diff options
Diffstat (limited to 'drivers/infiniband/ulp/ipoib/ipoib_ib.c')
| -rw-r--r-- | drivers/infiniband/ulp/ipoib/ipoib_ib.c | 195 |
1 files changed, 106 insertions, 89 deletions
diff --git a/drivers/infiniband/ulp/ipoib/ipoib_ib.c b/drivers/infiniband/ulp/ipoib/ipoib_ib.c index 72626c348174..63b92cbb29ad 100644 --- a/drivers/infiniband/ulp/ipoib/ipoib_ib.c +++ b/drivers/infiniband/ulp/ipoib/ipoib_ib.c @@ -94,39 +94,9 @@ void ipoib_free_ah(struct kref *kref) static void ipoib_ud_dma_unmap_rx(struct ipoib_dev_priv *priv, u64 mapping[IPOIB_UD_RX_SG]) { - if (ipoib_ud_need_sg(priv->max_ib_mtu)) { - ib_dma_unmap_single(priv->ca, mapping[0], IPOIB_UD_HEAD_SIZE, - DMA_FROM_DEVICE); - ib_dma_unmap_page(priv->ca, mapping[1], PAGE_SIZE, - DMA_FROM_DEVICE); - } else - ib_dma_unmap_single(priv->ca, mapping[0], - IPOIB_UD_BUF_SIZE(priv->max_ib_mtu), - DMA_FROM_DEVICE); -} - -static void ipoib_ud_skb_put_frags(struct ipoib_dev_priv *priv, - struct sk_buff *skb, - unsigned int length) -{ - if (ipoib_ud_need_sg(priv->max_ib_mtu)) { - skb_frag_t *frag = &skb_shinfo(skb)->frags[0]; - unsigned int size; - /* - * There is only two buffers needed for max_payload = 4K, - * first buf size is IPOIB_UD_HEAD_SIZE - */ - skb->tail += IPOIB_UD_HEAD_SIZE; - skb->len += length; - - size = length - IPOIB_UD_HEAD_SIZE; - - skb_frag_size_set(frag, size); - skb->data_len += size; - skb->truesize += PAGE_SIZE; - } else - skb_put(skb, length); - + ib_dma_unmap_single(priv->ca, mapping[0], + IPOIB_UD_BUF_SIZE(priv->max_ib_mtu), + DMA_FROM_DEVICE); } static int ipoib_ib_post_receive(struct net_device *dev, int id) @@ -156,18 +126,11 @@ static struct sk_buff *ipoib_alloc_rx_skb(struct net_device *dev, int id) struct ipoib_dev_priv *priv = netdev_priv(dev); struct sk_buff *skb; int buf_size; - int tailroom; u64 *mapping; - if (ipoib_ud_need_sg(priv->max_ib_mtu)) { - buf_size = IPOIB_UD_HEAD_SIZE; - tailroom = 128; /* reserve some tailroom for IP/TCP headers */ - } else { - buf_size = IPOIB_UD_BUF_SIZE(priv->max_ib_mtu); - tailroom = 0; - } + buf_size = IPOIB_UD_BUF_SIZE(priv->max_ib_mtu); - skb = dev_alloc_skb(buf_size + tailroom + 4); + skb = dev_alloc_skb(buf_size + IPOIB_ENCAP_LEN); if (unlikely(!skb)) return NULL; @@ -184,23 +147,8 @@ static struct sk_buff *ipoib_alloc_rx_skb(struct net_device *dev, int id) if (unlikely(ib_dma_mapping_error(priv->ca, mapping[0]))) goto error; - if (ipoib_ud_need_sg(priv->max_ib_mtu)) { - struct page *page = alloc_page(GFP_ATOMIC); - if (!page) - goto partial_error; - skb_fill_page_desc(skb, 0, page, 0, PAGE_SIZE); - mapping[1] = - ib_dma_map_page(priv->ca, page, - 0, PAGE_SIZE, DMA_FROM_DEVICE); - if (unlikely(ib_dma_mapping_error(priv->ca, mapping[1]))) - goto partial_error; - } - priv->rx_ring[id].skb = skb; return skb; - -partial_error: - ib_dma_unmap_single(priv->ca, mapping[0], buf_size, DMA_FROM_DEVICE); error: dev_kfree_skb_any(skb); return NULL; @@ -278,7 +226,8 @@ static void ipoib_ib_handle_rx_wc(struct net_device *dev, struct ib_wc *wc) wc->byte_len, wc->slid); ipoib_ud_dma_unmap_rx(priv, mapping); - ipoib_ud_skb_put_frags(priv, skb, wc->byte_len); + + skb_put(skb, wc->byte_len); /* First byte of dgid signals multicast when 0xff */ dgid = &((struct ib_grh *)skb->data)->dgid; @@ -296,6 +245,8 @@ static void ipoib_ib_handle_rx_wc(struct net_device *dev, struct ib_wc *wc) skb_reset_mac_header(skb); skb_pull(skb, IPOIB_ENCAP_LEN); + skb->truesize = SKB_TRUESIZE(skb->len); + ++dev->stats.rx_packets; dev->stats.rx_bytes += skb->len; @@ -376,6 +327,51 @@ static void ipoib_dma_unmap_tx(struct ib_device *ca, } } +/* + * As the result of a completion error the QP Can be transferred to SQE states. + * The function checks if the (send)QP is in SQE state and + * moves it back to RTS state, that in order to have it functional again. + */ +static void ipoib_qp_state_validate_work(struct work_struct *work) +{ + struct ipoib_qp_state_validate *qp_work = + container_of(work, struct ipoib_qp_state_validate, work); + + struct ipoib_dev_priv *priv = qp_work->priv; + struct ib_qp_attr qp_attr; + struct ib_qp_init_attr query_init_attr; + int ret; + + ret = ib_query_qp(priv->qp, &qp_attr, IB_QP_STATE, &query_init_attr); + if (ret) { + ipoib_warn(priv, "%s: Failed to query QP ret: %d\n", + __func__, ret); + goto free_res; + } + pr_info("%s: QP: 0x%x is in state: %d\n", + __func__, priv->qp->qp_num, qp_attr.qp_state); + + /* currently support only in SQE->RTS transition*/ + if (qp_attr.qp_state == IB_QPS_SQE) { + qp_attr.qp_state = IB_QPS_RTS; + + ret = ib_modify_qp(priv->qp, &qp_attr, IB_QP_STATE); + if (ret) { + pr_warn("failed(%d) modify QP:0x%x SQE->RTS\n", + ret, priv->qp->qp_num); + goto free_res; + } + pr_info("%s: QP: 0x%x moved from IB_QPS_SQE to IB_QPS_RTS\n", + __func__, priv->qp->qp_num); + } else { + pr_warn("QP (%d) will stay in state: %d\n", + priv->qp->qp_num, qp_attr.qp_state); + } + +free_res: + kfree(qp_work); +} + static void ipoib_ib_handle_tx_wc(struct net_device *dev, struct ib_wc *wc) { struct ipoib_dev_priv *priv = netdev_priv(dev); @@ -407,10 +403,22 @@ static void ipoib_ib_handle_tx_wc(struct net_device *dev, struct ib_wc *wc) netif_wake_queue(dev); if (wc->status != IB_WC_SUCCESS && - wc->status != IB_WC_WR_FLUSH_ERR) + wc->status != IB_WC_WR_FLUSH_ERR) { + struct ipoib_qp_state_validate *qp_work; ipoib_warn(priv, "failed send event " "(status=%d, wrid=%d vend_err %x)\n", wc->status, wr_id, wc->vendor_err); + qp_work = kzalloc(sizeof(*qp_work), GFP_ATOMIC); + if (!qp_work) { + ipoib_warn(priv, "%s Failed alloc ipoib_qp_state_validate for qp: 0x%x\n", + __func__, priv->qp->qp_num); + return; + } + + INIT_WORK(&qp_work->work, ipoib_qp_state_validate_work); + qp_work->priv = priv; + queue_work(priv->wq, &qp_work->work); + } } static int poll_tx(struct ipoib_dev_priv *priv) @@ -655,16 +663,33 @@ void ipoib_reap_ah(struct work_struct *work) __ipoib_reap_ah(dev); if (!test_bit(IPOIB_STOP_REAPER, &priv->flags)) - queue_delayed_work(ipoib_workqueue, &priv->ah_reap_task, + queue_delayed_work(priv->wq, &priv->ah_reap_task, round_jiffies_relative(HZ)); } +static void ipoib_flush_ah(struct net_device *dev) +{ + struct ipoib_dev_priv *priv = netdev_priv(dev); + + cancel_delayed_work(&priv->ah_reap_task); + flush_workqueue(priv->wq); + ipoib_reap_ah(&priv->ah_reap_task.work); +} + +static void ipoib_stop_ah(struct net_device *dev) +{ + struct ipoib_dev_priv *priv = netdev_priv(dev); + + set_bit(IPOIB_STOP_REAPER, &priv->flags); + ipoib_flush_ah(dev); +} + static void ipoib_ib_tx_timer_func(unsigned long ctx) { drain_tx_cq((struct net_device *)ctx); } -int ipoib_ib_dev_open(struct net_device *dev, int flush) +int ipoib_ib_dev_open(struct net_device *dev) { struct ipoib_dev_priv *priv = netdev_priv(dev); int ret; @@ -696,7 +721,7 @@ int ipoib_ib_dev_open(struct net_device *dev, int flush) } clear_bit(IPOIB_STOP_REAPER, &priv->flags); - queue_delayed_work(ipoib_workqueue, &priv->ah_reap_task, + queue_delayed_work(priv->wq, &priv->ah_reap_task, round_jiffies_relative(HZ)); if (!test_and_set_bit(IPOIB_FLAG_INITIALIZED, &priv->flags)) @@ -706,7 +731,7 @@ int ipoib_ib_dev_open(struct net_device *dev, int flush) dev_stop: if (!test_and_set_bit(IPOIB_FLAG_INITIALIZED, &priv->flags)) napi_enable(&priv->napi); - ipoib_ib_dev_stop(dev, flush); + ipoib_ib_dev_stop(dev); return -1; } @@ -738,7 +763,7 @@ int ipoib_ib_dev_up(struct net_device *dev) return ipoib_mcast_start_thread(dev); } -int ipoib_ib_dev_down(struct net_device *dev, int flush) +int ipoib_ib_dev_down(struct net_device *dev) { struct ipoib_dev_priv *priv = netdev_priv(dev); @@ -747,7 +772,7 @@ int ipoib_ib_dev_down(struct net_device *dev, int flush) clear_bit(IPOIB_FLAG_OPER_UP, &priv->flags); netif_carrier_off(dev); - ipoib_mcast_stop_thread(dev, flush); + ipoib_mcast_stop_thread(dev); ipoib_mcast_dev_flush(dev); ipoib_flush_paths(dev); @@ -807,7 +832,7 @@ void ipoib_drain_cq(struct net_device *dev) local_bh_enable(); } -int ipoib_ib_dev_stop(struct net_device *dev, int flush) +int ipoib_ib_dev_stop(struct net_device *dev) { struct ipoib_dev_priv *priv = netdev_priv(dev); struct ib_qp_attr qp_attr; @@ -877,24 +902,7 @@ timeout: if (ib_modify_qp(priv->qp, &qp_attr, IB_QP_STATE)) ipoib_warn(priv, "Failed to modify QP to RESET state\n"); - /* Wait for all AHs to be reaped */ - set_bit(IPOIB_STOP_REAPER, &priv->flags); - cancel_delayed_work(&priv->ah_reap_task); - if (flush) - flush_workqueue(ipoib_workqueue); - - begin = jiffies; - - while (!list_empty(&priv->dead_ahs)) { - __ipoib_reap_ah(dev); - - if (time_after(jiffies, begin + HZ)) { - ipoib_warn(priv, "timing out; will leak address handles\n"); - break; - } - - msleep(1); - } + ipoib_flush_ah(dev); ib_req_notify_cq(priv->recv_cq, IB_CQ_NEXT_COMP); @@ -918,7 +926,7 @@ int ipoib_ib_dev_init(struct net_device *dev, struct ib_device *ca, int port) (unsigned long) dev); if (dev->flags & IFF_UP) { - if (ipoib_ib_dev_open(dev, 1)) { + if (ipoib_ib_dev_open(dev)) { ipoib_transport_dev_cleanup(dev); return -ENODEV; } @@ -1037,15 +1045,16 @@ static void __ipoib_ib_dev_flush(struct ipoib_dev_priv *priv, if (level == IPOIB_FLUSH_LIGHT) { ipoib_mark_paths_invalid(dev); ipoib_mcast_dev_flush(dev); + ipoib_flush_ah(dev); } if (level >= IPOIB_FLUSH_NORMAL) - ipoib_ib_dev_down(dev, 0); + ipoib_ib_dev_down(dev); if (level == IPOIB_FLUSH_HEAVY) { if (test_bit(IPOIB_FLAG_INITIALIZED, &priv->flags)) - ipoib_ib_dev_stop(dev, 0); - if (ipoib_ib_dev_open(dev, 0) != 0) + ipoib_ib_dev_stop(dev); + if (ipoib_ib_dev_open(dev) != 0) return; if (netif_queue_stopped(dev)) netif_start_queue(dev); @@ -1097,9 +1106,17 @@ void ipoib_ib_dev_cleanup(struct net_device *dev) */ ipoib_flush_paths(dev); - ipoib_mcast_stop_thread(dev, 1); + ipoib_mcast_stop_thread(dev); ipoib_mcast_dev_flush(dev); + /* + * All of our ah references aren't free until after + * ipoib_mcast_dev_flush(), ipoib_flush_paths, and + * the neighbor garbage collection is stopped and reaped. + * That should all be done now, so make a final ah flush. + */ + ipoib_stop_ah(dev); + ipoib_transport_dev_cleanup(dev); } |