storage: Handle failure from refreshVol CVE-2015-5247-3
authorJohn Ferlan <jferlan@redhat.com>
Mon, 24 Aug 2015 16:38:13 +0000 (12:38 -0400)
committerJohn Ferlan <jferlan@redhat.com>
Wed, 2 Sep 2015 12:59:53 +0000 (08:59 -0400)
commitdb9277a39bc364806e8d3e08a08fc128d59b7094
tree28d9879b8d49854d0d7a1e32906923fa502eca71
parent691dd388aee99f8b06177540303b690586d5f5b3
storage: Handle failure from refreshVol

Commit id '155ca616' added the 'refreshVol' API. In an NFS root-squash
environment it was possible that if the just created volume from XML wasn't
properly created with the right uid/gid and/or mode, then the followup
refreshVol will fail to open the volume in order to get the allocation/
capacity values. This would leave the volume still on the server and
cause a libvirtd crash because 'voldef' would be in the pool list, but
the cleanup code would free it.
src/storage/storage_driver.c