[tor-bugs] #5072 [Obfsproxy]: obfsproxy assertion failure src/network.c:726: down->buffer == bev

Tor Bug Tracker & Wiki torproject-admin at torproject.org
Fri Feb 10 21:07:55 UTC 2012


#5072: obfsproxy assertion failure src/network.c:726: down->buffer == bev
-----------------------+----------------------------------------------------
 Reporter:  arma       |          Owner:  asn
     Type:  defect     |         Status:  new
 Priority:  major      |      Milestone:     
Component:  Obfsproxy  |        Version:     
 Keywords:             |         Parent:     
   Points:             |   Actualpoints:     
-----------------------+----------------------------------------------------
Changes (by asn):

  * priority:  normal => major


Comment:

 This log:
 {{{
 [warn] downstream_read_cb: bev: '0x1fec380' down->buffer:'0x2045660',
 down: '0x20459d0'
 [warn] bev == down->circuit->upstream->buffer: 0
 bev == down->circuit->downstream->buffer: 1
 down->buffer == down->circuit->upstream->buffer: 1
 down->buffer == down->circuit->downstream->buffer: 0
 }}}

 tells us that the `downstream_read_cb()` callback was triggered by the
 downstream bufferevent of a circuit (`0x1fec380`), but the callback
 argument was the '''upstream''' connection (`0x20459d0`). The callback
 argument of `downstream_read_cb()` must be the downstream connection,
 and that's why we assert out.

 Looking at the whole log, we see that `0x1fec380` is the downstream
 bufferevent of a connection established some lines above:
 {{{
 [warn] simple_server_listener_cb: conn: '0x2030120' conn->buffer:
 '0x1fec380'
 [warn] pending_conn_cb: circ->downstream: '0x2030120'
 circ->downstream->buffer: '0x1fec380'
 }}}

 I pushed another commit to my `bug5072` branch which might or might not
 help debug this issue. I'm also CCing nickm in case he has any good
 ideas.

-- 
Ticket URL: <https://trac.torproject.org/projects/tor/ticket/5072#comment:7>
Tor Bug Tracker & Wiki <https://trac.torproject.org/>
The Tor Project: anonymity online


More information about the tor-bugs mailing list