From: Sagi Grimberg Date: Thu, 3 Mar 2016 11:37:51 +0000 (+0200) Subject: IB/mlx5: Expose correct max_fast_reg_page_list_len X-Git-Url: http://git.cdn.openwrt.org/?a=commitdiff_plain;h=911f4331bc87f4589b9096f4fb24b335d4c2967d;p=openwrt%2Fstaging%2Fblogic.git IB/mlx5: Expose correct max_fast_reg_page_list_len While documentation indicates that the number of translation entries per memory key is unlimited, in practice, we can only fit a finite amount of translation entries in a single registration wqe (which is log_max_klm_list_size). Signed-off-by: Sagi Grimberg Signed-off-by: Doug Ledford --- diff --git a/drivers/infiniband/hw/mlx5/main.c b/drivers/infiniband/hw/mlx5/main.c index 63c3d21dd471..55fa5889d3b2 100644 --- a/drivers/infiniband/hw/mlx5/main.c +++ b/drivers/infiniband/hw/mlx5/main.c @@ -539,7 +539,8 @@ static int mlx5_ib_query_device(struct ib_device *ibdev, props->local_ca_ack_delay = MLX5_CAP_GEN(mdev, local_ca_ack_delay); props->max_res_rd_atom = props->max_qp_rd_atom * props->max_qp; props->max_srq_sge = max_rq_sg - 1; - props->max_fast_reg_page_list_len = (unsigned int)-1; + props->max_fast_reg_page_list_len = + 1 << MLX5_CAP_GEN(mdev, log_max_klm_list_size); get_atomic_caps(dev, props); props->masked_atomic_cap = IB_ATOMIC_NONE; props->max_mcast_grp = 1 << MLX5_CAP_GEN(mdev, log_max_mcg);