GNOME Bugzilla – Bug 678921
gvncviewer crashes when remote side restarts alsa
Last modified: 2013-01-31 10:43:23 UTC
This bug was originally reported downstream as: https://bugzilla.novell.com/show_bug.cgi?id=768568 ### when the remote side does rcalsasound restart gvncviewer segfaults Program received signal SIGSEGV, Segmentation fault. __memcpy_sse2 () at ../sysdeps/x86_64/memcpy.S:202 202 ../sysdeps/x86_64/memcpy.S: Datei oder Verzeichnis nicht gefunden. (gdb) bt
+ Trace 230441
$2 = (VncAudioSample *) 0x0 ###
Can you tell me what exact version of GTK-VNC you are using - without this info I cannot match up the line numbers in the stack trace to the code. Also, it would be helpful if you can re-run gvncviewer, passing the '--gtk-vnc-debug' argument & capture all output that is generated until the crash occurs & upload it
gtk-vnc version 0.5.0 remote side is running As for the extended output: I'll ask the original reporter to provide this information.
(In reply to comment #2) > gtk-vnc version 0.5.0 > remote side is running remote side is running openSUSE Factory in qemu-kvm -vnc :0 -soundhw ac97
The user reported: ### am I doing something wrong, that this does not give me more output? > rpm -q gtk-vnc-tools gtk-vnc-tools-0.5.0-2.1.x86_64 > gvncviewer --gtk-vnc-debug ... Connected to server Remote desktop size changed to 640x480 Connection initialized Assertion 'p' failed at pulse/simple.c:275, function pa_simple_write(). Aborting. Aborted this is bnc#766294 which might be related (https://bugzilla.novell.com/show_bug.cgi?id=766294) on 12.1 with gvncviewer-0.4.x it produced plenty output, but did not run into problems
Oh, what fun - 3 bugs in one. I suspect the reason why --gtk-vnc-debug did not produce output is due to the reporter having a GLib >= 2.32, which changed the default logging settings :-( I did a workaround for that here: commit 1435c6c8d117ebd06abfacc842ee4dd2ac94d71d Author: Daniel P. Berrange <berrange@redhat.com> Date: Wed Apr 25 10:16:36 2012 +0100 Ensure debugging works on GLib >= 2.32 Signed-off-by: Daniel P. Berrange <berrange@redhat.com> The assertion in pulse/simple.h looks like another bug in our audio support, but I believe it is separate from this issue, so please do file another gnome.org bug about that assertion, so we can track it separately.
(In reply to comment #5) > Oh, what fun - 3 bugs in one. Yeah.. sorry for that... and in plus I'm only proxying.. > I suspect the reason why --gtk-vnc-debug did not produce output is due to the > reporter having a GLib >= 2.32, which changed the default logging settings :-( Indeed, this system runs with glib 2.32.1 (openSUSE Factory / 12.2) I created a package for the user with your workaround included, so we can get more debug info... > The assertion in pulse/simple.h looks like another bug in our audio support, > but I believe it is separate from this issue, so please do file another > gnome.org bug about that assertion, so we can track it separately. Will open an additional bug, referencing the downstream bug.
(In reply to comment #6) > > The assertion in pulse/simple.h looks like another bug in our audio support, > > but I believe it is separate from this issue, so please do file another > > gnome.org bug about that assertion, so we can track it separately. > > Will open an additional bug, referencing the downstream bug. done, see bug 679283.
The crash bug is fixed and the other issues mentioned here are also addressed, so closing this