[tor-bugs] #5263 [Tor Relay]: Busy/infinite Libevent loops

[Tor Bug Tracker & Wiki]

#5263: Busy/infinite Libevent loops
------------------------+---------------------------------------------------
 Reporter:  robgjansen  |          Owner:                    
     Type:  defect      |         Status:  needs_review      
 Priority:  major       |      Milestone:  Tor: 0.2.3.x-final
Component:  Tor Relay   |        Version:                    
 Keywords:              |         Parent:                    
   Points:              |   Actualpoints:                    
------------------------+---------------------------------------------------

Comment(by robgjansen):

 Replying to [comment:19 nickm]:
 > Hm. What I said above about making the change in connection_read_to_buf:
 I think that's wrong; we don't want to stop writing every time we get a
 wantread: we only want to stop writing in the case you mentioned, where we
 try to read because connection_handle_read() failed, and we get a
 "wantread".  But we don't want to do that in just any case where
 read_result <= 0 : only in the case where it actually said wantread.
 Also, we need some way to make sure that once the operation eventually
 succeeds, we enable writing again, unless there already is such a thing.
 >

 Maybe some type of 'event source type' code or something would be helpful
 here.

 I merged master into a local copy of your bug5263 branch. The
 OR_CONN_STATE_TLS_SERVER_RENEGOTIATING and other issues no longer appear.
 I'm waiting for the bug log message. I recall the bug appearing after a
 few hours of an experiment.

-- 
Ticket URL: <https://trac.torproject.org/projects/tor/ticket/5263#comment:21>
Tor Bug Tracker & Wiki <https://trac.torproject.org/>
The Tor Project: anonymity online