net/vmxnet3: convert to new Rx offload API
authorLouis Luo <llouis@vmware.com>
Tue, 1 May 2018 21:22:56 +0000 (14:22 -0700)
committerFerruh Yigit <ferruh.yigit@intel.com>
Wed, 2 May 2018 17:28:48 +0000 (19:28 +0200)
Ethdev RX offloads API has changed since: commit ce17eddefc20
("ethdev: introduce Rx queue offloads API")

This patch adopts the new RX Offload API in vmxnet3 driver.

Signed-off-by: Louis Luo <llouis@vmware.com>
Acked-by: Yong Wang <yongwang@vmware.com>
drivers/net/vmxnet3/vmxnet3_ethdev.c

index b5fb39e..c850241 100644 (file)
 
 #define        VMXNET3_TX_MAX_SEG      UINT8_MAX
 
+#define VMXNET3_TX_OFFLOAD_CAP         \
+       (DEV_TX_OFFLOAD_VLAN_INSERT |   \
+        DEV_TX_OFFLOAD_IPV4_CKSUM |    \
+        DEV_TX_OFFLOAD_TCP_CKSUM |     \
+        DEV_TX_OFFLOAD_UDP_CKSUM |     \
+        DEV_TX_OFFLOAD_TCP_TSO |       \
+        DEV_TX_OFFLOAD_MULTI_SEGS)
+
+#define VMXNET3_RX_OFFLOAD_CAP         \
+       (DEV_RX_OFFLOAD_VLAN_STRIP |    \
+        DEV_RX_OFFLOAD_SCATTER |       \
+        DEV_RX_OFFLOAD_IPV4_CKSUM |    \
+        DEV_RX_OFFLOAD_UDP_CKSUM |     \
+        DEV_RX_OFFLOAD_TCP_CKSUM |     \
+        DEV_RX_OFFLOAD_TCP_LRO |       \
+        DEV_RX_OFFLOAD_JUMBO_FRAME)
+
 static int eth_vmxnet3_dev_init(struct rte_eth_dev *eth_dev);
 static int eth_vmxnet3_dev_uninit(struct rte_eth_dev *eth_dev);
 static int vmxnet3_dev_configure(struct rte_eth_dev *dev);
@@ -376,9 +393,25 @@ vmxnet3_dev_configure(struct rte_eth_dev *dev)
        const struct rte_memzone *mz;
        struct vmxnet3_hw *hw = dev->data->dev_private;
        size_t size;
+       uint64_t rx_offloads = dev->data->dev_conf.rxmode.offloads;
+       uint64_t tx_offloads = dev->data->dev_conf.txmode.offloads;
 
        PMD_INIT_FUNC_TRACE();
 
+       if ((rx_offloads & VMXNET3_RX_OFFLOAD_CAP) != rx_offloads) {
+               RTE_LOG(ERR, PMD, "Requested RX offloads 0x%" PRIx64
+                       " do not match supported 0x%" PRIx64,
+                       rx_offloads, (uint64_t)VMXNET3_RX_OFFLOAD_CAP);
+               return -ENOTSUP;
+       }
+
+       if ((tx_offloads & VMXNET3_TX_OFFLOAD_CAP) != tx_offloads) {
+               RTE_LOG(ERR, PMD, "Requested TX offloads 0x%" PRIx64
+                       " do not match supported 0x%" PRIx64,
+                       tx_offloads, (uint64_t)VMXNET3_TX_OFFLOAD_CAP);
+               return -ENOTSUP;
+       }
+
        if (dev->data->nb_tx_queues > VMXNET3_MAX_TX_QUEUES ||
            dev->data->nb_rx_queues > VMXNET3_MAX_RX_QUEUES) {
                PMD_INIT_LOG(ERR, "ERROR: Number of queues not supported");
@@ -567,6 +600,7 @@ vmxnet3_setup_driver_shared(struct rte_eth_dev *dev)
        uint32_t mtu = dev->data->mtu;
        Vmxnet3_DriverShared *shared = hw->shared;
        Vmxnet3_DSDevRead *devRead = &shared->devRead;
+       uint64_t rx_offloads = dev->data->dev_conf.rxmode.offloads;
        uint32_t i;
        int ret;
 
@@ -646,10 +680,10 @@ vmxnet3_setup_driver_shared(struct rte_eth_dev *dev)
        devRead->rxFilterConf.rxMode = 0;
 
        /* Setting up feature flags */
-       if (dev->data->dev_conf.rxmode.hw_ip_checksum)
+       if (rx_offloads & DEV_RX_OFFLOAD_CHECKSUM)
                devRead->misc.uptFeatures |= VMXNET3_F_RXCSUM;
 
-       if (dev->data->dev_conf.rxmode.enable_lro) {
+       if (rx_offloads & DEV_RX_OFFLOAD_TCP_LRO) {
                devRead->misc.uptFeatures |= VMXNET3_F_LRO;
                devRead->misc.maxNumRxSG = 0;
        }
@@ -1052,17 +1086,10 @@ vmxnet3_dev_info_get(struct rte_eth_dev *dev __rte_unused,
                .nb_mtu_seg_max = VMXNET3_MAX_TXD_PER_PKT,
        };
 
-       dev_info->rx_offload_capa =
-               DEV_RX_OFFLOAD_VLAN_STRIP |
-               DEV_RX_OFFLOAD_UDP_CKSUM |
-               DEV_RX_OFFLOAD_TCP_CKSUM |
-               DEV_RX_OFFLOAD_TCP_LRO;
-
-       dev_info->tx_offload_capa =
-               DEV_TX_OFFLOAD_VLAN_INSERT |
-               DEV_TX_OFFLOAD_TCP_CKSUM |
-               DEV_TX_OFFLOAD_UDP_CKSUM |
-               DEV_TX_OFFLOAD_TCP_TSO;
+       dev_info->rx_offload_capa = VMXNET3_RX_OFFLOAD_CAP;
+       dev_info->rx_queue_offload_capa = 0;
+       dev_info->tx_offload_capa = VMXNET3_TX_OFFLOAD_CAP;
+       dev_info->tx_queue_offload_capa = 0;
 }
 
 static const uint32_t *
@@ -1156,8 +1183,9 @@ vmxnet3_dev_promiscuous_disable(struct rte_eth_dev *dev)
 {
        struct vmxnet3_hw *hw = dev->data->dev_private;
        uint32_t *vf_table = hw->shared->devRead.rxFilterConf.vfTable;
+       uint64_t rx_offloads = dev->data->dev_conf.rxmode.offloads;
 
-       if (dev->data->dev_conf.rxmode.hw_vlan_filter)
+       if (rx_offloads & DEV_RX_OFFLOAD_VLAN_FILTER)
                memcpy(vf_table, hw->shadow_vfta, VMXNET3_VFT_TABLE_SIZE);
        else
                memset(vf_table, 0xff, VMXNET3_VFT_TABLE_SIZE);
@@ -1219,9 +1247,10 @@ vmxnet3_dev_vlan_offload_set(struct rte_eth_dev *dev, int mask)
        struct vmxnet3_hw *hw = dev->data->dev_private;
        Vmxnet3_DSDevRead *devRead = &hw->shared->devRead;
        uint32_t *vf_table = devRead->rxFilterConf.vfTable;
+       uint64_t rx_offloads = dev->data->dev_conf.rxmode.offloads;
 
        if (mask & ETH_VLAN_STRIP_MASK) {
-               if (dev->data->dev_conf.rxmode.hw_vlan_strip)
+               if (rx_offloads & DEV_RX_OFFLOAD_VLAN_STRIP)
                        devRead->misc.uptFeatures |= UPT1_F_RXVLAN;
                else
                        devRead->misc.uptFeatures &= ~UPT1_F_RXVLAN;
@@ -1231,7 +1260,7 @@ vmxnet3_dev_vlan_offload_set(struct rte_eth_dev *dev, int mask)
        }
 
        if (mask & ETH_VLAN_FILTER_MASK) {
-               if (dev->data->dev_conf.rxmode.hw_vlan_filter)
+               if (rx_offloads & DEV_RX_OFFLOAD_VLAN_FILTER)
                        memcpy(vf_table, hw->shadow_vfta, VMXNET3_VFT_TABLE_SIZE);
                else
                        memset(vf_table, 0xff, VMXNET3_VFT_TABLE_SIZE);