On Mon, Apr 13, 2020 at 04:20:28PM +0300, Leon Romanovsky wrote:
> From: Aharon Landau <aharonl@xxxxxxxxxxxx>
>
> GRH fields such as sgid_index, hop limit and etc. are set in the
> QP context when QP is created/modified.
>
> Currently, when query QP is performed, we fill the GRH fields only
> if the GRH bit is set in the QP context, but this bit is not set
> for RoCE. Adjust the check so we will set all relevant data for
> the RoCE too.
>
> Fixes: d8966fcd4c25 ("IB/core: Use rdma_ah_attr accessor functions")
> Signed-off-by: Aharon Landau <aharonl@xxxxxxxxxxxx>
> Reviewed-by: Maor Gottlieb <maorg@xxxxxxxxxxxx>
> Signed-off-by: Leon Romanovsky <leonro@xxxxxxxxxxxx>
> ---
> drivers/infiniband/hw/mlx5/qp.c | 10 +++++-----
> 1 file changed, 5 insertions(+), 5 deletions(-)
>
> diff --git a/drivers/infiniband/hw/mlx5/qp.c b/drivers/infiniband/hw/mlx5/qp.c
> index 1456db4b6295..a4f8e7c7ed24 100644
> --- a/drivers/infiniband/hw/mlx5/qp.c
> +++ b/drivers/infiniband/hw/mlx5/qp.c
> @@ -5558,13 +5558,13 @@ static void to_rdma_ah_attr(struct mlx5_ib_dev *ibdev,
> rdma_ah_set_path_bits(ah_attr, path->grh_mlid & 0x7f);
> rdma_ah_set_static_rate(ah_attr,
> path->static_rate ? path->static_rate - 5 : 0);
> - if (path->grh_mlid & (1 << 7)) {
> +
> + if (path->grh_mlid & (1 << 7) ||
> + ah_attr->type == RDMA_AH_ATTR_TYPE_ROCE) {
> u32 tc_fl = be32_to_cpu(path->tclass_flowlabel);
>
> - rdma_ah_set_grh(ah_attr, NULL,
> - tc_fl & 0xfffff,
> - path->mgid_index,
> - path->hop_limit,
> + rdma_ah_set_grh(ah_attr, NULL, tc_fl & 0xfffff,
> + path->mgid_index, path->hop_limit,
Why is whitespace reformatting in a -rc patch? I dropped this.
Applied to for-rc
Jason
