[Author Prev][Author Next][Thread Prev][Thread Next][Author Index][Thread Index]

Re: [tor-bugs] #1125 [Tor Client]: (*chp)->next assertion fail?



#1125: (*chp)->next assertion fail?
-------------------------+--------------------------------------------------
  Reporter:  arma        |       Owner:  nickm             
      Type:  defect      |      Status:  accepted          
  Priority:  major       |   Milestone:  Tor: 0.2.2.x-final
 Component:  Tor Client  |     Version:  0.2.2.5-alpha     
Resolution:  None        |    Keywords:                    
    Parent:              |  
-------------------------+--------------------------------------------------

Comment(by nickm):

 Robert Ransom solved this.  The problem is that buf_shrink_freelists() can
 call log_info, which can log to a controller, which can make the
 controller allocate buffer chunks, which can affect the freelists.

 {{{
 08:33 < rransom> See git://repo.or.cz/tor/rransom.git bug1125-master .
 08:34 < rransom> ( http://repo.or.cz/w/tor/rransom.git/ )
 08:40 < nickm> ok; my current inclination is to go with your moving of the
                log_info stuff, add the enable/disable_control_logging
 functions
                to wrap the whole function, and reinstate the log_warns.
 Maybe
                also, in master, rename the enable/disable_control_logging
                functions to something that implies fragile logging, or
                potentially reentrant logging, or callback-driven logging
 08:40 < nickm> then to backport to 0.2.1, and 0.2.2
 }}}

 It's not a huge amount of work, but let's do it so we can close this dumb
 thing and feel smart.

-- 
Ticket URL: <https://trac.torproject.org/projects/tor/ticket/1125#comment:20>
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