net/mlx5e: Fix TX Timeout to detect queues stuck on BQL

[cascardo/linux.git] / drivers / net / ethernet / mellanox / mlx5 / core / en_main.c

diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_main.c b/drivers/net/ethernet/mellanox/mlx5/core/en_main.c
index 38c1286..0cebc7e 100644 (file)
--- a/drivers/net/ethernet/mellanox/mlx5/core/en_main.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_main.c
@@ -81,10 +81,13 @@ static void mlx5e_update_carrier(struct mlx5e_priv *priv)
        port_state = mlx5_query_vport_state(mdev,
                MLX5_QUERY_VPORT_STATE_IN_OP_MOD_VNIC_VPORT, 0);
 
-       if (port_state == VPORT_STATE_UP)
+       if (port_state == VPORT_STATE_UP) {
+               netdev_info(priv->netdev, "Link up\n");
                netif_carrier_on(priv->netdev);
-       else
+       } else {
+               netdev_info(priv->netdev, "Link down\n");
                netif_carrier_off(priv->netdev);
+       }
 }
 
 static void mlx5e_update_carrier_work(struct work_struct *work)
@@ -332,6 +335,7 @@ static int mlx5e_create_rq(struct mlx5e_channel *c,
                }
                rq->handle_rx_cqe = mlx5e_handle_rx_cqe_mpwrq;
                rq->alloc_wqe = mlx5e_alloc_rx_mpwqe;
+               rq->dealloc_wqe = mlx5e_dealloc_rx_mpwqe;
 
                rq->mpwqe_stride_sz = BIT(priv->params.mpwqe_log_stride_sz);
                rq->mpwqe_num_strides = BIT(priv->params.mpwqe_log_num_strides);
@@ -347,6 +351,7 @@ static int mlx5e_create_rq(struct mlx5e_channel *c,
                }
                rq->handle_rx_cqe = mlx5e_handle_rx_cqe;
                rq->alloc_wqe = mlx5e_alloc_rx_wqe;
+               rq->dealloc_wqe = mlx5e_dealloc_rx_wqe;
 
                rq->wqe_sz = (priv->params.lro_en) ?
                                priv->params.lro_wqe_sz :
@@ -552,17 +557,25 @@ err_destroy_rq:
 
 static void mlx5e_close_rq(struct mlx5e_rq *rq)
 {
+       int tout = 0;
+       int err;
+
        clear_bit(MLX5E_RQ_STATE_POST_WQES_ENABLE, &rq->state);
        napi_synchronize(&rq->channel->napi); /* prevent mlx5e_post_rx_wqes */
 
-       mlx5e_modify_rq_state(rq, MLX5_RQC_STATE_RDY, MLX5_RQC_STATE_ERR);
-       while (!mlx5_wq_ll_is_empty(&rq->wq))
-               msleep(20);
+       err = mlx5e_modify_rq_state(rq, MLX5_RQC_STATE_RDY, MLX5_RQC_STATE_ERR);
+       while (!mlx5_wq_ll_is_empty(&rq->wq) && !err &&
+              tout++ < MLX5_EN_QP_FLUSH_MAX_ITER)
+               msleep(MLX5_EN_QP_FLUSH_MSLEEP_QUANT);
+
+       if (err || tout == MLX5_EN_QP_FLUSH_MAX_ITER)
+               set_bit(MLX5E_RQ_STATE_FLUSH_TIMEOUT, &rq->state);
 
        /* avoid destroying rq before mlx5e_poll_rx_cq() is done with it */
        napi_synchronize(&rq->channel->napi);
 
        mlx5e_disable_rq(rq);
+       mlx5e_free_rx_descs(rq);
        mlx5e_destroy_rq(rq);
 }
 
@@ -1697,8 +1710,11 @@ static void mlx5e_netdev_set_tcs(struct net_device *netdev)
 
        netdev_set_num_tc(netdev, ntc);
 
+       /* Map netdev TCs to offset 0
+        * We have our own UP to TXQ mapping for QoS
+        */
        for (tc = 0; tc < ntc; tc++)
-               netdev_set_tc_queue(netdev, tc, nch, tc * nch);
+               netdev_set_tc_queue(netdev, tc, nch, 0);
 }
 
 int mlx5e_open_locked(struct net_device *netdev)
@@ -2640,7 +2656,7 @@ static void mlx5e_tx_timeout(struct net_device *dev)
        for (i = 0; i < priv->params.num_channels * priv->params.num_tc; i++) {
                struct mlx5e_sq *sq = priv->txq_to_sq_map[i];
 
-               if (!netif_tx_queue_stopped(netdev_get_tx_queue(dev, i)))
+               if (!netif_xmit_stopped(netdev_get_tx_queue(dev, i)))
                        continue;
                sched_work = true;
                set_bit(MLX5E_SQ_STATE_TX_TIMEOUT, &sq->state);