On Fri, Feb 08, 2013 at 09:08:01PM +0800, Osier Yang wrote:
qemuProcessStart invokes qemuProcessStop when fails, to avoid
removing hash entry which belongs to other domain(s), this introduces
a new virBitmapPtr argument for qemuProcessStop. And qemuProcessStart
sets the bit for the disk only if it's successfully added into the
hash table. Thus if the argument is provided for qemuProcessStop, it
can't remove the hash entry belongs to other domain(s).
I find this approach rather dubious - IMHO it is a sign that you're
not recording enough information in the shared disk hash. eg perhaps
the hash ought to be recording the UUID of each VM that is holding
a reference. That way you're protected from qemuProcessStop() trying
todo something wrong.
Daniel
--
|:
http://berrange.com -o-
http://www.flickr.com/photos/dberrange/ :|
|:
http://libvirt.org -o-
http://virt-manager.org :|
|:
http://autobuild.org -o-
http://search.cpan.org/~danberr/ :|
|:
http://entangle-photo.org -o-
http://live.gnome.org/gtk-vnc :|