[tor-bugs] #19522 [Core Tor/Tor]: HS intro circuit retry logic fails when network interface is down
Tor Bug Tracker & Wiki
blackhole at torproject.org
Mon Aug 1 15:21:17 UTC 2016
#19522: HS intro circuit retry logic fails when network interface is down
--------------------------+--------------------------------
Reporter: asn | Owner:
Type: defect | Status: needs_revision
Priority: Medium | Milestone: Tor: 0.2.???
Component: Core Tor/Tor | Version:
Severity: Normal | Resolution:
Keywords: tor-hs | Actual Points:
Parent ID: #16387 | Points: 1.5
Reviewer: | Sponsor: SponsorR-can
--------------------------+--------------------------------
Changes (by dgoulet):
* status: needs_review => needs_revision
Comment:
Replying to [comment:4 asn]:
> Thanks for the testing timonh.
>
> I pushed a branch `bug19522` in my repo, so that people can test
further:
> https://gitweb.torproject.org/user/asn/tor.git
>
> FWIW, I read some code to verify the assumption of comment:1 and it
seems to be accurate. But more digging and testing is required to get
better confidence, as those functions were quite hairy.
The fix seems logical but adds some _powerful_ assumptions that
`rend_service_launch_establish_intro()` failure is always due to some
"local issues". As far as I can tell, it seems to be the case that
basically if we can't launch a circuit it's because we just can't get
packet out of the wire or we simply don't have enough information to be
able to do so (consensus for instance).
If you could, adding a comment to `rend_service_launch_establish_intro()`
documenting the returned value and in this case the `-1` being that it
doesn't mean the intro point is _bad_ per-se but rather the failing of
launching a circuit is due to "local reachability" or not enough
information to continue issues. This is _very_ important that we don't
mess it up I believe because retrying over and over a bad intro point is
also a bad thing.
--
Ticket URL: <https://trac.torproject.org/projects/tor/ticket/19522#comment:5>
Tor Bug Tracker & Wiki <https://trac.torproject.org/>
The Tor Project: anonymity online
More information about the tor-bugs
mailing list