On 08/25/2014 01:39 AM, Chris wrote: > Hello, > > I encountered segfaults with libvirt + libxl with this specific patch : > http://libvirt.org/git/?p=libvirt.git;a=commitdiff;h=b55cc5f4e31b488c4f9c3c8470c992c1f8f5d09c > > My Libvirt/Xen environment (on a slackware64-14.1) : > > virsh # version > Compiled against library: libvirt 1.2.6 > Using library: libvirt 1.2.6 > Using API: Xen 1.2.6 > Running hypervisor: Xen 4.4.0 > > > The VM starts fine with "virsh start" (cf attached libxl log), the VNC process is listening as requested on 0.0.0.0 but whenever I open a new virsh cli and send a command (e.g. "list"), libvirtd segfaults. I attached a gdb trace to this mail. > If I rebuild libvirt without the patch mentioned earlier, no more segfaults, but of course the VNC process is listening on 127.0.0.1 instead of requested 0.0.0.0 when I start the VM. > The trace shows SIGSEGV in malloc_consolidate, which usually means we touched some memory we shouldn't have and overwritten malloc's internal data. Could you run libvirtd under valgrind and see if it shows any invalid reads/writes? Jan
Attachment:
signature.asc
Description: OpenPGP digital signature
-- libvir-list mailing list libvir-list@xxxxxxxxxx https://www.redhat.com/mailman/listinfo/libvir-list