[tor-project] Meeting notes, network team meeting, 18 Dec

Nick Mathewson nickm at torproject.org
Mon Dec 18 20:28:11 UTC 2017

Hi!  Our transcript is available at
 .  I copy-paste our pad below:

Network team meeting pad, 18 December 2017

"Why," Leo's eyes widened with delight, "it's an engineering problem after all!"

       -- Lois McMaster Bujold, _Falling Free_

Welcome to our meeting!  Every Monday at 1700 UTC 1800 UTC on
#tor-meeting on OFTC.  (This channel is logged while meetings are in

Want to participate?  Awesome!  Here's what to do:
    1. If you have updates, enter them below, under your name.
    2. If you see anything you want to talk about in your updates, put
them in boldface!
    3. Show up to the IRC meeting and say hi!

Note the meeting location: #tor-meeting on OFTC!
 (See https://lists.torproject.org/pipermail/tor-project/2017-September/001459.html
for background.)

Meeting notes from last week:
    * https://lists.torproject.org/pipermail/tor-project/2017-December/001598.html

Old Announcements:
    - On the roadmap spreadsheet: Please take december/january tasks.
(If somebody else has already taken something you want, please talk to
them and/or add yourself too.)
    - Soon it will be It is now time to triage 0.3.3.x tickets.
Please make yourself the owner of tickets in that milestone that you
will do.
    - The meeting time is now 1800 UTC.
    - The _free() patch got merged so I would like to bring this to
everyone attention:

 doc/HACKING/CodingStandards.md -> Allocator conventions

    - Tor (or alpha) came out last week
    - Let's have some proposal discussions. Isis kicked off the process here:

    - We are planning a single-day pre-planned highly-focused hackfest
before the Rome Tor meeting
        - teor: Jon says we will have a venue for the pre-Rome hackfest
        - komlo: Should there be a Rust update/planning session?  What
should we prepare/do beforehand to help this be
    - review-group-27 is open, there are only 2 tickets left
    - No regularly scheduled network-team meetings on 25 Dec through
Jan 5. Next team meeting is 8 Jan 2018!
    - Please check in with the roadmap; december is almost over as a
working month!  What isn't getting done? How are we doing?

    - How shall we revise the 0.3.2.x schedule?
    - And should we revise the 0.3.3.x schedule accordingly?

    18:31 <+nickm> how about we aim for jan 8 for stable (if
possible), and delay
               feature-freeze by 1 week on 0.3.3?

    - strategy/plan for addressing current network issues.

       dgoulet, nickm, teor, isis try to figure it out.

    - Last week:
        - PrivCount fixes and performance improvements
        - Admin for collecting research data
        - Got an equal-area projection for metrics maps (thanks, cyperpunks!)
          It's been deployed at https://atlas.torproject.org/#map
          Gosh, the southern hemisphere is big.
          And we have almost no relays south of the tropic of cancer.
        - Various security stuff and code reviews and helping volunteers
        - Booked flights for Rome
        - Tried to kick off the fallback directory rebuild, but didn't
get time to review the tickets
    - This week:
        - It is a short week for me, I am on all-volunteer time /
holidays from Wednesday onwards
        - Work on getting the test network running new IPv6 consensus code
        - Revise some of the IPv6 patches
        - Rebuild fallback directory mirrors, along with whoever is
around at the time
          (pastly, Hello71, let's make a time to do fallbacks? Or talk
on tor-dev@?)
          How about this week's patch party? (pastly)
        - Review some other volunteer tickets

  Last week:
  - Reviewed mike's #13837 and #23101.
  - Restarted coding the prop247 simulator. Requires more work!!
  - Reviewed #23271 from ffmancera.
  - prop224 work: Develop and review on #24595. Worked on #24610.
  - Met with David Stainton, Moritz and others. Talked about relay load
    balancing and bandwidth dirauths. People are sad about the state of the
    network: some relays are overloaded while others idle, many overloaded
    relays cant even establish circuits to each other. Need to do something
    about it: deploy bwscanner and start thinking about peerflow. What about
    isis' bridge bandwidth scanner?
  This week:
  - Second round review of mike's #13837 and #23101.
  - More work on prop247 simulator.
  - Try to make a plan forward for the bw auth work. I'm really hoping we can
    find someone to work on this!!!

    Last week:
        - Reviewed a panoramix design thing for dawuud
        - Finished rust privcount proof-of-concept implementation
        - Read a bunch of rust documentation
        - Released
        - Made progress on diagnosing and fixing 23847 (Make sure Tor
can shut down via control port, and start again in same process).
It's not perfect now, but at least it doesn't automatically crash like
it used to!  It'll be cool if mobile devs try it out
        - Diagnosed 24367 (Changing pluggable transports (during
start-up) in Tor Browser is broken) and tried to fix it.  Fix is
either wrong, or incomplete.
        - Fix other issue encountered while looking at code for above.
        - Review, review; merge, merge.
    This week:
        - Try to help with dos issues, if I can
        - Try to finish a working fix for 24367
        - Review review; merge, merge.
        - Revise recent needs_Review tickets
        - Check-in with december roadmap; december is nearly done.

    Last week:
        - Updated #13837 and #23101 based on asn and dgoulet's code reviews
        - Reviewed #24613
        - Fixed a memory leak in the circuitstats unit tests.
          - What is the official valgrind options/leak checking thing?
valgrind --leak-check=full --show-leak-kinds=all has lots of noise.
    This week:
        - Continued #13837 and #23101 updates
        - #244847

    Last week:
        - met with nick, ahf and catalyst on sponsor8 Q4 work
        - was at Mozilla All Hands
    This week:
        - sent update to list about follow ups from Mozilla All Hands
        - isis - bugs #24636 and #24637 are blocking TB team on moat
integration [please prioritize these, the TB team deadline is getting
        - need to follow up on all sponsor8 discussion from last week
        - sync on .onion states experience w/ asn, antonela and geko
[wed at 1900]

        [asn: sounds good]

komlo (offline):
    - Last week:
        - Submitted #23881 (tor logging) for re-review.
    - This week:
        - Pick the next piece that will be helpful to improving
support for Rust in core tor. I wonder if #24249 (minimal automated
mechanism to keep C/Rust types in sync) will be a good next step?
Especially toward making future/current Rust work in core tor easier
to maintain. If there is other low-hanging fruit that will be more
helpful, let me know.
        - Helping on #24030, which someone picked up as their first
tor ticket (which is great)

    Last week:
        Sponsor 8:
            - Uploaded patches for event loop metrics (#24605).
            - Looked into testing if the monotime fixes have helped us on
              32-bit arm: https://people.torproject.org/~ahf/sponsor8/20171217/
              (#24613 + #24374)
            - Status meeting with Catalyst, Nick, and Isa!
            - Looking into the HSv3 service issue(s) that David is now digging
    This week:
        Sponsor 8:
            - Talk with Nick about next steps with CPU optimizations:
                - Remaining udivdi3 issues?
                - SHA3 ?
            - See if I can get Shadow to run on device/emulator to be able to
              run more deterministic experiments on Android.
            - Finish my script to `bench` tests individually with automatically
              generated graph+report.txt from Simpleperf.
            - Finish Q4 reporting (missing the benchmark info).
    Misc: Going to 34c3 between christmas and NYE, back Tuesday the 2
of Jan. 2018

    Last week (2017-W50):
        - sponsor 8 status meeting
        - helped review STACK patches some more
        - looked more at delaying descriptor downloads if we'd get
stuck without a live consensus where we need one
        - looked at failing offline Rust builds in Travis (thanks isis!)
        - upgraded my virtualization infrastructure (with minor obstacles)
        - fighting medical billing bureaucracy
    This week (2017-W51):
        - more work on delaying descriptor downloads

        - somehow need a cheap way to ask guard subsystem whether it
would need a live consensus to make progress; this would get called
from periodic callbacks so counting smartlist entries each time
probably isn't a great idea

- vacation 12/22 (also 12/25-1/1 along with most of TPI)

  Last week:
    - HS IPv6 work starting with the unify link specifier ticket #22781.
    - Some scheduler work on #24554.
    - Investigated #24011 but wasn't able to figured it out :S
    - Dirauth have been notified and they are now broadcasting the consensus
      param hsdir_spread_store=4. #24425
    - Investigated #24610 HS bug and tried to come up with a fix.
  This week:
    - Try to wrap up #22781 hopefully and continue the IPv6 HS work (see
      tor-dev@ thread about it that teor started).
    - Follow up on #24554.
    - Review group and ticktets for 032.
    - Do some more triaging on 033.

    last week:
        - proposal discussion meeting for prop#249
        - began writing a proposal to migrate to TLS 1.3
        - began writing a proposal to use ECDH+SIDH for relay-to-relay
TLS connections
        - worked on #23886 (fn pointers to ed25519-dalek),
        - #24608 (cargo issue causing build troubles for us),
        - several moat/spec bugs #24433
        - how do we want to do low-level external crypto dependencies
in our rust code? (i.e. sha2/sha3 getrandom())
            [nickm: i have thoughts here; let's talk on tor-dev after
the meeting? ]
    this week:
        - vacation starting friday (through NYE)

More information about the tor-project mailing list