static bool mlx4_en_process_tx_cq(struct ether *dev, struct mlx4_en_cq *cq) { struct mlx4_en_priv *priv = netdev_priv(dev); struct mlx4_cq *mcq = &cq->mcq; struct mlx4_en_tx_ring *ring = priv->tx_ring[cq->ring]; struct mlx4_cqe *cqe; uint16_t index; uint16_t new_index, ring_index, stamp_index; uint32_t txbbs_skipped = 0; uint32_t txbbs_stamp = 0; uint32_t cons_index = mcq->cons_index; int size = cq->size; uint32_t size_mask = ring->size_mask; struct mlx4_cqe *buf = cq->buf; uint32_t packets = 0; uint32_t bytes = 0; int factor = priv->cqe_factor; uint64_t timestamp = 0; int done = 0; int budget = priv->tx_work_limit; uint32_t last_nr_txbb; uint32_t ring_cons; if (!priv->port_up) return true; #if 0 // AKAROS_PORT netdev_txq_bql_complete_prefetchw(ring->tx_queue); #endif index = cons_index & size_mask; cqe = mlx4_en_get_cqe(buf, index, priv->cqe_size) + factor; last_nr_txbb = ACCESS_ONCE(ring->last_nr_txbb); ring_cons = ACCESS_ONCE(ring->cons); ring_index = ring_cons & size_mask; stamp_index = ring_index; /* Process all completed CQEs */ while (XNOR(cqe->owner_sr_opcode & MLX4_CQE_OWNER_MASK, cons_index & size) && (done < budget)) { /* * make sure we read the CQE after we read the * ownership bit */ bus_rmb(); if (unlikely((cqe->owner_sr_opcode & MLX4_CQE_OPCODE_MASK) == MLX4_CQE_OPCODE_ERROR)) { struct mlx4_err_cqe *cqe_err = (struct mlx4_err_cqe *)cqe; en_err(priv, "CQE error - vendor syndrome: 0x%x syndrome: 0x%x\n", cqe_err->vendor_err_syndrome, cqe_err->syndrome); } /* Skip over last polled CQE */ new_index = be16_to_cpu(cqe->wqe_index) & size_mask; do { txbbs_skipped += last_nr_txbb; ring_index = (ring_index + last_nr_txbb) & size_mask; if (ring->tx_info[ring_index].ts_requested) timestamp = mlx4_en_get_cqe_ts(cqe); /* free next descriptor */ last_nr_txbb = mlx4_en_free_tx_desc( priv, ring, ring_index, !!((ring_cons + txbbs_skipped) & ring->size), timestamp); mlx4_en_stamp_wqe(priv, ring, stamp_index, !!((ring_cons + txbbs_stamp) & ring->size)); stamp_index = ring_index; txbbs_stamp = txbbs_skipped; packets++; bytes += ring->tx_info[ring_index].nr_bytes; } while ((++done < budget) && (ring_index != new_index)); ++cons_index; index = cons_index & size_mask; cqe = mlx4_en_get_cqe(buf, index, priv->cqe_size) + factor; } /* * To prevent CQ overflow we first update CQ consumer and only then * the ring consumer. */ mcq->cons_index = cons_index; mlx4_cq_set_ci(mcq); wmb(); /* we want to dirty this cache line once */ ACCESS_ONCE(ring->last_nr_txbb) = last_nr_txbb; ACCESS_ONCE(ring->cons) = ring_cons + txbbs_skipped; #if 0 // AKAROS_PORT netdev_tx_completed_queue(ring->tx_queue, packets, bytes); /* * Wakeup Tx queue if this stopped, and at least 1 packet * was completed */ if (netif_tx_queue_stopped(ring->tx_queue) && txbbs_skipped > 0) { netif_tx_wake_queue(ring->tx_queue); ring->wake_queue++; } #endif return done < budget; }
u32 mlx4_en_recycle_tx_desc(struct mlx4_en_priv *priv, struct mlx4_en_tx_ring *ring, int index, u64 timestamp, int napi_mode) { struct mlx4_en_tx_info *tx_info = &ring->tx_info[index]; struct mlx4_en_rx_alloc frame = { .page = tx_info->page, .dma = tx_info->map0_dma, }; if (!mlx4_en_rx_recycle(ring->recycle_ring, &frame)) { dma_unmap_page(priv->ddev, tx_info->map0_dma, PAGE_SIZE, priv->dma_dir); put_page(tx_info->page); } return tx_info->nr_txbb; } int mlx4_en_free_tx_buf(struct net_device *dev, struct mlx4_en_tx_ring *ring) { struct mlx4_en_priv *priv = netdev_priv(dev); int cnt = 0; /* Skip last polled descriptor */ ring->cons += ring->last_nr_txbb; en_dbg(DRV, priv, "Freeing Tx buf - cons:0x%x prod:0x%x\n", ring->cons, ring->prod); if ((u32) (ring->prod - ring->cons) > ring->size) { if (netif_msg_tx_err(priv)) en_warn(priv, "Tx consumer passed producer!\n"); return 0; } while (ring->cons != ring->prod) { ring->last_nr_txbb = ring->free_tx_desc(priv, ring, ring->cons & ring->size_mask, 0, 0 /* Non-NAPI caller */); ring->cons += ring->last_nr_txbb; cnt++; } if (ring->tx_queue) netdev_tx_reset_queue(ring->tx_queue); if (cnt) en_dbg(DRV, priv, "Freed %d uncompleted tx descriptors\n", cnt); return cnt; } bool mlx4_en_process_tx_cq(struct net_device *dev, struct mlx4_en_cq *cq, int napi_budget) { struct mlx4_en_priv *priv = netdev_priv(dev); struct mlx4_cq *mcq = &cq->mcq; struct mlx4_en_tx_ring *ring = priv->tx_ring[cq->type][cq->ring]; struct mlx4_cqe *cqe; u16 index, ring_index, stamp_index; u32 txbbs_skipped = 0; u32 txbbs_stamp = 0; u32 cons_index = mcq->cons_index; int size = cq->size; u32 size_mask = ring->size_mask; struct mlx4_cqe *buf = cq->buf; u32 packets = 0; u32 bytes = 0; int factor = priv->cqe_factor; int done = 0; int budget = priv->tx_work_limit; u32 last_nr_txbb; u32 ring_cons; if (unlikely(!priv->port_up)) return true; netdev_txq_bql_complete_prefetchw(ring->tx_queue); index = cons_index & size_mask; cqe = mlx4_en_get_cqe(buf, index, priv->cqe_size) + factor; last_nr_txbb = READ_ONCE(ring->last_nr_txbb); ring_cons = READ_ONCE(ring->cons); ring_index = ring_cons & size_mask; stamp_index = ring_index; /* Process all completed CQEs */ while (XNOR(cqe->owner_sr_opcode & MLX4_CQE_OWNER_MASK, cons_index & size) && (done < budget)) { u16 new_index; /* * make sure we read the CQE after we read the * ownership bit */ dma_rmb(); if (unlikely((cqe->owner_sr_opcode & MLX4_CQE_OPCODE_MASK) == MLX4_CQE_OPCODE_ERROR)) { struct mlx4_err_cqe *cqe_err = (struct mlx4_err_cqe *)cqe; en_err(priv, "CQE error - vendor syndrome: 0x%x syndrome: 0x%x\n", cqe_err->vendor_err_syndrome, cqe_err->syndrome); } /* Skip over last polled CQE */ new_index = be16_to_cpu(cqe->wqe_index) & size_mask; do { u64 timestamp = 0; txbbs_skipped += last_nr_txbb; ring_index = (ring_index + last_nr_txbb) & size_mask; if (unlikely(ring->tx_info[ring_index].ts_requested)) timestamp = mlx4_en_get_cqe_ts(cqe); /* free next descriptor */ last_nr_txbb = ring->free_tx_desc( priv, ring, ring_index, timestamp, napi_budget); mlx4_en_stamp_wqe(priv, ring, stamp_index, !!((ring_cons + txbbs_stamp) & ring->size)); stamp_index = ring_index; txbbs_stamp = txbbs_skipped; packets++; bytes += ring->tx_info[ring_index].nr_bytes; } while ((++done < budget) && (ring_index != new_index)); ++cons_index; index = cons_index & size_mask; cqe = mlx4_en_get_cqe(buf, index, priv->cqe_size) + factor; } /* * To prevent CQ overflow we first update CQ consumer and only then * the ring consumer. */ mcq->cons_index = cons_index; mlx4_cq_set_ci(mcq); wmb(); /* we want to dirty this cache line once */ WRITE_ONCE(ring->last_nr_txbb, last_nr_txbb); WRITE_ONCE(ring->cons, ring_cons + txbbs_skipped); if (cq->type == TX_XDP) return done < budget; netdev_tx_completed_queue(ring->tx_queue, packets, bytes); /* Wakeup Tx queue if this stopped, and ring is not full. */ if (netif_tx_queue_stopped(ring->tx_queue) && !mlx4_en_is_tx_ring_full(ring)) { netif_tx_wake_queue(ring->tx_queue); ring->wake_queue++; } return done < budget; } void mlx4_en_tx_irq(struct mlx4_cq *mcq) { struct mlx4_en_cq *cq = container_of(mcq, struct mlx4_en_cq, mcq); struct mlx4_en_priv *priv = netdev_priv(cq->dev); if (likely(priv->port_up)) napi_schedule_irqoff(&cq->napi); else mlx4_en_arm_cq(priv, cq); } /* TX CQ polling - called by NAPI */ int mlx4_en_poll_tx_cq(struct napi_struct *napi, int budget) { struct mlx4_en_cq *cq = container_of(napi, struct mlx4_en_cq, napi); struct net_device *dev = cq->dev; struct mlx4_en_priv *priv = netdev_priv(dev); bool clean_complete; clean_complete = mlx4_en_process_tx_cq(dev, cq, budget); if (!clean_complete) return budget; napi_complete(napi); mlx4_en_arm_cq(priv, cq); return 0; } static struct mlx4_en_tx_desc *mlx4_en_bounce_to_desc(struct mlx4_en_priv *priv, struct mlx4_en_tx_ring *ring, u32 index, unsigned int desc_size) { u32 copy = (ring->size - index) << LOG_TXBB_SIZE; int i; for (i = desc_size - copy - 4; i >= 0; i -= 4) { if ((i & (TXBB_SIZE - 1)) == 0) wmb(); *((u32 *) (ring->buf + i)) = *((u32 *) (ring->bounce_buf + copy + i)); } for (i = copy - 4; i >= 4 ; i -= 4) { if ((i & (TXBB_SIZE - 1)) == 0) wmb(); *((u32 *)(ring->buf + (index << LOG_TXBB_SIZE) + i)) = *((u32 *) (ring->bounce_buf + i)); } /* Return real descriptor location */ return ring->buf + (index << LOG_TXBB_SIZE); }