Bug 886834 - [abrt] mutt-1.5.21-12.fc17: tunnel_socket_close: Process /usr/bin/mutt was killed by signal 11 (SIGSEGV)
Summary: [abrt] mutt-1.5.21-12.fc17: tunnel_socket_close: Process /usr/bin/mutt was ki...
Keywords:
Status: CLOSED INSUFFICIENT_DATA
Alias: None
Product: Fedora
Classification: Fedora
Component: mutt
Version: 17
Hardware: x86_64
OS: Unspecified
unspecified
unspecified
Target Milestone: ---
Assignee: Honza Horak
QA Contact: Fedora Extras Quality Assurance
URL:
Whiteboard: abrt_hash:14edb5e2d9731d325f1e28afea2...
Depends On:
Blocks:
TreeView+ depends on / blocked
 
Reported: 2012-12-13 10:09 UTC by Brian Kelly
Modified: 2013-06-27 10:04 UTC (History)
3 users (show)

Fixed In Version:
Clone Of:
Environment:
Last Closed: 2013-06-27 10:04:01 UTC
Type: ---
Embargoed:

Attachments (Terms of Use)
File: backtrace (16.47 KB, text/plain)
2012-12-13 10:09 UTC, Brian Kelly 		no flags Details
File: limits (1.29 KB, text/plain)
2012-12-13 10:09 UTC, Brian Kelly 		no flags Details
File: cgroup (133 bytes, text/plain)
2012-12-13 10:09 UTC, Brian Kelly 		no flags Details
File: smolt_data (2.47 KB, text/plain)
2012-12-13 10:09 UTC, Brian Kelly 		no flags Details
File: executable (13 bytes, text/plain)
2012-12-13 10:09 UTC, Brian Kelly 		no flags Details
File: maps (19.60 KB, text/plain)
2012-12-13 10:09 UTC, Brian Kelly 		no flags Details
File: dso_list (4.50 KB, text/plain)
2012-12-13 10:09 UTC, Brian Kelly 		no flags Details
File: proc_pid_status (921 bytes, text/plain)
2012-12-13 10:09 UTC, Brian Kelly 		no flags Details
File: var_log_messages (295 bytes, text/plain)
2012-12-13 10:09 UTC, Brian Kelly 		no flags Details
File: open_fds (384 bytes, text/plain)
2012-12-13 10:09 UTC, Brian Kelly 		no flags Details
View All

Description Brian Kelly 2012-12-13 10:09:13 UTC 
Description of problem:
Failed STARTTLS handshake
Briefly:
A connection was attempted over SMTP STARTTLS (briefly a message "gnutls_handshake: Error in the push function." is displayed, before "Could not negotiate TLS connection" and then "Segmentation fault (core dumped)" before crashing; the server advertises STARTTLS support, but actually achieving this is impossible due to tunnelling set up.



Version-Release number of selected component:
mutt-1.5.21-12.fc17

Additional info:
libreport version: 2.0.18
abrt_version:   2.0.18
backtrace_rating: 4
cmdline:        mutt
crash_function: tunnel_socket_close
kernel:         3.6.9-2.fc17.x86_64

truncated backtrace:
:Thread no. 1 (8 frames)
: #0 tunnel_socket_close at mutt_tunnel.c:134
: #1 mutt_socket_close at mutt_socket.c:81
: #2 mutt_smtp_send at smtp.c:311
: #3 send_message at send.c:1031
: #4 ci_send_message at send.c:1777
: #5 mutt_pager at pager.c:2536
: #6 mutt_display_message at commands.c:214
: #7 mutt_index_menu at curs_main.c:1201
Comment 1 Brian Kelly 2012-12-13 10:09:16 UTC 
Created attachment 662837 [details]
File: backtrace
Comment 2 Brian Kelly 2012-12-13 10:09:18 UTC 
Created attachment 662838 [details]
File: limits
Comment 3 Brian Kelly 2012-12-13 10:09:20 UTC 
Created attachment 662839 [details]
File: cgroup
Comment 4 Brian Kelly 2012-12-13 10:09:22 UTC 
Created attachment 662840 [details]
File: smolt_data
Comment 5 Brian Kelly 2012-12-13 10:09:24 UTC 
Created attachment 662841 [details]
File: executable
Comment 6 Brian Kelly 2012-12-13 10:09:27 UTC 
Created attachment 662842 [details]
File: maps
Comment 7 Brian Kelly 2012-12-13 10:09:29 UTC 
Created attachment 662843 [details]
File: dso_list
Comment 8 Brian Kelly 2012-12-13 10:09:31 UTC 
Created attachment 662844 [details]
File: proc_pid_status
Comment 9 Brian Kelly 2012-12-13 10:09:33 UTC 
Created attachment 662845 [details]
File: var_log_messages
Comment 10 Brian Kelly 2012-12-13 10:09:35 UTC 
Created attachment 662846 [details]
File: open_fds
Comment 11 Brian Kelly 2012-12-13 10:20:33 UTC 
The problem is essentially that mutt segfaults rather than failing gracefully (e.g. issuing a warning message and treating the message attempted to send as a postponed messaged)

The impact of this is very minor, and resultant of a bespoke set up.
Comment 12 Honza Horak 2012-12-13 11:35:15 UTC 
Thanks for the report, is that failure reproducible? I tried to reproduce it but I'm not successful.

If it is, I'm wondering how conn->sockdata (mutt_tunnel.c:131) could be NULL pointer. It could be closed twice, but I don't have a clue how that could happen.

If you could run mutt under GDB with checkpoint to tunnel_socket_close function and see if this function is called only once (at time it fails) or it is called twice -- then a backtrace of the first call would be really interesting.
Comment 13 Honza Horak 2013-06-27 10:04:01 UTC 
Closing since I'm not able to do anything without further information.
Note You need to log in before you can comment on or make changes to this bug.