[MPTCP] [Weekly meetings] MoM - 9th of January 2020

[MPTCP] [Weekly meetings] MoM - 9th of January 2020

[Matthieu Baerts]

[-- Attachment #1: Type: text/plain, Size: 7665 bytes --]

Hello,

We just had our 81st meeting with Mat, Peter and Ossama (Intel OTC), 
Christoph (Apple), Paolo, Florian and Davide (RedHat) and myself (Tessares).

Thanks again for this new good meeting!



Here are the minutes of the meeting:


Accepted patches:
     - The list of accepted patches can be seen on PatchWork:
       https://patchwork.ozlabs.org/project/mptcp/list/?state=3
     - By Florian Westphal, Mat Martineau, Matthieu Baerts, Paolo Abeni

1220414  [2/2] protocol: fix list corruption
1220413  [1/2] protocol: pm callback must be done after msk state is set
1219998  Squash-to: "tcp, ulp: Add clone operation to tcp_ulp_ops"
1219492  Re: Rebase request
1218978  [v2] Squash-to: "sock: Make sk_protocol a 16-bit value"



Pending patches:
     - The list of pending patches can be seen on PatchWork:
       https://patchwork.ozlabs.org/project/mptcp/list/?state=*
     - By Matthieu Baerts, Paolo Abeni, Peter Krystad

1194592: Changes Requested: [RFC,1/1] mptcp: Optimize struct mptcp_recei
1196109: RFC: [10/10,RFC] selftests:mptcp: decrease timeout to 100 sec
1220535: New: [v2,1/7] Squash-to: "mptcp: Handle MP_CAPABLE options for
1220536: New: [v2,2/7] Squash-to "mptcp: Create SUBFLOW socket for incom
1220534: New: [v2,3/7] Squash-to: "mptcp: Add key generation and token t
1220537: New: [v2,4/7] Squash-to: "mptcp: Write MPTCP DSS headers to out
1220541: New: [v2,5/7] Squash-to: "mptcp: Implement MPTCP receive path"
1220540: New: [v2,6/7] Squash-to: "mptcp: cope with later TCP fallback"
1220539: New: [v2,7/7] Squash-to: "mptcp: Add handling of incoming MP_JO



FYI: Current Roadmap:
     - Part 1 (mainly TCP changes, will be sent with Part 2):
         - Sent to Netdev, in review
     - Part 2 (minimum set for MPTCP, up to KSelftests, one subflow):
         - Sent to Netdev, in review, waiting for part 1 to be merged
     - Part 3 (MPTCPv1):
         - Ready to be sent to Netdev, waiting for part 2 to be merged
         - we propose to merge part 2 and 3, see below
     - Part 4 (to be sent before the end of the merge window or more 
probably at the next one)
         - Full DATA_FIN support [WIP]
         - Shared recv window (drop data received on other subflows) [TODO]
         - Active backup support [WIP]
         - Opti in TCP option structures (unions) [to be rebased)
     - Part 5 (to fully support MPTCP):
         - Shared recv window (full support)
         - IPv6 - IPv4 mapped support
         - not dropping MPTCP options (ADD_ADDR, etc.)
         - FAST_CLOSE
         - full MPTCP v1 support (reliable add_addr, etc.)
         - after a few attempts of failed MPTCP, we fallback to TCP 
(like TFO is doing)
     - Part 6 (extra needed for prod):
         - opti/perfs
         - TFO



Initial submission to net-next:
     - Part 1:
         - v6 has been sent at the beginning of this week
         - v7 has been sent just before this meeting
         - we might have to send a v8 at the end of the week

     - Part 2:
         - v3 will be sent when part 1 is merged, not even as an RFC 
except if asked

     - Part 3:
         - we propose to merge part 2 and 3, see below



Merge window:
     - the v5.6 kernel predictions: merge window closes on Sunday, 
2020-02-23 (and release on Sunday, 2020-04-12)

     - But net-next should close when v5.5 is released, around the 2nd 
of February (Sunday, 2020-02-02).

     - We don't want to have part 2 merged if part 3 is not merged in 
the same window (not to have MPTCPv0 in one release and v1 later)

     - Can we merge part 2 and 3 together? Part 2 is composed of 16 
patches, part 3 has 4 patches but Christoph has to send them.

     - Can we ask not to merge part 2 without part 3?

     - Decision:
         - Christoph will send part 2 and 3 together
         - → TODO: *Matth* adapt Signed-off
         - → TODO: *Matth* adapt part 2/3 in the notes



Other WIP items:
     - Active backup support:
         - Florian is working on it, patches have been shared

     - DATA_FIN:
         - Mat is working on it

     - Paolo is looking at TCP fallback (patches have been shared)

     - Some TODOs:
         - reviewing and comment about memory barriers in part 2 → we 
need to improve comment around them (would be nice to squash that in 
patches from part 2):
             - Florian can help looking at that

         - memleak if a request comes after a socket is closed:
             - Florian plans to look at that

         - reduce locks when mptcp_stream_accept (currently 3 locks) but 
we might need to add an helper in inet API



Should we have a kind of "retrospective" at some points?:\x10
     - to focus on stuff we could improve, remove, what's working well, etc.
     - We did it quickly:
         - Patchwork:
             - helpful
             - don't hesitate to update status

         - CI:
             - we should run BPF selftests (not at each run)
             - we should run other network selftests
             - could be good to have syzcaller (another server?)
             - → having the builds from Intel's kbuild? → TODO *Mat*
             - → could be good if we don't need to run all these tests 
all each time we modify our git repo

         - TopGit tree:
             - good to be able to change commit messages with patch
             - good to review (no squash)
             - we can continue to re-create the tree using a rebased 
branch (review is harder but quicker to do in some cases)

         - reviews:
             - seems working well

         - meetings:
             - once per week for max one hour seems OK

             - listing accepted/pending patches: summary for the 
accepted one, ask for questions in pending ones?

             - face-to-face meeting? could be good

         - bug tracking: not needed? (not discussed)

         - "roadmap": seems ok? (not discussed)

         - testing:
             - more to launch by the CI

             - packetdrill:
                 - looking good so far (on Davide's github 
https://github.com/dcaratti/packetdrill )
                 - would be good to upstream that
                 - before we have it upstreamed, we can use Davide's version
                 - we can also move this repo under multipath-tcp group 
on Github
                 - basic tests will be part of Davide's github
                 - others can be added "somewhere" (new repo/new branch)



Netdev 0x14:
     - Submission deadline 15-Jan-2020. 
https://netdevconf.info/0x14/submit-proposal.html
     - Conference 17-20 March
     - Recycled talks not accepted
     - Ideas?:
         - Practical MPTCP - interoperability and effects of v1 protocol 
choice, userspace API (converting apps, socket option 
limitations/details), current upstream feature set
         - Selftest & MPTCP
         - mptcpd (part of userspace API?)
     - Mat and Peter are not "far", they might go
     - harder for people in Europe



Next meeting:
     - We propose to have the next meeting on Thursday, the 16th of January.
     - Usual time: 17:00 UTC (9am PST, 6pm CET)
     - Still open to everyone!
     - https://annuel2.framapad.org/p/mptcp_upstreaming_20200116



Feel free to comment on these points and propose new ones for the next
meeting!

Talk to you next week,
Matt
-- 
Matthieu Baerts | R&D Engineer
matthieu.baerts(a)tessares.net
Tessares SA | Hybrid Access Solutions
www.tessares.net
1 Avenue Jean Monnet, 1348 Louvain-la-Neuve, Belgium