[PATCH v2 5/6] qemu: Allow setting per-disk snapshot name for RBD disks

[Peter Krempa]

On Wed, Feb 15, 2023 at 05:28:21 -0600, Or Ozeri wrote:
> This commit adds the option for setting per-disk snapshot name for
> RBD disks. All other disk types are still disallowed to use the
> snapshotName attribute.
> 
> Signed-off-by: Or Ozeri <oro at il.ibm.com>
> ---
>  src/qemu/qemu_snapshot.c | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
> 
> diff --git a/src/qemu/qemu_snapshot.c b/src/qemu/qemu_snapshot.c
> index a10bdf7bf2..c72bdb4723 100644
> --- a/src/qemu/qemu_snapshot.c
> +++ b/src/qemu/qemu_snapshot.c
> @@ -664,7 +664,7 @@ qemuSnapshotPrepare(virDomainObj *vm,
>                  return -1;
>              }
>  
> -            if (disk->snapshot_name) {
> +            if (disk->snapshot_name && !is_raw_rbd) {
>                  virReportError(VIR_ERR_CONFIG_UNSUPPORTED,
>                                 _("snapshot name setting for disk %s unsupported "
>                                   "for storage type %s"),

I don't see anything that updates the equivalent of
dom_disk->src->snapshot after the snapshot:

The 'snapshot' property is filled as the equivalent property when
formatting the backend definition for the 'rbd' disk.

In case when the 'snapshot' field is meant to actually mean label the
'old' state. You then must actually tweak the snapshot metadata to point
to it. That will allow proper reversion of the image.