[tor-bugs] #18365 [Core Tor/Tor]: Fined-grain timer implementation to support per-connection or per-circuit timers

[Tor Bug Tracker & Wiki]

#18365: Fined-grain timer implementation to support per-connection or per-circuit
timers
-------------------------------------------------+-------------------------
 Reporter:  nickm                                |          Owner:  nickm
     Type:  enhancement                          |         Status:  closed
 Priority:  Medium                               |      Milestone:  Tor:
Component:  Core Tor/Tor                         |  0.2.9.x-final
 Severity:  Normal                               |        Version:
 Keywords:  performance, backend,                |     Resolution:
  TorCoreTeam201605, TorCoreTeam-                |  implemented
  postponed-201604                               |  Actual Points:  medium
Parent ID:                                       |         Points:  medium
 Reviewer:  mikeperry,athena                     |        Sponsor:
-------------------------------------------------+-------------------------

Comment (by mikeperry):

 I made some comments on IRC last week that I still do not feel were
 addressed.

 These issues are:

 1. If time is cached, then we will always have a delta from when the
 timeval is cached from the libevent callback until this timer is
 scheduled. For second_elapsed_callback() all the way to
 channelpadding_decide_to_pad_channel(), this delta could be very large.

 2. I am still not sure that the fix in #18907 for "monotonic" time
 actually prevents stalling. Last I heard Nick was going to write tests to
 verify this. Did that happen?

 I do not believe that this code can be used by the netflow code until
 these issues are addressed. You can merge it if you want, but I'm not
 willing to use it.

--
Ticket URL: <https://trac.torproject.org/projects/tor/ticket/18365#comment:19>
Tor Bug Tracker & Wiki <https://trac.torproject.org/>
The Tor Project: anonymity online