[tor-bugs] #16825 [Core Tor/Tor]: client with explicit EntryNodes, no cached-* files never finds its entrynodes

Tor Bug Tracker & Wiki blackhole at torproject.org
Thu Jun 23 16:59:57 UTC 2016


#16825: client with explicit EntryNodes, no cached-* files never finds its
entrynodes
--------------------------------+------------------------------------
 Reporter:  starlight           |          Owner:
     Type:  defect              |         Status:  closed
 Priority:  Medium              |      Milestone:  Tor: 0.2.8.x-final
Component:  Core Tor/Tor        |        Version:  Tor: 0.2.6.10
 Severity:  Normal              |     Resolution:  fixed
 Keywords:  guards, 028-triage  |  Actual Points:
Parent ID:                      |         Points:  medium
 Reviewer:                      |        Sponsor:  SponsorU
--------------------------------+------------------------------------

Comment (by starlight):

 Experienced today a recurrence of a variation of this bug.  I believe
 #17147 also applies or is the same issue.

 Running 2.6.10 with the patch from this bug report applied in client-only
 bridge mode.  Image compiled with ASAN+UBSAN.  Uptime 72 days 12 hours.

 Got

 {{{
 Jun 23 06:23:57 Tor[]: new bridge descriptor 'omitted' (fresh): $[omitted]
 at [omitted]
 Jun 23 06:23:58 Tor[]: Delaying directory fetches: No running bridges
 }}}

 and Tor client refused to build any circuits.

 Put it in full debug logging and switched to a different bridge and it
 started working fine.  Then switched back to the normal bridge and the
 client daemon returned to the inoperative state.  Have debug log of this
 event as well.

 My belief is this bug / #17147 results from a logic flaw where micro-
 descriptors or full-descriptors are corrupted when converted to the
 internal binary structure representation.

 Killed the client daemon with a SEGV and successfully captured a fully
 gdb-usable core image.  Restarted and it came back running perfectly.

 If anyone has interest in trying to unravel the state of this core and/or
 examining the log file please let me know and the materials will be made
 available.  On the other hand I understand that the guard/descriptor logic
 will shortly be rewritten, so perhaps there is no point in chasing such a
 rarely exhibited bug.

--
Ticket URL: <https://trac.torproject.org/projects/tor/ticket/16825#comment:17>
Tor Bug Tracker & Wiki <https://trac.torproject.org/>
The Tor Project: anonymity online


More information about the tor-bugs mailing list