[tor-bugs] #5458 [Tor Client]: Clients should warn and disable guards responsible for excessive circuit failures
Tor Bug Tracker & Wiki
torproject-admin at torproject.org
Thu Jun 14 19:02:37 UTC 2012
#5458: Clients should warn and disable guards responsible for excessive circuit
failures
-----------------------------+----------------------------------------------
Reporter: mikeperry | Owner: mikeperry
Type: defect | Status: needs_revision
Priority: major | Milestone: Tor: 0.2.3.x-final
Component: Tor Client | Version:
Keywords: MikePerry201206 | Parent: #5456
Points: 6 | Actualpoints:
-----------------------------+----------------------------------------------
Comment(by mikeperry):
Replying to [comment:31 nickm]:
> {{{
> + if ((guard->first_hops % (unsigned)pathbias_get_min_circs(options))
== 0) {
> + entry_guards_changed();
> + }
> }}}
>
> That's a kludge, and I think it might be a needless one. We can just
say entry_guards_changed(), and that will cause a state write to happen in
10 minutes or an hour, depending. Am I wrong?
Yeah, I was just reluctant to cause the equivalent action of "Write your
state file every 10 minutes". I guess we don't care? I'll change
entry_guard_inc_first_hop_count() to call entry_guards_changed() every
time.
> Using "0" for the "never disable circuits" value seems a little
worrisome to me; it relies on us using < rather than <= for when we
compare the actual failure rate to the threshold. Should I worry?
I guess your worry is someone might change it later and leave an
opportunity to break it? Let's just futureproof the comparison with a
comment rather than try to support negative failure rates and have all
sorts of signed vs unsigned comparison issues. Does that satisfy your
concern, or is it something deeper? I don't see any opportunities for
integer overflow, if that's what you mean?
--
Ticket URL: <https://trac.torproject.org/projects/tor/ticket/5458#comment:33>
Tor Bug Tracker & Wiki <https://trac.torproject.org/>
The Tor Project: anonymity online
More information about the tor-bugs
mailing list