On Mon, Apr 23, 2018 at 10:44:10AM -0500, Steve Wise wrote:
>
> On 4/20/2018 1:09 PM, Jason Gunthorpe wrote:
> > On Fri, Mar 30, 2018 at 11:03:36AM -0700, Steve Wise wrote:
> >> diff --git a/drivers/infiniband/core/nldev.c b/drivers/infiniband/core/nldev.c
> >> index 884843e..1a680a3 100644
> >> +++ b/drivers/infiniband/core/nldev.c
> >> @@ -95,8 +95,25 @@
> >> [RDMA_NLDEV_ATTR_RES_PD_ENTRY] = { .type = NLA_NESTED },
> >> [RDMA_NLDEV_ATTR_RES_LOCAL_DMA_LKEY] = { .type = NLA_U32 },
> >> [RDMA_NLDEV_ATTR_RES_UNSAFE_GLOBAL_RKEY] = { .type = NLA_U32 },
> >> + [RDMA_NLDEV_ATTR_PROVIDER] = { .type = NLA_NESTED },
> >> + [RDMA_NLDEV_ATTR_PROVIDER_ENTRY] = { .type = NLA_NESTED },
> >> + [RDMA_NLDEV_ATTR_PROVIDER_STRING] = { .type = NLA_NUL_STRING,
> >> + .len = RDMA_NLDEV_ATTR_ENTRY_STRLEN },
> >> + [RDMA_NLDEV_ATTR_PROVIDER_PRINT_TYPE] = { .type = NLA_U8 },
> >> + [RDMA_NLDEV_ATTR_PROVIDER_S32] = { .type = NLA_S32 },
> >> + [RDMA_NLDEV_ATTR_PROVIDER_U32] = { .type = NLA_U32 },
> >> + [RDMA_NLDEV_ATTR_PROVIDER_S64] = { .type = NLA_S64 },
> >> + [RDMA_NLDEV_ATTR_PROVIDER_U64] = { .type = NLA_U64 },
> > Why do we need 64 and 32 bit version of this? Pass everything as s64
> > or u64?
> >
> >> +static int provider_fill_res_entry(struct rdma_restrack_root *resroot,
> >> + struct sk_buff *msg,
> >> + struct rdma_restrack_entry *res)
> >> +{
> >> + return resroot->fill_res_entry ?
> >> + resroot->fill_res_entry(msg, res) : 0;
> >> +}
> > Just init resroot->fill_res_entry to a function that returns 0 and
> > drop provider_fill_res_entry..
>
> Why is calling a function that returns 0 better than testing for the
> null function pointer?
I guess the reason to it that it will allow to us to call this
function unconditionally and get rid of provider_fill_res_entry()
entirely.
Thanks
>
> Steve.
Attachment:
signature.asc
Description: PGP signature
