Adding upstream version 6.1.76.upstream/6.1.76

Signed-off-by: Daniel Baumann <daniel.baumann@progress-linux.org>

12 files changed, 8900 insertions, 0 deletions

diff --git a/drivers/infiniband/ulp/ipoib/Kconfig b/drivers/infiniband/ulp/ipoib/Kconfig
new file mode 100644
index 000000000..254e31a90
--- /dev/null
+++ b/drivers/infiniband/ulp/ipoib/Kconfig
@@ -0,0 +1,50 @@
+# SPDX-License-Identifier: GPL-2.0-only
+config INFINIBAND_IPOIB
+	tristate "IP-over-InfiniBand"
+	depends on NETDEVICES && INET && (IPV6 || IPV6=n)
+	help
+	  Support for the IP-over-InfiniBand protocol (IPoIB). This
+	  transports IP packets over InfiniBand so you can use your IB
+	  device as a fancy NIC.
+
+	  See Documentation/infiniband/ipoib.rst for more information
+
+config INFINIBAND_IPOIB_CM
+	bool "IP-over-InfiniBand Connected Mode support"
+	depends on INFINIBAND_IPOIB
+	default n
+	help
+	  This option enables support for IPoIB connected mode.  After
+	  enabling this option, you need to switch to connected mode
+	  through /sys/class/net/ibXXX/mode to actually create
+	  connections, and then increase the interface MTU with
+	  e.g. ifconfig ib0 mtu 65520.
+
+	  WARNING: Enabling connected mode will trigger some packet
+	  drops for multicast and UD mode traffic from this interface,
+	  unless you limit mtu for these destinations to 2044.
+
+config INFINIBAND_IPOIB_DEBUG
+	bool "IP-over-InfiniBand debugging" if EXPERT
+	depends on INFINIBAND_IPOIB
+	default y
+	help
+	  This option causes debugging code to be compiled into the
+	  IPoIB driver.  The output can be turned on via the
+	  debug_level and mcast_debug_level module parameters (which
+	  can also be set after the driver is loaded through sysfs).
+
+	  This option also creates a directory tree under ipoib/ in
+	  debugfs, which contains files that expose debugging
+	  information about IB multicast groups used by the IPoIB
+	  driver.
+
+config INFINIBAND_IPOIB_DEBUG_DATA
+	bool "IP-over-InfiniBand data path debugging"
+	depends on INFINIBAND_IPOIB_DEBUG
+	help
+	  This option compiles debugging code into the data path
+	  of the IPoIB driver.  The output can be turned on via the
+	  data_debug_level module parameter; however, even with output
+	  turned off, this debugging code will have some performance
+	  impact.
diff --git a/drivers/infiniband/ulp/ipoib/Makefile b/drivers/infiniband/ulp/ipoib/Makefile
new file mode 100644
index 000000000..6ece857ed
--- /dev/null
+++ b/drivers/infiniband/ulp/ipoib/Makefile
@@ -0,0 +1,13 @@
+# SPDX-License-Identifier: GPL-2.0
+obj-$(CONFIG_INFINIBAND_IPOIB)			+= ib_ipoib.o
+
+ib_ipoib-y					:= ipoib_main.o \
+						   ipoib_ib.o \
+						   ipoib_multicast.o \
+						   ipoib_verbs.o \
+						   ipoib_vlan.o \
+						   ipoib_ethtool.o \
+						   ipoib_netlink.o
+ib_ipoib-$(CONFIG_INFINIBAND_IPOIB_CM)		+= ipoib_cm.o
+ib_ipoib-$(CONFIG_INFINIBAND_IPOIB_DEBUG)	+= ipoib_fs.o
+
diff --git a/drivers/infiniband/ulp/ipoib/ipoib.h b/drivers/infiniband/ulp/ipoib/ipoib.h
new file mode 100644
index 000000000..35e9c8a33
--- /dev/null
+++ b/drivers/infiniband/ulp/ipoib/ipoib.h
@@ -0,0 +1,843 @@
+/*
+ * Copyright (c) 2004, 2005 Topspin Communications.  All rights reserved.
+ * Copyright (c) 2005 Sun Microsystems, Inc. All rights reserved.
+ * Copyright (c) 2004 Voltaire, Inc. All rights reserved.
+ *
+ * This software is available to you under a choice of one of two
+ * licenses.  You may choose to be licensed under the terms of the GNU
+ * General Public License (GPL) Version 2, available from the file
+ * COPYING in the main directory of this source tree, or the
+ * OpenIB.org BSD license below:
+ *
+ *     Redistribution and use in source and binary forms, with or
+ *     without modification, are permitted provided that the following
+ *     conditions are met:
+ *
+ *      - Redistributions of source code must retain the above
+ *        copyright notice, this list of conditions and the following
+ *        disclaimer.
+ *
+ *      - Redistributions in binary form must reproduce the above
+ *        copyright notice, this list of conditions and the following
+ *        disclaimer in the documentation and/or other materials
+ *        provided with the distribution.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
+ * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
+ * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS
+ * BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN
+ * ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
+ * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+ * SOFTWARE.
+ */
+
+#ifndef _IPOIB_H
+#define _IPOIB_H
+
+#include <linux/list.h>
+#include <linux/skbuff.h>
+#include <linux/netdevice.h>
+#include <linux/workqueue.h>
+#include <linux/kref.h>
+#include <linux/if_infiniband.h>
+#include <linux/mutex.h>
+
+#include <net/neighbour.h>
+#include <net/sch_generic.h>
+
+#include <linux/atomic.h>
+
+#include <rdma/ib_verbs.h>
+#include <rdma/ib_pack.h>
+#include <rdma/ib_sa.h>
+#include <linux/sched.h>
+/* constants */
+
+enum ipoib_flush_level {
+	IPOIB_FLUSH_LIGHT,
+	IPOIB_FLUSH_NORMAL,
+	IPOIB_FLUSH_HEAVY
+};
+
+enum {
+	IPOIB_ENCAP_LEN		  = 4,
+	IPOIB_PSEUDO_LEN	  = 20,
+	IPOIB_HARD_LEN		  = IPOIB_ENCAP_LEN + IPOIB_PSEUDO_LEN,
+
+	IPOIB_UD_HEAD_SIZE	  = IB_GRH_BYTES + IPOIB_ENCAP_LEN,
+	IPOIB_UD_RX_SG		  = 2, /* max buffer needed for 4K mtu */
+
+	IPOIB_CM_MTU		  = 0x10000 - 0x10, /* padding to align header to 16 */
+	IPOIB_CM_BUF_SIZE	  = IPOIB_CM_MTU  + IPOIB_ENCAP_LEN,
+	IPOIB_CM_HEAD_SIZE	  = IPOIB_CM_BUF_SIZE % PAGE_SIZE,
+	IPOIB_CM_RX_SG		  = ALIGN(IPOIB_CM_BUF_SIZE, PAGE_SIZE) / PAGE_SIZE,
+	IPOIB_RX_RING_SIZE	  = 256,
+	IPOIB_TX_RING_SIZE	  = 128,
+	IPOIB_MAX_QUEUE_SIZE	  = 8192,
+	IPOIB_MIN_QUEUE_SIZE	  = 2,
+	IPOIB_CM_MAX_CONN_QP	  = 4096,
+
+	IPOIB_NUM_WC		  = 4,
+
+	IPOIB_MAX_PATH_REC_QUEUE  = 3,
+	IPOIB_MAX_MCAST_QUEUE	  = 64,
+
+	IPOIB_FLAG_OPER_UP	  = 0,
+	IPOIB_FLAG_INITIALIZED	  = 1,
+	IPOIB_FLAG_ADMIN_UP	  = 2,
+	IPOIB_PKEY_ASSIGNED	  = 3,
+	IPOIB_FLAG_SUBINTERFACE	  = 5,
+	IPOIB_STOP_REAPER	  = 7,
+	IPOIB_FLAG_ADMIN_CM	  = 9,
+	IPOIB_FLAG_UMCAST	  = 10,
+	IPOIB_NEIGH_TBL_FLUSH	  = 12,
+	IPOIB_FLAG_DEV_ADDR_SET	  = 13,
+	IPOIB_FLAG_DEV_ADDR_CTRL  = 14,
+
+	IPOIB_MAX_BACKOFF_SECONDS = 16,
+
+	IPOIB_MCAST_FLAG_FOUND	  = 0,	/* used in set_multicast_list */
+	IPOIB_MCAST_FLAG_SENDONLY = 1,
+	/*
+	 * For IPOIB_MCAST_FLAG_BUSY
+	 * When set, in flight join and mcast->mc is unreliable
+	 * When clear and mcast->mc IS_ERR_OR_NULL, need to restart or
+	 *   haven't started yet
+	 * When clear and mcast->mc is valid pointer, join was successful
+	 */
+	IPOIB_MCAST_FLAG_BUSY	  = 2,
+	IPOIB_MCAST_FLAG_ATTACHED = 3,
+
+	MAX_SEND_CQE		  = 64,
+	IPOIB_CM_COPYBREAK	  = 256,
+
+	IPOIB_NON_CHILD		  = 0,
+	IPOIB_LEGACY_CHILD	  = 1,
+	IPOIB_RTNL_CHILD	  = 2,
+};
+
+#define	IPOIB_OP_RECV   (1ul << 31)
+#ifdef CONFIG_INFINIBAND_IPOIB_CM
+#define	IPOIB_OP_CM     (1ul << 30)
+#else
+#define	IPOIB_OP_CM     (0)
+#endif
+
+#define IPOIB_QPN_MASK ((__force u32) cpu_to_be32(0xFFFFFF))
+
+/* structs */
+
+struct ipoib_header {
+	__be16	proto;
+	u16	reserved;
+};
+
+struct ipoib_pseudo_header {
+	u8	hwaddr[INFINIBAND_ALEN];
+};
+
+static inline void skb_add_pseudo_hdr(struct sk_buff *skb)
+{
+	char *data = skb_push(skb, IPOIB_PSEUDO_LEN);
+
+	/*
+	 * only the ipoib header is present now, make room for a dummy
+	 * pseudo header and set skb field accordingly
+	 */
+	memset(data, 0, IPOIB_PSEUDO_LEN);
+	skb_reset_mac_header(skb);
+	skb_pull(skb, IPOIB_HARD_LEN);
+}
+
+static inline struct ipoib_dev_priv *ipoib_priv(const struct net_device *dev)
+{
+	struct rdma_netdev *rn = netdev_priv(dev);
+
+	return rn->clnt_priv;
+}
+
+/* Used for all multicast joins (broadcast, IPv4 mcast and IPv6 mcast) */
+struct ipoib_mcast {
+	struct ib_sa_mcmember_rec mcmember;
+	struct ib_sa_multicast	 *mc;
+	struct ipoib_ah		 *ah;
+
+	struct rb_node    rb_node;
+	struct list_head  list;
+
+	unsigned long created;
+	unsigned long backoff;
+	unsigned long delay_until;
+
+	unsigned long flags;
+	unsigned char logcount;
+
+	struct list_head  neigh_list;
+
+	struct sk_buff_head pkt_queue;
+
+	struct net_device *dev;
+	struct completion done;
+};
+
+struct ipoib_rx_buf {
+	struct sk_buff *skb;
+	u64		mapping[IPOIB_UD_RX_SG];
+};
+
+struct ipoib_tx_buf {
+	struct sk_buff *skb;
+	u64		mapping[MAX_SKB_FRAGS + 1];
+};
+
+struct ib_cm_id;
+
+struct ipoib_cm_data {
+	__be32 qpn; /* High byte MUST be ignored on receive */
+	__be32 mtu;
+};
+
+/*
+ * Quoting 10.3.1 Queue Pair and EE Context States:
+ *
+ * Note, for QPs that are associated with an SRQ, the Consumer should take the
+ * QP through the Error State before invoking a Destroy QP or a Modify QP to the
+ * Reset State.  The Consumer may invoke the Destroy QP without first performing
+ * a Modify QP to the Error State and waiting for the Affiliated Asynchronous
+ * Last WQE Reached Event. However, if the Consumer does not wait for the
+ * Affiliated Asynchronous Last WQE Reached Event, then WQE and Data Segment
+ * leakage may occur. Therefore, it is good programming practice to tear down a
+ * QP that is associated with an SRQ by using the following process:
+ *
+ * - Put the QP in the Error State
+ * - Wait for the Affiliated Asynchronous Last WQE Reached Event;
+ * - either:
+ *       drain the CQ by invoking the Poll CQ verb and either wait for CQ
+ *       to be empty or the number of Poll CQ operations has exceeded
+ *       CQ capacity size;
+ * - or
+ *       post another WR that completes on the same CQ and wait for this
+ *       WR to return as a WC;
+ * - and then invoke a Destroy QP or Reset QP.
+ *
+ * We use the second option and wait for a completion on the
+ * same CQ before destroying QPs attached to our SRQ.
+ */
+
+enum ipoib_cm_state {
+	IPOIB_CM_RX_LIVE,
+	IPOIB_CM_RX_ERROR, /* Ignored by stale task */
+	IPOIB_CM_RX_FLUSH  /* Last WQE Reached event observed */
+};
+
+struct ipoib_cm_rx {
+	struct ib_cm_id	       *id;
+	struct ib_qp	       *qp;
+	struct ipoib_cm_rx_buf *rx_ring;
+	struct list_head	list;
+	struct net_device      *dev;
+	unsigned long		jiffies;
+	enum ipoib_cm_state	state;
+	int			recv_count;
+};
+
+struct ipoib_cm_tx {
+	struct ib_cm_id	    *id;
+	struct ib_qp	    *qp;
+	struct list_head     list;
+	struct net_device   *dev;
+	struct ipoib_neigh  *neigh;
+	struct ipoib_tx_buf *tx_ring;
+	unsigned int	     tx_head;
+	unsigned int	     tx_tail;
+	unsigned long	     flags;
+	u32		     mtu;
+	unsigned int         max_send_sge;
+};
+
+struct ipoib_cm_rx_buf {
+	struct sk_buff *skb;
+	u64 mapping[IPOIB_CM_RX_SG];
+};
+
+struct ipoib_cm_dev_priv {
+	struct ib_srq	       *srq;
+	struct ipoib_cm_rx_buf *srq_ring;
+	struct ib_cm_id	       *id;
+	struct list_head	passive_ids;   /* state: LIVE */
+	struct list_head	rx_error_list; /* state: ERROR */
+	struct list_head	rx_flush_list; /* state: FLUSH, drain not started */
+	struct list_head	rx_drain_list; /* state: FLUSH, drain started */
+	struct list_head	rx_reap_list;  /* state: FLUSH, drain done */
+	struct work_struct      start_task;
+	struct work_struct      reap_task;
+	struct work_struct      skb_task;
+	struct work_struct      rx_reap_task;
+	struct delayed_work     stale_task;
+	struct sk_buff_head     skb_queue;
+	struct list_head	start_list;
+	struct list_head	reap_list;
+	struct ib_wc		ibwc[IPOIB_NUM_WC];
+	struct ib_sge		rx_sge[IPOIB_CM_RX_SG];
+	struct ib_recv_wr       rx_wr;
+	int			nonsrq_conn_qp;
+	int			max_cm_mtu;
+	int			num_frags;
+};
+
+struct ipoib_ethtool_st {
+	u16     coalesce_usecs;
+	u16     max_coalesced_frames;
+};
+
+struct ipoib_neigh_table;
+
+struct ipoib_neigh_hash {
+	struct ipoib_neigh_table       *ntbl;
+	struct ipoib_neigh __rcu      **buckets;
+	struct rcu_head			rcu;
+	u32				mask;
+	u32				size;
+};
+
+struct ipoib_neigh_table {
+	struct ipoib_neigh_hash __rcu  *htbl;
+	atomic_t			entries;
+	struct completion		flushed;
+	struct completion		deleted;
+};
+
+struct ipoib_qp_state_validate {
+	struct work_struct work;
+	struct ipoib_dev_priv   *priv;
+};
+
+/*
+ * Device private locking: network stack tx_lock protects members used
+ * in TX fast path, lock protects everything else.  lock nests inside
+ * of tx_lock (ie tx_lock must be acquired first if needed).
+ */
+struct ipoib_dev_priv {
+	spinlock_t lock;
+
+	struct net_device *dev;
+	void (*next_priv_destructor)(struct net_device *dev);
+
+	struct napi_struct send_napi;
+	struct napi_struct recv_napi;
+
+	unsigned long flags;
+
+	/*
+	 * This protects access to the child_intfs list.
+	 * To READ from child_intfs the RTNL or vlan_rwsem read side must be
+	 * held.  To WRITE RTNL and the vlan_rwsem write side must be held (in
+	 * that order) This lock exists because we have a few contexts where
+	 * we need the child_intfs, but do not want to grab the RTNL.
+	 */
+	struct rw_semaphore vlan_rwsem;
+	struct mutex mcast_mutex;
+
+	struct rb_root  path_tree;
+	struct list_head path_list;
+
+	struct ipoib_neigh_table ntbl;
+
+	struct ipoib_mcast *broadcast;
+	struct list_head multicast_list;
+	struct rb_root multicast_tree;
+
+	struct workqueue_struct *wq;
+	struct delayed_work mcast_task;
+	struct work_struct carrier_on_task;
+	struct work_struct flush_light;
+	struct work_struct flush_normal;
+	struct work_struct flush_heavy;
+	struct work_struct restart_task;
+	struct delayed_work ah_reap_task;
+	struct delayed_work neigh_reap_task;
+	struct ib_device *ca;
+	u8		  port;
+	u16		  pkey;
+	u16		  pkey_index;
+	struct ib_pd	 *pd;
+	struct ib_cq	 *recv_cq;
+	struct ib_cq	 *send_cq;
+	struct ib_qp	 *qp;
+	u32		  qkey;
+
+	union ib_gid local_gid;
+	u32	     local_lid;
+
+	unsigned int admin_mtu;
+	unsigned int mcast_mtu;
+	unsigned int max_ib_mtu;
+
+	struct ipoib_rx_buf *rx_ring;
+
+	struct ipoib_tx_buf *tx_ring;
+	/* cyclic ring variables for managing tx_ring, for UD only */
+	unsigned int	     tx_head;
+	unsigned int	     tx_tail;
+	/* cyclic ring variables for counting overall outstanding send WRs */
+	unsigned int	     global_tx_head;
+	unsigned int	     global_tx_tail;
+	struct ib_sge	     tx_sge[MAX_SKB_FRAGS + 1];
+	struct ib_ud_wr      tx_wr;
+	struct ib_wc	     send_wc[MAX_SEND_CQE];
+
+	struct ib_recv_wr    rx_wr;
+	struct ib_sge	     rx_sge[IPOIB_UD_RX_SG];
+
+	struct ib_wc ibwc[IPOIB_NUM_WC];
+
+	struct list_head dead_ahs;
+
+	struct ib_event_handler event_handler;
+
+	struct net_device *parent;
+	struct list_head child_intfs;
+	struct list_head list;
+	int    child_type;
+
+#ifdef CONFIG_INFINIBAND_IPOIB_CM
+	struct ipoib_cm_dev_priv cm;
+#endif
+
+#ifdef CONFIG_INFINIBAND_IPOIB_DEBUG
+	struct list_head fs_list;
+	struct dentry *mcg_dentry;
+	struct dentry *path_dentry;
+#endif
+	u64	hca_caps;
+	u64	kernel_caps;
+	struct ipoib_ethtool_st ethtool;
+	unsigned int max_send_sge;
+	const struct net_device_ops	*rn_ops;
+};
+
+struct ipoib_ah {
+	struct net_device *dev;
+	struct ib_ah	  *ah;
+	struct list_head   list;
+	struct kref	   ref;
+	unsigned int	   last_send;
+	int  		   valid;
+};
+
+struct ipoib_path {
+	struct net_device    *dev;
+	struct sa_path_rec pathrec;
+	struct ipoib_ah      *ah;
+	struct sk_buff_head   queue;
+
+	struct list_head      neigh_list;
+
+	int		      query_id;
+	struct ib_sa_query   *query;
+	struct completion     done;
+
+	struct rb_node	      rb_node;
+	struct list_head      list;
+};
+
+struct ipoib_neigh {
+	struct ipoib_ah    *ah;
+#ifdef CONFIG_INFINIBAND_IPOIB_CM
+	struct ipoib_cm_tx *cm;
+#endif
+	u8     daddr[INFINIBAND_ALEN];
+	struct sk_buff_head queue;
+
+	struct net_device *dev;
+
+	struct list_head    list;
+	struct ipoib_neigh __rcu *hnext;
+	struct rcu_head     rcu;
+	refcount_t	    refcnt;
+	unsigned long       alive;
+};
+
+#define IPOIB_UD_MTU(ib_mtu)		(ib_mtu - IPOIB_ENCAP_LEN)
+#define IPOIB_UD_BUF_SIZE(ib_mtu)	(ib_mtu + IB_GRH_BYTES)
+
+void ipoib_neigh_dtor(struct ipoib_neigh *neigh);
+static inline void ipoib_neigh_put(struct ipoib_neigh *neigh)
+{
+	if (refcount_dec_and_test(&neigh->refcnt))
+		ipoib_neigh_dtor(neigh);
+}
+struct ipoib_neigh *ipoib_neigh_get(struct net_device *dev, u8 *daddr);
+struct ipoib_neigh *ipoib_neigh_alloc(u8 *daddr,
+				      struct net_device *dev);
+void ipoib_neigh_free(struct ipoib_neigh *neigh);
+void ipoib_del_neighs_by_gid(struct net_device *dev, u8 *gid);
+
+extern struct workqueue_struct *ipoib_workqueue;
+
+/* functions */
+
+int ipoib_rx_poll(struct napi_struct *napi, int budget);
+int ipoib_tx_poll(struct napi_struct *napi, int budget);
+void ipoib_ib_rx_completion(struct ib_cq *cq, void *ctx_ptr);
+void ipoib_ib_tx_completion(struct ib_cq *cq, void *ctx_ptr);
+
+struct ipoib_ah *ipoib_create_ah(struct net_device *dev,
+				 struct ib_pd *pd, struct rdma_ah_attr *attr);
+void ipoib_free_ah(struct kref *kref);
+static inline void ipoib_put_ah(struct ipoib_ah *ah)
+{
+	kref_put(&ah->ref, ipoib_free_ah);
+}
+int ipoib_open(struct net_device *dev);
+void ipoib_intf_free(struct net_device *dev);
+int ipoib_add_pkey_attr(struct net_device *dev);
+int ipoib_add_umcast_attr(struct net_device *dev);
+
+int ipoib_send(struct net_device *dev, struct sk_buff *skb,
+	       struct ib_ah *address, u32 dqpn);
+void ipoib_reap_ah(struct work_struct *work);
+
+struct ipoib_path *__path_find(struct net_device *dev, void *gid);
+void ipoib_mark_paths_invalid(struct net_device *dev);
+void ipoib_flush_paths(struct net_device *dev);
+struct net_device *ipoib_intf_alloc(struct ib_device *hca, u32 port,
+				    const char *format);
+int ipoib_intf_init(struct ib_device *hca, u32 port, const char *format,
+		    struct net_device *dev);
+void ipoib_ib_tx_timer_func(struct timer_list *t);
+void ipoib_ib_dev_flush_light(struct work_struct *work);
+void ipoib_ib_dev_flush_normal(struct work_struct *work);
+void ipoib_ib_dev_flush_heavy(struct work_struct *work);
+void ipoib_pkey_event(struct work_struct *work);
+void ipoib_ib_dev_cleanup(struct net_device *dev);
+
+int ipoib_ib_dev_open_default(struct net_device *dev);
+int ipoib_ib_dev_open(struct net_device *dev);
+void ipoib_ib_dev_stop(struct net_device *dev);
+void ipoib_ib_dev_up(struct net_device *dev);
+void ipoib_ib_dev_down(struct net_device *dev);
+int ipoib_ib_dev_stop_default(struct net_device *dev);
+void ipoib_pkey_dev_check_presence(struct net_device *dev);
+
+void ipoib_mcast_join_task(struct work_struct *work);
+void ipoib_mcast_carrier_on_task(struct work_struct *work);
+void ipoib_mcast_send(struct net_device *dev, u8 *daddr, struct sk_buff *skb);
+
+void ipoib_mcast_restart_task(struct work_struct *work);
+void ipoib_mcast_start_thread(struct net_device *dev);
+void ipoib_mcast_stop_thread(struct net_device *dev);
+
+void ipoib_mcast_dev_down(struct net_device *dev);
+void ipoib_mcast_dev_flush(struct net_device *dev);
+
+int ipoib_dma_map_tx(struct ib_device *ca, struct ipoib_tx_buf *tx_req);
+void ipoib_dma_unmap_tx(struct ipoib_dev_priv *priv,
+			struct ipoib_tx_buf *tx_req);
+
+struct rtnl_link_ops *ipoib_get_link_ops(void);
+
+static inline void ipoib_build_sge(struct ipoib_dev_priv *priv,
+				   struct ipoib_tx_buf *tx_req)
+{
+	int i, off;
+	struct sk_buff *skb = tx_req->skb;
+	skb_frag_t *frags = skb_shinfo(skb)->frags;
+	int nr_frags = skb_shinfo(skb)->nr_frags;
+	u64 *mapping = tx_req->mapping;
+
+	if (skb_headlen(skb)) {
+		priv->tx_sge[0].addr         = mapping[0];
+		priv->tx_sge[0].length       = skb_headlen(skb);
+		off = 1;
+	} else
+		off = 0;
+
+	for (i = 0; i < nr_frags; ++i) {
+		priv->tx_sge[i + off].addr = mapping[i + off];
+		priv->tx_sge[i + off].length = skb_frag_size(&frags[i]);
+	}
+	priv->tx_wr.wr.num_sge	     = nr_frags + off;
+}
+
+#ifdef CONFIG_INFINIBAND_IPOIB_DEBUG
+struct ipoib_mcast_iter *ipoib_mcast_iter_init(struct net_device *dev);
+int ipoib_mcast_iter_next(struct ipoib_mcast_iter *iter);
+void ipoib_mcast_iter_read(struct ipoib_mcast_iter *iter,
+				  union ib_gid *gid,
+				  unsigned long *created,
+				  unsigned int *queuelen,
+				  unsigned int *complete,
+				  unsigned int *send_only);
+
+struct ipoib_path_iter *ipoib_path_iter_init(struct net_device *dev);
+int ipoib_path_iter_next(struct ipoib_path_iter *iter);
+void ipoib_path_iter_read(struct ipoib_path_iter *iter,
+			  struct ipoib_path *path);
+#endif
+
+int ipoib_mcast_attach(struct net_device *dev, struct ib_device *hca,
+		       union ib_gid *mgid, u16 mlid, int set_qkey, u32 qkey);
+int ipoib_mcast_detach(struct net_device *dev, struct ib_device *hca,
+		       union ib_gid *mgid, u16 mlid);
+void ipoib_mcast_remove_list(struct list_head *remove_list);
+void ipoib_check_and_add_mcast_sendonly(struct ipoib_dev_priv *priv, u8 *mgid,
+				struct list_head *remove_list);
+
+int ipoib_init_qp(struct net_device *dev);
+int ipoib_transport_dev_init(struct net_device *dev, struct ib_device *ca);
+void ipoib_transport_dev_cleanup(struct net_device *dev);
+
+void ipoib_event(struct ib_event_handler *handler,
+		 struct ib_event *record);
+
+int ipoib_vlan_add(struct net_device *pdev, unsigned short pkey);
+int ipoib_vlan_delete(struct net_device *pdev, unsigned short pkey);
+
+int __ipoib_vlan_add(struct ipoib_dev_priv *ppriv, struct ipoib_dev_priv *priv,
+		     u16 pkey, int child_type);
+
+int  __init ipoib_netlink_init(void);
+void __exit ipoib_netlink_fini(void);
+
+void ipoib_set_umcast(struct net_device *ndev, int umcast_val);
+int  ipoib_set_mode(struct net_device *dev, const char *buf);
+
+void ipoib_setup_common(struct net_device *dev);
+
+void ipoib_pkey_open(struct ipoib_dev_priv *priv);
+void ipoib_drain_cq(struct net_device *dev);
+
+void ipoib_set_ethtool_ops(struct net_device *dev);
+
+#define IPOIB_FLAGS_RC		0x80
+#define IPOIB_FLAGS_UC		0x40
+
+/* We don't support UC connections at the moment */
+#define IPOIB_CM_SUPPORTED(ha)   (ha[0] & (IPOIB_FLAGS_RC))
+
+#ifdef CONFIG_INFINIBAND_IPOIB_CM
+
+extern int ipoib_max_conn_qp;
+
+static inline int ipoib_cm_admin_enabled(struct net_device *dev)
+{
+	struct ipoib_dev_priv *priv = ipoib_priv(dev);
+	return IPOIB_CM_SUPPORTED(dev->dev_addr) &&
+		test_bit(IPOIB_FLAG_ADMIN_CM, &priv->flags);
+}
+
+static inline int ipoib_cm_enabled(struct net_device *dev, u8 *hwaddr)
+{
+	struct ipoib_dev_priv *priv = ipoib_priv(dev);
+	return IPOIB_CM_SUPPORTED(hwaddr) &&
+		test_bit(IPOIB_FLAG_ADMIN_CM, &priv->flags);
+}
+
+static inline int ipoib_cm_up(struct ipoib_neigh *neigh)
+
+{
+	return test_bit(IPOIB_FLAG_OPER_UP, &neigh->cm->flags);
+}
+
+static inline struct ipoib_cm_tx *ipoib_cm_get(struct ipoib_neigh *neigh)
+{
+	return neigh->cm;
+}
+
+static inline void ipoib_cm_set(struct ipoib_neigh *neigh, struct ipoib_cm_tx *tx)
+{
+	neigh->cm = tx;
+}
+
+static inline int ipoib_cm_has_srq(struct net_device *dev)
+{
+	struct ipoib_dev_priv *priv = ipoib_priv(dev);
+	return !!priv->cm.srq;
+}
+
+static inline unsigned int ipoib_cm_max_mtu(struct net_device *dev)
+{
+	struct ipoib_dev_priv *priv = ipoib_priv(dev);
+	return priv->cm.max_cm_mtu;
+}
+
+void ipoib_cm_send(struct net_device *dev, struct sk_buff *skb, struct ipoib_cm_tx *tx);
+int ipoib_cm_dev_open(struct net_device *dev);
+void ipoib_cm_dev_stop(struct net_device *dev);
+int ipoib_cm_dev_init(struct net_device *dev);
+int ipoib_cm_add_mode_attr(struct net_device *dev);
+void ipoib_cm_dev_cleanup(struct net_device *dev);
+struct ipoib_cm_tx *ipoib_cm_create_tx(struct net_device *dev, struct ipoib_path *path,
+				    struct ipoib_neigh *neigh);
+void ipoib_cm_destroy_tx(struct ipoib_cm_tx *tx);
+void ipoib_cm_skb_too_long(struct net_device *dev, struct sk_buff *skb,
+			   unsigned int mtu);
+void ipoib_cm_handle_rx_wc(struct net_device *dev, struct ib_wc *wc);
+void ipoib_cm_handle_tx_wc(struct net_device *dev, struct ib_wc *wc);
+#else
+
+#define ipoib_max_conn_qp 0
+
+static inline int ipoib_cm_admin_enabled(struct net_device *dev)
+{
+	return 0;
+}
+static inline int ipoib_cm_enabled(struct net_device *dev, u8 *hwaddr)
+
+{
+	return 0;
+}
+
+static inline int ipoib_cm_up(struct ipoib_neigh *neigh)
+
+{
+	return 0;
+}
+
+static inline struct ipoib_cm_tx *ipoib_cm_get(struct ipoib_neigh *neigh)
+{
+	return NULL;
+}
+
+static inline void ipoib_cm_set(struct ipoib_neigh *neigh, struct ipoib_cm_tx *tx)
+{
+}
+
+static inline int ipoib_cm_has_srq(struct net_device *dev)
+{
+	return 0;
+}
+
+static inline unsigned int ipoib_cm_max_mtu(struct net_device *dev)
+{
+	return 0;
+}
+
+static inline
+void ipoib_cm_send(struct net_device *dev, struct sk_buff *skb, struct ipoib_cm_tx *tx)
+{
+	return;
+}
+
+static inline
+int ipoib_cm_dev_open(struct net_device *dev)
+{
+	return 0;
+}
+
+static inline
+void ipoib_cm_dev_stop(struct net_device *dev)
+{
+	return;
+}
+
+static inline
+int ipoib_cm_dev_init(struct net_device *dev)
+{
+	return -EOPNOTSUPP;
+}
+
+static inline
+void ipoib_cm_dev_cleanup(struct net_device *dev)
+{
+	return;
+}
+
+static inline
+struct ipoib_cm_tx *ipoib_cm_create_tx(struct net_device *dev, struct ipoib_path *path,
+				    struct ipoib_neigh *neigh)
+{
+	return NULL;
+}
+
+static inline
+void ipoib_cm_destroy_tx(struct ipoib_cm_tx *tx)
+{
+	return;
+}
+
+static inline
+int ipoib_cm_add_mode_attr(struct net_device *dev)
+{
+	return 0;
+}
+
+static inline void ipoib_cm_skb_too_long(struct net_device *dev, struct sk_buff *skb,
+					 unsigned int mtu)
+{
+	dev_kfree_skb_any(skb);
+}
+
+static inline void ipoib_cm_handle_rx_wc(struct net_device *dev, struct ib_wc *wc)
+{
+}
+
+static inline void ipoib_cm_handle_tx_wc(struct net_device *dev, struct ib_wc *wc)
+{
+}
+#endif
+
+#ifdef CONFIG_INFINIBAND_IPOIB_DEBUG
+void ipoib_create_debug_files(struct net_device *dev);
+void ipoib_delete_debug_files(struct net_device *dev);
+void ipoib_register_debugfs(void);
+void ipoib_unregister_debugfs(void);
+#else
+static inline void ipoib_create_debug_files(struct net_device *dev) { }
+static inline void ipoib_delete_debug_files(struct net_device *dev) { }
+static inline void ipoib_register_debugfs(void) { }
+static inline void ipoib_unregister_debugfs(void) { }
+#endif
+
+#define ipoib_printk(level, priv, format, arg...)	\
+	printk(level "%s: " format, ((struct ipoib_dev_priv *) priv)->dev->name , ## arg)
+#define ipoib_warn(priv, format, arg...)		\
+do {							\
+	static DEFINE_RATELIMIT_STATE(_rs,		\
+		10 * HZ /*10 seconds */,		\
+		100);		\
+	if (__ratelimit(&_rs))				\
+		ipoib_printk(KERN_WARNING, priv, format , ## arg);\
+} while (0)
+
+extern int ipoib_sendq_size;
+extern int ipoib_recvq_size;
+
+extern struct ib_sa_client ipoib_sa_client;
+
+#ifdef CONFIG_INFINIBAND_IPOIB_DEBUG
+extern int ipoib_debug_level;
+
+#define ipoib_dbg(priv, format, arg...)			\
+	do {						\
+		if (ipoib_debug_level > 0)			\
+			ipoib_printk(KERN_DEBUG, priv, format , ## arg); \
+	} while (0)
+#define ipoib_dbg_mcast(priv, format, arg...)		\
+	do {						\
+		if (mcast_debug_level > 0)		\
+			ipoib_printk(KERN_DEBUG, priv, format , ## arg); \
+	} while (0)
+#else /* CONFIG_INFINIBAND_IPOIB_DEBUG */
+#define ipoib_dbg(priv, format, arg...)			\
+	do { (void) (priv); } while (0)
+#define ipoib_dbg_mcast(priv, format, arg...)		\
+	do { (void) (priv); } while (0)
+#endif /* CONFIG_INFINIBAND_IPOIB_DEBUG */
+
+#ifdef CONFIG_INFINIBAND_IPOIB_DEBUG_DATA
+#define ipoib_dbg_data(priv, format, arg...)		\
+	do {						\
+		if (data_debug_level > 0)		\
+			ipoib_printk(KERN_DEBUG, priv, format , ## arg); \
+	} while (0)
+#else /* CONFIG_INFINIBAND_IPOIB_DEBUG_DATA */
+#define ipoib_dbg_data(priv, format, arg...)		\
+	do { (void) (priv); } while (0)
+#endif /* CONFIG_INFINIBAND_IPOIB_DEBUG_DATA */
+
+#define IPOIB_QPN(ha) (be32_to_cpup((__be32 *) ha) & 0xffffff)
+
+#endif /* _IPOIB_H */
diff --git a/drivers/infiniband/ulp/ipoib/ipoib_cm.c b/drivers/infiniband/ulp/ipoib/ipoib_cm.c
new file mode 100644
index 000000000..b610d3629
--- /dev/null
+++ b/drivers/infiniband/ulp/ipoib/ipoib_cm.c
@@ -0,0 +1,1661 @@
+/*
+ * Copyright (c) 2006 Mellanox Technologies. All rights reserved
+ *
+ * This software is available to you under a choice of one of two
+ * licenses.  You may choose to be licensed under the terms of the GNU
+ * General Public License (GPL) Version 2, available from the file
+ * COPYING in the main directory of this source tree, or the
+ * OpenIB.org BSD license below:
+ *
+ *     Redistribution and use in source and binary forms, with or
+ *     without modification, are permitted provided that the following
+ *     conditions are met:
+ *
+ *      - Redistributions of source code must retain the above
+ *        copyright notice, this list of conditions and the following
+ *        disclaimer.
+ *
+ *      - Redistributions in binary form must reproduce the above
+ *        copyright notice, this list of conditions and the following
+ *        disclaimer in the documentation and/or other materials
+ *        provided with the distribution.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
+ * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
+ * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS
+ * BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN
+ * ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
+ * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+ * SOFTWARE.
+ */
+
+#include <rdma/ib_cm.h>
+#include <net/dst.h>
+#include <net/icmp.h>
+#include <linux/icmpv6.h>
+#include <linux/delay.h>
+#include <linux/slab.h>
+#include <linux/vmalloc.h>
+#include <linux/moduleparam.h>
+#include <linux/sched/signal.h>
+#include <linux/sched/mm.h>
+
+#include "ipoib.h"
+
+int ipoib_max_conn_qp = 128;
+
+module_param_named(max_nonsrq_conn_qp, ipoib_max_conn_qp, int, 0444);
+MODULE_PARM_DESC(max_nonsrq_conn_qp,
+		 "Max number of connected-mode QPs per interface "
+		 "(applied only if shared receive queue is not available)");
+
+#ifdef CONFIG_INFINIBAND_IPOIB_DEBUG_DATA
+static int data_debug_level;
+
+module_param_named(cm_data_debug_level, data_debug_level, int, 0644);
+MODULE_PARM_DESC(cm_data_debug_level,
+		 "Enable data path debug tracing for connected mode if > 0");
+#endif
+
+#define IPOIB_CM_IETF_ID 0x1000000000000000ULL
+
+#define IPOIB_CM_RX_UPDATE_TIME (256 * HZ)
+#define IPOIB_CM_RX_TIMEOUT     (2 * 256 * HZ)
+#define IPOIB_CM_RX_DELAY       (3 * 256 * HZ)
+#define IPOIB_CM_RX_UPDATE_MASK (0x3)
+
+#define IPOIB_CM_RX_RESERVE     (ALIGN(IPOIB_HARD_LEN, 16) - IPOIB_ENCAP_LEN)
+
+static struct ib_qp_attr ipoib_cm_err_attr = {
+	.qp_state = IB_QPS_ERR
+};
+
+#define IPOIB_CM_RX_DRAIN_WRID 0xffffffff
+
+static struct ib_send_wr ipoib_cm_rx_drain_wr = {
+	.opcode = IB_WR_SEND,
+};
+
+static int ipoib_cm_tx_handler(struct ib_cm_id *cm_id,
+			       const struct ib_cm_event *event);
+
+static void ipoib_cm_dma_unmap_rx(struct ipoib_dev_priv *priv, int frags,
+				  u64 mapping[IPOIB_CM_RX_SG])
+{
+	int i;
+
+	ib_dma_unmap_single(priv->ca, mapping[0], IPOIB_CM_HEAD_SIZE, DMA_FROM_DEVICE);
+
+	for (i = 0; i < frags; ++i)
+		ib_dma_unmap_page(priv->ca, mapping[i + 1], PAGE_SIZE, DMA_FROM_DEVICE);
+}
+
+static int ipoib_cm_post_receive_srq(struct net_device *dev, int id)
+{
+	struct ipoib_dev_priv *priv = ipoib_priv(dev);
+	int i, ret;
+
+	priv->cm.rx_wr.wr_id = id | IPOIB_OP_CM | IPOIB_OP_RECV;
+
+	for (i = 0; i < priv->cm.num_frags; ++i)
+		priv->cm.rx_sge[i].addr = priv->cm.srq_ring[id].mapping[i];
+
+	ret = ib_post_srq_recv(priv->cm.srq, &priv->cm.rx_wr, NULL);
+	if (unlikely(ret)) {
+		ipoib_warn(priv, "post srq failed for buf %d (%d)\n", id, ret);
+		ipoib_cm_dma_unmap_rx(priv, priv->cm.num_frags - 1,
+				      priv->cm.srq_ring[id].mapping);
+		dev_kfree_skb_any(priv->cm.srq_ring[id].skb);
+		priv->cm.srq_ring[id].skb = NULL;
+	}
+
+	return ret;
+}
+
+static int ipoib_cm_post_receive_nonsrq(struct net_device *dev,
+					struct ipoib_cm_rx *rx,
+					struct ib_recv_wr *wr,
+					struct ib_sge *sge, int id)
+{
+	struct ipoib_dev_priv *priv = ipoib_priv(dev);
+	int i, ret;
+
+	wr->wr_id = id | IPOIB_OP_CM | IPOIB_OP_RECV;
+
+	for (i = 0; i < IPOIB_CM_RX_SG; ++i)
+		sge[i].addr = rx->rx_ring[id].mapping[i];
+
+	ret = ib_post_recv(rx->qp, wr, NULL);
+	if (unlikely(ret)) {
+		ipoib_warn(priv, "post recv failed for buf %d (%d)\n", id, ret);
+		ipoib_cm_dma_unmap_rx(priv, IPOIB_CM_RX_SG - 1,
+				      rx->rx_ring[id].mapping);
+		dev_kfree_skb_any(rx->rx_ring[id].skb);
+		rx->rx_ring[id].skb = NULL;
+	}
+
+	return ret;
+}
+
+static struct sk_buff *ipoib_cm_alloc_rx_skb(struct net_device *dev,
+					     struct ipoib_cm_rx_buf *rx_ring,
+					     int id, int frags,
+					     u64 mapping[IPOIB_CM_RX_SG],
+					     gfp_t gfp)
+{
+	struct ipoib_dev_priv *priv = ipoib_priv(dev);
+	struct sk_buff *skb;
+	int i;
+
+	skb = dev_alloc_skb(ALIGN(IPOIB_CM_HEAD_SIZE + IPOIB_PSEUDO_LEN, 16));
+	if (unlikely(!skb))
+		return NULL;
+
+	/*
+	 * IPoIB adds a IPOIB_ENCAP_LEN byte header, this will align the
+	 * IP header to a multiple of 16.
+	 */
+	skb_reserve(skb, IPOIB_CM_RX_RESERVE);
+
+	mapping[0] = ib_dma_map_single(priv->ca, skb->data, IPOIB_CM_HEAD_SIZE,
+				       DMA_FROM_DEVICE);
+	if (unlikely(ib_dma_mapping_error(priv->ca, mapping[0]))) {
+		dev_kfree_skb_any(skb);
+		return NULL;
+	}
+
+	for (i = 0; i < frags; i++) {
+		struct page *page = alloc_page(gfp);
+
+		if (!page)
+			goto partial_error;
+		skb_fill_page_desc(skb, i, page, 0, PAGE_SIZE);
+
+		mapping[i + 1] = ib_dma_map_page(priv->ca, page,
+						 0, PAGE_SIZE, DMA_FROM_DEVICE);
+		if (unlikely(ib_dma_mapping_error(priv->ca, mapping[i + 1])))
+			goto partial_error;
+	}
+
+	rx_ring[id].skb = skb;
+	return skb;
+
+partial_error:
+
+	ib_dma_unmap_single(priv->ca, mapping[0], IPOIB_CM_HEAD_SIZE, DMA_FROM_DEVICE);
+
+	for (; i > 0; --i)
+		ib_dma_unmap_page(priv->ca, mapping[i], PAGE_SIZE, DMA_FROM_DEVICE);
+
+	dev_kfree_skb_any(skb);
+	return NULL;
+}
+
+static void ipoib_cm_free_rx_ring(struct net_device *dev,
+				  struct ipoib_cm_rx_buf *rx_ring)
+{
+	struct ipoib_dev_priv *priv = ipoib_priv(dev);
+	int i;
+
+	for (i = 0; i < ipoib_recvq_size; ++i)
+		if (rx_ring[i].skb) {
+			ipoib_cm_dma_unmap_rx(priv, IPOIB_CM_RX_SG - 1,
+					      rx_ring[i].mapping);
+			dev_kfree_skb_any(rx_ring[i].skb);
+		}
+
+	vfree(rx_ring);
+}
+
+static void ipoib_cm_start_rx_drain(struct ipoib_dev_priv *priv)
+{
+	struct ipoib_cm_rx *p;
+
+	/* We only reserved 1 extra slot in CQ for drain WRs, so
+	 * make sure we have at most 1 outstanding WR. */
+	if (list_empty(&priv->cm.rx_flush_list) ||
+	    !list_empty(&priv->cm.rx_drain_list))
+		return;
+
+	/*
+	 * QPs on flush list are error state.  This way, a "flush
+	 * error" WC will be immediately generated for each WR we post.
+	 */
+	p = list_entry(priv->cm.rx_flush_list.next, typeof(*p), list);
+	ipoib_cm_rx_drain_wr.wr_id = IPOIB_CM_RX_DRAIN_WRID;
+	if (ib_post_send(p->qp, &ipoib_cm_rx_drain_wr, NULL))
+		ipoib_warn(priv, "failed to post drain wr\n");
+
+	list_splice_init(&priv->cm.rx_flush_list, &priv->cm.rx_drain_list);
+}
+
+static void ipoib_cm_rx_event_handler(struct ib_event *event, void *ctx)
+{
+	struct ipoib_cm_rx *p = ctx;
+	struct ipoib_dev_priv *priv = ipoib_priv(p->dev);
+	unsigned long flags;
+
+	if (event->event != IB_EVENT_QP_LAST_WQE_REACHED)
+		return;
+
+	spin_lock_irqsave(&priv->lock, flags);
+	list_move(&p->list, &priv->cm.rx_flush_list);
+	p->state = IPOIB_CM_RX_FLUSH;
+	ipoib_cm_start_rx_drain(priv);
+	spin_unlock_irqrestore(&priv->lock, flags);
+}
+
+static struct ib_qp *ipoib_cm_create_rx_qp(struct net_device *dev,
+					   struct ipoib_cm_rx *p)
+{
+	struct ipoib_dev_priv *priv = ipoib_priv(dev);
+	struct ib_qp_init_attr attr = {
+		.event_handler = ipoib_cm_rx_event_handler,
+		.send_cq = priv->recv_cq, /* For drain WR */
+		.recv_cq = priv->recv_cq,
+		.srq = priv->cm.srq,
+		.cap.max_send_wr = 1, /* For drain WR */
+		.cap.max_send_sge = 1, /* FIXME: 0 Seems not to work */
+		.sq_sig_type = IB_SIGNAL_ALL_WR,
+		.qp_type = IB_QPT_RC,
+		.qp_context = p,
+	};
+
+	if (!ipoib_cm_has_srq(dev)) {
+		attr.cap.max_recv_wr  = ipoib_recvq_size;
+		attr.cap.max_recv_sge = IPOIB_CM_RX_SG;
+	}
+
+	return ib_create_qp(priv->pd, &attr);
+}
+
+static int ipoib_cm_modify_rx_qp(struct net_device *dev,
+				 struct ib_cm_id *cm_id, struct ib_qp *qp,
+				 unsigned int psn)
+{
+	struct ipoib_dev_priv *priv = ipoib_priv(dev);
+	struct ib_qp_attr qp_attr;
+	int qp_attr_mask, ret;
+
+	qp_attr.qp_state = IB_QPS_INIT;
+	ret = ib_cm_init_qp_attr(cm_id, &qp_attr, &qp_attr_mask);
+	if (ret) {
+		ipoib_warn(priv, "failed to init QP attr for INIT: %d\n", ret);
+		return ret;
+	}
+	ret = ib_modify_qp(qp, &qp_attr, qp_attr_mask);
+	if (ret) {
+		ipoib_warn(priv, "failed to modify QP to INIT: %d\n", ret);
+		return ret;
+	}
+	qp_attr.qp_state = IB_QPS_RTR;
+	ret = ib_cm_init_qp_attr(cm_id, &qp_attr, &qp_attr_mask);
+	if (ret) {
+		ipoib_warn(priv, "failed to init QP attr for RTR: %d\n", ret);
+		return ret;
+	}
+	qp_attr.rq_psn = psn;
+	ret = ib_modify_qp(qp, &qp_attr, qp_attr_mask);
+	if (ret) {
+		ipoib_warn(priv, "failed to modify QP to RTR: %d\n", ret);
+		return ret;
+	}
+
+	/*
+	 * Current Mellanox HCA firmware won't generate completions
+	 * with error for drain WRs unless the QP has been moved to
+	 * RTS first. This work-around leaves a window where a QP has
+	 * moved to error asynchronously, but this will eventually get
+	 * fixed in firmware, so let's not error out if modify QP
+	 * fails.
+	 */
+	qp_attr.qp_state = IB_QPS_RTS;
+	ret = ib_cm_init_qp_attr(cm_id, &qp_attr, &qp_attr_mask);
+	if (ret) {
+		ipoib_warn(priv, "failed to init QP attr for RTS: %d\n", ret);
+		return 0;
+	}
+	ret = ib_modify_qp(qp, &qp_attr, qp_attr_mask);
+	if (ret) {
+		ipoib_warn(priv, "failed to modify QP to RTS: %d\n", ret);
+		return 0;
+	}
+
+	return 0;
+}
+
+static void ipoib_cm_init_rx_wr(struct net_device *dev,
+				struct ib_recv_wr *wr,
+				struct ib_sge *sge)
+{
+	struct ipoib_dev_priv *priv = ipoib_priv(dev);
+	int i;
+
+	for (i = 0; i < priv->cm.num_frags; ++i)
+		sge[i].lkey = priv->pd->local_dma_lkey;
+
+	sge[0].length = IPOIB_CM_HEAD_SIZE;
+	for (i = 1; i < priv->cm.num_frags; ++i)
+		sge[i].length = PAGE_SIZE;
+
+	wr->next    = NULL;
+	wr->sg_list = sge;
+	wr->num_sge = priv->cm.num_frags;
+}
+
+static int ipoib_cm_nonsrq_init_rx(struct net_device *dev, struct ib_cm_id *cm_id,
+				   struct ipoib_cm_rx *rx)
+{
+	struct ipoib_dev_priv *priv = ipoib_priv(dev);
+	struct {
+		struct ib_recv_wr wr;
+		struct ib_sge sge[IPOIB_CM_RX_SG];
+	} *t;
+	int ret;
+	int i;
+
+	rx->rx_ring = vzalloc(array_size(ipoib_recvq_size,
+					 sizeof(*rx->rx_ring)));
+	if (!rx->rx_ring)
+		return -ENOMEM;
+
+	t = kmalloc(sizeof(*t), GFP_KERNEL);
+	if (!t) {
+		ret = -ENOMEM;
+		goto err_free_1;
+	}
+
+	ipoib_cm_init_rx_wr(dev, &t->wr, t->sge);
+
+	spin_lock_irq(&priv->lock);
+
+	if (priv->cm.nonsrq_conn_qp >= ipoib_max_conn_qp) {
+		spin_unlock_irq(&priv->lock);
+		ib_send_cm_rej(cm_id, IB_CM_REJ_NO_QP, NULL, 0, NULL, 0);
+		ret = -EINVAL;
+		goto err_free;
+	} else
+		++priv->cm.nonsrq_conn_qp;
+
+	spin_unlock_irq(&priv->lock);
+
+	for (i = 0; i < ipoib_recvq_size; ++i) {
+		if (!ipoib_cm_alloc_rx_skb(dev, rx->rx_ring, i, IPOIB_CM_RX_SG - 1,
+					   rx->rx_ring[i].mapping,
+					   GFP_KERNEL)) {
+			ipoib_warn(priv, "failed to allocate receive buffer %d\n", i);
+			ret = -ENOMEM;
+			goto err_count;
+		}
+		ret = ipoib_cm_post_receive_nonsrq(dev, rx, &t->wr, t->sge, i);
+		if (ret) {
+			ipoib_warn(priv, "ipoib_cm_post_receive_nonsrq "
+				   "failed for buf %d\n", i);
+			ret = -EIO;
+			goto err_count;
+		}
+	}
+
+	rx->recv_count = ipoib_recvq_size;
+
+	kfree(t);
+
+	return 0;
+
+err_count:
+	spin_lock_irq(&priv->lock);
+	--priv->cm.nonsrq_conn_qp;
+	spin_unlock_irq(&priv->lock);
+
+err_free:
+	kfree(t);
+
+err_free_1:
+	ipoib_cm_free_rx_ring(dev, rx->rx_ring);
+
+	return ret;
+}
+
+static int ipoib_cm_send_rep(struct net_device *dev, struct ib_cm_id *cm_id,
+			     struct ib_qp *qp,
+			     const struct ib_cm_req_event_param *req,
+			     unsigned int psn)
+{
+	struct ipoib_dev_priv *priv = ipoib_priv(dev);
+	struct ipoib_cm_data data = {};
+	struct ib_cm_rep_param rep = {};
+
+	data.qpn = cpu_to_be32(priv->qp->qp_num);
+	data.mtu = cpu_to_be32(IPOIB_CM_BUF_SIZE);
+
+	rep.private_data = &data;
+	rep.private_data_len = sizeof(data);
+	rep.flow_control = 0;
+	rep.rnr_retry_count = req->rnr_retry_count;
+	rep.srq = ipoib_cm_has_srq(dev);
+	rep.qp_num = qp->qp_num;
+	rep.starting_psn = psn;
+	return ib_send_cm_rep(cm_id, &rep);
+}
+
+static int ipoib_cm_req_handler(struct ib_cm_id *cm_id,
+				const struct ib_cm_event *event)
+{
+	struct net_device *dev = cm_id->context;
+	struct ipoib_dev_priv *priv = ipoib_priv(dev);
+	struct ipoib_cm_rx *p;
+	unsigned int psn;
+	int ret;
+
+	ipoib_dbg(priv, "REQ arrived\n");
+	p = kzalloc(sizeof(*p), GFP_KERNEL);
+	if (!p)
+		return -ENOMEM;
+	p->dev = dev;
+	p->id = cm_id;
+	cm_id->context = p;
+	p->state = IPOIB_CM_RX_LIVE;
+	p->jiffies = jiffies;
+	INIT_LIST_HEAD(&p->list);
+
+	p->qp = ipoib_cm_create_rx_qp(dev, p);
+	if (IS_ERR(p->qp)) {
+		ret = PTR_ERR(p->qp);
+		goto err_qp;
+	}
+
+	psn = get_random_u32() & 0xffffff;
+	ret = ipoib_cm_modify_rx_qp(dev, cm_id, p->qp, psn);
+	if (ret)
+		goto err_modify;
+
+	if (!ipoib_cm_has_srq(dev)) {
+		ret = ipoib_cm_nonsrq_init_rx(dev, cm_id, p);
+		if (ret)
+			goto err_modify;
+	}
+
+	spin_lock_irq(&priv->lock);
+	queue_delayed_work(priv->wq,
+			   &priv->cm.stale_task, IPOIB_CM_RX_DELAY);
+	/* Add this entry to passive ids list head, but do not re-add it
+	 * if IB_EVENT_QP_LAST_WQE_REACHED has moved it to flush list. */
+	p->jiffies = jiffies;
+	if (p->state == IPOIB_CM_RX_LIVE)
+		list_move(&p->list, &priv->cm.passive_ids);
+	spin_unlock_irq(&priv->lock);
+
+	ret = ipoib_cm_send_rep(dev, cm_id, p->qp, &event->param.req_rcvd, psn);
+	if (ret) {
+		ipoib_warn(priv, "failed to send REP: %d\n", ret);
+		if (ib_modify_qp(p->qp, &ipoib_cm_err_attr, IB_QP_STATE))
+			ipoib_warn(priv, "unable to move qp to error state\n");
+	}
+	return 0;
+
+err_modify:
+	ib_destroy_qp(p->qp);
+err_qp:
+	kfree(p);
+	return ret;
+}
+
+static int ipoib_cm_rx_handler(struct ib_cm_id *cm_id,
+			       const struct ib_cm_event *event)
+{
+	struct ipoib_cm_rx *p;
+	struct ipoib_dev_priv *priv;
+
+	switch (event->event) {
+	case IB_CM_REQ_RECEIVED:
+		return ipoib_cm_req_handler(cm_id, event);
+	case IB_CM_DREQ_RECEIVED:
+		ib_send_cm_drep(cm_id, NULL, 0);
+		fallthrough;
+	case IB_CM_REJ_RECEIVED:
+		p = cm_id->context;
+		priv = ipoib_priv(p->dev);
+		if (ib_modify_qp(p->qp, &ipoib_cm_err_attr, IB_QP_STATE))
+			ipoib_warn(priv, "unable to move qp to error state\n");
+		fallthrough;
+	default:
+		return 0;
+	}
+}
+/* Adjust length of skb with fragments to match received data */
+static void skb_put_frags(struct sk_buff *skb, unsigned int hdr_space,
+			  unsigned int length, struct sk_buff *toskb)
+{
+	int i, num_frags;
+	unsigned int size;
+
+	/* put header into skb */
+	size = min(length, hdr_space);
+	skb->tail += size;
+	skb->len += size;
+	length -= size;
+
+	num_frags = skb_shinfo(skb)->nr_frags;
+	for (i = 0; i < num_frags; i++) {
+		skb_frag_t *frag = &skb_shinfo(skb)->frags[i];
+
+		if (length == 0) {
+			/* don't need this page */
+			skb_fill_page_desc(toskb, i, skb_frag_page(frag),
+					   0, PAGE_SIZE);
+			--skb_shinfo(skb)->nr_frags;
+		} else {
+			size = min_t(unsigned int, length, PAGE_SIZE);
+
+			skb_frag_size_set(frag, size);
+			skb->data_len += size;
+			skb->truesize += size;
+			skb->len += size;
+			length -= size;
+		}
+	}
+}
+
+void ipoib_cm_handle_rx_wc(struct net_device *dev, struct ib_wc *wc)
+{
+	struct ipoib_dev_priv *priv = ipoib_priv(dev);
+	struct ipoib_cm_rx_buf *rx_ring;
+	unsigned int wr_id = wc->wr_id & ~(IPOIB_OP_CM | IPOIB_OP_RECV);
+	struct sk_buff *skb, *newskb;
+	struct ipoib_cm_rx *p;
+	unsigned long flags;
+	u64 mapping[IPOIB_CM_RX_SG];
+	int frags;
+	int has_srq;
+	struct sk_buff *small_skb;
+
+	ipoib_dbg_data(priv, "cm recv completion: id %d, status: %d\n",
+		       wr_id, wc->status);
+
+	if (unlikely(wr_id >= ipoib_recvq_size)) {
+		if (wr_id == (IPOIB_CM_RX_DRAIN_WRID & ~(IPOIB_OP_CM | IPOIB_OP_RECV))) {
+			spin_lock_irqsave(&priv->lock, flags);
+			list_splice_init(&priv->cm.rx_drain_list, &priv->cm.rx_reap_list);
+			ipoib_cm_start_rx_drain(priv);
+			queue_work(priv->wq, &priv->cm.rx_reap_task);
+			spin_unlock_irqrestore(&priv->lock, flags);
+		} else
+			ipoib_warn(priv, "cm recv completion event with wrid %d (> %d)\n",
+				   wr_id, ipoib_recvq_size);
+		return;
+	}
+
+	p = wc->qp->qp_context;
+
+	has_srq = ipoib_cm_has_srq(dev);
+	rx_ring = has_srq ? priv->cm.srq_ring : p->rx_ring;
+
+	skb = rx_ring[wr_id].skb;
+
+	if (unlikely(wc->status != IB_WC_SUCCESS)) {
+		ipoib_dbg(priv,
+			  "cm recv error (status=%d, wrid=%d vend_err %#x)\n",
+			  wc->status, wr_id, wc->vendor_err);
+		++dev->stats.rx_dropped;
+		if (has_srq)
+			goto repost;
+		else {
+			if (!--p->recv_count) {
+				spin_lock_irqsave(&priv->lock, flags);
+				list_move(&p->list, &priv->cm.rx_reap_list);
+				spin_unlock_irqrestore(&priv->lock, flags);
+				queue_work(priv->wq, &priv->cm.rx_reap_task);
+			}
+			return;
+		}
+	}
+
+	if (unlikely(!(wr_id & IPOIB_CM_RX_UPDATE_MASK))) {
+		if (p && time_after_eq(jiffies, p->jiffies + IPOIB_CM_RX_UPDATE_TIME)) {
+			spin_lock_irqsave(&priv->lock, flags);
+			p->jiffies = jiffies;
+			/* Move this entry to list head, but do not re-add it
+			 * if it has been moved out of list. */
+			if (p->state == IPOIB_CM_RX_LIVE)
+				list_move(&p->list, &priv->cm.passive_ids);
+			spin_unlock_irqrestore(&priv->lock, flags);
+		}
+	}
+
+	if (wc->byte_len < IPOIB_CM_COPYBREAK) {
+		int dlen = wc->byte_len;
+
+		small_skb = dev_alloc_skb(dlen + IPOIB_CM_RX_RESERVE);
+		if (small_skb) {
+			skb_reserve(small_skb, IPOIB_CM_RX_RESERVE);
+			ib_dma_sync_single_for_cpu(priv->ca, rx_ring[wr_id].mapping[0],
+						   dlen, DMA_FROM_DEVICE);
+			skb_copy_from_linear_data(skb, small_skb->data, dlen);
+			ib_dma_sync_single_for_device(priv->ca, rx_ring[wr_id].mapping[0],
+						      dlen, DMA_FROM_DEVICE);
+			skb_put(small_skb, dlen);
+			skb = small_skb;
+			goto copied;
+		}
+	}
+
+	frags = PAGE_ALIGN(wc->byte_len -
+			   min_t(u32, wc->byte_len, IPOIB_CM_HEAD_SIZE)) /
+		PAGE_SIZE;
+
+	newskb = ipoib_cm_alloc_rx_skb(dev, rx_ring, wr_id, frags,
+				       mapping, GFP_ATOMIC);
+	if (unlikely(!newskb)) {
+		/*
+		 * If we can't allocate a new RX buffer, dump
+		 * this packet and reuse the old buffer.
+		 */
+		ipoib_dbg(priv, "failed to allocate receive buffer %d\n", wr_id);
+		++dev->stats.rx_dropped;
+		goto repost;
+	}
+
+	ipoib_cm_dma_unmap_rx(priv, frags, rx_ring[wr_id].mapping);
+	memcpy(rx_ring[wr_id].mapping, mapping, (frags + 1) * sizeof(*mapping));
+
+	ipoib_dbg_data(priv, "received %d bytes, SLID 0x%04x\n",
+		       wc->byte_len, wc->slid);
+
+	skb_put_frags(skb, IPOIB_CM_HEAD_SIZE, wc->byte_len, newskb);
+
+copied:
+	skb->protocol = ((struct ipoib_header *) skb->data)->proto;
+	skb_add_pseudo_hdr(skb);
+
+	++dev->stats.rx_packets;
+	dev->stats.rx_bytes += skb->len;
+
+	skb->dev = dev;
+	/* XXX get correct PACKET_ type here */
+	skb->pkt_type = PACKET_HOST;
+	netif_receive_skb(skb);
+
+repost:
+	if (has_srq) {
+		if (unlikely(ipoib_cm_post_receive_srq(dev, wr_id)))
+			ipoib_warn(priv, "ipoib_cm_post_receive_srq failed "
+				   "for buf %d\n", wr_id);
+	} else {
+		if (unlikely(ipoib_cm_post_receive_nonsrq(dev, p,
+							  &priv->cm.rx_wr,
+							  priv->cm.rx_sge,
+							  wr_id))) {
+			--p->recv_count;
+			ipoib_warn(priv, "ipoib_cm_post_receive_nonsrq failed "
+				   "for buf %d\n", wr_id);
+		}
+	}
+}
+
+static inline int post_send(struct ipoib_dev_priv *priv,
+			    struct ipoib_cm_tx *tx,
+			    unsigned int wr_id,
+			    struct ipoib_tx_buf *tx_req)
+{
+	ipoib_build_sge(priv, tx_req);
+
+	priv->tx_wr.wr.wr_id	= wr_id | IPOIB_OP_CM;
+
+	return ib_post_send(tx->qp, &priv->tx_wr.wr, NULL);
+}
+
+void ipoib_cm_send(struct net_device *dev, struct sk_buff *skb, struct ipoib_cm_tx *tx)
+{
+	struct ipoib_dev_priv *priv = ipoib_priv(dev);
+	struct ipoib_tx_buf *tx_req;
+	int rc;
+	unsigned int usable_sge = tx->max_send_sge - !!skb_headlen(skb);
+
+	if (unlikely(skb->len > tx->mtu)) {
+		ipoib_warn(priv, "packet len %d (> %d) too long to send, dropping\n",
+			   skb->len, tx->mtu);
+		++dev->stats.tx_dropped;
+		++dev->stats.tx_errors;
+		ipoib_cm_skb_too_long(dev, skb, tx->mtu - IPOIB_ENCAP_LEN);
+		return;
+	}
+	if (skb_shinfo(skb)->nr_frags > usable_sge) {
+		if (skb_linearize(skb) < 0) {
+			ipoib_warn(priv, "skb could not be linearized\n");
+			++dev->stats.tx_dropped;
+			++dev->stats.tx_errors;
+			dev_kfree_skb_any(skb);
+			return;
+		}
+		/* Does skb_linearize return ok without reducing nr_frags? */
+		if (skb_shinfo(skb)->nr_frags > usable_sge) {
+			ipoib_warn(priv, "too many frags after skb linearize\n");
+			++dev->stats.tx_dropped;
+			++dev->stats.tx_errors;
+			dev_kfree_skb_any(skb);
+			return;
+		}
+	}
+	ipoib_dbg_data(priv, "sending packet: head 0x%x length %d connection 0x%x\n",
+		       tx->tx_head, skb->len, tx->qp->qp_num);
+
+	/*
+	 * We put the skb into the tx_ring _before_ we call post_send()
+	 * because it's entirely possible that the completion handler will
+	 * run before we execute anything after the post_send().  That
+	 * means we have to make sure everything is properly recorded and
+	 * our state is consistent before we call post_send().
+	 */
+	tx_req = &tx->tx_ring[tx->tx_head & (ipoib_sendq_size - 1)];
+	tx_req->skb = skb;
+
+	if (unlikely(ipoib_dma_map_tx(priv->ca, tx_req))) {
+		++dev->stats.tx_errors;
+		dev_kfree_skb_any(skb);
+		return;
+	}
+
+	if ((priv->global_tx_head - priv->global_tx_tail) ==
+	    ipoib_sendq_size - 1) {
+		ipoib_dbg(priv, "TX ring 0x%x full, stopping kernel net queue\n",
+			  tx->qp->qp_num);
+		netif_stop_queue(dev);
+	}
+
+	skb_orphan(skb);
+	skb_dst_drop(skb);
+
+	if (netif_queue_stopped(dev)) {
+		rc = ib_req_notify_cq(priv->send_cq, IB_CQ_NEXT_COMP |
+				      IB_CQ_REPORT_MISSED_EVENTS);
+		if (unlikely(rc < 0))
+			ipoib_warn(priv, "IPoIB/CM:request notify on send CQ failed\n");
+		else if (rc)
+			napi_schedule(&priv->send_napi);
+	}
+
+	rc = post_send(priv, tx, tx->tx_head & (ipoib_sendq_size - 1), tx_req);
+	if (unlikely(rc)) {
+		ipoib_warn(priv, "IPoIB/CM:post_send failed, error %d\n", rc);
+		++dev->stats.tx_errors;
+		ipoib_dma_unmap_tx(priv, tx_req);
+		dev_kfree_skb_any(skb);
+
+		if (netif_queue_stopped(dev))
+			netif_wake_queue(dev);
+	} else {
+		netif_trans_update(dev);
+		++tx->tx_head;
+		++priv->global_tx_head;
+	}
+}
+
+void ipoib_cm_handle_tx_wc(struct net_device *dev, struct ib_wc *wc)
+{
+	struct ipoib_dev_priv *priv = ipoib_priv(dev);
+	struct ipoib_cm_tx *tx = wc->qp->qp_context;
+	unsigned int wr_id = wc->wr_id & ~IPOIB_OP_CM;
+	struct ipoib_tx_buf *tx_req;
+	unsigned long flags;
+
+	ipoib_dbg_data(priv, "cm send completion: id %d, status: %d\n",
+		       wr_id, wc->status);
+
+	if (unlikely(wr_id >= ipoib_sendq_size)) {
+		ipoib_warn(priv, "cm send completion event with wrid %d (> %d)\n",
+			   wr_id, ipoib_sendq_size);
+		return;
+	}
+
+	tx_req = &tx->tx_ring[wr_id];
+
+	ipoib_dma_unmap_tx(priv, tx_req);
+
+	/* FIXME: is this right? Shouldn't we only increment on success? */
+	++dev->stats.tx_packets;
+	dev->stats.tx_bytes += tx_req->skb->len;
+
+	dev_kfree_skb_any(tx_req->skb);
+
+	netif_tx_lock(dev);
+
+	++tx->tx_tail;
+	++priv->global_tx_tail;
+
+	if (unlikely(netif_queue_stopped(dev) &&
+		     ((priv->global_tx_head - priv->global_tx_tail) <=
+		      ipoib_sendq_size >> 1) &&
+		     test_bit(IPOIB_FLAG_ADMIN_UP, &priv->flags)))
+		netif_wake_queue(dev);
+
+	if (wc->status != IB_WC_SUCCESS &&
+	    wc->status != IB_WC_WR_FLUSH_ERR) {
+		struct ipoib_neigh *neigh;
+
+		/* IB_WC[_RNR]_RETRY_EXC_ERR error is part of the life cycle,
+		 * so don't make waves.
+		 */
+		if (wc->status == IB_WC_RNR_RETRY_EXC_ERR ||
+		    wc->status == IB_WC_RETRY_EXC_ERR)
+			ipoib_dbg(priv,
+				  "%s: failed cm send event (status=%d, wrid=%d vend_err %#x)\n",
+				   __func__, wc->status, wr_id, wc->vendor_err);
+		else
+			ipoib_warn(priv,
+				    "%s: failed cm send event (status=%d, wrid=%d vend_err %#x)\n",
+				   __func__, wc->status, wr_id, wc->vendor_err);
+
+		spin_lock_irqsave(&priv->lock, flags);
+		neigh = tx->neigh;
+
+		if (neigh) {
+			neigh->cm = NULL;
+			ipoib_neigh_free(neigh);
+
+			tx->neigh = NULL;
+		}
+
+		if (test_and_clear_bit(IPOIB_FLAG_INITIALIZED, &tx->flags)) {
+			list_move(&tx->list, &priv->cm.reap_list);
+			queue_work(priv->wq, &priv->cm.reap_task);
+		}
+
+		clear_bit(IPOIB_FLAG_OPER_UP, &tx->flags);
+
+		spin_unlock_irqrestore(&priv->lock, flags);
+	}
+
+	netif_tx_unlock(dev);
+}
+
+int ipoib_cm_dev_open(struct net_device *dev)
+{
+	struct ipoib_dev_priv *priv = ipoib_priv(dev);
+	int ret;
+
+	if (!IPOIB_CM_SUPPORTED(dev->dev_addr))
+		return 0;
+
+	priv->cm.id = ib_create_cm_id(priv->ca, ipoib_cm_rx_handler, dev);
+	if (IS_ERR(priv->cm.id)) {
+		pr_warn("%s: failed to create CM ID\n", priv->ca->name);
+		ret = PTR_ERR(priv->cm.id);
+		goto err_cm;
+	}
+
+	ret = ib_cm_listen(priv->cm.id,
+			   cpu_to_be64(IPOIB_CM_IETF_ID | priv->qp->qp_num));
+	if (ret) {
+		pr_warn("%s: failed to listen on ID 0x%llx\n", priv->ca->name,
+			IPOIB_CM_IETF_ID | priv->qp->qp_num);
+		goto err_listen;
+	}
+
+	return 0;
+
+err_listen:
+	ib_destroy_cm_id(priv->cm.id);
+err_cm:
+	priv->cm.id = NULL;
+	return ret;
+}
+
+static void ipoib_cm_free_rx_reap_list(struct net_device *dev)
+{
+	struct ipoib_dev_priv *priv = ipoib_priv(dev);
+	struct ipoib_cm_rx *rx, *n;
+	LIST_HEAD(list);
+
+	spin_lock_irq(&priv->lock);
+	list_splice_init(&priv->cm.rx_reap_list, &list);
+	spin_unlock_irq(&priv->lock);
+
+	list_for_each_entry_safe(rx, n, &list, list) {
+		ib_destroy_cm_id(rx->id);
+		ib_destroy_qp(rx->qp);
+		if (!ipoib_cm_has_srq(dev)) {
+			ipoib_cm_free_rx_ring(priv->dev, rx->rx_ring);
+			spin_lock_irq(&priv->lock);
+			--priv->cm.nonsrq_conn_qp;
+			spin_unlock_irq(&priv->lock);
+		}
+		kfree(rx);
+	}
+}
+
+void ipoib_cm_dev_stop(struct net_device *dev)
+{
+	struct ipoib_dev_priv *priv = ipoib_priv(dev);
+	struct ipoib_cm_rx *p;
+	unsigned long begin;
+	int ret;
+
+	if (!IPOIB_CM_SUPPORTED(dev->dev_addr) || !priv->cm.id)
+		return;
+
+	ib_destroy_cm_id(priv->cm.id);
+	priv->cm.id = NULL;
+
+	spin_lock_irq(&priv->lock);
+	while (!list_empty(&priv->cm.passive_ids)) {
+		p = list_entry(priv->cm.passive_ids.next, typeof(*p), list);
+		list_move(&p->list, &priv->cm.rx_error_list);
+		p->state = IPOIB_CM_RX_ERROR;
+		spin_unlock_irq(&priv->lock);
+		ret = ib_modify_qp(p->qp, &ipoib_cm_err_attr, IB_QP_STATE);
+		if (ret)
+			ipoib_warn(priv, "unable to move qp to error state: %d\n", ret);
+		spin_lock_irq(&priv->lock);
+	}
+
+	/* Wait for all RX to be drained */
+	begin = jiffies;
+
+	while (!list_empty(&priv->cm.rx_error_list) ||
+	       !list_empty(&priv->cm.rx_flush_list) ||
+	       !list_empty(&priv->cm.rx_drain_list)) {
+		if (time_after(jiffies, begin + 5 * HZ)) {
+			ipoib_warn(priv, "RX drain timing out\n");
+
+			/*
+			 * assume the HW is wedged and just free up everything.
+			 */
+			list_splice_init(&priv->cm.rx_flush_list,
+					 &priv->cm.rx_reap_list);
+			list_splice_init(&priv->cm.rx_error_list,
+					 &priv->cm.rx_reap_list);
+			list_splice_init(&priv->cm.rx_drain_list,
+					 &priv->cm.rx_reap_list);
+			break;
+		}
+		spin_unlock_irq(&priv->lock);
+		usleep_range(1000, 2000);
+		ipoib_drain_cq(dev);
+		spin_lock_irq(&priv->lock);
+	}
+
+	spin_unlock_irq(&priv->lock);
+
+	ipoib_cm_free_rx_reap_list(dev);
+
+	cancel_delayed_work(&priv->cm.stale_task);
+}
+
+static int ipoib_cm_rep_handler(struct ib_cm_id *cm_id,
+				const struct ib_cm_event *event)
+{
+	struct ipoib_cm_tx *p = cm_id->context;
+	struct ipoib_dev_priv *priv = ipoib_priv(p->dev);
+	struct ipoib_cm_data *data = event->private_data;
+	struct sk_buff_head skqueue;
+	struct ib_qp_attr qp_attr;
+	int qp_attr_mask, ret;
+	struct sk_buff *skb;
+
+	p->mtu = be32_to_cpu(data->mtu);
+
+	if (p->mtu <= IPOIB_ENCAP_LEN) {
+		ipoib_warn(priv, "Rejecting connection: mtu %d <= %d\n",
+			   p->mtu, IPOIB_ENCAP_LEN);
+		return -EINVAL;
+	}
+
+	qp_attr.qp_state = IB_QPS_RTR;
+	ret = ib_cm_init_qp_attr(cm_id, &qp_attr, &qp_attr_mask);
+	if (ret) {
+		ipoib_warn(priv, "failed to init QP attr for RTR: %d\n", ret);
+		return ret;
+	}
+
+	qp_attr.rq_psn = 0 /* FIXME */;
+	ret = ib_modify_qp(p->qp, &qp_attr, qp_attr_mask);
+	if (ret) {
+		ipoib_warn(priv, "failed to modify QP to RTR: %d\n", ret);
+		return ret;
+	}
+
+	qp_attr.qp_state = IB_QPS_RTS;
+	ret = ib_cm_init_qp_attr(cm_id, &qp_attr, &qp_attr_mask);
+	if (ret) {
+		ipoib_warn(priv, "failed to init QP attr for RTS: %d\n", ret);
+		return ret;
+	}
+	ret = ib_modify_qp(p->qp, &qp_attr, qp_attr_mask);
+	if (ret) {
+		ipoib_warn(priv, "failed to modify QP to RTS: %d\n", ret);
+		return ret;
+	}
+
+	skb_queue_head_init(&skqueue);
+
+	netif_tx_lock_bh(p->dev);
+	spin_lock_irq(&priv->lock);
+	set_bit(IPOIB_FLAG_OPER_UP, &p->flags);
+	if (p->neigh)
+		while ((skb = __skb_dequeue(&p->neigh->queue)))
+			__skb_queue_tail(&skqueue, skb);
+	spin_unlock_irq(&priv->lock);
+	netif_tx_unlock_bh(p->dev);
+
+	while ((skb = __skb_dequeue(&skqueue))) {
+		skb->dev = p->dev;
+		ret = dev_queue_xmit(skb);
+		if (ret)
+			ipoib_warn(priv, "%s:dev_queue_xmit failed to re-queue packet, ret:%d\n",
+				   __func__, ret);
+	}
+
+	ret = ib_send_cm_rtu(cm_id, NULL, 0);
+	if (ret) {
+		ipoib_warn(priv, "failed to send RTU: %d\n", ret);
+		return ret;
+	}
+	return 0;
+}
+
+static struct ib_qp *ipoib_cm_create_tx_qp(struct net_device *dev, struct ipoib_cm_tx *tx)
+{
+	struct ipoib_dev_priv *priv = ipoib_priv(dev);
+	struct ib_qp_init_attr attr = {
+		.send_cq		= priv->send_cq,
+		.recv_cq		= priv->recv_cq,
+		.srq			= priv->cm.srq,
+		.cap.max_send_wr	= ipoib_sendq_size,
+		.cap.max_send_sge	= 1,
+		.sq_sig_type		= IB_SIGNAL_ALL_WR,
+		.qp_type		= IB_QPT_RC,
+		.qp_context		= tx,
+		.create_flags		= 0
+	};
+	struct ib_qp *tx_qp;
+
+	if (dev->features & NETIF_F_SG)
+		attr.cap.max_send_sge = min_t(u32, priv->ca->attrs.max_send_sge,
+					      MAX_SKB_FRAGS + 1);
+
+	tx_qp = ib_create_qp(priv->pd, &attr);
+	tx->max_send_sge = attr.cap.max_send_sge;
+	return tx_qp;
+}
+
+static int ipoib_cm_send_req(struct net_device *dev,
+			     struct ib_cm_id *id, struct ib_qp *qp,
+			     u32 qpn,
+			     struct sa_path_rec *pathrec)
+{
+	struct ipoib_dev_priv *priv = ipoib_priv(dev);
+	struct ipoib_cm_data data = {};
+	struct ib_cm_req_param req = {};
+
+	data.qpn = cpu_to_be32(priv->qp->qp_num);
+	data.mtu = cpu_to_be32(IPOIB_CM_BUF_SIZE);
+
+	req.primary_path		= pathrec;
+	req.alternate_path		= NULL;
+	req.service_id			= cpu_to_be64(IPOIB_CM_IETF_ID | qpn);
+	req.qp_num			= qp->qp_num;
+	req.qp_type			= qp->qp_type;
+	req.private_data		= &data;
+	req.private_data_len		= sizeof(data);
+	req.flow_control		= 0;
+
+	req.starting_psn		= 0; /* FIXME */
+
+	/*
+	 * Pick some arbitrary defaults here; we could make these
+	 * module parameters if anyone cared about setting them.
+	 */
+	req.responder_resources		= 4;
+	req.remote_cm_response_timeout	= 20;
+	req.local_cm_response_timeout	= 20;
+	req.retry_count			= 0; /* RFC draft warns against retries */
+	req.rnr_retry_count		= 0; /* RFC draft warns against retries */
+	req.max_cm_retries		= 15;
+	req.srq				= ipoib_cm_has_srq(dev);
+	return ib_send_cm_req(id, &req);
+}
+
+static int ipoib_cm_modify_tx_init(struct net_device *dev,
+				  struct ib_cm_id *cm_id, struct ib_qp *qp)
+{
+	struct ipoib_dev_priv *priv = ipoib_priv(dev);
+	struct ib_qp_attr qp_attr;
+	int qp_attr_mask, ret;
+
+	qp_attr.pkey_index = priv->pkey_index;
+	qp_attr.qp_state = IB_QPS_INIT;
+	qp_attr.qp_access_flags = IB_ACCESS_LOCAL_WRITE;
+	qp_attr.port_num = priv->port;
+	qp_attr_mask = IB_QP_STATE | IB_QP_ACCESS_FLAGS | IB_QP_PKEY_INDEX | IB_QP_PORT;
+
+	ret = ib_modify_qp(qp, &qp_attr, qp_attr_mask);
+	if (ret) {
+		ipoib_warn(priv, "failed to modify tx QP to INIT: %d\n", ret);
+		return ret;
+	}
+	return 0;
+}
+
+static int ipoib_cm_tx_init(struct ipoib_cm_tx *p, u32 qpn,
+			    struct sa_path_rec *pathrec)
+{
+	struct ipoib_dev_priv *priv = ipoib_priv(p->dev);
+	unsigned int noio_flag;
+	int ret;
+
+	noio_flag = memalloc_noio_save();
+	p->tx_ring = vzalloc(array_size(ipoib_sendq_size, sizeof(*p->tx_ring)));
+	if (!p->tx_ring) {
+		memalloc_noio_restore(noio_flag);
+		ret = -ENOMEM;
+		goto err_tx;
+	}
+
+	p->qp = ipoib_cm_create_tx_qp(p->dev, p);
+	memalloc_noio_restore(noio_flag);
+	if (IS_ERR(p->qp)) {
+		ret = PTR_ERR(p->qp);
+		ipoib_warn(priv, "failed to create tx qp: %d\n", ret);
+		goto err_qp;
+	}
+
+	p->id = ib_create_cm_id(priv->ca, ipoib_cm_tx_handler, p);
+	if (IS_ERR(p->id)) {
+		ret = PTR_ERR(p->id);
+		ipoib_warn(priv, "failed to create tx cm id: %d\n", ret);
+		goto err_id;
+	}
+
+	ret = ipoib_cm_modify_tx_init(p->dev, p->id,  p->qp);
+	if (ret) {
+		ipoib_warn(priv, "failed to modify tx qp to rtr: %d\n", ret);
+		goto err_modify_send;
+	}
+
+	ret = ipoib_cm_send_req(p->dev, p->id, p->qp, qpn, pathrec);
+	if (ret) {
+		ipoib_warn(priv, "failed to send cm req: %d\n", ret);
+		goto err_modify_send;
+	}
+
+	ipoib_dbg(priv, "Request connection 0x%x for gid %pI6 qpn 0x%x\n",
+		  p->qp->qp_num, pathrec->dgid.raw, qpn);
+
+	return 0;
+
+err_modify_send:
+	ib_destroy_cm_id(p->id);
+err_id:
+	p->id = NULL;
+	ib_destroy_qp(p->qp);
+err_qp:
+	p->qp = NULL;
+	vfree(p->tx_ring);
+err_tx:
+	return ret;
+}
+
+static void ipoib_cm_tx_destroy(struct ipoib_cm_tx *p)
+{
+	struct ipoib_dev_priv *priv = ipoib_priv(p->dev);
+	struct ipoib_tx_buf *tx_req;
+	unsigned long begin;
+
+	ipoib_dbg(priv, "Destroy active connection 0x%x head 0x%x tail 0x%x\n",
+		  p->qp ? p->qp->qp_num : 0, p->tx_head, p->tx_tail);
+
+	if (p->id)
+		ib_destroy_cm_id(p->id);
+
+	if (p->tx_ring) {
+		/* Wait for all sends to complete */
+		begin = jiffies;
+		while ((int) p->tx_tail - (int) p->tx_head < 0) {
+			if (time_after(jiffies, begin + 5 * HZ)) {
+				ipoib_warn(priv, "timing out; %d sends not completed\n",
+					   p->tx_head - p->tx_tail);
+				goto timeout;
+			}
+
+			usleep_range(1000, 2000);
+		}
+	}
+
+timeout:
+
+	while ((int) p->tx_tail - (int) p->tx_head < 0) {
+		tx_req = &p->tx_ring[p->tx_tail & (ipoib_sendq_size - 1)];
+		ipoib_dma_unmap_tx(priv, tx_req);
+		dev_kfree_skb_any(tx_req->skb);
+		netif_tx_lock_bh(p->dev);
+		++p->tx_tail;
+		++priv->global_tx_tail;
+		if (unlikely((priv->global_tx_head - priv->global_tx_tail) <=
+			     ipoib_sendq_size >> 1) &&
+		    netif_queue_stopped(p->dev) &&
+		    test_bit(IPOIB_FLAG_ADMIN_UP, &priv->flags))
+			netif_wake_queue(p->dev);
+		netif_tx_unlock_bh(p->dev);
+	}
+
+	if (p->qp)
+		ib_destroy_qp(p->qp);
+
+	vfree(p->tx_ring);
+	kfree(p);
+}
+
+static int ipoib_cm_tx_handler(struct ib_cm_id *cm_id,
+			       const struct ib_cm_event *event)
+{
+	struct ipoib_cm_tx *tx = cm_id->context;
+	struct ipoib_dev_priv *priv = ipoib_priv(tx->dev);
+	struct net_device *dev = priv->dev;
+	struct ipoib_neigh *neigh;
+	unsigned long flags;
+	int ret;
+
+	switch (event->event) {
+	case IB_CM_DREQ_RECEIVED:
+		ipoib_dbg(priv, "DREQ received.\n");
+		ib_send_cm_drep(cm_id, NULL, 0);
+		break;
+	case IB_CM_REP_RECEIVED:
+		ipoib_dbg(priv, "REP received.\n");
+		ret = ipoib_cm_rep_handler(cm_id, event);
+		if (ret)
+			ib_send_cm_rej(cm_id, IB_CM_REJ_CONSUMER_DEFINED,
+				       NULL, 0, NULL, 0);
+		break;
+	case IB_CM_REQ_ERROR:
+	case IB_CM_REJ_RECEIVED:
+	case IB_CM_TIMEWAIT_EXIT:
+		ipoib_dbg(priv, "CM error %d.\n", event->event);
+		netif_tx_lock_bh(dev);
+		spin_lock_irqsave(&priv->lock, flags);
+		neigh = tx->neigh;
+
+		if (neigh) {
+			neigh->cm = NULL;
+			ipoib_neigh_free(neigh);
+
+			tx->neigh = NULL;
+		}
+
+		if (test_and_clear_bit(IPOIB_FLAG_INITIALIZED, &tx->flags)) {
+			list_move(&tx->list, &priv->cm.reap_list);
+			queue_work(priv->wq, &priv->cm.reap_task);
+		}
+
+		spin_unlock_irqrestore(&priv->lock, flags);
+		netif_tx_unlock_bh(dev);
+		break;
+	default:
+		break;
+	}
+
+	return 0;
+}
+
+struct ipoib_cm_tx *ipoib_cm_create_tx(struct net_device *dev, struct ipoib_path *path,
+				       struct ipoib_neigh *neigh)
+{
+	struct ipoib_dev_priv *priv = ipoib_priv(dev);
+	struct ipoib_cm_tx *tx;
+
+	tx = kzalloc(sizeof(*tx), GFP_ATOMIC);
+	if (!tx)
+		return NULL;
+
+	neigh->cm = tx;
+	tx->neigh = neigh;
+	tx->dev = dev;
+	list_add(&tx->list, &priv->cm.start_list);
+	set_bit(IPOIB_FLAG_INITIALIZED, &tx->flags);
+	queue_work(priv->wq, &priv->cm.start_task);
+	return tx;
+}
+
+void ipoib_cm_destroy_tx(struct ipoib_cm_tx *tx)
+{
+	struct ipoib_dev_priv *priv = ipoib_priv(tx->dev);
+	unsigned long flags;
+	if (test_and_clear_bit(IPOIB_FLAG_INITIALIZED, &tx->flags)) {
+		spin_lock_irqsave(&priv->lock, flags);
+		list_move(&tx->list, &priv->cm.reap_list);
+		queue_work(priv->wq, &priv->cm.reap_task);
+		ipoib_dbg(priv, "Reap connection for gid %pI6\n",
+			  tx->neigh->daddr + 4);
+		tx->neigh = NULL;
+		spin_unlock_irqrestore(&priv->lock, flags);
+	}
+}
+
+#define QPN_AND_OPTIONS_OFFSET	4
+
+static void ipoib_cm_tx_start(struct work_struct *work)
+{
+	struct ipoib_dev_priv *priv = container_of(work, struct ipoib_dev_priv,
+						   cm.start_task);
+	struct net_device *dev = priv->dev;
+	struct ipoib_neigh *neigh;
+	struct ipoib_cm_tx *p;
+	unsigned long flags;
+	struct ipoib_path *path;
+	int ret;
+
+	struct sa_path_rec pathrec;
+	u32 qpn;
+
+	netif_tx_lock_bh(dev);
+	spin_lock_irqsave(&priv->lock, flags);
+
+	while (!list_empty(&priv->cm.start_list)) {
+		p = list_entry(priv->cm.start_list.next, typeof(*p), list);
+		list_del_init(&p->list);
+		neigh = p->neigh;
+
+		qpn = IPOIB_QPN(neigh->daddr);
+		/*
+		 * As long as the search is with these 2 locks,
+		 * path existence indicates its validity.
+		 */
+		path = __path_find(dev, neigh->daddr + QPN_AND_OPTIONS_OFFSET);
+		if (!path) {
+			pr_info("%s ignore not valid path %pI6\n",
+				__func__,
+				neigh->daddr + QPN_AND_OPTIONS_OFFSET);
+			goto free_neigh;
+		}
+		memcpy(&pathrec, &path->pathrec, sizeof(pathrec));
+
+		spin_unlock_irqrestore(&priv->lock, flags);
+		netif_tx_unlock_bh(dev);
+
+		ret = ipoib_cm_tx_init(p, qpn, &pathrec);
+
+		netif_tx_lock_bh(dev);
+		spin_lock_irqsave(&priv->lock, flags);
+
+		if (ret) {
+free_neigh:
+			neigh = p->neigh;
+			if (neigh) {
+				neigh->cm = NULL;
+				ipoib_neigh_free(neigh);
+			}
+			list_del(&p->list);
+			kfree(p);
+		}
+	}
+
+	spin_unlock_irqrestore(&priv->lock, flags);
+	netif_tx_unlock_bh(dev);
+}
+
+static void ipoib_cm_tx_reap(struct work_struct *work)
+{
+	struct ipoib_dev_priv *priv = container_of(work, struct ipoib_dev_priv,
+						   cm.reap_task);
+	struct net_device *dev = priv->dev;
+	struct ipoib_cm_tx *p;
+	unsigned long flags;
+
+	netif_tx_lock_bh(dev);
+	spin_lock_irqsave(&priv->lock, flags);
+
+	while (!list_empty(&priv->cm.reap_list)) {
+		p = list_entry(priv->cm.reap_list.next, typeof(*p), list);
+		list_del_init(&p->list);
+		spin_unlock_irqrestore(&priv->lock, flags);
+		netif_tx_unlock_bh(dev);
+		ipoib_cm_tx_destroy(p);
+		netif_tx_lock_bh(dev);
+		spin_lock_irqsave(&priv->lock, flags);
+	}
+
+	spin_unlock_irqrestore(&priv->lock, flags);
+	netif_tx_unlock_bh(dev);
+}
+
+static void ipoib_cm_skb_reap(struct work_struct *work)
+{
+	struct ipoib_dev_priv *priv = container_of(work, struct ipoib_dev_priv,
+						   cm.skb_task);
+	struct net_device *dev = priv->dev;
+	struct sk_buff *skb;
+	unsigned long flags;
+	unsigned int mtu = priv->mcast_mtu;
+
+	netif_tx_lock_bh(dev);
+	spin_lock_irqsave(&priv->lock, flags);
+
+	while ((skb = skb_dequeue(&priv->cm.skb_queue))) {
+		spin_unlock_irqrestore(&priv->lock, flags);
+		netif_tx_unlock_bh(dev);
+
+		if (skb->protocol == htons(ETH_P_IP)) {
+			memset(IPCB(skb), 0, sizeof(*IPCB(skb)));
+			icmp_send(skb, ICMP_DEST_UNREACH, ICMP_FRAG_NEEDED, htonl(mtu));
+		}
+#if IS_ENABLED(CONFIG_IPV6)
+		else if (skb->protocol == htons(ETH_P_IPV6)) {
+			memset(IP6CB(skb), 0, sizeof(*IP6CB(skb)));
+			icmpv6_send(skb, ICMPV6_PKT_TOOBIG, 0, mtu);
+		}
+#endif
+		dev_kfree_skb_any(skb);
+
+		netif_tx_lock_bh(dev);
+		spin_lock_irqsave(&priv->lock, flags);
+	}
+
+	spin_unlock_irqrestore(&priv->lock, flags);
+	netif_tx_unlock_bh(dev);
+}
+
+void ipoib_cm_skb_too_long(struct net_device *dev, struct sk_buff *skb,
+			   unsigned int mtu)
+{
+	struct ipoib_dev_priv *priv = ipoib_priv(dev);
+	int e = skb_queue_empty(&priv->cm.skb_queue);
+
+	skb_dst_update_pmtu(skb, mtu);
+
+	skb_queue_tail(&priv->cm.skb_queue, skb);
+	if (e)
+		queue_work(priv->wq, &priv->cm.skb_task);
+}
+
+static void ipoib_cm_rx_reap(struct work_struct *work)
+{
+	ipoib_cm_free_rx_reap_list(container_of(work, struct ipoib_dev_priv,
+						cm.rx_reap_task)->dev);
+}
+
+static void ipoib_cm_stale_task(struct work_struct *work)
+{
+	struct ipoib_dev_priv *priv = container_of(work, struct ipoib_dev_priv,
+						   cm.stale_task.work);
+	struct ipoib_cm_rx *p;
+	int ret;
+
+	spin_lock_irq(&priv->lock);
+	while (!list_empty(&priv->cm.passive_ids)) {
+		/* List is sorted by LRU, start from tail,
+		 * stop when we see a recently used entry */
+		p = list_entry(priv->cm.passive_ids.prev, typeof(*p), list);
+		if (time_before_eq(jiffies, p->jiffies + IPOIB_CM_RX_TIMEOUT))
+			break;
+		list_move(&p->list, &priv->cm.rx_error_list);
+		p->state = IPOIB_CM_RX_ERROR;
+		spin_unlock_irq(&priv->lock);
+		ret = ib_modify_qp(p->qp, &ipoib_cm_err_attr, IB_QP_STATE);
+		if (ret)
+			ipoib_warn(priv, "unable to move qp to error state: %d\n", ret);
+		spin_lock_irq(&priv->lock);
+	}
+
+	if (!list_empty(&priv->cm.passive_ids))
+		queue_delayed_work(priv->wq,
+				   &priv->cm.stale_task, IPOIB_CM_RX_DELAY);
+	spin_unlock_irq(&priv->lock);
+}
+
+static ssize_t mode_show(struct device *d, struct device_attribute *attr,
+			 char *buf)
+{
+	struct net_device *dev = to_net_dev(d);
+	struct ipoib_dev_priv *priv = ipoib_priv(dev);
+
+	if (test_bit(IPOIB_FLAG_ADMIN_CM, &priv->flags))
+		return sysfs_emit(buf, "connected\n");
+	else
+		return sysfs_emit(buf, "datagram\n");
+}
+
+static ssize_t mode_store(struct device *d, struct device_attribute *attr,
+			  const char *buf, size_t count)
+{
+	struct net_device *dev = to_net_dev(d);
+	int ret;
+
+	if (!rtnl_trylock()) {
+		return restart_syscall();
+	}
+
+	if (dev->reg_state != NETREG_REGISTERED) {
+		rtnl_unlock();
+		return -EPERM;
+	}
+
+	ret = ipoib_set_mode(dev, buf);
+
+	/* The assumption is that the function ipoib_set_mode returned
+	 * with the rtnl held by it, if not the value -EBUSY returned,
+	 * then no need to rtnl_unlock
+	 */
+	if (ret != -EBUSY)
+		rtnl_unlock();
+
+	return (!ret || ret == -EBUSY) ? count : ret;
+}
+
+static DEVICE_ATTR_RW(mode);
+
+int ipoib_cm_add_mode_attr(struct net_device *dev)
+{
+	return device_create_file(&dev->dev, &dev_attr_mode);
+}
+
+static void ipoib_cm_create_srq(struct net_device *dev, int max_sge)
+{
+	struct ipoib_dev_priv *priv = ipoib_priv(dev);
+	struct ib_srq_init_attr srq_init_attr = {
+		.srq_type = IB_SRQT_BASIC,
+		.attr = {
+			.max_wr  = ipoib_recvq_size,
+			.max_sge = max_sge
+		}
+	};
+
+	priv->cm.srq = ib_create_srq(priv->pd, &srq_init_attr);
+	if (IS_ERR(priv->cm.srq)) {
+		if (PTR_ERR(priv->cm.srq) != -EOPNOTSUPP)
+			pr_warn("%s: failed to allocate SRQ, error %ld\n",
+			       priv->ca->name, PTR_ERR(priv->cm.srq));
+		priv->cm.srq = NULL;
+		return;
+	}
+
+	priv->cm.srq_ring = vzalloc(array_size(ipoib_recvq_size,
+					       sizeof(*priv->cm.srq_ring)));
+	if (!priv->cm.srq_ring) {
+		ib_destroy_srq(priv->cm.srq);
+		priv->cm.srq = NULL;
+		return;
+	}
+
+}
+
+int ipoib_cm_dev_init(struct net_device *dev)
+{
+	struct ipoib_dev_priv *priv = ipoib_priv(dev);
+	int max_srq_sge, i;
+	u8 addr;
+
+	INIT_LIST_HEAD(&priv->cm.passive_ids);
+	INIT_LIST_HEAD(&priv->cm.reap_list);
+	INIT_LIST_HEAD(&priv->cm.start_list);
+	INIT_LIST_HEAD(&priv->cm.rx_error_list);
+	INIT_LIST_HEAD(&priv->cm.rx_flush_list);
+	INIT_LIST_HEAD(&priv->cm.rx_drain_list);
+	INIT_LIST_HEAD(&priv->cm.rx_reap_list);
+	INIT_WORK(&priv->cm.start_task, ipoib_cm_tx_start);
+	INIT_WORK(&priv->cm.reap_task, ipoib_cm_tx_reap);
+	INIT_WORK(&priv->cm.skb_task, ipoib_cm_skb_reap);
+	INIT_WORK(&priv->cm.rx_reap_task, ipoib_cm_rx_reap);
+	INIT_DELAYED_WORK(&priv->cm.stale_task, ipoib_cm_stale_task);
+
+	skb_queue_head_init(&priv->cm.skb_queue);
+
+	ipoib_dbg(priv, "max_srq_sge=%d\n", priv->ca->attrs.max_srq_sge);
+
+	max_srq_sge = min_t(int, IPOIB_CM_RX_SG, priv->ca->attrs.max_srq_sge);
+	ipoib_cm_create_srq(dev, max_srq_sge);
+	if (ipoib_cm_has_srq(dev)) {
+		priv->cm.max_cm_mtu = max_srq_sge * PAGE_SIZE - 0x10;
+		priv->cm.num_frags  = max_srq_sge;
+		ipoib_dbg(priv, "max_cm_mtu = 0x%x, num_frags=%d\n",
+			  priv->cm.max_cm_mtu, priv->cm.num_frags);
+	} else {
+		priv->cm.max_cm_mtu = IPOIB_CM_MTU;
+		priv->cm.num_frags  = IPOIB_CM_RX_SG;
+	}
+
+	ipoib_cm_init_rx_wr(dev, &priv->cm.rx_wr, priv->cm.rx_sge);
+
+	if (ipoib_cm_has_srq(dev)) {
+		for (i = 0; i < ipoib_recvq_size; ++i) {
+			if (!ipoib_cm_alloc_rx_skb(dev, priv->cm.srq_ring, i,
+						   priv->cm.num_frags - 1,
+						   priv->cm.srq_ring[i].mapping,
+						   GFP_KERNEL)) {
+				ipoib_warn(priv, "failed to allocate "
+					   "receive buffer %d\n", i);
+				ipoib_cm_dev_cleanup(dev);
+				return -ENOMEM;
+			}
+
+			if (ipoib_cm_post_receive_srq(dev, i)) {
+				ipoib_warn(priv, "ipoib_cm_post_receive_srq "
+					   "failed for buf %d\n", i);
+				ipoib_cm_dev_cleanup(dev);
+				return -EIO;
+			}
+		}
+	}
+
+	addr = IPOIB_FLAGS_RC;
+	dev_addr_mod(dev, 0, &addr, 1);
+	return 0;
+}
+
+void ipoib_cm_dev_cleanup(struct net_device *dev)
+{
+	struct ipoib_dev_priv *priv = ipoib_priv(dev);
+
+	if (!priv->cm.srq)
+		return;
+
+	ipoib_dbg(priv, "Cleanup ipoib connected mode.\n");
+
+	ib_destroy_srq(priv->cm.srq);
+	priv->cm.srq = NULL;
+	if (!priv->cm.srq_ring)
+		return;
+
+	ipoib_cm_free_rx_ring(dev, priv->cm.srq_ring);
+	priv->cm.srq_ring = NULL;
+}
diff --git a/drivers/infiniband/ulp/ipoib/ipoib_ethtool.c b/drivers/infiniband/ulp/ipoib/ipoib_ethtool.c
new file mode 100644
index 000000000..8af99b18d
--- /dev/null
+++ b/drivers/infiniband/ulp/ipoib/ipoib_ethtool.c
@@ -0,0 +1,236 @@
+/*
+ * Copyright (c) 2007 Mellanox Technologies. All rights reserved.
+ *
+ * This software is available to you under a choice of one of two
+ * licenses.  You may choose to be licensed under the terms of the GNU
+ * General Public License (GPL) Version 2, available from the file
+ * COPYING in the main directory of this source tree, or the
+ * OpenIB.org BSD license below:
+ *
+ *     Redistribution and use in source and binary forms, with or
+ *     without modification, are permitted provided that the following
+ *     conditions are met:
+ *
+ *      - Redistributions of source code must retain the above
+ *        copyright notice, this list of conditions and the following
+ *        disclaimer.
+ *
+ *      - Redistributions in binary form must reproduce the above
+ *        copyright notice, this list of conditions and the following
+ *        disclaimer in the documentation and/or other materials
+ *        provided with the distribution.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
+ * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
+ * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS
+ * BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN
+ * ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
+ * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+ * SOFTWARE.
+ */
+
+#include <linux/kernel.h>
+#include <linux/ethtool.h>
+#include <linux/netdevice.h>
+
+#include "ipoib.h"
+
+struct ipoib_stats {
+	char stat_string[ETH_GSTRING_LEN];
+	int stat_offset;
+};
+
+#define IPOIB_NETDEV_STAT(m) { \
+		.stat_string = #m, \
+		.stat_offset = offsetof(struct rtnl_link_stats64, m) }
+
+static const struct ipoib_stats ipoib_gstrings_stats[] = {
+	IPOIB_NETDEV_STAT(rx_packets),
+	IPOIB_NETDEV_STAT(tx_packets),
+	IPOIB_NETDEV_STAT(rx_bytes),
+	IPOIB_NETDEV_STAT(tx_bytes),
+	IPOIB_NETDEV_STAT(tx_errors),
+	IPOIB_NETDEV_STAT(rx_dropped),
+	IPOIB_NETDEV_STAT(tx_dropped),
+	IPOIB_NETDEV_STAT(multicast),
+};
+
+#define IPOIB_GLOBAL_STATS_LEN	ARRAY_SIZE(ipoib_gstrings_stats)
+
+static void ipoib_get_drvinfo(struct net_device *netdev,
+			      struct ethtool_drvinfo *drvinfo)
+{
+	struct ipoib_dev_priv *priv = ipoib_priv(netdev);
+
+	ib_get_device_fw_str(priv->ca, drvinfo->fw_version);
+
+	strscpy(drvinfo->bus_info, dev_name(priv->ca->dev.parent),
+		sizeof(drvinfo->bus_info));
+
+	strscpy(drvinfo->driver, "ib_ipoib", sizeof(drvinfo->driver));
+}
+
+static int ipoib_get_coalesce(struct net_device *dev,
+			      struct ethtool_coalesce *coal,
+			      struct kernel_ethtool_coalesce *kernel_coal,
+			      struct netlink_ext_ack *extack)
+{
+	struct ipoib_dev_priv *priv = ipoib_priv(dev);
+
+	coal->rx_coalesce_usecs = priv->ethtool.coalesce_usecs;
+	coal->rx_max_coalesced_frames = priv->ethtool.max_coalesced_frames;
+
+	return 0;
+}
+
+static int ipoib_set_coalesce(struct net_device *dev,
+			      struct ethtool_coalesce *coal,
+			      struct kernel_ethtool_coalesce *kernel_coal,
+			      struct netlink_ext_ack *extack)
+{
+	struct ipoib_dev_priv *priv = ipoib_priv(dev);
+	int ret;
+
+	/*
+	 * These values are saved in the private data and returned
+	 * when ipoib_get_coalesce() is called
+	 */
+	if (coal->rx_coalesce_usecs       > 0xffff ||
+	    coal->rx_max_coalesced_frames > 0xffff)
+		return -EINVAL;
+
+	ret = rdma_set_cq_moderation(priv->recv_cq,
+				     coal->rx_max_coalesced_frames,
+				     coal->rx_coalesce_usecs);
+	if (ret && ret != -EOPNOTSUPP) {
+		ipoib_warn(priv, "failed modifying CQ (%d)\n", ret);
+		return ret;
+	}
+
+	priv->ethtool.coalesce_usecs       = coal->rx_coalesce_usecs;
+	priv->ethtool.max_coalesced_frames = coal->rx_max_coalesced_frames;
+
+	return 0;
+}
+static void ipoib_get_ethtool_stats(struct net_device *dev,
+				    struct ethtool_stats __always_unused *stats,
+				    u64 *data)
+{
+	int i;
+	struct net_device_stats *net_stats = &dev->stats;
+	u8 *p = (u8 *)net_stats;
+
+	for (i = 0; i < IPOIB_GLOBAL_STATS_LEN; i++)
+		data[i] = *(u64 *)(p + ipoib_gstrings_stats[i].stat_offset);
+
+}
+static void ipoib_get_strings(struct net_device __always_unused *dev,
+			      u32 stringset, u8 *data)
+{
+	u8 *p = data;
+	int i;
+
+	switch (stringset) {
+	case ETH_SS_STATS:
+		for (i = 0; i < IPOIB_GLOBAL_STATS_LEN; i++) {
+			memcpy(p, ipoib_gstrings_stats[i].stat_string,
+				ETH_GSTRING_LEN);
+			p += ETH_GSTRING_LEN;
+		}
+		break;
+	default:
+		break;
+	}
+}
+static int ipoib_get_sset_count(struct net_device __always_unused *dev,
+				 int sset)
+{
+	switch (sset) {
+	case ETH_SS_STATS:
+		return IPOIB_GLOBAL_STATS_LEN;
+	default:
+		break;
+	}
+	return -EOPNOTSUPP;
+}
+
+/* Return lane speed in unit of 1e6 bit/sec */
+static inline int ib_speed_enum_to_int(int speed)
+{
+	switch (speed) {
+	case IB_SPEED_SDR:
+		return SPEED_2500;
+	case IB_SPEED_DDR:
+		return SPEED_5000;
+	case IB_SPEED_QDR:
+	case IB_SPEED_FDR10:
+		return SPEED_10000;
+	case IB_SPEED_FDR:
+		return SPEED_14000;
+	case IB_SPEED_EDR:
+		return SPEED_25000;
+	case IB_SPEED_HDR:
+		return SPEED_50000;
+	case IB_SPEED_NDR:
+		return SPEED_100000;
+	}
+
+	return SPEED_UNKNOWN;
+}
+
+static int ipoib_get_link_ksettings(struct net_device *netdev,
+				    struct ethtool_link_ksettings *cmd)
+{
+	struct ipoib_dev_priv *priv = ipoib_priv(netdev);
+	struct ib_port_attr attr;
+	int ret, speed, width;
+
+	if (!netif_carrier_ok(netdev)) {
+		cmd->base.speed = SPEED_UNKNOWN;
+		cmd->base.duplex = DUPLEX_UNKNOWN;
+		return 0;
+	}
+
+	ret = ib_query_port(priv->ca, priv->port, &attr);
+	if (ret < 0)
+		return -EINVAL;
+
+	speed = ib_speed_enum_to_int(attr.active_speed);
+	width = ib_width_enum_to_int(attr.active_width);
+
+	if (speed < 0 || width < 0)
+		return -EINVAL;
+
+	/* Except the following are set, the other members of
+	 * the struct ethtool_link_settings are initialized to
+	 * zero in the function __ethtool_get_link_ksettings.
+	 */
+	cmd->base.speed		 = speed * width;
+	cmd->base.duplex	 = DUPLEX_FULL;
+
+	cmd->base.phy_address	 = 0xFF;
+
+	cmd->base.autoneg	 = AUTONEG_ENABLE;
+	cmd->base.port		 = PORT_OTHER;
+
+	return 0;
+}
+
+static const struct ethtool_ops ipoib_ethtool_ops = {
+	.supported_coalesce_params = ETHTOOL_COALESCE_RX_USECS |
+				     ETHTOOL_COALESCE_RX_MAX_FRAMES,
+	.get_link_ksettings	= ipoib_get_link_ksettings,
+	.get_drvinfo		= ipoib_get_drvinfo,
+	.get_coalesce		= ipoib_get_coalesce,
+	.set_coalesce		= ipoib_set_coalesce,
+	.get_strings		= ipoib_get_strings,
+	.get_ethtool_stats	= ipoib_get_ethtool_stats,
+	.get_sset_count		= ipoib_get_sset_count,
+	.get_link		= ethtool_op_get_link,
+};
+
+void ipoib_set_ethtool_ops(struct net_device *dev)
+{
+	dev->ethtool_ops = &ipoib_ethtool_ops;
+}
diff --git a/drivers/infiniband/ulp/ipoib/ipoib_fs.c b/drivers/infiniband/ulp/ipoib/ipoib_fs.c
new file mode 100644
index 000000000..12ba7a0fe
--- /dev/null
+++ b/drivers/infiniband/ulp/ipoib/ipoib_fs.c
@@ -0,0 +1,251 @@
+/*
+ * Copyright (c) 2004 Topspin Communications.  All rights reserved.
+ *
+ * This software is available to you under a choice of one of two
+ * licenses.  You may choose to be licensed under the terms of the GNU
+ * General Public License (GPL) Version 2, available from the file
+ * COPYING in the main directory of this source tree, or the
+ * OpenIB.org BSD license below:
+ *
+ *     Redistribution and use in source and binary forms, with or
+ *     without modification, are permitted provided that the following
+ *     conditions are met:
+ *
+ *      - Redistributions of source code must retain the above
+ *        copyright notice, this list of conditions and the following
+ *        disclaimer.
+ *
+ *      - Redistributions in binary form must reproduce the above
+ *        copyright notice, this list of conditions and the following
+ *        disclaimer in the documentation and/or other materials
+ *        provided with the distribution.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
+ * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
+ * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS
+ * BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN
+ * ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
+ * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+ * SOFTWARE.
+ */
+
+#include <linux/err.h>
+#include <linux/seq_file.h>
+#include <linux/slab.h>
+
+struct file_operations;
+
+#include <linux/debugfs.h>
+#include <linux/export.h>
+
+#include "ipoib.h"
+
+static struct dentry *ipoib_root;
+
+static void format_gid(union ib_gid *gid, char *buf)
+{
+	int i, n;
+
+	for (n = 0, i = 0; i < 8; ++i) {
+		n += sprintf(buf + n, "%x",
+			     be16_to_cpu(((__be16 *) gid->raw)[i]));
+		if (i < 7)
+			buf[n++] = ':';
+	}
+}
+
+static void *ipoib_mcg_seq_start(struct seq_file *file, loff_t *pos)
+{
+	struct ipoib_mcast_iter *iter;
+	loff_t n = *pos;
+
+	iter = ipoib_mcast_iter_init(file->private);
+	if (!iter)
+		return NULL;
+
+	while (n--) {
+		if (ipoib_mcast_iter_next(iter)) {
+			kfree(iter);
+			return NULL;
+		}
+	}
+
+	return iter;
+}
+
+static void *ipoib_mcg_seq_next(struct seq_file *file, void *iter_ptr,
+				   loff_t *pos)
+{
+	struct ipoib_mcast_iter *iter = iter_ptr;
+
+	(*pos)++;
+
+	if (ipoib_mcast_iter_next(iter)) {
+		kfree(iter);
+		return NULL;
+	}
+
+	return iter;
+}
+
+static void ipoib_mcg_seq_stop(struct seq_file *file, void *iter_ptr)
+{
+	/* nothing for now */
+}
+
+static int ipoib_mcg_seq_show(struct seq_file *file, void *iter_ptr)
+{
+	struct ipoib_mcast_iter *iter = iter_ptr;
+	char gid_buf[sizeof "ffff:ffff:ffff:ffff:ffff:ffff:ffff:ffff"];
+	union ib_gid mgid;
+	unsigned long created;
+	unsigned int queuelen, complete, send_only;
+
+	if (!iter)
+		return 0;
+
+	ipoib_mcast_iter_read(iter, &mgid, &created, &queuelen,
+			      &complete, &send_only);
+
+	format_gid(&mgid, gid_buf);
+
+	seq_printf(file,
+		   "GID: %s\n"
+		   "  created: %10ld\n"
+		   "  queuelen: %9d\n"
+		   "  complete: %9s\n"
+		   "  send_only: %8s\n"
+		   "\n",
+		   gid_buf, created, queuelen,
+		   complete ? "yes" : "no",
+		   send_only ? "yes" : "no");
+
+	return 0;
+}
+
+static const struct seq_operations ipoib_mcg_sops = {
+	.start = ipoib_mcg_seq_start,
+	.next  = ipoib_mcg_seq_next,
+	.stop  = ipoib_mcg_seq_stop,
+	.show  = ipoib_mcg_seq_show,
+};
+
+DEFINE_SEQ_ATTRIBUTE(ipoib_mcg);
+
+static void *ipoib_path_seq_start(struct seq_file *file, loff_t *pos)
+{
+	struct ipoib_path_iter *iter;
+	loff_t n = *pos;
+
+	iter = ipoib_path_iter_init(file->private);
+	if (!iter)
+		return NULL;
+
+	while (n--) {
+		if (ipoib_path_iter_next(iter)) {
+			kfree(iter);
+			return NULL;
+		}
+	}
+
+	return iter;
+}
+
+static void *ipoib_path_seq_next(struct seq_file *file, void *iter_ptr,
+				   loff_t *pos)
+{
+	struct ipoib_path_iter *iter = iter_ptr;
+
+	(*pos)++;
+
+	if (ipoib_path_iter_next(iter)) {
+		kfree(iter);
+		return NULL;
+	}
+
+	return iter;
+}
+
+static void ipoib_path_seq_stop(struct seq_file *file, void *iter_ptr)
+{
+	/* nothing for now */
+}
+
+static int ipoib_path_seq_show(struct seq_file *file, void *iter_ptr)
+{
+	struct ipoib_path_iter *iter = iter_ptr;
+	char gid_buf[sizeof "ffff:ffff:ffff:ffff:ffff:ffff:ffff:ffff"];
+	struct ipoib_path path;
+	int rate;
+
+	if (!iter)
+		return 0;
+
+	ipoib_path_iter_read(iter, &path);
+
+	format_gid(&path.pathrec.dgid, gid_buf);
+
+	seq_printf(file,
+		   "GID: %s\n"
+		   "  complete: %6s\n",
+		   gid_buf, sa_path_get_dlid(&path.pathrec) ? "yes" : "no");
+
+	if (sa_path_get_dlid(&path.pathrec)) {
+		rate = ib_rate_to_mbps(path.pathrec.rate);
+
+		seq_printf(file,
+			   "  DLID:     0x%04x\n"
+			   "  SL: %12d\n"
+			   "  rate: %8d.%d Gb/sec\n",
+			   be32_to_cpu(sa_path_get_dlid(&path.pathrec)),
+			   path.pathrec.sl,
+			   rate / 1000, rate % 1000);
+	}
+
+	seq_putc(file, '\n');
+
+	return 0;
+}
+
+static const struct seq_operations ipoib_path_sops = {
+	.start = ipoib_path_seq_start,
+	.next  = ipoib_path_seq_next,
+	.stop  = ipoib_path_seq_stop,
+	.show  = ipoib_path_seq_show,
+};
+
+DEFINE_SEQ_ATTRIBUTE(ipoib_path);
+
+void ipoib_create_debug_files(struct net_device *dev)
+{
+	struct ipoib_dev_priv *priv = ipoib_priv(dev);
+	char name[IFNAMSIZ + sizeof("_path")];
+
+	snprintf(name, sizeof(name), "%s_mcg", dev->name);
+	priv->mcg_dentry = debugfs_create_file(name, S_IFREG | S_IRUGO,
+					       ipoib_root, dev, &ipoib_mcg_fops);
+
+	snprintf(name, sizeof(name), "%s_path", dev->name);
+	priv->path_dentry = debugfs_create_file(name, S_IFREG | S_IRUGO,
+						ipoib_root, dev, &ipoib_path_fops);
+}
+
+void ipoib_delete_debug_files(struct net_device *dev)
+{
+	struct ipoib_dev_priv *priv = ipoib_priv(dev);
+
+	debugfs_remove(priv->mcg_dentry);
+	debugfs_remove(priv->path_dentry);
+	priv->mcg_dentry = priv->path_dentry = NULL;
+}
+
+void ipoib_register_debugfs(void)
+{
+	ipoib_root = debugfs_create_dir("ipoib", NULL);
+}
+
+void ipoib_unregister_debugfs(void)
+{
+	debugfs_remove(ipoib_root);
+}
diff --git a/drivers/infiniband/ulp/ipoib/ipoib_ib.c b/drivers/infiniband/ulp/ipoib/ipoib_ib.c
new file mode 100644
index 000000000..ed25061fa
--- /dev/null
+++ b/drivers/infiniband/ulp/ipoib/ipoib_ib.c
@@ -0,0 +1,1311 @@
+/*
+ * Copyright (c) 2004, 2005 Topspin Communications.  All rights reserved.
+ * Copyright (c) 2005 Sun Microsystems, Inc. All rights reserved.
+ * Copyright (c) 2005 Mellanox Technologies. All rights reserved.
+ * Copyright (c) 2004, 2005 Voltaire, Inc. All rights reserved.
+ *
+ * This software is available to you under a choice of one of two
+ * licenses.  You may choose to be licensed under the terms of the GNU
+ * General Public License (GPL) Version 2, available from the file
+ * COPYING in the main directory of this source tree, or the
+ * OpenIB.org BSD license below:
+ *
+ *     Redistribution and use in source and binary forms, with or
+ *     without modification, are permitted provided that the following
+ *     conditions are met:
+ *
+ *      - Redistributions of source code must retain the above
+ *        copyright notice, this list of conditions and the following
+ *        disclaimer.
+ *
+ *      - Redistributions in binary form must reproduce the above
+ *        copyright notice, this list of conditions and the following
+ *        disclaimer in the documentation and/or other materials
+ *        provided with the distribution.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
+ * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
+ * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS
+ * BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN
+ * ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
+ * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+ * SOFTWARE.
+ */
+
+#include <linux/delay.h>
+#include <linux/moduleparam.h>
+#include <linux/dma-mapping.h>
+#include <linux/slab.h>
+
+#include <linux/ip.h>
+#include <linux/tcp.h>
+#include <rdma/ib_cache.h>
+
+#include "ipoib.h"
+
+#ifdef CONFIG_INFINIBAND_IPOIB_DEBUG_DATA
+static int data_debug_level;
+
+module_param(data_debug_level, int, 0644);
+MODULE_PARM_DESC(data_debug_level,
+		 "Enable data path debug tracing if > 0");
+#endif
+
+struct ipoib_ah *ipoib_create_ah(struct net_device *dev,
+				 struct ib_pd *pd, struct rdma_ah_attr *attr)
+{
+	struct ipoib_ah *ah;
+	struct ib_ah *vah;
+
+	ah = kmalloc(sizeof(*ah), GFP_KERNEL);
+	if (!ah)
+		return ERR_PTR(-ENOMEM);
+
+	ah->dev       = dev;
+	ah->last_send = 0;
+	kref_init(&ah->ref);
+
+	vah = rdma_create_ah(pd, attr, RDMA_CREATE_AH_SLEEPABLE);
+	if (IS_ERR(vah)) {
+		kfree(ah);
+		ah = (struct ipoib_ah *)vah;
+	} else {
+		ah->ah = vah;
+		ipoib_dbg(ipoib_priv(dev), "Created ah %p\n", ah->ah);
+	}
+
+	return ah;
+}
+
+void ipoib_free_ah(struct kref *kref)
+{
+	struct ipoib_ah *ah = container_of(kref, struct ipoib_ah, ref);
+	struct ipoib_dev_priv *priv = ipoib_priv(ah->dev);
+
+	unsigned long flags;
+
+	spin_lock_irqsave(&priv->lock, flags);
+	list_add_tail(&ah->list, &priv->dead_ahs);
+	spin_unlock_irqrestore(&priv->lock, flags);
+}
+
+static void ipoib_ud_dma_unmap_rx(struct ipoib_dev_priv *priv,
+				  u64 mapping[IPOIB_UD_RX_SG])
+{
+	ib_dma_unmap_single(priv->ca, mapping[0],
+			    IPOIB_UD_BUF_SIZE(priv->max_ib_mtu),
+			    DMA_FROM_DEVICE);
+}
+
+static int ipoib_ib_post_receive(struct net_device *dev, int id)
+{
+	struct ipoib_dev_priv *priv = ipoib_priv(dev);
+	int ret;
+
+	priv->rx_wr.wr_id   = id | IPOIB_OP_RECV;
+	priv->rx_sge[0].addr = priv->rx_ring[id].mapping[0];
+	priv->rx_sge[1].addr = priv->rx_ring[id].mapping[1];
+
+
+	ret = ib_post_recv(priv->qp, &priv->rx_wr, NULL);
+	if (unlikely(ret)) {
+		ipoib_warn(priv, "receive failed for buf %d (%d)\n", id, ret);
+		ipoib_ud_dma_unmap_rx(priv, priv->rx_ring[id].mapping);
+		dev_kfree_skb_any(priv->rx_ring[id].skb);
+		priv->rx_ring[id].skb = NULL;
+	}
+
+	return ret;
+}
+
+static struct sk_buff *ipoib_alloc_rx_skb(struct net_device *dev, int id)
+{
+	struct ipoib_dev_priv *priv = ipoib_priv(dev);
+	struct sk_buff *skb;
+	int buf_size;
+	u64 *mapping;
+
+	buf_size = IPOIB_UD_BUF_SIZE(priv->max_ib_mtu);
+
+	skb = dev_alloc_skb(buf_size + IPOIB_HARD_LEN);
+	if (unlikely(!skb))
+		return NULL;
+
+	/*
+	 * the IP header will be at IPOIP_HARD_LEN + IB_GRH_BYTES, that is
+	 * 64 bytes aligned
+	 */
+	skb_reserve(skb, sizeof(struct ipoib_pseudo_header));
+
+	mapping = priv->rx_ring[id].mapping;
+	mapping[0] = ib_dma_map_single(priv->ca, skb->data, buf_size,
+				       DMA_FROM_DEVICE);
+	if (unlikely(ib_dma_mapping_error(priv->ca, mapping[0])))
+		goto error;
+
+	priv->rx_ring[id].skb = skb;
+	return skb;
+error:
+	dev_kfree_skb_any(skb);
+	return NULL;
+}
+
+static int ipoib_ib_post_receives(struct net_device *dev)
+{
+	struct ipoib_dev_priv *priv = ipoib_priv(dev);
+	int i;
+
+	for (i = 0; i < ipoib_recvq_size; ++i) {
+		if (!ipoib_alloc_rx_skb(dev, i)) {
+			ipoib_warn(priv, "failed to allocate receive buffer %d\n", i);
+			return -ENOMEM;
+		}
+		if (ipoib_ib_post_receive(dev, i)) {
+			ipoib_warn(priv, "ipoib_ib_post_receive failed for buf %d\n", i);
+			return -EIO;
+		}
+	}
+
+	return 0;
+}
+
+static void ipoib_ib_handle_rx_wc(struct net_device *dev, struct ib_wc *wc)
+{
+	struct ipoib_dev_priv *priv = ipoib_priv(dev);
+	unsigned int wr_id = wc->wr_id & ~IPOIB_OP_RECV;
+	struct sk_buff *skb;
+	u64 mapping[IPOIB_UD_RX_SG];
+	union ib_gid *dgid;
+	union ib_gid *sgid;
+
+	ipoib_dbg_data(priv, "recv completion: id %d, status: %d\n",
+		       wr_id, wc->status);
+
+	if (unlikely(wr_id >= ipoib_recvq_size)) {
+		ipoib_warn(priv, "recv completion event with wrid %d (> %d)\n",
+			   wr_id, ipoib_recvq_size);
+		return;
+	}
+
+	skb  = priv->rx_ring[wr_id].skb;
+
+	if (unlikely(wc->status != IB_WC_SUCCESS)) {
+		if (wc->status != IB_WC_WR_FLUSH_ERR)
+			ipoib_warn(priv,
+				   "failed recv event (status=%d, wrid=%d vend_err %#x)\n",
+				   wc->status, wr_id, wc->vendor_err);
+		ipoib_ud_dma_unmap_rx(priv, priv->rx_ring[wr_id].mapping);
+		dev_kfree_skb_any(skb);
+		priv->rx_ring[wr_id].skb = NULL;
+		return;
+	}
+
+	memcpy(mapping, priv->rx_ring[wr_id].mapping,
+	       IPOIB_UD_RX_SG * sizeof(*mapping));
+
+	/*
+	 * If we can't allocate a new RX buffer, dump
+	 * this packet and reuse the old buffer.
+	 */
+	if (unlikely(!ipoib_alloc_rx_skb(dev, wr_id))) {
+		++dev->stats.rx_dropped;
+		goto repost;
+	}
+
+	ipoib_dbg_data(priv, "received %d bytes, SLID 0x%04x\n",
+		       wc->byte_len, wc->slid);
+
+	ipoib_ud_dma_unmap_rx(priv, mapping);
+
+	skb_put(skb, wc->byte_len);
+
+	/* First byte of dgid signals multicast when 0xff */
+	dgid = &((struct ib_grh *)skb->data)->dgid;
+
+	if (!(wc->wc_flags & IB_WC_GRH) || dgid->raw[0] != 0xff)
+		skb->pkt_type = PACKET_HOST;
+	else if (memcmp(dgid, dev->broadcast + 4, sizeof(union ib_gid)) == 0)
+		skb->pkt_type = PACKET_BROADCAST;
+	else
+		skb->pkt_type = PACKET_MULTICAST;
+
+	sgid = &((struct ib_grh *)skb->data)->sgid;
+
+	/*
+	 * Drop packets that this interface sent, ie multicast packets
+	 * that the HCA has replicated.
+	 */
+	if (wc->slid == priv->local_lid && wc->src_qp == priv->qp->qp_num) {
+		int need_repost = 1;
+
+		if ((wc->wc_flags & IB_WC_GRH) &&
+		    sgid->global.interface_id != priv->local_gid.global.interface_id)
+			need_repost = 0;
+
+		if (need_repost) {
+			dev_kfree_skb_any(skb);
+			goto repost;
+		}
+	}
+
+	skb_pull(skb, IB_GRH_BYTES);
+
+	skb->protocol = ((struct ipoib_header *) skb->data)->proto;
+	skb_add_pseudo_hdr(skb);
+
+	++dev->stats.rx_packets;
+	dev->stats.rx_bytes += skb->len;
+	if (skb->pkt_type == PACKET_MULTICAST)
+		dev->stats.multicast++;
+
+	skb->dev = dev;
+	if ((dev->features & NETIF_F_RXCSUM) &&
+			likely(wc->wc_flags & IB_WC_IP_CSUM_OK))
+		skb->ip_summed = CHECKSUM_UNNECESSARY;
+
+	napi_gro_receive(&priv->recv_napi, skb);
+
+repost:
+	if (unlikely(ipoib_ib_post_receive(dev, wr_id)))
+		ipoib_warn(priv, "ipoib_ib_post_receive failed "
+			   "for buf %d\n", wr_id);
+}
+
+int ipoib_dma_map_tx(struct ib_device *ca, struct ipoib_tx_buf *tx_req)
+{
+	struct sk_buff *skb = tx_req->skb;
+	u64 *mapping = tx_req->mapping;
+	int i;
+	int off;
+
+	if (skb_headlen(skb)) {
+		mapping[0] = ib_dma_map_single(ca, skb->data, skb_headlen(skb),
+					       DMA_TO_DEVICE);
+		if (unlikely(ib_dma_mapping_error(ca, mapping[0])))
+			return -EIO;
+
+		off = 1;
+	} else
+		off = 0;
+
+	for (i = 0; i < skb_shinfo(skb)->nr_frags; ++i) {
+		const skb_frag_t *frag = &skb_shinfo(skb)->frags[i];
+		mapping[i + off] = ib_dma_map_page(ca,
+						 skb_frag_page(frag),
+						 skb_frag_off(frag),
+						 skb_frag_size(frag),
+						 DMA_TO_DEVICE);
+		if (unlikely(ib_dma_mapping_error(ca, mapping[i + off])))
+			goto partial_error;
+	}
+	return 0;
+
+partial_error:
+	for (; i > 0; --i) {
+		const skb_frag_t *frag = &skb_shinfo(skb)->frags[i - 1];
+
+		ib_dma_unmap_page(ca, mapping[i - !off], skb_frag_size(frag), DMA_TO_DEVICE);
+	}
+
+	if (off)
+		ib_dma_unmap_single(ca, mapping[0], skb_headlen(skb), DMA_TO_DEVICE);
+
+	return -EIO;
+}
+
+void ipoib_dma_unmap_tx(struct ipoib_dev_priv *priv,
+			struct ipoib_tx_buf *tx_req)
+{
+	struct sk_buff *skb = tx_req->skb;
+	u64 *mapping = tx_req->mapping;
+	int i;
+	int off;
+
+	if (skb_headlen(skb)) {
+		ib_dma_unmap_single(priv->ca, mapping[0], skb_headlen(skb),
+				    DMA_TO_DEVICE);
+		off = 1;
+	} else
+		off = 0;
+
+	for (i = 0; i < skb_shinfo(skb)->nr_frags; ++i) {
+		const skb_frag_t *frag = &skb_shinfo(skb)->frags[i];
+
+		ib_dma_unmap_page(priv->ca, mapping[i + off],
+				  skb_frag_size(frag), DMA_TO_DEVICE);
+	}
+}
+
+/*
+ * As the result of a completion error the QP Can be transferred to SQE states.
+ * The function checks if the (send)QP is in SQE state and
+ * moves it back to RTS state, that in order to have it functional again.
+ */
+static void ipoib_qp_state_validate_work(struct work_struct *work)
+{
+	struct ipoib_qp_state_validate *qp_work =
+		container_of(work, struct ipoib_qp_state_validate, work);
+
+	struct ipoib_dev_priv *priv = qp_work->priv;
+	struct ib_qp_attr qp_attr;
+	struct ib_qp_init_attr query_init_attr;
+	int ret;
+
+	ret = ib_query_qp(priv->qp, &qp_attr, IB_QP_STATE, &query_init_attr);
+	if (ret) {
+		ipoib_warn(priv, "%s: Failed to query QP ret: %d\n",
+			   __func__, ret);
+		goto free_res;
+	}
+	pr_info("%s: QP: 0x%x is in state: %d\n",
+		__func__, priv->qp->qp_num, qp_attr.qp_state);
+
+	/* currently support only in SQE->RTS transition*/
+	if (qp_attr.qp_state == IB_QPS_SQE) {
+		qp_attr.qp_state = IB_QPS_RTS;
+
+		ret = ib_modify_qp(priv->qp, &qp_attr, IB_QP_STATE);
+		if (ret) {
+			pr_warn("failed(%d) modify QP:0x%x SQE->RTS\n",
+				ret, priv->qp->qp_num);
+			goto free_res;
+		}
+		pr_info("%s: QP: 0x%x moved from IB_QPS_SQE to IB_QPS_RTS\n",
+			__func__, priv->qp->qp_num);
+	} else {
+		pr_warn("QP (%d) will stay in state: %d\n",
+			priv->qp->qp_num, qp_attr.qp_state);
+	}
+
+free_res:
+	kfree(qp_work);
+}
+
+static void ipoib_ib_handle_tx_wc(struct net_device *dev, struct ib_wc *wc)
+{
+	struct ipoib_dev_priv *priv = ipoib_priv(dev);
+	unsigned int wr_id = wc->wr_id;
+	struct ipoib_tx_buf *tx_req;
+
+	ipoib_dbg_data(priv, "send completion: id %d, status: %d\n",
+		       wr_id, wc->status);
+
+	if (unlikely(wr_id >= ipoib_sendq_size)) {
+		ipoib_warn(priv, "send completion event with wrid %d (> %d)\n",
+			   wr_id, ipoib_sendq_size);
+		return;
+	}
+
+	tx_req = &priv->tx_ring[wr_id];
+
+	ipoib_dma_unmap_tx(priv, tx_req);
+
+	++dev->stats.tx_packets;
+	dev->stats.tx_bytes += tx_req->skb->len;
+
+	dev_kfree_skb_any(tx_req->skb);
+
+	++priv->tx_tail;
+	++priv->global_tx_tail;
+
+	if (unlikely(netif_queue_stopped(dev) &&
+		     ((priv->global_tx_head - priv->global_tx_tail) <=
+		      ipoib_sendq_size >> 1) &&
+		     test_bit(IPOIB_FLAG_ADMIN_UP, &priv->flags)))
+		netif_wake_queue(dev);
+
+	if (wc->status != IB_WC_SUCCESS &&
+	    wc->status != IB_WC_WR_FLUSH_ERR) {
+		struct ipoib_qp_state_validate *qp_work;
+		ipoib_warn(priv,
+			   "failed send event (status=%d, wrid=%d vend_err %#x)\n",
+			   wc->status, wr_id, wc->vendor_err);
+		qp_work = kzalloc(sizeof(*qp_work), GFP_ATOMIC);
+		if (!qp_work)
+			return;
+
+		INIT_WORK(&qp_work->work, ipoib_qp_state_validate_work);
+		qp_work->priv = priv;
+		queue_work(priv->wq, &qp_work->work);
+	}
+}
+
+static int poll_tx(struct ipoib_dev_priv *priv)
+{
+	int n, i;
+	struct ib_wc *wc;
+
+	n = ib_poll_cq(priv->send_cq, MAX_SEND_CQE, priv->send_wc);
+	for (i = 0; i < n; ++i) {
+		wc = priv->send_wc + i;
+		if (wc->wr_id & IPOIB_OP_CM)
+			ipoib_cm_handle_tx_wc(priv->dev, priv->send_wc + i);
+		else
+			ipoib_ib_handle_tx_wc(priv->dev, priv->send_wc + i);
+	}
+	return n == MAX_SEND_CQE;
+}
+
+int ipoib_rx_poll(struct napi_struct *napi, int budget)
+{
+	struct ipoib_dev_priv *priv =
+		container_of(napi, struct ipoib_dev_priv, recv_napi);
+	struct net_device *dev = priv->dev;
+	int done;
+	int t;
+	int n, i;
+
+	done  = 0;
+
+poll_more:
+	while (done < budget) {
+		int max = (budget - done);
+
+		t = min(IPOIB_NUM_WC, max);
+		n = ib_poll_cq(priv->recv_cq, t, priv->ibwc);
+
+		for (i = 0; i < n; i++) {
+			struct ib_wc *wc = priv->ibwc + i;
+
+			if (wc->wr_id & IPOIB_OP_RECV) {
+				++done;
+				if (wc->wr_id & IPOIB_OP_CM)
+					ipoib_cm_handle_rx_wc(dev, wc);
+				else
+					ipoib_ib_handle_rx_wc(dev, wc);
+			} else {
+				pr_warn("%s: Got unexpected wqe id\n", __func__);
+			}
+		}
+
+		if (n != t)
+			break;
+	}
+
+	if (done < budget) {
+		napi_complete(napi);
+		if (unlikely(ib_req_notify_cq(priv->recv_cq,
+					      IB_CQ_NEXT_COMP |
+					      IB_CQ_REPORT_MISSED_EVENTS)) &&
+		    napi_reschedule(napi))
+			goto poll_more;
+	}
+
+	return done;
+}
+
+int ipoib_tx_poll(struct napi_struct *napi, int budget)
+{
+	struct ipoib_dev_priv *priv = container_of(napi, struct ipoib_dev_priv,
+						   send_napi);
+	struct net_device *dev = priv->dev;
+	int n, i;
+	struct ib_wc *wc;
+
+poll_more:
+	n = ib_poll_cq(priv->send_cq, MAX_SEND_CQE, priv->send_wc);
+
+	for (i = 0; i < n; i++) {
+		wc = priv->send_wc + i;
+		if (wc->wr_id & IPOIB_OP_CM)
+			ipoib_cm_handle_tx_wc(dev, wc);
+		else
+			ipoib_ib_handle_tx_wc(dev, wc);
+	}
+
+	if (n < budget) {
+		napi_complete(napi);
+		if (unlikely(ib_req_notify_cq(priv->send_cq, IB_CQ_NEXT_COMP |
+					      IB_CQ_REPORT_MISSED_EVENTS)) &&
+		    napi_reschedule(napi))
+			goto poll_more;
+	}
+	return n < 0 ? 0 : n;
+}
+
+void ipoib_ib_rx_completion(struct ib_cq *cq, void *ctx_ptr)
+{
+	struct ipoib_dev_priv *priv = ctx_ptr;
+
+	napi_schedule(&priv->recv_napi);
+}
+
+void ipoib_ib_tx_completion(struct ib_cq *cq, void *ctx_ptr)
+{
+	struct ipoib_dev_priv *priv = ctx_ptr;
+
+	napi_schedule(&priv->send_napi);
+}
+
+static inline int post_send(struct ipoib_dev_priv *priv,
+			    unsigned int wr_id,
+			    struct ib_ah *address, u32 dqpn,
+			    struct ipoib_tx_buf *tx_req,
+			    void *head, int hlen)
+{
+	struct sk_buff *skb = tx_req->skb;
+
+	ipoib_build_sge(priv, tx_req);
+
+	priv->tx_wr.wr.wr_id	= wr_id;
+	priv->tx_wr.remote_qpn	= dqpn;
+	priv->tx_wr.ah		= address;
+
+	if (head) {
+		priv->tx_wr.mss		= skb_shinfo(skb)->gso_size;
+		priv->tx_wr.header	= head;
+		priv->tx_wr.hlen	= hlen;
+		priv->tx_wr.wr.opcode	= IB_WR_LSO;
+	} else
+		priv->tx_wr.wr.opcode	= IB_WR_SEND;
+
+	return ib_post_send(priv->qp, &priv->tx_wr.wr, NULL);
+}
+
+int ipoib_send(struct net_device *dev, struct sk_buff *skb,
+	       struct ib_ah *address, u32 dqpn)
+{
+	struct ipoib_dev_priv *priv = ipoib_priv(dev);
+	struct ipoib_tx_buf *tx_req;
+	int hlen, rc;
+	void *phead;
+	unsigned int usable_sge = priv->max_send_sge - !!skb_headlen(skb);
+
+	if (skb_is_gso(skb)) {
+		hlen = skb_tcp_all_headers(skb);
+		phead = skb->data;
+		if (unlikely(!skb_pull(skb, hlen))) {
+			ipoib_warn(priv, "linear data too small\n");
+			++dev->stats.tx_dropped;
+			++dev->stats.tx_errors;
+			dev_kfree_skb_any(skb);
+			return -1;
+		}
+	} else {
+		if (unlikely(skb->len > priv->mcast_mtu + IPOIB_ENCAP_LEN)) {
+			ipoib_warn(priv, "packet len %d (> %d) too long to send, dropping\n",
+				   skb->len, priv->mcast_mtu + IPOIB_ENCAP_LEN);
+			++dev->stats.tx_dropped;
+			++dev->stats.tx_errors;
+			ipoib_cm_skb_too_long(dev, skb, priv->mcast_mtu);
+			return -1;
+		}
+		phead = NULL;
+		hlen  = 0;
+	}
+	if (skb_shinfo(skb)->nr_frags > usable_sge) {
+		if (skb_linearize(skb) < 0) {
+			ipoib_warn(priv, "skb could not be linearized\n");
+			++dev->stats.tx_dropped;
+			++dev->stats.tx_errors;
+			dev_kfree_skb_any(skb);
+			return -1;
+		}
+		/* Does skb_linearize return ok without reducing nr_frags? */
+		if (skb_shinfo(skb)->nr_frags > usable_sge) {
+			ipoib_warn(priv, "too many frags after skb linearize\n");
+			++dev->stats.tx_dropped;
+			++dev->stats.tx_errors;
+			dev_kfree_skb_any(skb);
+			return -1;
+		}
+	}
+
+	ipoib_dbg_data(priv,
+		       "sending packet, length=%d address=%p dqpn=0x%06x\n",
+		       skb->len, address, dqpn);
+
+	/*
+	 * We put the skb into the tx_ring _before_ we call post_send()
+	 * because it's entirely possible that the completion handler will
+	 * run before we execute anything after the post_send().  That
+	 * means we have to make sure everything is properly recorded and
+	 * our state is consistent before we call post_send().
+	 */
+	tx_req = &priv->tx_ring[priv->tx_head & (ipoib_sendq_size - 1)];
+	tx_req->skb = skb;
+	if (unlikely(ipoib_dma_map_tx(priv->ca, tx_req))) {
+		++dev->stats.tx_errors;
+		dev_kfree_skb_any(skb);
+		return -1;
+	}
+
+	if (skb->ip_summed == CHECKSUM_PARTIAL)
+		priv->tx_wr.wr.send_flags |= IB_SEND_IP_CSUM;
+	else
+		priv->tx_wr.wr.send_flags &= ~IB_SEND_IP_CSUM;
+	/* increase the tx_head after send success, but use it for queue state */
+	if ((priv->global_tx_head - priv->global_tx_tail) ==
+	    ipoib_sendq_size - 1) {
+		ipoib_dbg(priv, "TX ring full, stopping kernel net queue\n");
+		netif_stop_queue(dev);
+	}
+
+	skb_orphan(skb);
+	skb_dst_drop(skb);
+
+	if (netif_queue_stopped(dev))
+		if (ib_req_notify_cq(priv->send_cq, IB_CQ_NEXT_COMP |
+				     IB_CQ_REPORT_MISSED_EVENTS) < 0)
+			ipoib_warn(priv, "request notify on send CQ failed\n");
+
+	rc = post_send(priv, priv->tx_head & (ipoib_sendq_size - 1),
+		       address, dqpn, tx_req, phead, hlen);
+	if (unlikely(rc)) {
+		ipoib_warn(priv, "post_send failed, error %d\n", rc);
+		++dev->stats.tx_errors;
+		ipoib_dma_unmap_tx(priv, tx_req);
+		dev_kfree_skb_any(skb);
+		if (netif_queue_stopped(dev))
+			netif_wake_queue(dev);
+		rc = 0;
+	} else {
+		netif_trans_update(dev);
+
+		rc = priv->tx_head;
+		++priv->tx_head;
+		++priv->global_tx_head;
+	}
+	return rc;
+}
+
+static void ipoib_reap_dead_ahs(struct ipoib_dev_priv *priv)
+{
+	struct ipoib_ah *ah, *tah;
+	unsigned long flags;
+
+	netif_tx_lock_bh(priv->dev);
+	spin_lock_irqsave(&priv->lock, flags);
+
+	list_for_each_entry_safe(ah, tah, &priv->dead_ahs, list)
+		if ((int) priv->tx_tail - (int) ah->last_send >= 0) {
+			list_del(&ah->list);
+			rdma_destroy_ah(ah->ah, 0);
+			kfree(ah);
+		}
+
+	spin_unlock_irqrestore(&priv->lock, flags);
+	netif_tx_unlock_bh(priv->dev);
+}
+
+void ipoib_reap_ah(struct work_struct *work)
+{
+	struct ipoib_dev_priv *priv =
+		container_of(work, struct ipoib_dev_priv, ah_reap_task.work);
+
+	ipoib_reap_dead_ahs(priv);
+
+	if (!test_bit(IPOIB_STOP_REAPER, &priv->flags))
+		queue_delayed_work(priv->wq, &priv->ah_reap_task,
+				   round_jiffies_relative(HZ));
+}
+
+static void ipoib_start_ah_reaper(struct ipoib_dev_priv *priv)
+{
+	clear_bit(IPOIB_STOP_REAPER, &priv->flags);
+	queue_delayed_work(priv->wq, &priv->ah_reap_task,
+			   round_jiffies_relative(HZ));
+}
+
+static void ipoib_stop_ah_reaper(struct ipoib_dev_priv *priv)
+{
+	set_bit(IPOIB_STOP_REAPER, &priv->flags);
+	cancel_delayed_work(&priv->ah_reap_task);
+	/*
+	 * After ipoib_stop_ah_reaper() we always go through
+	 * ipoib_reap_dead_ahs() which ensures the work is really stopped and
+	 * does a final flush out of the dead_ah's list
+	 */
+}
+
+static int recvs_pending(struct net_device *dev)
+{
+	struct ipoib_dev_priv *priv = ipoib_priv(dev);
+	int pending = 0;
+	int i;
+
+	for (i = 0; i < ipoib_recvq_size; ++i)
+		if (priv->rx_ring[i].skb)
+			++pending;
+
+	return pending;
+}
+
+static void check_qp_movement_and_print(struct ipoib_dev_priv *priv,
+					struct ib_qp *qp,
+					enum ib_qp_state new_state)
+{
+	struct ib_qp_attr qp_attr;
+	struct ib_qp_init_attr query_init_attr;
+	int ret;
+
+	ret = ib_query_qp(qp, &qp_attr, IB_QP_STATE, &query_init_attr);
+	if (ret) {
+		ipoib_warn(priv, "%s: Failed to query QP\n", __func__);
+		return;
+	}
+	/* print according to the new-state and the previous state.*/
+	if (new_state == IB_QPS_ERR && qp_attr.qp_state == IB_QPS_RESET)
+		ipoib_dbg(priv, "Failed modify QP, IB_QPS_RESET to IB_QPS_ERR, acceptable\n");
+	else
+		ipoib_warn(priv, "Failed to modify QP to state: %d from state: %d\n",
+			   new_state, qp_attr.qp_state);
+}
+
+static void ipoib_napi_enable(struct net_device *dev)
+{
+	struct ipoib_dev_priv *priv = ipoib_priv(dev);
+
+	napi_enable(&priv->recv_napi);
+	napi_enable(&priv->send_napi);
+}
+
+static void ipoib_napi_disable(struct net_device *dev)
+{
+	struct ipoib_dev_priv *priv = ipoib_priv(dev);
+
+	napi_disable(&priv->recv_napi);
+	napi_disable(&priv->send_napi);
+}
+
+int ipoib_ib_dev_stop_default(struct net_device *dev)
+{
+	struct ipoib_dev_priv *priv = ipoib_priv(dev);
+	struct ib_qp_attr qp_attr;
+	unsigned long begin;
+	struct ipoib_tx_buf *tx_req;
+	int i;
+
+	if (test_bit(IPOIB_FLAG_INITIALIZED, &priv->flags))
+		ipoib_napi_disable(dev);
+
+	ipoib_cm_dev_stop(dev);
+
+	/*
+	 * Move our QP to the error state and then reinitialize in
+	 * when all work requests have completed or have been flushed.
+	 */
+	qp_attr.qp_state = IB_QPS_ERR;
+	if (ib_modify_qp(priv->qp, &qp_attr, IB_QP_STATE))
+		check_qp_movement_and_print(priv, priv->qp, IB_QPS_ERR);
+
+	/* Wait for all sends and receives to complete */
+	begin = jiffies;
+
+	while (priv->tx_head != priv->tx_tail || recvs_pending(dev)) {
+		if (time_after(jiffies, begin + 5 * HZ)) {
+			ipoib_warn(priv,
+				   "timing out; %d sends %d receives not completed\n",
+				   priv->tx_head - priv->tx_tail,
+				   recvs_pending(dev));
+
+			/*
+			 * assume the HW is wedged and just free up
+			 * all our pending work requests.
+			 */
+			while ((int)priv->tx_tail - (int)priv->tx_head < 0) {
+				tx_req = &priv->tx_ring[priv->tx_tail &
+							(ipoib_sendq_size - 1)];
+				ipoib_dma_unmap_tx(priv, tx_req);
+				dev_kfree_skb_any(tx_req->skb);
+				++priv->tx_tail;
+				++priv->global_tx_tail;
+			}
+
+			for (i = 0; i < ipoib_recvq_size; ++i) {
+				struct ipoib_rx_buf *rx_req;
+
+				rx_req = &priv->rx_ring[i];
+				if (!rx_req->skb)
+					continue;
+				ipoib_ud_dma_unmap_rx(priv,
+						      priv->rx_ring[i].mapping);
+				dev_kfree_skb_any(rx_req->skb);
+				rx_req->skb = NULL;
+			}
+
+			goto timeout;
+		}
+
+		ipoib_drain_cq(dev);
+
+		usleep_range(1000, 2000);
+	}
+
+	ipoib_dbg(priv, "All sends and receives done.\n");
+
+timeout:
+	qp_attr.qp_state = IB_QPS_RESET;
+	if (ib_modify_qp(priv->qp, &qp_attr, IB_QP_STATE))
+		ipoib_warn(priv, "Failed to modify QP to RESET state\n");
+
+	ib_req_notify_cq(priv->recv_cq, IB_CQ_NEXT_COMP);
+
+	return 0;
+}
+
+int ipoib_ib_dev_open_default(struct net_device *dev)
+{
+	struct ipoib_dev_priv *priv = ipoib_priv(dev);
+	int ret;
+
+	ret = ipoib_init_qp(dev);
+	if (ret) {
+		ipoib_warn(priv, "ipoib_init_qp returned %d\n", ret);
+		return -1;
+	}
+
+	ret = ipoib_ib_post_receives(dev);
+	if (ret) {
+		ipoib_warn(priv, "ipoib_ib_post_receives returned %d\n", ret);
+		goto out;
+	}
+
+	ret = ipoib_cm_dev_open(dev);
+	if (ret) {
+		ipoib_warn(priv, "ipoib_cm_dev_open returned %d\n", ret);
+		goto out;
+	}
+
+	if (!test_bit(IPOIB_FLAG_INITIALIZED, &priv->flags))
+		ipoib_napi_enable(dev);
+
+	return 0;
+out:
+	return -1;
+}
+
+int ipoib_ib_dev_open(struct net_device *dev)
+{
+	struct ipoib_dev_priv *priv = ipoib_priv(dev);
+
+	ipoib_pkey_dev_check_presence(dev);
+
+	if (!test_bit(IPOIB_PKEY_ASSIGNED, &priv->flags)) {
+		ipoib_warn(priv, "P_Key 0x%04x is %s\n", priv->pkey,
+			   (!(priv->pkey & 0x7fff) ? "Invalid" : "not found"));
+		return -1;
+	}
+
+	ipoib_start_ah_reaper(priv);
+	if (priv->rn_ops->ndo_open(dev)) {
+		pr_warn("%s: Failed to open dev\n", dev->name);
+		goto dev_stop;
+	}
+
+	set_bit(IPOIB_FLAG_INITIALIZED, &priv->flags);
+
+	return 0;
+
+dev_stop:
+	ipoib_stop_ah_reaper(priv);
+	return -1;
+}
+
+void ipoib_ib_dev_stop(struct net_device *dev)
+{
+	struct ipoib_dev_priv *priv = ipoib_priv(dev);
+
+	priv->rn_ops->ndo_stop(dev);
+
+	clear_bit(IPOIB_FLAG_INITIALIZED, &priv->flags);
+	ipoib_stop_ah_reaper(priv);
+}
+
+void ipoib_pkey_dev_check_presence(struct net_device *dev)
+{
+	struct ipoib_dev_priv *priv = ipoib_priv(dev);
+	struct rdma_netdev *rn = netdev_priv(dev);
+
+	if (!(priv->pkey & 0x7fff) ||
+	    ib_find_pkey(priv->ca, priv->port, priv->pkey,
+			 &priv->pkey_index)) {
+		clear_bit(IPOIB_PKEY_ASSIGNED, &priv->flags);
+	} else {
+		if (rn->set_id)
+			rn->set_id(dev, priv->pkey_index);
+		set_bit(IPOIB_PKEY_ASSIGNED, &priv->flags);
+	}
+}
+
+void ipoib_ib_dev_up(struct net_device *dev)
+{
+	struct ipoib_dev_priv *priv = ipoib_priv(dev);
+
+	ipoib_pkey_dev_check_presence(dev);
+
+	if (!test_bit(IPOIB_PKEY_ASSIGNED, &priv->flags)) {
+		ipoib_dbg(priv, "PKEY is not assigned.\n");
+		return;
+	}
+
+	set_bit(IPOIB_FLAG_OPER_UP, &priv->flags);
+
+	ipoib_mcast_start_thread(dev);
+}
+
+void ipoib_ib_dev_down(struct net_device *dev)
+{
+	struct ipoib_dev_priv *priv = ipoib_priv(dev);
+
+	ipoib_dbg(priv, "downing ib_dev\n");
+
+	clear_bit(IPOIB_FLAG_OPER_UP, &priv->flags);
+	netif_carrier_off(dev);
+
+	ipoib_mcast_stop_thread(dev);
+	ipoib_mcast_dev_flush(dev);
+
+	ipoib_flush_paths(dev);
+}
+
+void ipoib_drain_cq(struct net_device *dev)
+{
+	struct ipoib_dev_priv *priv = ipoib_priv(dev);
+	int i, n;
+
+	/*
+	 * We call completion handling routines that expect to be
+	 * called from the BH-disabled NAPI poll context, so disable
+	 * BHs here too.
+	 */
+	local_bh_disable();
+
+	do {
+		n = ib_poll_cq(priv->recv_cq, IPOIB_NUM_WC, priv->ibwc);
+		for (i = 0; i < n; ++i) {
+			/*
+			 * Convert any successful completions to flush
+			 * errors to avoid passing packets up the
+			 * stack after bringing the device down.
+			 */
+			if (priv->ibwc[i].status == IB_WC_SUCCESS)
+				priv->ibwc[i].status = IB_WC_WR_FLUSH_ERR;
+
+			if (priv->ibwc[i].wr_id & IPOIB_OP_RECV) {
+				if (priv->ibwc[i].wr_id & IPOIB_OP_CM)
+					ipoib_cm_handle_rx_wc(dev, priv->ibwc + i);
+				else
+					ipoib_ib_handle_rx_wc(dev, priv->ibwc + i);
+			} else {
+				pr_warn("%s: Got unexpected wqe id\n", __func__);
+			}
+		}
+	} while (n == IPOIB_NUM_WC);
+
+	while (poll_tx(priv))
+		; /* nothing */
+
+	local_bh_enable();
+}
+
+/*
+ * Takes whatever value which is in pkey index 0 and updates priv->pkey
+ * returns 0 if the pkey value was changed.
+ */
+static inline int update_parent_pkey(struct ipoib_dev_priv *priv)
+{
+	int result;
+	u16 prev_pkey;
+
+	prev_pkey = priv->pkey;
+	result = ib_query_pkey(priv->ca, priv->port, 0, &priv->pkey);
+	if (result) {
+		ipoib_warn(priv, "ib_query_pkey port %d failed (ret = %d)\n",
+			   priv->port, result);
+		return result;
+	}
+
+	priv->pkey |= 0x8000;
+
+	if (prev_pkey != priv->pkey) {
+		ipoib_dbg(priv, "pkey changed from 0x%x to 0x%x\n",
+			  prev_pkey, priv->pkey);
+		/*
+		 * Update the pkey in the broadcast address, while making sure to set
+		 * the full membership bit, so that we join the right broadcast group.
+		 */
+		priv->dev->broadcast[8] = priv->pkey >> 8;
+		priv->dev->broadcast[9] = priv->pkey & 0xff;
+		return 0;
+	}
+
+	return 1;
+}
+/*
+ * returns 0 if pkey value was found in a different slot.
+ */
+static inline int update_child_pkey(struct ipoib_dev_priv *priv)
+{
+	u16 old_index = priv->pkey_index;
+
+	priv->pkey_index = 0;
+	ipoib_pkey_dev_check_presence(priv->dev);
+
+	if (test_bit(IPOIB_PKEY_ASSIGNED, &priv->flags) &&
+	    (old_index == priv->pkey_index))
+		return 1;
+	return 0;
+}
+
+/*
+ * returns true if the device address of the ipoib interface has changed and the
+ * new address is a valid one (i.e in the gid table), return false otherwise.
+ */
+static bool ipoib_dev_addr_changed_valid(struct ipoib_dev_priv *priv)
+{
+	union ib_gid search_gid;
+	union ib_gid gid0;
+	int err;
+	u16 index;
+	u32 port;
+	bool ret = false;
+
+	if (rdma_query_gid(priv->ca, priv->port, 0, &gid0))
+		return false;
+
+	netif_addr_lock_bh(priv->dev);
+
+	/* The subnet prefix may have changed, update it now so we won't have
+	 * to do it later
+	 */
+	priv->local_gid.global.subnet_prefix = gid0.global.subnet_prefix;
+	dev_addr_mod(priv->dev, 4, (u8 *)&gid0.global.subnet_prefix,
+		     sizeof(gid0.global.subnet_prefix));
+	search_gid.global.subnet_prefix = gid0.global.subnet_prefix;
+
+	search_gid.global.interface_id = priv->local_gid.global.interface_id;
+
+	netif_addr_unlock_bh(priv->dev);
+
+	err = ib_find_gid(priv->ca, &search_gid, &port, &index);
+
+	netif_addr_lock_bh(priv->dev);
+
+	if (search_gid.global.interface_id !=
+	    priv->local_gid.global.interface_id)
+		/* There was a change while we were looking up the gid, bail
+		 * here and let the next work sort this out
+		 */
+		goto out;
+
+	/* The next section of code needs some background:
+	 * Per IB spec the port GUID can't change if the HCA is powered on.
+	 * port GUID is the basis for GID at index 0 which is the basis for
+	 * the default device address of a ipoib interface.
+	 *
+	 * so it seems the flow should be:
+	 * if user_changed_dev_addr && gid in gid tbl
+	 *	set bit dev_addr_set
+	 *	return true
+	 * else
+	 *	return false
+	 *
+	 * The issue is that there are devices that don't follow the spec,
+	 * they change the port GUID when the HCA is powered, so in order
+	 * not to break userspace applications, We need to check if the
+	 * user wanted to control the device address and we assume that
+	 * if he sets the device address back to be based on GID index 0,
+	 * he no longer wishs to control it.
+	 *
+	 * If the user doesn't control the device address,
+	 * IPOIB_FLAG_DEV_ADDR_SET is set and ib_find_gid failed it means
+	 * the port GUID has changed and GID at index 0 has changed
+	 * so we need to change priv->local_gid and priv->dev->dev_addr
+	 * to reflect the new GID.
+	 */
+	if (!test_bit(IPOIB_FLAG_DEV_ADDR_SET, &priv->flags)) {
+		if (!err && port == priv->port) {
+			set_bit(IPOIB_FLAG_DEV_ADDR_SET, &priv->flags);
+			if (index == 0)
+				clear_bit(IPOIB_FLAG_DEV_ADDR_CTRL,
+					  &priv->flags);
+			else
+				set_bit(IPOIB_FLAG_DEV_ADDR_CTRL, &priv->flags);
+			ret = true;
+		} else {
+			ret = false;
+		}
+	} else {
+		if (!err && port == priv->port) {
+			ret = true;
+		} else {
+			if (!test_bit(IPOIB_FLAG_DEV_ADDR_CTRL, &priv->flags)) {
+				memcpy(&priv->local_gid, &gid0,
+				       sizeof(priv->local_gid));
+				dev_addr_mod(priv->dev, 4, (u8 *)&gid0,
+					     sizeof(priv->local_gid));
+				ret = true;
+			}
+		}
+	}
+
+out:
+	netif_addr_unlock_bh(priv->dev);
+
+	return ret;
+}
+
+static void __ipoib_ib_dev_flush(struct ipoib_dev_priv *priv,
+				enum ipoib_flush_level level,
+				int nesting)
+{
+	struct ipoib_dev_priv *cpriv;
+	struct net_device *dev = priv->dev;
+	int result;
+
+	down_read_nested(&priv->vlan_rwsem, nesting);
+
+	/*
+	 * Flush any child interfaces too -- they might be up even if
+	 * the parent is down.
+	 */
+	list_for_each_entry(cpriv, &priv->child_intfs, list)
+		__ipoib_ib_dev_flush(cpriv, level, nesting + 1);
+
+	up_read(&priv->vlan_rwsem);
+
+	if (!test_bit(IPOIB_FLAG_INITIALIZED, &priv->flags) &&
+	    level != IPOIB_FLUSH_HEAVY) {
+		/* Make sure the dev_addr is set even if not flushing */
+		if (level == IPOIB_FLUSH_LIGHT)
+			ipoib_dev_addr_changed_valid(priv);
+		ipoib_dbg(priv, "Not flushing - IPOIB_FLAG_INITIALIZED not set.\n");
+		return;
+	}
+
+	if (!test_bit(IPOIB_FLAG_ADMIN_UP, &priv->flags)) {
+		/* interface is down. update pkey and leave. */
+		if (level == IPOIB_FLUSH_HEAVY) {
+			if (!test_bit(IPOIB_FLAG_SUBINTERFACE, &priv->flags))
+				update_parent_pkey(priv);
+			else
+				update_child_pkey(priv);
+		} else if (level == IPOIB_FLUSH_LIGHT)
+			ipoib_dev_addr_changed_valid(priv);
+		ipoib_dbg(priv, "Not flushing - IPOIB_FLAG_ADMIN_UP not set.\n");
+		return;
+	}
+
+	if (level == IPOIB_FLUSH_HEAVY) {
+		/* child devices chase their origin pkey value, while non-child
+		 * (parent) devices should always takes what present in pkey index 0
+		 */
+		if (test_bit(IPOIB_FLAG_SUBINTERFACE, &priv->flags)) {
+			result = update_child_pkey(priv);
+			if (result) {
+				/* restart QP only if P_Key index is changed */
+				ipoib_dbg(priv, "Not flushing - P_Key index not changed.\n");
+				return;
+			}
+
+		} else {
+			result = update_parent_pkey(priv);
+			/* restart QP only if P_Key value changed */
+			if (result) {
+				ipoib_dbg(priv, "Not flushing - P_Key value not changed.\n");
+				return;
+			}
+		}
+	}
+
+	if (level == IPOIB_FLUSH_LIGHT) {
+		int oper_up;
+		ipoib_mark_paths_invalid(dev);
+		/* Set IPoIB operation as down to prevent races between:
+		 * the flush flow which leaves MCG and on the fly joins
+		 * which can happen during that time. mcast restart task
+		 * should deal with join requests we missed.
+		 */
+		oper_up = test_and_clear_bit(IPOIB_FLAG_OPER_UP, &priv->flags);
+		ipoib_mcast_dev_flush(dev);
+		if (oper_up)
+			set_bit(IPOIB_FLAG_OPER_UP, &priv->flags);
+		ipoib_reap_dead_ahs(priv);
+	}
+
+	if (level >= IPOIB_FLUSH_NORMAL)
+		ipoib_ib_dev_down(dev);
+
+	if (level == IPOIB_FLUSH_HEAVY) {
+		if (test_bit(IPOIB_FLAG_INITIALIZED, &priv->flags))
+			ipoib_ib_dev_stop(dev);
+
+		if (ipoib_ib_dev_open(dev))
+			return;
+
+		if (netif_queue_stopped(dev))
+			netif_start_queue(dev);
+	}
+
+	/*
+	 * The device could have been brought down between the start and when
+	 * we get here, don't bring it back up if it's not configured up
+	 */
+	if (test_bit(IPOIB_FLAG_ADMIN_UP, &priv->flags)) {
+		if (level >= IPOIB_FLUSH_NORMAL)
+			ipoib_ib_dev_up(dev);
+		if (ipoib_dev_addr_changed_valid(priv))
+			ipoib_mcast_restart_task(&priv->restart_task);
+	}
+}
+
+void ipoib_ib_dev_flush_light(struct work_struct *work)
+{
+	struct ipoib_dev_priv *priv =
+		container_of(work, struct ipoib_dev_priv, flush_light);
+
+	__ipoib_ib_dev_flush(priv, IPOIB_FLUSH_LIGHT, 0);
+}
+
+void ipoib_ib_dev_flush_normal(struct work_struct *work)
+{
+	struct ipoib_dev_priv *priv =
+		container_of(work, struct ipoib_dev_priv, flush_normal);
+
+	__ipoib_ib_dev_flush(priv, IPOIB_FLUSH_NORMAL, 0);
+}
+
+void ipoib_ib_dev_flush_heavy(struct work_struct *work)
+{
+	struct ipoib_dev_priv *priv =
+		container_of(work, struct ipoib_dev_priv, flush_heavy);
+
+	rtnl_lock();
+	__ipoib_ib_dev_flush(priv, IPOIB_FLUSH_HEAVY, 0);
+	rtnl_unlock();
+}
+
+void ipoib_ib_dev_cleanup(struct net_device *dev)
+{
+	struct ipoib_dev_priv *priv = ipoib_priv(dev);
+
+	ipoib_dbg(priv, "cleaning up ib_dev\n");
+	/*
+	 * We must make sure there are no more (path) completions
+	 * that may wish to touch priv fields that are no longer valid
+	 */
+	ipoib_flush_paths(dev);
+
+	ipoib_mcast_stop_thread(dev);
+	ipoib_mcast_dev_flush(dev);
+
+	/*
+	 * All of our ah references aren't free until after
+	 * ipoib_mcast_dev_flush(), ipoib_flush_paths, and
+	 * the neighbor garbage collection is stopped and reaped.
+	 * That should all be done now, so make a final ah flush.
+	 */
+	ipoib_reap_dead_ahs(priv);
+
+	clear_bit(IPOIB_PKEY_ASSIGNED, &priv->flags);
+
+	priv->rn_ops->ndo_uninit(dev);
+
+	if (priv->pd) {
+		ib_dealloc_pd(priv->pd);
+		priv->pd = NULL;
+	}
+}
+
diff --git a/drivers/infiniband/ulp/ipoib/ipoib_main.c b/drivers/infiniband/ulp/ipoib/ipoib_main.c
new file mode 100644
index 000000000..f10d4bcf8
--- /dev/null
+++ b/drivers/infiniband/ulp/ipoib/ipoib_main.c
@@ -0,0 +1,2698 @@
+/*
+ * Copyright (c) 2004 Topspin Communications.  All rights reserved.
+ * Copyright (c) 2005 Sun Microsystems, Inc. All rights reserved.
+ * Copyright (c) 2004 Voltaire, Inc. All rights reserved.
+ *
+ * This software is available to you under a choice of one of two
+ * licenses.  You may choose to be licensed under the terms of the GNU
+ * General Public License (GPL) Version 2, available from the file
+ * COPYING in the main directory of this source tree, or the
+ * OpenIB.org BSD license below:
+ *
+ *     Redistribution and use in source and binary forms, with or
+ *     without modification, are permitted provided that the following
+ *     conditions are met:
+ *
+ *      - Redistributions of source code must retain the above
+ *        copyright notice, this list of conditions and the following
+ *        disclaimer.
+ *
+ *      - Redistributions in binary form must reproduce the above
+ *        copyright notice, this list of conditions and the following
+ *        disclaimer in the documentation and/or other materials
+ *        provided with the distribution.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
+ * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
+ * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS
+ * BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN
+ * ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
+ * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+ * SOFTWARE.
+ */
+
+#include "ipoib.h"
+
+#include <linux/module.h>
+
+#include <linux/init.h>
+#include <linux/slab.h>
+#include <linux/kernel.h>
+#include <linux/vmalloc.h>
+
+#include <linux/if_arp.h>	/* For ARPHRD_xxx */
+
+#include <linux/ip.h>
+#include <linux/in.h>
+
+#include <linux/jhash.h>
+#include <net/arp.h>
+#include <net/addrconf.h>
+#include <linux/inetdevice.h>
+#include <rdma/ib_cache.h>
+
+MODULE_AUTHOR("Roland Dreier");
+MODULE_DESCRIPTION("IP-over-InfiniBand net driver");
+MODULE_LICENSE("Dual BSD/GPL");
+
+int ipoib_sendq_size __read_mostly = IPOIB_TX_RING_SIZE;
+int ipoib_recvq_size __read_mostly = IPOIB_RX_RING_SIZE;
+
+module_param_named(send_queue_size, ipoib_sendq_size, int, 0444);
+MODULE_PARM_DESC(send_queue_size, "Number of descriptors in send queue");
+module_param_named(recv_queue_size, ipoib_recvq_size, int, 0444);
+MODULE_PARM_DESC(recv_queue_size, "Number of descriptors in receive queue");
+
+#ifdef CONFIG_INFINIBAND_IPOIB_DEBUG
+int ipoib_debug_level;
+
+module_param_named(debug_level, ipoib_debug_level, int, 0644);
+MODULE_PARM_DESC(debug_level, "Enable debug tracing if > 0");
+#endif
+
+struct ipoib_path_iter {
+	struct net_device *dev;
+	struct ipoib_path  path;
+};
+
+static const u8 ipv4_bcast_addr[] = {
+	0x00, 0xff, 0xff, 0xff,
+	0xff, 0x12, 0x40, 0x1b,	0x00, 0x00, 0x00, 0x00,
+	0x00, 0x00, 0x00, 0x00,	0xff, 0xff, 0xff, 0xff
+};
+
+struct workqueue_struct *ipoib_workqueue;
+
+struct ib_sa_client ipoib_sa_client;
+
+static int ipoib_add_one(struct ib_device *device);
+static void ipoib_remove_one(struct ib_device *device, void *client_data);
+static void ipoib_neigh_reclaim(struct rcu_head *rp);
+static struct net_device *ipoib_get_net_dev_by_params(
+		struct ib_device *dev, u32 port, u16 pkey,
+		const union ib_gid *gid, const struct sockaddr *addr,
+		void *client_data);
+static int ipoib_set_mac(struct net_device *dev, void *addr);
+static int ipoib_ioctl(struct net_device *dev, struct ifreq *ifr,
+		       int cmd);
+
+static struct ib_client ipoib_client = {
+	.name   = "ipoib",
+	.add    = ipoib_add_one,
+	.remove = ipoib_remove_one,
+	.get_net_dev_by_params = ipoib_get_net_dev_by_params,
+};
+
+#ifdef CONFIG_INFINIBAND_IPOIB_DEBUG
+static int ipoib_netdev_event(struct notifier_block *this,
+			      unsigned long event, void *ptr)
+{
+	struct netdev_notifier_info *ni = ptr;
+	struct net_device *dev = ni->dev;
+
+	if (dev->netdev_ops->ndo_open != ipoib_open)
+		return NOTIFY_DONE;
+
+	switch (event) {
+	case NETDEV_REGISTER:
+		ipoib_create_debug_files(dev);
+		break;
+	case NETDEV_CHANGENAME:
+		ipoib_delete_debug_files(dev);
+		ipoib_create_debug_files(dev);
+		break;
+	case NETDEV_UNREGISTER:
+		ipoib_delete_debug_files(dev);
+		break;
+	}
+
+	return NOTIFY_DONE;
+}
+#endif
+
+int ipoib_open(struct net_device *dev)
+{
+	struct ipoib_dev_priv *priv = ipoib_priv(dev);
+
+	ipoib_dbg(priv, "bringing up interface\n");
+
+	netif_carrier_off(dev);
+
+	set_bit(IPOIB_FLAG_ADMIN_UP, &priv->flags);
+
+	if (ipoib_ib_dev_open(dev)) {
+		if (!test_bit(IPOIB_PKEY_ASSIGNED, &priv->flags))
+			return 0;
+		goto err_disable;
+	}
+
+	ipoib_ib_dev_up(dev);
+
+	if (!test_bit(IPOIB_FLAG_SUBINTERFACE, &priv->flags)) {
+		struct ipoib_dev_priv *cpriv;
+
+		/* Bring up any child interfaces too */
+		down_read(&priv->vlan_rwsem);
+		list_for_each_entry(cpriv, &priv->child_intfs, list) {
+			int flags;
+
+			flags = cpriv->dev->flags;
+			if (flags & IFF_UP)
+				continue;
+
+			dev_change_flags(cpriv->dev, flags | IFF_UP, NULL);
+		}
+		up_read(&priv->vlan_rwsem);
+	} else if (priv->parent) {
+		struct ipoib_dev_priv *ppriv = ipoib_priv(priv->parent);
+
+		if (!test_bit(IPOIB_FLAG_ADMIN_UP, &ppriv->flags))
+			ipoib_dbg(priv, "parent device %s is not up, so child device may be not functioning.\n",
+				  ppriv->dev->name);
+	}
+	netif_start_queue(dev);
+
+	return 0;
+
+err_disable:
+	clear_bit(IPOIB_FLAG_ADMIN_UP, &priv->flags);
+
+	return -EINVAL;
+}
+
+static int ipoib_stop(struct net_device *dev)
+{
+	struct ipoib_dev_priv *priv = ipoib_priv(dev);
+
+	ipoib_dbg(priv, "stopping interface\n");
+
+	clear_bit(IPOIB_FLAG_ADMIN_UP, &priv->flags);
+
+	netif_stop_queue(dev);
+
+	ipoib_ib_dev_down(dev);
+	ipoib_ib_dev_stop(dev);
+
+	if (!test_bit(IPOIB_FLAG_SUBINTERFACE, &priv->flags)) {
+		struct ipoib_dev_priv *cpriv;
+
+		/* Bring down any child interfaces too */
+		down_read(&priv->vlan_rwsem);
+		list_for_each_entry(cpriv, &priv->child_intfs, list) {
+			int flags;
+
+			flags = cpriv->dev->flags;
+			if (!(flags & IFF_UP))
+				continue;
+
+			dev_change_flags(cpriv->dev, flags & ~IFF_UP, NULL);
+		}
+		up_read(&priv->vlan_rwsem);
+	}
+
+	return 0;
+}
+
+static netdev_features_t ipoib_fix_features(struct net_device *dev, netdev_features_t features)
+{
+	struct ipoib_dev_priv *priv = ipoib_priv(dev);
+
+	if (test_bit(IPOIB_FLAG_ADMIN_CM, &priv->flags))
+		features &= ~(NETIF_F_IP_CSUM | NETIF_F_TSO);
+
+	return features;
+}
+
+static int ipoib_change_mtu(struct net_device *dev, int new_mtu)
+{
+	struct ipoib_dev_priv *priv = ipoib_priv(dev);
+	int ret = 0;
+
+	/* dev->mtu > 2K ==> connected mode */
+	if (ipoib_cm_admin_enabled(dev)) {
+		if (new_mtu > ipoib_cm_max_mtu(dev))
+			return -EINVAL;
+
+		if (new_mtu > priv->mcast_mtu)
+			ipoib_warn(priv, "mtu > %d will cause multicast packet drops.\n",
+				   priv->mcast_mtu);
+
+		dev->mtu = new_mtu;
+		return 0;
+	}
+
+	if (new_mtu < (ETH_MIN_MTU + IPOIB_ENCAP_LEN) ||
+	    new_mtu > IPOIB_UD_MTU(priv->max_ib_mtu))
+		return -EINVAL;
+
+	priv->admin_mtu = new_mtu;
+
+	if (priv->mcast_mtu < priv->admin_mtu)
+		ipoib_dbg(priv, "MTU must be smaller than the underlying "
+				"link layer MTU - 4 (%u)\n", priv->mcast_mtu);
+
+	new_mtu = min(priv->mcast_mtu, priv->admin_mtu);
+
+	if (priv->rn_ops->ndo_change_mtu) {
+		bool carrier_status = netif_carrier_ok(dev);
+
+		netif_carrier_off(dev);
+
+		/* notify lower level on the real mtu */
+		ret = priv->rn_ops->ndo_change_mtu(dev, new_mtu);
+
+		if (carrier_status)
+			netif_carrier_on(dev);
+	} else {
+		dev->mtu = new_mtu;
+	}
+
+	return ret;
+}
+
+static void ipoib_get_stats(struct net_device *dev,
+			    struct rtnl_link_stats64 *stats)
+{
+	struct ipoib_dev_priv *priv = ipoib_priv(dev);
+
+	if (priv->rn_ops->ndo_get_stats64)
+		priv->rn_ops->ndo_get_stats64(dev, stats);
+	else
+		netdev_stats_to_stats64(stats, &dev->stats);
+}
+
+/* Called with an RCU read lock taken */
+static bool ipoib_is_dev_match_addr_rcu(const struct sockaddr *addr,
+					struct net_device *dev)
+{
+	struct net *net = dev_net(dev);
+	struct in_device *in_dev;
+	struct sockaddr_in *addr_in = (struct sockaddr_in *)addr;
+	struct sockaddr_in6 *addr_in6 = (struct sockaddr_in6 *)addr;
+	__be32 ret_addr;
+
+	switch (addr->sa_family) {
+	case AF_INET:
+		in_dev = in_dev_get(dev);
+		if (!in_dev)
+			return false;
+
+		ret_addr = inet_confirm_addr(net, in_dev, 0,
+					     addr_in->sin_addr.s_addr,
+					     RT_SCOPE_HOST);
+		in_dev_put(in_dev);
+		if (ret_addr)
+			return true;
+
+		break;
+	case AF_INET6:
+		if (IS_ENABLED(CONFIG_IPV6) &&
+		    ipv6_chk_addr(net, &addr_in6->sin6_addr, dev, 1))
+			return true;
+
+		break;
+	}
+	return false;
+}
+
+/*
+ * Find the master net_device on top of the given net_device.
+ * @dev: base IPoIB net_device
+ *
+ * Returns the master net_device with a reference held, or the same net_device
+ * if no master exists.
+ */
+static struct net_device *ipoib_get_master_net_dev(struct net_device *dev)
+{
+	struct net_device *master;
+
+	rcu_read_lock();
+	master = netdev_master_upper_dev_get_rcu(dev);
+	if (master)
+		dev_hold(master);
+	rcu_read_unlock();
+
+	if (master)
+		return master;
+
+	dev_hold(dev);
+	return dev;
+}
+
+struct ipoib_walk_data {
+	const struct sockaddr *addr;
+	struct net_device *result;
+};
+
+static int ipoib_upper_walk(struct net_device *upper,
+			    struct netdev_nested_priv *priv)
+{
+	struct ipoib_walk_data *data = (struct ipoib_walk_data *)priv->data;
+	int ret = 0;
+
+	if (ipoib_is_dev_match_addr_rcu(data->addr, upper)) {
+		dev_hold(upper);
+		data->result = upper;
+		ret = 1;
+	}
+
+	return ret;
+}
+
+/**
+ * ipoib_get_net_dev_match_addr - Find a net_device matching
+ * the given address, which is an upper device of the given net_device.
+ *
+ * @addr: IP address to look for.
+ * @dev: base IPoIB net_device
+ *
+ * If found, returns the net_device with a reference held. Otherwise return
+ * NULL.
+ */
+static struct net_device *ipoib_get_net_dev_match_addr(
+		const struct sockaddr *addr, struct net_device *dev)
+{
+	struct netdev_nested_priv priv;
+	struct ipoib_walk_data data = {
+		.addr = addr,
+	};
+
+	priv.data = (void *)&data;
+	rcu_read_lock();
+	if (ipoib_is_dev_match_addr_rcu(addr, dev)) {
+		dev_hold(dev);
+		data.result = dev;
+		goto out;
+	}
+
+	netdev_walk_all_upper_dev_rcu(dev, ipoib_upper_walk, &priv);
+out:
+	rcu_read_unlock();
+	return data.result;
+}
+
+/* returns the number of IPoIB netdevs on top a given ipoib device matching a
+ * pkey_index and address, if one exists.
+ *
+ * @found_net_dev: contains a matching net_device if the return value >= 1,
+ * with a reference held. */
+static int ipoib_match_gid_pkey_addr(struct ipoib_dev_priv *priv,
+				     const union ib_gid *gid,
+				     u16 pkey_index,
+				     const struct sockaddr *addr,
+				     int nesting,
+				     struct net_device **found_net_dev)
+{
+	struct ipoib_dev_priv *child_priv;
+	struct net_device *net_dev = NULL;
+	int matches = 0;
+
+	if (priv->pkey_index == pkey_index &&
+	    (!gid || !memcmp(gid, &priv->local_gid, sizeof(*gid)))) {
+		if (!addr) {
+			net_dev = ipoib_get_master_net_dev(priv->dev);
+		} else {
+			/* Verify the net_device matches the IP address, as
+			 * IPoIB child devices currently share a GID. */
+			net_dev = ipoib_get_net_dev_match_addr(addr, priv->dev);
+		}
+		if (net_dev) {
+			if (!*found_net_dev)
+				*found_net_dev = net_dev;
+			else
+				dev_put(net_dev);
+			++matches;
+		}
+	}
+
+	/* Check child interfaces */
+	down_read_nested(&priv->vlan_rwsem, nesting);
+	list_for_each_entry(child_priv, &priv->child_intfs, list) {
+		matches += ipoib_match_gid_pkey_addr(child_priv, gid,
+						    pkey_index, addr,
+						    nesting + 1,
+						    found_net_dev);
+		if (matches > 1)
+			break;
+	}
+	up_read(&priv->vlan_rwsem);
+
+	return matches;
+}
+
+/* Returns the number of matching net_devs found (between 0 and 2). Also
+ * return the matching net_device in the @net_dev parameter, holding a
+ * reference to the net_device, if the number of matches >= 1 */
+static int __ipoib_get_net_dev_by_params(struct list_head *dev_list, u32 port,
+					 u16 pkey_index,
+					 const union ib_gid *gid,
+					 const struct sockaddr *addr,
+					 struct net_device **net_dev)
+{
+	struct ipoib_dev_priv *priv;
+	int matches = 0;
+
+	*net_dev = NULL;
+
+	list_for_each_entry(priv, dev_list, list) {
+		if (priv->port != port)
+			continue;
+
+		matches += ipoib_match_gid_pkey_addr(priv, gid, pkey_index,
+						     addr, 0, net_dev);
+		if (matches > 1)
+			break;
+	}
+
+	return matches;
+}
+
+static struct net_device *ipoib_get_net_dev_by_params(
+		struct ib_device *dev, u32 port, u16 pkey,
+		const union ib_gid *gid, const struct sockaddr *addr,
+		void *client_data)
+{
+	struct net_device *net_dev;
+	struct list_head *dev_list = client_data;
+	u16 pkey_index;
+	int matches;
+	int ret;
+
+	if (!rdma_protocol_ib(dev, port))
+		return NULL;
+
+	ret = ib_find_cached_pkey(dev, port, pkey, &pkey_index);
+	if (ret)
+		return NULL;
+
+	/* See if we can find a unique device matching the L2 parameters */
+	matches = __ipoib_get_net_dev_by_params(dev_list, port, pkey_index,
+						gid, NULL, &net_dev);
+
+	switch (matches) {
+	case 0:
+		return NULL;
+	case 1:
+		return net_dev;
+	}
+
+	dev_put(net_dev);
+
+	/* Couldn't find a unique device with L2 parameters only. Use L3
+	 * address to uniquely match the net device */
+	matches = __ipoib_get_net_dev_by_params(dev_list, port, pkey_index,
+						gid, addr, &net_dev);
+	switch (matches) {
+	case 0:
+		return NULL;
+	default:
+		dev_warn_ratelimited(&dev->dev,
+				     "duplicate IP address detected\n");
+		fallthrough;
+	case 1:
+		return net_dev;
+	}
+}
+
+int ipoib_set_mode(struct net_device *dev, const char *buf)
+{
+	struct ipoib_dev_priv *priv = ipoib_priv(dev);
+
+	if ((test_bit(IPOIB_FLAG_ADMIN_CM, &priv->flags) &&
+	     !strcmp(buf, "connected\n")) ||
+	     (!test_bit(IPOIB_FLAG_ADMIN_CM, &priv->flags) &&
+	     !strcmp(buf, "datagram\n"))) {
+		return 0;
+	}
+
+	/* flush paths if we switch modes so that connections are restarted */
+	if (IPOIB_CM_SUPPORTED(dev->dev_addr) && !strcmp(buf, "connected\n")) {
+		set_bit(IPOIB_FLAG_ADMIN_CM, &priv->flags);
+		ipoib_warn(priv, "enabling connected mode "
+			   "will cause multicast packet drops\n");
+		netdev_update_features(dev);
+		dev_set_mtu(dev, ipoib_cm_max_mtu(dev));
+		netif_set_real_num_tx_queues(dev, 1);
+		rtnl_unlock();
+		priv->tx_wr.wr.send_flags &= ~IB_SEND_IP_CSUM;
+
+		ipoib_flush_paths(dev);
+		return (!rtnl_trylock()) ? -EBUSY : 0;
+	}
+
+	if (!strcmp(buf, "datagram\n")) {
+		clear_bit(IPOIB_FLAG_ADMIN_CM, &priv->flags);
+		netdev_update_features(dev);
+		dev_set_mtu(dev, min(priv->mcast_mtu, dev->mtu));
+		netif_set_real_num_tx_queues(dev, dev->num_tx_queues);
+		rtnl_unlock();
+		ipoib_flush_paths(dev);
+		return (!rtnl_trylock()) ? -EBUSY : 0;
+	}
+
+	return -EINVAL;
+}
+
+struct ipoib_path *__path_find(struct net_device *dev, void *gid)
+{
+	struct ipoib_dev_priv *priv = ipoib_priv(dev);
+	struct rb_node *n = priv->path_tree.rb_node;
+	struct ipoib_path *path;
+	int ret;
+
+	while (n) {
+		path = rb_entry(n, struct ipoib_path, rb_node);
+
+		ret = memcmp(gid, path->pathrec.dgid.raw,
+			     sizeof (union ib_gid));
+
+		if (ret < 0)
+			n = n->rb_left;
+		else if (ret > 0)
+			n = n->rb_right;
+		else
+			return path;
+	}
+
+	return NULL;
+}
+
+static int __path_add(struct net_device *dev, struct ipoib_path *path)
+{
+	struct ipoib_dev_priv *priv = ipoib_priv(dev);
+	struct rb_node **n = &priv->path_tree.rb_node;
+	struct rb_node *pn = NULL;
+	struct ipoib_path *tpath;
+	int ret;
+
+	while (*n) {
+		pn = *n;
+		tpath = rb_entry(pn, struct ipoib_path, rb_node);
+
+		ret = memcmp(path->pathrec.dgid.raw, tpath->pathrec.dgid.raw,
+			     sizeof (union ib_gid));
+		if (ret < 0)
+			n = &pn->rb_left;
+		else if (ret > 0)
+			n = &pn->rb_right;
+		else
+			return -EEXIST;
+	}
+
+	rb_link_node(&path->rb_node, pn, n);
+	rb_insert_color(&path->rb_node, &priv->path_tree);
+
+	list_add_tail(&path->list, &priv->path_list);
+
+	return 0;
+}
+
+static void path_free(struct net_device *dev, struct ipoib_path *path)
+{
+	struct sk_buff *skb;
+
+	while ((skb = __skb_dequeue(&path->queue)))
+		dev_kfree_skb_irq(skb);
+
+	ipoib_dbg(ipoib_priv(dev), "%s\n", __func__);
+
+	/* remove all neigh connected to this path */
+	ipoib_del_neighs_by_gid(dev, path->pathrec.dgid.raw);
+
+	if (path->ah)
+		ipoib_put_ah(path->ah);
+
+	kfree(path);
+}
+
+#ifdef CONFIG_INFINIBAND_IPOIB_DEBUG
+
+struct ipoib_path_iter *ipoib_path_iter_init(struct net_device *dev)
+{
+	struct ipoib_path_iter *iter;
+
+	iter = kmalloc(sizeof(*iter), GFP_KERNEL);
+	if (!iter)
+		return NULL;
+
+	iter->dev = dev;
+	memset(iter->path.pathrec.dgid.raw, 0, 16);
+
+	if (ipoib_path_iter_next(iter)) {
+		kfree(iter);
+		return NULL;
+	}
+
+	return iter;
+}
+
+int ipoib_path_iter_next(struct ipoib_path_iter *iter)
+{
+	struct ipoib_dev_priv *priv = ipoib_priv(iter->dev);
+	struct rb_node *n;
+	struct ipoib_path *path;
+	int ret = 1;
+
+	spin_lock_irq(&priv->lock);
+
+	n = rb_first(&priv->path_tree);
+
+	while (n) {
+		path = rb_entry(n, struct ipoib_path, rb_node);
+
+		if (memcmp(iter->path.pathrec.dgid.raw, path->pathrec.dgid.raw,
+			   sizeof (union ib_gid)) < 0) {
+			iter->path = *path;
+			ret = 0;
+			break;
+		}
+
+		n = rb_next(n);
+	}
+
+	spin_unlock_irq(&priv->lock);
+
+	return ret;
+}
+
+void ipoib_path_iter_read(struct ipoib_path_iter *iter,
+			  struct ipoib_path *path)
+{
+	*path = iter->path;
+}
+
+#endif /* CONFIG_INFINIBAND_IPOIB_DEBUG */
+
+void ipoib_mark_paths_invalid(struct net_device *dev)
+{
+	struct ipoib_dev_priv *priv = ipoib_priv(dev);
+	struct ipoib_path *path, *tp;
+
+	spin_lock_irq(&priv->lock);
+
+	list_for_each_entry_safe(path, tp, &priv->path_list, list) {
+		ipoib_dbg(priv, "mark path LID 0x%08x GID %pI6 invalid\n",
+			  be32_to_cpu(sa_path_get_dlid(&path->pathrec)),
+			  path->pathrec.dgid.raw);
+		if (path->ah)
+			path->ah->valid = 0;
+	}
+
+	spin_unlock_irq(&priv->lock);
+}
+
+static void push_pseudo_header(struct sk_buff *skb, const char *daddr)
+{
+	struct ipoib_pseudo_header *phdr;
+
+	phdr = skb_push(skb, sizeof(*phdr));
+	memcpy(phdr->hwaddr, daddr, INFINIBAND_ALEN);
+}
+
+void ipoib_flush_paths(struct net_device *dev)
+{
+	struct ipoib_dev_priv *priv = ipoib_priv(dev);
+	struct ipoib_path *path, *tp;
+	LIST_HEAD(remove_list);
+	unsigned long flags;
+
+	netif_tx_lock_bh(dev);
+	spin_lock_irqsave(&priv->lock, flags);
+
+	list_splice_init(&priv->path_list, &remove_list);
+
+	list_for_each_entry(path, &remove_list, list)
+		rb_erase(&path->rb_node, &priv->path_tree);
+
+	list_for_each_entry_safe(path, tp, &remove_list, list) {
+		if (path->query)
+			ib_sa_cancel_query(path->query_id, path->query);
+		spin_unlock_irqrestore(&priv->lock, flags);
+		netif_tx_unlock_bh(dev);
+		wait_for_completion(&path->done);
+		path_free(dev, path);
+		netif_tx_lock_bh(dev);
+		spin_lock_irqsave(&priv->lock, flags);
+	}
+
+	spin_unlock_irqrestore(&priv->lock, flags);
+	netif_tx_unlock_bh(dev);
+}
+
+static void path_rec_completion(int status,
+				struct sa_path_rec *pathrec,
+				int num_prs, void *path_ptr)
+{
+	struct ipoib_path *path = path_ptr;
+	struct net_device *dev = path->dev;
+	struct ipoib_dev_priv *priv = ipoib_priv(dev);
+	struct ipoib_ah *ah = NULL;
+	struct ipoib_ah *old_ah = NULL;
+	struct ipoib_neigh *neigh, *tn;
+	struct sk_buff_head skqueue;
+	struct sk_buff *skb;
+	unsigned long flags;
+
+	if (!status)
+		ipoib_dbg(priv, "PathRec LID 0x%04x for GID %pI6\n",
+			  be32_to_cpu(sa_path_get_dlid(pathrec)),
+			  pathrec->dgid.raw);
+	else
+		ipoib_dbg(priv, "PathRec status %d for GID %pI6\n",
+			  status, path->pathrec.dgid.raw);
+
+	skb_queue_head_init(&skqueue);
+
+	if (!status) {
+		struct rdma_ah_attr av;
+
+		if (!ib_init_ah_attr_from_path(priv->ca, priv->port,
+					       pathrec, &av, NULL)) {
+			ah = ipoib_create_ah(dev, priv->pd, &av);
+			rdma_destroy_ah_attr(&av);
+		}
+	}
+
+	spin_lock_irqsave(&priv->lock, flags);
+
+	if (!IS_ERR_OR_NULL(ah)) {
+		/*
+		 * pathrec.dgid is used as the database key from the LLADDR,
+		 * it must remain unchanged even if the SA returns a different
+		 * GID to use in the AH.
+		 */
+		if (memcmp(pathrec->dgid.raw, path->pathrec.dgid.raw,
+			   sizeof(union ib_gid))) {
+			ipoib_dbg(
+				priv,
+				"%s got PathRec for gid %pI6 while asked for %pI6\n",
+				dev->name, pathrec->dgid.raw,
+				path->pathrec.dgid.raw);
+			memcpy(pathrec->dgid.raw, path->pathrec.dgid.raw,
+			       sizeof(union ib_gid));
+		}
+
+		path->pathrec = *pathrec;
+
+		old_ah   = path->ah;
+		path->ah = ah;
+
+		ipoib_dbg(priv, "created address handle %p for LID 0x%04x, SL %d\n",
+			  ah, be32_to_cpu(sa_path_get_dlid(pathrec)),
+			  pathrec->sl);
+
+		while ((skb = __skb_dequeue(&path->queue)))
+			__skb_queue_tail(&skqueue, skb);
+
+		list_for_each_entry_safe(neigh, tn, &path->neigh_list, list) {
+			if (neigh->ah) {
+				WARN_ON(neigh->ah != old_ah);
+				/*
+				 * Dropping the ah reference inside
+				 * priv->lock is safe here, because we
+				 * will hold one more reference from
+				 * the original value of path->ah (ie
+				 * old_ah).
+				 */
+				ipoib_put_ah(neigh->ah);
+			}
+			kref_get(&path->ah->ref);
+			neigh->ah = path->ah;
+
+			if (ipoib_cm_enabled(dev, neigh->daddr)) {
+				if (!ipoib_cm_get(neigh))
+					ipoib_cm_set(neigh, ipoib_cm_create_tx(dev,
+									       path,
+									       neigh));
+				if (!ipoib_cm_get(neigh)) {
+					ipoib_neigh_free(neigh);
+					continue;
+				}
+			}
+
+			while ((skb = __skb_dequeue(&neigh->queue)))
+				__skb_queue_tail(&skqueue, skb);
+		}
+		path->ah->valid = 1;
+	}
+
+	path->query = NULL;
+	complete(&path->done);
+
+	spin_unlock_irqrestore(&priv->lock, flags);
+
+	if (IS_ERR_OR_NULL(ah))
+		ipoib_del_neighs_by_gid(dev, path->pathrec.dgid.raw);
+
+	if (old_ah)
+		ipoib_put_ah(old_ah);
+
+	while ((skb = __skb_dequeue(&skqueue))) {
+		int ret;
+		skb->dev = dev;
+		ret = dev_queue_xmit(skb);
+		if (ret)
+			ipoib_warn(priv, "%s: dev_queue_xmit failed to re-queue packet, ret:%d\n",
+				   __func__, ret);
+	}
+}
+
+static void init_path_rec(struct ipoib_dev_priv *priv, struct ipoib_path *path,
+			  void *gid)
+{
+	path->dev = priv->dev;
+
+	if (rdma_cap_opa_ah(priv->ca, priv->port))
+		path->pathrec.rec_type = SA_PATH_REC_TYPE_OPA;
+	else
+		path->pathrec.rec_type = SA_PATH_REC_TYPE_IB;
+
+	memcpy(path->pathrec.dgid.raw, gid, sizeof(union ib_gid));
+	path->pathrec.sgid	    = priv->local_gid;
+	path->pathrec.pkey	    = cpu_to_be16(priv->pkey);
+	path->pathrec.numb_path     = 1;
+	path->pathrec.traffic_class = priv->broadcast->mcmember.traffic_class;
+}
+
+static struct ipoib_path *path_rec_create(struct net_device *dev, void *gid)
+{
+	struct ipoib_dev_priv *priv = ipoib_priv(dev);
+	struct ipoib_path *path;
+
+	if (!priv->broadcast)
+		return NULL;
+
+	path = kzalloc(sizeof(*path), GFP_ATOMIC);
+	if (!path)
+		return NULL;
+
+	skb_queue_head_init(&path->queue);
+
+	INIT_LIST_HEAD(&path->neigh_list);
+
+	init_path_rec(priv, path, gid);
+
+	return path;
+}
+
+static int path_rec_start(struct net_device *dev,
+			  struct ipoib_path *path)
+{
+	struct ipoib_dev_priv *priv = ipoib_priv(dev);
+
+	ipoib_dbg(priv, "Start path record lookup for %pI6\n",
+		  path->pathrec.dgid.raw);
+
+	init_completion(&path->done);
+
+	path->query_id =
+		ib_sa_path_rec_get(&ipoib_sa_client, priv->ca, priv->port,
+				   &path->pathrec,
+				   IB_SA_PATH_REC_DGID		|
+				   IB_SA_PATH_REC_SGID		|
+				   IB_SA_PATH_REC_NUMB_PATH	|
+				   IB_SA_PATH_REC_TRAFFIC_CLASS |
+				   IB_SA_PATH_REC_PKEY,
+				   1000, GFP_ATOMIC,
+				   path_rec_completion,
+				   path, &path->query);
+	if (path->query_id < 0) {
+		ipoib_warn(priv, "ib_sa_path_rec_get failed: %d\n", path->query_id);
+		path->query = NULL;
+		complete(&path->done);
+		return path->query_id;
+	}
+
+	return 0;
+}
+
+static void neigh_refresh_path(struct ipoib_neigh *neigh, u8 *daddr,
+			       struct net_device *dev)
+{
+	struct ipoib_dev_priv *priv = ipoib_priv(dev);
+	struct ipoib_path *path;
+	unsigned long flags;
+
+	spin_lock_irqsave(&priv->lock, flags);
+
+	path = __path_find(dev, daddr + 4);
+	if (!path)
+		goto out;
+	if (!path->query)
+		path_rec_start(dev, path);
+out:
+	spin_unlock_irqrestore(&priv->lock, flags);
+}
+
+static struct ipoib_neigh *neigh_add_path(struct sk_buff *skb, u8 *daddr,
+					  struct net_device *dev)
+{
+	struct ipoib_dev_priv *priv = ipoib_priv(dev);
+	struct rdma_netdev *rn = netdev_priv(dev);
+	struct ipoib_path *path;
+	struct ipoib_neigh *neigh;
+	unsigned long flags;
+
+	spin_lock_irqsave(&priv->lock, flags);
+	neigh = ipoib_neigh_alloc(daddr, dev);
+	if (!neigh) {
+		spin_unlock_irqrestore(&priv->lock, flags);
+		++dev->stats.tx_dropped;
+		dev_kfree_skb_any(skb);
+		return NULL;
+	}
+
+	/* To avoid race condition, make sure that the
+	 * neigh will be added only once.
+	 */
+	if (unlikely(!list_empty(&neigh->list))) {
+		spin_unlock_irqrestore(&priv->lock, flags);
+		return neigh;
+	}
+
+	path = __path_find(dev, daddr + 4);
+	if (!path) {
+		path = path_rec_create(dev, daddr + 4);
+		if (!path)
+			goto err_path;
+
+		__path_add(dev, path);
+	}
+
+	list_add_tail(&neigh->list, &path->neigh_list);
+
+	if (path->ah && path->ah->valid) {
+		kref_get(&path->ah->ref);
+		neigh->ah = path->ah;
+
+		if (ipoib_cm_enabled(dev, neigh->daddr)) {
+			if (!ipoib_cm_get(neigh))
+				ipoib_cm_set(neigh, ipoib_cm_create_tx(dev, path, neigh));
+			if (!ipoib_cm_get(neigh)) {
+				ipoib_neigh_free(neigh);
+				goto err_drop;
+			}
+			if (skb_queue_len(&neigh->queue) <
+			    IPOIB_MAX_PATH_REC_QUEUE) {
+				push_pseudo_header(skb, neigh->daddr);
+				__skb_queue_tail(&neigh->queue, skb);
+			} else {
+				ipoib_warn(priv, "queue length limit %d. Packet drop.\n",
+					   skb_queue_len(&neigh->queue));
+				goto err_drop;
+			}
+		} else {
+			spin_unlock_irqrestore(&priv->lock, flags);
+			path->ah->last_send = rn->send(dev, skb, path->ah->ah,
+						       IPOIB_QPN(daddr));
+			ipoib_neigh_put(neigh);
+			return NULL;
+		}
+	} else {
+		neigh->ah  = NULL;
+
+		if (!path->query && path_rec_start(dev, path))
+			goto err_path;
+		if (skb_queue_len(&neigh->queue) < IPOIB_MAX_PATH_REC_QUEUE) {
+			push_pseudo_header(skb, neigh->daddr);
+			__skb_queue_tail(&neigh->queue, skb);
+		} else {
+			goto err_drop;
+		}
+	}
+
+	spin_unlock_irqrestore(&priv->lock, flags);
+	ipoib_neigh_put(neigh);
+	return NULL;
+
+err_path:
+	ipoib_neigh_free(neigh);
+err_drop:
+	++dev->stats.tx_dropped;
+	dev_kfree_skb_any(skb);
+
+	spin_unlock_irqrestore(&priv->lock, flags);
+	ipoib_neigh_put(neigh);
+
+	return NULL;
+}
+
+static void unicast_arp_send(struct sk_buff *skb, struct net_device *dev,
+			     struct ipoib_pseudo_header *phdr)
+{
+	struct ipoib_dev_priv *priv = ipoib_priv(dev);
+	struct rdma_netdev *rn = netdev_priv(dev);
+	struct ipoib_path *path;
+	unsigned long flags;
+
+	spin_lock_irqsave(&priv->lock, flags);
+
+	/* no broadcast means that all paths are (going to be) not valid */
+	if (!priv->broadcast)
+		goto drop_and_unlock;
+
+	path = __path_find(dev, phdr->hwaddr + 4);
+	if (!path || !path->ah || !path->ah->valid) {
+		if (!path) {
+			path = path_rec_create(dev, phdr->hwaddr + 4);
+			if (!path)
+				goto drop_and_unlock;
+			__path_add(dev, path);
+		} else {
+			/*
+			 * make sure there are no changes in the existing
+			 * path record
+			 */
+			init_path_rec(priv, path, phdr->hwaddr + 4);
+		}
+		if (!path->query && path_rec_start(dev, path)) {
+			goto drop_and_unlock;
+		}
+
+		if (skb_queue_len(&path->queue) < IPOIB_MAX_PATH_REC_QUEUE) {
+			push_pseudo_header(skb, phdr->hwaddr);
+			__skb_queue_tail(&path->queue, skb);
+			goto unlock;
+		} else {
+			goto drop_and_unlock;
+		}
+	}
+
+	spin_unlock_irqrestore(&priv->lock, flags);
+	ipoib_dbg(priv, "Send unicast ARP to %08x\n",
+		  be32_to_cpu(sa_path_get_dlid(&path->pathrec)));
+	path->ah->last_send = rn->send(dev, skb, path->ah->ah,
+				       IPOIB_QPN(phdr->hwaddr));
+	return;
+
+drop_and_unlock:
+	++dev->stats.tx_dropped;
+	dev_kfree_skb_any(skb);
+unlock:
+	spin_unlock_irqrestore(&priv->lock, flags);
+}
+
+static netdev_tx_t ipoib_start_xmit(struct sk_buff *skb, struct net_device *dev)
+{
+	struct ipoib_dev_priv *priv = ipoib_priv(dev);
+	struct rdma_netdev *rn = netdev_priv(dev);
+	struct ipoib_neigh *neigh;
+	struct ipoib_pseudo_header *phdr;
+	struct ipoib_header *header;
+	unsigned long flags;
+
+	phdr = (struct ipoib_pseudo_header *) skb->data;
+	skb_pull(skb, sizeof(*phdr));
+	header = (struct ipoib_header *) skb->data;
+
+	if (unlikely(phdr->hwaddr[4] == 0xff)) {
+		/* multicast, arrange "if" according to probability */
+		if ((header->proto != htons(ETH_P_IP)) &&
+		    (header->proto != htons(ETH_P_IPV6)) &&
+		    (header->proto != htons(ETH_P_ARP)) &&
+		    (header->proto != htons(ETH_P_RARP)) &&
+		    (header->proto != htons(ETH_P_TIPC))) {
+			/* ethertype not supported by IPoIB */
+			++dev->stats.tx_dropped;
+			dev_kfree_skb_any(skb);
+			return NETDEV_TX_OK;
+		}
+		/* Add in the P_Key for multicast*/
+		phdr->hwaddr[8] = (priv->pkey >> 8) & 0xff;
+		phdr->hwaddr[9] = priv->pkey & 0xff;
+
+		neigh = ipoib_neigh_get(dev, phdr->hwaddr);
+		if (likely(neigh))
+			goto send_using_neigh;
+		ipoib_mcast_send(dev, phdr->hwaddr, skb);
+		return NETDEV_TX_OK;
+	}
+
+	/* unicast, arrange "switch" according to probability */
+	switch (header->proto) {
+	case htons(ETH_P_IP):
+	case htons(ETH_P_IPV6):
+	case htons(ETH_P_TIPC):
+		neigh = ipoib_neigh_get(dev, phdr->hwaddr);
+		if (unlikely(!neigh)) {
+			neigh = neigh_add_path(skb, phdr->hwaddr, dev);
+			if (likely(!neigh))
+				return NETDEV_TX_OK;
+		}
+		break;
+	case htons(ETH_P_ARP):
+	case htons(ETH_P_RARP):
+		/* for unicast ARP and RARP should always perform path find */
+		unicast_arp_send(skb, dev, phdr);
+		return NETDEV_TX_OK;
+	default:
+		/* ethertype not supported by IPoIB */
+		++dev->stats.tx_dropped;
+		dev_kfree_skb_any(skb);
+		return NETDEV_TX_OK;
+	}
+
+send_using_neigh:
+	/* note we now hold a ref to neigh */
+	if (ipoib_cm_get(neigh)) {
+		if (ipoib_cm_up(neigh)) {
+			ipoib_cm_send(dev, skb, ipoib_cm_get(neigh));
+			goto unref;
+		}
+	} else if (neigh->ah && neigh->ah->valid) {
+		neigh->ah->last_send = rn->send(dev, skb, neigh->ah->ah,
+						IPOIB_QPN(phdr->hwaddr));
+		goto unref;
+	} else if (neigh->ah) {
+		neigh_refresh_path(neigh, phdr->hwaddr, dev);
+	}
+
+	if (skb_queue_len(&neigh->queue) < IPOIB_MAX_PATH_REC_QUEUE) {
+		push_pseudo_header(skb, phdr->hwaddr);
+		spin_lock_irqsave(&priv->lock, flags);
+		__skb_queue_tail(&neigh->queue, skb);
+		spin_unlock_irqrestore(&priv->lock, flags);
+	} else {
+		++dev->stats.tx_dropped;
+		dev_kfree_skb_any(skb);
+	}
+
+unref:
+	ipoib_neigh_put(neigh);
+
+	return NETDEV_TX_OK;
+}
+
+static void ipoib_timeout(struct net_device *dev, unsigned int txqueue)
+{
+	struct ipoib_dev_priv *priv = ipoib_priv(dev);
+	struct rdma_netdev *rn = netdev_priv(dev);
+
+	if (rn->tx_timeout) {
+		rn->tx_timeout(dev, txqueue);
+		return;
+	}
+	ipoib_warn(priv, "transmit timeout: latency %d msecs\n",
+		   jiffies_to_msecs(jiffies - dev_trans_start(dev)));
+	ipoib_warn(priv,
+		   "queue stopped %d, tx_head %u, tx_tail %u, global_tx_head %u, global_tx_tail %u\n",
+		   netif_queue_stopped(dev), priv->tx_head, priv->tx_tail,
+		   priv->global_tx_head, priv->global_tx_tail);
+
+	/* XXX reset QP, etc. */
+}
+
+static int ipoib_hard_header(struct sk_buff *skb,
+			     struct net_device *dev,
+			     unsigned short type,
+			     const void *daddr,
+			     const void *saddr,
+			     unsigned int len)
+{
+	struct ipoib_header *header;
+
+	header = skb_push(skb, sizeof(*header));
+
+	header->proto = htons(type);
+	header->reserved = 0;
+
+	/*
+	 * we don't rely on dst_entry structure,  always stuff the
+	 * destination address into skb hard header so we can figure out where
+	 * to send the packet later.
+	 */
+	push_pseudo_header(skb, daddr);
+
+	return IPOIB_HARD_LEN;
+}
+
+static void ipoib_set_mcast_list(struct net_device *dev)
+{
+	struct ipoib_dev_priv *priv = ipoib_priv(dev);
+
+	if (!test_bit(IPOIB_FLAG_OPER_UP, &priv->flags)) {
+		ipoib_dbg(priv, "IPOIB_FLAG_OPER_UP not set");
+		return;
+	}
+
+	queue_work(priv->wq, &priv->restart_task);
+}
+
+static int ipoib_get_iflink(const struct net_device *dev)
+{
+	struct ipoib_dev_priv *priv = ipoib_priv(dev);
+
+	/* parent interface */
+	if (!test_bit(IPOIB_FLAG_SUBINTERFACE, &priv->flags))
+		return dev->ifindex;
+
+	/* child/vlan interface */
+	return priv->parent->ifindex;
+}
+
+static u32 ipoib_addr_hash(struct ipoib_neigh_hash *htbl, u8 *daddr)
+{
+	/*
+	 * Use only the address parts that contributes to spreading
+	 * The subnet prefix is not used as one can not connect to
+	 * same remote port (GUID) using the same remote QPN via two
+	 * different subnets.
+	 */
+	 /* qpn octets[1:4) & port GUID octets[12:20) */
+	u32 *d32 = (u32 *) daddr;
+	u32 hv;
+
+	hv = jhash_3words(d32[3], d32[4], IPOIB_QPN_MASK & d32[0], 0);
+	return hv & htbl->mask;
+}
+
+struct ipoib_neigh *ipoib_neigh_get(struct net_device *dev, u8 *daddr)
+{
+	struct ipoib_dev_priv *priv = ipoib_priv(dev);
+	struct ipoib_neigh_table *ntbl = &priv->ntbl;
+	struct ipoib_neigh_hash *htbl;
+	struct ipoib_neigh *neigh = NULL;
+	u32 hash_val;
+
+	rcu_read_lock_bh();
+
+	htbl = rcu_dereference_bh(ntbl->htbl);
+
+	if (!htbl)
+		goto out_unlock;
+
+	hash_val = ipoib_addr_hash(htbl, daddr);
+	for (neigh = rcu_dereference_bh(htbl->buckets[hash_val]);
+	     neigh != NULL;
+	     neigh = rcu_dereference_bh(neigh->hnext)) {
+		if (memcmp(daddr, neigh->daddr, INFINIBAND_ALEN) == 0) {
+			/* found, take one ref on behalf of the caller */
+			if (!refcount_inc_not_zero(&neigh->refcnt)) {
+				/* deleted */
+				neigh = NULL;
+				goto out_unlock;
+			}
+
+			if (likely(skb_queue_len(&neigh->queue) < IPOIB_MAX_PATH_REC_QUEUE))
+				neigh->alive = jiffies;
+			goto out_unlock;
+		}
+	}
+
+out_unlock:
+	rcu_read_unlock_bh();
+	return neigh;
+}
+
+static void __ipoib_reap_neigh(struct ipoib_dev_priv *priv)
+{
+	struct ipoib_neigh_table *ntbl = &priv->ntbl;
+	struct ipoib_neigh_hash *htbl;
+	unsigned long neigh_obsolete;
+	unsigned long dt;
+	unsigned long flags;
+	int i;
+	LIST_HEAD(remove_list);
+
+	spin_lock_irqsave(&priv->lock, flags);
+
+	htbl = rcu_dereference_protected(ntbl->htbl,
+					 lockdep_is_held(&priv->lock));
+
+	if (!htbl)
+		goto out_unlock;
+
+	/* neigh is obsolete if it was idle for two GC periods */
+	dt = 2 * arp_tbl.gc_interval;
+	neigh_obsolete = jiffies - dt;
+
+	for (i = 0; i < htbl->size; i++) {
+		struct ipoib_neigh *neigh;
+		struct ipoib_neigh __rcu **np = &htbl->buckets[i];
+
+		while ((neigh = rcu_dereference_protected(*np,
+							  lockdep_is_held(&priv->lock))) != NULL) {
+			/* was the neigh idle for two GC periods */
+			if (time_after(neigh_obsolete, neigh->alive)) {
+
+				ipoib_check_and_add_mcast_sendonly(priv, neigh->daddr + 4, &remove_list);
+
+				rcu_assign_pointer(*np,
+						   rcu_dereference_protected(neigh->hnext,
+									     lockdep_is_held(&priv->lock)));
+				/* remove from path/mc list */
+				list_del_init(&neigh->list);
+				call_rcu(&neigh->rcu, ipoib_neigh_reclaim);
+			} else {
+				np = &neigh->hnext;
+			}
+
+		}
+	}
+
+out_unlock:
+	spin_unlock_irqrestore(&priv->lock, flags);
+	ipoib_mcast_remove_list(&remove_list);
+}
+
+static void ipoib_reap_neigh(struct work_struct *work)
+{
+	struct ipoib_dev_priv *priv =
+		container_of(work, struct ipoib_dev_priv, neigh_reap_task.work);
+
+	__ipoib_reap_neigh(priv);
+
+	queue_delayed_work(priv->wq, &priv->neigh_reap_task,
+			   arp_tbl.gc_interval);
+}
+
+
+static struct ipoib_neigh *ipoib_neigh_ctor(u8 *daddr,
+				      struct net_device *dev)
+{
+	struct ipoib_neigh *neigh;
+
+	neigh = kzalloc(sizeof(*neigh), GFP_ATOMIC);
+	if (!neigh)
+		return NULL;
+
+	neigh->dev = dev;
+	memcpy(&neigh->daddr, daddr, sizeof(neigh->daddr));
+	skb_queue_head_init(&neigh->queue);
+	INIT_LIST_HEAD(&neigh->list);
+	ipoib_cm_set(neigh, NULL);
+	/* one ref on behalf of the caller */
+	refcount_set(&neigh->refcnt, 1);
+
+	return neigh;
+}
+
+struct ipoib_neigh *ipoib_neigh_alloc(u8 *daddr,
+				      struct net_device *dev)
+{
+	struct ipoib_dev_priv *priv = ipoib_priv(dev);
+	struct ipoib_neigh_table *ntbl = &priv->ntbl;
+	struct ipoib_neigh_hash *htbl;
+	struct ipoib_neigh *neigh;
+	u32 hash_val;
+
+	htbl = rcu_dereference_protected(ntbl->htbl,
+					 lockdep_is_held(&priv->lock));
+	if (!htbl) {
+		neigh = NULL;
+		goto out_unlock;
+	}
+
+	/* need to add a new neigh, but maybe some other thread succeeded?
+	 * recalc hash, maybe hash resize took place so we do a search
+	 */
+	hash_val = ipoib_addr_hash(htbl, daddr);
+	for (neigh = rcu_dereference_protected(htbl->buckets[hash_val],
+					       lockdep_is_held(&priv->lock));
+	     neigh != NULL;
+	     neigh = rcu_dereference_protected(neigh->hnext,
+					       lockdep_is_held(&priv->lock))) {
+		if (memcmp(daddr, neigh->daddr, INFINIBAND_ALEN) == 0) {
+			/* found, take one ref on behalf of the caller */
+			if (!refcount_inc_not_zero(&neigh->refcnt)) {
+				/* deleted */
+				neigh = NULL;
+				break;
+			}
+			neigh->alive = jiffies;
+			goto out_unlock;
+		}
+	}
+
+	neigh = ipoib_neigh_ctor(daddr, dev);
+	if (!neigh)
+		goto out_unlock;
+
+	/* one ref on behalf of the hash table */
+	refcount_inc(&neigh->refcnt);
+	neigh->alive = jiffies;
+	/* put in hash */
+	rcu_assign_pointer(neigh->hnext,
+			   rcu_dereference_protected(htbl->buckets[hash_val],
+						     lockdep_is_held(&priv->lock)));
+	rcu_assign_pointer(htbl->buckets[hash_val], neigh);
+	atomic_inc(&ntbl->entries);
+
+out_unlock:
+
+	return neigh;
+}
+
+void ipoib_neigh_dtor(struct ipoib_neigh *neigh)
+{
+	/* neigh reference count was dropprd to zero */
+	struct net_device *dev = neigh->dev;
+	struct ipoib_dev_priv *priv = ipoib_priv(dev);
+	struct sk_buff *skb;
+	if (neigh->ah)
+		ipoib_put_ah(neigh->ah);
+	while ((skb = __skb_dequeue(&neigh->queue))) {
+		++dev->stats.tx_dropped;
+		dev_kfree_skb_any(skb);
+	}
+	if (ipoib_cm_get(neigh))
+		ipoib_cm_destroy_tx(ipoib_cm_get(neigh));
+	ipoib_dbg(ipoib_priv(dev),
+		  "neigh free for %06x %pI6\n",
+		  IPOIB_QPN(neigh->daddr),
+		  neigh->daddr + 4);
+	kfree(neigh);
+	if (atomic_dec_and_test(&priv->ntbl.entries)) {
+		if (test_bit(IPOIB_NEIGH_TBL_FLUSH, &priv->flags))
+			complete(&priv->ntbl.flushed);
+	}
+}
+
+static void ipoib_neigh_reclaim(struct rcu_head *rp)
+{
+	/* Called as a result of removal from hash table */
+	struct ipoib_neigh *neigh = container_of(rp, struct ipoib_neigh, rcu);
+	/* note TX context may hold another ref */
+	ipoib_neigh_put(neigh);
+}
+
+void ipoib_neigh_free(struct ipoib_neigh *neigh)
+{
+	struct net_device *dev = neigh->dev;
+	struct ipoib_dev_priv *priv = ipoib_priv(dev);
+	struct ipoib_neigh_table *ntbl = &priv->ntbl;
+	struct ipoib_neigh_hash *htbl;
+	struct ipoib_neigh __rcu **np;
+	struct ipoib_neigh *n;
+	u32 hash_val;
+
+	htbl = rcu_dereference_protected(ntbl->htbl,
+					lockdep_is_held(&priv->lock));
+	if (!htbl)
+		return;
+
+	hash_val = ipoib_addr_hash(htbl, neigh->daddr);
+	np = &htbl->buckets[hash_val];
+	for (n = rcu_dereference_protected(*np,
+					    lockdep_is_held(&priv->lock));
+	     n != NULL;
+	     n = rcu_dereference_protected(*np,
+					lockdep_is_held(&priv->lock))) {
+		if (n == neigh) {
+			/* found */
+			rcu_assign_pointer(*np,
+					   rcu_dereference_protected(neigh->hnext,
+								     lockdep_is_held(&priv->lock)));
+			/* remove from parent list */
+			list_del_init(&neigh->list);
+			call_rcu(&neigh->rcu, ipoib_neigh_reclaim);
+			return;
+		} else {
+			np = &n->hnext;
+		}
+	}
+}
+
+static int ipoib_neigh_hash_init(struct ipoib_dev_priv *priv)
+{
+	struct ipoib_neigh_table *ntbl = &priv->ntbl;
+	struct ipoib_neigh_hash *htbl;
+	struct ipoib_neigh __rcu **buckets;
+	u32 size;
+
+	clear_bit(IPOIB_NEIGH_TBL_FLUSH, &priv->flags);
+	ntbl->htbl = NULL;
+	htbl = kzalloc(sizeof(*htbl), GFP_KERNEL);
+	if (!htbl)
+		return -ENOMEM;
+	size = roundup_pow_of_two(arp_tbl.gc_thresh3);
+	buckets = kvcalloc(size, sizeof(*buckets), GFP_KERNEL);
+	if (!buckets) {
+		kfree(htbl);
+		return -ENOMEM;
+	}
+	htbl->size = size;
+	htbl->mask = (size - 1);
+	htbl->buckets = buckets;
+	RCU_INIT_POINTER(ntbl->htbl, htbl);
+	htbl->ntbl = ntbl;
+	atomic_set(&ntbl->entries, 0);
+
+	/* start garbage collection */
+	queue_delayed_work(priv->wq, &priv->neigh_reap_task,
+			   arp_tbl.gc_interval);
+
+	return 0;
+}
+
+static void neigh_hash_free_rcu(struct rcu_head *head)
+{
+	struct ipoib_neigh_hash *htbl = container_of(head,
+						    struct ipoib_neigh_hash,
+						    rcu);
+	struct ipoib_neigh __rcu **buckets = htbl->buckets;
+	struct ipoib_neigh_table *ntbl = htbl->ntbl;
+
+	kvfree(buckets);
+	kfree(htbl);
+	complete(&ntbl->deleted);
+}
+
+void ipoib_del_neighs_by_gid(struct net_device *dev, u8 *gid)
+{
+	struct ipoib_dev_priv *priv = ipoib_priv(dev);
+	struct ipoib_neigh_table *ntbl = &priv->ntbl;
+	struct ipoib_neigh_hash *htbl;
+	unsigned long flags;
+	int i;
+
+	/* remove all neigh connected to a given path or mcast */
+	spin_lock_irqsave(&priv->lock, flags);
+
+	htbl = rcu_dereference_protected(ntbl->htbl,
+					 lockdep_is_held(&priv->lock));
+
+	if (!htbl)
+		goto out_unlock;
+
+	for (i = 0; i < htbl->size; i++) {
+		struct ipoib_neigh *neigh;
+		struct ipoib_neigh __rcu **np = &htbl->buckets[i];
+
+		while ((neigh = rcu_dereference_protected(*np,
+							  lockdep_is_held(&priv->lock))) != NULL) {
+			/* delete neighs belong to this parent */
+			if (!memcmp(gid, neigh->daddr + 4, sizeof (union ib_gid))) {
+				rcu_assign_pointer(*np,
+						   rcu_dereference_protected(neigh->hnext,
+									     lockdep_is_held(&priv->lock)));
+				/* remove from parent list */
+				list_del_init(&neigh->list);
+				call_rcu(&neigh->rcu, ipoib_neigh_reclaim);
+			} else {
+				np = &neigh->hnext;
+			}
+
+		}
+	}
+out_unlock:
+	spin_unlock_irqrestore(&priv->lock, flags);
+}
+
+static void ipoib_flush_neighs(struct ipoib_dev_priv *priv)
+{
+	struct ipoib_neigh_table *ntbl = &priv->ntbl;
+	struct ipoib_neigh_hash *htbl;
+	unsigned long flags;
+	int i, wait_flushed = 0;
+
+	init_completion(&priv->ntbl.flushed);
+	set_bit(IPOIB_NEIGH_TBL_FLUSH, &priv->flags);
+
+	spin_lock_irqsave(&priv->lock, flags);
+
+	htbl = rcu_dereference_protected(ntbl->htbl,
+					lockdep_is_held(&priv->lock));
+	if (!htbl)
+		goto out_unlock;
+
+	wait_flushed = atomic_read(&priv->ntbl.entries);
+	if (!wait_flushed)
+		goto free_htbl;
+
+	for (i = 0; i < htbl->size; i++) {
+		struct ipoib_neigh *neigh;
+		struct ipoib_neigh __rcu **np = &htbl->buckets[i];
+
+		while ((neigh = rcu_dereference_protected(*np,
+				       lockdep_is_held(&priv->lock))) != NULL) {
+			rcu_assign_pointer(*np,
+					   rcu_dereference_protected(neigh->hnext,
+								     lockdep_is_held(&priv->lock)));
+			/* remove from path/mc list */
+			list_del_init(&neigh->list);
+			call_rcu(&neigh->rcu, ipoib_neigh_reclaim);
+		}
+	}
+
+free_htbl:
+	rcu_assign_pointer(ntbl->htbl, NULL);
+	call_rcu(&htbl->rcu, neigh_hash_free_rcu);
+
+out_unlock:
+	spin_unlock_irqrestore(&priv->lock, flags);
+	if (wait_flushed)
+		wait_for_completion(&priv->ntbl.flushed);
+}
+
+static void ipoib_neigh_hash_uninit(struct net_device *dev)
+{
+	struct ipoib_dev_priv *priv = ipoib_priv(dev);
+
+	ipoib_dbg(priv, "%s\n", __func__);
+	init_completion(&priv->ntbl.deleted);
+
+	cancel_delayed_work_sync(&priv->neigh_reap_task);
+
+	ipoib_flush_neighs(priv);
+
+	wait_for_completion(&priv->ntbl.deleted);
+}
+
+static void ipoib_napi_add(struct net_device *dev)
+{
+	struct ipoib_dev_priv *priv = ipoib_priv(dev);
+
+	netif_napi_add_weight(dev, &priv->recv_napi, ipoib_rx_poll,
+			      IPOIB_NUM_WC);
+	netif_napi_add_weight(dev, &priv->send_napi, ipoib_tx_poll,
+			      MAX_SEND_CQE);
+}
+
+static void ipoib_napi_del(struct net_device *dev)
+{
+	struct ipoib_dev_priv *priv = ipoib_priv(dev);
+
+	netif_napi_del(&priv->recv_napi);
+	netif_napi_del(&priv->send_napi);
+}
+
+static void ipoib_dev_uninit_default(struct net_device *dev)
+{
+	struct ipoib_dev_priv *priv = ipoib_priv(dev);
+
+	ipoib_transport_dev_cleanup(dev);
+
+	ipoib_napi_del(dev);
+
+	ipoib_cm_dev_cleanup(dev);
+
+	kfree(priv->rx_ring);
+	vfree(priv->tx_ring);
+
+	priv->rx_ring = NULL;
+	priv->tx_ring = NULL;
+}
+
+static int ipoib_dev_init_default(struct net_device *dev)
+{
+	struct ipoib_dev_priv *priv = ipoib_priv(dev);
+	u8 addr_mod[3];
+
+	ipoib_napi_add(dev);
+
+	/* Allocate RX/TX "rings" to hold queued skbs */
+	priv->rx_ring =	kcalloc(ipoib_recvq_size,
+				       sizeof(*priv->rx_ring),
+				       GFP_KERNEL);
+	if (!priv->rx_ring)
+		goto out;
+
+	priv->tx_ring = vzalloc(array_size(ipoib_sendq_size,
+					   sizeof(*priv->tx_ring)));
+	if (!priv->tx_ring) {
+		pr_warn("%s: failed to allocate TX ring (%d entries)\n",
+			priv->ca->name, ipoib_sendq_size);
+		goto out_rx_ring_cleanup;
+	}
+
+	/* priv->tx_head, tx_tail and global_tx_tail/head are already 0 */
+
+	if (ipoib_transport_dev_init(dev, priv->ca)) {
+		pr_warn("%s: ipoib_transport_dev_init failed\n",
+			priv->ca->name);
+		goto out_tx_ring_cleanup;
+	}
+
+	/* after qp created set dev address */
+	addr_mod[0] = (priv->qp->qp_num >> 16) & 0xff;
+	addr_mod[1] = (priv->qp->qp_num >>  8) & 0xff;
+	addr_mod[2] = (priv->qp->qp_num) & 0xff;
+	dev_addr_mod(priv->dev, 1, addr_mod, sizeof(addr_mod));
+
+	return 0;
+
+out_tx_ring_cleanup:
+	vfree(priv->tx_ring);
+
+out_rx_ring_cleanup:
+	kfree(priv->rx_ring);
+
+out:
+	ipoib_napi_del(dev);
+	return -ENOMEM;
+}
+
+static int ipoib_ioctl(struct net_device *dev, struct ifreq *ifr,
+		       int cmd)
+{
+	struct ipoib_dev_priv *priv = ipoib_priv(dev);
+
+	if (!priv->rn_ops->ndo_eth_ioctl)
+		return -EOPNOTSUPP;
+
+	return priv->rn_ops->ndo_eth_ioctl(dev, ifr, cmd);
+}
+
+static int ipoib_dev_init(struct net_device *dev)
+{
+	struct ipoib_dev_priv *priv = ipoib_priv(dev);
+	int ret = -ENOMEM;
+
+	priv->qp = NULL;
+
+	/*
+	 * the various IPoIB tasks assume they will never race against
+	 * themselves, so always use a single thread workqueue
+	 */
+	priv->wq = alloc_ordered_workqueue("ipoib_wq", WQ_MEM_RECLAIM);
+	if (!priv->wq) {
+		pr_warn("%s: failed to allocate device WQ\n", dev->name);
+		goto out;
+	}
+
+	/* create pd, which used both for control and datapath*/
+	priv->pd = ib_alloc_pd(priv->ca, 0);
+	if (IS_ERR(priv->pd)) {
+		pr_warn("%s: failed to allocate PD\n", priv->ca->name);
+		goto clean_wq;
+	}
+
+	ret = priv->rn_ops->ndo_init(dev);
+	if (ret) {
+		pr_warn("%s failed to init HW resource\n", dev->name);
+		goto out_free_pd;
+	}
+
+	ret = ipoib_neigh_hash_init(priv);
+	if (ret) {
+		pr_warn("%s failed to init neigh hash\n", dev->name);
+		goto out_dev_uninit;
+	}
+
+	if (dev->flags & IFF_UP) {
+		if (ipoib_ib_dev_open(dev)) {
+			pr_warn("%s failed to open device\n", dev->name);
+			ret = -ENODEV;
+			goto out_hash_uninit;
+		}
+	}
+
+	return 0;
+
+out_hash_uninit:
+	ipoib_neigh_hash_uninit(dev);
+
+out_dev_uninit:
+	ipoib_ib_dev_cleanup(dev);
+
+out_free_pd:
+	if (priv->pd) {
+		ib_dealloc_pd(priv->pd);
+		priv->pd = NULL;
+	}
+
+clean_wq:
+	if (priv->wq) {
+		destroy_workqueue(priv->wq);
+		priv->wq = NULL;
+	}
+
+out:
+	return ret;
+}
+
+/*
+ * This must be called before doing an unregister_netdev on a parent device to
+ * shutdown the IB event handler.
+ */
+static void ipoib_parent_unregister_pre(struct net_device *ndev)
+{
+	struct ipoib_dev_priv *priv = ipoib_priv(ndev);
+
+	/*
+	 * ipoib_set_mac checks netif_running before pushing work, clearing
+	 * running ensures the it will not add more work.
+	 */
+	rtnl_lock();
+	dev_change_flags(priv->dev, priv->dev->flags & ~IFF_UP, NULL);
+	rtnl_unlock();
+
+	/* ipoib_event() cannot be running once this returns */
+	ib_unregister_event_handler(&priv->event_handler);
+
+	/*
+	 * Work on the queue grabs the rtnl lock, so this cannot be done while
+	 * also holding it.
+	 */
+	flush_workqueue(ipoib_workqueue);
+}
+
+static void ipoib_set_dev_features(struct ipoib_dev_priv *priv)
+{
+	priv->hca_caps = priv->ca->attrs.device_cap_flags;
+	priv->kernel_caps = priv->ca->attrs.kernel_cap_flags;
+
+	if (priv->hca_caps & IB_DEVICE_UD_IP_CSUM) {
+		priv->dev->hw_features |= NETIF_F_IP_CSUM | NETIF_F_RXCSUM;
+
+		if (priv->kernel_caps & IBK_UD_TSO)
+			priv->dev->hw_features |= NETIF_F_TSO;
+
+		priv->dev->features |= priv->dev->hw_features;
+	}
+}
+
+static int ipoib_parent_init(struct net_device *ndev)
+{
+	struct ipoib_dev_priv *priv = ipoib_priv(ndev);
+	struct ib_port_attr attr;
+	int result;
+
+	result = ib_query_port(priv->ca, priv->port, &attr);
+	if (result) {
+		pr_warn("%s: ib_query_port %d failed\n", priv->ca->name,
+			priv->port);
+		return result;
+	}
+	priv->max_ib_mtu = rdma_mtu_from_attr(priv->ca, priv->port, &attr);
+
+	result = ib_query_pkey(priv->ca, priv->port, 0, &priv->pkey);
+	if (result) {
+		pr_warn("%s: ib_query_pkey port %d failed (ret = %d)\n",
+			priv->ca->name, priv->port, result);
+		return result;
+	}
+
+	result = rdma_query_gid(priv->ca, priv->port, 0, &priv->local_gid);
+	if (result) {
+		pr_warn("%s: rdma_query_gid port %d failed (ret = %d)\n",
+			priv->ca->name, priv->port, result);
+		return result;
+	}
+	dev_addr_mod(priv->dev, 4, priv->local_gid.raw, sizeof(union ib_gid));
+
+	SET_NETDEV_DEV(priv->dev, priv->ca->dev.parent);
+	priv->dev->dev_port = priv->port - 1;
+	/* Let's set this one too for backwards compatibility. */
+	priv->dev->dev_id = priv->port - 1;
+
+	return 0;
+}
+
+static void ipoib_child_init(struct net_device *ndev)
+{
+	struct ipoib_dev_priv *priv = ipoib_priv(ndev);
+	struct ipoib_dev_priv *ppriv = ipoib_priv(priv->parent);
+
+	priv->max_ib_mtu = ppriv->max_ib_mtu;
+	set_bit(IPOIB_FLAG_SUBINTERFACE, &priv->flags);
+	if (memchr_inv(priv->dev->dev_addr, 0, INFINIBAND_ALEN))
+		memcpy(&priv->local_gid, priv->dev->dev_addr + 4,
+		       sizeof(priv->local_gid));
+	else {
+		__dev_addr_set(priv->dev, ppriv->dev->dev_addr,
+			       INFINIBAND_ALEN);
+		memcpy(&priv->local_gid, &ppriv->local_gid,
+		       sizeof(priv->local_gid));
+	}
+}
+
+static int ipoib_ndo_init(struct net_device *ndev)
+{
+	struct ipoib_dev_priv *priv = ipoib_priv(ndev);
+	int rc;
+	struct rdma_netdev *rn = netdev_priv(ndev);
+
+	if (priv->parent) {
+		ipoib_child_init(ndev);
+	} else {
+		rc = ipoib_parent_init(ndev);
+		if (rc)
+			return rc;
+	}
+
+	/* MTU will be reset when mcast join happens */
+	ndev->mtu = IPOIB_UD_MTU(priv->max_ib_mtu);
+	priv->mcast_mtu = priv->admin_mtu = ndev->mtu;
+	rn->mtu = priv->mcast_mtu;
+	ndev->max_mtu = IPOIB_CM_MTU;
+
+	ndev->neigh_priv_len = sizeof(struct ipoib_neigh);
+
+	/*
+	 * Set the full membership bit, so that we join the right
+	 * broadcast group, etc.
+	 */
+	priv->pkey |= 0x8000;
+
+	ndev->broadcast[8] = priv->pkey >> 8;
+	ndev->broadcast[9] = priv->pkey & 0xff;
+	set_bit(IPOIB_FLAG_DEV_ADDR_SET, &priv->flags);
+
+	ipoib_set_dev_features(priv);
+
+	rc = ipoib_dev_init(ndev);
+	if (rc) {
+		pr_warn("%s: failed to initialize device: %s port %d (ret = %d)\n",
+			priv->ca->name, priv->dev->name, priv->port, rc);
+		return rc;
+	}
+
+	if (priv->parent) {
+		struct ipoib_dev_priv *ppriv = ipoib_priv(priv->parent);
+
+		dev_hold(priv->parent);
+
+		down_write(&ppriv->vlan_rwsem);
+		list_add_tail(&priv->list, &ppriv->child_intfs);
+		up_write(&ppriv->vlan_rwsem);
+	}
+
+	return 0;
+}
+
+static void ipoib_ndo_uninit(struct net_device *dev)
+{
+	struct ipoib_dev_priv *priv = ipoib_priv(dev);
+
+	ASSERT_RTNL();
+
+	/*
+	 * ipoib_remove_one guarantees the children are removed before the
+	 * parent, and that is the only place where a parent can be removed.
+	 */
+	WARN_ON(!list_empty(&priv->child_intfs));
+
+	if (priv->parent) {
+		struct ipoib_dev_priv *ppriv = ipoib_priv(priv->parent);
+
+		down_write(&ppriv->vlan_rwsem);
+		list_del(&priv->list);
+		up_write(&ppriv->vlan_rwsem);
+	}
+
+	ipoib_neigh_hash_uninit(dev);
+
+	ipoib_ib_dev_cleanup(dev);
+
+	/* no more works over the priv->wq */
+	if (priv->wq) {
+		/* See ipoib_mcast_carrier_on_task() */
+		WARN_ON(test_bit(IPOIB_FLAG_OPER_UP, &priv->flags));
+		destroy_workqueue(priv->wq);
+		priv->wq = NULL;
+	}
+
+	if (priv->parent)
+		dev_put(priv->parent);
+}
+
+static int ipoib_set_vf_link_state(struct net_device *dev, int vf, int link_state)
+{
+	struct ipoib_dev_priv *priv = ipoib_priv(dev);
+
+	return ib_set_vf_link_state(priv->ca, vf, priv->port, link_state);
+}
+
+static int ipoib_get_vf_config(struct net_device *dev, int vf,
+			       struct ifla_vf_info *ivf)
+{
+	struct ipoib_dev_priv *priv = ipoib_priv(dev);
+	int err;
+
+	err = ib_get_vf_config(priv->ca, vf, priv->port, ivf);
+	if (err)
+		return err;
+
+	ivf->vf = vf;
+	memcpy(ivf->mac, dev->dev_addr, dev->addr_len);
+
+	return 0;
+}
+
+static int ipoib_set_vf_guid(struct net_device *dev, int vf, u64 guid, int type)
+{
+	struct ipoib_dev_priv *priv = ipoib_priv(dev);
+
+	if (type != IFLA_VF_IB_NODE_GUID && type != IFLA_VF_IB_PORT_GUID)
+		return -EINVAL;
+
+	return ib_set_vf_guid(priv->ca, vf, priv->port, guid, type);
+}
+
+static int ipoib_get_vf_guid(struct net_device *dev, int vf,
+			     struct ifla_vf_guid *node_guid,
+			     struct ifla_vf_guid *port_guid)
+{
+	struct ipoib_dev_priv *priv = ipoib_priv(dev);
+
+	return ib_get_vf_guid(priv->ca, vf, priv->port, node_guid, port_guid);
+}
+
+static int ipoib_get_vf_stats(struct net_device *dev, int vf,
+			      struct ifla_vf_stats *vf_stats)
+{
+	struct ipoib_dev_priv *priv = ipoib_priv(dev);
+
+	return ib_get_vf_stats(priv->ca, vf, priv->port, vf_stats);
+}
+
+static const struct header_ops ipoib_header_ops = {
+	.create	= ipoib_hard_header,
+};
+
+static const struct net_device_ops ipoib_netdev_ops_pf = {
+	.ndo_init		 = ipoib_ndo_init,
+	.ndo_uninit		 = ipoib_ndo_uninit,
+	.ndo_open		 = ipoib_open,
+	.ndo_stop		 = ipoib_stop,
+	.ndo_change_mtu		 = ipoib_change_mtu,
+	.ndo_fix_features	 = ipoib_fix_features,
+	.ndo_start_xmit		 = ipoib_start_xmit,
+	.ndo_tx_timeout		 = ipoib_timeout,
+	.ndo_set_rx_mode	 = ipoib_set_mcast_list,
+	.ndo_get_iflink		 = ipoib_get_iflink,
+	.ndo_set_vf_link_state	 = ipoib_set_vf_link_state,
+	.ndo_get_vf_config	 = ipoib_get_vf_config,
+	.ndo_get_vf_stats	 = ipoib_get_vf_stats,
+	.ndo_get_vf_guid	 = ipoib_get_vf_guid,
+	.ndo_set_vf_guid	 = ipoib_set_vf_guid,
+	.ndo_set_mac_address	 = ipoib_set_mac,
+	.ndo_get_stats64	 = ipoib_get_stats,
+	.ndo_eth_ioctl		 = ipoib_ioctl,
+};
+
+static const struct net_device_ops ipoib_netdev_ops_vf = {
+	.ndo_init		 = ipoib_ndo_init,
+	.ndo_uninit		 = ipoib_ndo_uninit,
+	.ndo_open		 = ipoib_open,
+	.ndo_stop		 = ipoib_stop,
+	.ndo_change_mtu		 = ipoib_change_mtu,
+	.ndo_fix_features	 = ipoib_fix_features,
+	.ndo_start_xmit	 	 = ipoib_start_xmit,
+	.ndo_tx_timeout		 = ipoib_timeout,
+	.ndo_set_rx_mode	 = ipoib_set_mcast_list,
+	.ndo_get_iflink		 = ipoib_get_iflink,
+	.ndo_get_stats64	 = ipoib_get_stats,
+	.ndo_eth_ioctl		 = ipoib_ioctl,
+};
+
+static const struct net_device_ops ipoib_netdev_default_pf = {
+	.ndo_init		 = ipoib_dev_init_default,
+	.ndo_uninit		 = ipoib_dev_uninit_default,
+	.ndo_open		 = ipoib_ib_dev_open_default,
+	.ndo_stop		 = ipoib_ib_dev_stop_default,
+};
+
+void ipoib_setup_common(struct net_device *dev)
+{
+	dev->header_ops		 = &ipoib_header_ops;
+	dev->netdev_ops          = &ipoib_netdev_default_pf;
+
+	ipoib_set_ethtool_ops(dev);
+
+	dev->watchdog_timeo	 = HZ;
+
+	dev->flags		|= IFF_BROADCAST | IFF_MULTICAST;
+
+	dev->hard_header_len	 = IPOIB_HARD_LEN;
+	dev->addr_len		 = INFINIBAND_ALEN;
+	dev->type		 = ARPHRD_INFINIBAND;
+	dev->tx_queue_len	 = ipoib_sendq_size * 2;
+	dev->features		 = (NETIF_F_VLAN_CHALLENGED	|
+				    NETIF_F_HIGHDMA);
+	netif_keep_dst(dev);
+
+	memcpy(dev->broadcast, ipv4_bcast_addr, INFINIBAND_ALEN);
+
+	/*
+	 * unregister_netdev always frees the netdev, we use this mode
+	 * consistently to unify all the various unregister paths, including
+	 * those connected to rtnl_link_ops which require it.
+	 */
+	dev->needs_free_netdev = true;
+}
+
+static void ipoib_build_priv(struct net_device *dev)
+{
+	struct ipoib_dev_priv *priv = ipoib_priv(dev);
+
+	priv->dev = dev;
+	spin_lock_init(&priv->lock);
+	init_rwsem(&priv->vlan_rwsem);
+	mutex_init(&priv->mcast_mutex);
+
+	INIT_LIST_HEAD(&priv->path_list);
+	INIT_LIST_HEAD(&priv->child_intfs);
+	INIT_LIST_HEAD(&priv->dead_ahs);
+	INIT_LIST_HEAD(&priv->multicast_list);
+
+	INIT_DELAYED_WORK(&priv->mcast_task,   ipoib_mcast_join_task);
+	INIT_WORK(&priv->carrier_on_task, ipoib_mcast_carrier_on_task);
+	INIT_WORK(&priv->flush_light,   ipoib_ib_dev_flush_light);
+	INIT_WORK(&priv->flush_normal,   ipoib_ib_dev_flush_normal);
+	INIT_WORK(&priv->flush_heavy,   ipoib_ib_dev_flush_heavy);
+	INIT_WORK(&priv->restart_task, ipoib_mcast_restart_task);
+	INIT_DELAYED_WORK(&priv->ah_reap_task, ipoib_reap_ah);
+	INIT_DELAYED_WORK(&priv->neigh_reap_task, ipoib_reap_neigh);
+}
+
+static struct net_device *ipoib_alloc_netdev(struct ib_device *hca, u32 port,
+					     const char *name)
+{
+	struct net_device *dev;
+
+	dev = rdma_alloc_netdev(hca, port, RDMA_NETDEV_IPOIB, name,
+				NET_NAME_UNKNOWN, ipoib_setup_common);
+	if (!IS_ERR(dev) || PTR_ERR(dev) != -EOPNOTSUPP)
+		return dev;
+
+	dev = alloc_netdev(sizeof(struct rdma_netdev), name, NET_NAME_UNKNOWN,
+			   ipoib_setup_common);
+	if (!dev)
+		return ERR_PTR(-ENOMEM);
+	return dev;
+}
+
+int ipoib_intf_init(struct ib_device *hca, u32 port, const char *name,
+		    struct net_device *dev)
+{
+	struct rdma_netdev *rn = netdev_priv(dev);
+	struct ipoib_dev_priv *priv;
+	int rc;
+
+	priv = kzalloc(sizeof(*priv), GFP_KERNEL);
+	if (!priv)
+		return -ENOMEM;
+
+	priv->ca = hca;
+	priv->port = port;
+
+	rc = rdma_init_netdev(hca, port, RDMA_NETDEV_IPOIB, name,
+			      NET_NAME_UNKNOWN, ipoib_setup_common, dev);
+	if (rc) {
+		if (rc != -EOPNOTSUPP)
+			goto out;
+
+		rn->send = ipoib_send;
+		rn->attach_mcast = ipoib_mcast_attach;
+		rn->detach_mcast = ipoib_mcast_detach;
+		rn->hca = hca;
+
+		rc = netif_set_real_num_tx_queues(dev, 1);
+		if (rc)
+			goto out;
+
+		rc = netif_set_real_num_rx_queues(dev, 1);
+		if (rc)
+			goto out;
+	}
+
+	priv->rn_ops = dev->netdev_ops;
+
+	if (hca->attrs.kernel_cap_flags & IBK_VIRTUAL_FUNCTION)
+		dev->netdev_ops	= &ipoib_netdev_ops_vf;
+	else
+		dev->netdev_ops	= &ipoib_netdev_ops_pf;
+
+	rn->clnt_priv = priv;
+	/*
+	 * Only the child register_netdev flows can handle priv_destructor
+	 * being set, so we force it to NULL here and handle manually until it
+	 * is safe to turn on.
+	 */
+	priv->next_priv_destructor = dev->priv_destructor;
+	dev->priv_destructor = NULL;
+
+	ipoib_build_priv(dev);
+
+	return 0;
+
+out:
+	kfree(priv);
+	return rc;
+}
+
+struct net_device *ipoib_intf_alloc(struct ib_device *hca, u32 port,
+				    const char *name)
+{
+	struct net_device *dev;
+	int rc;
+
+	dev = ipoib_alloc_netdev(hca, port, name);
+	if (IS_ERR(dev))
+		return dev;
+
+	rc = ipoib_intf_init(hca, port, name, dev);
+	if (rc) {
+		free_netdev(dev);
+		return ERR_PTR(rc);
+	}
+
+	/*
+	 * Upon success the caller must ensure ipoib_intf_free is called or
+	 * register_netdevice succeed'd and priv_destructor is set to
+	 * ipoib_intf_free.
+	 */
+	return dev;
+}
+
+void ipoib_intf_free(struct net_device *dev)
+{
+	struct ipoib_dev_priv *priv = ipoib_priv(dev);
+	struct rdma_netdev *rn = netdev_priv(dev);
+
+	dev->priv_destructor = priv->next_priv_destructor;
+	if (dev->priv_destructor)
+		dev->priv_destructor(dev);
+
+	/*
+	 * There are some error flows around register_netdev failing that may
+	 * attempt to call priv_destructor twice, prevent that from happening.
+	 */
+	dev->priv_destructor = NULL;
+
+	/* unregister/destroy is very complicated. Make bugs more obvious. */
+	rn->clnt_priv = NULL;
+
+	kfree(priv);
+}
+
+static ssize_t pkey_show(struct device *dev, struct device_attribute *attr,
+			 char *buf)
+{
+	struct net_device *ndev = to_net_dev(dev);
+	struct ipoib_dev_priv *priv = ipoib_priv(ndev);
+
+	return sysfs_emit(buf, "0x%04x\n", priv->pkey);
+}
+static DEVICE_ATTR_RO(pkey);
+
+static ssize_t umcast_show(struct device *dev, struct device_attribute *attr,
+			   char *buf)
+{
+	struct net_device *ndev = to_net_dev(dev);
+	struct ipoib_dev_priv *priv = ipoib_priv(ndev);
+
+	return sysfs_emit(buf, "%d\n",
+			  test_bit(IPOIB_FLAG_UMCAST, &priv->flags));
+}
+
+void ipoib_set_umcast(struct net_device *ndev, int umcast_val)
+{
+	struct ipoib_dev_priv *priv = ipoib_priv(ndev);
+
+	if (umcast_val > 0) {
+		set_bit(IPOIB_FLAG_UMCAST, &priv->flags);
+		ipoib_warn(priv, "ignoring multicast groups joined directly "
+				"by userspace\n");
+	} else
+		clear_bit(IPOIB_FLAG_UMCAST, &priv->flags);
+}
+
+static ssize_t umcast_store(struct device *dev, struct device_attribute *attr,
+			    const char *buf, size_t count)
+{
+	unsigned long umcast_val = simple_strtoul(buf, NULL, 0);
+
+	ipoib_set_umcast(to_net_dev(dev), umcast_val);
+
+	return count;
+}
+static DEVICE_ATTR_RW(umcast);
+
+int ipoib_add_umcast_attr(struct net_device *dev)
+{
+	return device_create_file(&dev->dev, &dev_attr_umcast);
+}
+
+static void set_base_guid(struct ipoib_dev_priv *priv, union ib_gid *gid)
+{
+	struct ipoib_dev_priv *child_priv;
+	struct net_device *netdev = priv->dev;
+
+	netif_addr_lock_bh(netdev);
+
+	memcpy(&priv->local_gid.global.interface_id,
+	       &gid->global.interface_id,
+	       sizeof(gid->global.interface_id));
+	dev_addr_mod(netdev, 4, (u8 *)&priv->local_gid, sizeof(priv->local_gid));
+	clear_bit(IPOIB_FLAG_DEV_ADDR_SET, &priv->flags);
+
+	netif_addr_unlock_bh(netdev);
+
+	if (!test_bit(IPOIB_FLAG_SUBINTERFACE, &priv->flags)) {
+		down_read(&priv->vlan_rwsem);
+		list_for_each_entry(child_priv, &priv->child_intfs, list)
+			set_base_guid(child_priv, gid);
+		up_read(&priv->vlan_rwsem);
+	}
+}
+
+static int ipoib_check_lladdr(struct net_device *dev,
+			      struct sockaddr_storage *ss)
+{
+	union ib_gid *gid = (union ib_gid *)(ss->__data + 4);
+	int ret = 0;
+
+	netif_addr_lock_bh(dev);
+
+	/* Make sure the QPN, reserved and subnet prefix match the current
+	 * lladdr, it also makes sure the lladdr is unicast.
+	 */
+	if (memcmp(dev->dev_addr, ss->__data,
+		   4 + sizeof(gid->global.subnet_prefix)) ||
+	    gid->global.interface_id == 0)
+		ret = -EINVAL;
+
+	netif_addr_unlock_bh(dev);
+
+	return ret;
+}
+
+static int ipoib_set_mac(struct net_device *dev, void *addr)
+{
+	struct ipoib_dev_priv *priv = ipoib_priv(dev);
+	struct sockaddr_storage *ss = addr;
+	int ret;
+
+	if (!(dev->priv_flags & IFF_LIVE_ADDR_CHANGE) && netif_running(dev))
+		return -EBUSY;
+
+	ret = ipoib_check_lladdr(dev, ss);
+	if (ret)
+		return ret;
+
+	set_base_guid(priv, (union ib_gid *)(ss->__data + 4));
+
+	queue_work(ipoib_workqueue, &priv->flush_light);
+
+	return 0;
+}
+
+static ssize_t create_child_store(struct device *dev,
+				  struct device_attribute *attr,
+				  const char *buf, size_t count)
+{
+	int pkey;
+	int ret;
+
+	if (sscanf(buf, "%i", &pkey) != 1)
+		return -EINVAL;
+
+	if (pkey <= 0 || pkey > 0xffff || pkey == 0x8000)
+		return -EINVAL;
+
+	ret = ipoib_vlan_add(to_net_dev(dev), pkey);
+
+	return ret ? ret : count;
+}
+static DEVICE_ATTR_WO(create_child);
+
+static ssize_t delete_child_store(struct device *dev,
+				  struct device_attribute *attr,
+				  const char *buf, size_t count)
+{
+	int pkey;
+	int ret;
+
+	if (sscanf(buf, "%i", &pkey) != 1)
+		return -EINVAL;
+
+	if (pkey < 0 || pkey > 0xffff)
+		return -EINVAL;
+
+	ret = ipoib_vlan_delete(to_net_dev(dev), pkey);
+
+	return ret ? ret : count;
+
+}
+static DEVICE_ATTR_WO(delete_child);
+
+int ipoib_add_pkey_attr(struct net_device *dev)
+{
+	return device_create_file(&dev->dev, &dev_attr_pkey);
+}
+
+/*
+ * We erroneously exposed the iface's port number in the dev_id
+ * sysfs field long after dev_port was introduced for that purpose[1],
+ * and we need to stop everyone from relying on that.
+ * Let's overload the shower routine for the dev_id file here
+ * to gently bring the issue up.
+ *
+ * [1] https://www.spinics.net/lists/netdev/msg272123.html
+ */
+static ssize_t dev_id_show(struct device *dev,
+			   struct device_attribute *attr, char *buf)
+{
+	struct net_device *ndev = to_net_dev(dev);
+
+	/*
+	 * ndev->dev_port will be equal to 0 in old kernel prior to commit
+	 * 9b8b2a323008 ("IB/ipoib: Use dev_port to expose network interface
+	 * port numbers") Zero was chosen as special case for user space
+	 * applications to fallback and query dev_id to check if it has
+	 * different value or not.
+	 *
+	 * Don't print warning in such scenario.
+	 *
+	 * https://github.com/systemd/systemd/blob/master/src/udev/udev-builtin-net_id.c#L358
+	 */
+	if (ndev->dev_port && ndev->dev_id == ndev->dev_port)
+		netdev_info_once(ndev,
+			"\"%s\" wants to know my dev_id. Should it look at dev_port instead? See Documentation/ABI/testing/sysfs-class-net for more info.\n",
+			current->comm);
+
+	return sysfs_emit(buf, "%#x\n", ndev->dev_id);
+}
+static DEVICE_ATTR_RO(dev_id);
+
+static int ipoib_intercept_dev_id_attr(struct net_device *dev)
+{
+	device_remove_file(&dev->dev, &dev_attr_dev_id);
+	return device_create_file(&dev->dev, &dev_attr_dev_id);
+}
+
+static struct net_device *ipoib_add_port(const char *format,
+					 struct ib_device *hca, u32 port)
+{
+	struct rtnl_link_ops *ops = ipoib_get_link_ops();
+	struct rdma_netdev_alloc_params params;
+	struct ipoib_dev_priv *priv;
+	struct net_device *ndev;
+	int result;
+
+	ndev = ipoib_intf_alloc(hca, port, format);
+	if (IS_ERR(ndev)) {
+		pr_warn("%s, %d: ipoib_intf_alloc failed %ld\n", hca->name, port,
+			PTR_ERR(ndev));
+		return ndev;
+	}
+	priv = ipoib_priv(ndev);
+
+	INIT_IB_EVENT_HANDLER(&priv->event_handler,
+			      priv->ca, ipoib_event);
+	ib_register_event_handler(&priv->event_handler);
+
+	/* call event handler to ensure pkey in sync */
+	queue_work(ipoib_workqueue, &priv->flush_heavy);
+
+	ndev->rtnl_link_ops = ipoib_get_link_ops();
+
+	result = register_netdev(ndev);
+	if (result) {
+		pr_warn("%s: couldn't register ipoib port %d; error %d\n",
+			hca->name, port, result);
+
+		ipoib_parent_unregister_pre(ndev);
+		ipoib_intf_free(ndev);
+		free_netdev(ndev);
+
+		return ERR_PTR(result);
+	}
+
+	if (hca->ops.rdma_netdev_get_params) {
+		int rc = hca->ops.rdma_netdev_get_params(hca, port,
+						     RDMA_NETDEV_IPOIB,
+						     &params);
+
+		if (!rc && ops->priv_size < params.sizeof_priv)
+			ops->priv_size = params.sizeof_priv;
+	}
+	/*
+	 * We cannot set priv_destructor before register_netdev because we
+	 * need priv to be always valid during the error flow to execute
+	 * ipoib_parent_unregister_pre(). Instead handle it manually and only
+	 * enter priv_destructor mode once we are completely registered.
+	 */
+	ndev->priv_destructor = ipoib_intf_free;
+
+	if (ipoib_intercept_dev_id_attr(ndev))
+		goto sysfs_failed;
+	if (ipoib_cm_add_mode_attr(ndev))
+		goto sysfs_failed;
+	if (ipoib_add_pkey_attr(ndev))
+		goto sysfs_failed;
+	if (ipoib_add_umcast_attr(ndev))
+		goto sysfs_failed;
+	if (device_create_file(&ndev->dev, &dev_attr_create_child))
+		goto sysfs_failed;
+	if (device_create_file(&ndev->dev, &dev_attr_delete_child))
+		goto sysfs_failed;
+
+	return ndev;
+
+sysfs_failed:
+	ipoib_parent_unregister_pre(ndev);
+	unregister_netdev(ndev);
+	return ERR_PTR(-ENOMEM);
+}
+
+static int ipoib_add_one(struct ib_device *device)
+{
+	struct list_head *dev_list;
+	struct net_device *dev;
+	struct ipoib_dev_priv *priv;
+	unsigned int p;
+	int count = 0;
+
+	dev_list = kmalloc(sizeof(*dev_list), GFP_KERNEL);
+	if (!dev_list)
+		return -ENOMEM;
+
+	INIT_LIST_HEAD(dev_list);
+
+	rdma_for_each_port (device, p) {
+		if (!rdma_protocol_ib(device, p))
+			continue;
+		dev = ipoib_add_port("ib%d", device, p);
+		if (!IS_ERR(dev)) {
+			priv = ipoib_priv(dev);
+			list_add_tail(&priv->list, dev_list);
+			count++;
+		}
+	}
+
+	if (!count) {
+		kfree(dev_list);
+		return -EOPNOTSUPP;
+	}
+
+	ib_set_client_data(device, &ipoib_client, dev_list);
+	return 0;
+}
+
+static void ipoib_remove_one(struct ib_device *device, void *client_data)
+{
+	struct ipoib_dev_priv *priv, *tmp, *cpriv, *tcpriv;
+	struct list_head *dev_list = client_data;
+
+	list_for_each_entry_safe(priv, tmp, dev_list, list) {
+		LIST_HEAD(head);
+		ipoib_parent_unregister_pre(priv->dev);
+
+		rtnl_lock();
+
+		list_for_each_entry_safe(cpriv, tcpriv, &priv->child_intfs,
+					 list)
+			unregister_netdevice_queue(cpriv->dev, &head);
+		unregister_netdevice_queue(priv->dev, &head);
+		unregister_netdevice_many(&head);
+
+		rtnl_unlock();
+	}
+
+	kfree(dev_list);
+}
+
+#ifdef CONFIG_INFINIBAND_IPOIB_DEBUG
+static struct notifier_block ipoib_netdev_notifier = {
+	.notifier_call = ipoib_netdev_event,
+};
+#endif
+
+static int __init ipoib_init_module(void)
+{
+	int ret;
+
+	ipoib_recvq_size = roundup_pow_of_two(ipoib_recvq_size);
+	ipoib_recvq_size = min(ipoib_recvq_size, IPOIB_MAX_QUEUE_SIZE);
+	ipoib_recvq_size = max(ipoib_recvq_size, IPOIB_MIN_QUEUE_SIZE);
+
+	ipoib_sendq_size = roundup_pow_of_two(ipoib_sendq_size);
+	ipoib_sendq_size = min(ipoib_sendq_size, IPOIB_MAX_QUEUE_SIZE);
+	ipoib_sendq_size = max3(ipoib_sendq_size, 2 * MAX_SEND_CQE, IPOIB_MIN_QUEUE_SIZE);
+#ifdef CONFIG_INFINIBAND_IPOIB_CM
+	ipoib_max_conn_qp = min(ipoib_max_conn_qp, IPOIB_CM_MAX_CONN_QP);
+	ipoib_max_conn_qp = max(ipoib_max_conn_qp, 0);
+#endif
+
+	/*
+	 * When copying small received packets, we only copy from the
+	 * linear data part of the SKB, so we rely on this condition.
+	 */
+	BUILD_BUG_ON(IPOIB_CM_COPYBREAK > IPOIB_CM_HEAD_SIZE);
+
+	ipoib_register_debugfs();
+
+	/*
+	 * We create a global workqueue here that is used for all flush
+	 * operations.  However, if you attempt to flush a workqueue
+	 * from a task on that same workqueue, it deadlocks the system.
+	 * We want to be able to flush the tasks associated with a
+	 * specific net device, so we also create a workqueue for each
+	 * netdevice.  We queue up the tasks for that device only on
+	 * its private workqueue, and we only queue up flush events
+	 * on our global flush workqueue.  This avoids the deadlocks.
+	 */
+	ipoib_workqueue = alloc_ordered_workqueue("ipoib_flush", 0);
+	if (!ipoib_workqueue) {
+		ret = -ENOMEM;
+		goto err_fs;
+	}
+
+	ib_sa_register_client(&ipoib_sa_client);
+
+	ret = ib_register_client(&ipoib_client);
+	if (ret)
+		goto err_sa;
+
+	ret = ipoib_netlink_init();
+	if (ret)
+		goto err_client;
+
+#ifdef CONFIG_INFINIBAND_IPOIB_DEBUG
+	register_netdevice_notifier(&ipoib_netdev_notifier);
+#endif
+	return 0;
+
+err_client:
+	ib_unregister_client(&ipoib_client);
+
+err_sa:
+	ib_sa_unregister_client(&ipoib_sa_client);
+	destroy_workqueue(ipoib_workqueue);
+
+err_fs:
+	ipoib_unregister_debugfs();
+
+	return ret;
+}
+
+static void __exit ipoib_cleanup_module(void)
+{
+#ifdef CONFIG_INFINIBAND_IPOIB_DEBUG
+	unregister_netdevice_notifier(&ipoib_netdev_notifier);
+#endif
+	ipoib_netlink_fini();
+	ib_unregister_client(&ipoib_client);
+	ib_sa_unregister_client(&ipoib_sa_client);
+	ipoib_unregister_debugfs();
+	destroy_workqueue(ipoib_workqueue);
+}
+
+module_init(ipoib_init_module);
+module_exit(ipoib_cleanup_module);
diff --git a/drivers/infiniband/ulp/ipoib/ipoib_multicast.c b/drivers/infiniband/ulp/ipoib/ipoib_multicast.c
new file mode 100644
index 000000000..5b3154503
--- /dev/null
+++ b/drivers/infiniband/ulp/ipoib/ipoib_multicast.c
@@ -0,0 +1,1051 @@
+/*
+ * Copyright (c) 2004, 2005 Topspin Communications.  All rights reserved.
+ * Copyright (c) 2005 Sun Microsystems, Inc. All rights reserved.
+ * Copyright (c) 2004 Voltaire, Inc. All rights reserved.
+ *
+ * This software is available to you under a choice of one of two
+ * licenses.  You may choose to be licensed under the terms of the GNU
+ * General Public License (GPL) Version 2, available from the file
+ * COPYING in the main directory of this source tree, or the
+ * OpenIB.org BSD license below:
+ *
+ *     Redistribution and use in source and binary forms, with or
+ *     without modification, are permitted provided that the following
+ *     conditions are met:
+ *
+ *      - Redistributions of source code must retain the above
+ *        copyright notice, this list of conditions and the following
+ *        disclaimer.
+ *
+ *      - Redistributions in binary form must reproduce the above
+ *        copyright notice, this list of conditions and the following
+ *        disclaimer in the documentation and/or other materials
+ *        provided with the distribution.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
+ * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
+ * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS
+ * BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN
+ * ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
+ * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+ * SOFTWARE.
+ */
+
+#include <linux/skbuff.h>
+#include <linux/rtnetlink.h>
+#include <linux/moduleparam.h>
+#include <linux/ip.h>
+#include <linux/in.h>
+#include <linux/igmp.h>
+#include <linux/inetdevice.h>
+#include <linux/delay.h>
+#include <linux/completion.h>
+#include <linux/slab.h>
+
+#include <net/dst.h>
+
+#include "ipoib.h"
+
+#ifdef CONFIG_INFINIBAND_IPOIB_DEBUG
+static int mcast_debug_level;
+
+module_param(mcast_debug_level, int, 0644);
+MODULE_PARM_DESC(mcast_debug_level,
+		 "Enable multicast debug tracing if > 0");
+#endif
+
+struct ipoib_mcast_iter {
+	struct net_device *dev;
+	union ib_gid       mgid;
+	unsigned long      created;
+	unsigned int       queuelen;
+	unsigned int       complete;
+	unsigned int       send_only;
+};
+
+/* join state that allows creating mcg with sendonly member request */
+#define SENDONLY_FULLMEMBER_JOIN	8
+
+/*
+ * This should be called with the priv->lock held
+ */
+static void __ipoib_mcast_schedule_join_thread(struct ipoib_dev_priv *priv,
+					       struct ipoib_mcast *mcast,
+					       bool delay)
+{
+	if (!test_bit(IPOIB_FLAG_OPER_UP, &priv->flags))
+		return;
+
+	/*
+	 * We will be scheduling *something*, so cancel whatever is
+	 * currently scheduled first
+	 */
+	cancel_delayed_work(&priv->mcast_task);
+	if (mcast && delay) {
+		/*
+		 * We had a failure and want to schedule a retry later
+		 */
+		mcast->backoff *= 2;
+		if (mcast->backoff > IPOIB_MAX_BACKOFF_SECONDS)
+			mcast->backoff = IPOIB_MAX_BACKOFF_SECONDS;
+		mcast->delay_until = jiffies + (mcast->backoff * HZ);
+		/*
+		 * Mark this mcast for its delay, but restart the
+		 * task immediately.  The join task will make sure to
+		 * clear out all entries without delays, and then
+		 * schedule itself to run again when the earliest
+		 * delay expires
+		 */
+		queue_delayed_work(priv->wq, &priv->mcast_task, 0);
+	} else if (delay) {
+		/*
+		 * Special case of retrying after a failure to
+		 * allocate the broadcast multicast group, wait
+		 * 1 second and try again
+		 */
+		queue_delayed_work(priv->wq, &priv->mcast_task, HZ);
+	} else
+		queue_delayed_work(priv->wq, &priv->mcast_task, 0);
+}
+
+static void ipoib_mcast_free(struct ipoib_mcast *mcast)
+{
+	struct net_device *dev = mcast->dev;
+	int tx_dropped = 0;
+
+	ipoib_dbg_mcast(ipoib_priv(dev), "deleting multicast group %pI6\n",
+			mcast->mcmember.mgid.raw);
+
+	/* remove all neigh connected to this mcast */
+	ipoib_del_neighs_by_gid(dev, mcast->mcmember.mgid.raw);
+
+	if (mcast->ah)
+		ipoib_put_ah(mcast->ah);
+
+	while (!skb_queue_empty(&mcast->pkt_queue)) {
+		++tx_dropped;
+		dev_kfree_skb_any(skb_dequeue(&mcast->pkt_queue));
+	}
+
+	netif_tx_lock_bh(dev);
+	dev->stats.tx_dropped += tx_dropped;
+	netif_tx_unlock_bh(dev);
+
+	kfree(mcast);
+}
+
+static struct ipoib_mcast *ipoib_mcast_alloc(struct net_device *dev)
+{
+	struct ipoib_mcast *mcast;
+
+	mcast = kzalloc(sizeof(*mcast), GFP_ATOMIC);
+	if (!mcast)
+		return NULL;
+
+	mcast->dev = dev;
+	mcast->created = jiffies;
+	mcast->delay_until = jiffies;
+	mcast->backoff = 1;
+
+	INIT_LIST_HEAD(&mcast->list);
+	INIT_LIST_HEAD(&mcast->neigh_list);
+	skb_queue_head_init(&mcast->pkt_queue);
+
+	return mcast;
+}
+
+static struct ipoib_mcast *__ipoib_mcast_find(struct net_device *dev, void *mgid)
+{
+	struct ipoib_dev_priv *priv = ipoib_priv(dev);
+	struct rb_node *n = priv->multicast_tree.rb_node;
+
+	while (n) {
+		struct ipoib_mcast *mcast;
+		int ret;
+
+		mcast = rb_entry(n, struct ipoib_mcast, rb_node);
+
+		ret = memcmp(mgid, mcast->mcmember.mgid.raw,
+			     sizeof (union ib_gid));
+		if (ret < 0)
+			n = n->rb_left;
+		else if (ret > 0)
+			n = n->rb_right;
+		else
+			return mcast;
+	}
+
+	return NULL;
+}
+
+static int __ipoib_mcast_add(struct net_device *dev, struct ipoib_mcast *mcast)
+{
+	struct ipoib_dev_priv *priv = ipoib_priv(dev);
+	struct rb_node **n = &priv->multicast_tree.rb_node, *pn = NULL;
+
+	while (*n) {
+		struct ipoib_mcast *tmcast;
+		int ret;
+
+		pn = *n;
+		tmcast = rb_entry(pn, struct ipoib_mcast, rb_node);
+
+		ret = memcmp(mcast->mcmember.mgid.raw, tmcast->mcmember.mgid.raw,
+			     sizeof (union ib_gid));
+		if (ret < 0)
+			n = &pn->rb_left;
+		else if (ret > 0)
+			n = &pn->rb_right;
+		else
+			return -EEXIST;
+	}
+
+	rb_link_node(&mcast->rb_node, pn, n);
+	rb_insert_color(&mcast->rb_node, &priv->multicast_tree);
+
+	return 0;
+}
+
+static int ipoib_mcast_join_finish(struct ipoib_mcast *mcast,
+				   struct ib_sa_mcmember_rec *mcmember)
+{
+	struct net_device *dev = mcast->dev;
+	struct ipoib_dev_priv *priv = ipoib_priv(dev);
+	struct rdma_netdev *rn = netdev_priv(dev);
+	struct ipoib_ah *ah;
+	struct rdma_ah_attr av;
+	int ret;
+	int set_qkey = 0;
+	int mtu;
+
+	mcast->mcmember = *mcmember;
+
+	/* Set the multicast MTU and cached Q_Key before we attach if it's
+	 * the broadcast group.
+	 */
+	if (!memcmp(mcast->mcmember.mgid.raw, priv->dev->broadcast + 4,
+		    sizeof (union ib_gid))) {
+		spin_lock_irq(&priv->lock);
+		if (!priv->broadcast) {
+			spin_unlock_irq(&priv->lock);
+			return -EAGAIN;
+		}
+		/*update priv member according to the new mcast*/
+		priv->broadcast->mcmember.qkey = mcmember->qkey;
+		priv->broadcast->mcmember.mtu = mcmember->mtu;
+		priv->broadcast->mcmember.traffic_class = mcmember->traffic_class;
+		priv->broadcast->mcmember.rate = mcmember->rate;
+		priv->broadcast->mcmember.sl = mcmember->sl;
+		priv->broadcast->mcmember.flow_label = mcmember->flow_label;
+		priv->broadcast->mcmember.hop_limit = mcmember->hop_limit;
+		/* assume if the admin and the mcast are the same both can be changed */
+		mtu = rdma_mtu_enum_to_int(priv->ca,  priv->port,
+					   priv->broadcast->mcmember.mtu);
+		if (priv->mcast_mtu == priv->admin_mtu)
+			priv->admin_mtu = IPOIB_UD_MTU(mtu);
+		priv->mcast_mtu = IPOIB_UD_MTU(mtu);
+		rn->mtu = priv->mcast_mtu;
+
+		priv->qkey = be32_to_cpu(priv->broadcast->mcmember.qkey);
+		spin_unlock_irq(&priv->lock);
+		priv->tx_wr.remote_qkey = priv->qkey;
+		set_qkey = 1;
+	}
+
+	if (!test_bit(IPOIB_MCAST_FLAG_SENDONLY, &mcast->flags)) {
+		if (test_and_set_bit(IPOIB_MCAST_FLAG_ATTACHED, &mcast->flags)) {
+			ipoib_warn(priv, "multicast group %pI6 already attached\n",
+				   mcast->mcmember.mgid.raw);
+
+			return 0;
+		}
+
+		ret = rn->attach_mcast(dev, priv->ca, &mcast->mcmember.mgid,
+				       be16_to_cpu(mcast->mcmember.mlid),
+				       set_qkey, priv->qkey);
+		if (ret < 0) {
+			ipoib_warn(priv, "couldn't attach QP to multicast group %pI6\n",
+				   mcast->mcmember.mgid.raw);
+
+			clear_bit(IPOIB_MCAST_FLAG_ATTACHED, &mcast->flags);
+			return ret;
+		}
+	}
+
+	memset(&av, 0, sizeof(av));
+	av.type = rdma_ah_find_type(priv->ca, priv->port);
+	rdma_ah_set_dlid(&av, be16_to_cpu(mcast->mcmember.mlid));
+	rdma_ah_set_port_num(&av, priv->port);
+	rdma_ah_set_sl(&av, mcast->mcmember.sl);
+	rdma_ah_set_static_rate(&av, mcast->mcmember.rate);
+
+	rdma_ah_set_grh(&av, &mcast->mcmember.mgid,
+			be32_to_cpu(mcast->mcmember.flow_label),
+			0, mcast->mcmember.hop_limit,
+			mcast->mcmember.traffic_class);
+
+	ah = ipoib_create_ah(dev, priv->pd, &av);
+	if (IS_ERR(ah)) {
+		ipoib_warn(priv, "ib_address_create failed %ld\n",
+			   -PTR_ERR(ah));
+		/* use original error */
+		return PTR_ERR(ah);
+	}
+	spin_lock_irq(&priv->lock);
+	mcast->ah = ah;
+	spin_unlock_irq(&priv->lock);
+
+	ipoib_dbg_mcast(priv, "MGID %pI6 AV %p, LID 0x%04x, SL %d\n",
+			mcast->mcmember.mgid.raw,
+			mcast->ah->ah,
+			be16_to_cpu(mcast->mcmember.mlid),
+			mcast->mcmember.sl);
+
+	/* actually send any queued packets */
+	netif_tx_lock_bh(dev);
+	while (!skb_queue_empty(&mcast->pkt_queue)) {
+		struct sk_buff *skb = skb_dequeue(&mcast->pkt_queue);
+
+		netif_tx_unlock_bh(dev);
+
+		skb->dev = dev;
+
+		ret = dev_queue_xmit(skb);
+		if (ret)
+			ipoib_warn(priv, "%s:dev_queue_xmit failed to re-queue packet, ret:%d\n",
+				   __func__, ret);
+		netif_tx_lock_bh(dev);
+	}
+	netif_tx_unlock_bh(dev);
+
+	return 0;
+}
+
+void ipoib_mcast_carrier_on_task(struct work_struct *work)
+{
+	struct ipoib_dev_priv *priv = container_of(work, struct ipoib_dev_priv,
+						   carrier_on_task);
+	struct ib_port_attr attr;
+
+	if (ib_query_port(priv->ca, priv->port, &attr) ||
+	    attr.state != IB_PORT_ACTIVE) {
+		ipoib_dbg(priv, "Keeping carrier off until IB port is active\n");
+		return;
+	}
+	/*
+	 * Take rtnl_lock to avoid racing with ipoib_stop() and
+	 * turning the carrier back on while a device is being
+	 * removed.  However, ipoib_stop() will attempt to flush
+	 * the workqueue while holding the rtnl lock, so loop
+	 * on trylock until either we get the lock or we see
+	 * FLAG_OPER_UP go away as that signals that we are bailing
+	 * and can safely ignore the carrier on work.
+	 */
+	while (!rtnl_trylock()) {
+		if (!test_bit(IPOIB_FLAG_OPER_UP, &priv->flags))
+			return;
+		else
+			msleep(20);
+	}
+	if (!ipoib_cm_admin_enabled(priv->dev))
+		dev_set_mtu(priv->dev, min(priv->mcast_mtu, priv->admin_mtu));
+	netif_carrier_on(priv->dev);
+	rtnl_unlock();
+}
+
+static int ipoib_mcast_join_complete(int status,
+				     struct ib_sa_multicast *multicast)
+{
+	struct ipoib_mcast *mcast = multicast->context;
+	struct net_device *dev = mcast->dev;
+	struct ipoib_dev_priv *priv = ipoib_priv(dev);
+
+	ipoib_dbg_mcast(priv, "%sjoin completion for %pI6 (status %d)\n",
+			test_bit(IPOIB_MCAST_FLAG_SENDONLY, &mcast->flags) ?
+			"sendonly " : "",
+			mcast->mcmember.mgid.raw, status);
+
+	/* We trap for port events ourselves. */
+	if (status == -ENETRESET) {
+		status = 0;
+		goto out;
+	}
+
+	if (!status)
+		status = ipoib_mcast_join_finish(mcast, &multicast->rec);
+
+	if (!status) {
+		mcast->backoff = 1;
+		mcast->delay_until = jiffies;
+
+		/*
+		 * Defer carrier on work to priv->wq to avoid a
+		 * deadlock on rtnl_lock here.  Requeue our multicast
+		 * work too, which will end up happening right after
+		 * our carrier on task work and will allow us to
+		 * send out all of the non-broadcast joins
+		 */
+		if (mcast == priv->broadcast) {
+			spin_lock_irq(&priv->lock);
+			queue_work(priv->wq, &priv->carrier_on_task);
+			__ipoib_mcast_schedule_join_thread(priv, NULL, 0);
+			goto out_locked;
+		}
+	} else {
+		bool silent_fail =
+		    test_bit(IPOIB_MCAST_FLAG_SENDONLY, &mcast->flags) &&
+		    status == -EINVAL;
+
+		if (mcast->logcount < 20) {
+			if (status == -ETIMEDOUT || status == -EAGAIN ||
+			    silent_fail) {
+				ipoib_dbg_mcast(priv, "%smulticast join failed for %pI6, status %d\n",
+						test_bit(IPOIB_MCAST_FLAG_SENDONLY, &mcast->flags) ? "sendonly " : "",
+						mcast->mcmember.mgid.raw, status);
+			} else {
+				ipoib_warn(priv, "%smulticast join failed for %pI6, status %d\n",
+						test_bit(IPOIB_MCAST_FLAG_SENDONLY, &mcast->flags) ? "sendonly " : "",
+					   mcast->mcmember.mgid.raw, status);
+			}
+
+			if (!silent_fail)
+				mcast->logcount++;
+		}
+
+		if (test_bit(IPOIB_MCAST_FLAG_SENDONLY, &mcast->flags) &&
+		    mcast->backoff >= 2) {
+			/*
+			 * We only retry sendonly joins once before we drop
+			 * the packet and quit trying to deal with the
+			 * group.  However, we leave the group in the
+			 * mcast list as an unjoined group.  If we want to
+			 * try joining again, we simply queue up a packet
+			 * and restart the join thread.  The empty queue
+			 * is why the join thread ignores this group.
+			 */
+			mcast->backoff = 1;
+			netif_tx_lock_bh(dev);
+			while (!skb_queue_empty(&mcast->pkt_queue)) {
+				++dev->stats.tx_dropped;
+				dev_kfree_skb_any(skb_dequeue(&mcast->pkt_queue));
+			}
+			netif_tx_unlock_bh(dev);
+		} else {
+			spin_lock_irq(&priv->lock);
+			/* Requeue this join task with a backoff delay */
+			__ipoib_mcast_schedule_join_thread(priv, mcast, 1);
+			goto out_locked;
+		}
+	}
+out:
+	spin_lock_irq(&priv->lock);
+out_locked:
+	/*
+	 * Make sure to set mcast->mc before we clear the busy flag to avoid
+	 * racing with code that checks for BUSY before checking mcast->mc
+	 */
+	if (status)
+		mcast->mc = NULL;
+	else
+		mcast->mc = multicast;
+	clear_bit(IPOIB_MCAST_FLAG_BUSY, &mcast->flags);
+	spin_unlock_irq(&priv->lock);
+	complete(&mcast->done);
+
+	return status;
+}
+
+/*
+ * Caller must hold 'priv->lock'
+ */
+static int ipoib_mcast_join(struct net_device *dev, struct ipoib_mcast *mcast)
+{
+	struct ipoib_dev_priv *priv = ipoib_priv(dev);
+	struct ib_sa_multicast *multicast;
+	struct ib_sa_mcmember_rec rec = {
+		.join_state = 1
+	};
+	ib_sa_comp_mask comp_mask;
+	int ret = 0;
+
+	if (!priv->broadcast ||
+	    !test_bit(IPOIB_FLAG_OPER_UP, &priv->flags))
+		return -EINVAL;
+
+	init_completion(&mcast->done);
+	set_bit(IPOIB_MCAST_FLAG_BUSY, &mcast->flags);
+
+	ipoib_dbg_mcast(priv, "joining MGID %pI6\n", mcast->mcmember.mgid.raw);
+
+	rec.mgid     = mcast->mcmember.mgid;
+	rec.port_gid = priv->local_gid;
+	rec.pkey     = cpu_to_be16(priv->pkey);
+
+	comp_mask =
+		IB_SA_MCMEMBER_REC_MGID		|
+		IB_SA_MCMEMBER_REC_PORT_GID	|
+		IB_SA_MCMEMBER_REC_PKEY		|
+		IB_SA_MCMEMBER_REC_JOIN_STATE;
+
+	if (mcast != priv->broadcast) {
+		/*
+		 * RFC 4391:
+		 *  The MGID MUST use the same P_Key, Q_Key, SL, MTU,
+		 *  and HopLimit as those used in the broadcast-GID.  The rest
+		 *  of attributes SHOULD follow the values used in the
+		 *  broadcast-GID as well.
+		 */
+		comp_mask |=
+			IB_SA_MCMEMBER_REC_QKEY			|
+			IB_SA_MCMEMBER_REC_MTU_SELECTOR		|
+			IB_SA_MCMEMBER_REC_MTU			|
+			IB_SA_MCMEMBER_REC_TRAFFIC_CLASS	|
+			IB_SA_MCMEMBER_REC_RATE_SELECTOR	|
+			IB_SA_MCMEMBER_REC_RATE			|
+			IB_SA_MCMEMBER_REC_SL			|
+			IB_SA_MCMEMBER_REC_FLOW_LABEL		|
+			IB_SA_MCMEMBER_REC_HOP_LIMIT;
+
+		rec.qkey	  = priv->broadcast->mcmember.qkey;
+		rec.mtu_selector  = IB_SA_EQ;
+		rec.mtu		  = priv->broadcast->mcmember.mtu;
+		rec.traffic_class = priv->broadcast->mcmember.traffic_class;
+		rec.rate_selector = IB_SA_EQ;
+		rec.rate	  = priv->broadcast->mcmember.rate;
+		rec.sl		  = priv->broadcast->mcmember.sl;
+		rec.flow_label	  = priv->broadcast->mcmember.flow_label;
+		rec.hop_limit	  = priv->broadcast->mcmember.hop_limit;
+
+		/*
+		 * Send-only IB Multicast joins work at the core IB layer but
+		 * require specific SM support.
+		 * We can use such joins here only if the current SM supports that feature.
+		 * However, if not, we emulate an Ethernet multicast send,
+		 * which does not require a multicast subscription and will
+		 * still send properly. The most appropriate thing to
+		 * do is to create the group if it doesn't exist as that
+		 * most closely emulates the behavior, from a user space
+		 * application perspective, of Ethernet multicast operation.
+		 */
+		if (test_bit(IPOIB_MCAST_FLAG_SENDONLY, &mcast->flags))
+			rec.join_state = SENDONLY_FULLMEMBER_JOIN;
+	}
+	spin_unlock_irq(&priv->lock);
+
+	multicast = ib_sa_join_multicast(&ipoib_sa_client, priv->ca, priv->port,
+					 &rec, comp_mask, GFP_KERNEL,
+					 ipoib_mcast_join_complete, mcast);
+	spin_lock_irq(&priv->lock);
+	if (IS_ERR(multicast)) {
+		ret = PTR_ERR(multicast);
+		ipoib_warn(priv, "ib_sa_join_multicast failed, status %d\n", ret);
+		/* Requeue this join task with a backoff delay */
+		__ipoib_mcast_schedule_join_thread(priv, mcast, 1);
+		clear_bit(IPOIB_MCAST_FLAG_BUSY, &mcast->flags);
+		spin_unlock_irq(&priv->lock);
+		complete(&mcast->done);
+		spin_lock_irq(&priv->lock);
+	}
+	return 0;
+}
+
+void ipoib_mcast_join_task(struct work_struct *work)
+{
+	struct ipoib_dev_priv *priv =
+		container_of(work, struct ipoib_dev_priv, mcast_task.work);
+	struct net_device *dev = priv->dev;
+	struct ib_port_attr port_attr;
+	unsigned long delay_until = 0;
+	struct ipoib_mcast *mcast = NULL;
+
+	if (!test_bit(IPOIB_FLAG_OPER_UP, &priv->flags))
+		return;
+
+	if (ib_query_port(priv->ca, priv->port, &port_attr)) {
+		ipoib_dbg(priv, "ib_query_port() failed\n");
+		return;
+	}
+	if (port_attr.state != IB_PORT_ACTIVE) {
+		ipoib_dbg(priv, "port state is not ACTIVE (state = %d) suspending join task\n",
+			  port_attr.state);
+		return;
+	}
+	priv->local_lid = port_attr.lid;
+	netif_addr_lock_bh(dev);
+
+	if (!test_bit(IPOIB_FLAG_DEV_ADDR_SET, &priv->flags)) {
+		netif_addr_unlock_bh(dev);
+		return;
+	}
+	netif_addr_unlock_bh(dev);
+
+	spin_lock_irq(&priv->lock);
+	if (!test_bit(IPOIB_FLAG_OPER_UP, &priv->flags))
+		goto out;
+
+	if (!priv->broadcast) {
+		struct ipoib_mcast *broadcast;
+
+		broadcast = ipoib_mcast_alloc(dev);
+		if (!broadcast) {
+			ipoib_warn(priv, "failed to allocate broadcast group\n");
+			/*
+			 * Restart us after a 1 second delay to retry
+			 * creating our broadcast group and attaching to
+			 * it.  Until this succeeds, this ipoib dev is
+			 * completely stalled (multicast wise).
+			 */
+			__ipoib_mcast_schedule_join_thread(priv, NULL, 1);
+			goto out;
+		}
+
+		memcpy(broadcast->mcmember.mgid.raw, priv->dev->broadcast + 4,
+		       sizeof (union ib_gid));
+		priv->broadcast = broadcast;
+
+		__ipoib_mcast_add(dev, priv->broadcast);
+	}
+
+	if (!test_bit(IPOIB_MCAST_FLAG_ATTACHED, &priv->broadcast->flags)) {
+		if (IS_ERR_OR_NULL(priv->broadcast->mc) &&
+		    !test_bit(IPOIB_MCAST_FLAG_BUSY, &priv->broadcast->flags)) {
+			mcast = priv->broadcast;
+			if (mcast->backoff > 1 &&
+			    time_before(jiffies, mcast->delay_until)) {
+				delay_until = mcast->delay_until;
+				mcast = NULL;
+			}
+		}
+		goto out;
+	}
+
+	/*
+	 * We'll never get here until the broadcast group is both allocated
+	 * and attached
+	 */
+	list_for_each_entry(mcast, &priv->multicast_list, list) {
+		if (IS_ERR_OR_NULL(mcast->mc) &&
+		    !test_bit(IPOIB_MCAST_FLAG_BUSY, &mcast->flags) &&
+		    (!test_bit(IPOIB_MCAST_FLAG_SENDONLY, &mcast->flags) ||
+		     !skb_queue_empty(&mcast->pkt_queue))) {
+			if (mcast->backoff == 1 ||
+			    time_after_eq(jiffies, mcast->delay_until)) {
+				/* Found the next unjoined group */
+				if (ipoib_mcast_join(dev, mcast)) {
+					spin_unlock_irq(&priv->lock);
+					return;
+				}
+			} else if (!delay_until ||
+				 time_before(mcast->delay_until, delay_until))
+				delay_until = mcast->delay_until;
+		}
+	}
+
+	mcast = NULL;
+	ipoib_dbg_mcast(priv, "successfully started all multicast joins\n");
+
+out:
+	if (delay_until) {
+		cancel_delayed_work(&priv->mcast_task);
+		queue_delayed_work(priv->wq, &priv->mcast_task,
+				   delay_until - jiffies);
+	}
+	if (mcast)
+		ipoib_mcast_join(dev, mcast);
+
+	spin_unlock_irq(&priv->lock);
+}
+
+void ipoib_mcast_start_thread(struct net_device *dev)
+{
+	struct ipoib_dev_priv *priv = ipoib_priv(dev);
+	unsigned long flags;
+
+	ipoib_dbg_mcast(priv, "starting multicast thread\n");
+
+	spin_lock_irqsave(&priv->lock, flags);
+	__ipoib_mcast_schedule_join_thread(priv, NULL, 0);
+	spin_unlock_irqrestore(&priv->lock, flags);
+}
+
+void ipoib_mcast_stop_thread(struct net_device *dev)
+{
+	struct ipoib_dev_priv *priv = ipoib_priv(dev);
+
+	ipoib_dbg_mcast(priv, "stopping multicast thread\n");
+
+	cancel_delayed_work_sync(&priv->mcast_task);
+}
+
+static int ipoib_mcast_leave(struct net_device *dev, struct ipoib_mcast *mcast)
+{
+	struct ipoib_dev_priv *priv = ipoib_priv(dev);
+	struct rdma_netdev *rn = netdev_priv(dev);
+	int ret = 0;
+
+	if (test_and_clear_bit(IPOIB_MCAST_FLAG_BUSY, &mcast->flags))
+		ipoib_warn(priv, "ipoib_mcast_leave on an in-flight join\n");
+
+	if (!IS_ERR_OR_NULL(mcast->mc))
+		ib_sa_free_multicast(mcast->mc);
+
+	if (test_and_clear_bit(IPOIB_MCAST_FLAG_ATTACHED, &mcast->flags)) {
+		ipoib_dbg_mcast(priv, "leaving MGID %pI6\n",
+				mcast->mcmember.mgid.raw);
+
+		/* Remove ourselves from the multicast group */
+		ret = rn->detach_mcast(dev, priv->ca, &mcast->mcmember.mgid,
+				       be16_to_cpu(mcast->mcmember.mlid));
+		if (ret)
+			ipoib_warn(priv, "ib_detach_mcast failed (result = %d)\n", ret);
+	} else if (!test_bit(IPOIB_MCAST_FLAG_SENDONLY, &mcast->flags))
+		ipoib_dbg(priv, "leaving with no mcmember but not a "
+			  "SENDONLY join\n");
+
+	return 0;
+}
+
+/*
+ * Check if the multicast group is sendonly. If so remove it from the maps
+ * and add to the remove list
+ */
+void ipoib_check_and_add_mcast_sendonly(struct ipoib_dev_priv *priv, u8 *mgid,
+				struct list_head *remove_list)
+{
+	/* Is this multicast ? */
+	if (*mgid == 0xff) {
+		struct ipoib_mcast *mcast = __ipoib_mcast_find(priv->dev, mgid);
+
+		if (mcast && test_bit(IPOIB_MCAST_FLAG_SENDONLY, &mcast->flags)) {
+			list_del(&mcast->list);
+			rb_erase(&mcast->rb_node, &priv->multicast_tree);
+			list_add_tail(&mcast->list, remove_list);
+		}
+	}
+}
+
+void ipoib_mcast_remove_list(struct list_head *remove_list)
+{
+	struct ipoib_mcast *mcast, *tmcast;
+
+	/*
+	 * make sure the in-flight joins have finished before we attempt
+	 * to leave
+	 */
+	list_for_each_entry_safe(mcast, tmcast, remove_list, list)
+		if (test_bit(IPOIB_MCAST_FLAG_BUSY, &mcast->flags))
+			wait_for_completion(&mcast->done);
+
+	list_for_each_entry_safe(mcast, tmcast, remove_list, list) {
+		ipoib_mcast_leave(mcast->dev, mcast);
+		ipoib_mcast_free(mcast);
+	}
+}
+
+void ipoib_mcast_send(struct net_device *dev, u8 *daddr, struct sk_buff *skb)
+{
+	struct ipoib_dev_priv *priv = ipoib_priv(dev);
+	struct rdma_netdev *rn = netdev_priv(dev);
+	struct ipoib_mcast *mcast;
+	unsigned long flags;
+	void *mgid = daddr + 4;
+
+	spin_lock_irqsave(&priv->lock, flags);
+
+	if (!test_bit(IPOIB_FLAG_OPER_UP, &priv->flags)		||
+	    !priv->broadcast					||
+	    !test_bit(IPOIB_MCAST_FLAG_ATTACHED, &priv->broadcast->flags)) {
+		++dev->stats.tx_dropped;
+		dev_kfree_skb_any(skb);
+		goto unlock;
+	}
+
+	mcast = __ipoib_mcast_find(dev, mgid);
+	if (!mcast || !mcast->ah) {
+		if (!mcast) {
+			/* Let's create a new send only group now */
+			ipoib_dbg_mcast(priv, "setting up send only multicast group for %pI6\n",
+					mgid);
+
+			mcast = ipoib_mcast_alloc(dev);
+			if (!mcast) {
+				ipoib_warn(priv, "unable to allocate memory "
+					   "for multicast structure\n");
+				++dev->stats.tx_dropped;
+				dev_kfree_skb_any(skb);
+				goto unlock;
+			}
+
+			set_bit(IPOIB_MCAST_FLAG_SENDONLY, &mcast->flags);
+			memcpy(mcast->mcmember.mgid.raw, mgid,
+			       sizeof (union ib_gid));
+			__ipoib_mcast_add(dev, mcast);
+			list_add_tail(&mcast->list, &priv->multicast_list);
+		}
+		if (skb_queue_len(&mcast->pkt_queue) < IPOIB_MAX_MCAST_QUEUE) {
+			/* put pseudoheader back on for next time */
+			skb_push(skb, sizeof(struct ipoib_pseudo_header));
+			skb_queue_tail(&mcast->pkt_queue, skb);
+		} else {
+			++dev->stats.tx_dropped;
+			dev_kfree_skb_any(skb);
+		}
+		if (!test_bit(IPOIB_MCAST_FLAG_BUSY, &mcast->flags)) {
+			__ipoib_mcast_schedule_join_thread(priv, NULL, 0);
+		}
+	} else {
+		struct ipoib_neigh *neigh;
+
+		spin_unlock_irqrestore(&priv->lock, flags);
+		neigh = ipoib_neigh_get(dev, daddr);
+		spin_lock_irqsave(&priv->lock, flags);
+		if (!neigh) {
+			neigh = ipoib_neigh_alloc(daddr, dev);
+			/* Make sure that the neigh will be added only
+			 * once to mcast list.
+			 */
+			if (neigh && list_empty(&neigh->list)) {
+				kref_get(&mcast->ah->ref);
+				neigh->ah	= mcast->ah;
+				neigh->ah->valid = 1;
+				list_add_tail(&neigh->list, &mcast->neigh_list);
+			}
+		}
+		spin_unlock_irqrestore(&priv->lock, flags);
+		mcast->ah->last_send = rn->send(dev, skb, mcast->ah->ah,
+						IB_MULTICAST_QPN);
+		if (neigh)
+			ipoib_neigh_put(neigh);
+		return;
+	}
+
+unlock:
+	spin_unlock_irqrestore(&priv->lock, flags);
+}
+
+void ipoib_mcast_dev_flush(struct net_device *dev)
+{
+	struct ipoib_dev_priv *priv = ipoib_priv(dev);
+	LIST_HEAD(remove_list);
+	struct ipoib_mcast *mcast, *tmcast;
+	unsigned long flags;
+
+	mutex_lock(&priv->mcast_mutex);
+	ipoib_dbg_mcast(priv, "flushing multicast list\n");
+
+	spin_lock_irqsave(&priv->lock, flags);
+
+	list_for_each_entry_safe(mcast, tmcast, &priv->multicast_list, list) {
+		list_del(&mcast->list);
+		rb_erase(&mcast->rb_node, &priv->multicast_tree);
+		list_add_tail(&mcast->list, &remove_list);
+	}
+
+	if (priv->broadcast) {
+		rb_erase(&priv->broadcast->rb_node, &priv->multicast_tree);
+		list_add_tail(&priv->broadcast->list, &remove_list);
+		priv->broadcast = NULL;
+	}
+
+	spin_unlock_irqrestore(&priv->lock, flags);
+
+	ipoib_mcast_remove_list(&remove_list);
+	mutex_unlock(&priv->mcast_mutex);
+}
+
+static int ipoib_mcast_addr_is_valid(const u8 *addr, const u8 *broadcast)
+{
+	/* reserved QPN, prefix, scope */
+	if (memcmp(addr, broadcast, 6))
+		return 0;
+	/* signature lower, pkey */
+	if (memcmp(addr + 7, broadcast + 7, 3))
+		return 0;
+	return 1;
+}
+
+void ipoib_mcast_restart_task(struct work_struct *work)
+{
+	struct ipoib_dev_priv *priv =
+		container_of(work, struct ipoib_dev_priv, restart_task);
+	struct net_device *dev = priv->dev;
+	struct netdev_hw_addr *ha;
+	struct ipoib_mcast *mcast, *tmcast;
+	LIST_HEAD(remove_list);
+	struct ib_sa_mcmember_rec rec;
+
+	if (!test_bit(IPOIB_FLAG_OPER_UP, &priv->flags))
+		/*
+		 * shortcut...on shutdown flush is called next, just
+		 * let it do all the work
+		 */
+		return;
+
+	ipoib_dbg_mcast(priv, "restarting multicast task\n");
+
+	netif_addr_lock_bh(dev);
+	spin_lock_irq(&priv->lock);
+
+	/*
+	 * Unfortunately, the networking core only gives us a list of all of
+	 * the multicast hardware addresses. We need to figure out which ones
+	 * are new and which ones have been removed
+	 */
+
+	/* Clear out the found flag */
+	list_for_each_entry(mcast, &priv->multicast_list, list)
+		clear_bit(IPOIB_MCAST_FLAG_FOUND, &mcast->flags);
+
+	/* Mark all of the entries that are found or don't exist */
+	netdev_for_each_mc_addr(ha, dev) {
+		union ib_gid mgid;
+
+		if (!ipoib_mcast_addr_is_valid(ha->addr, dev->broadcast))
+			continue;
+
+		memcpy(mgid.raw, ha->addr + 4, sizeof(mgid));
+
+		mcast = __ipoib_mcast_find(dev, &mgid);
+		if (!mcast || test_bit(IPOIB_MCAST_FLAG_SENDONLY, &mcast->flags)) {
+			struct ipoib_mcast *nmcast;
+
+			/* ignore group which is directly joined by userspace */
+			if (test_bit(IPOIB_FLAG_UMCAST, &priv->flags) &&
+			    !ib_sa_get_mcmember_rec(priv->ca, priv->port, &mgid, &rec)) {
+				ipoib_dbg_mcast(priv, "ignoring multicast entry for mgid %pI6\n",
+						mgid.raw);
+				continue;
+			}
+
+			/* Not found or send-only group, let's add a new entry */
+			ipoib_dbg_mcast(priv, "adding multicast entry for mgid %pI6\n",
+					mgid.raw);
+
+			nmcast = ipoib_mcast_alloc(dev);
+			if (!nmcast) {
+				ipoib_warn(priv, "unable to allocate memory for multicast structure\n");
+				continue;
+			}
+
+			set_bit(IPOIB_MCAST_FLAG_FOUND, &nmcast->flags);
+
+			nmcast->mcmember.mgid = mgid;
+
+			if (mcast) {
+				/* Destroy the send only entry */
+				list_move_tail(&mcast->list, &remove_list);
+
+				rb_replace_node(&mcast->rb_node,
+						&nmcast->rb_node,
+						&priv->multicast_tree);
+			} else
+				__ipoib_mcast_add(dev, nmcast);
+
+			list_add_tail(&nmcast->list, &priv->multicast_list);
+		}
+
+		if (mcast)
+			set_bit(IPOIB_MCAST_FLAG_FOUND, &mcast->flags);
+	}
+
+	/* Remove all of the entries don't exist anymore */
+	list_for_each_entry_safe(mcast, tmcast, &priv->multicast_list, list) {
+		if (!test_bit(IPOIB_MCAST_FLAG_FOUND, &mcast->flags) &&
+		    !test_bit(IPOIB_MCAST_FLAG_SENDONLY, &mcast->flags)) {
+			ipoib_dbg_mcast(priv, "deleting multicast group %pI6\n",
+					mcast->mcmember.mgid.raw);
+
+			rb_erase(&mcast->rb_node, &priv->multicast_tree);
+
+			/* Move to the remove list */
+			list_move_tail(&mcast->list, &remove_list);
+		}
+	}
+
+	spin_unlock_irq(&priv->lock);
+	netif_addr_unlock_bh(dev);
+
+	ipoib_mcast_remove_list(&remove_list);
+
+	/*
+	 * Double check that we are still up
+	 */
+	if (test_bit(IPOIB_FLAG_OPER_UP, &priv->flags)) {
+		spin_lock_irq(&priv->lock);
+		__ipoib_mcast_schedule_join_thread(priv, NULL, 0);
+		spin_unlock_irq(&priv->lock);
+	}
+}
+
+#ifdef CONFIG_INFINIBAND_IPOIB_DEBUG
+
+struct ipoib_mcast_iter *ipoib_mcast_iter_init(struct net_device *dev)
+{
+	struct ipoib_mcast_iter *iter;
+
+	iter = kmalloc(sizeof(*iter), GFP_KERNEL);
+	if (!iter)
+		return NULL;
+
+	iter->dev = dev;
+	memset(iter->mgid.raw, 0, 16);
+
+	if (ipoib_mcast_iter_next(iter)) {
+		kfree(iter);
+		return NULL;
+	}
+
+	return iter;
+}
+
+int ipoib_mcast_iter_next(struct ipoib_mcast_iter *iter)
+{
+	struct ipoib_dev_priv *priv = ipoib_priv(iter->dev);
+	struct rb_node *n;
+	struct ipoib_mcast *mcast;
+	int ret = 1;
+
+	spin_lock_irq(&priv->lock);
+
+	n = rb_first(&priv->multicast_tree);
+
+	while (n) {
+		mcast = rb_entry(n, struct ipoib_mcast, rb_node);
+
+		if (memcmp(iter->mgid.raw, mcast->mcmember.mgid.raw,
+			   sizeof (union ib_gid)) < 0) {
+			iter->mgid      = mcast->mcmember.mgid;
+			iter->created   = mcast->created;
+			iter->queuelen  = skb_queue_len(&mcast->pkt_queue);
+			iter->complete  = !!mcast->ah;
+			iter->send_only = !!(mcast->flags & (1 << IPOIB_MCAST_FLAG_SENDONLY));
+
+			ret = 0;
+
+			break;
+		}
+
+		n = rb_next(n);
+	}
+
+	spin_unlock_irq(&priv->lock);
+
+	return ret;
+}
+
+void ipoib_mcast_iter_read(struct ipoib_mcast_iter *iter,
+			   union ib_gid *mgid,
+			   unsigned long *created,
+			   unsigned int *queuelen,
+			   unsigned int *complete,
+			   unsigned int *send_only)
+{
+	*mgid      = iter->mgid;
+	*created   = iter->created;
+	*queuelen  = iter->queuelen;
+	*complete  = iter->complete;
+	*send_only = iter->send_only;
+}
+
+#endif /* CONFIG_INFINIBAND_IPOIB_DEBUG */
diff --git a/drivers/infiniband/ulp/ipoib/ipoib_netlink.c b/drivers/infiniband/ulp/ipoib/ipoib_netlink.c
new file mode 100644
index 000000000..9ad8d9856
--- /dev/null
+++ b/drivers/infiniband/ulp/ipoib/ipoib_netlink.c
@@ -0,0 +1,199 @@
+/*
+ * Copyright (c) 2012 Mellanox Technologies. -  All rights reserved.
+ *
+ * This software is available to you under a choice of one of two
+ * licenses.  You may choose to be licensed under the terms of the GNU
+ * General Public License (GPL) Version 2, available from the file
+ * COPYING in the main directory of this source tree, or the
+ * OpenIB.org BSD license below:
+ *
+ *     Redistribution and use in source and binary forms, with or
+ *     without modification, are permitted provided that the following
+ *     conditions are met:
+ *
+ *      - Redistributions of source code must retain the above
+ *        copyright notice, this list of conditions and the following
+ *        disclaimer.
+ *
+ *      - Redistributions in binary form must reproduce the above
+ *        copyright notice, this list of conditions and the following
+ *        disclaimer in the documentation and/or other materials
+ *        provided with the distribution.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
+ * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
+ * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS
+ * BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN
+ * ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
+ * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+ * SOFTWARE.
+ */
+
+#include <linux/netdevice.h>
+#include <linux/if_arp.h>      /* For ARPHRD_xxx */
+#include <net/rtnetlink.h>
+#include "ipoib.h"
+
+static const struct nla_policy ipoib_policy[IFLA_IPOIB_MAX + 1] = {
+	[IFLA_IPOIB_PKEY]	= { .type = NLA_U16 },
+	[IFLA_IPOIB_MODE]	= { .type = NLA_U16 },
+	[IFLA_IPOIB_UMCAST]	= { .type = NLA_U16 },
+};
+
+static unsigned int ipoib_get_max_num_queues(void)
+{
+	return min_t(unsigned int, num_possible_cpus(), 128);
+}
+
+static int ipoib_fill_info(struct sk_buff *skb, const struct net_device *dev)
+{
+	struct ipoib_dev_priv *priv = ipoib_priv(dev);
+	u16 val;
+
+	if (nla_put_u16(skb, IFLA_IPOIB_PKEY, priv->pkey))
+		goto nla_put_failure;
+
+	val = test_bit(IPOIB_FLAG_ADMIN_CM, &priv->flags);
+	if (nla_put_u16(skb, IFLA_IPOIB_MODE, val))
+		goto nla_put_failure;
+
+	val = test_bit(IPOIB_FLAG_UMCAST, &priv->flags);
+	if (nla_put_u16(skb, IFLA_IPOIB_UMCAST, val))
+		goto nla_put_failure;
+
+	return 0;
+
+nla_put_failure:
+	return -EMSGSIZE;
+}
+
+static int ipoib_changelink(struct net_device *dev, struct nlattr *tb[],
+			    struct nlattr *data[],
+			    struct netlink_ext_ack *extack)
+{
+	u16 mode, umcast;
+	int ret = 0;
+
+	if (data[IFLA_IPOIB_MODE]) {
+		mode  = nla_get_u16(data[IFLA_IPOIB_MODE]);
+		if (mode == IPOIB_MODE_DATAGRAM)
+			ret = ipoib_set_mode(dev, "datagram\n");
+		else if (mode == IPOIB_MODE_CONNECTED)
+			ret = ipoib_set_mode(dev, "connected\n");
+		else
+			ret = -EINVAL;
+
+		if (ret < 0)
+			goto out_err;
+	}
+
+	if (data[IFLA_IPOIB_UMCAST]) {
+		umcast = nla_get_u16(data[IFLA_IPOIB_UMCAST]);
+		ipoib_set_umcast(dev, umcast);
+	}
+
+out_err:
+	return ret;
+}
+
+static int ipoib_new_child_link(struct net *src_net, struct net_device *dev,
+				struct nlattr *tb[], struct nlattr *data[],
+				struct netlink_ext_ack *extack)
+{
+	struct net_device *pdev;
+	struct ipoib_dev_priv *ppriv;
+	u16 child_pkey;
+	int err;
+
+	if (!tb[IFLA_LINK])
+		return -EINVAL;
+
+	pdev = __dev_get_by_index(src_net, nla_get_u32(tb[IFLA_LINK]));
+	if (!pdev || pdev->type != ARPHRD_INFINIBAND)
+		return -ENODEV;
+
+	ppriv = ipoib_priv(pdev);
+
+	if (test_bit(IPOIB_FLAG_SUBINTERFACE, &ppriv->flags)) {
+		ipoib_warn(ppriv, "child creation disallowed for child devices\n");
+		return -EINVAL;
+	}
+
+	if (!data || !data[IFLA_IPOIB_PKEY]) {
+		ipoib_dbg(ppriv, "no pkey specified, using parent pkey\n");
+		child_pkey  = ppriv->pkey;
+	} else
+		child_pkey  = nla_get_u16(data[IFLA_IPOIB_PKEY]);
+
+	err = ipoib_intf_init(ppriv->ca, ppriv->port, dev->name, dev);
+	if (err) {
+		ipoib_warn(ppriv, "failed to initialize pkey device\n");
+		return err;
+	}
+
+	err = __ipoib_vlan_add(ppriv, ipoib_priv(dev),
+			       child_pkey, IPOIB_RTNL_CHILD);
+	if (err)
+		return err;
+
+	if (data) {
+		err = ipoib_changelink(dev, tb, data, extack);
+		if (err) {
+			unregister_netdevice(dev);
+			return err;
+		}
+	}
+
+	return 0;
+}
+
+static void ipoib_del_child_link(struct net_device *dev, struct list_head *head)
+{
+	struct ipoib_dev_priv *priv = ipoib_priv(dev);
+
+	if (!priv->parent)
+		return;
+
+	unregister_netdevice_queue(dev, head);
+}
+
+static size_t ipoib_get_size(const struct net_device *dev)
+{
+	return nla_total_size(2) +	/* IFLA_IPOIB_PKEY   */
+		nla_total_size(2) +	/* IFLA_IPOIB_MODE   */
+		nla_total_size(2);	/* IFLA_IPOIB_UMCAST */
+}
+
+static struct rtnl_link_ops ipoib_link_ops __read_mostly = {
+	.kind		= "ipoib",
+	.netns_refund   = true,
+	.maxtype	= IFLA_IPOIB_MAX,
+	.policy		= ipoib_policy,
+	.priv_size	= sizeof(struct ipoib_dev_priv),
+	.setup		= ipoib_setup_common,
+	.newlink	= ipoib_new_child_link,
+	.dellink	= ipoib_del_child_link,
+	.changelink	= ipoib_changelink,
+	.get_size	= ipoib_get_size,
+	.fill_info	= ipoib_fill_info,
+	.get_num_rx_queues = ipoib_get_max_num_queues,
+	.get_num_tx_queues = ipoib_get_max_num_queues,
+};
+
+struct rtnl_link_ops *ipoib_get_link_ops(void)
+{
+	return &ipoib_link_ops;
+}
+
+int __init ipoib_netlink_init(void)
+{
+	return rtnl_link_register(&ipoib_link_ops);
+}
+
+void __exit ipoib_netlink_fini(void)
+{
+	rtnl_link_unregister(&ipoib_link_ops);
+}
+
+MODULE_ALIAS_RTNL_LINK("ipoib");
diff --git a/drivers/infiniband/ulp/ipoib/ipoib_verbs.c b/drivers/infiniband/ulp/ipoib/ipoib_verbs.c
new file mode 100644
index 000000000..368e5d774
--- /dev/null
+++ b/drivers/infiniband/ulp/ipoib/ipoib_verbs.c
@@ -0,0 +1,294 @@
+/*
+ * Copyright (c) 2004, 2005 Topspin Communications.  All rights reserved.
+ * Copyright (c) 2005 Mellanox Technologies. All rights reserved.
+ *
+ * This software is available to you under a choice of one of two
+ * licenses.  You may choose to be licensed under the terms of the GNU
+ * General Public License (GPL) Version 2, available from the file
+ * COPYING in the main directory of this source tree, or the
+ * OpenIB.org BSD license below:
+ *
+ *     Redistribution and use in source and binary forms, with or
+ *     without modification, are permitted provided that the following
+ *     conditions are met:
+ *
+ *      - Redistributions of source code must retain the above
+ *        copyright notice, this list of conditions and the following
+ *        disclaimer.
+ *
+ *      - Redistributions in binary form must reproduce the above
+ *        copyright notice, this list of conditions and the following
+ *        disclaimer in the documentation and/or other materials
+ *        provided with the distribution.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
+ * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
+ * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS
+ * BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN
+ * ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
+ * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+ * SOFTWARE.
+ */
+
+#include <linux/slab.h>
+
+#include "ipoib.h"
+
+int ipoib_mcast_attach(struct net_device *dev, struct ib_device *hca,
+		       union ib_gid *mgid, u16 mlid, int set_qkey, u32 qkey)
+{
+	struct ipoib_dev_priv *priv = ipoib_priv(dev);
+	struct ib_qp_attr *qp_attr = NULL;
+	int ret;
+	u16 pkey_index;
+
+	if (ib_find_pkey(priv->ca, priv->port, priv->pkey, &pkey_index)) {
+		clear_bit(IPOIB_PKEY_ASSIGNED, &priv->flags);
+		ret = -ENXIO;
+		goto out;
+	}
+	set_bit(IPOIB_PKEY_ASSIGNED, &priv->flags);
+
+	if (set_qkey) {
+		ret = -ENOMEM;
+		qp_attr = kmalloc(sizeof(*qp_attr), GFP_KERNEL);
+		if (!qp_attr)
+			goto out;
+
+		/* set correct QKey for QP */
+		qp_attr->qkey = qkey;
+		ret = ib_modify_qp(priv->qp, qp_attr, IB_QP_QKEY);
+		if (ret) {
+			ipoib_warn(priv, "failed to modify QP, ret = %d\n", ret);
+			goto out;
+		}
+	}
+
+	/* attach QP to multicast group */
+	ret = ib_attach_mcast(priv->qp, mgid, mlid);
+	if (ret)
+		ipoib_warn(priv, "failed to attach to multicast group, ret = %d\n", ret);
+
+out:
+	kfree(qp_attr);
+	return ret;
+}
+
+int ipoib_mcast_detach(struct net_device *dev, struct ib_device *hca,
+		       union ib_gid *mgid, u16 mlid)
+{
+	struct ipoib_dev_priv *priv = ipoib_priv(dev);
+	int ret;
+
+	ret = ib_detach_mcast(priv->qp, mgid, mlid);
+
+	return ret;
+}
+
+int ipoib_init_qp(struct net_device *dev)
+{
+	struct ipoib_dev_priv *priv = ipoib_priv(dev);
+	int ret;
+	struct ib_qp_attr qp_attr;
+	int attr_mask;
+
+	if (!test_bit(IPOIB_PKEY_ASSIGNED, &priv->flags))
+		return -1;
+
+	qp_attr.qp_state = IB_QPS_INIT;
+	qp_attr.qkey = 0;
+	qp_attr.port_num = priv->port;
+	qp_attr.pkey_index = priv->pkey_index;
+	attr_mask =
+	    IB_QP_QKEY |
+	    IB_QP_PORT |
+	    IB_QP_PKEY_INDEX |
+	    IB_QP_STATE;
+	ret = ib_modify_qp(priv->qp, &qp_attr, attr_mask);
+	if (ret) {
+		ipoib_warn(priv, "failed to modify QP to init, ret = %d\n", ret);
+		goto out_fail;
+	}
+
+	qp_attr.qp_state = IB_QPS_RTR;
+	/* Can't set this in a INIT->RTR transition */
+	attr_mask &= ~IB_QP_PORT;
+	ret = ib_modify_qp(priv->qp, &qp_attr, attr_mask);
+	if (ret) {
+		ipoib_warn(priv, "failed to modify QP to RTR, ret = %d\n", ret);
+		goto out_fail;
+	}
+
+	qp_attr.qp_state = IB_QPS_RTS;
+	qp_attr.sq_psn = 0;
+	attr_mask |= IB_QP_SQ_PSN;
+	attr_mask &= ~IB_QP_PKEY_INDEX;
+	ret = ib_modify_qp(priv->qp, &qp_attr, attr_mask);
+	if (ret) {
+		ipoib_warn(priv, "failed to modify QP to RTS, ret = %d\n", ret);
+		goto out_fail;
+	}
+
+	return 0;
+
+out_fail:
+	qp_attr.qp_state = IB_QPS_RESET;
+	if (ib_modify_qp(priv->qp, &qp_attr, IB_QP_STATE))
+		ipoib_warn(priv, "Failed to modify QP to RESET state\n");
+
+	return ret;
+}
+
+int ipoib_transport_dev_init(struct net_device *dev, struct ib_device *ca)
+{
+	struct ipoib_dev_priv *priv = ipoib_priv(dev);
+	struct ib_qp_init_attr init_attr = {
+		.cap = {
+			.max_send_wr  = ipoib_sendq_size,
+			.max_recv_wr  = ipoib_recvq_size,
+			.max_send_sge = min_t(u32, priv->ca->attrs.max_send_sge,
+					      MAX_SKB_FRAGS + 1),
+			.max_recv_sge = IPOIB_UD_RX_SG
+		},
+		.sq_sig_type = IB_SIGNAL_ALL_WR,
+		.qp_type     = IB_QPT_UD
+	};
+	struct ib_cq_init_attr cq_attr = {};
+
+	int ret, size, req_vec;
+	int i;
+	static atomic_t counter;
+
+	size = ipoib_recvq_size + 1;
+	ret = ipoib_cm_dev_init(dev);
+	if (!ret) {
+		size += ipoib_sendq_size;
+		if (ipoib_cm_has_srq(dev))
+			size += ipoib_recvq_size + 1; /* 1 extra for rx_drain_qp */
+		else
+			size += ipoib_recvq_size * ipoib_max_conn_qp;
+	} else
+		if (ret != -EOPNOTSUPP)
+			return ret;
+
+	req_vec = atomic_inc_return(&counter) * 2;
+	cq_attr.cqe = size;
+	cq_attr.comp_vector = req_vec % priv->ca->num_comp_vectors;
+	priv->recv_cq = ib_create_cq(priv->ca, ipoib_ib_rx_completion, NULL,
+				     priv, &cq_attr);
+	if (IS_ERR(priv->recv_cq)) {
+		pr_warn("%s: failed to create receive CQ\n", ca->name);
+		goto out_cm_dev_cleanup;
+	}
+
+	cq_attr.cqe = ipoib_sendq_size;
+	cq_attr.comp_vector = (req_vec + 1) % priv->ca->num_comp_vectors;
+	priv->send_cq = ib_create_cq(priv->ca, ipoib_ib_tx_completion, NULL,
+				     priv, &cq_attr);
+	if (IS_ERR(priv->send_cq)) {
+		pr_warn("%s: failed to create send CQ\n", ca->name);
+		goto out_free_recv_cq;
+	}
+
+	if (ib_req_notify_cq(priv->recv_cq, IB_CQ_NEXT_COMP))
+		goto out_free_send_cq;
+
+	init_attr.send_cq = priv->send_cq;
+	init_attr.recv_cq = priv->recv_cq;
+
+	if (priv->kernel_caps & IBK_UD_TSO)
+		init_attr.create_flags |= IB_QP_CREATE_IPOIB_UD_LSO;
+
+	if (priv->kernel_caps & IBK_BLOCK_MULTICAST_LOOPBACK)
+		init_attr.create_flags |= IB_QP_CREATE_BLOCK_MULTICAST_LOOPBACK;
+
+	if (priv->hca_caps & IB_DEVICE_MANAGED_FLOW_STEERING)
+		init_attr.create_flags |= IB_QP_CREATE_NETIF_QP;
+
+	if (priv->kernel_caps & IBK_RDMA_NETDEV_OPA)
+		init_attr.create_flags |= IB_QP_CREATE_NETDEV_USE;
+
+	priv->qp = ib_create_qp(priv->pd, &init_attr);
+	if (IS_ERR(priv->qp)) {
+		pr_warn("%s: failed to create QP\n", ca->name);
+		goto out_free_send_cq;
+	}
+
+	if (ib_req_notify_cq(priv->send_cq, IB_CQ_NEXT_COMP))
+		goto out_free_send_cq;
+
+	for (i = 0; i < MAX_SKB_FRAGS + 1; ++i)
+		priv->tx_sge[i].lkey = priv->pd->local_dma_lkey;
+
+	priv->tx_wr.wr.opcode		= IB_WR_SEND;
+	priv->tx_wr.wr.sg_list		= priv->tx_sge;
+	priv->tx_wr.wr.send_flags	= IB_SEND_SIGNALED;
+
+	priv->rx_sge[0].lkey = priv->pd->local_dma_lkey;
+
+	priv->rx_sge[0].length = IPOIB_UD_BUF_SIZE(priv->max_ib_mtu);
+	priv->rx_wr.num_sge = 1;
+
+	priv->rx_wr.next = NULL;
+	priv->rx_wr.sg_list = priv->rx_sge;
+
+	if (init_attr.cap.max_send_sge > 1)
+		dev->features |= NETIF_F_SG;
+
+	priv->max_send_sge = init_attr.cap.max_send_sge;
+
+	return 0;
+
+out_free_send_cq:
+	ib_destroy_cq(priv->send_cq);
+
+out_free_recv_cq:
+	ib_destroy_cq(priv->recv_cq);
+
+out_cm_dev_cleanup:
+	ipoib_cm_dev_cleanup(dev);
+
+	return -ENODEV;
+}
+
+void ipoib_transport_dev_cleanup(struct net_device *dev)
+{
+	struct ipoib_dev_priv *priv = ipoib_priv(dev);
+
+	if (priv->qp) {
+		if (ib_destroy_qp(priv->qp))
+			ipoib_warn(priv, "ib_qp_destroy failed\n");
+
+		priv->qp = NULL;
+	}
+
+	ib_destroy_cq(priv->send_cq);
+	ib_destroy_cq(priv->recv_cq);
+}
+
+void ipoib_event(struct ib_event_handler *handler,
+		 struct ib_event *record)
+{
+	struct ipoib_dev_priv *priv =
+		container_of(handler, struct ipoib_dev_priv, event_handler);
+
+	if (record->element.port_num != priv->port)
+		return;
+
+	ipoib_dbg(priv, "Event %d on device %s port %d\n", record->event,
+		  dev_name(&record->device->dev), record->element.port_num);
+
+	if (record->event == IB_EVENT_CLIENT_REREGISTER) {
+		queue_work(ipoib_workqueue, &priv->flush_light);
+	} else if (record->event == IB_EVENT_PORT_ERR ||
+		   record->event == IB_EVENT_PORT_ACTIVE ||
+		   record->event == IB_EVENT_LID_CHANGE) {
+		queue_work(ipoib_workqueue, &priv->flush_normal);
+	} else if (record->event == IB_EVENT_PKEY_CHANGE) {
+		queue_work(ipoib_workqueue, &priv->flush_heavy);
+	} else if (record->event == IB_EVENT_GID_CHANGE &&
+		   !test_bit(IPOIB_FLAG_DEV_ADDR_SET, &priv->flags)) {
+		queue_work(ipoib_workqueue, &priv->flush_light);
+	}
+}
diff --git a/drivers/infiniband/ulp/ipoib/ipoib_vlan.c b/drivers/infiniband/ulp/ipoib/ipoib_vlan.c
new file mode 100644
index 000000000..4bd161e86
--- /dev/null
+++ b/drivers/infiniband/ulp/ipoib/ipoib_vlan.c
@@ -0,0 +1,293 @@
+/*
+ * Copyright (c) 2004 Topspin Communications.  All rights reserved.
+ *
+ * This software is available to you under a choice of one of two
+ * licenses.  You may choose to be licensed under the terms of the GNU
+ * General Public License (GPL) Version 2, available from the file
+ * COPYING in the main directory of this source tree, or the
+ * OpenIB.org BSD license below:
+ *
+ *     Redistribution and use in source and binary forms, with or
+ *     without modification, are permitted provided that the following
+ *     conditions are met:
+ *
+ *      - Redistributions of source code must retain the above
+ *        copyright notice, this list of conditions and the following
+ *        disclaimer.
+ *
+ *      - Redistributions in binary form must reproduce the above
+ *        copyright notice, this list of conditions and the following
+ *        disclaimer in the documentation and/or other materials
+ *        provided with the distribution.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
+ * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
+ * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS
+ * BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN
+ * ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
+ * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+ * SOFTWARE.
+ */
+
+#include <linux/sched/signal.h>
+
+#include <linux/init.h>
+#include <linux/seq_file.h>
+
+#include <linux/uaccess.h>
+
+#include "ipoib.h"
+
+static ssize_t parent_show(struct device *d, struct device_attribute *attr,
+			   char *buf)
+{
+	struct net_device *dev = to_net_dev(d);
+	struct ipoib_dev_priv *priv = ipoib_priv(dev);
+
+	return sysfs_emit(buf, "%s\n", priv->parent->name);
+}
+static DEVICE_ATTR_RO(parent);
+
+static bool is_child_unique(struct ipoib_dev_priv *ppriv,
+			    struct ipoib_dev_priv *priv)
+{
+	struct ipoib_dev_priv *tpriv;
+
+	ASSERT_RTNL();
+
+	/*
+	 * Since the legacy sysfs interface uses pkey for deletion it cannot
+	 * support more than one interface with the same pkey, it creates
+	 * ambiguity.  The RTNL interface deletes using the netdev so it does
+	 * not have a problem to support duplicated pkeys.
+	 */
+	if (priv->child_type != IPOIB_LEGACY_CHILD)
+		return true;
+
+	/*
+	 * First ensure this isn't a duplicate. We check the parent device and
+	 * then all of the legacy child interfaces to make sure the Pkey
+	 * doesn't match.
+	 */
+	if (ppriv->pkey == priv->pkey)
+		return false;
+
+	list_for_each_entry(tpriv, &ppriv->child_intfs, list) {
+		if (tpriv->pkey == priv->pkey &&
+		    tpriv->child_type == IPOIB_LEGACY_CHILD)
+			return false;
+	}
+
+	return true;
+}
+
+/*
+ * NOTE: If this function fails then the priv->dev will remain valid, however
+ * priv will have been freed and must not be touched by caller in the error
+ * case.
+ *
+ * If (ndev->reg_state == NETREG_UNINITIALIZED) then it is up to the caller to
+ * free the net_device (just as rtnl_newlink does) otherwise the net_device
+ * will be freed when the rtnl is unlocked.
+ */
+int __ipoib_vlan_add(struct ipoib_dev_priv *ppriv, struct ipoib_dev_priv *priv,
+		     u16 pkey, int type)
+{
+	struct net_device *ndev = priv->dev;
+	int result;
+	struct rdma_netdev *rn = netdev_priv(ndev);
+
+	ASSERT_RTNL();
+
+	/*
+	 * We do not need to touch priv if register_netdevice fails, so just
+	 * always use this flow.
+	 */
+	ndev->priv_destructor = ipoib_intf_free;
+
+	/*
+	 * Racing with unregister of the parent must be prevented by the
+	 * caller.
+	 */
+	WARN_ON(ppriv->dev->reg_state != NETREG_REGISTERED);
+
+	if (pkey == 0 || pkey == 0x8000) {
+		result = -EINVAL;
+		goto out_early;
+	}
+
+	rn->mtu = priv->mcast_mtu;
+
+	priv->parent = ppriv->dev;
+	priv->pkey = pkey;
+	priv->child_type = type;
+
+	if (!is_child_unique(ppriv, priv)) {
+		result = -ENOTUNIQ;
+		goto out_early;
+	}
+
+	result = register_netdevice(ndev);
+	if (result) {
+		ipoib_warn(priv, "failed to initialize; error %i", result);
+
+		/*
+		 * register_netdevice sometimes calls priv_destructor,
+		 * sometimes not. Make sure it was done.
+		 */
+		goto out_early;
+	}
+
+	/* RTNL childs don't need proprietary sysfs entries */
+	if (type == IPOIB_LEGACY_CHILD) {
+		if (ipoib_cm_add_mode_attr(ndev))
+			goto sysfs_failed;
+		if (ipoib_add_pkey_attr(ndev))
+			goto sysfs_failed;
+		if (ipoib_add_umcast_attr(ndev))
+			goto sysfs_failed;
+
+		if (device_create_file(&ndev->dev, &dev_attr_parent))
+			goto sysfs_failed;
+	}
+
+	return 0;
+
+sysfs_failed:
+	unregister_netdevice(priv->dev);
+	return -ENOMEM;
+
+out_early:
+	if (ndev->priv_destructor)
+		ndev->priv_destructor(ndev);
+	return result;
+}
+
+int ipoib_vlan_add(struct net_device *pdev, unsigned short pkey)
+{
+	struct ipoib_dev_priv *ppriv, *priv;
+	char intf_name[IFNAMSIZ];
+	struct net_device *ndev;
+	int result;
+
+	if (!capable(CAP_NET_ADMIN))
+		return -EPERM;
+
+	if (!rtnl_trylock())
+		return restart_syscall();
+
+	if (pdev->reg_state != NETREG_REGISTERED) {
+		rtnl_unlock();
+		return -EPERM;
+	}
+
+	ppriv = ipoib_priv(pdev);
+
+	snprintf(intf_name, sizeof(intf_name), "%s.%04x",
+		 ppriv->dev->name, pkey);
+
+	ndev = ipoib_intf_alloc(ppriv->ca, ppriv->port, intf_name);
+	if (IS_ERR(ndev)) {
+		result = PTR_ERR(ndev);
+		goto out;
+	}
+	priv = ipoib_priv(ndev);
+
+	ndev->rtnl_link_ops = ipoib_get_link_ops();
+
+	result = __ipoib_vlan_add(ppriv, priv, pkey, IPOIB_LEGACY_CHILD);
+
+	if (result && ndev->reg_state == NETREG_UNINITIALIZED)
+		free_netdev(ndev);
+
+out:
+	rtnl_unlock();
+
+	return result;
+}
+
+struct ipoib_vlan_delete_work {
+	struct work_struct work;
+	struct net_device *dev;
+};
+
+/*
+ * sysfs callbacks of a netdevice cannot obtain the rtnl lock as
+ * unregister_netdev ultimately deletes the sysfs files while holding the rtnl
+ * lock. This deadlocks the system.
+ *
+ * A callback can use rtnl_trylock to avoid the deadlock but it cannot call
+ * unregister_netdev as that internally takes and releases the rtnl_lock.  So
+ * instead we find the netdev to unregister and then do the actual unregister
+ * from the global work queue where we can obtain the rtnl_lock safely.
+ */
+static void ipoib_vlan_delete_task(struct work_struct *work)
+{
+	struct ipoib_vlan_delete_work *pwork =
+		container_of(work, struct ipoib_vlan_delete_work, work);
+	struct net_device *dev = pwork->dev;
+
+	rtnl_lock();
+
+	/* Unregistering tasks can race with another task or parent removal */
+	if (dev->reg_state == NETREG_REGISTERED) {
+		struct ipoib_dev_priv *priv = ipoib_priv(dev);
+		struct ipoib_dev_priv *ppriv = ipoib_priv(priv->parent);
+
+		ipoib_dbg(ppriv, "delete child vlan %s\n", dev->name);
+		unregister_netdevice(dev);
+	}
+
+	rtnl_unlock();
+
+	kfree(pwork);
+}
+
+int ipoib_vlan_delete(struct net_device *pdev, unsigned short pkey)
+{
+	struct ipoib_dev_priv *ppriv, *priv, *tpriv;
+	int rc;
+
+	if (!capable(CAP_NET_ADMIN))
+		return -EPERM;
+
+	if (!rtnl_trylock())
+		return restart_syscall();
+
+	if (pdev->reg_state != NETREG_REGISTERED) {
+		rtnl_unlock();
+		return -EPERM;
+	}
+
+	ppriv = ipoib_priv(pdev);
+
+	rc = -ENODEV;
+	list_for_each_entry_safe(priv, tpriv, &ppriv->child_intfs, list) {
+		if (priv->pkey == pkey &&
+		    priv->child_type == IPOIB_LEGACY_CHILD) {
+			struct ipoib_vlan_delete_work *work;
+
+			work = kmalloc(sizeof(*work), GFP_KERNEL);
+			if (!work) {
+				rc = -ENOMEM;
+				goto out;
+			}
+
+			down_write(&ppriv->vlan_rwsem);
+			list_del_init(&priv->list);
+			up_write(&ppriv->vlan_rwsem);
+			work->dev = priv->dev;
+			INIT_WORK(&work->work, ipoib_vlan_delete_task);
+			queue_work(ipoib_workqueue, &work->work);
+
+			rc = 0;
+			break;
+		}
+	}
+
+out:
+	rtnl_unlock();
+
+	return rc;
+}