Arch Linux

Please read this before reporting a bug:
https://wiki.archlinux.org/title/Bug_reporting_guidelines

Do NOT report bugs when a package is just outdated, or it is in the AUR. Use the 'flag out of date' link on the package page, or the Mailing List.

REPEAT: Do NOT report bugs for outdated packages!
Tasklist

FS#15387 - [ntop] 3.3.10-1 segfaults and aborts

Attached to Project: Arch Linux
Opened by Jan Alexander Steffens (heftig) - Sunday, 05 July 2009, 18:06 GMT
Last edited by Jan de Groot (JGC) - Monday, 21 December 2009, 14:22 GMT
Task Type Bug Report
Category Upstream Bugs
Status Closed
Assigned To No-one
Architecture All
Severity Medium
Priority Normal
Reported Version
Due in Version Undecided
Due Date Undecided
Percent Complete 100%
Votes 1
Private No

Details

Description:
After a few seconds to minutes of being started, ntop crashes and quits, either with a SIGSEGV or a SIGABRT.

It seems this problem was introduced by the last update to ntop. ntop 3.3.9-2 seems to work fine.

Additional info:
I use i686 testing.

extra/ntop 3.3.10-1
core/libpcap 1.0.0-1
testing/gd 2.0.35-2
extra/glib 1.2.10-7
extra/libxml2 2.7.3-2
core/openssl 0.9.8k-2
core/gdbm 1.8.3-6
extra/rrdtool 1.3.8-1
core/pcre 7.9-1
extra/geoip 1.4.6-1
extra/lua 5.1.4-4
core/glibc 2.10.1-2

Steps to reproduce:
1. Start ntop
2. Wait a while
   crash1.txt (11.7 KiB)
   crash2.txt (11.9 KiB)
   crash3.txt (11.1 KiB)
This task depends upon

Closed by  Jan de Groot (JGC)
Monday, 21 December 2009, 14:22 GMT
Reason for closing:  Fixed
Additional comments about closing:  libevent bug, fixed in 1.4.12.
Comment by Roman Kyrylych (Romashka) - Sunday, 05 July 2009, 18:31 GMT
@Allan: may be libevent bug, so I added you too.
Comment by Allan McRae (Allan) - Friday, 10 July 2009, 04:47 GMT
Well, I don't actually maintain libevent, but given no-one else seems to...

Try with the updated libevent-1.4.11-1 from [testing]. It is safe to install just that package from the [testing] repo.
Comment by Jan Alexander Steffens (heftig) - Friday, 10 July 2009, 18:30 GMT
Still crashes with libevent 1.4.11-1.

#0 0xb768a273 in nameserver_read () from /usr/lib/libevent-1.4.so.2
#1 0xb767f9a8 in event_base_loop () from /usr/lib/libevent-1.4.so.2
#2 0xb767fb19 in event_loop () from /usr/lib/libevent-1.4.so.2
#3 0x0804b181 in main ()
Comment by Allan McRae (Allan) - Saturday, 11 July 2009, 09:57 GMT
@heftig: can you test compiling the current ntop and the version that works for you against the current libevent. That will let us know whether this is an ntop or libevent issue.
Comment by Jan Alexander Steffens (heftig) - Saturday, 11 July 2009, 16:20 GMT
Rebuilt 3.3.10 still crashes:

Program received signal SIGSEGV, Segmentation fault.
[Switching to Thread 0xb6cff6c0 (LWP 5759)]
0xb753b273 in nameserver_read () from /usr/lib/libevent-1.4.so.2
(gdb) bt
#0 0xb753b273 in nameserver_read () from /usr/lib/libevent-1.4.so.2
#1 0xb75309a8 in event_base_loop () from /usr/lib/libevent-1.4.so.2
#2 0xb7530b19 in event_loop () from /usr/lib/libevent-1.4.so.2
#3 0x0804b189 in main ()

Program received signal SIGSEGV, Segmentation fault.
[Switching to Thread 0xb6db76c0 (LWP 5795)]
0xb75e7d59 in event_queue_insert () from /usr/lib/libevent-1.4.so.2
(gdb) bt
#0 0xb75e7d59 in event_queue_insert () from /usr/lib/libevent-1.4.so.2
#1 0xb75e881e in event_base_loop () from /usr/lib/libevent-1.4.so.2
#2 0xb75e8b19 in event_loop () from /usr/lib/libevent-1.4.so.2
#3 0x0804b189 in main ()

Rebuilt 3.3.9 crashes as well, but differently:

Program received signal SIGSEGV, Segmentation fault.
[Switching to Thread 0xadd64b70 (LWP 7730)]
0xb7712e55 in handleHTTPSession () from /usr/lib/libntop-3.3.9.so
(gdb) bt
#0 0xb7712e55 in handleHTTPSession () from /usr/lib/libntop-3.3.9.so
#1 0xb7716cee in handleTCPSession () from /usr/lib/libntop-3.3.9.so
#2 0xb7718e86 in handleSession () from /usr/lib/libntop-3.3.9.so
#3 0xb7700bab in processIpPkt () from /usr/lib/libntop-3.3.9.so
#4 0xb7702c4e in processPacket () from /usr/lib/libntop-3.3.9.so
#5 0xb7708328 in queuePacket () from /usr/lib/libntop-3.3.9.so
#6 0xb7464916 in pcap_read_linux_mmap () from /usr/lib/libpcap.so.1
#7 0xb7468797 in pcap_dispatch () from /usr/lib/libpcap.so.1
#8 0xb76f3afd in pcapDispatch () from /usr/lib/libntop-3.3.9.so
#9 0xb73d368c in start_thread () from /lib/libpthread.so.0
#10 0xb7355e2e in clone () from /lib/libc.so.6

Program received signal SIGSEGV, Segmentation fault.
[Switching to Thread 0xae67fb70 (LWP 7786)]
0xb782ce55 in handleHTTPSession () from /usr/lib/libntop-3.3.9.so
(gdb) bt
#0 0xb782ce55 in handleHTTPSession () from /usr/lib/libntop-3.3.9.so
#1 0xb7830cee in handleTCPSession () from /usr/lib/libntop-3.3.9.so
#2 0xb7832e86 in handleSession () from /usr/lib/libntop-3.3.9.so
#3 0xb781abab in processIpPkt () from /usr/lib/libntop-3.3.9.so
#4 0xb781e666 in processPacket () from /usr/lib/libntop-3.3.9.so
#5 0xb7822328 in queuePacket () from /usr/lib/libntop-3.3.9.so
#6 0xb757e916 in pcap_read_linux_mmap () from /usr/lib/libpcap.so.1
#7 0xb7582797 in pcap_dispatch () from /usr/lib/libpcap.so.1
#8 0xb780dafd in pcapDispatch () from /usr/lib/libntop-3.3.9.so
#9 0xb74ed68c in start_thread () from /lib/libpthread.so.0
#10 0xb746fe2e in clone () from /lib/libc.so.6
Comment by Allan McRae (Allan) - Sunday, 12 July 2009, 02:13 GMT
Strange. Could try with a downgraded libevent. I suggest version 1.4.9 as this seems when ntop last worked...
Comment by Jan Alexander Steffens (heftig) - Tuesday, 14 July 2009, 18:17 GMT
Rebuilt 3.3.10 with libevent 1.4.9 segfaults:
#0 0xb750d210 in transaction_id_pick () from /usr/lib/libevent-1.4.so.2
#1 0xb750e12c in evdns_requests_pump_waiting_queue ()
from /usr/lib/libevent-1.4.so.2
#2 0xb7510da5 in reply_handle () from /usr/lib/libevent-1.4.so.2
#3 0xb7511594 in nameserver_read () from /usr/lib/libevent-1.4.so.2
#4 0xb75068f8 in event_base_loop () from /usr/lib/libevent-1.4.so.2
#5 0xb7506a69 in event_loop () from /usr/lib/libevent-1.4.so.2
#6 0x0804b189 in main ()

#0 0xb755a210 in transaction_id_pick () from /usr/lib/libevent-1.4.so.2
#1 0xb755daa5 in request_new () from /usr/lib/libevent-1.4.so.2
#2 0xb755e98c in evdns_resolve_reverse () from /usr/lib/libevent-1.4.so.2
#3 0xb776c479 in ipaddr2str () from /usr/lib/libntop-3.3.10.so
#4 0xb77741da in _lookupHost () from /usr/lib/libntop-3.3.10.so
#5 0xb77837ff in processIpPkt () from /usr/lib/libntop-3.3.10.so
#6 0xb778c7ba in processPacket () from /usr/lib/libntop-3.3.10.so
#7 0xb77904e8 in queuePacket () from /usr/lib/libntop-3.3.10.so
#8 0xb74d7916 in pcap_read_linux_mmap () from /usr/lib/libpcap.so.1
#9 0xb74db797 in pcap_dispatch () from /usr/lib/libpcap.so.1
#10 0xb777be0d in pcapDispatch () from /usr/lib/libntop-3.3.10.so
#11 0xb744668c in start_thread () from /lib/libpthread.so.0
#12 0xb73c7e2e in clone () from /lib/libc.so.6

Rebuilt 3.3.9 with libevent 1.4.9 seems to be stable.
Comment by Jan Alexander Steffens (heftig) - Tuesday, 14 July 2009, 18:18 GMT
Never mind, 3.3.9 segfaults as well:
#0 0xb784de55 in handleHTTPSession () from /usr/lib/libntop-3.3.9.so
#1 0xb7851cee in handleTCPSession () from /usr/lib/libntop-3.3.9.so
#2 0xb7853e86 in handleSession () from /usr/lib/libntop-3.3.9.so
#3 0xb783bbab in processIpPkt () from /usr/lib/libntop-3.3.9.so
#4 0xb783f666 in processPacket () from /usr/lib/libntop-3.3.9.so
#5 0xb7843328 in queuePacket () from /usr/lib/libntop-3.3.9.so
#6 0xb759f916 in pcap_read_linux_mmap () from /usr/lib/libpcap.so.1
#7 0xb75a3797 in pcap_dispatch () from /usr/lib/libpcap.so.1
#8 0xb782eafd in pcapDispatch () from /usr/lib/libntop-3.3.9.so
#9 0xb750e68c in start_thread () from /lib/libpthread.so.0
#10 0xb7490e2e in clone () from /lib/libc.so.6
Comment by Allan McRae (Allan) - Wednesday, 15 July 2009, 02:10 GMT
As far as I can see, this is an ntop bug and not libevent so removing myself from assignees.

I suggest you follow this up with ntop upstream. I had a quick search and other people seem to report crashes too but there was very little follow-up.
Comment by Laszlo Papp (djszapi) - Sunday, 06 December 2009, 18:47 GMT
Status of it ?
Comment by Thomas Dziedzic (tomd123) - Sunday, 06 December 2009, 19:10 GMT
I can't reproduce the segfault. Using i686 without testing.
Comment by Laszlo Papp (djszapi) - Sunday, 06 December 2009, 19:27 GMT
The daemon works here too, but if I start it with simply 'ntop' command, and I send a SIGINT I get segfault now.

* ntop 3.3.10-1
* libevent 1.4.11-1
Comment by Jan Alexander Steffens (heftig) - Monday, 21 December 2009, 12:58 GMT
Can't reproduce anymore with libevent 1.4.12-1.
Reappears when downgrading to 1.4.11-1.

Seems it was a libevent bug afterall. Changelog of 1.4.12 mentions:
- Fix an obscure timing-dependent, allocator-dependent crash in the evdns code.

Loading...