[tor-bugs] #14216 [Tor]: We only "optimistically retry connections" once if we have other bridges descs

Tor Bug Tracker & Wiki blackhole at torproject.org
Sat Jan 31 08:44:11 UTC 2015


#14216: We only "optimistically retry connections" once if we have other bridges
descs
------------------------+--------------------------------------
     Reporter:  arma    |      Owner:
         Type:  defect  |     Status:  needs_review
     Priority:  normal  |  Milestone:  Tor: 0.2.6.x-final
    Component:  Tor     |    Version:
   Resolution:          |   Keywords:  tor-bridge, 025-backport
Actual Points:          |  Parent ID:
       Points:          |
------------------------+--------------------------------------

Comment (by sysrqb):

 Replying to [comment:3 arma]:
 > If you merge this patch, stuff gets better, but not completely better.
 People who switch between multiple PTs, and who try to use PTs that
 shoehorn their network design into our bridge design (currently meek and
 flashproxy, with their "I'll just say there's a bridge on 127.0.0.1:9060
 and never you mind that there's a whole internet behind that one bridge"),
 will become much happier when they're on not-perfectly-stable network
 connections.

 I think this is a good enough fix for 0.2.6. I started writing a more
 comprehensive one, but I think this will do fine for now. I also wonder if
 we should also check !e->bad_since
 {{{
 -          node_is_bridge(node) == need_bridges) {
 +          node_is_bridge(node) == need_bridges &&
 +          (!need_bridges || (!e->bad_since &&
                               node_is_a_configured_bridge(node)))) {
 }}}

 As far as I can tell, we only set bad_since on a bridge's entry_guard_t
 when we used it and then it was remove from the configuration, so it's
 marked as bad/unlisted.
 >
 > People who are in the above situation, who have two bridges in their
 state file which offer different PTs but which use the same bridge
 identity fingerprint, will not see an increase in happiness from this
 patch. That's because node_is_a_configured_bridge(node) says "yes" if
 node's identity fingerprint is a configured bridge, even if the PT offered
 by node isn't the one that's configured.

 Created #14581 for this.

--
Ticket URL: <https://trac.torproject.org/projects/tor/ticket/14216#comment:5>
Tor Bug Tracker & Wiki <https://trac.torproject.org/>
The Tor Project: anonymity online


More information about the tor-bugs mailing list