Re: [tor-bugs] #7086 [Tor]: moria1 asserts on startup

["Tor Bug Tracker & Wiki" <torproject-admin@xxxxxxxxxxxxxx>]

#7086: moria1 asserts on startup
----------------------+-----------------------------------------------------
 Reporter:  arma      |          Owner:                    
     Type:  defect    |         Status:  new               
 Priority:  major     |      Milestone:  Tor: 0.2.4.x-final
Component:  Tor       |        Version:                    
 Keywords:  tor-auth  |         Parent:                    
   Points:            |   Actualpoints:                    
----------------------+-----------------------------------------------------

Comment(by andrea):

 Okay, I think I understand this one now.  There are three places new
 channels get created: from circuitbuild.c in
 channel_connect_for_circuit(), which then immediately calls
 command_setup_channel() on them; incoming connections from a channel
 listener get passed to the handler, command_handle_incoming_channel() in
 command.c, which does the same.  The other case is dirserv reachability
 testing dirserv_single_reachability_test() of dirserv.c, which doesn't
 install incoming cell handlers.  Whoops!  The channel gets a cell from the
 remote end that stays stuck forever and then closes with a non-empty
 queue.

 Obvious fix: make it call command_setup_channel() too so those channels
 route cells to the right place.  Query, though: if no handlers are
 available to drain the queue when closing the channel, is asserting in the
 case that it has leftover cells the right thing?

-- 
Ticket URL: <https://trac.torproject.org/projects/tor/ticket/7086#comment:7>
Tor Bug Tracker & Wiki <https://trac.torproject.org/>
The Tor Project: anonymity online
_______________________________________________
tor-bugs mailing list
tor-bugs@xxxxxxxxxxxxxxxxxxxx
https://lists.torproject.org/cgi-bin/mailman/listinfo/tor-bugs