On 05/26/2017 08:56 AM, Peter Krempa wrote:
On Fri, May 26, 2017 at 07:59:08 -0400, John Ferlan wrote:
> For now it'll just call the virInterfaceObjUnlock, but a future adjustment
> will do something different.
>
> The virInterfaceObjUnlock is now private to virinterfaceobj.c with a
> short term forward reference.
>
> Additionally, make virInterfaceObjLock private since it's only used in
> virinterfaceobj anyway. For now this will involved creating a forward
> reference, but this will go away soon too.
>
> Signed-off-by: John Ferlan <jferlan(a)redhat.com>
> ---
> src/conf/virinterfaceobj.c | 17 +++++++++++++++--
> src/conf/virinterfaceobj.h | 9 +++------
> src/libvirt_private.syms | 3 +--
> src/test/test_driver.c | 15 +++++++--------
> 4 files changed, 26 insertions(+), 18 deletions(-)
[...]
>
> @@ -82,6 +85,16 @@ virInterfaceObjUnlock(virInterfaceObjPtr obj)
>
>
> void
> +virInterfaceObjEndAPI(virInterfaceObjPtr *obj)
> +{
> + if (!*obj)
> + return;
> +
> + virInterfaceObjUnlock(*obj);
The double pointer is really pointless in this function. Did you forget
to set it to NULL after the unlock as we do in the other EndAPI
functions?
Well... It was a way to avoid having more changes in the "next" patch.
The reality is this and the next one were one at one time, but I figured
I could extract out one and make the following one a bit easier to read.
I can combine them into one if that's desired.
Tks -
John