fm10k: Add support for MACVLAN acceleration
This patch adds support for L2 MACVLAN by making use of the fact that the
RRC provides a unique tag per filter called a Global Resource Tag, or GLORT.
In the case of this offload what I have done is assigned a linear block of
these so that each GLORT represents one of the MACVLAN netdevs. By doing
this I can share the Rx queues and Tx queues for all of the MACVLAN netdevs
while allowing them to be demuxed in the Rx cleanup path.
Signed-off-by: Alexander Duyck <alexander.h.duyck@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
diff --git a/drivers/net/ethernet/intel/fm10k/fm10k_netdev.c b/drivers/net/ethernet/intel/fm10k/fm10k_netdev.c
index 5dbb4d7..268966b 100644
--- a/drivers/net/ethernet/intel/fm10k/fm10k_netdev.c
+++ b/drivers/net/ethernet/intel/fm10k/fm10k_netdev.c
@@ -1148,6 +1148,155 @@
return 0;
}
+static void fm10k_assign_l2_accel(struct fm10k_intfc *interface,
+ struct fm10k_l2_accel *l2_accel)
+{
+ struct fm10k_ring *ring;
+ int i;
+
+ for (i = 0; i < interface->num_rx_queues; i++) {
+ ring = interface->rx_ring[i];
+ rcu_assign_pointer(ring->l2_accel, l2_accel);
+ }
+
+ interface->l2_accel = l2_accel;
+}
+
+static void *fm10k_dfwd_add_station(struct net_device *dev,
+ struct net_device *sdev)
+{
+ struct fm10k_intfc *interface = netdev_priv(dev);
+ struct fm10k_l2_accel *l2_accel = interface->l2_accel;
+ struct fm10k_l2_accel *old_l2_accel = NULL;
+ struct fm10k_dglort_cfg dglort = { 0 };
+ struct fm10k_hw *hw = &interface->hw;
+ int size = 0, i;
+ u16 glort;
+
+ /* allocate l2 accel structure if it is not available */
+ if (!l2_accel) {
+ /* verify there is enough free GLORTs to support l2_accel */
+ if (interface->glort_count < 7)
+ return ERR_PTR(-EBUSY);
+
+ size = offsetof(struct fm10k_l2_accel, macvlan[7]);
+ l2_accel = kzalloc(size, GFP_KERNEL);
+ if (!l2_accel)
+ return ERR_PTR(-ENOMEM);
+
+ l2_accel->size = 7;
+ l2_accel->dglort = interface->glort;
+
+ /* update pointers */
+ fm10k_assign_l2_accel(interface, l2_accel);
+ /* do not expand if we are at our limit */
+ } else if ((l2_accel->count == FM10K_MAX_STATIONS) ||
+ (l2_accel->count == (interface->glort_count - 1))) {
+ return ERR_PTR(-EBUSY);
+ /* expand if we have hit the size limit */
+ } else if (l2_accel->count == l2_accel->size) {
+ old_l2_accel = l2_accel;
+ size = offsetof(struct fm10k_l2_accel,
+ macvlan[(l2_accel->size * 2) + 1]);
+ l2_accel = kzalloc(size, GFP_KERNEL);
+ if (!l2_accel)
+ return ERR_PTR(-ENOMEM);
+
+ memcpy(l2_accel, old_l2_accel,
+ offsetof(struct fm10k_l2_accel,
+ macvlan[old_l2_accel->size]));
+
+ l2_accel->size = (old_l2_accel->size * 2) + 1;
+
+ /* update pointers */
+ fm10k_assign_l2_accel(interface, l2_accel);
+ kfree_rcu(old_l2_accel, rcu);
+ }
+
+ /* add macvlan to accel table, and record GLORT for position */
+ for (i = 0; i < l2_accel->size; i++) {
+ if (!l2_accel->macvlan[i])
+ break;
+ }
+
+ /* record station */
+ l2_accel->macvlan[i] = sdev;
+ l2_accel->count++;
+
+ /* configure default DGLORT mapping for RSS/DCB */
+ dglort.idx = fm10k_dglort_pf_rss;
+ dglort.inner_rss = 1;
+ dglort.rss_l = fls(interface->ring_feature[RING_F_RSS].mask);
+ dglort.pc_l = fls(interface->ring_feature[RING_F_QOS].mask);
+ dglort.glort = interface->glort;
+ dglort.shared_l = fls(l2_accel->size);
+ hw->mac.ops.configure_dglort_map(hw, &dglort);
+
+ /* Add rules for this specific dglort to the switch */
+ fm10k_mbx_lock(interface);
+
+ glort = l2_accel->dglort + 1 + i;
+ hw->mac.ops.update_xcast_mode(hw, glort, FM10K_XCAST_MODE_MULTI);
+ hw->mac.ops.update_uc_addr(hw, glort, sdev->dev_addr, 0, true, 0);
+
+ fm10k_mbx_unlock(interface);
+
+ return sdev;
+}
+
+static void fm10k_dfwd_del_station(struct net_device *dev, void *priv)
+{
+ struct fm10k_intfc *interface = netdev_priv(dev);
+ struct fm10k_l2_accel *l2_accel = ACCESS_ONCE(interface->l2_accel);
+ struct fm10k_dglort_cfg dglort = { 0 };
+ struct fm10k_hw *hw = &interface->hw;
+ struct net_device *sdev = priv;
+ int i;
+ u16 glort;
+
+ if (!l2_accel)
+ return;
+
+ /* search table for matching interface */
+ for (i = 0; i < l2_accel->size; i++) {
+ if (l2_accel->macvlan[i] == sdev)
+ break;
+ }
+
+ /* exit if macvlan not found */
+ if (i == l2_accel->size)
+ return;
+
+ /* Remove any rules specific to this dglort */
+ fm10k_mbx_lock(interface);
+
+ glort = l2_accel->dglort + 1 + i;
+ hw->mac.ops.update_xcast_mode(hw, glort, FM10K_XCAST_MODE_NONE);
+ hw->mac.ops.update_uc_addr(hw, glort, sdev->dev_addr, 0, false, 0);
+
+ fm10k_mbx_unlock(interface);
+
+ /* record removal */
+ l2_accel->macvlan[i] = NULL;
+ l2_accel->count--;
+
+ /* configure default DGLORT mapping for RSS/DCB */
+ dglort.idx = fm10k_dglort_pf_rss;
+ dglort.inner_rss = 1;
+ dglort.rss_l = fls(interface->ring_feature[RING_F_RSS].mask);
+ dglort.pc_l = fls(interface->ring_feature[RING_F_QOS].mask);
+ dglort.glort = interface->glort;
+ if (l2_accel)
+ dglort.shared_l = fls(l2_accel->size);
+ hw->mac.ops.configure_dglort_map(hw, &dglort);
+
+ /* If table is empty remove it */
+ if (l2_accel->count == 0) {
+ fm10k_assign_l2_accel(interface, NULL);
+ kfree_rcu(l2_accel, rcu);
+ }
+}
+
static const struct net_device_ops fm10k_netdev_ops = {
.ndo_open = fm10k_open,
.ndo_stop = fm10k_close,
@@ -1163,6 +1312,8 @@
.ndo_setup_tc = fm10k_setup_tc,
.ndo_add_vxlan_port = fm10k_add_vxlan_port,
.ndo_del_vxlan_port = fm10k_del_vxlan_port,
+ .ndo_dfwd_add_station = fm10k_dfwd_add_station,
+ .ndo_dfwd_del_station = fm10k_dfwd_del_station,
};
#define DEFAULT_DEBUG_LEVEL_SHIFT 3
@@ -1198,6 +1349,9 @@
/* all features defined to this point should be changeable */
dev->hw_features |= dev->features;
+ /* allow user to enable L2 forwarding acceleration */
+ dev->hw_features |= NETIF_F_HW_L2FW_DOFFLOAD;
+
/* configure VLAN features */
dev->vlan_features |= dev->features;