net/ena: fix per-queue offload capabilities
authorMichal Krawczyk <mk@semihalf.com>
Tue, 19 Oct 2021 10:56:25 +0000 (12:56 +0200)
committerFerruh Yigit <ferruh.yigit@intel.com>
Tue, 19 Oct 2021 13:04:17 +0000 (15:04 +0200)
As ENA currently doesn't support offloads which could be configured
per-queue, only per-port flags should be set.

In addition, to make the code cleaner, parsing appropriate offload
flags is encapsulated into helper functions, in a similar matter it's
done by the other PMDs.

[1] https://doc.dpdk.org/guides/prog_guide/
    poll_mode_drv.html?highlight=offloads#hardware-offload

Fixes: 7369f88f88c0 ("net/ena: convert to new Rx offloads API")
Fixes: 56b8b9b7e5d2 ("net/ena: convert to new Tx offloads API")
Cc: stable@dpdk.org
Signed-off-by: Michal Krawczyk <mk@semihalf.com>
Reviewed-by: Igor Chauskin <igorch@amazon.com>
Reviewed-by: Shai Brandes <shaibran@amazon.com>
drivers/net/ena/ena_ethdev.c

index fe9bac8..655c53b 100644 (file)
@@ -223,6 +223,10 @@ static int ena_queue_start(struct rte_eth_dev *dev, struct ena_ring *ring);
 static int ena_queue_start_all(struct rte_eth_dev *dev,
                               enum ena_ring_type ring_type);
 static void ena_stats_restart(struct rte_eth_dev *dev);
+static uint64_t ena_get_rx_port_offloads(struct ena_adapter *adapter);
+static uint64_t ena_get_tx_port_offloads(struct ena_adapter *adapter);
+static uint64_t ena_get_rx_queue_offloads(struct ena_adapter *adapter);
+static uint64_t ena_get_tx_queue_offloads(struct ena_adapter *adapter);
 static int ena_infos_get(struct rte_eth_dev *dev,
                         struct rte_eth_dev_info *dev_info);
 static void ena_interrupt_handler_rte(void *cb_arg);
@@ -1947,12 +1951,63 @@ static void ena_init_rings(struct ena_adapter *adapter,
        }
 }
 
+static uint64_t ena_get_rx_port_offloads(struct ena_adapter *adapter)
+{
+       uint64_t port_offloads = 0;
+
+       if (adapter->offloads.rx_offloads & ENA_L3_IPV4_CSUM)
+               port_offloads |= DEV_RX_OFFLOAD_IPV4_CKSUM;
+
+       if (adapter->offloads.rx_offloads &
+           (ENA_L4_IPV4_CSUM | ENA_L4_IPV6_CSUM))
+               port_offloads |=
+                       DEV_RX_OFFLOAD_UDP_CKSUM | DEV_RX_OFFLOAD_TCP_CKSUM;
+
+       if (adapter->offloads.rx_offloads & ENA_RX_RSS_HASH)
+               port_offloads |= DEV_RX_OFFLOAD_RSS_HASH;
+
+       return port_offloads;
+}
+
+static uint64_t ena_get_tx_port_offloads(struct ena_adapter *adapter)
+{
+       uint64_t port_offloads = 0;
+
+       if (adapter->offloads.tx_offloads & ENA_IPV4_TSO)
+               port_offloads |= DEV_TX_OFFLOAD_TCP_TSO;
+
+       if (adapter->offloads.tx_offloads & ENA_L3_IPV4_CSUM)
+               port_offloads |= DEV_TX_OFFLOAD_IPV4_CKSUM;
+       if (adapter->offloads.tx_offloads &
+           (ENA_L4_IPV4_CSUM_PARTIAL | ENA_L4_IPV4_CSUM |
+            ENA_L4_IPV6_CSUM | ENA_L4_IPV6_CSUM_PARTIAL))
+               port_offloads |=
+                       DEV_TX_OFFLOAD_UDP_CKSUM | DEV_TX_OFFLOAD_TCP_CKSUM;
+
+       port_offloads |= DEV_TX_OFFLOAD_MULTI_SEGS;
+
+       return port_offloads;
+}
+
+static uint64_t ena_get_rx_queue_offloads(struct ena_adapter *adapter)
+{
+       RTE_SET_USED(adapter);
+
+       return 0;
+}
+
+static uint64_t ena_get_tx_queue_offloads(struct ena_adapter *adapter)
+{
+       RTE_SET_USED(adapter);
+
+       return 0;
+}
+
 static int ena_infos_get(struct rte_eth_dev *dev,
                          struct rte_eth_dev_info *dev_info)
 {
        struct ena_adapter *adapter;
        struct ena_com_dev *ena_dev;
-       uint64_t rx_feat = 0, tx_feat = 0;
 
        ena_assert_msg(dev->data != NULL, "Uninitialized device\n");
        ena_assert_msg(dev->data->dev_private != NULL, "Uninitialized device\n");
@@ -1971,32 +2026,11 @@ static int ena_infos_get(struct rte_eth_dev *dev,
                        ETH_LINK_SPEED_50G  |
                        ETH_LINK_SPEED_100G;
 
-       /* Set Tx & Rx features available for device */
-       if (adapter->offloads.tx_offloads & ENA_IPV4_TSO)
-               tx_feat |= DEV_TX_OFFLOAD_TCP_TSO;
-
-       if (adapter->offloads.tx_offloads & ENA_L3_IPV4_CSUM)
-               tx_feat |= DEV_TX_OFFLOAD_IPV4_CKSUM;
-       if (adapter->offloads.tx_offloads &
-           (ENA_L4_IPV4_CSUM_PARTIAL | ENA_L4_IPV4_CSUM |
-            ENA_L4_IPV6_CSUM | ENA_L4_IPV6_CSUM_PARTIAL))
-               tx_feat |= DEV_TX_OFFLOAD_UDP_CKSUM | DEV_TX_OFFLOAD_TCP_CKSUM;
-
-       if (adapter->offloads.rx_offloads & ENA_L3_IPV4_CSUM)
-               rx_feat |= DEV_RX_OFFLOAD_IPV4_CKSUM;
-       if (adapter->offloads.rx_offloads &
-           (ENA_L4_IPV4_CSUM | ENA_L4_IPV6_CSUM))
-               rx_feat |= DEV_RX_OFFLOAD_UDP_CKSUM | DEV_RX_OFFLOAD_TCP_CKSUM;
-
-       tx_feat |= DEV_TX_OFFLOAD_MULTI_SEGS;
-
        /* Inform framework about available features */
-       dev_info->rx_offload_capa = rx_feat;
-       if (adapter->offloads.rx_offloads & ENA_RX_RSS_HASH)
-               dev_info->rx_offload_capa |= DEV_RX_OFFLOAD_RSS_HASH;
-       dev_info->rx_queue_offload_capa = rx_feat;
-       dev_info->tx_offload_capa = tx_feat;
-       dev_info->tx_queue_offload_capa = tx_feat;
+       dev_info->rx_offload_capa = ena_get_rx_port_offloads(adapter);
+       dev_info->tx_offload_capa = ena_get_tx_port_offloads(adapter);
+       dev_info->rx_queue_offload_capa = ena_get_rx_queue_offloads(adapter);
+       dev_info->tx_queue_offload_capa = ena_get_tx_queue_offloads(adapter);
 
        dev_info->flow_type_rss_offloads = ENA_ALL_RSS_HF;
        dev_info->hash_key_size = ENA_HASH_KEY_SIZE;
@@ -2012,9 +2046,6 @@ static int ena_infos_get(struct rte_eth_dev *dev,
        dev_info->max_tx_queues = adapter->max_num_io_queues;
        dev_info->reta_size = ENA_RX_RSS_TABLE_SIZE;
 
-       adapter->tx_supported_offloads = tx_feat;
-       adapter->rx_supported_offloads = rx_feat;
-
        dev_info->rx_desc_lim.nb_max = adapter->max_rx_ring_size;
        dev_info->rx_desc_lim.nb_min = ENA_MIN_RING_DESC;
        dev_info->rx_desc_lim.nb_seg_max = RTE_MIN(ENA_PKT_MAX_BUFS,