[tor-project] Network team meeting notes, 4 Dec 2017

Nick Mathewson nickm at torproject.org
Mon Dec 4 19:33:09 UTC 2017


Meeting transcript available here:

Notes below:

Network team meeting pad, 4 December 2017

    `This must be the wood, she said thoughtfully to herself, `where
    things have no names. I wonder what'll become of MY name when I
    go in? I shouldn't like to lose it at all -- because they'd have
    to give me another, and it would be almost certain to be an ugly
    one. But then the fun would be, trying to find the creature that
    had got my old name! [...] Just fancy calling everything you
    met "Alice," till one of them answered! Only they wouldn't answer
    at all, if they were wise.'

-- Through the Looking Glass

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-November/001578.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.)
    - Please review stuff in review-group-26
    - Soon it will be 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.

Discussion Topics:
    - Do we want any network team hackfest days before the March
meeting in Rome?
      (The meeting is from Sunday 11 March to Thursday 15 March, and
teor needs to book flights soon)

    - Last week:
        - Implemented PrivCount country and AS features for the next release
        - Revised some of the client code for IPv6 in microdesc
consensus (prop#283)
        - So many security issues. I helped out a little.
        - Tried to help out with the 10.000 Tor Hidden Service ticket (#15251)
        - Sketched some experimental privcount measurements for HSDirs (#24468)
    - This week:
        - PrivCount release and deployment
        - Admin for collecting research data
        - Start implementing onion service client counts for
experimental PrivCount?
        - Book flights for Rome - do we want a hackfest before Rome?
          (Discussion: Yes, that seems like a good idea to people.
Perhaps one day only. Can somebody ask Jon et al to make arrangments?)

  Last week:
  - Reviewed various SponsorV guard-discovery tickets: #23100 and
#23114! Also #13837!
  - Did some thinking on #23247 (have another UX+onion meeting this wednesday)
  - Thought about some HSDir-related statistics: #24468 and #24425
  - Backported #23862 to 031 and 030.
  - Thought a bit about tor_free code-style and inserted my opinion in #24337.
    Not planning to spend more time on this debate.
  - Opened ticket about ripping out guardfraction: #24456
  This week:
  - More review on prop247 related tickets by mike.
  - Some prop224 work on #24346 and #23603.
  - Start rebooting prop247 (see above). Sent email to get prop247
    back in track and would appreciate reply from Mike:
  - Started re-reading my prop247 simulator code.

  * Last week:
    - Tracing cell timings in tor from inbuf to outbuf with KIST. Trying to
      find issues and contentions. So far so good. Hopefully will try to post
      some results soon. (Branch is "cell-tracing")
    - Over the weekend, I hit #24346 on one of my v3 HS and turns out after
      investigation that it is due to #23603 which is back in 032 and High
    - Reviewed 032 tickets in review-group-26.
    - Some patches for 032 milestone as well for HSv3. (#24346 and #24425)
    - Worked on the some of the TROVEs' advisories and patches.
    - Investigate a bit more #23696 and the solution for now is to downgrade
      the warning to info since we will have platforms where the monotonic time
      is not true.
  * This week:
    - Attempt a 033 triage of my assigned, HS and scheduler tickets.
    - Have all my 032 tickest in needs_review (especially #23696 and #23603).
    - I should start looking at HS + IPv6 which is planned for 033. Sync up
      with what teor has been doing so far with IPv6.
    - Feature free of 033 is Jan 15th so I really need to continue working on
      033 tor-sched tickets.
     Most scheduler fixes heavily depend on #23709 being merged so would be
     neat to have a reviewer soon-ish for it. It ain't small nor simple but it
     is a lot of code removal. Thanks!

    * Last week:
        - Worked on security patches a lot.
        - Released a new alpha and security releases.
        - Worked on a rust client-side privcount implementation.
        - Released several proposals
        - Talked with a namecoin dev
        - Talked with mozilla people about specification work
    * This week:
        - Review merge_ready stuff for 0.3.2 and Reveiew group 26.
        - Review other stuff for Review group 26.
        - Ideally, merge some version of my _free() patch [#24377].
Can we decide on which of the several proposed variants we like? IMO
they are all okay.
        - Improve coarse_time speed on channels and osx and ios.
        - Open RG27, I hope.
        - Decide about a sandbox strategy for future Tor releases.
        - Follow-up about utf8 proposal
        - Follow-up about privcount proposal
        - Follow-up about hibernation proposal
        - Finish rust client-side privcount proof-of-concept hacking
        - Other tickets as time permits
        - Review 23709.

    Last week:
        - Fixed up #23114 based on asn's review
        - Implemented #13837 and #23101
    This week:
        - Write up stuf for Prop247 experiments
        - Review #20699 (I should be looking at v3 control port stuff
for prop247 anyhow).

    Last week:
        Sponsor 8:
            - Talked with Hans and Nathan about build-system integration for
              mobile Tor code if we want it in tor.git.
              - Looked into Android.mk / autotools integration.
            - Updated test scripts to fetch .onion as well as exit node tests.
            - Looked into whether #24374 is helping with udivdi usage
(missing soem work: see #1 in this week).
            - Looked at the current tracing code for Tor: Android's tracing
              framework expects "enter function"/"leave function" and
              tor_trace() probably wont be suitable for that.
            - Wrapped my head around libevent / event loop code in Tor.
            - Meeting with Mozilla about specification work.
    This week:
        Sponsor 8:
            - Catch up some loose ends from last week:
                1. Have an easier way to compare `simpleperf` sample reports.
                Right now the #24374 patches looks like they yield a
                performance enhancement, but it is hard to compare the relative
                data in the reports.
                2. Collect data from event loop data on device.
            - Create a way for other people to easily (via Orbot UI) to get tor
              bench/test results on the device for debugging (if they
are available).
            - Discuss december optimization work with Nick.
            - Bug triage rotation duty.

    last week:
        - Investigated how offline rust dependencies work and
documented it #22907
        - Debugged the meek←→moat tunnelling and redirects not working #22432
        - Fixed a couple moat bugs #24433 #24443
        - Removed some of the other (untested, unused) cruft from
bridgedb's codebase #3015
        - Updated the BridgeAuth and enabled IPv6 bridge reachability
testing #24264
    this week:
        - actually finish the proposal writeups for prop#249 prop#269
and prop#270 and the new one for XE5/hila5

    Last week (2017-W48):
        - looked at #24377 (*_free() nulling macros) -- leaning away
from uppercasing them for now
        - bug triage
        - still looking into clock skew / reasonbly live consensus
issues. structure of event handling code for connections is very
    This week (2017-W49):
        - following up on #24377 (and any related stuff)
        - work on a reasonable way to defer descriptor fetches if we
wouldn't be able to initialize our guard state due to expired
consensus (#2878, #23605)
        - maybe raise severity of untrusted clock skew warnings to
NOTICE so they're more visible?
        - catching up from roof monster attack
        - vacation 12/8 (Friday); also vacation on 12/15 and 12/22


Following topics are important lately:

 - security updates for dirauths/public list of those not updated yet

 - More difficult to host exit nodes with 80/443 port combination

 - Limit streams originating from one circuit?


Last week:

 - Got replies from testers - they also told me where they were
testing it from and so far is like 2/3 wifi 1/3 mobile

 - Finished deadline stuff :) Also participated on a nice panel about
privacy,surveillance within the political context now (it was a closed
meeting with private foundations) I spoke a lot about ways .onion
services can help the resistance (will email use cases to asn)

This week:

 - sync on wed with asn, geko and antonela - follow up on:#23247

 - start reviewing sponsor8 quarter work - to see how we are doing in general

 - work on a new set of reports (none related to network team tho)

 - prep for mozilla all hands next week

 - will be in seattle wed - sat

 - Sent out Tor Launcher weekly check in email (ping for isis) asking
folks to send their updates for this week

More information about the tor-project mailing list