On Wed, Aug 12, 2009 at 05:27:35PM -0700, Kaitlin Rupert wrote:
Hi,
I'm trying to track down a seg fault I'm seeing with the
libvirt-0.7.0-3.fc12.x86_64 rawhide version of libvirt.
I'm having trouble finding the exact set of steps, as I see the behavior
when interacting with the libvirt-cim providers.
Here's the stack trace I was able to get. Looks like the seg fault
actually occurs in netcf, which I haven't looked at much. Has anyone
seen this?
Program received signal SIGSEGV, Segmentation fault.
0x00007fffe4004383 in ?? ()
(gdb) bt
#0 0x00007fffe4004383 in ?? ()
#1 0x00007ffff6a5cbfc in xmlHashScanFull__internal_alias (
table=0x7fffe4005f50, f=<value optimized out>, data=<value optimized
out>)
at hash.c:849
#2 0x00007ffff6a5cc4c in xmlHashScan__internal_alias
(table=0x7fffe400ef10,
f=<value optimized out>, data=<value optimized out>) at hash.c:818
#3 0x00007ffff6ff5104 in xsltShutdownExts () from
/usr/lib64/libxslt.so.1
#4 0x00007ffff6fe3e3d in xsltFreeStylesheet () from
/usr/lib64/libxslt.so.1
#5 0x00007ffff7452961 in ?? () from /usr/lib64/libnetcf.so.1
#6 0x00007ffff74501d1 in ncf_close () from /usr/lib64/libnetcf.so.1
OMG ! You have no idea where it comes from ...
I'm afraid it's related to
http://bugzilla.gnome.org/show_bug.cgi?id=305913
a global variable, used in libxslt, itself used in netcf, itself used by
libvirt ....
There is probably a workaround possible at the netcf level forcing
locking when XSLT transformations are run to avoid the multithread issue
but the real fix should be in libxslt, I'm working on libxml2/libxslt
bug fixes this week and will try to get this fixed in the next libxslt
release, hopefully by next week ... I will keep you posted !
Daniel
--
Daniel Veillard | libxml Gnome XML XSLT toolkit
http://xmlsoft.org/
daniel(a)veillard.com | Rpmfind RPM search engine
http://rpmfind.net/
http://veillard.com/ | virtualization library
http://libvirt.org/