net/vhost: convert to new Rx/Tx offload API

Signed-off-by: Maxime Coquelin <maxime.coquelin@redhat.com>
Reviewed-by: Tiwei Bie <tiwei.bie@intel.com>

diff --git a/drivers/net/vhost/rte_eth_vhost.c b/drivers/net/vhost/rte_eth_vhost.c
index f473bbb..ba9d768 100644 (file)
--- a/drivers/net/vhost/rte_eth_vhost.c
+++ b/drivers/net/vhost/rte_eth_vhost.c
@@ -491,12 +491,7 @@ eth_dev_configure(struct rte_eth_dev *dev __rte_unused)
        struct pmd_internal *internal = dev->data->dev_private;
        const struct rte_eth_rxmode *rxmode = &dev->data->dev_conf.rxmode;
 
-       internal->vlan_strip = rxmode->hw_vlan_strip;
-
-       if (rxmode->hw_vlan_filter)
-               VHOST_LOG(WARNING,
-                       "vhost(%s): vlan filtering not available\n",
-                       internal->dev_name);
+       internal->vlan_strip = !!(rxmode->offloads & DEV_RX_OFFLOAD_VLAN_STRIP);
 
        return 0;
 }
@@ -1072,6 +1067,10 @@ eth_dev_info(struct rte_eth_dev *dev,
        dev_info->max_rx_queues = internal->max_queues;
        dev_info->max_tx_queues = internal->max_queues;
        dev_info->min_rx_bufsize = 0;
+
+       dev_info->tx_offload_capa = DEV_TX_OFFLOAD_MULTI_SEGS |
+                               DEV_TX_OFFLOAD_VLAN_INSERT;
+       dev_info->rx_offload_capa = DEV_RX_OFFLOAD_VLAN_STRIP;
 }
 
 static int