net/mlx5: DR, Enable QP retransmission

[ Upstream commit ec449ed8230cd30769de3cb70ee0fce293047372 ]

Under high stress, SW steering might get stuck on polling for completion
that never comes.
For such cases QP needs to have protocol retransmission mechanism enabled.
Currently the retransmission timeout is defined as 0 (unlimited). Fix this
by defining a real timeout.

Signed-off-by: Yevgeny Kliteynik <kliteyn@nvidia.com>
Reviewed-by: Alex Vesker <valex@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

diff --git a/drivers/net/ethernet/mellanox/mlx5/core/steering/dr_send.c b/drivers/net/ethernet/mellanox/mlx5/core/steering/dr_send.c
index f012aac83b10e1cfccb4d56a058ef9b4073f8c73..401564b94eb100863dd696d3b2d6f71e90e97cf3 100644 (file)
--- a/drivers/net/ethernet/mellanox/mlx5/core/steering/dr_send.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/steering/dr_send.c
@@ -603,6 +603,7 @@ static int dr_cmd_modify_qp_rtr2rts(struct mlx5_core_dev *mdev,
        MLX5_SET(qpc, qpc, log_ack_req_freq, 0);
        MLX5_SET(qpc, qpc, retry_count, attr->retry_cnt);
        MLX5_SET(qpc, qpc, rnr_retry, attr->rnr_retry);
+       MLX5_SET(qpc, qpc, primary_address_path.ack_timeout, 0x8); /* ~1ms */
 
        return mlx5_core_qp_modify(mdev, MLX5_CMD_OP_RTR2RTS_QP, 0, qpc,
                                   &dr_qp->mqp);