[tor-bugs] #3000 [Tor Client]: Clients do not clear HS descriptor cache on SIGNAL NEWNYM

Tor Bug Tracker & Wiki torproject-admin at torproject.org
Tue Apr 26 21:57:57 UTC 2011


#3000: Clients do not clear HS descriptor cache on SIGNAL NEWNYM
------------------------+---------------------------------------------------
 Reporter:  rransom     |          Owner:  rransom           
     Type:  defect      |         Status:  needs_review      
 Priority:  major       |      Milestone:  Tor: 0.2.1.x-final
Component:  Tor Client  |        Version:                    
 Keywords:              |         Parent:                    
   Points:              |   Actualpoints:                    
------------------------+---------------------------------------------------

Comment(by rransom):

 Replying to [comment:3 nickm]:
 > Reviewing bug3000-021, it looks superfically okay, meaning that if there
 are any issues, it's not issues in the code but rather in the code's
 ramifications in other parts of Tor.  We should meditate on that.
 >
 > Already noted to you in IRC: on e05cdf9b8dfea30 , I think you need a
 "return -1" after tor_fragile_assert(): tor_fragile_assert() can be
 defined to not  assert at all, in which case you'd run off the bottom of
 the function.
 >
 > What have you done to test this out so far?

 I compiled and ran the tree as of ‘Ignore SIGNAL NEWNYM on relay-only Tor
 instances’ in order to make sure I didn't get the conditionals in that
 patch or ‘Don't allow v0 HS auths to act as clients’ backwards, and then
 send SIGNAL NEWNYM while trying to connect to
 [http://56apzofkmsmgb3yr.onion/ the archive.tpo HS] to see what I broke by
 nuking the HS descriptor cache.

 Then, many changes later, I tested the tree as of ‘Fix bug 1930’ by
 repeatedly sending SIGNAL NEWNYM while repeatedly trying to connect to the
 archive.tpo HS again.  Nothing broke, and I was eventually able to view
 the HS with Firefox 3.6 without Polipo or the Firefox SOCKS timeout patch.
 I am still running that build as my Tor client.

 Then, to test the merges, I tried compiling the source files after
 performing each merge.  (I was unable to link Tor on the maint-0.2.2
 branches due to idiocy in the FreeBSD libevent2 port.)

-- 
Ticket URL: <https://trac.torproject.org/projects/tor/ticket/3000#comment:4>
Tor Bug Tracker & Wiki <https://trac.torproject.org/>
The Tor Project: anonymity online


More information about the tor-bugs mailing list