net/vhost: emulate device start/stop behavior
authorChas Williams <ciwillia@brocade.com>
Tue, 3 Jan 2017 16:22:43 +0000 (11:22 -0500)
committerYuanhan Liu <yuanhan.liu@linux.intel.com>
Tue, 17 Jan 2017 08:20:18 +0000 (09:20 +0100)
.dev_start()/.dev_stop() roughly corresponds to the local device's port
being ready.  This is different from the remote client being connected
which is roughly link up or down.  Emulate the device start/stop behavior
by separately tracking the start/stop state to determine if we should
allow packets to be queued to/from the remote client.

Signed-off-by: Chas Williams <ciwillia@brocade.com>
Acked-by: Yuanhan Liu <yuanhan.liu@linux.intel.com>
drivers/net/vhost/rte_eth_vhost.c

index 79e4b33..410bfd8 100644 (file)
@@ -112,9 +112,11 @@ struct vhost_queue {
 };
 
 struct pmd_internal {
+       rte_atomic32_t dev_attached;
        char *dev_name;
        char *iface_name;
        uint16_t max_queues;
+       rte_atomic32_t started;
 };
 
 struct internal_list {
@@ -494,6 +496,38 @@ find_internal_resource(char *ifname)
        return list;
 }
 
+static void
+update_queuing_status(struct rte_eth_dev *dev)
+{
+       struct pmd_internal *internal = dev->data->dev_private;
+       struct vhost_queue *vq;
+       unsigned int i;
+       int allow_queuing = 1;
+
+       if (rte_atomic32_read(&internal->started) == 0 ||
+           rte_atomic32_read(&internal->dev_attached) == 0)
+               allow_queuing = 0;
+
+       /* Wait until rx/tx_pkt_burst stops accessing vhost device */
+       for (i = 0; i < dev->data->nb_rx_queues; i++) {
+               vq = dev->data->rx_queues[i];
+               if (vq == NULL)
+                       continue;
+               rte_atomic32_set(&vq->allow_queuing, allow_queuing);
+               while (rte_atomic32_read(&vq->while_queuing))
+                       rte_pause();
+       }
+
+       for (i = 0; i < dev->data->nb_tx_queues; i++) {
+               vq = dev->data->tx_queues[i];
+               if (vq == NULL)
+                       continue;
+               rte_atomic32_set(&vq->allow_queuing, allow_queuing);
+               while (rte_atomic32_read(&vq->while_queuing))
+                       rte_pause();
+       }
+}
+
 static int
 new_device(int vid)
 {
@@ -545,18 +579,8 @@ new_device(int vid)
 
        eth_dev->data->dev_link.link_status = ETH_LINK_UP;
 
-       for (i = 0; i < eth_dev->data->nb_rx_queues; i++) {
-               vq = eth_dev->data->rx_queues[i];
-               if (vq == NULL)
-                       continue;
-               rte_atomic32_set(&vq->allow_queuing, 1);
-       }
-       for (i = 0; i < eth_dev->data->nb_tx_queues; i++) {
-               vq = eth_dev->data->tx_queues[i];
-               if (vq == NULL)
-                       continue;
-               rte_atomic32_set(&vq->allow_queuing, 1);
-       }
+       rte_atomic32_set(&internal->dev_attached, 1);
+       update_queuing_status(eth_dev);
 
        RTE_LOG(INFO, PMD, "New connection established\n");
 
@@ -569,6 +593,7 @@ static void
 destroy_device(int vid)
 {
        struct rte_eth_dev *eth_dev;
+       struct pmd_internal *internal;
        struct vhost_queue *vq;
        struct internal_list *list;
        char ifname[PATH_MAX];
@@ -582,24 +607,10 @@ destroy_device(int vid)
                return;
        }
        eth_dev = list->eth_dev;
+       internal = eth_dev->data->dev_private;
 
-       /* Wait until rx/tx_pkt_burst stops accessing vhost device */
-       for (i = 0; i < eth_dev->data->nb_rx_queues; i++) {
-               vq = eth_dev->data->rx_queues[i];
-               if (vq == NULL)
-                       continue;
-               rte_atomic32_set(&vq->allow_queuing, 0);
-               while (rte_atomic32_read(&vq->while_queuing))
-                       rte_pause();
-       }
-       for (i = 0; i < eth_dev->data->nb_tx_queues; i++) {
-               vq = eth_dev->data->tx_queues[i];
-               if (vq == NULL)
-                       continue;
-               rte_atomic32_set(&vq->allow_queuing, 0);
-               while (rte_atomic32_read(&vq->while_queuing))
-                       rte_pause();
-       }
+       rte_atomic32_set(&internal->dev_attached, 0);
+       update_queuing_status(eth_dev);
 
        eth_dev->data->dev_link.link_status = ETH_LINK_DOWN;
 
@@ -773,14 +784,23 @@ vhost_driver_session_stop(void)
 }
 
 static int
-eth_dev_start(struct rte_eth_dev *dev __rte_unused)
+eth_dev_start(struct rte_eth_dev *dev)
 {
+       struct pmd_internal *internal = dev->data->dev_private;
+
+       rte_atomic32_set(&internal->started, 1);
+       update_queuing_status(dev);
+
        return 0;
 }
 
 static void
-eth_dev_stop(struct rte_eth_dev *dev __rte_unused)
+eth_dev_stop(struct rte_eth_dev *dev)
 {
+       struct pmd_internal *internal = dev->data->dev_private;
+
+       rte_atomic32_set(&internal->started, 0);
+       update_queuing_status(dev);
 }
 
 static int