[tor-bugs] #6799 [Tor]: Don't expire unused relay-to-relay TLS conns so quickly

[Tor Bug Tracker & Wiki]

#6799: Don't expire unused relay-to-relay TLS conns so quickly
-------------------------+-------------------------------------------------
     Reporter:  arma     |      Owner:
         Type:  defect   |     Status:  needs_review
     Priority:  major    |  Milestone:  Tor: 0.2.5.x-final
    Component:  Tor      |    Version:
   Resolution:           |   Keywords:  tor-relay anonymity-attack
Actual Points:           |  025-triaged 024-backport andrea-review-0255
       Points:           |  Parent ID:
-------------------------+-------------------------------------------------

Comment (by andrea):

 Code review for the ticket6799_024_v2 branch:

 051d599b4adba70312e23148f5e208075b673bae:
  - I think this is the case, but just to double-check: is the only visible
    behavior which depends on or_conn->idle_timeout closing the connection?
    This defense depends on there not being any way for the attacker to
 learn
    the randomized timeout.
  - All the code here looks correct to me modulo the bug fixed in
    add7c7c50c2ba7357d1bf22132f8b9985060f4b0

 add7c7c50c2ba7357d1bf22132f8b9985060f4b0:
  - This looks fine to me.

 b9919b7bae75f831d31ae5d3d11bb0b721bb9aab:
  - I think this patch is correct, but is this another case where things
 might
    break if the clock jumps and we should use CLOCK_MONOTONIC if
 available?
    That may be a bug in the old code too.
  - We run run_connection_housekeeping() once a second, so this has the
 same
    granularity as the randomized timeouts and doesn't reduce the effective
    entropy, but that's a bit by coincidence.  Perhaps a comment to that
    effect somewhere?

 fbc964b41b0bff4e55e90a1245dc65744abbebc2:
  - Isn't have_any_circuits = 0; redundant, since it's initialized to zero?

--
Ticket URL: <https://trac.torproject.org/projects/tor/ticket/6799#comment:25>
Tor Bug Tracker & Wiki <https://trac.torproject.org/>
The Tor Project: anonymity online