IB/mlx4: Add support for IBoE

Add support for IBoE to mlx4_ib.  The bulk of the code is handling the
new address vector fields; mlx4 needs the MAC address of a remote node
to include it in a WQE (for datagrams) or in the QP context (for
connected QPs).  Address resolution is done by assuming all unicast
GIDs are either link-local IPv6 addresses.

Multicast group attach/detach needs to update the NIC's multicast
filters; but since attaching a QP to a multicast group can be done
before the QP is bound to a port, for IBoE we need to keep track of
all multicast groups that a QP is attached too before it transitions
from INIT to RTR (since it does not have a port in the INIT state).

Signed-off-by: Eli Cohen <eli@mellanox.co.il>

[ Many things cleaned up and otherwise monkeyed with; hope I didn't
  introduce too many bugs.  - Roland ]

Signed-off-by: Roland Dreier <rolandd@cisco.com>

1 files changed, 412 insertions, 36 deletions

diff --git a/drivers/infiniband/hw/mlx4/main.c b/drivers/infiniband/hw/mlx4/main.c
index 4e94e360e43b..e65db73fc277 100644
--- a/drivers/infiniband/hw/mlx4/main.c
+++ b/drivers/infiniband/hw/mlx4/main.c
@@ -35,9 +35,13 @@
 #include <linux/init.h>
 #include <linux/slab.h>
 #include <linux/errno.h>
+#include <linux/netdevice.h>
+#include <linux/inetdevice.h>
+#include <linux/rtnetlink.h>
 
 #include <rdma/ib_smi.h>
 #include <rdma/ib_user_verbs.h>
+#include <rdma/ib_addr.h>
 
 #include <linux/mlx4/driver.h>
 #include <linux/mlx4/cmd.h>
@@ -58,6 +62,15 @@ static const char mlx4_ib_version[] =
 	DRV_NAME ": Mellanox ConnectX InfiniBand driver v"
 	DRV_VERSION " (" DRV_RELDATE ")\n";
 
+struct update_gid_work {
+	struct work_struct	work;
+	union ib_gid		gids[128];
+	struct mlx4_ib_dev     *dev;
+	int			port;
+};
+
+static struct workqueue_struct *wq;
+
 static void init_query_mad(struct ib_smp *mad)
 {
 	mad->base_version  = 1;
@@ -154,28 +167,19 @@ out:
 	return err;
 }
 
-static int mlx4_ib_query_port(struct ib_device *ibdev, u8 port,
-			      struct ib_port_attr *props)
+static enum rdma_link_layer
+mlx4_ib_port_link_layer(struct ib_device *device, u8 port_num)
 {
-	struct ib_smp *in_mad  = NULL;
-	struct ib_smp *out_mad = NULL;
-	int err = -ENOMEM;
+	struct mlx4_dev *dev = to_mdev(device)->dev;
 
-	in_mad  = kzalloc(sizeof *in_mad, GFP_KERNEL);
-	out_mad = kmalloc(sizeof *out_mad, GFP_KERNEL);
-	if (!in_mad || !out_mad)
-		goto out;
-
-	memset(props, 0, sizeof *props);
-
-	init_query_mad(in_mad);
-	in_mad->attr_id  = IB_SMP_ATTR_PORT_INFO;
-	in_mad->attr_mod = cpu_to_be32(port);
-
-	err = mlx4_MAD_IFC(to_mdev(ibdev), 1, 1, port, NULL, NULL, in_mad, out_mad);
-	if (err)
-		goto out;
+	return dev->caps.port_mask & (1 << (port_num - 1)) ?
+		IB_LINK_LAYER_INFINIBAND : IB_LINK_LAYER_ETHERNET;
+}
 
+static int ib_link_query_port(struct ib_device *ibdev, u8 port,
+			      struct ib_port_attr *props,
+			      struct ib_smp *out_mad)
+{
 	props->lid		= be16_to_cpup((__be16 *) (out_mad->data + 16));
 	props->lmc		= out_mad->data[34] & 0x7;
 	props->sm_lid		= be16_to_cpup((__be16 *) (out_mad->data + 18));
@@ -196,6 +200,80 @@ static int mlx4_ib_query_port(struct ib_device *ibdev, u8 port,
 	props->max_vl_num	= out_mad->data[37] >> 4;
 	props->init_type_reply	= out_mad->data[41] >> 4;
 
+	return 0;
+}
+
+static u8 state_to_phys_state(enum ib_port_state state)
+{
+	return state == IB_PORT_ACTIVE ? 5 : 3;
+}
+
+static int eth_link_query_port(struct ib_device *ibdev, u8 port,
+			       struct ib_port_attr *props,
+			       struct ib_smp *out_mad)
+{
+	struct mlx4_ib_iboe *iboe = &to_mdev(ibdev)->iboe;
+	struct net_device *ndev;
+	enum ib_mtu tmp;
+
+	props->active_width	= IB_WIDTH_4X;
+	props->active_speed	= 4;
+	props->port_cap_flags	= IB_PORT_CM_SUP;
+	props->gid_tbl_len	= to_mdev(ibdev)->dev->caps.gid_table_len[port];
+	props->max_msg_sz	= to_mdev(ibdev)->dev->caps.max_msg_sz;
+	props->pkey_tbl_len	= 1;
+	props->bad_pkey_cntr	= be16_to_cpup((__be16 *) (out_mad->data + 46));
+	props->qkey_viol_cntr	= be16_to_cpup((__be16 *) (out_mad->data + 48));
+	props->max_mtu		= IB_MTU_2048;
+	props->subnet_timeout	= 0;
+	props->max_vl_num	= out_mad->data[37] >> 4;
+	props->init_type_reply	= 0;
+	props->state		= IB_PORT_DOWN;
+	props->phys_state	= state_to_phys_state(props->state);
+	props->active_mtu	= IB_MTU_256;
+	spin_lock(&iboe->lock);
+	ndev = iboe->netdevs[port - 1];
+	if (!ndev)
+		goto out;
+
+	tmp = iboe_get_mtu(ndev->mtu);
+	props->active_mtu = tmp ? min(props->max_mtu, tmp) : IB_MTU_256;
+
+	props->state		= netif_running(ndev) &&  netif_oper_up(ndev) ?
+					IB_PORT_ACTIVE : IB_PORT_DOWN;
+	props->phys_state	= state_to_phys_state(props->state);
+
+out:
+	spin_unlock(&iboe->lock);
+	return 0;
+}
+
+static int mlx4_ib_query_port(struct ib_device *ibdev, u8 port,
+			      struct ib_port_attr *props)
+{
+	struct ib_smp *in_mad  = NULL;
+	struct ib_smp *out_mad = NULL;
+	int err = -ENOMEM;
+
+	in_mad  = kzalloc(sizeof *in_mad, GFP_KERNEL);
+	out_mad = kmalloc(sizeof *out_mad, GFP_KERNEL);
+	if (!in_mad || !out_mad)
+		goto out;
+
+	memset(props, 0, sizeof *props);
+
+	init_query_mad(in_mad);
+	in_mad->attr_id  = IB_SMP_ATTR_PORT_INFO;
+	in_mad->attr_mod = cpu_to_be32(port);
+
+	err = mlx4_MAD_IFC(to_mdev(ibdev), 1, 1, port, NULL, NULL, in_mad, out_mad);
+	if (err)
+		goto out;
+
+	err = mlx4_ib_port_link_layer(ibdev, port) == IB_LINK_LAYER_INFINIBAND ?
+		ib_link_query_port(ibdev, port, props, out_mad) :
+		eth_link_query_port(ibdev, port, props, out_mad);
+
 out:
 	kfree(in_mad);
 	kfree(out_mad);
@@ -203,8 +281,8 @@ out:
 	return err;
 }
 
-static int mlx4_ib_query_gid(struct ib_device *ibdev, u8 port, int index,
-			     union ib_gid *gid)
+static int __mlx4_ib_query_gid(struct ib_device *ibdev, u8 port, int index,
+			       union ib_gid *gid)
 {
 	struct ib_smp *in_mad  = NULL;
 	struct ib_smp *out_mad = NULL;
@@ -241,6 +319,25 @@ out:
 	return err;
 }
 
+static int iboe_query_gid(struct ib_device *ibdev, u8 port, int index,
+			  union ib_gid *gid)
+{
+	struct mlx4_ib_dev *dev = to_mdev(ibdev);
+
+	*gid = dev->iboe.gid_table[port - 1][index];
+
+	return 0;
+}
+
+static int mlx4_ib_query_gid(struct ib_device *ibdev, u8 port, int index,
+			     union ib_gid *gid)
+{
+	if (rdma_port_get_link_layer(ibdev, port) == IB_LINK_LAYER_INFINIBAND)
+		return __mlx4_ib_query_gid(ibdev, port, index, gid);
+	else
+		return iboe_query_gid(ibdev, port, index, gid);
+}
+
 static int mlx4_ib_query_pkey(struct ib_device *ibdev, u8 port, u16 index,
 			      u16 *pkey)
 {
@@ -289,6 +386,7 @@ static int mlx4_SET_PORT(struct mlx4_ib_dev *dev, u8 port, int reset_qkey_viols,
 {
 	struct mlx4_cmd_mailbox *mailbox;
 	int err;
+	u8 is_eth = dev->dev->caps.port_type[port] == MLX4_PORT_TYPE_ETH;
 
 	mailbox = mlx4_alloc_cmd_mailbox(dev->dev);
 	if (IS_ERR(mailbox))
@@ -304,7 +402,7 @@ static int mlx4_SET_PORT(struct mlx4_ib_dev *dev, u8 port, int reset_qkey_viols,
 		((__be32 *) mailbox->buf)[1] = cpu_to_be32(cap_mask);
 	}
 
-	err = mlx4_cmd(dev->dev, mailbox->dma, port, 0, MLX4_CMD_SET_PORT,
+	err = mlx4_cmd(dev->dev, mailbox->dma, port, is_eth, MLX4_CMD_SET_PORT,
 		       MLX4_CMD_TIME_CLASS_B);
 
 	mlx4_free_cmd_mailbox(dev->dev, mailbox);
@@ -447,18 +545,132 @@ static int mlx4_ib_dealloc_pd(struct ib_pd *pd)
 	return 0;
 }
 
+static int add_gid_entry(struct ib_qp *ibqp, union ib_gid *gid)
+{
+	struct mlx4_ib_qp *mqp = to_mqp(ibqp);
+	struct mlx4_ib_dev *mdev = to_mdev(ibqp->device);
+	struct mlx4_ib_gid_entry *ge;
+
+	ge = kzalloc(sizeof *ge, GFP_KERNEL);
+	if (!ge)
+		return -ENOMEM;
+
+	ge->gid = *gid;
+	if (mlx4_ib_add_mc(mdev, mqp, gid)) {
+		ge->port = mqp->port;
+		ge->added = 1;
+	}
+
+	mutex_lock(&mqp->mutex);
+	list_add_tail(&ge->list, &mqp->gid_list);
+	mutex_unlock(&mqp->mutex);
+
+	return 0;
+}
+
+int mlx4_ib_add_mc(struct mlx4_ib_dev *mdev, struct mlx4_ib_qp *mqp,
+		   union ib_gid *gid)
+{
+	u8 mac[6];
+	struct net_device *ndev;
+	int ret = 0;
+
+	if (!mqp->port)
+		return 0;
+
+	spin_lock(&mdev->iboe.lock);
+	ndev = mdev->iboe.netdevs[mqp->port - 1];
+	if (ndev)
+		dev_hold(ndev);
+	spin_unlock(&mdev->iboe.lock);
+
+	if (ndev) {
+		rdma_get_mcast_mac((struct in6_addr *)gid, mac);
+		rtnl_lock();
+		dev_mc_add(mdev->iboe.netdevs[mqp->port - 1], mac);
+		ret = 1;
+		rtnl_unlock();
+		dev_put(ndev);
+	}
+
+	return ret;
+}
+
 static int mlx4_ib_mcg_attach(struct ib_qp *ibqp, union ib_gid *gid, u16 lid)
 {
-	return mlx4_multicast_attach(to_mdev(ibqp->device)->dev,
-				     &to_mqp(ibqp)->mqp, gid->raw,
-				     !!(to_mqp(ibqp)->flags &
-					MLX4_IB_QP_BLOCK_MULTICAST_LOOPBACK));
+	int err;
+	struct mlx4_ib_dev *mdev = to_mdev(ibqp->device);
+	struct mlx4_ib_qp *mqp = to_mqp(ibqp);
+
+	err = mlx4_multicast_attach(mdev->dev, &mqp->mqp, gid->raw, !!(mqp->flags &
+				    MLX4_IB_QP_BLOCK_MULTICAST_LOOPBACK));
+	if (err)
+		return err;
+
+	err = add_gid_entry(ibqp, gid);
+	if (err)
+		goto err_add;
+
+	return 0;
+
+err_add:
+	mlx4_multicast_detach(mdev->dev, &mqp->mqp, gid->raw);
+	return err;
+}
+
+static struct mlx4_ib_gid_entry *find_gid_entry(struct mlx4_ib_qp *qp, u8 *raw)
+{
+	struct mlx4_ib_gid_entry *ge;
+	struct mlx4_ib_gid_entry *tmp;
+	struct mlx4_ib_gid_entry *ret = NULL;
+
+	list_for_each_entry_safe(ge, tmp, &qp->gid_list, list) {
+		if (!memcmp(raw, ge->gid.raw, 16)) {
+			ret = ge;
+			break;
+		}
+	}
+
+	return ret;
 }
 
 static int mlx4_ib_mcg_detach(struct ib_qp *ibqp, union ib_gid *gid, u16 lid)
 {
-	return mlx4_multicast_detach(to_mdev(ibqp->device)->dev,
-				     &to_mqp(ibqp)->mqp, gid->raw);
+	int err;
+	struct mlx4_ib_dev *mdev = to_mdev(ibqp->device);
+	struct mlx4_ib_qp *mqp = to_mqp(ibqp);
+	u8 mac[6];
+	struct net_device *ndev;
+	struct mlx4_ib_gid_entry *ge;
+
+	err = mlx4_multicast_detach(mdev->dev,
+				    &mqp->mqp, gid->raw);
+	if (err)
+		return err;
+
+	mutex_lock(&mqp->mutex);
+	ge = find_gid_entry(mqp, gid->raw);
+	if (ge) {
+		spin_lock(&mdev->iboe.lock);
+		ndev = ge->added ? mdev->iboe.netdevs[ge->port - 1] : NULL;
+		if (ndev)
+			dev_hold(ndev);
+		spin_unlock(&mdev->iboe.lock);
+		rdma_get_mcast_mac((struct in6_addr *)gid, mac);
+		if (ndev) {
+			rtnl_lock();
+			dev_mc_del(mdev->iboe.netdevs[ge->port - 1], mac);
+			rtnl_unlock();
+			dev_put(ndev);
+		}
+		list_del(&ge->list);
+		kfree(ge);
+	} else
+		printk(KERN_WARNING "could not find mgid entry\n");
+
+	mutex_unlock(&mqp->mutex);
+
+	return 0;
 }
 
 static int init_node_data(struct mlx4_ib_dev *dev)
@@ -543,15 +755,143 @@ static struct device_attribute *mlx4_class_attributes[] = {
 	&dev_attr_board_id
 };
 
+static void mlx4_addrconf_ifid_eui48(u8 *eui, struct net_device *dev)
+{
+	memcpy(eui, dev->dev_addr, 3);
+	memcpy(eui + 5, dev->dev_addr + 3, 3);
+	eui[3] = 0xFF;
+	eui[4] = 0xFE;
+	eui[0] ^= 2;
+}
+
+static void update_gids_task(struct work_struct *work)
+{
+	struct update_gid_work *gw = container_of(work, struct update_gid_work, work);
+	struct mlx4_cmd_mailbox *mailbox;
+	union ib_gid *gids;
+	int err;
+	struct mlx4_dev	*dev = gw->dev->dev;
+	struct ib_event event;
+
+	mailbox = mlx4_alloc_cmd_mailbox(dev);
+	if (IS_ERR(mailbox)) {
+		printk(KERN_WARNING "update gid table failed %ld\n", PTR_ERR(mailbox));
+		return;
+	}
+
+	gids = mailbox->buf;
+	memcpy(gids, gw->gids, sizeof gw->gids);
+
+	err = mlx4_cmd(dev, mailbox->dma, MLX4_SET_PORT_GID_TABLE << 8 | gw->port,
+		       1, MLX4_CMD_SET_PORT, MLX4_CMD_TIME_CLASS_B);
+	if (err)
+		printk(KERN_WARNING "set port command failed\n");
+	else {
+		memcpy(gw->dev->iboe.gid_table[gw->port - 1], gw->gids, sizeof gw->gids);
+		event.device = &gw->dev->ib_dev;
+		event.element.port_num = gw->port;
+		event.event    = IB_EVENT_LID_CHANGE;
+		ib_dispatch_event(&event);
+	}
+
+	mlx4_free_cmd_mailbox(dev, mailbox);
+	kfree(gw);
+}
+
+static int update_ipv6_gids(struct mlx4_ib_dev *dev, int port, int clear)
+{
+	struct net_device *ndev = dev->iboe.netdevs[port - 1];
+	struct update_gid_work *work;
+
+	work = kzalloc(sizeof *work, GFP_ATOMIC);
+	if (!work)
+		return -ENOMEM;
+
+	if (!clear) {
+		mlx4_addrconf_ifid_eui48(&work->gids[0].raw[8], ndev);
+		work->gids[0].global.subnet_prefix = cpu_to_be64(0xfe80000000000000LL);
+	}
+
+	INIT_WORK(&work->work, update_gids_task);
+	work->port = port;
+	work->dev = dev;
+	queue_work(wq, &work->work);
+
+	return 0;
+}
+
+static void handle_en_event(struct mlx4_ib_dev *dev, int port, unsigned long event)
+{
+	switch (event) {
+	case NETDEV_UP:
+		update_ipv6_gids(dev, port, 0);
+		break;
+
+	case NETDEV_DOWN:
+		update_ipv6_gids(dev, port, 1);
+		dev->iboe.netdevs[port - 1] = NULL;
+	}
+}
+
+static void netdev_added(struct mlx4_ib_dev *dev, int port)
+{
+	update_ipv6_gids(dev, port, 0);
+}
+
+static void netdev_removed(struct mlx4_ib_dev *dev, int port)
+{
+	update_ipv6_gids(dev, port, 1);
+}
+
+static int mlx4_ib_netdev_event(struct notifier_block *this, unsigned long event,
+				void *ptr)
+{
+	struct net_device *dev = ptr;
+	struct mlx4_ib_dev *ibdev;
+	struct net_device *oldnd;
+	struct mlx4_ib_iboe *iboe;
+	int port;
+
+	if (!net_eq(dev_net(dev), &init_net))
+		return NOTIFY_DONE;
+
+	ibdev = container_of(this, struct mlx4_ib_dev, iboe.nb);
+	iboe = &ibdev->iboe;
+
+	spin_lock(&iboe->lock);
+	mlx4_foreach_ib_transport_port(port, ibdev->dev) {
+		oldnd = iboe->netdevs[port - 1];
+		iboe->netdevs[port - 1] =
+			mlx4_get_protocol_dev(ibdev->dev, MLX4_PROTOCOL_EN, port);
+		if (oldnd != iboe->netdevs[port - 1]) {
+			if (iboe->netdevs[port - 1])
+				netdev_added(ibdev, port);
+			else
+				netdev_removed(ibdev, port);
+		}
+	}
+
+	if (dev == iboe->netdevs[0])
+		handle_en_event(ibdev, 1, event);
+	else if (dev == iboe->netdevs[1])
+		handle_en_event(ibdev, 2, event);
+
+	spin_unlock(&iboe->lock);
+
+	return NOTIFY_DONE;
+}
+
 static void *mlx4_ib_add(struct mlx4_dev *dev)
 {
 	struct mlx4_ib_dev *ibdev;
 	int num_ports = 0;
 	int i;
+	int err;
+	struct mlx4_ib_iboe *iboe;
 
 	printk_once(KERN_INFO "%s", mlx4_ib_version);
 
-	mlx4_foreach_port(i, dev, MLX4_PORT_TYPE_IB)
+	mlx4_foreach_ib_transport_port(i, dev)
 		num_ports++;
 
 	/* No point in registering a device with no ports... */
@@ -564,6 +904,8 @@ static void *mlx4_ib_add(struct mlx4_dev *dev)
 		return NULL;
 	}
 
+	iboe = &ibdev->iboe;
+
 	if (mlx4_pd_alloc(dev, &ibdev->priv_pdn))
 		goto err_dealloc;
 
@@ -612,6 +954,7 @@ static void *mlx4_ib_add(struct mlx4_dev *dev)
 
 	ibdev->ib_dev.query_device	= mlx4_ib_query_device;
 	ibdev->ib_dev.query_port	= mlx4_ib_query_port;
+	ibdev->ib_dev.get_link_layer	= mlx4_ib_port_link_layer;
 	ibdev->ib_dev.query_gid		= mlx4_ib_query_gid;
 	ibdev->ib_dev.query_pkey	= mlx4_ib_query_pkey;
 	ibdev->ib_dev.modify_device	= mlx4_ib_modify_device;
@@ -656,6 +999,8 @@ static void *mlx4_ib_add(struct mlx4_dev *dev)
 	ibdev->ib_dev.unmap_fmr		= mlx4_ib_unmap_fmr;
 	ibdev->ib_dev.dealloc_fmr	= mlx4_ib_fmr_dealloc;
 
+	spin_lock_init(&iboe->lock);
+
 	if (init_node_data(ibdev))
 		goto err_map;
 
@@ -668,16 +1013,28 @@ static void *mlx4_ib_add(struct mlx4_dev *dev)
 	if (mlx4_ib_mad_init(ibdev))
 		goto err_reg;
 
+	if (dev->caps.flags & MLX4_DEV_CAP_FLAG_IBOE && !iboe->nb.notifier_call) {
+		iboe->nb.notifier_call = mlx4_ib_netdev_event;
+		err = register_netdevice_notifier(&iboe->nb);
+		if (err)
+			goto err_reg;
+	}
+
 	for (i = 0; i < ARRAY_SIZE(mlx4_class_attributes); ++i) {
 		if (device_create_file(&ibdev->ib_dev.dev,
 				       mlx4_class_attributes[i]))
-			goto err_reg;
+			goto err_notif;
 	}
 
 	ibdev->ib_active = true;
 
 	return ibdev;
 
+err_notif:
+	if (unregister_netdevice_notifier(&ibdev->iboe.nb))
+		printk(KERN_WARNING "failure unregistering notifier\n");
+	flush_workqueue(wq);
+
 err_reg:
 	ib_unregister_device(&ibdev->ib_dev);
 
@@ -703,11 +1060,16 @@ static void mlx4_ib_remove(struct mlx4_dev *dev, void *ibdev_ptr)
 
 	mlx4_ib_mad_cleanup(ibdev);
 	ib_unregister_device(&ibdev->ib_dev);
+	if (ibdev->iboe.nb.notifier_call) {
+		if (unregister_netdevice_notifier(&ibdev->iboe.nb))
+			printk(KERN_WARNING "failure unregistering notifier\n");
+		ibdev->iboe.nb.notifier_call = NULL;
+	}
+	iounmap(ibdev->uar_map);
 
-	for (p = 1; p <= ibdev->num_ports; ++p)
+	mlx4_foreach_port(p, dev, MLX4_PORT_TYPE_IB)
 		mlx4_CLOSE_PORT(dev, p);
 
-	iounmap(ibdev->uar_map);
 	mlx4_uar_free(dev, &ibdev->priv_uar);
 	mlx4_pd_free(dev, ibdev->priv_pdn);
 	ib_dealloc_device(&ibdev->ib_dev);
@@ -747,19 +1109,33 @@ static void mlx4_ib_event(struct mlx4_dev *dev, void *ibdev_ptr,
 }
 
 static struct mlx4_interface mlx4_ib_interface = {
-	.add	= mlx4_ib_add,
-	.remove	= mlx4_ib_remove,
-	.event	= mlx4_ib_event
+	.add		= mlx4_ib_add,
+	.remove		= mlx4_ib_remove,
+	.event		= mlx4_ib_event,
+	.protocol	= MLX4_PROTOCOL_IB
 };
 
 static int __init mlx4_ib_init(void)
 {
-	return mlx4_register_interface(&mlx4_ib_interface);
+	int err;
+
+	wq = create_singlethread_workqueue("mlx4_ib");
+	if (!wq)
+		return -ENOMEM;
+
+	err = mlx4_register_interface(&mlx4_ib_interface);
+	if (err) {
+		destroy_workqueue(wq);
+		return err;
+	}
+
+	return 0;
 }
 
 static void __exit mlx4_ib_cleanup(void)
 {
 	mlx4_unregister_interface(&mlx4_ib_interface);
+	destroy_workqueue(wq);
 }
 
 module_init(mlx4_ib_init);