Re: [PATCH for-next] RDMA/providers: Fix return value when QP type isn't supported

[Jason Gunthorpe <jgg@ziepe.ca>]

On Thu, Jan 30, 2020 at 10:20:49AM +0200, Kamal Heib wrote:
> The proper return code is "-EOPNOTSUPP" when the requested QP type is
> not supported by the provider.
> 
> Signed-off-by: Kamal Heib <kamalheib1@gmail.com>
> ---
>  drivers/infiniband/hw/bnxt_re/ib_verbs.c     | 2 +-
>  drivers/infiniband/hw/cxgb4/qp.c             | 2 +-
>  drivers/infiniband/hw/hns/hns_roce_qp.c      | 2 +-
>  drivers/infiniband/hw/i40iw/i40iw_verbs.c    | 2 +-
>  drivers/infiniband/hw/mlx4/qp.c              | 2 +-
>  drivers/infiniband/hw/mlx5/qp.c              | 2 +-
>  drivers/infiniband/hw/mthca/mthca_provider.c | 2 +-
>  drivers/infiniband/hw/ocrdma/ocrdma_verbs.c  | 2 +-
>  drivers/infiniband/hw/qedr/verbs.c           | 2 +-
>  drivers/infiniband/hw/usnic/usnic_ib_verbs.c | 2 +-
>  drivers/infiniband/hw/vmw_pvrdma/pvrdma_qp.c | 2 +-
>  drivers/infiniband/sw/rdmavt/qp.c            | 2 +-
>  drivers/infiniband/sw/siw/siw_verbs.c        | 2 +-
>  13 files changed, 13 insertions(+), 13 deletions(-)

Well, it looks like we already have providers returning EOPNOTSUPP for
various cases under create_qp, so all callers already need to handle
it.

However, there are still lots of cases even within the create_qp flow
where the return codes are wrong. Below is what I thought for mlx5, so
this problem space needs a lot more attention.

But, fixing the obvious places that check qp_type seems like a
reasonable step, so applied to for-next

diff --git a/drivers/infiniband/hw/mlx5/qp.c b/drivers/infiniband/hw/mlx5/qp.c
index a4f8e703078718..daa1b6b370e17b 100644
--- a/drivers/infiniband/hw/mlx5/qp.c
+++ b/drivers/infiniband/hw/mlx5/qp.c
@@ -2023,7 +2023,7 @@ static int create_qp_common(struct mlx5_ib_dev *dev, struct ib_pd *pd,
 
 	if (init_attr->rwq_ind_tbl) {
 		if (!udata)
-			return -ENOSYS;
+			return -EOPNOTSUPP;
 
 		err = create_rss_raw_qp_tir(dev, qp, pd, init_attr, udata);
 		return err;
@@ -2032,7 +2032,7 @@ static int create_qp_common(struct mlx5_ib_dev *dev, struct ib_pd *pd,
 	if (init_attr->create_flags & IB_QP_CREATE_BLOCK_MULTICAST_LOOPBACK) {
 		if (!MLX5_CAP_GEN(mdev, block_lb_mc)) {
 			mlx5_ib_dbg(dev, "block multicast loopback isn't supported\n");
-			return -EINVAL;
+			return -EOPNOTSUPP;
 		} else {
 			qp->flags |= MLX5_IB_QP_BLOCK_MULTICAST_LOOPBACK;
 		}
@@ -2044,7 +2044,7 @@ static int create_qp_common(struct mlx5_ib_dev *dev, struct ib_pd *pd,
 			 IB_QP_CREATE_MANAGED_RECV)) {
 		if (!MLX5_CAP_GEN(mdev, cd)) {
 			mlx5_ib_dbg(dev, "cross-channel isn't supported\n");
-			return -EINVAL;
+			return -EOPNOTSUPP;
 		}
 		if (init_attr->create_flags & IB_QP_CREATE_CROSS_CHANNEL)
 			qp->flags |= MLX5_IB_QP_CROSS_CHANNEL;
@@ -2693,7 +2693,7 @@ struct ib_qp *mlx5_ib_create_qp(struct ib_pd *pd,
 		if (init_attr->qp_type == IB_QPT_RAW_PACKET) {
 			if (!ucontext) {
 				mlx5_ib_dbg(dev, "Raw Packet QP is not supported for kernel consumers\n");
-				return ERR_PTR(-EINVAL);
+				return ERR_PTR(-EOPNOTSUPP;
 			} else if (!ucontext->cqe_version) {
 				mlx5_ib_dbg(dev, "Raw Packet QP is only supported for CQE version > 0\n");
 				return ERR_PTR(-EINVAL);
@@ -2735,7 +2735,7 @@ struct ib_qp *mlx5_ib_create_qp(struct ib_pd *pd,
 	case IB_QPT_XRC_INI:
 		if (!MLX5_CAP_GEN(dev->mdev, xrc)) {
 			mlx5_ib_dbg(dev, "XRC not supported\n");
-			return ERR_PTR(-ENOSYS);
+			return ERR_PTR(-EOPNOTSUPP);
 		}
 		init_attr->recv_cq = NULL;
 		if (init_attr->qp_type == IB_QPT_XRC_TGT) {
@@ -2789,7 +2789,7 @@ struct ib_qp *mlx5_ib_create_qp(struct ib_pd *pd,
 		mlx5_ib_dbg(dev, "unsupported qp type %d\n",
 			    init_attr->qp_type);
 		/* Don't support raw QPs */
-		return ERR_PTR(-EINVAL);
+		return ERR_PTR(-EOPNOTSUPP);
 	}
 
 	if (verbs_init_attr->qp_type == IB_QPT_DRIVER)