Re: [libvirt] [PATCH v3 2/5] storage: Use virSecretGetSecretString
On 06/06/2016 03:32 AM, Peter Krempa wrote:
On Fri, Jun 03, 2016 at 06:52:50 -0400, John Ferlan wrote:
> Rather than inline code secret lookup for rbd/iscsi, use the common function.
>
> Signed-off-by: John Ferlan <jferlan(a)redhat.com>
> ---
> src/Makefile.am | 1 +
> src/storage/storage_backend_iscsi.c | 50 +++++--------------------------------
> src/storage/storage_backend_rbd.c | 48 +++--------------------------------
> 3 files changed, 10 insertions(+), 89 deletions(-)
>
> diff --git a/src/Makefile.am b/src/Makefile.am
> index f3c9a14..019242b 100644
> --- a/src/Makefile.am
> +++ b/src/Makefile.am
> @@ -1615,6 +1615,7 @@ libvirt_driver_storage_impl_la_SOURCES =
> libvirt_driver_storage_impl_la_CFLAGS = \
> -I$(srcdir)/access \
> -I$(srcdir)/conf \
> + -I$(srcdir)/secret \
Similarly to my complaint in 5/5 this is breaking the boundaries between
the storage driver and the secret driver. This is not the first case of
this though so we just need to thoroughly check that the appropriate
bits are linked in to the storage driver.
I'll need to check this.
Otherwise looks good.
Peter
Do you mean boundaries that weren't broken before? I would think 5/5 is
a different case/issue. Ironcially, RBD/iSCSI would use virSecret*
libvirt API's to get the lookup type and then conn->secretDriver
knowledge to perform the actual lookup.
This is the one that had already been ACK'd and essentially follows
commit id 2844de6f4 done for libvirt_driver_qemu_impl_la and
libvirt_driver_libxl_impl_la. It's adding access to one function and one
enum.
John