On Thu, Jan 25, 2018 at 11:39:59AM -0500, Doug Ledford wrote:
> On Thu, 2018-01-25 at 17:12 +0200, Leon Romanovsky wrote:
> > + if (rdma_restrack_init(&device->res)) {
> > + kfree(device);
> > + return NULL;
> > + }
> > +
>
> [ snip ]
>
> > +int rdma_restrack_init(struct rdma_restrack_root *res)
> > +{
> > + init_rwsem(&res->rwsem);
> > + return 0;
> > +}
> > +
>
> Is there a reason you are doing it this way? Right now, the init can't
> fail and the test for failure is a no-op. I looked through the
> remaining patches and nothing in this patchset changes that (at least
> that I caught). Is there a future intent here? If so, I can see
> leaving this as is (although if there is a respin, fixing it and then
> changing it back when you add the actual failure opportunity would make
> sense). But, if not, the init function should be void and there should
> be no failure test when calling it.
I don't use the returned value, it is leftover.
>
> --
> Doug Ledford <dledford@xxxxxxxxxx>
> GPG KeyID: B826A3330E572FDD
> Key fingerprint = AE6B 1BDA 122B 23B4 265B 1274 B826 A333 0E57 2FDD
Attachment:
signature.asc
Description: PGP signature
