On 07/04/2014 05:29 AM, Peter Krempa wrote:
Refactor the existing code to allow re-using it for the per-image label restore too. --- src/security/security_dac.c | 59 ++++++++++++++++++++++++--------------------- 1 file changed, 31 insertions(+), 28 deletions(-)
- secdef = virDomainDefGetSecurityLabelDef(def, SECURITY_DAC_NAME); + /* Don't restore labels on readoly/shared disks, because other VMs may + * still be accessing these Alternatively we could iterate over all running
Pre-existing typo, but s/these/these./
+ * domains and try to figure out if it is in use, but this would not work + * for clustered filesystems, since we can't see running VMs using the file + * on other nodes Safest bet is thus to skip the restore step.
s/nodes/nodes./
- /* If we have a shared FS & doing migrated, we must not - * change ownership, because that kills access on the - * destination host which is sub-optimal for the guest - * VM's I/O attempts :-) + /* If we have a shared FS & doing migrated, we must not change ownership,
s/doing migrated/are doing migration/
+ * because that kills access on the destination host which is sub-optimal + * for the guest VM's I/O attempts :-)
Technically, changing SELinux labels kills guests (that takes effect immediately); but per POSIX, calling chmod() should not (as chmod only affects future open() calls, but not already-open fds). But then you throw in non-POSIX NFS, and there we have a case where chmod() can kill access to open fds (naughty NFS). So this comment is true, but only because of botched file systems :) ACK with comment cleanups. -- Eric Blake eblake redhat com +1-919-301-3266 Libvirt virtualization library http://libvirt.org