net/ixgbe: move set VF functions from the ethdev
authorBernard Iremonger <bernard.iremonger@intel.com>
Tue, 13 Dec 2016 11:40:35 +0000 (11:40 +0000)
committerFerruh Yigit <ferruh.yigit@intel.com>
Tue, 17 Jan 2017 18:39:28 +0000 (19:39 +0100)
Move the following functions from eth_dev_ops to the ixgbe PMD and rename:

ixgbe_set_pool_rx_mode
ixgbe_set_pool_rx
ixgbe_set_pool_tx
ixgbe_set_pool_vlan_filter
ixgbe_set_vf_rate_limit

Rename the functions to the following:

rte_pmd_ixgbe_set_vf_rxmode
rte_pmd_ixgbe_set_vf_rx
rte_pmd_ixgbe_set_vf_tx
rte_pmd_ixgbe_set_vf_vlan_filter
rte_pmd_ixgbe_set_vf_rate_limit

Use public function internally

Signed-off-by: Bernard Iremonger <bernard.iremonger@intel.com>
drivers/net/ixgbe/ixgbe_ethdev.c
drivers/net/ixgbe/rte_pmd_ixgbe.h
drivers/net/ixgbe/rte_pmd_ixgbe_version.map

index 224c947..f93b951 100644 (file)
@@ -2348,7 +2348,8 @@ ixgbe_dev_start(struct rte_eth_dev *dev)
                for (vf = 0; vf < pci_dev->max_vfs; vf++)
                        for (idx = 0; idx < IXGBE_MAX_QUEUE_NUM_PER_VF; idx++)
                                if (vfinfo[vf].tx_rate[idx] != 0)
-                                       ixgbe_set_vf_rate_limit(dev, vf,
+                                       rte_pmd_ixgbe_set_vf_rate_limit(
+                                               dev->data->port_id, vf,
                                                vfinfo[vf].tx_rate[idx],
                                                1 << idx);
        }
@@ -5072,6 +5073,271 @@ rte_pmd_ixgbe_set_vf_vlan_stripq(uint8_t port, uint16_t vf, uint8_t on)
        return 0;
 }
 
+int
+rte_pmd_ixgbe_set_vf_rxmode(uint8_t port, uint16_t vf, uint16_t rx_mask, uint8_t on)
+{
+       int val = 0;
+       struct rte_eth_dev *dev;
+       struct rte_eth_dev_info dev_info;
+       struct ixgbe_hw *hw;
+       uint32_t vmolr;
+
+       RTE_ETH_VALID_PORTID_OR_ERR_RET(port, -ENODEV);
+
+       dev = &rte_eth_devices[port];
+       rte_eth_dev_info_get(port, &dev_info);
+
+       if (strstr(dev_info.driver_name, "ixgbe_vf"))
+               return -ENOTSUP;
+
+       if (vf >= dev_info.max_vfs)
+               return -EINVAL;
+
+       if (on > 1)
+               return -EINVAL;
+
+       hw = IXGBE_DEV_PRIVATE_TO_HW(dev->data->dev_private);
+       vmolr = IXGBE_READ_REG(hw, IXGBE_VMOLR(vf));
+
+       if (hw->mac.type == ixgbe_mac_82598EB) {
+               PMD_INIT_LOG(ERR, "setting VF receive mode set should be done"
+                            " on 82599 hardware and newer");
+               return -ENOTSUP;
+       }
+       if (ixgbe_vmdq_mode_check(hw) < 0)
+               return -ENOTSUP;
+
+       val = ixgbe_convert_vm_rx_mask_to_val(rx_mask, val);
+
+       if (on)
+               vmolr |= val;
+       else
+               vmolr &= ~val;
+
+       IXGBE_WRITE_REG(hw, IXGBE_VMOLR(vf), vmolr);
+
+       return 0;
+}
+
+int
+rte_pmd_ixgbe_set_vf_rx(uint8_t port, uint16_t vf, uint8_t on)
+{
+       struct rte_eth_dev *dev;
+       struct rte_eth_dev_info dev_info;
+       uint32_t reg, addr;
+       uint32_t val;
+       const uint8_t bit1 = 0x1;
+       struct ixgbe_hw *hw;
+
+       RTE_ETH_VALID_PORTID_OR_ERR_RET(port, -ENODEV);
+
+       dev = &rte_eth_devices[port];
+       rte_eth_dev_info_get(port, &dev_info);
+
+       if (strstr(dev_info.driver_name, "ixgbe_vf"))
+               return -ENOTSUP;
+
+       if (vf >= dev_info.max_vfs)
+               return -EINVAL;
+
+       if (on > 1)
+               return -EINVAL;
+
+       hw = IXGBE_DEV_PRIVATE_TO_HW(dev->data->dev_private);
+
+       if (ixgbe_vmdq_mode_check(hw) < 0)
+               return -ENOTSUP;
+
+       /* for vf >= 32, set bit in PFVFRE[1], otherwise PFVFRE[0] */
+       if (vf >= 32) {
+               addr = IXGBE_VFRE(1);
+               val = bit1 << (vf - 32);
+       } else {
+               addr = IXGBE_VFRE(0);
+               val = bit1 << vf;
+       }
+
+       reg = IXGBE_READ_REG(hw, addr);
+
+       if (on)
+               reg |= val;
+       else
+               reg &= ~val;
+
+       IXGBE_WRITE_REG(hw, addr, reg);
+
+       return 0;
+}
+
+int
+rte_pmd_ixgbe_set_vf_tx(uint8_t port, uint16_t vf, uint8_t on)
+{
+       struct rte_eth_dev *dev;
+       struct rte_eth_dev_info dev_info;
+       uint32_t reg, addr;
+       uint32_t val;
+       const uint8_t bit1 = 0x1;
+
+       struct ixgbe_hw *hw;
+
+       RTE_ETH_VALID_PORTID_OR_ERR_RET(port, -ENODEV);
+
+       dev = &rte_eth_devices[port];
+       rte_eth_dev_info_get(port, &dev_info);
+
+       if (strstr(dev_info.driver_name, "ixgbe_vf"))
+               return -ENOTSUP;
+
+       if (vf >= dev_info.max_vfs)
+               return -EINVAL;
+
+       if (on > 1)
+               return -EINVAL;
+
+       hw = IXGBE_DEV_PRIVATE_TO_HW(dev->data->dev_private);
+       if (ixgbe_vmdq_mode_check(hw) < 0)
+               return -ENOTSUP;
+
+       /* for vf >= 32, set bit in PFVFTE[1], otherwise PFVFTE[0] */
+       if (vf >= 32) {
+               addr = IXGBE_VFTE(1);
+               val = bit1 << (vf - 32);
+       } else {
+               addr = IXGBE_VFTE(0);
+               val = bit1 << vf;
+       }
+
+       reg = IXGBE_READ_REG(hw, addr);
+
+       if (on)
+               reg |= val;
+       else
+               reg &= ~val;
+
+       IXGBE_WRITE_REG(hw, addr, reg);
+
+       return 0;
+}
+
+int
+rte_pmd_ixgbe_set_vf_vlan_filter(uint8_t port, uint16_t vlan,
+                       uint64_t vf_mask, uint8_t vlan_on)
+{
+       struct rte_eth_dev *dev;
+       struct rte_eth_dev_info dev_info;
+       int ret = 0;
+       uint16_t vf_idx;
+       struct ixgbe_hw *hw;
+
+       RTE_ETH_VALID_PORTID_OR_ERR_RET(port, -ENODEV);
+
+       dev = &rte_eth_devices[port];
+       rte_eth_dev_info_get(port, &dev_info);
+
+       if (strstr(dev_info.driver_name, "ixgbe_vf"))
+               return -ENOTSUP;
+
+       if ((vlan > ETHER_MAX_VLAN_ID) || (vf_mask == 0))
+               return -EINVAL;
+
+       hw = IXGBE_DEV_PRIVATE_TO_HW(dev->data->dev_private);
+       if (ixgbe_vmdq_mode_check(hw) < 0)
+               return -ENOTSUP;
+
+       for (vf_idx = 0; vf_idx < 64; vf_idx++) {
+               if (vf_mask & ((uint64_t)(1ULL << vf_idx))) {
+                       ret = hw->mac.ops.set_vfta(hw, vlan, vf_idx,
+                                                  vlan_on, false);
+                       if (ret < 0)
+                               return ret;
+               }
+       }
+
+       return ret;
+}
+
+int rte_pmd_ixgbe_set_vf_rate_limit(uint8_t port, uint16_t vf,
+       uint16_t tx_rate, uint64_t q_msk)
+{
+       struct rte_eth_dev *dev;
+       struct rte_eth_dev_info dev_info;
+       struct ixgbe_hw *hw;
+       struct ixgbe_vf_info *vfinfo;
+       struct rte_eth_link link;
+       uint8_t  nb_q_per_pool;
+       uint32_t queue_stride;
+       uint32_t queue_idx, idx = 0, vf_idx;
+       uint32_t queue_end;
+       uint16_t total_rate = 0;
+       struct rte_pci_device *pci_dev;
+
+       RTE_ETH_VALID_PORTID_OR_ERR_RET(port, -ENODEV);
+
+       dev = &rte_eth_devices[port];
+       rte_eth_dev_info_get(port, &dev_info);
+       rte_eth_link_get_nowait(port, &link);
+
+       if (strstr(dev_info.driver_name, "ixgbe_vf"))
+               return -ENOTSUP;
+
+       if (vf >= dev_info.max_vfs)
+               return -EINVAL;
+
+       if (tx_rate > link.link_speed)
+               return -EINVAL;
+
+       if (q_msk == 0)
+               return 0;
+
+       pci_dev = IXGBE_DEV_TO_PCI(dev);
+       hw = IXGBE_DEV_PRIVATE_TO_HW(dev->data->dev_private);
+       vfinfo = *(IXGBE_DEV_PRIVATE_TO_P_VFDATA(dev->data->dev_private));
+       nb_q_per_pool = RTE_ETH_DEV_SRIOV(dev).nb_q_per_pool;
+       queue_stride = IXGBE_MAX_RX_QUEUE_NUM / RTE_ETH_DEV_SRIOV(dev).active;
+       queue_idx = vf * queue_stride;
+       queue_end = queue_idx + nb_q_per_pool - 1;
+       if (queue_end >= hw->mac.max_tx_queues)
+               return -EINVAL;
+
+       if (vfinfo) {
+               for (vf_idx = 0; vf_idx < pci_dev->max_vfs; vf_idx++) {
+                       if (vf_idx == vf)
+                               continue;
+                       for (idx = 0; idx < RTE_DIM(vfinfo[vf_idx].tx_rate);
+                               idx++)
+                               total_rate += vfinfo[vf_idx].tx_rate[idx];
+               }
+       } else {
+               return -EINVAL;
+       }
+
+       /* Store tx_rate for this vf. */
+       for (idx = 0; idx < nb_q_per_pool; idx++) {
+               if (((uint64_t)0x1 << idx) & q_msk) {
+                       if (vfinfo[vf].tx_rate[idx] != tx_rate)
+                               vfinfo[vf].tx_rate[idx] = tx_rate;
+                       total_rate += tx_rate;
+               }
+       }
+
+       if (total_rate > dev->data->dev_link.link_speed) {
+               /* Reset stored TX rate of the VF if it causes exceed
+                * link speed.
+                */
+               memset(vfinfo[vf].tx_rate, 0, sizeof(vfinfo[vf].tx_rate));
+               return -EINVAL;
+       }
+
+       /* Set RTTBCNRC of each queue/pool for vf X  */
+       for (; queue_idx <= queue_end; queue_idx++) {
+               if (0x1 & q_msk)
+                       ixgbe_set_queue_rate_limit(dev, queue_idx, tx_rate);
+               q_msk = q_msk >> 1;
+       }
+
+       return 0;
+}
+
 #define IXGBE_MRCTL_VPME  0x01 /* Virtual Pool Mirroring. */
 #define IXGBE_MRCTL_UPME  0x02 /* Uplink Port Mirroring. */
 #define IXGBE_MRCTL_DPME  0x04 /* Downlink Port Mirroring. */
index 22bc4c7..dade57a 100644 (file)
@@ -282,6 +282,110 @@ int rte_pmd_ixgbe_macsec_select_txsa(uint8_t port, uint8_t idx, uint8_t an,
 int rte_pmd_ixgbe_macsec_select_rxsa(uint8_t port, uint8_t idx, uint8_t an,
                uint32_t pn, uint8_t *key);
 
+/**
+* Set RX L2 Filtering mode of a VF of an Ethernet device.
+*
+* @param port
+*   The port identifier of the Ethernet device.
+* @param vf
+*   VF id.
+* @param rx_mode
+*    The RX mode mask, which is one or more of accepting Untagged Packets,
+*    packets that match the PFUTA table, Broadcast and Multicast Promiscuous.
+*    ETH_VMDQ_ACCEPT_UNTAG,ETH_VMDQ_ACCEPT_HASH_UC,
+*    ETH_VMDQ_ACCEPT_BROADCAST and ETH_VMDQ_ACCEPT_MULTICAST will be used
+*    in rx_mode.
+* @param on
+*    1 - Enable a VF RX mode.
+*    0 - Disable a VF RX mode.
+* @return
+*   - (0) if successful.
+*   - (-ENOTSUP) if hardware doesn't support.
+*   - (-ENODEV) if *port_id* invalid.
+*   - (-EINVAL) if bad parameter.
+*/
+int
+rte_pmd_ixgbe_set_vf_rxmode(uint8_t port, uint16_t vf, uint16_t rx_mask, uint8_t on);
+
+/**
+* Enable or disable a VF traffic receive of an Ethernet device.
+*
+* @param port
+*   The port identifier of the Ethernet device.
+* @param vf
+*   VF id.
+* @param on
+*    1 - Enable a VF traffic receive.
+*    0 - Disable a VF traffic receive.
+* @return
+*   - (0) if successful.
+*   - (-ENOTSUP) if hardware doesn't support.
+*   - (-ENODEV) if *port_id* invalid.
+*   - (-EINVAL) if bad parameter.
+*/
+int
+rte_pmd_ixgbe_set_vf_rx(uint8_t port, uint16_t vf, uint8_t on);
+
+/**
+* Enable or disable a VF traffic transmit of the Ethernet device.
+*
+* @param port
+*   The port identifier of the Ethernet device.
+* @param vf
+*   VF id.
+* @param on
+*    1 - Enable a VF traffic transmit.
+*    0 - Disable a VF traffic transmit.
+* @return
+*   - (0) if successful.
+*   - (-ENODEV) if *port_id* invalid.
+*   - (-ENOTSUP) if hardware doesn't support.
+*   - (-EINVAL) if bad parameter.
+*/
+int
+rte_pmd_ixgbe_set_vf_tx(uint8_t port, uint16_t vf, uint8_t on);
+
+/**
+* Enable/Disable hardware VF VLAN filtering by an Ethernet device of
+* received VLAN packets tagged with a given VLAN Tag Identifier.
+*
+* @param port id
+*   The port identifier of the Ethernet device.
+* @param vlan_id
+*   The VLAN Tag Identifier whose filtering must be enabled or disabled.
+* @param vf_mask
+*    Bitmap listing which VFs participate in the VLAN filtering.
+* @param vlan_on
+*    1 - Enable VFs VLAN filtering.
+*    0 - Disable VFs VLAN filtering.
+* @return
+*   - (0) if successful.
+*   - (-ENOTSUP) if hardware doesn't support.
+*   - (-ENODEV) if *port_id* invalid.
+*   - (-EINVAL) if bad parameter.
+*/
+int
+rte_pmd_ixgbe_set_vf_vlan_filter(uint8_t port, uint16_t vlan, uint64_t vf_mask, uint8_t vlan_on);
+
+/**
+ * Set the rate limitation for a vf on an Ethernet device.
+ *
+ * @param port_id
+ *   The port identifier of the Ethernet device.
+ * @param vf
+ *   VF id.
+ * @param tx_rate
+ *   The tx rate allocated from the total link speed for this VF id.
+ * @param q_msk
+ *   The queue mask which need to set the rate.
+ * @return
+ *   - (0) if successful.
+ *   - (-ENOTSUP) if hardware doesn't support this feature.
+ *   - (-ENODEV) if *port_id* invalid.
+ *   - (-EINVAL) if bad parameter.
+ */
+int rte_pmd_ixgbe_set_vf_rate_limit(uint8_t port, uint16_t vf, uint16_t tx_rate, uint64_t q_msk);
+
 /**
  * Response sent back to ixgbe driver from user app after callback
  */
index 6d68934..a992dfd 100644 (file)
@@ -19,10 +19,15 @@ DPDK_16.11 {
 DPDK_17.02 {
        global:
 
-       rte_pmd_ixgbe_macsec_enable;
-       rte_pmd_ixgbe_macsec_disable;
-       rte_pmd_ixgbe_macsec_config_txsc;
        rte_pmd_ixgbe_macsec_config_rxsc;
-       rte_pmd_ixgbe_macsec_select_txsa;
+       rte_pmd_ixgbe_macsec_config_txsc;
+       rte_pmd_ixgbe_macsec_disable;
+       rte_pmd_ixgbe_macsec_enable;
        rte_pmd_ixgbe_macsec_select_rxsa;
+       rte_pmd_ixgbe_macsec_select_txsa;
+       rte_pmd_ixgbe_set_vf_rate_limit;
+       rte_pmd_ixgbe_set_vf_rx;
+       rte_pmd_ixgbe_set_vf_rxmode;
+       rte_pmd_ixgbe_set_vf_tx;
+       rte_pmd_ixgbe_set_vf_vlan_filter;
 } DPDK_16.11;