On Tue, Mar 30, 2021 at 06:14:13PM +0200, Cornelia Huck wrote:
> On Fri, 26 Mar 2021 08:53:50 -0300
> Jason Gunthorpe <jgg@xxxxxxxxxx> wrote:
>
> > On Fri, Mar 26, 2021 at 03:53:10AM +0000, Tian, Kevin wrote:
> >
> > > > @@ -58,12 +58,11 @@ void mdev_release_parent(struct kref *kref)
> > > > /* Caller must hold parent unreg_sem read or write lock */
> > > > static void mdev_device_remove_common(struct mdev_device *mdev)
> > > > {
> > > > - struct mdev_parent *parent;
> > > > + struct mdev_parent *parent = mdev->type->parent;
> > >
> > > What about having a wrapper here, like mdev_parent_dev? For
> > > readability it's not necessary to show that the parent is indirectly
> > > retrieved through mdev_type.
> >
> > I think that is too much wrappering, we only have three usages of the
> > mdev->type->parent sequence and two are already single line inlines.
>
> I'm counting more of those in this patch... or do you mean at the end
> of the series?
I'm thinking of either moving them to inlines earlier or moving them
all as a final patch. Given how far we've come review wise I'm
inclined to do the latter
Thanks,
Jason
