dpdk.org, dts@dpdk.org Content-Type: multipart/alternative; boundary="00000000000057d782061212dfa4" 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 --00000000000057d782061212dfa4 Content-Type: text/plain; charset="UTF-8" Content-Transfer-Encoding: quoted-printable February 22, 2024 ##################################################################### Attendees 1. Patrick Robb 2. Ali Alnubani 3. Paul Szczepanek 4. Juraj Linke=C5=A1 5. Luca Vizzarro ##################################################################### 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 * Aaron is polling the tech board for feedback on what server hardware is most needed in the community lab going forward. Some ideas are: * AMD CPUs * RISC-V CPUs * Better PCI-E generation slots which will allow us to use newer NICs * He is visiting UNH today so we can work on starting a proposal and pla= n * 24.03 RC1 has been released this morning * Retest framework: Email has been sent out with the proposed syntax and approach for retests in which users want to request their patch be re-applied on tip of branch =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=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 * Hardware Refresh: * NVIDIA CX7: * Had some minor improvements in the performance results, but still debugging with Ali and NVIDIA * Can possibly replaced the cx5 with another cx7, and do forwarding between 2 nics, solving 1 bandwidth bottleneck * Will work on resolving the lower speeds seen on the CX6 first - this could be related to using a different board and CPU with different clock rate etc. * This server has a broadwell CPU * Patrick Robb Make sure that Ali and Gal are included for the initial feedback thread for server refresh and what is most needed * Bringing the NVIDIA testbed offline today for a few hours so Ali and Bing can do some debugging on the mlx5 failure on the CX5 yesterday * Arm IPSEC-MB Library: Had to move to running from tip of main on the ARM ipsec repo to enablev1.4 - just storing the commit hash right now but ARM will publish a new tag for a known good version soon. * Wathsala will be doing the new tag soon --------------------------------------------------------------------- Intel Lab * None --------------------------------------------------------------------- Github Actions * We plan to have a maintenance window either Thus the 22nd, or sometime next week to cover migrating to the original server. This will involve upgrading the base os for both the host and the VM. Michael will send out the notice on the day it happens letting everyone know of the downtime. * We don't expect that the downtime will last too long, less than a day. We will likely recover the workflows shortly after that. --------------------------------------------------------------------- Loongarch Lab * Patrick pinged Zhoumin about adding retest framework support to the Loongson lab * UNH willing to assist - not sure right now what is possible/not possible based on how the loongson lab stood up their automation. They do use the dpdk-ci repo tools. =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=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 * Dockerfile patch can be merged - Thomas has been pinged about this on slack * Scattered packets patch: * Patrick tested this on a bnxt_en NIC, and it worked fine * When gathering device capabilities, the scatter capability is always off on mlnx nics, even when including the scatter offload flag with testpmd * Juraj is going to send to Patrick and Jermey a summary of what he has learned about passing the scatter flag and how DPDK handles it. And what he has learned about querying for this capability. * For now, not including the scatter offload flag testcase with this testsuite, only submitting the testcase which is a direct port from old dts * Luca is going to review this today. He is also running it on a MLNX nic. * Juraj: Interestingly, on the Intel NICs it is on whether you include the --enable-scatter flag or not, but the Mellanox NICs don't have it set to on in either case * Ali: Have you tried "--enable-scatter --tx-offloads=3D0x8000" * Jeremy Spewockwill try this * For next Wednesday, We are going to have to have discussions for putting together the 24.07 DTS roadmap since Juraj will be on vacation in March and we won=E2=80=99t be able to have the conversation then. * Patrick will put together some ethdev suite ideas which the new DTS employees at UNH can start on * We will also review bugzilla tickets then, assign more tickets if needed * Patch for the testcast blocking: * There is a bug (it does not include the smoke tests in the list of suites to run), so Juraj will be sending a new version * Gregory reached out to see whether his framework=E2=80=99s approach could= be used for simple DTS cases. Juraj and Patrick read the code. There are some good ideas we are bringing into the framework, but not the phase based yaml approach which translates scapy and testpmd commands to testsuites. * XMLRPC Server: Jeremy found that there is a python dependency for the TG server which was not known (must be python 3.10+). Or else you will have an issue with a XMLRPC server function not included until verson x. * Should turn this into a bugzilla ticket, and can discuss on Wednesday. * Scapy version on TG node. Need to make sure we are on x version. Should be 2.4.4 or 2.5.0? Any other business =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D * Next Meeting: March 7, 2024 --00000000000057d782061212dfa4 Content-Type: text/html; charset="UTF-8" Content-Transfer-Encoding: quoted-printable

February 22, 2024

#############################= ########################################
Attendees
1. Patrick Robb2. Ali Alnubani
3. Paul Szczepanek
4. Juraj Linke=C5=A1
5. Luca V= izzarro

############################################################= #########
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
* Aaron is polling the tech board = for feedback on what server hardware is most needed in the community lab go= ing forward. Some ideas are:
=C2=A0 =C2=A0* AMD CPUs
=C2=A0 =C2=A0* R= ISC-V CPUs
=C2=A0 =C2=A0* Better PCI-E generation slots which will allow= us to use newer NICs
=C2=A0 =C2=A0* He is visiting UNH today so we can = work on starting a proposal and plan
* 24.03 RC1 has been released this = morning
* Retest framework: Email has been sent out with the proposed sy= ntax and approach for retests in which users want to request their patch be= re-applied on tip of branch

=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=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
* Hardware Re= fresh:
=C2=A0 =C2=A0* NVIDIA CX7:
=C2=A0 =C2=A0 =C2=A0 * Had some mi= nor improvements in the performance results, but still debugging with Ali a= nd NVIDIA
=C2=A0 =C2=A0 =C2=A0 * Can possibly replaced the cx5 with anot= her cx7, and do forwarding between 2 nics, solving 1 bandwidth bottleneck=C2=A0 =C2=A0 =C2=A0 * Will work on resolving the lower speeds seen on th= e CX6 first - this could be related to using a different board and CPU with= different clock rate etc.
=C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0* This ser= ver has a broadwell CPU
=C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0* Patrick Robb= Make sure that Ali and Gal are included for the initial feedback thread fo= r server refresh and what is most needed
* Bringing the NVIDIA testbed o= ffline today for a few hours so Ali and Bing can do some debugging on the m= lx5 failure on the CX5 yesterday
* Arm IPSEC-MB Library: Had to move to = running from tip of main on the ARM ipsec repo to enablev1.4 - just storing= the commit hash right now but ARM will publish a new tag for a known good = version soon.
=C2=A0 =C2=A0* Wathsala will be doing the new tag soon
= =C2=A0 =C2=A0
----------------------------------------------------------= -----------
Intel Lab
* None

---------------------------------= ------------------------------------
Github Actions
* We plan to have= a maintenance window either Thus the 22nd, or sometime next week to cover = migrating to the original server.=C2=A0 This will involve upgrading the bas= e os for both the host and the VM.=C2=A0 Michael =C2=A0will send out the no= tice on the day it happens letting everyone know =C2=A0of the downtime.
= =C2=A0 =C2=A0* We don't expect that the downtime will last too long, le= ss than a day.=C2=A0 We will likely recover the workflows shortly after tha= t.

-----------------------------------------------------------------= ----
Loongarch Lab
* Patrick pinged Zhoumin about adding retest frame= work support to the Loongson lab
=C2=A0 =C2=A0* UNH willing to assist - = not sure right now what is possible/not possible based on how the loongson = lab stood up their automation. They do use the dpdk-ci repo tools.
=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=3DDTS Improvements & Test Development
* Dockerfile patch can be merge= d - Thomas has been pinged about this on slack
* Scattered packets patch= :
=C2=A0 =C2=A0* Patrick tested this on a bnxt_en NIC, and it worked fin= e
=C2=A0 =C2=A0* When gathering device capabilities, the scatter capabil= ity is always off on mlnx nics, even when including the scatter offload fla= g with testpmd
=C2=A0 =C2=A0 =C2=A0 * Juraj is going to send to Patrick = and Jermey a summary of what he has learned about passing the scatter flag = and how DPDK handles it. And what he has learned about querying for this ca= pability.
=C2=A0 =C2=A0* For now, not including the scatter offload flag= testcase with this testsuite, only submitting the testcase which is a dire= ct port from old dts
=C2=A0 =C2=A0* Luca is going to review this today. = He is also running it on a MLNX nic.
=C2=A0 =C2=A0* Juraj: Interestingl= y, on the Intel NICs it is on whether you include the --enable-scatter flag= or not, but the Mellanox NICs don't have it set to on in either case=C2=A0 =C2=A0* Ali: Have you tried "--enable-scatter --tx-offloads= =3D0x8000"
=C2=A0 =C2=A0 =C2=A0 * Jeremy Spewockwill try this
* = For next Wednesday, We are going to have to have discussions for putting to= gether the 24.07 DTS roadmap since Juraj will be on vacation in March and w= e won=E2=80=99t be able to have the conversation then.
=C2=A0 =C2=A0* Pa= trick will put together some ethdev suite ideas which the new DTS employees= at UNH can start on
=C2=A0 =C2=A0* We will also review bugzilla tickets= then, assign more tickets if needed
* Patch for the testcast blocking:<= br>=C2=A0 =C2=A0* There is a bug (it does not include the smoke tests in th= e list of suites to run), so Juraj will be sending a new version
* Grego= ry reached out to see whether his framework=E2=80=99s approach could be use= d for simple DTS cases. Juraj and Patrick read the code. There are some goo= d ideas we are bringing into the framework, but not the phase based yaml ap= proach which translates scapy and testpmd commands to testsuites.
* XMLR= PC Server: Jeremy found that there is a python dependency for the TG server= which was not known (must be python 3.10+). Or else you will have an issue= with a XMLRPC server function not included until verson x.
=C2=A0 =C2= =A0* Should turn this into a bugzilla ticket, and can discuss on Wednesday.=
* Scapy version on TG node. Need to make sure we are on x version. Shou= ld be 2.4.4 or 2.5.0?
Any other business

=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D
* Next Meeting: March 7, 2024
--00000000000057d782061212dfa4--