[MPTCP] [Weekly meetings] MoM - 11th of June 2020

[MPTCP] [Weekly meetings] MoM - 11th of June 2020

[Matthieu Baerts]

[-- Attachment #1: Type: text/plain, Size: 8709 bytes --]

Hello,

Yesterday, we had our 103rd meeting with Mat, Ossama and Todd (Intel 
OTC), Christoph (Apple), Paolo 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

     netdev (if mptcp ML is in cc) (by: Geliang Tang, Paolo Abeni):

1306666  [net] mptcp: don't leak msk in token container
1306664  [net] mptcp: fix races between shutdown and recvmsg
1305085  [net,v2] mptcp: bugfix for RM_ADDR option parsing
1303633  [net] inet_connection_sock: clear inet_num out of destroy helpe

     our repo (by: Davide Caratti, Florian Westphal, Paolo Abeni):

1306843  [mptcp-next,2/2] mptcp: add receive buffer auto-tuning
1306841  [mptcp-next,1/2] selftests: mptcp: add option to specify size o
1305430  [v2,3/3] mptcp: do nonce initialization at subflow creation tim
1305429  [v2,2/3] Squash-to: "net: mptcp: improve fallback to TCP"
1305428  [v2,1/3] Squash-to: "mptcp: refactor token container."
1303988  [mptcp-next,v2] net: mptcp: improve fallback to TCP
1303654  [4/4,DO-NOT-MERGE] mptcp: use kmalloc on kasan build
1303655  [3/4] mptcp: don't leak msk in token container
1303656  [2/4] Squash-to: "mptcp: introduce token KUNIT self-tests"
1303652  [1/4] Squash-to: "mptcp: refactor token container."
1292673  [RFC,net-next,2/2] mptcp: fix races between shutdown and recvms



Pending patches:
     - The list of pending patches can be seen on PatchWork:
       https://patchwork.ozlabs.org/project/mptcp/list/?state=*

     netdev (if mptcp ML is in cc) (by: /):

/

     our repo (by: Davide Caratti, Paolo Abeni):

1307088: Awaiting Upstream: mptcp: close poll() races:
     - reviewed by Mat
     - *Matth* : TODO: can be applied

1307588: New: [net-next] mptcp: fallback in case of simultaneous connect:
     - can be reviewed
     - should address all comments
     - Christoph reported another splat but another patch will be needed 
for this one (different)



Issues on Github:
     https://github.com/multipath-tcp/mptcp_net-next/issues/

     open: (latest from last week: 35)

Note: you can use some tags to automatically close issues:

 
https://help.github.com/en/enterprise/2.16/user/github/managing-your-work-on-github/closing-issues-using-keywords#about-issue-references

   39  [syzkaller] WARNING in subflow_data_ready (v2.0):
       - /

   36  [syzkaller] WARNING in mptcp_token_destroy:
       - /

   35  [syzkaller] WARNING in subflow_data_ready 3 @dcaratti:
       - will be closed when "mptcp: fallback in case of simultaneous 
connect" is applied

   34  WARNING: Bad mapping: ssn=1 map_seq=498340137 map_data_len=79:
       - maybe fixed with recent modifications on last export branch
       - Paolo was not able to reproduce it (but not sure if he was able 
to reproduce before)
       - *Christoph* will check if it is still detected by syzkaller

   33  weighttp-test on 1KB file with mptcp.org-client in "ndiffports"-mode
   31  Allow MPTCP + SYN_COOKIES [enhancement]
   30  ssh restart does not work

   26  Unstable packetdrill tests:
       - there is an issue (endianess issue but dack4 is now correctly 
understood), *Davide* will check
       - CI tests are more stable now

   25  mptcp_connect: Connection reset by peer during a test:
       - can certainly be closed
       - *Matth* : TODO check that

   24  Revisit layout of struct mptcp_subflow_context [enhancement]
   21  sort-out {set,get}sockopt handling [enhancement]
   20  implement msk diag interface [enhancement]:
       - could be possible now that the token container has been refactor
       - there is an API issue with IPPROTO_MPTCP
       - "Florian said that the diag handler for IPPROTO_RAW already 
uses it"
       - adding something in '/proc/net/mptcp'?:
           - instead of a real diag interface
           - *Matth* : TODO: we said something about that at the LPC, 
check that

   19  let PM netlink update live sockets on local addresses list change 
[enhancement]
   18  allow 'force to MPTCP' mode [enhancement]
   17  audit 'backup' flag usage  [enhancement]
   15  reduce mptcp_out_option struct size [enhancement]
    6  loss and delay without reordering causes very slow transfer
    4  keep a single work struct in mptcp socket [enhancement]

     recently closed:

   38  [syzkaller] INFO: task hung in lock_sock_nested
   37  [syzkaller] general protection fault in mptcp_stream_connect
   32  [syzkaller] INFO: rcu detected stall in ip_rcv
   29  [syzkaller] INFO: task hung in lock_sock_nested
   28  [syzkaller] WARNING in __mptcp_move_skbs_from_subflow
   23  refactor token container
   22  fix covscan issues 1 @dcaratti
   16  get rid of mptcp hooking in tcp_check_req()
   11  fix fallback to TCP  1 @dcaratti



FYI: Current Roadmap:
     - Part 4 (next merge window):
         - Fix bugs reported on Github: 
https://github.com/multipath-tcp/mptcp_net-next/issues/
         - 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)
         - PM server (more advanced)
         - Active backup support
         - sending "simultaneously" on different subflows (multiple non 
backup subflows)
         - Full DATA_FIN support [WIP by Mat]
         - ADD_ADDR for MPTCPv1: echo bit [WIP by Peter]
         - Opti in TCP option structures (unions) [DONE by Paolo]
         - Shared recv window (full support) [DONE by Florian]

     - Part 5 (extra needed for prod):
         - opti/perfs
         - TFO
         - PM netlink
         - PM bpf
         - Scheduler bpf
         - syncookies
         - [gs]etsockopt per subflow
         - notify the userspace when a subflow is added/removed → cmsg



Other in progress fixes for -net:
     - Paolo is currently at failures with AB: it gets stuck after a few 
connections:
         - the client tries to look at adding new subflows while the 
connection is closing
         - shutdown is blocking
         - maybe linked to issue 33



Part 4: new features:
     - news about "Full DATA_FIN support"?:
         - /

     - news about "ADD_ADDR for MPTCPv1: echo bit"?:
         - /

     - Florian might look soon at issue 21 (sort-out {set,get}sockopt 
handling)



Extra tests:
     - news about Syzkaller? (Christoph):
         - https://github.com/google/syzkaller/pull/1815
         - upstreaming what was used to test the PM
         - already got a review!

     - news about interop with mptcp.org? (Christoph):
         - /

     - news about Intel's kbuild? (Mat):
         - strangely no new update

     - packetdrill (Davide & Todd):
         - we can have 2 branches:
             - one for the CI (to test patches in mptcp_net-next)
             - one for the new features (stuff not in mptcp_net-next)
         - *Davide* will look at that
         - Todd continues working on MP_JOIN support. Might have less 
time but his work is on Github

     - CI (Matth):
         - should we test more combinations of mptcp_connect.sh?
             - there are many possible combinations
             - feel free to share interesting ones!

         - run the C reproducers from syzkaller:
             - to check regressions
             - *Matth: TODO: regroup them in a repo and test them

         - run AB tests:
             - when stable (Paolo working on it)
             - can be done in one VM with different ns
             - using https://github.com/pabeni/mptcp-tools/tree/master/mpab
             - maybe re-use the setup from mptcp_connect.sh
             - but it would only be internal: AB cannot be integrated in 
selftests for upstream

     - Coverity (Davide):
         - no news from coverity:
             - 2 cases are still open:
                 - one is fixed
                 - the other one is a false positive



Next meeting:
     - We propose to have the next meeting on Thursday, the 18th of June.
     - Usual time: 16:00 UTC (9am PDT, 6pm CEST)
     - Still open to everyone!
     - https://annuel2.framapad.org/p/mptcp_upstreaming_20200618



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

Talk to you next week,
Matt
-- 
Tessares | Belgium | Hybrid Access Solutions
www.tessares.net