On 8/28/24 4:27 AM, Shen Lichuan wrote:
> As opposed to open-code, using the ERR_CAST macro clearly indicates that
> this is a pointer to an error value and a type conversion was performed.
>
> Signed-off-by: Shen Lichuan <shenlichuan@xxxxxxxx>
> ---
> drivers/infiniband/sw/rdmavt/mr.c | 6 +++---
> 1 file changed, 3 insertions(+), 3 deletions(-)
>
> diff --git a/drivers/infiniband/sw/rdmavt/mr.c b/drivers/infiniband/sw/rdmavt/mr.c
> index 7a9afd5231d5..5ed5cfc2b280 100644
> --- a/drivers/infiniband/sw/rdmavt/mr.c
> +++ b/drivers/infiniband/sw/rdmavt/mr.c
> @@ -348,13 +348,13 @@ struct ib_mr *rvt_reg_user_mr(struct ib_pd *pd, u64 start, u64 length,
>
> umem = ib_umem_get(pd->device, start, length, mr_access_flags);
> if (IS_ERR(umem))
> - return (void *)umem;
> + return ERR_CAST(umem);
>
> n = ib_umem_num_pages(umem);
>
> mr = __rvt_alloc_mr(n, pd);
> if (IS_ERR(mr)) {
> - ret = (struct ib_mr *)mr;
> + ret = ERR_CAST(mr);
> goto bail_umem;
> }
>
> @@ -542,7 +542,7 @@ struct ib_mr *rvt_alloc_mr(struct ib_pd *pd, enum ib_mr_type mr_type,
>
> mr = __rvt_alloc_mr(max_num_sg, pd);
> if (IS_ERR(mr))
> - return (struct ib_mr *)mr;
> + return ERR_CAST(mr);
>
> return &mr->ibmr;
> }
I don't think this is really necessary. You are not making the code more
readable. It doesn't simplify things. So I'm not going to Ack it, but I won't
Nak either.
-Denny
