On Fri, Jul 01, 2016 at 03:33:07PM +0100, Daniel P. Berrange wrote:
On Fri, Jul 01, 2016 at 04:30:44PM +0200, Andrea Bolognani wrote:
audit is using cap-ng itself since version 2.6.1, and it gets to initialize the internal state of the cap-ng library before we call fork(). Because of that, our own use of cap-ng in the child process ends up attempting to change the capabilities of the parent process instead of the child process, which fails.
Calling capng_setpid() before using any other cap-ng API ensures we're altering the capabilities of the right process.
I really don't want to do any hack workaround in libvirt until we've got feedback from the cap-ng maintainers, as I think it is really their bug to fix.
FYI, I attached a patch for libcap-ng whic uses pthread_atfork to reset their global state https://bugzilla.redhat.com/show_bug.cgi?id=1352064 It works with my demo program, but haven't tested with libvirt yet 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 :|