GNOME Bugzilla – Bug 150317
Segfault after remote user transfers call.
Last modified: 2004-12-22 21:47:04 UTC
100% repeatable sequence.... Start a audio/video connection (56k spec settings, but DSL does the same) with a local lan user, who then transfers the call to the outside world. NAT forwarding on firewall is to the local machine. Transfer is to 'Johnny Strom' test stream. Both local and remote (other lan user) general history report transfer as successful - shows closing audio/video RX/TX, and opening audio/video RX/TX codecs for new connection. Console from which GM was run echos (sequence number changes):- Xlib: unexpected async reply (sequence 0x1a23)! The stats line shows a briefly very silly number for audio bandwidth and then settles to show normal (and changing) audio TX/RX rates, but no audio is heard. Video TX is 0, RX is normal (and changing) but no incoming video is displayed (instead I get the static gnomemeeting logo). Connection icon (next to callto:) is showing unconnected. First attempt to do anything with any of gnomemeeting windows cause gnome bug report to start. Back trace is as follows: --- Backtrace was generated from '/usr/bin/gnomemeeting' (no debugging symbols found)...Using host libthread_db library "/lib/libthread_db.so.1". (no debugging symbols found)...(no debugging symbols found)...(no debugging symbols found)...(no debugging symbols found)...(no debugging symbols found)...(no debugging symbols found)...(no debugging symbols found)...(no debugging symbols found)...(no debugging symbols found)...(no debugging symbols found)...(no debugging symbols found)...(no debugging symbols found)...(no debugging symbols found)...(no debugging symbols found)...(no debugging symbols found)...(no debugging symbols found)...(no debugging symbols found)...(no debugging symbols found)...(no debugging symbols found)...(no debugging symbols found)...(no debugging symbols found)...(no debugging symbols found)...(no debugging symbols found)...(no debugging symbols found)...(no debugging symbols found)...(no debugging symbols found)...(no debugging symbols found)...(no debugging symbols found)...(no debugging symbols found)...(no debugging symbols found)...(no debugging symbols found)...(no debugging symbols found)...(no debugging symbols found)...(no debugging symbols found)...[Thread debugging using libthread_db enabled] [New Thread 16384 (LWP 1622)] [New Thread 32769 (LWP 1627)] [New Thread 16386 (LWP 1628)] [New Thread 32771 (LWP 1630)] [New Thread 49156 (LWP 1631)] [New Thread 65541 (LWP 1632)] [New Thread 393222 (LWP 1896)] [New Thread 262151 (LWP 1869)] [New Thread 294920 (LWP 1871)] [New Thread 311305 (LWP 1872)] [New Thread 327690 (LWP 1873)] [New Thread 344075 (LWP 1874)] [New Thread 360460 (LWP 1878)] [New Thread 245773 (LWP 1868)] (no debugging symbols found)...(no debugging symbols found)...(no debugging symbols found)...(no debugging symbols found)...(no debugging symbols found)...(no debugging symbols found)...(no debugging symbols found)...(no debugging symbols found)...(no debugging symbols found)...(no debugging symbols found)...(no debugging symbols found)...(no debugging symbols found)...(no debugging symbols found)...(no debugging symbols found)...(no debugging symbols found)...(no debugging symbols found)...(no debugging symbols found)...(no debugging symbols found)...(no debugging symbols found)...(no debugging symbols found)...(no debugging symbols found)...(no debugging symbols found)...(no debugging symbols found)...(no debugging symbols found)...(no debugging symbols found)...(no debugging symbols found)...(no debugging symbols found)...(no debugging symbols found)...(no debugging symbols found)...(no debugging symbols found)...(no debugging symbols found)...(no debugging symbols found)...(no debugging symbols found)...(no debugging symbols found)...(no debugging symbols found)...(no debugging symbols found)...(no debugging symbols found)...(no debugging symbols found)...(no debugging symbols found)...(no debugging symbols found)...(no debugging symbols found)...(no debugging symbols found)...(no debugging symbols found)...(no debugging symbols found)...(no debugging symbols found)...(no debugging symbols found)...(no debugging symbols found)...(no debugging symbols found)...(no debugging symbols found)...(no debugging symbols found)...(no debugging symbols found)...(no debugging symbols found)...(no debugging symbols found)...(no debugging symbols found)...(no debugging symbols found)...(no debugging symbols found)...(no debugging symbols found)...(no debugging symbols found)...0x4154cc0b in waitpid () from /lib/libpthread.so.0
+ Trace 49382
Thread 13 (Thread 360460 (LWP 1878))
Is it reproducable if you run gnomemeeting without your patches? They are not protected by the GDK lock, so that could explain the XLib async reply.
This segfault occurs with the "unstable" Debian package "gnomemeeting-1.0.2-5" and is repeatable today. The remote machine is also Debian Unstable, but is running on an underpowered laptop. As stated I'm running on a NAT'ed cable connection (Linksys Cable Route BEFSR41). This call is changing from a not NAT'ed (remote user on local LAN) to NAT'ed (remote user in 'real' world), this may have something to do with the segfault. Also I tried a couple of transfers to others on the Seconix ILS server, these apeared to work. But transfers to 'Johnny Strom' still fail - I note that he's using GM-1.1.0!!! Anyone else with 1.1.0 mind if I test with them????
The backtrace seems to indicate a crash in the H.261 codec (video) either encoding or decoding, which would be a bug in openh323 and more precisely in the VIC encoder. If you want to try with other 1.1.0 and debug this, youo should go on irc.gnome.org,#gnomemeeting Thanks! The only difference between johnny and the rest is that he is in auto-answer mode. That could cause the crash. So perhaps you can try with another person who is in auto-answer?
This does appear to be related to the having auto-answer (Free to Chat) enabled on the target of the transfer. i.e. I get a segfault whilst transfering to another user who was 'Free to Chat'. The problem occurs on transfers to a (third) machine on the local LAN and to a user in the outside world.
I think that should be fixed. Please test and report back. CVS packs are at snapshots.seconix.com, they are rebuilding every night.