[tor-bugs] #31653 [Core Tor/Tor]: Padding cells sent with 0ms delay cause circuit failures

[Tor Bug Tracker & Wiki]

#31653: Padding cells sent with 0ms delay cause circuit failures
-------------------------------------------------+-------------------------
 Reporter:  pulls                                |          Owner:
                                                 |  mikeperry
     Type:  defect                               |         Status:
                                                 |  assigned
 Priority:  Medium                               |      Milestone:  Tor:
                                                 |  0.4.2.x-final
Component:  Core Tor/Tor                         |        Version:  Tor:
                                                 |  0.4.1.5
 Severity:  Normal                               |     Resolution:
 Keywords:  wtf-pad circpad-researchers-want     |  Actual Points:
  042-should                                     |
Parent ID:                                       |         Points:
 Reviewer:                                       |        Sponsor:
-------------------------------------------------+-------------------------

Comment (by mikeperry):

 Tobias: FYI I have noticed a perf issue with this solution. Going back
 into libevent for the callback introduces anywhere from 0-10ms delay, at
 random, on just a client. On relays, it may be much worse.. Or maybe
 better, if they are not building circuits (client path construction can
 block the event loop for a long time).

 For origin/master, this means we need to fast-path the 0-delay case still
 somehow without callbacks, and also warn about this in the developer doc.
 I bet trying to compose a packet train to fake a burst that actually has
 9ms delay between packets is going to get seen by classifiers pretty
 easily :/.

--
Ticket URL: <https://trac.torproject.org/projects/tor/ticket/31653#comment:14>
Tor Bug Tracker & Wiki <https://trac.torproject.org/>
The Tor Project: anonymity online