Re: [dpdk-dev] [dpdk-stable] [PATCH] net/i40e: fix queue related exception handling

[Ye Xiaolong <xiaolong.ye@intel.com>]

On 05/13, Qiming Yang wrote:
>There should have different behavior in queue start fail and stop fail case.
>When queue start fail, all the next actions should be terminated and then
>started queues should be cleared. But for queue stop stage, one queue stop
>fail should not end other queues stop. This patch fixed that issue in PF
>and VF.
>
>Fixes: b6583ee40265 ("i40e: full VMDQ pools support")
>Fixes: 3f6a696f1054 ("i40evf: queue start and stop")
>
>Signed-off-by: Qiming Yang <qiming.yang@intel.com>
>---
> drivers/net/i40e/i40e_ethdev.c    | 116 ++++++++------------------------------
> drivers/net/i40e/i40e_ethdev_vf.c |   2 -
> drivers/net/i40e/i40e_rxtx.c      |  28 +++++++++
> 3 files changed, 53 insertions(+), 93 deletions(-)
>
>diff --git a/drivers/net/i40e/i40e_ethdev.c b/drivers/net/i40e/i40e_ethdev.c
>index 9fbda1c..024178d 100644
>--- a/drivers/net/i40e/i40e_ethdev.c
>+++ b/drivers/net/i40e/i40e_ethdev.c
>@@ -2276,6 +2276,7 @@ i40e_dev_start(struct rte_eth_dev *dev)
> 	struct rte_intr_handle *intr_handle = &pci_dev->intr_handle;
> 	uint32_t intr_vector = 0;
> 	struct i40e_vsi *vsi;
>+	uint16_t nb_rxq, nb_txq;
> 
> 	hw->adapter_stopped = 0;
> 
>@@ -2307,7 +2308,7 @@ i40e_dev_start(struct rte_eth_dev *dev)
> 	ret = i40e_dev_rxtx_init(pf);
> 	if (ret != I40E_SUCCESS) {
> 		PMD_DRV_LOG(ERR, "Failed to init rx/tx queues");
>-		goto err_up;
>+		return ret;
> 	}
> 
> 	/* Map queues with MSIX interrupt */
>@@ -2332,10 +2333,16 @@ i40e_dev_start(struct rte_eth_dev *dev)
> 	}
> 
> 	/* Enable all queues which have been configured */
>-	ret = i40e_dev_switch_queues(pf, TRUE);
>-	if (ret != I40E_SUCCESS) {
>-		PMD_DRV_LOG(ERR, "Failed to enable VSI");
>-		goto err_up;
>+	for (nb_rxq = 0; nb_rxq < dev->data->nb_rx_queues; nb_rxq++) {
>+		ret = i40e_dev_rx_queue_start(dev, nb_rxq);
>+		if (ret)
>+			goto rx_err;
>+	}
>+
>+	for (nb_txq = 0; nb_txq < dev->data->nb_tx_queues; nb_txq++) {
>+		ret = i40e_dev_tx_queue_start(dev, nb_txq);
>+		if (ret)
>+			goto tx_err;
> 	}
> 
> 	/* Enable receiving broadcast packets */
>@@ -2365,7 +2372,7 @@ i40e_dev_start(struct rte_eth_dev *dev)
> 		ret = i40e_aq_set_lb_modes(hw, dev->data->dev_conf.lpbk_mode, NULL);
> 		if (ret != I40E_SUCCESS) {
> 			PMD_DRV_LOG(ERR, "fail to set loopback link");
>-			goto err_up;
>+			goto tx_err;
> 		}
> 	}
> 
>@@ -2373,7 +2380,7 @@ i40e_dev_start(struct rte_eth_dev *dev)
> 	ret = i40e_apply_link_speed(dev);
> 	if (I40E_SUCCESS != ret) {
> 		PMD_DRV_LOG(ERR, "Fail to apply link setting");
>-		goto err_up;
>+		goto tx_err;
> 	}
> 
> 	if (!rte_intr_allow_others(intr_handle)) {
>@@ -2416,9 +2423,12 @@ i40e_dev_start(struct rte_eth_dev *dev)
> 
> 	return I40E_SUCCESS;
> 
>-err_up:
>-	i40e_dev_switch_queues(pf, FALSE);
>-	i40e_dev_clear_queues(dev);
>+tx_err:
>+	for (i = 0; i < nb_txq; i++)
>+		i40e_dev_tx_queue_stop(dev, i);
>+rx_err:
>+	for (i = 0; i < nb_rxq; i++)
>+		i40e_dev_rx_queue_stop(dev, i);

I think we still need to clear queues in the error handling.

> 
> 	return ret;
> }
>@@ -2442,7 +2452,11 @@ i40e_dev_stop(struct rte_eth_dev *dev)
> 	}
> 
> 	/* Disable all queues */
>-	i40e_dev_switch_queues(pf, FALSE);
>+	for (i = 0; i < dev->data->nb_tx_queues; i++)
>+		i40e_dev_tx_queue_stop(dev, i);
>+
>+	for (i = 0; i < dev->data->nb_rx_queues; i++)
>+		i40e_dev_rx_queue_stop(dev, i);
> 
> 	/* un-map queues with interrupt registers */
> 	i40e_vsi_disable_queues_intr(main_vsi);

[snip]

>diff --git a/drivers/net/i40e/i40e_rxtx.c b/drivers/net/i40e/i40e_rxtx.c
>index f6d23c9..d0bada9 100644
>--- a/drivers/net/i40e/i40e_rxtx.c
>+++ b/drivers/net/i40e/i40e_rxtx.c
>@@ -1570,6 +1570,15 @@ i40e_dev_rx_queue_start(struct rte_eth_dev *dev, uint16_t rx_queue_id)
> 	PMD_INIT_FUNC_TRACE();
> 
> 	rxq = dev->data->rx_queues[rx_queue_id];
>+	if (!rxq || !rxq->q_set) {
>+		PMD_DRV_LOG(ERR, "RX queue %u not available or setup",
>+			    rx_queue_id);
>+		return -EINVAL;
>+	}
>+
>+	if (rxq->rx_deferred_start)
>+		PMD_DRV_LOG(ERR, "RX queue %u is deferrd start",
>+			    rx_queue_id);

Do we need to take any action if rx_deferred_start is set?
Just print an ERR log doesn't make sense to me.

> 
> 	err = i40e_alloc_rx_queue_mbufs(rxq);
> 	if (err) {
>@@ -1602,6 +1611,11 @@ i40e_dev_rx_queue_stop(struct rte_eth_dev *dev, uint16_t rx_queue_id)
> 	struct i40e_hw *hw = I40E_DEV_PRIVATE_TO_HW(dev->data->dev_private);
> 
> 	rxq = dev->data->rx_queues[rx_queue_id];
>+	if (!rxq || !rxq->q_set) {
>+		PMD_DRV_LOG(ERR, "RX queue %u not available or setup",
>+				rx_queue_id);
>+		return -EINVAL;
>+	}
> 
> 	/*
> 	 * rx_queue_id is queue id application refers to, while
>@@ -1630,6 +1644,15 @@ i40e_dev_tx_queue_start(struct rte_eth_dev *dev, uint16_t tx_queue_id)
> 	PMD_INIT_FUNC_TRACE();
> 
> 	txq = dev->data->tx_queues[tx_queue_id];
>+	if (!txq || !txq->q_set) {
>+		PMD_DRV_LOG(ERR, "TX queue %u is not available or setup",
>+			    tx_queue_id);
>+		return -EINVAL;
>+	}
>+
>+	if (txq->tx_deferred_start)
>+		PMD_DRV_LOG(ERR, "TX queue %u is deferrd start",
>+			    tx_queue_id);

Ditto.

Thanks,
Xiaolong

> 
> 	/*
> 	 * tx_queue_id is queue id application refers to, while
>@@ -1654,6 +1677,11 @@ i40e_dev_tx_queue_stop(struct rte_eth_dev *dev, uint16_t tx_queue_id)
> 	struct i40e_hw *hw = I40E_DEV_PRIVATE_TO_HW(dev->data->dev_private);
> 
> 	txq = dev->data->tx_queues[tx_queue_id];
>+	if (!txq || !txq->q_set) {
>+		PMD_DRV_LOG(ERR, "TX queue %u is not available or setup",
>+			tx_queue_id);
>+		return -EINVAL;
>+	}
> 
> 	/*
> 	 * tx_queue_id is queue id application refers to, while
>-- 
>2.9.5
>