net/enic: support multicast filtering
authorHyong Youb Kim <hyonkim@cisco.com>
Mon, 10 Dec 2018 18:28:55 +0000 (10:28 -0800)
committerFerruh Yigit <ferruh.yigit@intel.com>
Thu, 13 Dec 2018 17:52:54 +0000 (17:52 +0000)
The VIC hardware has 64 MAC filters per vNIC, which can be either
unicast or multicast. Use one half for unicast and the other half for
multicast, as the VIC kernel drivers for Linux and Windows do.

Signed-off-by: Hyong Youb Kim <hyonkim@cisco.com>
Reviewed-by: John Daley <johndale@cisco.com>
doc/guides/nics/features/enic.ini
doc/guides/rel_notes/release_19_02.rst
drivers/net/enic/enic.h
drivers/net/enic/enic_ethdev.c
drivers/net/enic/enic_main.c

index 020b75c..c6d3749 100644 (file)
@@ -15,7 +15,7 @@ TSO                  = Y
 Promiscuous mode     = Y
 Allmulticast mode    = Y
 Unicast MAC filter   = Y
-Multicast MAC filter =
+Multicast MAC filter = Y
 RSS hash             = Y
 RSS key update       = Y
 RSS reta update      = Y
index a0be086..e86ef95 100644 (file)
@@ -58,6 +58,7 @@ New Features
 
   * Added support for ``RTE_ETH_DEV_CLOSE_REMOVE`` flag.
   * Added the handler to get firmware version string.
+  * Added support for multicast filtering.
 
 
 Removed Items
index 7bca3ca..6c497e9 100644 (file)
@@ -25,8 +25,6 @@
 #define DRV_DESCRIPTION                "Cisco VIC Ethernet NIC Poll-mode Driver"
 #define DRV_COPYRIGHT          "Copyright 2008-2015 Cisco Systems, Inc"
 
-#define ENIC_MAX_MAC_ADDR      64
-
 #define VLAN_ETH_HLEN           18
 
 #define ENICPMD_SETTING(enic, f) ((enic->config.flags & VENETF_##f) ? 1 : 0)
@@ -196,6 +194,10 @@ struct enic {
        uint64_t tx_offload_capa; /* DEV_TX_OFFLOAD flags */
        uint64_t tx_queue_offload_capa; /* DEV_TX_OFFLOAD flags */
        uint64_t tx_offload_mask; /* PKT_TX flags accepted */
+
+       /* Multicast MAC addresses added to the NIC */
+       uint32_t mc_count;
+       struct ether_addr mc_addrs[ENIC_MULTICAST_PERFECT_FILTERS];
 };
 
 /* Compute ethdev's max packet size from MTU */
index 9b206e8..fbbec22 100644 (file)
@@ -367,6 +367,7 @@ static int enicpmd_dev_configure(struct rte_eth_dev *eth_dev)
                return ret;
        }
 
+       enic->mc_count = 0;
        enic->hw_ip_checksum = !!(eth_dev->data->dev_conf.rxmode.offloads &
                                  DEV_RX_OFFLOAD_CHECKSUM);
        /* All vlan offload masks to apply the current settings */
@@ -473,7 +474,7 @@ static void enicpmd_dev_info_get(struct rte_eth_dev *eth_dev,
         * ignoring vNIC mtu.
         */
        device_info->max_rx_pktlen = enic_mtu_to_max_rx_pktlen(enic->max_mtu);
-       device_info->max_mac_addrs = ENIC_MAX_MAC_ADDR;
+       device_info->max_mac_addrs = ENIC_UNICAST_PERFECT_FILTERS;
        device_info->rx_offload_capa = enic->rx_offload_capa;
        device_info->tx_offload_capa = enic->tx_offload_capa;
        device_info->tx_queue_offload_capa = enic->tx_queue_offload_capa;
@@ -643,6 +644,98 @@ static int enicpmd_set_mac_addr(struct rte_eth_dev *eth_dev,
        return enic_set_mac_address(enic, addr->addr_bytes);
 }
 
+static void debug_log_add_del_addr(struct ether_addr *addr, bool add)
+{
+       char mac_str[ETHER_ADDR_FMT_SIZE];
+       if (rte_log_get_level(enicpmd_logtype_init) == RTE_LOG_DEBUG) {
+               ether_format_addr(mac_str, ETHER_ADDR_FMT_SIZE, addr);
+               PMD_INIT_LOG(ERR, " %s address %s\n",
+                            add ? "add" : "remove", mac_str);
+       }
+}
+
+static int enicpmd_set_mc_addr_list(struct rte_eth_dev *eth_dev,
+                                   struct ether_addr *mc_addr_set,
+                                   uint32_t nb_mc_addr)
+{
+       struct enic *enic = pmd_priv(eth_dev);
+       char mac_str[ETHER_ADDR_FMT_SIZE];
+       struct ether_addr *addr;
+       uint32_t i, j;
+       int ret;
+
+       ENICPMD_FUNC_TRACE();
+
+       /* Validate the given addresses first */
+       for (i = 0; i < nb_mc_addr && mc_addr_set != NULL; i++) {
+               addr = &mc_addr_set[i];
+               if (!is_multicast_ether_addr(addr) ||
+                   is_broadcast_ether_addr(addr)) {
+                       ether_format_addr(mac_str, ETHER_ADDR_FMT_SIZE, addr);
+                       PMD_INIT_LOG(ERR, " invalid multicast address %s\n",
+                                    mac_str);
+                       return -EINVAL;
+               }
+       }
+
+       /* Flush all if requested */
+       if (nb_mc_addr == 0 || mc_addr_set == NULL) {
+               PMD_INIT_LOG(DEBUG, " flush multicast addresses\n");
+               for (i = 0; i < enic->mc_count; i++) {
+                       addr = &enic->mc_addrs[i];
+                       debug_log_add_del_addr(addr, false);
+                       ret = vnic_dev_del_addr(enic->vdev, addr->addr_bytes);
+                       if (ret)
+                               return ret;
+               }
+               enic->mc_count = 0;
+               return 0;
+       }
+
+       if (nb_mc_addr > ENIC_MULTICAST_PERFECT_FILTERS) {
+               PMD_INIT_LOG(ERR, " too many multicast addresses: max=%d\n",
+                            ENIC_MULTICAST_PERFECT_FILTERS);
+               return -ENOSPC;
+       }
+       /*
+        * devcmd is slow, so apply the difference instead of flushing and
+        * adding everything.
+        * 1. Delete addresses on the NIC but not on the host
+        */
+       for (i = 0; i < enic->mc_count; i++) {
+               addr = &enic->mc_addrs[i];
+               for (j = 0; j < nb_mc_addr; j++) {
+                       if (is_same_ether_addr(addr, &mc_addr_set[j]))
+                               break;
+               }
+               if (j < nb_mc_addr)
+                       continue;
+               debug_log_add_del_addr(addr, false);
+               ret = vnic_dev_del_addr(enic->vdev, addr->addr_bytes);
+               if (ret)
+                       return ret;
+       }
+       /* 2. Add addresses on the host but not on the NIC */
+       for (i = 0; i < nb_mc_addr; i++) {
+               addr = &mc_addr_set[i];
+               for (j = 0; j < enic->mc_count; j++) {
+                       if (is_same_ether_addr(addr, &enic->mc_addrs[j]))
+                               break;
+               }
+               if (j < enic->mc_count)
+                       continue;
+               debug_log_add_del_addr(addr, true);
+               ret = vnic_dev_add_addr(enic->vdev, addr->addr_bytes);
+               if (ret)
+                       return ret;
+       }
+       /* Keep a copy so we can flush/apply later on.. */
+       memcpy(enic->mc_addrs, mc_addr_set,
+              nb_mc_addr * sizeof(struct ether_addr));
+       enic->mc_count = nb_mc_addr;
+       return 0;
+}
+
 static int enicpmd_mtu_set(struct rte_eth_dev *eth_dev, uint16_t mtu)
 {
        struct enic *enic = pmd_priv(eth_dev);
@@ -951,6 +1044,7 @@ static const struct eth_dev_ops enicpmd_eth_dev_ops = {
        .mac_addr_add         = enicpmd_add_mac_addr,
        .mac_addr_remove      = enicpmd_remove_mac_addr,
        .mac_addr_set         = enicpmd_set_mac_addr,
+       .set_mc_addr_list     = enicpmd_set_mc_addr_list,
        .filter_ctrl          = enicpmd_dev_filter_ctrl,
        .reta_query           = enicpmd_dev_rss_reta_query,
        .reta_update          = enicpmd_dev_rss_reta_update,
index f457173..621a984 100644 (file)
@@ -1667,8 +1667,9 @@ static int enic_dev_init(struct enic *enic)
        /* Get the supported filters */
        enic_fdir_info(enic);
 
-       eth_dev->data->mac_addrs = rte_zmalloc("enic_mac_addr", ETH_ALEN
-                                               * ENIC_MAX_MAC_ADDR, 0);
+       eth_dev->data->mac_addrs = rte_zmalloc("enic_mac_addr",
+                                       sizeof(struct ether_addr) *
+                                       ENIC_UNICAST_PERFECT_FILTERS, 0);
        if (!eth_dev->data->mac_addrs) {
                dev_err(enic, "mac addr storage alloc failed, aborting.\n");
                return -1;