[tor-bugs] #24898 [Core Tor/Tor]: We have two conflicting notions of channel_is_client()

Tor Bug Tracker & Wiki blackhole at torproject.org
Wed Jan 31 04:45:20 UTC 2018

#24898: We have two conflicting notions of channel_is_client()
 Reporter:  arma                                 |          Owner:  (none)
     Type:  defect                               |         Status:
                                                 |  needs_revision
 Priority:  Medium                               |      Milestone:  Tor:
                                                 |  0.2.9.x-final
Component:  Core Tor/Tor                         |        Version:  Tor:
 Severity:  Normal                               |     Resolution:
 Keywords:  tor-relay, 032-backport,             |  Actual Points:
  031-backport, 030-backport-maybe-with-21406,   |
  029-backport-maybe-with-21406                  |
Parent ID:  #24902                               |         Points:  0.5
 Reviewer:  teor                                 |        Sponsor:
Changes (by teor):

 * status:  merge_ready => needs_revision


 Replying to [comment:19 arma]:
 > Replying to [comment:18 teor]:
 > > Do we still use CREATE_FAST to mark clients in 029? Or did we backport
 that fix as well?
 > We still do it. That is, anybody who sends you a create-fast cell when
 you're running 029 or 030 now gets counted as a client, meaning the DoS
 defenses will be triggered against them (once #24902 gets backported), and
 meaning when we consider whether a channel can satisfy an extend request,
 this one can't.
 > Why, are we going to regret that? :) See also the discussions in #22805.

 I think we are going to regret that, and we should backport the parts of
 #22805 that remove the CREATE_FAST mark client.

 We should also take them out on general principles, because it reduces the
 number of connections between relays.

Ticket URL: <https://trac.torproject.org/projects/tor/ticket/24898#comment:20>
Tor Bug Tracker & Wiki <https://trac.torproject.org/>
The Tor Project: anonymity online

More information about the tor-bugs mailing list