
On Wed, Apr 02, 2014 at 07:00:09AM -0600, Eric Blake wrote:
On 04/02/2014 06:44 AM, Ján Tomko wrote:
Every security driver has domainGenSecurityLabel defined.
As currently written. But Dan wrote the manager to be flexible to future drivers that omit obvious functions.
This patch makes sense for silencing Coverity, but I think it is incomplete unless you also fix the registration with the manager to forcefully require that all drivers supply callback functions that we are going to blindly assume exist, rather than the current status quo of allowing a driver to omit callbacks even if none of them do. That is, virSecurityManagerNewDriver() should be taught to require drv->domainGenSecurityLabel is non-NULL.
Further this change makes domainGenSecurityLabel an exception to the rule now, since all others have checks for NULL. I don't object to removing the checks for NULL and mandating non-NULL, but we should do it for all the callbacks, or none of the callbacks. Regards, 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 :|