On Thu, Dec 22, 2016 at 09:12:43AM -0500, John Ferlan wrote:
https://bugzilla.redhat.com/show_bug.cgi?id=1405269
If a secret was not provided for what was determined to be a LUKS
encrypted disk (during virStorageFileGetMetadata processing when
called from qemuDomainDetermineDiskChain as a result of hotplug
attach qemuDomainAttachDeviceDiskLive), then do not attempt to
look it up (avoiding a libvirtd crash) and do not alter the format
to "luks" when adding the disk; otherwise, the device_add would
fail with a message such as:
"unable to execute QEMU command 'device_add': Property
'scsi-hd.drive'
can't find value 'drive-scsi0-0-0-0'"
because of assumptions that when the format=luks that libvirt would have
provided the secret to decrypt the volume.
Access to unlock the volume will thus be left to the application.
Signed-off-by: John Ferlan <jferlan(a)redhat.com>
---
I could have spread things over 2 patches (one to create the helper
and one to add the extra check for nsecrets > 0), but just going with
one to fix the issue just felt cleaner (besides makes any possible
backports a bit simpler).
NB: The "extra" !virStorageSourceIsEmpty check for the qemu_command
path is a no-op essentially.
src/qemu/qemu_command.c | 3 +--
src/qemu/qemu_domain.c | 15 +++++++++++++--
src/qemu/qemu_domain.h | 3 +++
src/qemu/qemu_hotplug.c | 3 ++-
4 files changed, 19 insertions(+), 5 deletions(-)
ACK
diff --git a/src/qemu/qemu_hotplug.c b/src/qemu/qemu_hotplug.c
index 92a2e73..6b10e63 100644
--- a/src/qemu/qemu_hotplug.c
+++ b/src/qemu/qemu_hotplug.c
@@ -3584,7 +3584,8 @@ qemuDomainRemoveDiskDevice(virQEMUDriverPtr driver,
* can remove the luks object password too
*/
if (!virStorageSourceIsEmpty(disk->src) && disk->src->encryption
&&
- disk->src->encryption->format == VIR_STORAGE_ENCRYPTION_FORMAT_LUKS) {
+ disk->src->encryption->format == VIR_STORAGE_ENCRYPTION_FORMAT_LUKS
&&
+ disk->src->encryption->nsecrets > 0) {
Why doesn't this code also use the helper?
Jan