All of lore.kernel.org
 help / color / mirror / Atom feed
From: Aaron Conole <aconole@redhat.com>
To: techboard@dpdk.org, dev@dpdk.org
Subject: Tech Board Meeting Minutes - Sept. 23, 2024
Date: Thu, 24 Oct 2024 09:41:52 -0400	[thread overview]
Message-ID: <f7t5xph7h3z.fsf@redhat.com> (raw)


Attendees
---------
Aaron Conole
Bruce Richardson
David Marchand
Hemant Agrawal
Honnappa Nagarahalli
Kevin Traynor
Konstantin Ananyev
Maxime Coquelin
Morten Brorup
Robin Jarry
Stephen Hemminger
Thomas Monjalon


Notes
-----

The Tech Board met at a f2f session in Montreal, prior to the DPDK
Summit.  This meeting went much longer than a typical Tech Board
meeting.


Agenda
------
* Prep. for the summit
  ** Make sure people helping to monitor live stream know what they are
     doing
  ** Discussion about board q&a
  ** Discussion about workshops
     ** Will need people to help do leading at the workshop sessions

* Code challenge
  ** Idea: identify and promote tools to help detect issues early
  *** Tech board needs to give some guidance for this effort, since it
      isn't well known how to move forward.
  *** Does it have some CI component?  Can't just turn it on unless
      everything is passing, otherwise it will put expected failures.
  ** Idea: offer some rewards for these efforts
  *** What kinds of rewards to offer?
  ** Idea: can we also hire someone?

* Events discussion
  ** Techboard doesn't have much transparency on how the decision for
     venue, location, etc. is chosen.
  *** Ex: reasoning for venues and locations isn't well known
  *** Ex: Montreal had difficult visa process which resulted in some
          missed in-person attendance.
  ** Techboard would like to be able to give some input on the process
  *** Ex: hotels for the events have been costly
  ** We would like some of the maintainers to also be sponsored for
     attendance:
  *** David Marchand
  *** Akhil Goyal
  *** Ferruh Yigit
  ** Remote / hybrid speaking is going to have to be a way of life.  We
     should reconsider the remote spoeaker statuses.  We needed to
     scramble at the last minute to accommodate remote speakers.
  *** We also want to continue to give preference to in-person.
  ** Awards, gifts, etc.
  *** We used to give awards for different areas, and we would like to
      return to doing that.
  *** We want to give speaker gifts, if possible.
  *** Maybe also do some recognition
  **** Thomas to do slides for 'hall of fame' kind of recognition to
       contributors.

* Marketing discussion
  ** We aren't sure what the marketing goals are
  ** We feel that the marketing for the events was quite good
  ** Maybe we can have a session with marketing to understand
     what the other goals of marketing are.
  ** Is there a way to have marketing show the value of investing in
     DPDK to the users (ie: rather than just taking DPDK code and
     disappearing)?
  ** Reminder to interview with Ben during the conference

* Testing
  ** Intel lab having issues
  *** Need to sync with John McNamara and others about the system
  *** Seems that the Intel CI is in a barely maintained state.
  **** Ex: Testing old OSes, but not newer ones, etc.
  *** Bruce will send a message.
  ** Ask for the lab to test grout
  *** Aaron to sync with Patrick about it
  ** DTS
  *** Basic testing for ethdev to complete by the end of this year
  *** What are we targeting next?
  **** Ex: maybe rte_flow?
  *** Need guides for how to run DTS.  Maybe youtube videos?
  **** Presentation at the conference might cover some of this.
  **** More should be done - blog posts, documentation, etc.
  ** Testing gaps we might have
  *** Kernel has Syzbot fuzzer that just randomly hits APIs with
      garbage data to try and break things
  **** Can we have something like this in DPDK?
  **** Needs to be automated
  **** Do we need reference implementations on the behavior to test
       against?  At least if there aren't crashes that is good.
  **** How can we implement it?
  ** Can we add some kind of testing initiatives to the code challenge?

* Gov. Board
  ** We haven't gotten many meeting minutes
  ** We like the rotating tech board representative
  ** Gov. board gives initiative requests to tech board, but there
     aren't any kind of resources attached.
  *** would be good to have guarantees about spending time on these
      initiatives.
  ** Need more transparency where possible.

* Tech board
  ** Need to do more outreach for contributions.

* Feature / Bug tracking
  ** Make sure maintainers are on the bug database and getting mails
  *** They are supposed to be
  ** Kevin will push on bugs to come to resolution.
  ** Roadmap on the website is out of date
  *** Goal is to avoid duplicating work across organizations
  *** The roadmap shouldn't be used to have some feature requests, but
      should really show what someone is working on.
  *** Work needs to be tied to specific individuals.
  *** Need to remove any outdated details / rework the roadmap display.
  ** Feature requests
  *** are they useful?  who will work on them?
  *** Formal feature request process might give a false hope that
      someone will work on something.
  *** Better to be on the mailing list rather than bugzilla so that
      others can have involvement with discussions
      
* TOO MANY PATCHES
  ** We have a huge patch backlog
  *** Stephen can take another pass through the patchwork backlog, but
      the whole thing doesn't scale
  ** Need more reviewers
  *** Getting quality reviews where someone will use 'Reviewed-by' vs.
      merely acknowledging that some idea is reasonable to accept with
      'Acked-by'
  *** How to get more people motivated to review
  *** Reviews are a large burden right now on David and Thomas
      primarily.
  ** Can we auto-merge patches?
  *** Other communities do this, and merge if someone hasn't commented
      in a few days' time.
  *** If someone breaks code, then it will incentivize people to do more
      reviewing... maybe?
  ** Can we find another person to do the kind of maintenance that David
     and Thomas do?
  *** Plan is to use the maintainers meeting to discuss
  ** Ask Gov Board to ask their respective comanies for more reviewers

* Branches
  ** Maybe restart the short-term stable releases
  *** Difficult to see the value
  *** Already exists a way to get point releases, someone just needs to
      step up for those, but no one does

* Grout
  ** project aims to replicate CNF/VNF use cases

* discussion about where apps fit
  ** testpmd and apps naming don't seem right?

* Cloud initiative
  ** Amazon
  *** Wants to participate in CI
  ** Microsoft
  *** Discussions will start soon
  ** Who else to reach out to?
  *** Ali Baba?
  *** others?

* Security
  ** Status doc to be published
  ** Protocols we want to see?
  *** Wireguard?
  *** psp?

* AI
  ** Use as a tool in Patch Reviews?
  *** Aaron has some code that kindof does it
  *** Needs robust training data
  *** Costly to integrate
  ** Do we need more publicity of mldev?


                 reply	other threads:[~2024-10-24 13:41 UTC|newest]

Thread overview: [no followups] expand[flat|nested]  mbox.gz  Atom feed

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=f7t5xph7h3z.fsf@redhat.com \
    --to=aconole@redhat.com \
    --cc=dev@dpdk.org \
    --cc=techboard@dpdk.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.