Merge Linus' tree to be be to apply submitted patches to newer code than

current trivial.git base

65 files changed, 3091 insertions, 1523 deletions

diff --git a/drivers/infiniband/core/agent.c b/drivers/infiniband/core/agent.c
index 2bc7f5af64f4..f6d29614cb01 100644
--- a/drivers/infiniband/core/agent.c
+++ b/drivers/infiniband/core/agent.c
@@ -94,14 +94,14 @@ void agent_send_response(struct ib_mad *mad, struct ib_grh *grh,
 		port_priv = ib_get_agent_port(device, port_num);
 
 	if (!port_priv) {
-		printk(KERN_ERR SPFX "Unable to find port agent\n");
+		dev_err(&device->dev, "Unable to find port agent\n");
 		return;
 	}
 
 	agent = port_priv->agent[qpn];
 	ah = ib_create_ah_from_wc(agent->qp->pd, wc, grh, port_num);
 	if (IS_ERR(ah)) {
-		printk(KERN_ERR SPFX "ib_create_ah_from_wc error %ld\n",
+		dev_err(&device->dev, "ib_create_ah_from_wc error %ld\n",
 			PTR_ERR(ah));
 		return;
 	}
@@ -110,7 +110,7 @@ void agent_send_response(struct ib_mad *mad, struct ib_grh *grh,
 				      IB_MGMT_MAD_HDR, IB_MGMT_MAD_DATA,
 				      GFP_KERNEL);
 	if (IS_ERR(send_buf)) {
-		printk(KERN_ERR SPFX "ib_create_send_mad error\n");
+		dev_err(&device->dev, "ib_create_send_mad error\n");
 		goto err1;
 	}
 
@@ -125,7 +125,7 @@ void agent_send_response(struct ib_mad *mad, struct ib_grh *grh,
 	}
 
 	if (ib_post_send_mad(send_buf, NULL)) {
-		printk(KERN_ERR SPFX "ib_post_send_mad error\n");
+		dev_err(&device->dev, "ib_post_send_mad error\n");
 		goto err2;
 	}
 	return;
@@ -151,7 +151,7 @@ int ib_agent_port_open(struct ib_device *device, int port_num)
 	/* Create new device info */
 	port_priv = kzalloc(sizeof *port_priv, GFP_KERNEL);
 	if (!port_priv) {
-		printk(KERN_ERR SPFX "No memory for ib_agent_port_private\n");
+		dev_err(&device->dev, "No memory for ib_agent_port_private\n");
 		ret = -ENOMEM;
 		goto error1;
 	}
@@ -161,7 +161,7 @@ int ib_agent_port_open(struct ib_device *device, int port_num)
 		port_priv->agent[0] = ib_register_mad_agent(device, port_num,
 							    IB_QPT_SMI, NULL, 0,
 							    &agent_send_handler,
-							    NULL, NULL);
+							    NULL, NULL, 0);
 		if (IS_ERR(port_priv->agent[0])) {
 			ret = PTR_ERR(port_priv->agent[0]);
 			goto error2;
@@ -172,7 +172,7 @@ int ib_agent_port_open(struct ib_device *device, int port_num)
 	port_priv->agent[1] = ib_register_mad_agent(device, port_num,
 						    IB_QPT_GSI, NULL, 0,
 						    &agent_send_handler,
-						    NULL, NULL);
+						    NULL, NULL, 0);
 	if (IS_ERR(port_priv->agent[1])) {
 		ret = PTR_ERR(port_priv->agent[1]);
 		goto error3;
@@ -202,7 +202,7 @@ int ib_agent_port_close(struct ib_device *device, int port_num)
 	port_priv = __ib_get_agent_port(device, port_num);
 	if (port_priv == NULL) {
 		spin_unlock_irqrestore(&ib_agent_port_list_lock, flags);
-		printk(KERN_ERR SPFX "Port %d not found\n", port_num);
+		dev_err(&device->dev, "Port %d not found\n", port_num);
 		return -ENODEV;
 	}
 	list_del(&port_priv->port_list);
diff --git a/drivers/infiniband/core/cm.c b/drivers/infiniband/core/cm.c
index c3239170d8b7..e28a494e2a3a 100644
--- a/drivers/infiniband/core/cm.c
+++ b/drivers/infiniband/core/cm.c
@@ -3753,7 +3753,7 @@ static void cm_add_one(struct ib_device *ib_device)
 	struct cm_port *port;
 	struct ib_mad_reg_req reg_req = {
 		.mgmt_class = IB_MGMT_CLASS_CM,
-		.mgmt_class_version = IB_CM_CLASS_VERSION
+		.mgmt_class_version = IB_CM_CLASS_VERSION,
 	};
 	struct ib_port_modify port_modify = {
 		.set_port_cap_mask = IB_PORT_CM_SUP
@@ -3801,7 +3801,8 @@ static void cm_add_one(struct ib_device *ib_device)
 							0,
 							cm_send_handler,
 							cm_recv_handler,
-							port);
+							port,
+							0);
 		if (IS_ERR(port->mad_agent))
 			goto error2;
 
diff --git a/drivers/infiniband/core/iwcm.c b/drivers/infiniband/core/iwcm.c
index 3d2e489ab732..ff9163dc1596 100644
--- a/drivers/infiniband/core/iwcm.c
+++ b/drivers/infiniband/core/iwcm.c
@@ -46,6 +46,7 @@
 #include <linux/completion.h>
 #include <linux/slab.h>
 #include <linux/module.h>
+#include <linux/sysctl.h>
 
 #include <rdma/iw_cm.h>
 #include <rdma/ib_addr.h>
@@ -65,6 +66,20 @@ struct iwcm_work {
 	struct list_head free_list;
 };
 
+static unsigned int default_backlog = 256;
+
+static struct ctl_table_header *iwcm_ctl_table_hdr;
+static struct ctl_table iwcm_ctl_table[] = {
+	{
+		.procname	= "default_backlog",
+		.data		= &default_backlog,
+		.maxlen		= sizeof(default_backlog),
+		.mode		= 0644,
+		.proc_handler	= proc_dointvec,
+	},
+	{ }
+};
+
 /*
  * The following services provide a mechanism for pre-allocating iwcm_work
  * elements.  The design pre-allocates them  based on the cm_id type:
@@ -425,6 +440,9 @@ int iw_cm_listen(struct iw_cm_id *cm_id, int backlog)
 
 	cm_id_priv = container_of(cm_id, struct iwcm_id_private, id);
 
+	if (!backlog)
+		backlog = default_backlog;
+
 	ret = alloc_work_entries(cm_id_priv, backlog);
 	if (ret)
 		return ret;
@@ -1030,11 +1048,20 @@ static int __init iw_cm_init(void)
 	if (!iwcm_wq)
 		return -ENOMEM;
 
+	iwcm_ctl_table_hdr = register_net_sysctl(&init_net, "net/iw_cm",
+						 iwcm_ctl_table);
+	if (!iwcm_ctl_table_hdr) {
+		pr_err("iw_cm: couldn't register sysctl paths\n");
+		destroy_workqueue(iwcm_wq);
+		return -ENOMEM;
+	}
+
 	return 0;
 }
 
 static void __exit iw_cm_cleanup(void)
 {
+	unregister_net_sysctl_table(iwcm_ctl_table_hdr);
 	destroy_workqueue(iwcm_wq);
 }
 
diff --git a/drivers/infiniband/core/mad.c b/drivers/infiniband/core/mad.c
index ab31f136d04b..74c30f4c557e 100644
--- a/drivers/infiniband/core/mad.c
+++ b/drivers/infiniband/core/mad.c
@@ -33,6 +33,9 @@
  * SOFTWARE.
  *
  */
+
+#define pr_fmt(fmt) KBUILD_MODNAME ": " fmt
+
 #include <linux/dma-mapping.h>
 #include <linux/slab.h>
 #include <linux/module.h>
@@ -195,7 +198,8 @@ struct ib_mad_agent *ib_register_mad_agent(struct ib_device *device,
 					   u8 rmpp_version,
 					   ib_mad_send_handler send_handler,
 					   ib_mad_recv_handler recv_handler,
-					   void *context)
+					   void *context,
+					   u32 registration_flags)
 {
 	struct ib_mad_port_private *port_priv;
 	struct ib_mad_agent *ret = ERR_PTR(-EINVAL);
@@ -211,68 +215,109 @@ struct ib_mad_agent *ib_register_mad_agent(struct ib_device *device,
 
 	/* Validate parameters */
 	qpn = get_spl_qp_index(qp_type);
-	if (qpn == -1)
+	if (qpn == -1) {
+		dev_notice(&device->dev,
+			   "ib_register_mad_agent: invalid QP Type %d\n",
+			   qp_type);
 		goto error1;
+	}
 
-	if (rmpp_version && rmpp_version != IB_MGMT_RMPP_VERSION)
+	if (rmpp_version && rmpp_version != IB_MGMT_RMPP_VERSION) {
+		dev_notice(&device->dev,
+			   "ib_register_mad_agent: invalid RMPP Version %u\n",
+			   rmpp_version);
 		goto error1;
+	}
 
 	/* Validate MAD registration request if supplied */
 	if (mad_reg_req) {
-		if (mad_reg_req->mgmt_class_version >= MAX_MGMT_VERSION)
+		if (mad_reg_req->mgmt_class_version >= MAX_MGMT_VERSION) {
+			dev_notice(&device->dev,
+				   "ib_register_mad_agent: invalid Class Version %u\n",
+				   mad_reg_req->mgmt_class_version);
 			goto error1;
-		if (!recv_handler)
+		}
+		if (!recv_handler) {
+			dev_notice(&device->dev,
+				   "ib_register_mad_agent: no recv_handler\n");
 			goto error1;
+		}
 		if (mad_reg_req->mgmt_class >= MAX_MGMT_CLASS) {
 			/*
 			 * IB_MGMT_CLASS_SUBN_DIRECTED_ROUTE is the only
 			 * one in this range currently allowed
 			 */
 			if (mad_reg_req->mgmt_class !=
-			    IB_MGMT_CLASS_SUBN_DIRECTED_ROUTE)
+			    IB_MGMT_CLASS_SUBN_DIRECTED_ROUTE) {
+				dev_notice(&device->dev,
+					   "ib_register_mad_agent: Invalid Mgmt Class 0x%x\n",
+					   mad_reg_req->mgmt_class);
 				goto error1;
+			}
 		} else if (mad_reg_req->mgmt_class == 0) {
 			/*
 			 * Class 0 is reserved in IBA and is used for
 			 * aliasing of IB_MGMT_CLASS_SUBN_DIRECTED_ROUTE
 			 */
+			dev_notice(&device->dev,
+				   "ib_register_mad_agent: Invalid Mgmt Class 0\n");
 			goto error1;
 		} else if (is_vendor_class(mad_reg_req->mgmt_class)) {
 			/*
 			 * If class is in "new" vendor range,
 			 * ensure supplied OUI is not zero
 			 */
-			if (!is_vendor_oui(mad_reg_req->oui))
+			if (!is_vendor_oui(mad_reg_req->oui)) {
+				dev_notice(&device->dev,
+					   "ib_register_mad_agent: No OUI specified for class 0x%x\n",
+					   mad_reg_req->mgmt_class);
 				goto error1;
+			}
 		}
 		/* Make sure class supplied is consistent with RMPP */
 		if (!ib_is_mad_class_rmpp(mad_reg_req->mgmt_class)) {
-			if (rmpp_version)
+			if (rmpp_version) {
+				dev_notice(&device->dev,
+					   "ib_register_mad_agent: RMPP version for non-RMPP class 0x%x\n",
+					   mad_reg_req->mgmt_class);
 				goto error1;
+			}
 		}
+
 		/* Make sure class supplied is consistent with QP type */
 		if (qp_type == IB_QPT_SMI) {
 			if ((mad_reg_req->mgmt_class !=
 					IB_MGMT_CLASS_SUBN_LID_ROUTED) &&
 			    (mad_reg_req->mgmt_class !=
-					IB_MGMT_CLASS_SUBN_DIRECTED_ROUTE))
+					IB_MGMT_CLASS_SUBN_DIRECTED_ROUTE)) {
+				dev_notice(&device->dev,
+					   "ib_register_mad_agent: Invalid SM QP type: class 0x%x\n",
+					   mad_reg_req->mgmt_class);
 				goto error1;
+			}
 		} else {
 			if ((mad_reg_req->mgmt_class ==
 					IB_MGMT_CLASS_SUBN_LID_ROUTED) ||
 			    (mad_reg_req->mgmt_class ==
-					IB_MGMT_CLASS_SUBN_DIRECTED_ROUTE))
+					IB_MGMT_CLASS_SUBN_DIRECTED_ROUTE)) {
+				dev_notice(&device->dev,
+					   "ib_register_mad_agent: Invalid GS QP type: class 0x%x\n",
+					   mad_reg_req->mgmt_class);
 				goto error1;
+			}
 		}
 	} else {
 		/* No registration request supplied */
 		if (!send_handler)
 			goto error1;
+		if (registration_flags & IB_MAD_USER_RMPP)
+			goto error1;
 	}
 
 	/* Validate device and port */
 	port_priv = ib_get_mad_port(device, port_num);
 	if (!port_priv) {
+		dev_notice(&device->dev, "ib_register_mad_agent: Invalid port\n");
 		ret = ERR_PTR(-ENODEV);
 		goto error1;
 	}
@@ -280,6 +325,8 @@ struct ib_mad_agent *ib_register_mad_agent(struct ib_device *device,
 	/* Verify the QP requested is supported.  For example, Ethernet devices
 	 * will not have QP0 */
 	if (!port_priv->qp_info[qpn].qp) {
+		dev_notice(&device->dev,
+			   "ib_register_mad_agent: QP %d not supported\n", qpn);
 		ret = ERR_PTR(-EPROTONOSUPPORT);
 		goto error1;
 	}
@@ -316,6 +363,7 @@ struct ib_mad_agent *ib_register_mad_agent(struct ib_device *device,
 	mad_agent_priv->agent.context = context;
 	mad_agent_priv->agent.qp = port_priv->qp_info[qpn].qp;
 	mad_agent_priv->agent.port_num = port_num;
+	mad_agent_priv->agent.flags = registration_flags;
 	spin_lock_init(&mad_agent_priv->lock);
 	INIT_LIST_HEAD(&mad_agent_priv->send_list);
 	INIT_LIST_HEAD(&mad_agent_priv->wait_list);
@@ -706,7 +754,7 @@ static int handle_outgoing_dr_smp(struct ib_mad_agent_private *mad_agent_priv,
 	     smi_handle_dr_smp_send(smp, device->node_type, port_num) ==
 	     IB_SMI_DISCARD) {
 		ret = -EINVAL;
-		printk(KERN_ERR PFX "Invalid directed route\n");
+		dev_err(&device->dev, "Invalid directed route\n");
 		goto out;
 	}
 
@@ -718,7 +766,7 @@ static int handle_outgoing_dr_smp(struct ib_mad_agent_private *mad_agent_priv,
 	local = kmalloc(sizeof *local, GFP_ATOMIC);
 	if (!local) {
 		ret = -ENOMEM;
-		printk(KERN_ERR PFX "No memory for ib_mad_local_private\n");
+		dev_err(&device->dev, "No memory for ib_mad_local_private\n");
 		goto out;
 	}
 	local->mad_priv = NULL;
@@ -726,7 +774,7 @@ static int handle_outgoing_dr_smp(struct ib_mad_agent_private *mad_agent_priv,
 	mad_priv = kmem_cache_alloc(ib_mad_cache, GFP_ATOMIC);
 	if (!mad_priv) {
 		ret = -ENOMEM;
-		printk(KERN_ERR PFX "No memory for local response MAD\n");
+		dev_err(&device->dev, "No memory for local response MAD\n");
 		kfree(local);
 		goto out;
 	}
@@ -837,9 +885,9 @@ static int alloc_send_rmpp_list(struct ib_mad_send_wr_private *send_wr,
 	for (left = send_buf->data_len + pad; left > 0; left -= seg_size) {
 		seg = kmalloc(sizeof (*seg) + seg_size, gfp_mask);
 		if (!seg) {
-			printk(KERN_ERR "alloc_send_rmpp_segs: RMPP mem "
-			       "alloc failed for len %zd, gfp %#x\n",
-			       sizeof (*seg) + seg_size, gfp_mask);
+			dev_err(&send_buf->mad_agent->device->dev,
+				"alloc_send_rmpp_segs: RMPP mem alloc failed for len %zd, gfp %#x\n",
+				sizeof (*seg) + seg_size, gfp_mask);
 			free_send_rmpp_list(send_wr);
 			return -ENOMEM;
 		}
@@ -862,6 +910,12 @@ static int alloc_send_rmpp_list(struct ib_mad_send_wr_private *send_wr,
 	return 0;
 }
 
+int ib_mad_kernel_rmpp_agent(struct ib_mad_agent *agent)
+{
+	return agent->rmpp_version && !(agent->flags & IB_MAD_USER_RMPP);
+}
+EXPORT_SYMBOL(ib_mad_kernel_rmpp_agent);
+
 struct ib_mad_send_buf * ib_create_send_mad(struct ib_mad_agent *mad_agent,
 					    u32 remote_qpn, u16 pkey_index,
 					    int rmpp_active,
@@ -878,10 +932,12 @@ struct ib_mad_send_buf * ib_create_send_mad(struct ib_mad_agent *mad_agent,
 	pad = get_pad_size(hdr_len, data_len);
 	message_size = hdr_len + data_len + pad;
 
-	if ((!mad_agent->rmpp_version &&
-	     (rmpp_active || message_size > sizeof(struct ib_mad))) ||
-	    (!rmpp_active && message_size > sizeof(struct ib_mad)))
-		return ERR_PTR(-EINVAL);
+	if (ib_mad_kernel_rmpp_agent(mad_agent)) {
+		if (!rmpp_active && message_size > sizeof(struct ib_mad))
+			return ERR_PTR(-EINVAL);
+	} else
+		if (rmpp_active || message_size > sizeof(struct ib_mad))
+			return ERR_PTR(-EINVAL);
 
 	size = rmpp_active ? hdr_len : sizeof(struct ib_mad);
 	buf = kzalloc(sizeof *mad_send_wr + size, gfp_mask);
@@ -1135,7 +1191,7 @@ int ib_post_send_mad(struct ib_mad_send_buf *send_buf,
 			      &mad_agent_priv->send_list);
 		spin_unlock_irqrestore(&mad_agent_priv->lock, flags);
 
-		if (mad_agent_priv->agent.rmpp_version) {
+		if (ib_mad_kernel_rmpp_agent(&mad_agent_priv->agent)) {
 			ret = ib_send_rmpp_mad(mad_send_wr);
 			if (ret >= 0 && ret != IB_RMPP_RESULT_CONSUMED)
 				ret = ib_send_mad(mad_send_wr);
@@ -1199,7 +1255,8 @@ EXPORT_SYMBOL(ib_redirect_mad_qp);
 int ib_process_mad_wc(struct ib_mad_agent *mad_agent,
 		      struct ib_wc *wc)
 {
-	printk(KERN_ERR PFX "ib_process_mad_wc() not implemented yet\n");
+	dev_err(&mad_agent->device->dev,
+		"ib_process_mad_wc() not implemented yet\n");
 	return 0;
 }
 EXPORT_SYMBOL(ib_process_mad_wc);
@@ -1211,7 +1268,7 @@ static int method_in_use(struct ib_mad_mgmt_method_table **method,
 
 	for_each_set_bit(i, mad_reg_req->method_mask, IB_MGMT_MAX_METHODS) {
 		if ((*method)->agent[i]) {
-			printk(KERN_ERR PFX "Method %d already in use\n", i);
+			pr_err("Method %d already in use\n", i);
 			return -EINVAL;
 		}
 	}
@@ -1223,8 +1280,7 @@ static int allocate_method_table(struct ib_mad_mgmt_method_table **method)
 	/* Allocate management method table */
 	*method = kzalloc(sizeof **method, GFP_ATOMIC);
 	if (!*method) {
-		printk(KERN_ERR PFX "No memory for "
-		       "ib_mad_mgmt_method_table\n");
+		pr_err("No memory for ib_mad_mgmt_method_table\n");
 		return -ENOMEM;
 	}
 
@@ -1319,8 +1375,8 @@ static int add_nonoui_reg_req(struct ib_mad_reg_req *mad_reg_req,
 		/* Allocate management class table for "new" class version */
 		*class = kzalloc(sizeof **class, GFP_ATOMIC);
 		if (!*class) {
-			printk(KERN_ERR PFX "No memory for "
-			       "ib_mad_mgmt_class_table\n");
+			dev_err(&agent_priv->agent.device->dev,
+				"No memory for ib_mad_mgmt_class_table\n");
 			ret = -ENOMEM;
 			goto error1;
 		}
@@ -1386,8 +1442,8 @@ static int add_oui_reg_req(struct ib_mad_reg_req *mad_reg_req,
 		/* Allocate mgmt vendor class table for "new" class version */
 		vendor = kzalloc(sizeof *vendor, GFP_ATOMIC);
 		if (!vendor) {
-			printk(KERN_ERR PFX "No memory for "
-			       "ib_mad_mgmt_vendor_class_table\n");
+			dev_err(&agent_priv->agent.device->dev,
+				"No memory for ib_mad_mgmt_vendor_class_table\n");
 			goto error1;
 		}
 
@@ -1397,8 +1453,8 @@ static int add_oui_reg_req(struct ib_mad_reg_req *mad_reg_req,
 		/* Allocate table for this management vendor class */
 		vendor_class = kzalloc(sizeof *vendor_class, GFP_ATOMIC);
 		if (!vendor_class) {
-			printk(KERN_ERR PFX "No memory for "
-			       "ib_mad_mgmt_vendor_class\n");
+			dev_err(&agent_priv->agent.device->dev,
+				"No memory for ib_mad_mgmt_vendor_class\n");
 			goto error2;
 		}
 
@@ -1429,7 +1485,7 @@ static int add_oui_reg_req(struct ib_mad_reg_req *mad_reg_req,
 			goto check_in_use;
 		}
 	}
-	printk(KERN_ERR PFX "All OUI slots in use\n");
+	dev_err(&agent_priv->agent.device->dev, "All OUI slots in use\n");
 	goto error3;
 
 check_in_use:
@@ -1640,9 +1696,9 @@ find_mad_agent(struct ib_mad_port_private *port_priv,
 		if (mad_agent->agent.recv_handler)
 			atomic_inc(&mad_agent->refcount);
 		else {
-			printk(KERN_NOTICE PFX "No receive handler for client "
-			       "%p on port %d\n",
-			       &mad_agent->agent, port_priv->port_num);
+			dev_notice(&port_priv->device->dev,
+				   "No receive handler for client %p on port %d\n",
+				   &mad_agent->agent, port_priv->port_num);
 			mad_agent = NULL;
 		}
 	}
@@ -1658,8 +1714,8 @@ static int validate_mad(struct ib_mad *mad, u32 qp_num)
 
 	/* Make sure MAD base version is understood */
 	if (mad->mad_hdr.base_version != IB_MGMT_BASE_VERSION) {
-		printk(KERN_ERR PFX "MAD received with unsupported base "
-		       "version %d\n", mad->mad_hdr.base_version);
+		pr_err("MAD received with unsupported base version %d\n",
+			mad->mad_hdr.base_version);
 		goto out;
 	}
 
@@ -1685,6 +1741,7 @@ static int is_data_mad(struct ib_mad_agent_private *mad_agent_priv,
 
 	rmpp_mad = (struct ib_rmpp_mad *)mad_hdr;
 	return !mad_agent_priv->agent.rmpp_version ||
+		!ib_mad_kernel_rmpp_agent(&mad_agent_priv->agent) ||
 		!(ib_get_rmpp_flags(&rmpp_mad->rmpp_hdr) &
 				    IB_MGMT_RMPP_FLAG_ACTIVE) ||
 		(rmpp_mad->rmpp_hdr.rmpp_type == IB_MGMT_RMPP_TYPE_DATA);
@@ -1812,7 +1869,7 @@ static void ib_mad_complete_recv(struct ib_mad_agent_private *mad_agent_priv,
 
 	INIT_LIST_HEAD(&mad_recv_wc->rmpp_list);
 	list_add(&mad_recv_wc->recv_buf.list, &mad_recv_wc->rmpp_list);
-	if (mad_agent_priv->agent.rmpp_version) {
+	if (ib_mad_kernel_rmpp_agent(&mad_agent_priv->agent)) {
 		mad_recv_wc = ib_process_rmpp_recv_wc(mad_agent_priv,
 						      mad_recv_wc);
 		if (!mad_recv_wc) {
@@ -1827,23 +1884,39 @@ static void ib_mad_complete_recv(struct ib_mad_agent_private *mad_agent_priv,
 		mad_send_wr = ib_find_send_mad(mad_agent_priv, mad_recv_wc);
 		if (!mad_send_wr) {
 			spin_unlock_irqrestore(&mad_agent_priv->lock, flags);
-			ib_free_recv_mad(mad_recv_wc);
-			deref_mad_agent(mad_agent_priv);
-			return;
-		}
-		ib_mark_mad_done(mad_send_wr);
-		spin_unlock_irqrestore(&mad_agent_priv->lock, flags);
+			if (!ib_mad_kernel_rmpp_agent(&mad_agent_priv->agent)
+			   && ib_is_mad_class_rmpp(mad_recv_wc->recv_buf.mad->mad_hdr.mgmt_class)
+			   && (ib_get_rmpp_flags(&((struct ib_rmpp_mad *)mad_recv_wc->recv_buf.mad)->rmpp_hdr)
+					& IB_MGMT_RMPP_FLAG_ACTIVE)) {
+				/* user rmpp is in effect
+				 * and this is an active RMPP MAD
+				 */
+				mad_recv_wc->wc->wr_id = 0;
+				mad_agent_priv->agent.recv_handler(&mad_agent_priv->agent,
+								   mad_recv_wc);
+				atomic_dec(&mad_agent_priv->refcount);
+			} else {
+				/* not user rmpp, revert to normal behavior and
+				 * drop the mad */
+				ib_free_recv_mad(mad_recv_wc);
+				deref_mad_agent(mad_agent_priv);
+				return;
+			}
+		} else {
+			ib_mark_mad_done(mad_send_wr);
+			spin_unlock_irqrestore(&mad_agent_priv->lock, flags);
 
-		/* Defined behavior is to complete response before request */
-		mad_recv_wc->wc->wr_id = (unsigned long) &mad_send_wr->send_buf;
-		mad_agent_priv->agent.recv_handler(&mad_agent_priv->agent,
-						   mad_recv_wc);
-		atomic_dec(&mad_agent_priv->refcount);
+			/* Defined behavior is to complete response before request */
+			mad_recv_wc->wc->wr_id = (unsigned long) &mad_send_wr->send_buf;
+			mad_agent_priv->agent.recv_handler(&mad_agent_priv->agent,
+							   mad_recv_wc);
+			atomic_dec(&mad_agent_priv->refcount);
 
-		mad_send_wc.status = IB_WC_SUCCESS;
-		mad_send_wc.vendor_err = 0;
-		mad_send_wc.send_buf = &mad_send_wr->send_buf;
-		ib_mad_complete_send_wr(mad_send_wr, &mad_send_wc);
+			mad_send_wc.status = IB_WC_SUCCESS;
+			mad_send_wc.vendor_err = 0;
+			mad_send_wc.send_buf = &mad_send_wr->send_buf;
+			ib_mad_complete_send_wr(mad_send_wr, &mad_send_wc);
+		}
 	} else {
 		mad_agent_priv->agent.recv_handler(&mad_agent_priv->agent,
 						   mad_recv_wc);
@@ -1911,8 +1984,8 @@ static void ib_mad_recv_done_handler(struct ib_mad_port_private *port_priv,
 
 	response = kmem_cache_alloc(ib_mad_cache, GFP_KERNEL);
 	if (!response) {
-		printk(KERN_ERR PFX "ib_mad_recv_done_handler no memory "
-		       "for response buffer\n");
+		dev_err(&port_priv->device->dev,
+			"ib_mad_recv_done_handler no memory for response buffer\n");
 		goto out;
 	}
 
@@ -2083,7 +2156,7 @@ void ib_mad_complete_send_wr(struct ib_mad_send_wr_private *mad_send_wr,
 
 	mad_agent_priv = mad_send_wr->mad_agent_priv;
 	spin_lock_irqsave(&mad_agent_priv->lock, flags);
-	if (mad_agent_priv->agent.rmpp_version) {
+	if (ib_mad_kernel_rmpp_agent(&mad_agent_priv->agent)) {
 		ret = ib_process_rmpp_send_wc(mad_send_wr, mad_send_wc);
 		if (ret == IB_RMPP_RESULT_CONSUMED)
 			goto done;
@@ -2176,7 +2249,8 @@ retry:
 		ret = ib_post_send(qp_info->qp, &queued_send_wr->send_wr,
 				   &bad_send_wr);
 		if (ret) {
-			printk(KERN_ERR PFX "ib_post_send failed: %d\n", ret);
+			dev_err(&port_priv->device->dev,
+				"ib_post_send failed: %d\n", ret);
 			mad_send_wr = queued_send_wr;
 			wc->status = IB_WC_LOC_QP_OP_ERR;
 			goto retry;
@@ -2248,8 +2322,9 @@ static void mad_error_handler(struct ib_mad_port_private *port_priv,
 					   IB_QP_STATE | IB_QP_CUR_STATE);
 			kfree(attr);
 			if (ret)
-				printk(KERN_ERR PFX "mad_error_handler - "
-				       "ib_modify_qp to RTS : %d\n", ret);
+				dev_err(&port_priv->device->dev,
+					"mad_error_handler - ib_modify_qp to RTS : %d\n",
+					ret);
 			else
 				mark_sends_for_retry(qp_info);
 		}
@@ -2408,7 +2483,8 @@ static void local_completions(struct work_struct *work)
 		if (local->mad_priv) {
 			recv_mad_agent = local->recv_mad_agent;
 			if (!recv_mad_agent) {
-				printk(KERN_ERR PFX "No receive MAD agent for local completion\n");
+				dev_err(&mad_agent_priv->agent.device->dev,
+					"No receive MAD agent for local completion\n");
 				free_mad = 1;
 				goto local_send_completion;
 			}
@@ -2476,7 +2552,7 @@ static int retry_send(struct ib_mad_send_wr_private *mad_send_wr)
 
 	mad_send_wr->timeout = msecs_to_jiffies(mad_send_wr->send_buf.timeout_ms);
 
-	if (mad_send_wr->mad_agent_priv->agent.rmpp_version) {
+	if (ib_mad_kernel_rmpp_agent(&mad_send_wr->mad_agent_priv->agent)) {
 		ret = ib_retry_rmpp(mad_send_wr);
 		switch (ret) {
 		case IB_RMPP_RESULT_UNHANDLED:
@@ -2589,7 +2665,8 @@ static int ib_mad_post_receive_mads(struct ib_mad_qp_info *qp_info,
 		} else {
 			mad_priv = kmem_cache_alloc(ib_mad_cache, GFP_KERNEL);
 			if (!mad_priv) {
-				printk(KERN_ERR PFX "No memory for receive buffer\n");
+				dev_err(&qp_info->port_priv->device->dev,
+					"No memory for receive buffer\n");
 				ret = -ENOMEM;
 				break;
 			}
@@ -2625,7 +2702,8 @@ static int ib_mad_post_receive_mads(struct ib_mad_qp_info *qp_info,
 					      sizeof mad_priv->header,
 					    DMA_FROM_DEVICE);
 			kmem_cache_free(ib_mad_cache, mad_priv);
-			printk(KERN_ERR PFX "ib_post_recv failed: %d\n", ret);
+			dev_err(&qp_info->port_priv->device->dev,
+				"ib_post_recv failed: %d\n", ret);
 			break;
 		}
 	} while (post);
@@ -2681,7 +2759,8 @@ static int ib_mad_port_start(struct ib_mad_port_private *port_priv)
 
 	attr = kmalloc(sizeof *attr, GFP_KERNEL);
 	if (!attr) {
-		printk(KERN_ERR PFX "Couldn't kmalloc ib_qp_attr\n");
+		dev_err(&port_priv->device->dev,
+			"Couldn't kmalloc ib_qp_attr\n");
 		return -ENOMEM;
 	}
 
@@ -2705,16 +2784,18 @@ static int ib_mad_port_start(struct ib_mad_port_private *port_priv)
 		ret = ib_modify_qp(qp, attr, IB_QP_STATE |
 					     IB_QP_PKEY_INDEX | IB_QP_QKEY);
 		if (ret) {
-			printk(KERN_ERR PFX "Couldn't change QP%d state to "
-			       "INIT: %d\n", i, ret);
+			dev_err(&port_priv->device->dev,
+				"Couldn't change QP%d state to INIT: %d\n",
+				i, ret);
 			goto out;
 		}
 
 		attr->qp_state = IB_QPS_RTR;
 		ret = ib_modify_qp(qp, attr, IB_QP_STATE);
 		if (ret) {
-			printk(KERN_ERR PFX "Couldn't change QP%d state to "
-			       "RTR: %d\n", i, ret);
+			dev_err(&port_priv->device->dev,
+				"Couldn't change QP%d state to RTR: %d\n",
+				i, ret);
 			goto out;
 		}
 
@@ -2722,16 +2803,18 @@ static int ib_mad_port_start(struct ib_mad_port_private *port_priv)
 		attr->sq_psn = IB_MAD_SEND_Q_PSN;
 		ret = ib_modify_qp(qp, attr, IB_QP_STATE | IB_QP_SQ_PSN);
 		if (ret) {
-			printk(KERN_ERR PFX "Couldn't change QP%d state to "
-			       "RTS: %d\n", i, ret);
+			dev_err(&port_priv->device->dev,
+				"Couldn't change QP%d state to RTS: %d\n",
+				i, ret);
 			goto out;
 		}
 	}
 
 	ret = ib_req_notify_cq(port_priv->cq, IB_CQ_NEXT_COMP);
 	if (ret) {
-		printk(KERN_ERR PFX "Failed to request completion "
-		       "notification: %d\n", ret);
+		dev_err(&port_priv->device->dev,
+			"Failed to request completion notification: %d\n",
+			ret);
 		goto out;
 	}
 
@@ -2741,7 +2824,8 @@ static int ib_mad_port_start(struct ib_mad_port_private *port_priv)
 
 		ret = ib_mad_post_receive_mads(&port_priv->qp_info[i], NULL);
 		if (ret) {
-			printk(KERN_ERR PFX "Couldn't post receive WRs\n");
+			dev_err(&port_priv->device->dev,
+				"Couldn't post receive WRs\n");
 			goto out;
 		}
 	}
@@ -2755,7 +2839,8 @@ static void qp_event_handler(struct ib_event *event, void *qp_context)
 	struct ib_mad_qp_info	*qp_info = qp_context;
 
 	/* It's worse than that! He's dead, Jim! */
-	printk(KERN_ERR PFX "Fatal error (%d) on MAD QP (%d)\n",
+	dev_err(&qp_info->port_priv->device->dev,
+		"Fatal error (%d) on MAD QP (%d)\n",
 		event->event, qp_info->qp->qp_num);
 }
 
@@ -2801,8 +2886,9 @@ static int create_mad_qp(struct ib_mad_qp_info *qp_info,
 	qp_init_attr.event_handler = qp_event_handler;
 	qp_info->qp = ib_create_qp(qp_info->port_priv->pd, &qp_init_attr);
 	if (IS_ERR(qp_info->qp)) {
-		printk(KERN_ERR PFX "Couldn't create ib_mad QP%d\n",
-		       get_spl_qp_index(qp_type));
+		dev_err(&qp_info->port_priv->device->dev,
+			"Couldn't create ib_mad QP%d\n",
+			get_spl_qp_index(qp_type));
 		ret = PTR_ERR(qp_info->qp);
 		goto error;
 	}
@@ -2840,7 +2926,7 @@ static int ib_mad_port_open(struct ib_device *device,
 	/* Create new device info */
 	port_priv = kzalloc(sizeof *port_priv, GFP_KERNEL);
 	if (!port_priv) {
-		printk(KERN_ERR PFX "No memory for ib_mad_port_private\n");
+		dev_err(&device->dev, "No memory for ib_mad_port_private\n");
 		return -ENOMEM;
 	}
 
@@ -2860,21 +2946,21 @@ static int ib_mad_port_open(struct ib_device *device,
 				     ib_mad_thread_completion_handler,
 				     NULL, port_priv, cq_size, 0);
 	if (IS_ERR(port_priv->cq)) {
-		printk(KERN_ERR PFX "Couldn't create ib_mad CQ\n");
+		dev_err(&device->dev, "Couldn't create ib_mad CQ\n");
 		ret = PTR_ERR(port_priv->cq);
 		goto error3;
 	}
 
 	port_priv->pd = ib_alloc_pd(device);
 	if (IS_ERR(port_priv->pd)) {
-		printk(KERN_ERR PFX "Couldn't create ib_mad PD\n");
+		dev_err(&device->dev, "Couldn't create ib_mad PD\n");
 		ret = PTR_ERR(port_priv->pd);
 		goto error4;
 	}
 
 	port_priv->mr = ib_get_dma_mr(port_priv->pd, IB_ACCESS_LOCAL_WRITE);
 	if (IS_ERR(port_priv->mr)) {
-		printk(KERN_ERR PFX "Couldn't get ib_mad DMA MR\n");
+		dev_err(&device->dev, "Couldn't get ib_mad DMA MR\n");
 		ret = PTR_ERR(port_priv->mr);
 		goto error5;
 	}
@@ -2902,7 +2988,7 @@ static int ib_mad_port_open(struct ib_device *device,
 
 	ret = ib_mad_port_start(port_priv);
 	if (ret) {
-		printk(KERN_ERR PFX "Couldn't start port\n");
+		dev_err(&device->dev, "Couldn't start port\n");
 		goto error9;
 	}
 
@@ -2946,7 +3032,7 @@ static int ib_mad_port_close(struct ib_device *device, int port_num)
 	port_priv = __ib_get_mad_port(device, port_num);
 	if (port_priv == NULL) {
 		spin_unlock_irqrestore(&ib_mad_port_list_lock, flags);
-		printk(KERN_ERR PFX "Port %d not found\n", port_num);
+		dev_err(&device->dev, "Port %d not found\n", port_num);
 		return -ENODEV;
 	}
 	list_del_init(&port_priv->port_list);
@@ -2984,14 +3070,12 @@ static void ib_mad_init_device(struct ib_device *device)
 
 	for (i = start; i <= end; i++) {
 		if (ib_mad_port_open(device, i)) {
-			printk(KERN_ERR PFX "Couldn't open %s port %d\n",
-			       device->name, i);
+			dev_err(&device->dev, "Couldn't open port %d\n", i);
 			goto error;
 		}
 		if (ib_agent_port_open(device, i)) {
-			printk(KERN_ERR PFX "Couldn't open %s port %d "
-			       "for agents\n",
-			       device->name, i);
+			dev_err(&device->dev,
+				"Couldn't open port %d for agents\n", i);
 			goto error_agent;
 		}
 	}
@@ -2999,20 +3083,17 @@ static void ib_mad_init_device(struct ib_device *device)
 
 error_agent:
 	if (ib_mad_port_close(device, i))
-		printk(KERN_ERR PFX "Couldn't close %s port %d\n",
-		       device->name, i);
+		dev_err(&device->dev, "Couldn't close port %d\n", i);
 
 error:
 	i--;
 
 	while (i >= start) {
 		if (ib_agent_port_close(device, i))
-			printk(KERN_ERR PFX "Couldn't close %s port %d "
-			       "for agents\n",
-			       device->name, i);
+			dev_err(&device->dev,
+				"Couldn't close port %d for agents\n", i);
 		if (ib_mad_port_close(device, i))
-			printk(KERN_ERR PFX "Couldn't close %s port %d\n",
-			       device->name, i);
+			dev_err(&device->dev, "Couldn't close port %d\n", i);
 		i--;
 	}
 }
@@ -3033,12 +3114,12 @@ static void ib_mad_remove_device(struct ib_device *device)
 	}
 	for (i = 0; i < num_ports; i++, cur_port++) {
 		if (ib_agent_port_close(device, cur_port))
-			printk(KERN_ERR PFX "Couldn't close %s port %d "
-			       "for agents\n",
-			       device->name, cur_port);
+			dev_err(&device->dev,
+				"Couldn't close port %d for agents\n",
+				cur_port);
 		if (ib_mad_port_close(device, cur_port))
-			printk(KERN_ERR PFX "Couldn't close %s port %d\n",
-			       device->name, cur_port);
+			dev_err(&device->dev, "Couldn't close port %d\n",
+				cur_port);
 	}
 }
 
@@ -3064,7 +3145,7 @@ static int __init ib_mad_init_module(void)
 					 SLAB_HWCACHE_ALIGN,
 					 NULL);
 	if (!ib_mad_cache) {
-		printk(KERN_ERR PFX "Couldn't create ib_mad cache\n");
+		pr_err("Couldn't create ib_mad cache\n");
 		ret = -ENOMEM;
 		goto error1;
 	}
@@ -3072,7 +3153,7 @@ static int __init ib_mad_init_module(void)
 	INIT_LIST_HEAD(&ib_mad_port_list);
 
 	if (ib_register_client(&mad_client)) {
-		printk(KERN_ERR PFX "Couldn't register ib_mad client\n");
+		pr_err("Couldn't register ib_mad client\n");
 		ret = -EINVAL;
 		goto error2;
 	}
diff --git a/drivers/infiniband/core/mad_priv.h b/drivers/infiniband/core/mad_priv.h
index 9430ab4969c5..d1a0b0ee9444 100644
--- a/drivers/infiniband/core/mad_priv.h
+++ b/drivers/infiniband/core/mad_priv.h
@@ -42,9 +42,6 @@
 #include <rdma/ib_mad.h>
 #include <rdma/ib_smi.h>
 
-
-#define PFX "ib_mad: "
-
 #define IB_MAD_QPS_CORE		2 /* Always QP0 and QP1 as a minimum */
 
 /* QP and CQ parameters */
diff --git a/drivers/infiniband/core/sa_query.c b/drivers/infiniband/core/sa_query.c
index 233eaf541f55..c38f030f0dc9 100644
--- a/drivers/infiniband/core/sa_query.c
+++ b/drivers/infiniband/core/sa_query.c
@@ -1184,7 +1184,7 @@ static void ib_sa_add_one(struct ib_device *device)
 		sa_dev->port[i].agent =
 			ib_register_mad_agent(device, i + s, IB_QPT_GSI,
 					      NULL, 0, send_handler,
-					      recv_handler, sa_dev);
+					      recv_handler, sa_dev, 0);
 		if (IS_ERR(sa_dev->port[i].agent))
 			goto err;
 
diff --git a/drivers/infiniband/core/umem.c b/drivers/infiniband/core/umem.c
index a3a2e9c1639b..df0c4f605a21 100644
--- a/drivers/infiniband/core/umem.c
+++ b/drivers/infiniband/core/umem.c
@@ -105,6 +105,7 @@ struct ib_umem *ib_umem_get(struct ib_ucontext *context, unsigned long addr,
 	umem->length    = size;
 	umem->offset    = addr & ~PAGE_MASK;
 	umem->page_size = PAGE_SIZE;
+	umem->pid       = get_task_pid(current, PIDTYPE_PID);
 	/*
 	 * We ask for writable memory if any access flags other than
 	 * "remote read" are set.  "Local write" and "remote write"
@@ -198,6 +199,7 @@ out:
 	if (ret < 0) {
 		if (need_release)
 			__ib_umem_release(context->device, umem, 0);
+		put_pid(umem->pid);
 		kfree(umem);
 	} else
 		current->mm->pinned_vm = locked;
@@ -230,15 +232,19 @@ void ib_umem_release(struct ib_umem *umem)
 {
 	struct ib_ucontext *context = umem->context;
 	struct mm_struct *mm;
+	struct task_struct *task;
 	unsigned long diff;
 
 	__ib_umem_release(umem->context->device, umem, 1);
 
-	mm = get_task_mm(current);
-	if (!mm) {
-		kfree(umem);
-		return;
-	}
+	task = get_pid_task(umem->pid, PIDTYPE_PID);
+	put_pid(umem->pid);
+	if (!task)
+		goto out;
+	mm = get_task_mm(task);
+	put_task_struct(task);
+	if (!mm)
+		goto out;
 
 	diff = PAGE_ALIGN(umem->length + umem->offset) >> PAGE_SHIFT;
 
@@ -262,9 +268,10 @@ void ib_umem_release(struct ib_umem *umem)
 	} else
 		down_write(&mm->mmap_sem);
 
-	current->mm->pinned_vm -= diff;
+	mm->pinned_vm -= diff;
 	up_write(&mm->mmap_sem);
 	mmput(mm);
+out:
 	kfree(umem);
 }
 EXPORT_SYMBOL(ib_umem_release);
diff --git a/drivers/infiniband/core/user_mad.c b/drivers/infiniband/core/user_mad.c
index 1acb99100556..928cdd20e2d1 100644
--- a/drivers/infiniband/core/user_mad.c
+++ b/drivers/infiniband/core/user_mad.c
@@ -33,6 +33,8 @@
  * SOFTWARE.
  */
 
+#define pr_fmt(fmt) "user_mad: " fmt
+
 #include <linux/module.h>
 #include <linux/init.h>
 #include <linux/device.h>
@@ -504,13 +506,15 @@ static ssize_t ib_umad_write(struct file *filp, const char __user *buf,
 
 	rmpp_mad = (struct ib_rmpp_mad *) packet->mad.data;
 	hdr_len = ib_get_mad_data_offset(rmpp_mad->mad_hdr.mgmt_class);
-	if (!ib_is_mad_class_rmpp(rmpp_mad->mad_hdr.mgmt_class)) {
-		copy_offset = IB_MGMT_MAD_HDR;
-		rmpp_active = 0;
-	} else {
+
+	if (ib_is_mad_class_rmpp(rmpp_mad->mad_hdr.mgmt_class)
+	    && ib_mad_kernel_rmpp_agent(agent)) {
 		copy_offset = IB_MGMT_RMPP_HDR;
 		rmpp_active = ib_get_rmpp_flags(&rmpp_mad->rmpp_hdr) &
-			      IB_MGMT_RMPP_FLAG_ACTIVE;
+						IB_MGMT_RMPP_FLAG_ACTIVE;
+	} else {
+		copy_offset = IB_MGMT_MAD_HDR;
+		rmpp_active = 0;
 	}
 
 	data_len = count - hdr_size(file) - hdr_len;
@@ -556,14 +560,22 @@ static ssize_t ib_umad_write(struct file *filp, const char __user *buf,
 		rmpp_mad->mad_hdr.tid = *tid;
 	}
 
-	spin_lock_irq(&file->send_lock);
-	ret = is_duplicate(file, packet);
-	if (!ret)
+	if (!ib_mad_kernel_rmpp_agent(agent)
+	   && ib_is_mad_class_rmpp(rmpp_mad->mad_hdr.mgmt_class)
+	   && (ib_get_rmpp_flags(&rmpp_mad->rmpp_hdr) & IB_MGMT_RMPP_FLAG_ACTIVE)) {
+		spin_lock_irq(&file->send_lock);
 		list_add_tail(&packet->list, &file->send_list);
-	spin_unlock_irq(&file->send_lock);
-	if (ret) {
-		ret = -EINVAL;
-		goto err_msg;
+		spin_unlock_irq(&file->send_lock);
+	} else {
+		spin_lock_irq(&file->send_lock);
+		ret = is_duplicate(file, packet);
+		if (!ret)
+			list_add_tail(&packet->list, &file->send_list);
+		spin_unlock_irq(&file->send_lock);
+		if (ret) {
+			ret = -EINVAL;
+			goto err_msg;
+		}
 	}
 
 	ret = ib_post_send_mad(packet->msg, NULL);
@@ -614,6 +626,8 @@ static int ib_umad_reg_agent(struct ib_umad_file *file, void __user *arg,
 	mutex_lock(&file->mutex);
 
 	if (!file->port->ib_dev) {
+		dev_notice(file->port->dev,
+			   "ib_umad_reg_agent: invalid device\n");
 		ret = -EPIPE;
 		goto out;
 	}
@@ -624,6 +638,9 @@ static int ib_umad_reg_agent(struct ib_umad_file *file, void __user *arg,
 	}
 
 	if (ureq.qpn != 0 && ureq.qpn != 1) {
+		dev_notice(file->port->dev,
+			   "ib_umad_reg_agent: invalid QPN %d specified\n",
+			   ureq.qpn);
 		ret = -EINVAL;
 		goto out;
 	}
@@ -632,11 +649,15 @@ static int ib_umad_reg_agent(struct ib_umad_file *file, void __user *arg,
 		if (!__get_agent(file, agent_id))
 			goto found;
 
+	dev_notice(file->port->dev,
+		   "ib_umad_reg_agent: Max Agents (%u) reached\n",
+		   IB_UMAD_MAX_AGENTS);
 	ret = -ENOMEM;
 	goto out;
 
 found:
 	if (ureq.mgmt_class) {
+		memset(&req, 0, sizeof(req));
 		req.mgmt_class         = ureq.mgmt_class;
 		req.mgmt_class_version = ureq.mgmt_class_version;
 		memcpy(req.oui, ureq.oui, sizeof req.oui);
@@ -657,7 +678,7 @@ found:
 				      ureq.qpn ? IB_QPT_GSI : IB_QPT_SMI,
 				      ureq.mgmt_class ? &req : NULL,
 				      ureq.rmpp_version,
-				      send_handler, recv_handler, file);
+				      send_handler, recv_handler, file, 0);
 	if (IS_ERR(agent)) {
 		ret = PTR_ERR(agent);
 		agent = NULL;
@@ -673,10 +694,11 @@ found:
 	if (!file->already_used) {
 		file->already_used = 1;
 		if (!file->use_pkey_index) {
-			printk(KERN_WARNING "user_mad: process %s did not enable "
-			       "P_Key index support.\n", current->comm);
-			printk(KERN_WARNING "user_mad:   Documentation/infiniband/user_mad.txt "
-			       "has info on the new ABI.\n");
+			dev_warn(file->port->dev,
+				"process %s did not enable P_Key index support.\n",
+				current->comm);
+			dev_warn(file->port->dev,
+				"   Documentation/infiniband/user_mad.txt has info on the new ABI.\n");
 		}
 	}
 
@@ -694,6 +716,119 @@ out:
 	return ret;
 }
 
+static int ib_umad_reg_agent2(struct ib_umad_file *file, void __user *arg)
+{
+	struct ib_user_mad_reg_req2 ureq;
+	struct ib_mad_reg_req req;
+	struct ib_mad_agent *agent = NULL;
+	int agent_id;
+	int ret;
+
+	mutex_lock(&file->port->file_mutex);
+	mutex_lock(&file->mutex);
+
+	if (!file->port->ib_dev) {
+		dev_notice(file->port->dev,
+			   "ib_umad_reg_agent2: invalid device\n");
+		ret = -EPIPE;
+		goto out;
+	}
+
+	if (copy_from_user(&ureq, arg, sizeof(ureq))) {
+		ret = -EFAULT;
+		goto out;
+	}
+
+	if (ureq.qpn != 0 && ureq.qpn != 1) {
+		dev_notice(file->port->dev,
+			   "ib_umad_reg_agent2: invalid QPN %d specified\n",
+			   ureq.qpn);
+		ret = -EINVAL;
+		goto out;
+	}
+
+	if (ureq.flags & ~IB_USER_MAD_REG_FLAGS_CAP) {
+		dev_notice(file->port->dev,
+			   "ib_umad_reg_agent2 failed: invalid registration flags specified 0x%x; supported 0x%x\n",
+			   ureq.flags, IB_USER_MAD_REG_FLAGS_CAP);
+		ret = -EINVAL;
+
+		if (put_user((u32)IB_USER_MAD_REG_FLAGS_CAP,
+				(u32 __user *) (arg + offsetof(struct
+				ib_user_mad_reg_req2, flags))))
+			ret = -EFAULT;
+
+		goto out;
+	}
+
+	for (agent_id = 0; agent_id < IB_UMAD_MAX_AGENTS; ++agent_id)
+		if (!__get_agent(file, agent_id))
+			goto found;
+
+	dev_notice(file->port->dev,
+		   "ib_umad_reg_agent2: Max Agents (%u) reached\n",
+		   IB_UMAD_MAX_AGENTS);
+	ret = -ENOMEM;
+	goto out;
+
+found:
+	if (ureq.mgmt_class) {
+		memset(&req, 0, sizeof(req));
+		req.mgmt_class         = ureq.mgmt_class;
+		req.mgmt_class_version = ureq.mgmt_class_version;
+		if (ureq.oui & 0xff000000) {
+			dev_notice(file->port->dev,
+				   "ib_umad_reg_agent2 failed: oui invalid 0x%08x\n",
+				   ureq.oui);
+			ret = -EINVAL;
+			goto out;
+		}
+		req.oui[2] =  ureq.oui & 0x0000ff;
+		req.oui[1] = (ureq.oui & 0x00ff00) >> 8;
+		req.oui[0] = (ureq.oui & 0xff0000) >> 16;
+		memcpy(req.method_mask, ureq.method_mask,
+			sizeof(req.method_mask));
+	}
+
+	agent = ib_register_mad_agent(file->port->ib_dev, file->port->port_num,
+				      ureq.qpn ? IB_QPT_GSI : IB_QPT_SMI,
+				      ureq.mgmt_class ? &req : NULL,
+				      ureq.rmpp_version,
+				      send_handler, recv_handler, file,
+				      ureq.flags);
+	if (IS_ERR(agent)) {
+		ret = PTR_ERR(agent);
+		agent = NULL;
+		goto out;
+	}
+
+	if (put_user(agent_id,
+		     (u32 __user *)(arg +
+				offsetof(struct ib_user_mad_reg_req2, id)))) {
+		ret = -EFAULT;
+		goto out;
+	}
+
+	if (!file->already_used) {
+		file->already_used = 1;
+		file->use_pkey_index = 1;
+	}
+
+	file->agent[agent_id] = agent;
+	ret = 0;
+
+out:
+	mutex_unlock(&file->mutex);
+
+	if (ret && agent)
+		ib_unregister_mad_agent(agent);
+
+	mutex_unlock(&file->port->file_mutex);
+
+	return ret;
+}
+
+
 static int ib_umad_unreg_agent(struct ib_umad_file *file, u32 __user *arg)
 {
 	struct ib_mad_agent *agent = NULL;
@@ -749,6 +884,8 @@ static long ib_umad_ioctl(struct file *filp, unsigned int cmd,
 		return ib_umad_unreg_agent(filp->private_data, (__u32 __user *) arg);
 	case IB_USER_MAD_ENABLE_PKEY:
 		return ib_umad_enable_pkey(filp->private_data);
+	case IB_USER_MAD_REGISTER_AGENT2:
+		return ib_umad_reg_agent2(filp->private_data, (void __user *) arg);
 	default:
 		return -ENOIOCTLCMD;
 	}
@@ -765,6 +902,8 @@ static long ib_umad_compat_ioctl(struct file *filp, unsigned int cmd,
 		return ib_umad_unreg_agent(filp->private_data, compat_ptr(arg));
 	case IB_USER_MAD_ENABLE_PKEY:
 		return ib_umad_enable_pkey(filp->private_data);
+	case IB_USER_MAD_REGISTER_AGENT2:
+		return ib_umad_reg_agent2(filp->private_data, compat_ptr(arg));
 	default:
 		return -ENOIOCTLCMD;
 	}
@@ -983,7 +1122,7 @@ static CLASS_ATTR_STRING(abi_version, S_IRUGO,
 
 static dev_t overflow_maj;
 static DECLARE_BITMAP(overflow_map, IB_UMAD_MAX_PORTS);
-static int find_overflow_devnum(void)
+static int find_overflow_devnum(struct ib_device *device)
 {
 	int ret;
 
@@ -991,7 +1130,8 @@ static int find_overflow_devnum(void)
 		ret = alloc_chrdev_region(&overflow_maj, 0, IB_UMAD_MAX_PORTS * 2,
 					  "infiniband_mad");
 		if (ret) {
-			printk(KERN_ERR "user_mad: couldn't register dynamic device number\n");
+			dev_err(&device->dev,
+				"couldn't register dynamic device number\n");
 			return ret;
 		}
 	}
@@ -1014,7 +1154,7 @@ static int ib_umad_init_port(struct ib_device *device, int port_num,
 	devnum = find_first_zero_bit(dev_map, IB_UMAD_MAX_PORTS);
 	if (devnum >= IB_UMAD_MAX_PORTS) {
 		spin_unlock(&port_lock);
-		devnum = find_overflow_devnum();
+		devnum = find_overflow_devnum(device);
 		if (devnum < 0)
 			return -1;
 
@@ -1200,14 +1340,14 @@ static int __init ib_umad_init(void)
 	ret = register_chrdev_region(base_dev, IB_UMAD_MAX_PORTS * 2,
 				     "infiniband_mad");
 	if (ret) {
-		printk(KERN_ERR "user_mad: couldn't register device number\n");
+		pr_err("couldn't register device number\n");
 		goto out;
 	}
 
 	umad_class = class_create(THIS_MODULE, "infiniband_mad");
 	if (IS_ERR(umad_class)) {
 		ret = PTR_ERR(umad_class);
-		printk(KERN_ERR "user_mad: couldn't create class infiniband_mad\n");
+		pr_err("couldn't create class infiniband_mad\n");
 		goto out_chrdev;
 	}
 
@@ -1215,13 +1355,13 @@ static int __init ib_umad_init(void)
 
 	ret = class_create_file(umad_class, &class_attr_abi_version.attr);
 	if (ret) {
-		printk(KERN_ERR "user_mad: couldn't create abi_version attribute\n");
+		pr_err("couldn't create abi_version attribute\n");
 		goto out_class;
 	}
 
 	ret = ib_register_client(&umad_client);
 	if (ret) {
-		printk(KERN_ERR "user_mad: couldn't register ib_umad client\n");
+		pr_err("couldn't register ib_umad client\n");
 		goto out_class;
 	}
 
diff --git a/drivers/infiniband/core/uverbs.h b/drivers/infiniband/core/uverbs.h
index a283274a5a09..643c08a025a5 100644
--- a/drivers/infiniband/core/uverbs.h
+++ b/drivers/infiniband/core/uverbs.h
@@ -221,6 +221,7 @@ IB_UVERBS_DECLARE_CMD(query_port);
 IB_UVERBS_DECLARE_CMD(alloc_pd);
 IB_UVERBS_DECLARE_CMD(dealloc_pd);
 IB_UVERBS_DECLARE_CMD(reg_mr);
+IB_UVERBS_DECLARE_CMD(rereg_mr);
 IB_UVERBS_DECLARE_CMD(dereg_mr);
 IB_UVERBS_DECLARE_CMD(alloc_mw);
 IB_UVERBS_DECLARE_CMD(dealloc_mw);
diff --git a/drivers/infiniband/core/uverbs_cmd.c b/drivers/infiniband/core/uverbs_cmd.c
index ea6203ee7bcc..5ba2a86aab6a 100644
--- a/drivers/infiniband/core/uverbs_cmd.c
+++ b/drivers/infiniband/core/uverbs_cmd.c
@@ -1002,6 +1002,99 @@ err_free:
 	return ret;
 }
 
+ssize_t ib_uverbs_rereg_mr(struct ib_uverbs_file *file,
+			   const char __user *buf, int in_len,
+			   int out_len)
+{
+	struct ib_uverbs_rereg_mr      cmd;
+	struct ib_uverbs_rereg_mr_resp resp;
+	struct ib_udata              udata;
+	struct ib_pd                *pd = NULL;
+	struct ib_mr                *mr;
+	struct ib_pd		    *old_pd;
+	int                          ret;
+	struct ib_uobject	    *uobj;
+
+	if (out_len < sizeof(resp))
+		return -ENOSPC;
+
+	if (copy_from_user(&cmd, buf, sizeof(cmd)))
+		return -EFAULT;
+
+	INIT_UDATA(&udata, buf + sizeof(cmd),
+		   (unsigned long) cmd.response + sizeof(resp),
+		   in_len - sizeof(cmd), out_len - sizeof(resp));
+
+	if (cmd.flags & ~IB_MR_REREG_SUPPORTED || !cmd.flags)
+		return -EINVAL;
+
+	if ((cmd.flags & IB_MR_REREG_TRANS) &&
+	    (!cmd.start || !cmd.hca_va || 0 >= cmd.length ||
+	     (cmd.start & ~PAGE_MASK) != (cmd.hca_va & ~PAGE_MASK)))
+			return -EINVAL;
+
+	uobj = idr_write_uobj(&ib_uverbs_mr_idr, cmd.mr_handle,
+			      file->ucontext);
+
+	if (!uobj)
+		return -EINVAL;
+
+	mr = uobj->object;
+
+	if (cmd.flags & IB_MR_REREG_ACCESS) {
+		ret = ib_check_mr_access(cmd.access_flags);
+		if (ret)
+			goto put_uobjs;
+	}
+
+	if (cmd.flags & IB_MR_REREG_PD) {
+		pd = idr_read_pd(cmd.pd_handle, file->ucontext);
+		if (!pd) {
+			ret = -EINVAL;
+			goto put_uobjs;
+		}
+	}
+
+	if (atomic_read(&mr->usecnt)) {
+		ret = -EBUSY;
+		goto put_uobj_pd;
+	}
+
+	old_pd = mr->pd;
+	ret = mr->device->rereg_user_mr(mr, cmd.flags, cmd.start,
+					cmd.length, cmd.hca_va,
+					cmd.access_flags, pd, &udata);
+	if (!ret) {
+		if (cmd.flags & IB_MR_REREG_PD) {
+			atomic_inc(&pd->usecnt);
+			mr->pd = pd;
+			atomic_dec(&old_pd->usecnt);
+		}
+	} else {
+		goto put_uobj_pd;
+	}
+
+	memset(&resp, 0, sizeof(resp));
+	resp.lkey      = mr->lkey;
+	resp.rkey      = mr->rkey;
+
+	if (copy_to_user((void __user *)(unsigned long)cmd.response,
+			 &resp, sizeof(resp)))
+		ret = -EFAULT;
+	else
+		ret = in_len;
+
+put_uobj_pd:
+	if (cmd.flags & IB_MR_REREG_PD)
+		put_pd_read(pd);
+
+put_uobjs:
+
+	put_uobj_write(mr->uobject);
+
+	return ret;
+}
+
 ssize_t ib_uverbs_dereg_mr(struct ib_uverbs_file *file,
 			   const char __user *buf, int in_len,
 			   int out_len)
@@ -2425,6 +2518,8 @@ ssize_t ib_uverbs_create_ah(struct ib_uverbs_file *file,
 	attr.grh.sgid_index    = cmd.attr.grh.sgid_index;
 	attr.grh.hop_limit     = cmd.attr.grh.hop_limit;
 	attr.grh.traffic_class = cmd.attr.grh.traffic_class;
+	attr.vlan_id           = 0;
+	memset(&attr.dmac, 0, sizeof(attr.dmac));
 	memcpy(attr.grh.dgid.raw, cmd.attr.grh.dgid, 16);
 
 	ah = ib_create_ah(pd, &attr);
diff --git a/drivers/infiniband/core/uverbs_main.c b/drivers/infiniband/core/uverbs_main.c
index 08219fb3338b..71ab83fde472 100644
--- a/drivers/infiniband/core/uverbs_main.c
+++ b/drivers/infiniband/core/uverbs_main.c
@@ -87,6 +87,7 @@ static ssize_t (*uverbs_cmd_table[])(struct ib_uverbs_file *file,
 	[IB_USER_VERBS_CMD_ALLOC_PD]		= ib_uverbs_alloc_pd,
 	[IB_USER_VERBS_CMD_DEALLOC_PD]		= ib_uverbs_dealloc_pd,
 	[IB_USER_VERBS_CMD_REG_MR]		= ib_uverbs_reg_mr,
+	[IB_USER_VERBS_CMD_REREG_MR]		= ib_uverbs_rereg_mr,
 	[IB_USER_VERBS_CMD_DEREG_MR]		= ib_uverbs_dereg_mr,
 	[IB_USER_VERBS_CMD_ALLOC_MW]		= ib_uverbs_alloc_mw,
 	[IB_USER_VERBS_CMD_DEALLOC_MW]		= ib_uverbs_dealloc_mw,
@@ -476,6 +477,7 @@ static void ib_uverbs_async_handler(struct ib_uverbs_file *file,
 
 	entry->desc.async.element    = element;
 	entry->desc.async.event_type = event;
+	entry->desc.async.reserved   = 0;
 	entry->counter               = counter;
 
 	list_add_tail(&entry->list, &file->async_file->event_list);
@@ -501,6 +503,10 @@ void ib_uverbs_qp_event_handler(struct ib_event *event, void *context_ptr)
 {
 	struct ib_uevent_object *uobj;
 
+	/* for XRC target qp's, check that qp is live */
+	if (!event->element.qp->uobject || !event->element.qp->uobject->live)
+		return;
+
 	uobj = container_of(event->element.qp->uobject,
 			    struct ib_uevent_object, uobject);
 
diff --git a/drivers/infiniband/core/uverbs_marshall.c b/drivers/infiniband/core/uverbs_marshall.c
index e7bee46868d1..abd97247443e 100644
--- a/drivers/infiniband/core/uverbs_marshall.c
+++ b/drivers/infiniband/core/uverbs_marshall.c
@@ -140,5 +140,9 @@ void ib_copy_path_rec_from_user(struct ib_sa_path_rec *dst,
 	dst->packet_life_time	= src->packet_life_time;
 	dst->preference		= src->preference;
 	dst->packet_life_time_selector = src->packet_life_time_selector;
+
+	memset(dst->smac, 0, sizeof(dst->smac));
+	memset(dst->dmac, 0, sizeof(dst->dmac));
+	dst->vlan_id = 0xffff;
 }
 EXPORT_SYMBOL(ib_copy_path_rec_from_user);
diff --git a/drivers/infiniband/hw/amso1100/c2.c b/drivers/infiniband/hw/amso1100/c2.c
index 00400c352c1a..766a71ccefed 100644
--- a/drivers/infiniband/hw/amso1100/c2.c
+++ b/drivers/infiniband/hw/amso1100/c2.c
@@ -604,16 +604,14 @@ static int c2_up(struct net_device *netdev)
 	tx_size = c2_port->tx_ring.count * sizeof(struct c2_tx_desc);
 
 	c2_port->mem_size = tx_size + rx_size;
-	c2_port->mem = pci_alloc_consistent(c2dev->pcidev, c2_port->mem_size,
-					    &c2_port->dma);
+	c2_port->mem = pci_zalloc_consistent(c2dev->pcidev, c2_port->mem_size,
+					     &c2_port->dma);
 	if (c2_port->mem == NULL) {
 		pr_debug("Unable to allocate memory for "
 			"host descriptor rings\n");
 		return -ENOMEM;
 	}
 
-	memset(c2_port->mem, 0, c2_port->mem_size);
-
 	/* Create the Rx host descriptor ring */
 	if ((ret =
 	     c2_rx_ring_alloc(&c2_port->rx_ring, c2_port->mem, c2_port->dma,
diff --git a/drivers/infiniband/hw/amso1100/c2_cq.c b/drivers/infiniband/hw/amso1100/c2_cq.c
index 49e0e8533f74..1b63185b4ad4 100644
--- a/drivers/infiniband/hw/amso1100/c2_cq.c
+++ b/drivers/infiniband/hw/amso1100/c2_cq.c
@@ -260,11 +260,14 @@ static void c2_free_cq_buf(struct c2_dev *c2dev, struct c2_mq *mq)
 			  mq->msg_pool.host, dma_unmap_addr(mq, mapping));
 }
 
-static int c2_alloc_cq_buf(struct c2_dev *c2dev, struct c2_mq *mq, int q_size,
-			   int msg_size)
+static int c2_alloc_cq_buf(struct c2_dev *c2dev, struct c2_mq *mq,
+			   size_t q_size, size_t msg_size)
 {
 	u8 *pool_start;
 
+	if (q_size > SIZE_MAX / msg_size)
+		return -EINVAL;
+
 	pool_start = dma_alloc_coherent(&c2dev->pcidev->dev, q_size * msg_size,
 					&mq->host_dma, GFP_KERNEL);
 	if (!pool_start)
diff --git a/drivers/infiniband/hw/cxgb4/cm.c b/drivers/infiniband/hw/cxgb4/cm.c
index c2fb71c182a8..fb61f6685809 100644
--- a/drivers/infiniband/hw/cxgb4/cm.c
+++ b/drivers/infiniband/hw/cxgb4/cm.c
@@ -236,10 +236,12 @@ static void release_tid(struct c4iw_rdev *rdev, u32 hwtid, struct sk_buff *skb)
 static void set_emss(struct c4iw_ep *ep, u16 opt)
 {
 	ep->emss = ep->com.dev->rdev.lldi.mtus[GET_TCPOPT_MSS(opt)] -
-		   sizeof(struct iphdr) - sizeof(struct tcphdr);
+		   ((AF_INET == ep->com.remote_addr.ss_family) ?
+		    sizeof(struct iphdr) : sizeof(struct ipv6hdr)) -
+		   sizeof(struct tcphdr);
 	ep->mss = ep->emss;
 	if (GET_TCPOPT_TSTAMP(opt))
-		ep->emss -= 12;
+		ep->emss -= round_up(TCPOLEN_TIMESTAMP, 4);
 	if (ep->emss < 128)
 		ep->emss = 128;
 	if (ep->emss & 7)
@@ -415,6 +417,7 @@ static struct dst_entry *find_route(struct c4iw_dev *dev, __be32 local_ip,
 		return NULL;
 	if (!our_interface(dev, n->dev) &&
 	    !(n->dev->flags & IFF_LOOPBACK)) {
+		neigh_release(n);
 		dst_release(&rt->dst);
 		return NULL;
 	}
@@ -581,11 +584,14 @@ static void c4iw_record_pm_msg(struct c4iw_ep *ep,
 }
 
 static void best_mtu(const unsigned short *mtus, unsigned short mtu,
-		     unsigned int *idx, int use_ts)
+		     unsigned int *idx, int use_ts, int ipv6)
 {
-	unsigned short hdr_size = sizeof(struct iphdr) +
+	unsigned short hdr_size = (ipv6 ?
+				   sizeof(struct ipv6hdr) :
+				   sizeof(struct iphdr)) +
 				  sizeof(struct tcphdr) +
-				  (use_ts ? 12 : 0);
+				  (use_ts ?
+				   round_up(TCPOLEN_TIMESTAMP, 4) : 0);
 	unsigned short data_size = mtu - hdr_size;
 
 	cxgb4_best_aligned_mtu(mtus, hdr_size, data_size, 8, idx);
@@ -634,7 +640,8 @@ static int send_connect(struct c4iw_ep *ep)
 	set_wr_txq(skb, CPL_PRIORITY_SETUP, ep->ctrlq_idx);
 
 	best_mtu(ep->com.dev->rdev.lldi.mtus, ep->mtu, &mtu_idx,
-		 enable_tcp_timestamps);
+		 enable_tcp_timestamps,
+		 (AF_INET == ep->com.remote_addr.ss_family) ? 0 : 1);
 	wscale = compute_wscale(rcv_win);
 
 	/*
@@ -668,6 +675,7 @@ static int send_connect(struct c4iw_ep *ep)
 	if (is_t5(ep->com.dev->rdev.lldi.adapter_type)) {
 		opt2 |= T5_OPT_2_VALID;
 		opt2 |= V_CONG_CNTRL(CONG_ALG_TAHOE);
+		opt2 |= CONG_CNTRL_VALID; /* OPT_2_ISS for T5 */
 	}
 	t4_set_arp_err_handler(skb, ep, act_open_req_arp_failure);
 
@@ -713,8 +721,6 @@ static int send_connect(struct c4iw_ep *ep)
 	} else {
 		u32 isn = (prandom_u32() & ~7UL) - 1;
 
-		opt2 |= T5_OPT_2_VALID;
-		opt2 |= CONG_CNTRL_VALID; /* OPT_2_ISS for T5 */
 		if (peer2peer)
 			isn += 4;
 
@@ -756,10 +762,10 @@ static int send_connect(struct c4iw_ep *ep)
 			t5_req6->peer_ip_lo = *((__be64 *)
 						(ra6->sin6_addr.s6_addr + 8));
 			t5_req6->opt0 = cpu_to_be64(opt0);
-			t5_req6->params = (__force __be64)cpu_to_be32(
+			t5_req6->params = cpu_to_be64(V_FILTER_TUPLE(
 							cxgb4_select_ntuple(
 						ep->com.dev->rdev.lldi.ports[0],
-						ep->l2t));
+						ep->l2t)));
 			t5_req6->rsvd = cpu_to_be32(isn);
 			PDBG("%s snd_isn %u\n", __func__,
 			     be32_to_cpu(t5_req6->rsvd));
@@ -1763,7 +1769,8 @@ static void send_fw_act_open_req(struct c4iw_ep *ep, unsigned int atid)
 	req->tcb.tx_max = (__force __be32) jiffies;
 	req->tcb.rcv_adv = htons(1);
 	best_mtu(ep->com.dev->rdev.lldi.mtus, ep->mtu, &mtu_idx,
-		 enable_tcp_timestamps);
+		 enable_tcp_timestamps,
+		 (AF_INET == ep->com.remote_addr.ss_family) ? 0 : 1);
 	wscale = compute_wscale(rcv_win);
 
 	/*
@@ -2162,7 +2169,8 @@ static void accept_cr(struct c4iw_ep *ep, struct sk_buff *skb,
 						    ep->hwtid));
 
 	best_mtu(ep->com.dev->rdev.lldi.mtus, ep->mtu, &mtu_idx,
-		 enable_tcp_timestamps && req->tcpopt.tstamp);
+		 enable_tcp_timestamps && req->tcpopt.tstamp,
+		 (AF_INET == ep->com.remote_addr.ss_family) ? 0 : 1);
 	wscale = compute_wscale(rcv_win);
 
 	/*
diff --git a/drivers/infiniband/hw/cxgb4/device.c b/drivers/infiniband/hw/cxgb4/device.c
index f25df5276c22..72f1f052e88c 100644
--- a/drivers/infiniband/hw/cxgb4/device.c
+++ b/drivers/infiniband/hw/cxgb4/device.c
@@ -60,7 +60,7 @@ int c4iw_wr_log = 0;
 module_param(c4iw_wr_log, int, 0444);
 MODULE_PARM_DESC(c4iw_wr_log, "Enables logging of work request timing data.");
 
-int c4iw_wr_log_size_order = 12;
+static int c4iw_wr_log_size_order = 12;
 module_param(c4iw_wr_log_size_order, int, 0444);
 MODULE_PARM_DESC(c4iw_wr_log_size_order,
 		 "Number of entries (log2) in the work request timing log.");
diff --git a/drivers/infiniband/hw/cxgb4/ev.c b/drivers/infiniband/hw/cxgb4/ev.c
index fbe6051af254..c9df0549f51d 100644
--- a/drivers/infiniband/hw/cxgb4/ev.c
+++ b/drivers/infiniband/hw/cxgb4/ev.c
@@ -227,6 +227,7 @@ int c4iw_ev_handler(struct c4iw_dev *dev, u32 qid)
 
 	chp = get_chp(dev, qid);
 	if (chp) {
+		t4_clear_cq_armed(&chp->cq);
 		spin_lock_irqsave(&chp->comp_handler_lock, flag);
 		(*chp->ibcq.comp_handler)(&chp->ibcq, chp->ibcq.cq_context);
 		spin_unlock_irqrestore(&chp->comp_handler_lock, flag);
diff --git a/drivers/infiniband/hw/cxgb4/qp.c b/drivers/infiniband/hw/cxgb4/qp.c
index c158fcc02bca..41cd6882b648 100644
--- a/drivers/infiniband/hw/cxgb4/qp.c
+++ b/drivers/infiniband/hw/cxgb4/qp.c
@@ -1105,7 +1105,7 @@ static void __flush_qp(struct c4iw_qp *qhp, struct c4iw_cq *rchp,
 		       struct c4iw_cq *schp)
 {
 	int count;
-	int flushed;
+	int rq_flushed, sq_flushed;
 	unsigned long flag;
 
 	PDBG("%s qhp %p rchp %p schp %p\n", __func__, qhp, rchp, schp);
@@ -1123,27 +1123,40 @@ static void __flush_qp(struct c4iw_qp *qhp, struct c4iw_cq *rchp,
 
 	c4iw_flush_hw_cq(rchp);
 	c4iw_count_rcqes(&rchp->cq, &qhp->wq, &count);
-	flushed = c4iw_flush_rq(&qhp->wq, &rchp->cq, count);
+	rq_flushed = c4iw_flush_rq(&qhp->wq, &rchp->cq, count);
 	spin_unlock(&qhp->lock);
 	spin_unlock_irqrestore(&rchp->lock, flag);
-	if (flushed) {
-		spin_lock_irqsave(&rchp->comp_handler_lock, flag);
-		(*rchp->ibcq.comp_handler)(&rchp->ibcq, rchp->ibcq.cq_context);
-		spin_unlock_irqrestore(&rchp->comp_handler_lock, flag);
-	}
 
 	/* locking hierarchy: cq lock first, then qp lock. */
 	spin_lock_irqsave(&schp->lock, flag);
 	spin_lock(&qhp->lock);
 	if (schp != rchp)
 		c4iw_flush_hw_cq(schp);
-	flushed = c4iw_flush_sq(qhp);
+	sq_flushed = c4iw_flush_sq(qhp);
 	spin_unlock(&qhp->lock);
 	spin_unlock_irqrestore(&schp->lock, flag);
-	if (flushed) {
-		spin_lock_irqsave(&schp->comp_handler_lock, flag);
-		(*schp->ibcq.comp_handler)(&schp->ibcq, schp->ibcq.cq_context);
-		spin_unlock_irqrestore(&schp->comp_handler_lock, flag);
+
+	if (schp == rchp) {
+		if (t4_clear_cq_armed(&rchp->cq) &&
+		    (rq_flushed || sq_flushed)) {
+			spin_lock_irqsave(&rchp->comp_handler_lock, flag);
+			(*rchp->ibcq.comp_handler)(&rchp->ibcq,
+						   rchp->ibcq.cq_context);
+			spin_unlock_irqrestore(&rchp->comp_handler_lock, flag);
+		}
+	} else {
+		if (t4_clear_cq_armed(&rchp->cq) && rq_flushed) {
+			spin_lock_irqsave(&rchp->comp_handler_lock, flag);
+			(*rchp->ibcq.comp_handler)(&rchp->ibcq,
+						   rchp->ibcq.cq_context);
+			spin_unlock_irqrestore(&rchp->comp_handler_lock, flag);
+		}
+		if (t4_clear_cq_armed(&schp->cq) && sq_flushed) {
+			spin_lock_irqsave(&schp->comp_handler_lock, flag);
+			(*schp->ibcq.comp_handler)(&schp->ibcq,
+						   schp->ibcq.cq_context);
+			spin_unlock_irqrestore(&schp->comp_handler_lock, flag);
+		}
 	}
 }
 
diff --git a/drivers/infiniband/hw/cxgb4/t4.h b/drivers/infiniband/hw/cxgb4/t4.h
index df5edfa31a8f..c04e5134b30c 100644
--- a/drivers/infiniband/hw/cxgb4/t4.h
+++ b/drivers/infiniband/hw/cxgb4/t4.h
@@ -524,6 +524,10 @@ static inline int t4_wq_db_enabled(struct t4_wq *wq)
 	return !wq->rq.queue[wq->rq.size].status.db_off;
 }
 
+enum t4_cq_flags {
+	CQ_ARMED	= 1,
+};
+
 struct t4_cq {
 	struct t4_cqe *queue;
 	dma_addr_t dma_addr;
@@ -544,12 +548,19 @@ struct t4_cq {
 	u16 cidx_inc;
 	u8 gen;
 	u8 error;
+	unsigned long flags;
 };
 
+static inline int t4_clear_cq_armed(struct t4_cq *cq)
+{
+	return test_and_clear_bit(CQ_ARMED, &cq->flags);
+}
+
 static inline int t4_arm_cq(struct t4_cq *cq, int se)
 {
 	u32 val;
 
+	set_bit(CQ_ARMED, &cq->flags);
 	while (cq->cidx_inc > CIDXINC_MASK) {
 		val = SEINTARM(0) | CIDXINC(CIDXINC_MASK) | TIMERREG(7) |
 		      INGRESSQID(cq->cqid);
diff --git a/drivers/infiniband/hw/ipath/ipath_fs.c b/drivers/infiniband/hw/ipath/ipath_fs.c
index e0c404bdc4a8..4977082e081f 100644
--- a/drivers/infiniband/hw/ipath/ipath_fs.c
+++ b/drivers/infiniband/hw/ipath/ipath_fs.c
@@ -82,7 +82,6 @@ static int create_file(const char *name, umode_t mode,
 {
 	int error;
 
-	*dentry = NULL;
 	mutex_lock(&parent->d_inode->i_mutex);
 	*dentry = lookup_one_len(name, parent, strlen(name));
 	if (!IS_ERR(*dentry))
diff --git a/drivers/infiniband/hw/ipath/ipath_mad.c b/drivers/infiniband/hw/ipath/ipath_mad.c
index 43f2d0424d4f..e890e5ba0e01 100644
--- a/drivers/infiniband/hw/ipath/ipath_mad.c
+++ b/drivers/infiniband/hw/ipath/ipath_mad.c
@@ -726,7 +726,7 @@ bail:
  * @dd: the infinipath device
  * @pkeys: the PKEY table
  */
-static int set_pkeys(struct ipath_devdata *dd, u16 *pkeys)
+static int set_pkeys(struct ipath_devdata *dd, u16 *pkeys, u8 port)
 {
 	struct ipath_portdata *pd;
 	int i;
@@ -759,6 +759,7 @@ static int set_pkeys(struct ipath_devdata *dd, u16 *pkeys)
 	}
 	if (changed) {
 		u64 pkey;
+		struct ib_event event;
 
 		pkey = (u64) dd->ipath_pkeys[0] |
 			((u64) dd->ipath_pkeys[1] << 16) |
@@ -768,12 +769,17 @@ static int set_pkeys(struct ipath_devdata *dd, u16 *pkeys)
 			   (unsigned long long) pkey);
 		ipath_write_kreg(dd, dd->ipath_kregs->kr_partitionkey,
 				 pkey);
+
+		event.event = IB_EVENT_PKEY_CHANGE;
+		event.device = &dd->verbs_dev->ibdev;
+		event.element.port_num = port;
+		ib_dispatch_event(&event);
 	}
 	return 0;
 }
 
 static int recv_subn_set_pkeytable(struct ib_smp *smp,
-				   struct ib_device *ibdev)
+				   struct ib_device *ibdev, u8 port)
 {
 	u32 startpx = 32 * (be32_to_cpu(smp->attr_mod) & 0xffff);
 	__be16 *p = (__be16 *) smp->data;
@@ -784,7 +790,7 @@ static int recv_subn_set_pkeytable(struct ib_smp *smp,
 	for (i = 0; i < n; i++)
 		q[i] = be16_to_cpu(p[i]);
 
-	if (startpx != 0 || set_pkeys(dev->dd, q) != 0)
+	if (startpx != 0 || set_pkeys(dev->dd, q, port) != 0)
 		smp->status |= IB_SMP_INVALID_FIELD;
 
 	return recv_subn_get_pkeytable(smp, ibdev);
@@ -1342,7 +1348,7 @@ static int process_subn(struct ib_device *ibdev, int mad_flags,
 			ret = recv_subn_set_portinfo(smp, ibdev, port_num);
 			goto bail;
 		case IB_SMP_ATTR_PKEY_TABLE:
-			ret = recv_subn_set_pkeytable(smp, ibdev);
+			ret = recv_subn_set_pkeytable(smp, ibdev, port_num);
 			goto bail;
 		case IB_SMP_ATTR_SM_INFO:
 			if (dev->port_cap_flags & IB_PORT_SM_DISABLED) {
diff --git a/drivers/infiniband/hw/ipath/ipath_user_pages.c b/drivers/infiniband/hw/ipath/ipath_user_pages.c
index dc66c4506916..1da1252dcdb3 100644
--- a/drivers/infiniband/hw/ipath/ipath_user_pages.c
+++ b/drivers/infiniband/hw/ipath/ipath_user_pages.c
@@ -54,7 +54,7 @@ static void __ipath_release_user_pages(struct page **p, size_t num_pages,
 
 /* call with current->mm->mmap_sem held */
 static int __ipath_get_user_pages(unsigned long start_page, size_t num_pages,
-				  struct page **p, struct vm_area_struct **vma)
+				  struct page **p)
 {
 	unsigned long lock_limit;
 	size_t got;
@@ -74,7 +74,7 @@ static int __ipath_get_user_pages(unsigned long start_page, size_t num_pages,
 		ret = get_user_pages(current, current->mm,
 				     start_page + got * PAGE_SIZE,
 				     num_pages - got, 1, 1,
-				     p + got, vma);
+				     p + got, NULL);
 		if (ret < 0)
 			goto bail_release;
 	}
@@ -165,7 +165,7 @@ int ipath_get_user_pages(unsigned long start_page, size_t num_pages,
 
 	down_write(&current->mm->mmap_sem);
 
-	ret = __ipath_get_user_pages(start_page, num_pages, p, NULL);
+	ret = __ipath_get_user_pages(start_page, num_pages, p);
 
 	up_write(&current->mm->mmap_sem);
 
diff --git a/drivers/infiniband/hw/mlx4/mad.c b/drivers/infiniband/hw/mlx4/mad.c
index 287ad0564acd..82a7dd87089b 100644
--- a/drivers/infiniband/hw/mlx4/mad.c
+++ b/drivers/infiniband/hw/mlx4/mad.c
@@ -891,7 +891,7 @@ int mlx4_ib_mad_init(struct mlx4_ib_dev *dev)
 				agent = ib_register_mad_agent(&dev->ib_dev, p + 1,
 							      q ? IB_QPT_GSI : IB_QPT_SMI,
 							      NULL, 0, send_handler,
-							      NULL, NULL);
+							      NULL, NULL, 0);
 				if (IS_ERR(agent)) {
 					ret = PTR_ERR(agent);
 					goto err;
diff --git a/drivers/infiniband/hw/mlx4/main.c b/drivers/infiniband/hw/mlx4/main.c
index 0f7027e7db13..8b72cf392b34 100644
--- a/drivers/infiniband/hw/mlx4/main.c
+++ b/drivers/infiniband/hw/mlx4/main.c
@@ -59,6 +59,7 @@
 
 #define MLX4_IB_FLOW_MAX_PRIO 0xFFF
 #define MLX4_IB_FLOW_QPN_MASK 0xFFFFFF
+#define MLX4_IB_CARD_REV_A0   0xA0
 
 MODULE_AUTHOR("Roland Dreier");
 MODULE_DESCRIPTION("Mellanox ConnectX HCA InfiniBand driver");
@@ -119,6 +120,17 @@ static int check_flow_steering_support(struct mlx4_dev *dev)
 	return dmfs;
 }
 
+static int num_ib_ports(struct mlx4_dev *dev)
+{
+	int ib_ports = 0;
+	int i;
+
+	mlx4_foreach_port(i, dev, MLX4_PORT_TYPE_IB)
+		ib_ports++;
+
+	return ib_ports;
+}
+
 static int mlx4_ib_query_device(struct ib_device *ibdev,
 				struct ib_device_attr *props)
 {
@@ -126,6 +138,7 @@ static int mlx4_ib_query_device(struct ib_device *ibdev,
 	struct ib_smp *in_mad  = NULL;
 	struct ib_smp *out_mad = NULL;
 	int err = -ENOMEM;
+	int have_ib_ports;
 
 	in_mad  = kzalloc(sizeof *in_mad, GFP_KERNEL);
 	out_mad = kmalloc(sizeof *out_mad, GFP_KERNEL);
@@ -142,6 +155,8 @@ static int mlx4_ib_query_device(struct ib_device *ibdev,
 
 	memset(props, 0, sizeof *props);
 
+	have_ib_ports = num_ib_ports(dev->dev);
+
 	props->fw_ver = dev->dev->caps.fw_ver;
 	props->device_cap_flags    = IB_DEVICE_CHANGE_PHY_PORT |
 		IB_DEVICE_PORT_ACTIVE_EVENT		|
@@ -152,13 +167,15 @@ static int mlx4_ib_query_device(struct ib_device *ibdev,
 		props->device_cap_flags |= IB_DEVICE_BAD_PKEY_CNTR;
 	if (dev->dev->caps.flags & MLX4_DEV_CAP_FLAG_BAD_QKEY_CNTR)
 		props->device_cap_flags |= IB_DEVICE_BAD_QKEY_CNTR;
-	if (dev->dev->caps.flags & MLX4_DEV_CAP_FLAG_APM)
+	if (dev->dev->caps.flags & MLX4_DEV_CAP_FLAG_APM && have_ib_ports)
 		props->device_cap_flags |= IB_DEVICE_AUTO_PATH_MIG;
 	if (dev->dev->caps.flags & MLX4_DEV_CAP_FLAG_UD_AV_PORT)
 		props->device_cap_flags |= IB_DEVICE_UD_AV_PORT_ENFORCE;
 	if (dev->dev->caps.flags & MLX4_DEV_CAP_FLAG_IPOIB_CSUM)
 		props->device_cap_flags |= IB_DEVICE_UD_IP_CSUM;
-	if (dev->dev->caps.max_gso_sz && dev->dev->caps.flags & MLX4_DEV_CAP_FLAG_BLH)
+	if (dev->dev->caps.max_gso_sz &&
+	    (dev->dev->rev_id != MLX4_IB_CARD_REV_A0) &&
+	    (dev->dev->caps.flags & MLX4_DEV_CAP_FLAG_BLH))
 		props->device_cap_flags |= IB_DEVICE_UD_TSO;
 	if (dev->dev->caps.bmme_flags & MLX4_BMME_FLAG_RESERVED_LKEY)
 		props->device_cap_flags |= IB_DEVICE_LOCAL_DMA_LKEY;
@@ -357,7 +374,7 @@ static int eth_link_query_port(struct ib_device *ibdev, u8 port,
 	props->state		= IB_PORT_DOWN;
 	props->phys_state	= state_to_phys_state(props->state);
 	props->active_mtu	= IB_MTU_256;
-	spin_lock(&iboe->lock);
+	spin_lock_bh(&iboe->lock);
 	ndev = iboe->netdevs[port - 1];
 	if (!ndev)
 		goto out_unlock;
@@ -369,7 +386,7 @@ static int eth_link_query_port(struct ib_device *ibdev, u8 port,
 					IB_PORT_ACTIVE : IB_PORT_DOWN;
 	props->phys_state	= state_to_phys_state(props->state);
 out_unlock:
-	spin_unlock(&iboe->lock);
+	spin_unlock_bh(&iboe->lock);
 out:
 	mlx4_free_cmd_mailbox(mdev->dev, mailbox);
 	return err;
@@ -811,11 +828,11 @@ int mlx4_ib_add_mc(struct mlx4_ib_dev *mdev, struct mlx4_ib_qp *mqp,
 	if (!mqp->port)
 		return 0;
 
-	spin_lock(&mdev->iboe.lock);
+	spin_lock_bh(&mdev->iboe.lock);
 	ndev = mdev->iboe.netdevs[mqp->port - 1];
 	if (ndev)
 		dev_hold(ndev);
-	spin_unlock(&mdev->iboe.lock);
+	spin_unlock_bh(&mdev->iboe.lock);
 
 	if (ndev) {
 		ret = 1;
@@ -910,8 +927,7 @@ static int __mlx4_ib_default_rules_match(struct ib_qp *qp,
 	const struct default_rules *pdefault_rules = default_table;
 	u8 link_layer = rdma_port_get_link_layer(qp->device, flow_attr->port);
 
-	for (i = 0; i < sizeof(default_table)/sizeof(default_table[0]); i++,
-	     pdefault_rules++) {
+	for (i = 0; i < ARRAY_SIZE(default_table); i++, pdefault_rules++) {
 		__u32 field_types[IB_FLOW_SPEC_SUPPORT_LAYERS];
 		memset(&field_types, 0, sizeof(field_types));
 
@@ -965,8 +981,7 @@ static int __mlx4_ib_create_default_rules(
 	int size = 0;
 	int i;
 
-	for (i = 0; i < sizeof(pdefault_rules->rules_create_list)/
-			sizeof(pdefault_rules->rules_create_list[0]); i++) {
+	for (i = 0; i < ARRAY_SIZE(pdefault_rules->rules_create_list); i++) {
 		int ret;
 		union ib_flow_spec ib_spec;
 		switch (pdefault_rules->rules_create_list[i]) {
@@ -1091,6 +1106,30 @@ static int __mlx4_ib_destroy_flow(struct mlx4_dev *dev, u64 reg_id)
 	return err;
 }
 
+static int mlx4_ib_tunnel_steer_add(struct ib_qp *qp, struct ib_flow_attr *flow_attr,
+				    u64 *reg_id)
+{
+	void *ib_flow;
+	union ib_flow_spec *ib_spec;
+	struct mlx4_dev	*dev = to_mdev(qp->device)->dev;
+	int err = 0;
+
+	if (dev->caps.tunnel_offload_mode != MLX4_TUNNEL_OFFLOAD_MODE_VXLAN)
+		return 0; /* do nothing */
+
+	ib_flow = flow_attr + 1;
+	ib_spec = (union ib_flow_spec *)ib_flow;
+
+	if (ib_spec->type !=  IB_FLOW_SPEC_ETH || flow_attr->num_of_specs != 1)
+		return 0; /* do nothing */
+
+	err = mlx4_tunnel_steer_add(to_mdev(qp->device)->dev, ib_spec->eth.val.dst_mac,
+				    flow_attr->port, qp->qp_num,
+				    MLX4_DOMAIN_UVERBS | (flow_attr->priority & 0xff),
+				    reg_id);
+	return err;
+}
+
 static struct ib_flow *mlx4_ib_create_flow(struct ib_qp *qp,
 				    struct ib_flow_attr *flow_attr,
 				    int domain)
@@ -1134,12 +1173,24 @@ static struct ib_flow *mlx4_ib_create_flow(struct ib_qp *qp,
 		err = __mlx4_ib_create_flow(qp, flow_attr, domain, type[i],
 					    &mflow->reg_id[i]);
 		if (err)
-			goto err_free;
+			goto err_create_flow;
+		i++;
+	}
+
+	if (i < ARRAY_SIZE(type) && flow_attr->type == IB_FLOW_ATTR_NORMAL) {
+		err = mlx4_ib_tunnel_steer_add(qp, flow_attr, &mflow->reg_id[i]);
+		if (err)
+			goto err_create_flow;
 		i++;
 	}
 
 	return &mflow->ibflow;
 
+err_create_flow:
+	while (i) {
+		(void)__mlx4_ib_destroy_flow(to_mdev(qp->device)->dev, mflow->reg_id[i]);
+		i--;
+	}
 err_free:
 	kfree(mflow);
 	return ERR_PTR(err);
@@ -1264,11 +1315,11 @@ static int mlx4_ib_mcg_detach(struct ib_qp *ibqp, union ib_gid *gid, u16 lid)
 	mutex_lock(&mqp->mutex);
 	ge = find_gid_entry(mqp, gid->raw);
 	if (ge) {
-		spin_lock(&mdev->iboe.lock);
+		spin_lock_bh(&mdev->iboe.lock);
 		ndev = ge->added ? mdev->iboe.netdevs[ge->port - 1] : NULL;
 		if (ndev)
 			dev_hold(ndev);
-		spin_unlock(&mdev->iboe.lock);
+		spin_unlock_bh(&mdev->iboe.lock);
 		if (ndev)
 			dev_put(ndev);
 		list_del(&ge->list);
@@ -1389,6 +1440,9 @@ static void update_gids_task(struct work_struct *work)
 	int err;
 	struct mlx4_dev	*dev = gw->dev->dev;
 
+	if (!gw->dev->ib_active)
+		return;
+
 	mailbox = mlx4_alloc_cmd_mailbox(dev);
 	if (IS_ERR(mailbox)) {
 		pr_warn("update gid table failed %ld\n", PTR_ERR(mailbox));
@@ -1419,6 +1473,9 @@ static void reset_gids_task(struct work_struct *work)
 	int err;
 	struct mlx4_dev	*dev = gw->dev->dev;
 
+	if (!gw->dev->ib_active)
+		return;
+
 	mailbox = mlx4_alloc_cmd_mailbox(dev);
 	if (IS_ERR(mailbox)) {
 		pr_warn("reset gid table failed\n");
@@ -1553,7 +1610,7 @@ static int mlx4_ib_addr_event(int event, struct net_device *event_netdev,
 		return 0;
 
 	iboe = &ibdev->iboe;
-	spin_lock(&iboe->lock);
+	spin_lock_bh(&iboe->lock);
 
 	for (port = 1; port <= ibdev->dev->caps.num_ports; ++port)
 		if ((netif_is_bond_master(real_dev) &&
@@ -1563,7 +1620,7 @@ static int mlx4_ib_addr_event(int event, struct net_device *event_netdev,
 			update_gid_table(ibdev, port, gid,
 					 event == NETDEV_DOWN, 0);
 
-	spin_unlock(&iboe->lock);
+	spin_unlock_bh(&iboe->lock);
 	return 0;
 
 }
@@ -1636,13 +1693,21 @@ static void mlx4_ib_update_qps(struct mlx4_ib_dev *ibdev,
 	new_smac = mlx4_mac_to_u64(dev->dev_addr);
 	read_unlock(&dev_base_lock);
 
+	atomic64_set(&ibdev->iboe.mac[port - 1], new_smac);
+
+	/* no need for update QP1 and mac registration in non-SRIOV */
+	if (!mlx4_is_mfunc(ibdev->dev))
+		return;
+
 	mutex_lock(&ibdev->qp1_proxy_lock[port - 1]);
 	qp = ibdev->qp1_proxy[port - 1];
 	if (qp) {
 		int new_smac_index;
-		u64 old_smac = qp->pri.smac;
+		u64 old_smac;
 		struct mlx4_update_qp_params update_params;
 
+		mutex_lock(&qp->mutex);
+		old_smac = qp->pri.smac;
 		if (new_smac == old_smac)
 			goto unlock;
 
@@ -1652,22 +1717,25 @@ static void mlx4_ib_update_qps(struct mlx4_ib_dev *ibdev,
 			goto unlock;
 
 		update_params.smac_index = new_smac_index;
-		if (mlx4_update_qp(ibdev->dev, &qp->mqp, MLX4_UPDATE_QP_SMAC,
+		if (mlx4_update_qp(ibdev->dev, qp->mqp.qpn, MLX4_UPDATE_QP_SMAC,
 				   &update_params)) {
 			release_mac = new_smac;
 			goto unlock;
 		}
-
+		/* if old port was zero, no mac was yet registered for this QP */
+		if (qp->pri.smac_port)
+			release_mac = old_smac;
 		qp->pri.smac = new_smac;
+		qp->pri.smac_port = port;
 		qp->pri.smac_index = new_smac_index;
-
-		release_mac = old_smac;
 	}
 
 unlock:
-	mutex_unlock(&ibdev->qp1_proxy_lock[port - 1]);
 	if (release_mac != MLX4_IB_INVALID_MAC)
 		mlx4_unregister_mac(ibdev->dev, port, release_mac);
+	if (qp)
+		mutex_unlock(&qp->mutex);
+	mutex_unlock(&ibdev->qp1_proxy_lock[port - 1]);
 }
 
 static void mlx4_ib_get_dev_addr(struct net_device *dev,
@@ -1678,6 +1746,7 @@ static void mlx4_ib_get_dev_addr(struct net_device *dev,
 	struct inet6_dev *in6_dev;
 	union ib_gid  *pgid;
 	struct inet6_ifaddr *ifp;
+	union ib_gid default_gid;
 #endif
 	union ib_gid gid;
 
@@ -1698,12 +1767,15 @@ static void mlx4_ib_get_dev_addr(struct net_device *dev,
 		in_dev_put(in_dev);
 	}
 #if IS_ENABLED(CONFIG_IPV6)
+	mlx4_make_default_gid(dev, &default_gid);
 	/* IPv6 gids */
 	in6_dev = in6_dev_get(dev);
 	if (in6_dev) {
 		read_lock_bh(&in6_dev->lock);
 		list_for_each_entry(ifp, &in6_dev->addr_list, if_list) {
 			pgid = (union ib_gid *)&ifp->addr;
+			if (!memcmp(pgid, &default_gid, sizeof(*pgid)))
+				continue;
 			update_gid_table(ibdev, port, pgid, 0, 0);
 		}
 		read_unlock_bh(&in6_dev->lock);
@@ -1725,24 +1797,33 @@ static int mlx4_ib_init_gid_table(struct mlx4_ib_dev *ibdev)
 	struct	net_device *dev;
 	struct mlx4_ib_iboe *iboe = &ibdev->iboe;
 	int i;
+	int err = 0;
 
-	for (i = 1; i <= ibdev->num_ports; ++i)
-		if (reset_gid_table(ibdev, i))
-			return -1;
+	for (i = 1; i <= ibdev->num_ports; ++i) {
+		if (rdma_port_get_link_layer(&ibdev->ib_dev, i) ==
+		    IB_LINK_LAYER_ETHERNET) {
+			err = reset_gid_table(ibdev, i);
+			if (err)
+				goto out;
+		}
+	}
 
 	read_lock(&dev_base_lock);
-	spin_lock(&iboe->lock);
+	spin_lock_bh(&iboe->lock);
 
 	for_each_netdev(&init_net, dev) {
 		u8 port = mlx4_ib_get_dev_port(dev, ibdev);
-		if (port)
+		/* port will be non-zero only for ETH ports */
+		if (port) {
+			mlx4_ib_set_default_gid(ibdev, dev, port);
 			mlx4_ib_get_dev_addr(dev, ibdev, port);
+		}
 	}
 
-	spin_unlock(&iboe->lock);
+	spin_unlock_bh(&iboe->lock);
 	read_unlock(&dev_base_lock);
-
-	return 0;
+out:
+	return err;
 }
 
 static void mlx4_ib_scan_netdevs(struct mlx4_ib_dev *ibdev,
@@ -1756,7 +1837,7 @@ static void mlx4_ib_scan_netdevs(struct mlx4_ib_dev *ibdev,
 
 	iboe = &ibdev->iboe;
 
-	spin_lock(&iboe->lock);
+	spin_lock_bh(&iboe->lock);
 	mlx4_foreach_ib_transport_port(port, ibdev->dev) {
 		enum ib_port_state	port_state = IB_PORT_NOP;
 		struct net_device *old_master = iboe->masters[port - 1];
@@ -1788,35 +1869,47 @@ static void mlx4_ib_scan_netdevs(struct mlx4_ib_dev *ibdev,
 			port_state = (netif_running(curr_netdev) && netif_carrier_ok(curr_netdev)) ?
 						IB_PORT_ACTIVE : IB_PORT_DOWN;
 			mlx4_ib_set_default_gid(ibdev, curr_netdev, port);
-		} else {
-			reset_gid_table(ibdev, port);
-		}
-		/* if using bonding/team and a slave port is down, we don't the bond IP
-		 * based gids in the table since flows that select port by gid may get
-		 * the down port.
-		 */
-		if (curr_master && (port_state == IB_PORT_DOWN)) {
-			reset_gid_table(ibdev, port);
-			mlx4_ib_set_default_gid(ibdev, curr_netdev, port);
-		}
-		/* if bonding is used it is possible that we add it to masters
-		 * only after IP address is assigned to the net bonding
-		 * interface.
-		*/
-		if (curr_master && (old_master != curr_master)) {
-			reset_gid_table(ibdev, port);
-			mlx4_ib_set_default_gid(ibdev, curr_netdev, port);
-			mlx4_ib_get_dev_addr(curr_master, ibdev, port);
-		}
+			if (curr_master) {
+				/* if using bonding/team and a slave port is down, we
+				 * don't want the bond IP based gids in the table since
+				 * flows that select port by gid may get the down port.
+				*/
+				if (port_state == IB_PORT_DOWN) {
+					reset_gid_table(ibdev, port);
+					mlx4_ib_set_default_gid(ibdev,
+								curr_netdev,
+								port);
+				} else {
+					/* gids from the upper dev (bond/team)
+					 * should appear in port's gid table
+					*/
+					mlx4_ib_get_dev_addr(curr_master,
+							     ibdev, port);
+				}
+			}
+			/* if bonding is used it is possible that we add it to
+			 * masters only after IP address is assigned to the
+			 * net bonding interface.
+			*/
+			if (curr_master && (old_master != curr_master)) {
+				reset_gid_table(ibdev, port);
+				mlx4_ib_set_default_gid(ibdev,
+							curr_netdev, port);
+				mlx4_ib_get_dev_addr(curr_master, ibdev, port);
+			}
 
-		if (!curr_master && (old_master != curr_master)) {
+			if (!curr_master && (old_master != curr_master)) {
+				reset_gid_table(ibdev, port);
+				mlx4_ib_set_default_gid(ibdev,
+							curr_netdev, port);
+				mlx4_ib_get_dev_addr(curr_netdev, ibdev, port);
+			}
+		} else {
 			reset_gid_table(ibdev, port);
-			mlx4_ib_set_default_gid(ibdev, curr_netdev, port);
-			mlx4_ib_get_dev_addr(curr_netdev, ibdev, port);
 		}
 	}
 
-	spin_unlock(&iboe->lock);
+	spin_unlock_bh(&iboe->lock);
 
 	if (update_qps_port > 0)
 		mlx4_ib_update_qps(ibdev, dev, update_qps_port);
@@ -2007,6 +2100,7 @@ static void *mlx4_ib_add(struct mlx4_dev *dev)
 		(1ull << IB_USER_VERBS_CMD_ALLOC_PD)		|
 		(1ull << IB_USER_VERBS_CMD_DEALLOC_PD)		|
 		(1ull << IB_USER_VERBS_CMD_REG_MR)		|
+		(1ull << IB_USER_VERBS_CMD_REREG_MR)		|
 		(1ull << IB_USER_VERBS_CMD_DEREG_MR)		|
 		(1ull << IB_USER_VERBS_CMD_CREATE_COMP_CHANNEL)	|
 		(1ull << IB_USER_VERBS_CMD_CREATE_CQ)		|
@@ -2059,6 +2153,7 @@ static void *mlx4_ib_add(struct mlx4_dev *dev)
 	ibdev->ib_dev.req_notify_cq	= mlx4_ib_arm_cq;
 	ibdev->ib_dev.get_dma_mr	= mlx4_ib_get_dma_mr;
 	ibdev->ib_dev.reg_user_mr	= mlx4_ib_reg_user_mr;
+	ibdev->ib_dev.rereg_user_mr	= mlx4_ib_rereg_user_mr;
 	ibdev->ib_dev.dereg_mr		= mlx4_ib_dereg_mr;
 	ibdev->ib_dev.alloc_fast_reg_mr = mlx4_ib_alloc_fast_reg_mr;
 	ibdev->ib_dev.alloc_fast_reg_page_list = mlx4_ib_alloc_fast_reg_page_list;
@@ -2156,6 +2251,9 @@ static void *mlx4_ib_add(struct mlx4_dev *dev)
 			goto err_steer_free_bitmap;
 	}
 
+	for (j = 1; j <= ibdev->dev->caps.num_ports; j++)
+		atomic64_set(&iboe->mac[j - 1], ibdev->dev->caps.def_mac[j]);
+
 	if (ib_register_device(&ibdev->ib_dev, NULL))
 		goto err_steer_free_bitmap;
 
@@ -2192,12 +2290,8 @@ static void *mlx4_ib_add(struct mlx4_dev *dev)
 			}
 		}
 #endif
-		for (i = 1 ; i <= ibdev->num_ports ; ++i)
-			reset_gid_table(ibdev, i);
-		rtnl_lock();
-		mlx4_ib_scan_netdevs(ibdev, NULL, 0);
-		rtnl_unlock();
-		mlx4_ib_init_gid_table(ibdev);
+		if (mlx4_ib_init_gid_table(ibdev))
+			goto err_notif;
 	}
 
 	for (j = 0; j < ARRAY_SIZE(mlx4_class_attributes); ++j) {
@@ -2345,6 +2439,9 @@ static void mlx4_ib_remove(struct mlx4_dev *dev, void *ibdev_ptr)
 	struct mlx4_ib_dev *ibdev = ibdev_ptr;
 	int p;
 
+	ibdev->ib_active = false;
+	flush_workqueue(wq);
+
 	mlx4_ib_close_sriov(ibdev);
 	mlx4_ib_mad_cleanup(ibdev);
 	ib_unregister_device(&ibdev->ib_dev);
diff --git a/drivers/infiniband/hw/mlx4/mlx4_ib.h b/drivers/infiniband/hw/mlx4/mlx4_ib.h
index 369da3ca5d64..6eb743f65f6f 100644
--- a/drivers/infiniband/hw/mlx4/mlx4_ib.h
+++ b/drivers/infiniband/hw/mlx4/mlx4_ib.h
@@ -451,6 +451,7 @@ struct mlx4_ib_iboe {
 	spinlock_t		lock;
 	struct net_device      *netdevs[MLX4_MAX_PORTS];
 	struct net_device      *masters[MLX4_MAX_PORTS];
+	atomic64_t		mac[MLX4_MAX_PORTS];
 	struct notifier_block 	nb;
 	struct notifier_block	nb_inet;
 	struct notifier_block	nb_inet6;
@@ -788,5 +789,9 @@ int mlx4_ib_steer_qp_alloc(struct mlx4_ib_dev *dev, int count, int *qpn);
 void mlx4_ib_steer_qp_free(struct mlx4_ib_dev *dev, u32 qpn, int count);
 int mlx4_ib_steer_qp_reg(struct mlx4_ib_dev *mdev, struct mlx4_ib_qp *mqp,
 			 int is_attach);
+int mlx4_ib_rereg_user_mr(struct ib_mr *mr, int flags,
+			  u64 start, u64 length, u64 virt_addr,
+			  int mr_access_flags, struct ib_pd *pd,
+			  struct ib_udata *udata);
 
 #endif /* MLX4_IB_H */
diff --git a/drivers/infiniband/hw/mlx4/mr.c b/drivers/infiniband/hw/mlx4/mr.c
index cb2a8727f3fb..8f9325cfc85d 100644
--- a/drivers/infiniband/hw/mlx4/mr.c
+++ b/drivers/infiniband/hw/mlx4/mr.c
@@ -144,8 +144,10 @@ struct ib_mr *mlx4_ib_reg_user_mr(struct ib_pd *pd, u64 start, u64 length,
 	if (!mr)
 		return ERR_PTR(-ENOMEM);
 
+	/* Force registering the memory as writable. */
+	/* Used for memory re-registeration. HCA protects the access */
 	mr->umem = ib_umem_get(pd->uobject->context, start, length,
-			       access_flags, 0);
+			       access_flags | IB_ACCESS_LOCAL_WRITE, 0);
 	if (IS_ERR(mr->umem)) {
 		err = PTR_ERR(mr->umem);
 		goto err_free;
@@ -183,6 +185,93 @@ err_free:
 	return ERR_PTR(err);
 }
 
+int mlx4_ib_rereg_user_mr(struct ib_mr *mr, int flags,
+			  u64 start, u64 length, u64 virt_addr,
+			  int mr_access_flags, struct ib_pd *pd,
+			  struct ib_udata *udata)
+{
+	struct mlx4_ib_dev *dev = to_mdev(mr->device);
+	struct mlx4_ib_mr *mmr = to_mmr(mr);
+	struct mlx4_mpt_entry *mpt_entry;
+	struct mlx4_mpt_entry **pmpt_entry = &mpt_entry;
+	int err;
+
+	/* Since we synchronize this call and mlx4_ib_dereg_mr via uverbs,
+	 * we assume that the calls can't run concurrently. Otherwise, a
+	 * race exists.
+	 */
+	err =  mlx4_mr_hw_get_mpt(dev->dev, &mmr->mmr, &pmpt_entry);
+
+	if (err)
+		return err;
+
+	if (flags & IB_MR_REREG_PD) {
+		err = mlx4_mr_hw_change_pd(dev->dev, *pmpt_entry,
+					   to_mpd(pd)->pdn);
+
+		if (err)
+			goto release_mpt_entry;
+	}
+
+	if (flags & IB_MR_REREG_ACCESS) {
+		err = mlx4_mr_hw_change_access(dev->dev, *pmpt_entry,
+					       convert_access(mr_access_flags));
+
+		if (err)
+			goto release_mpt_entry;
+	}
+
+	if (flags & IB_MR_REREG_TRANS) {
+		int shift;
+		int err;
+		int n;
+
+		mlx4_mr_rereg_mem_cleanup(dev->dev, &mmr->mmr);
+		ib_umem_release(mmr->umem);
+		mmr->umem = ib_umem_get(mr->uobject->context, start, length,
+					mr_access_flags |
+					IB_ACCESS_LOCAL_WRITE,
+					0);
+		if (IS_ERR(mmr->umem)) {
+			err = PTR_ERR(mmr->umem);
+			/* Prevent mlx4_ib_dereg_mr from free'ing invalid pointer */
+			mmr->umem = NULL;
+			goto release_mpt_entry;
+		}
+		n = ib_umem_page_count(mmr->umem);
+		shift = ilog2(mmr->umem->page_size);
+
+		err = mlx4_mr_rereg_mem_write(dev->dev, &mmr->mmr,
+					      virt_addr, length, n, shift,
+					      *pmpt_entry);
+		if (err) {
+			ib_umem_release(mmr->umem);
+			goto release_mpt_entry;
+		}
+		mmr->mmr.iova       = virt_addr;
+		mmr->mmr.size       = length;
+
+		err = mlx4_ib_umem_write_mtt(dev, &mmr->mmr.mtt, mmr->umem);
+		if (err) {
+			mlx4_mr_rereg_mem_cleanup(dev->dev, &mmr->mmr);
+			ib_umem_release(mmr->umem);
+			goto release_mpt_entry;
+		}
+	}
+
+	/* If we couldn't transfer the MR to the HCA, just remember to
+	 * return a failure. But dereg_mr will free the resources.
+	 */
+	err = mlx4_mr_hw_write_mpt(dev->dev, &mmr->mmr, pmpt_entry);
+	if (!err && flags & IB_MR_REREG_ACCESS)
+		mmr->mmr.access = mr_access_flags;
+
+release_mpt_entry:
+	mlx4_mr_hw_put_mpt(dev->dev, pmpt_entry);
+
+	return err;
+}
+
 int mlx4_ib_dereg_mr(struct ib_mr *ibmr)
 {
 	struct mlx4_ib_mr *mr = to_mmr(ibmr);
diff --git a/drivers/infiniband/hw/mlx4/qp.c b/drivers/infiniband/hw/mlx4/qp.c
index 67780452f0cf..9c5150c3cb31 100644
--- a/drivers/infiniband/hw/mlx4/qp.c
+++ b/drivers/infiniband/hw/mlx4/qp.c
@@ -964,9 +964,10 @@ static void destroy_qp_common(struct mlx4_ib_dev *dev, struct mlx4_ib_qp *qp,
 				   MLX4_QP_STATE_RST, NULL, 0, 0, &qp->mqp))
 			pr_warn("modify QP %06x to RESET failed.\n",
 			       qp->mqp.qpn);
-		if (qp->pri.smac) {
+		if (qp->pri.smac || (!qp->pri.smac && qp->pri.smac_port)) {
 			mlx4_unregister_mac(dev->dev, qp->pri.smac_port, qp->pri.smac);
 			qp->pri.smac = 0;
+			qp->pri.smac_port = 0;
 		}
 		if (qp->alt.smac) {
 			mlx4_unregister_mac(dev->dev, qp->alt.smac_port, qp->alt.smac);
@@ -1325,7 +1326,8 @@ static int _mlx4_set_path(struct mlx4_ib_dev *dev, const struct ib_ah_attr *ah,
 		 * If one was already assigned, but the new mac differs,
 		 * unregister the old one and register the new one.
 		*/
-		if (!smac_info->smac || smac_info->smac != smac) {
+		if ((!smac_info->smac && !smac_info->smac_port) ||
+		    smac_info->smac != smac) {
 			/* register candidate now, unreg if needed, after success */
 			smac_index = mlx4_register_mac(dev->dev, port, smac);
 			if (smac_index >= 0) {
@@ -1390,21 +1392,13 @@ static void update_mcg_macs(struct mlx4_ib_dev *dev, struct mlx4_ib_qp *qp)
 static int handle_eth_ud_smac_index(struct mlx4_ib_dev *dev, struct mlx4_ib_qp *qp, u8 *smac,
 				    struct mlx4_qp_context *context)
 {
-	struct net_device *ndev;
 	u64 u64_mac;
 	int smac_index;
 
-
-	ndev = dev->iboe.netdevs[qp->port - 1];
-	if (ndev) {
-		smac = ndev->dev_addr;
-		u64_mac = mlx4_mac_to_u64(smac);
-	} else {
-		u64_mac = dev->dev->caps.def_mac[qp->port];
-	}
+	u64_mac = atomic64_read(&dev->iboe.mac[qp->port - 1]);
 
 	context->pri_path.sched_queue = MLX4_IB_DEFAULT_SCHED_QUEUE | ((qp->port - 1) << 6);
-	if (!qp->pri.smac) {
+	if (!qp->pri.smac && !qp->pri.smac_port) {
 		smac_index = mlx4_register_mac(dev->dev, qp->port, u64_mac);
 		if (smac_index >= 0) {
 			qp->pri.candidate_smac_index = smac_index;
@@ -1432,6 +1426,12 @@ static int __mlx4_ib_modify_qp(struct ib_qp *ibqp,
 	int steer_qp = 0;
 	int err = -EINVAL;
 
+	/* APM is not supported under RoCE */
+	if (attr_mask & IB_QP_ALT_PATH &&
+	    rdma_port_get_link_layer(&dev->ib_dev, qp->port) ==
+	    IB_LINK_LAYER_ETHERNET)
+		return -ENOTSUPP;
+
 	context = kzalloc(sizeof *context, GFP_KERNEL);
 	if (!context)
 		return -ENOMEM;
@@ -1677,9 +1677,15 @@ static int __mlx4_ib_modify_qp(struct ib_qp *ibqp,
 		}
 	}
 
-	if (qp->ibqp.qp_type == IB_QPT_RAW_PACKET)
+	if (qp->ibqp.qp_type == IB_QPT_RAW_PACKET) {
 		context->pri_path.ackto = (context->pri_path.ackto & 0xf8) |
 					MLX4_IB_LINK_TYPE_ETH;
+		if (dev->dev->caps.tunnel_offload_mode ==  MLX4_TUNNEL_OFFLOAD_MODE_VXLAN) {
+			/* set QP to receive both tunneled & non-tunneled packets */
+			if (!(context->flags & cpu_to_be32(1 << MLX4_RSS_QPC_FLAG_OFFSET)))
+				context->srqn = cpu_to_be32(7 << 28);
+		}
+	}
 
 	if (ibqp->qp_type == IB_QPT_UD && (new_state == IB_QPS_RTR)) {
 		int is_eth = rdma_port_get_link_layer(
@@ -1780,9 +1786,10 @@ static int __mlx4_ib_modify_qp(struct ib_qp *ibqp,
 			if (qp->flags & MLX4_IB_QP_NETIF)
 				mlx4_ib_steer_qp_reg(dev, qp, 0);
 		}
-		if (qp->pri.smac) {
+		if (qp->pri.smac || (!qp->pri.smac && qp->pri.smac_port)) {
 			mlx4_unregister_mac(dev->dev, qp->pri.smac_port, qp->pri.smac);
 			qp->pri.smac = 0;
+			qp->pri.smac_port = 0;
 		}
 		if (qp->alt.smac) {
 			mlx4_unregister_mac(dev->dev, qp->alt.smac_port, qp->alt.smac);
@@ -1806,11 +1813,12 @@ out:
 	if (err && steer_qp)
 		mlx4_ib_steer_qp_reg(dev, qp, 0);
 	kfree(context);
-	if (qp->pri.candidate_smac) {
+	if (qp->pri.candidate_smac ||
+	    (!qp->pri.candidate_smac && qp->pri.candidate_smac_port)) {
 		if (err) {
 			mlx4_unregister_mac(dev->dev, qp->pri.candidate_smac_port, qp->pri.candidate_smac);
 		} else {
-			if (qp->pri.smac)
+			if (qp->pri.smac || (!qp->pri.smac && qp->pri.smac_port))
 				mlx4_unregister_mac(dev->dev, qp->pri.smac_port, qp->pri.smac);
 			qp->pri.smac = qp->pri.candidate_smac;
 			qp->pri.smac_index = qp->pri.candidate_smac_index;
@@ -2083,6 +2091,16 @@ static int build_sriov_qp0_header(struct mlx4_ib_sqp *sqp,
 	return 0;
 }
 
+static void mlx4_u64_to_smac(u8 *dst_mac, u64 src_mac)
+{
+	int i;
+
+	for (i = ETH_ALEN; i; i--) {
+		dst_mac[i - 1] = src_mac & 0xff;
+		src_mac >>= 8;
+	}
+}
+
 static int build_mlx_header(struct mlx4_ib_sqp *sqp, struct ib_send_wr *wr,
 			    void *wqe, unsigned *mlx_seg_len)
 {
@@ -2197,7 +2215,6 @@ static int build_mlx_header(struct mlx4_ib_sqp *sqp, struct ib_send_wr *wr,
 	}
 
 	if (is_eth) {
-		u8 *smac;
 		struct in6_addr in6;
 
 		u16 pcp = (be32_to_cpu(ah->av.ib.sl_tclass_flowlabel) >> 29) << 13;
@@ -2210,12 +2227,17 @@ static int build_mlx_header(struct mlx4_ib_sqp *sqp, struct ib_send_wr *wr,
 		memcpy(&ctrl->imm, ah->av.eth.mac + 2, 4);
 		memcpy(&in6, sgid.raw, sizeof(in6));
 
-		if (!mlx4_is_mfunc(to_mdev(ib_dev)->dev))
-			smac = to_mdev(sqp->qp.ibqp.device)->
-				iboe.netdevs[sqp->qp.port - 1]->dev_addr;
-		else	/* use the src mac of the tunnel */
-			smac = ah->av.eth.s_mac;
-		memcpy(sqp->ud_header.eth.smac_h, smac, 6);
+		if (!mlx4_is_mfunc(to_mdev(ib_dev)->dev)) {
+			u64 mac = atomic64_read(&to_mdev(ib_dev)->iboe.mac[sqp->qp.port - 1]);
+			u8 smac[ETH_ALEN];
+
+			mlx4_u64_to_smac(smac, mac);
+			memcpy(sqp->ud_header.eth.smac_h, smac, ETH_ALEN);
+		} else {
+			/* use the src mac of the tunnel */
+			memcpy(sqp->ud_header.eth.smac_h, ah->av.eth.s_mac, ETH_ALEN);
+		}
+
 		if (!memcmp(sqp->ud_header.eth.smac_h, sqp->ud_header.eth.dmac_h, 6))
 			mlx->flags |= cpu_to_be32(MLX4_WQE_CTRL_FORCE_LOOPBACK);
 		if (!is_vlan) {
diff --git a/drivers/infiniband/hw/mlx5/cq.c b/drivers/infiniband/hw/mlx5/cq.c
index e4056279166d..10cfce5119a9 100644
--- a/drivers/infiniband/hw/mlx5/cq.c
+++ b/drivers/infiniband/hw/mlx5/cq.c
@@ -752,7 +752,7 @@ struct ib_cq *mlx5_ib_create_cq(struct ib_device *ibdev, int entries,
 		return ERR_PTR(-EINVAL);
 
 	entries = roundup_pow_of_two(entries + 1);
-	if (entries > dev->mdev->caps.max_cqes)
+	if (entries > dev->mdev->caps.gen.max_cqes)
 		return ERR_PTR(-EINVAL);
 
 	cq = kzalloc(sizeof(*cq), GFP_KERNEL);
@@ -919,7 +919,7 @@ int mlx5_ib_modify_cq(struct ib_cq *cq, u16 cq_count, u16 cq_period)
 	int err;
 	u32 fsel;
 
-	if (!(dev->mdev->caps.flags & MLX5_DEV_CAP_FLAG_CQ_MODER))
+	if (!(dev->mdev->caps.gen.flags & MLX5_DEV_CAP_FLAG_CQ_MODER))
 		return -ENOSYS;
 
 	in = kzalloc(sizeof(*in), GFP_KERNEL);
@@ -1074,7 +1074,7 @@ int mlx5_ib_resize_cq(struct ib_cq *ibcq, int entries, struct ib_udata *udata)
 	int uninitialized_var(cqe_size);
 	unsigned long flags;
 
-	if (!(dev->mdev->caps.flags & MLX5_DEV_CAP_FLAG_RESIZE_CQ)) {
+	if (!(dev->mdev->caps.gen.flags & MLX5_DEV_CAP_FLAG_RESIZE_CQ)) {
 		pr_info("Firmware does not support resize CQ\n");
 		return -ENOSYS;
 	}
@@ -1083,7 +1083,7 @@ int mlx5_ib_resize_cq(struct ib_cq *ibcq, int entries, struct ib_udata *udata)
 		return -EINVAL;
 
 	entries = roundup_pow_of_two(entries + 1);
-	if (entries > dev->mdev->caps.max_cqes + 1)
+	if (entries > dev->mdev->caps.gen.max_cqes + 1)
 		return -EINVAL;
 
 	if (entries == ibcq->cqe + 1)
diff --git a/drivers/infiniband/hw/mlx5/mad.c b/drivers/infiniband/hw/mlx5/mad.c
index b514bbb5610f..657af9a1167c 100644
--- a/drivers/infiniband/hw/mlx5/mad.c
+++ b/drivers/infiniband/hw/mlx5/mad.c
@@ -129,7 +129,7 @@ int mlx5_query_ext_port_caps(struct mlx5_ib_dev *dev, u8 port)
 
 	packet_error = be16_to_cpu(out_mad->status);
 
-	dev->mdev->caps.ext_port_cap[port - 1] = (!err && !packet_error) ?
+	dev->mdev->caps.gen.ext_port_cap[port - 1] = (!err && !packet_error) ?
 		MLX_EXT_PORT_CAP_FLAG_EXTENDED_PORT_INFO : 0;
 
 out:
diff --git a/drivers/infiniband/hw/mlx5/main.c b/drivers/infiniband/hw/mlx5/main.c
index d8907b20522a..1ba6c42e4df8 100644
--- a/drivers/infiniband/hw/mlx5/main.c
+++ b/drivers/infiniband/hw/mlx5/main.c
@@ -157,11 +157,13 @@ static int mlx5_ib_query_device(struct ib_device *ibdev,
 	struct mlx5_ib_dev *dev = to_mdev(ibdev);
 	struct ib_smp *in_mad  = NULL;
 	struct ib_smp *out_mad = NULL;
+	struct mlx5_general_caps *gen;
 	int err = -ENOMEM;
 	int max_rq_sg;
 	int max_sq_sg;
 	u64 flags;
 
+	gen = &dev->mdev->caps.gen;
 	in_mad  = kzalloc(sizeof(*in_mad), GFP_KERNEL);
 	out_mad = kmalloc(sizeof(*out_mad), GFP_KERNEL);
 	if (!in_mad || !out_mad)
@@ -183,7 +185,7 @@ static int mlx5_ib_query_device(struct ib_device *ibdev,
 		IB_DEVICE_PORT_ACTIVE_EVENT		|
 		IB_DEVICE_SYS_IMAGE_GUID		|
 		IB_DEVICE_RC_RNR_NAK_GEN;
-	flags = dev->mdev->caps.flags;
+	flags = gen->flags;
 	if (flags & MLX5_DEV_CAP_FLAG_BAD_PKEY_CNTR)
 		props->device_cap_flags |= IB_DEVICE_BAD_PKEY_CNTR;
 	if (flags & MLX5_DEV_CAP_FLAG_BAD_QKEY_CNTR)
@@ -213,30 +215,31 @@ static int mlx5_ib_query_device(struct ib_device *ibdev,
 	memcpy(&props->sys_image_guid, out_mad->data +	4, 8);
 
 	props->max_mr_size	   = ~0ull;
-	props->page_size_cap	   = dev->mdev->caps.min_page_sz;
-	props->max_qp		   = 1 << dev->mdev->caps.log_max_qp;
-	props->max_qp_wr	   = dev->mdev->caps.max_wqes;
-	max_rq_sg = dev->mdev->caps.max_rq_desc_sz / sizeof(struct mlx5_wqe_data_seg);
-	max_sq_sg = (dev->mdev->caps.max_sq_desc_sz - sizeof(struct mlx5_wqe_ctrl_seg)) /
+	props->page_size_cap	   = gen->min_page_sz;
+	props->max_qp		   = 1 << gen->log_max_qp;
+	props->max_qp_wr	   = gen->max_wqes;
+	max_rq_sg = gen->max_rq_desc_sz / sizeof(struct mlx5_wqe_data_seg);
+	max_sq_sg = (gen->max_sq_desc_sz - sizeof(struct mlx5_wqe_ctrl_seg)) /
 		sizeof(struct mlx5_wqe_data_seg);
 	props->max_sge = min(max_rq_sg, max_sq_sg);
-	props->max_cq		   = 1 << dev->mdev->caps.log_max_cq;
-	props->max_cqe		   = dev->mdev->caps.max_cqes - 1;
-	props->max_mr		   = 1 << dev->mdev->caps.log_max_mkey;
-	props->max_pd		   = 1 << dev->mdev->caps.log_max_pd;
-	props->max_qp_rd_atom	   = dev->mdev->caps.max_ra_req_qp;
-	props->max_qp_init_rd_atom = dev->mdev->caps.max_ra_res_qp;
+	props->max_cq		   = 1 << gen->log_max_cq;
+	props->max_cqe		   = gen->max_cqes - 1;
+	props->max_mr		   = 1 << gen->log_max_mkey;
+	props->max_pd		   = 1 << gen->log_max_pd;
+	props->max_qp_rd_atom	   = 1 << gen->log_max_ra_req_qp;
+	props->max_qp_init_rd_atom = 1 << gen->log_max_ra_res_qp;
+	props->max_srq		   = 1 << gen->log_max_srq;
+	props->max_srq_wr	   = gen->max_srq_wqes - 1;
+	props->local_ca_ack_delay  = gen->local_ca_ack_delay;
 	props->max_res_rd_atom	   = props->max_qp_rd_atom * props->max_qp;
-	props->max_srq		   = 1 << dev->mdev->caps.log_max_srq;
-	props->max_srq_wr	   = dev->mdev->caps.max_srq_wqes - 1;
 	props->max_srq_sge	   = max_rq_sg - 1;
 	props->max_fast_reg_page_list_len = (unsigned int)-1;
-	props->local_ca_ack_delay  = dev->mdev->caps.local_ca_ack_delay;
+	props->local_ca_ack_delay  = gen->local_ca_ack_delay;
 	props->atomic_cap	   = IB_ATOMIC_NONE;
 	props->masked_atomic_cap   = IB_ATOMIC_NONE;
 	props->max_pkeys	   = be16_to_cpup((__be16 *)(out_mad->data + 28));
-	props->max_mcast_grp	   = 1 << dev->mdev->caps.log_max_mcg;
-	props->max_mcast_qp_attach = dev->mdev->caps.max_qp_mcg;
+	props->max_mcast_grp	   = 1 << gen->log_max_mcg;
+	props->max_mcast_qp_attach = gen->max_qp_mcg;
 	props->max_total_mcast_qp_attach = props->max_mcast_qp_attach *
 					   props->max_mcast_grp;
 	props->max_map_per_fmr = INT_MAX; /* no limit in ConnectIB */
@@ -254,10 +257,12 @@ int mlx5_ib_query_port(struct ib_device *ibdev, u8 port,
 	struct mlx5_ib_dev *dev = to_mdev(ibdev);
 	struct ib_smp *in_mad  = NULL;
 	struct ib_smp *out_mad = NULL;
+	struct mlx5_general_caps *gen;
 	int ext_active_speed;
 	int err = -ENOMEM;
 
-	if (port < 1 || port > dev->mdev->caps.num_ports) {
+	gen = &dev->mdev->caps.gen;
+	if (port < 1 || port > gen->num_ports) {
 		mlx5_ib_warn(dev, "invalid port number %d\n", port);
 		return -EINVAL;
 	}
@@ -288,8 +293,8 @@ int mlx5_ib_query_port(struct ib_device *ibdev, u8 port,
 	props->phys_state	= out_mad->data[33] >> 4;
 	props->port_cap_flags	= be32_to_cpup((__be32 *)(out_mad->data + 20));
 	props->gid_tbl_len	= out_mad->data[50];
-	props->max_msg_sz	= 1 << to_mdev(ibdev)->mdev->caps.log_max_msg;
-	props->pkey_tbl_len	= to_mdev(ibdev)->mdev->caps.port[port - 1].pkey_table_len;
+	props->max_msg_sz	= 1 << gen->log_max_msg;
+	props->pkey_tbl_len	= gen->port[port - 1].pkey_table_len;
 	props->bad_pkey_cntr	= be16_to_cpup((__be16 *)(out_mad->data + 46));
 	props->qkey_viol_cntr	= be16_to_cpup((__be16 *)(out_mad->data + 48));
 	props->active_width	= out_mad->data[31] & 0xf;
@@ -316,7 +321,7 @@ int mlx5_ib_query_port(struct ib_device *ibdev, u8 port,
 
 	/* If reported active speed is QDR, check if is FDR-10 */
 	if (props->active_speed == 4) {
-		if (dev->mdev->caps.ext_port_cap[port - 1] &
+		if (gen->ext_port_cap[port - 1] &
 		    MLX_EXT_PORT_CAP_FLAG_EXTENDED_PORT_INFO) {
 			init_query_mad(in_mad);
 			in_mad->attr_id = MLX5_ATTR_EXTENDED_PORT_INFO;
@@ -470,6 +475,7 @@ static struct ib_ucontext *mlx5_ib_alloc_ucontext(struct ib_device *ibdev,
 	struct mlx5_ib_alloc_ucontext_req_v2 req;
 	struct mlx5_ib_alloc_ucontext_resp resp;
 	struct mlx5_ib_ucontext *context;
+	struct mlx5_general_caps *gen;
 	struct mlx5_uuar_info *uuari;
 	struct mlx5_uar *uars;
 	int gross_uuars;
@@ -480,6 +486,7 @@ static struct ib_ucontext *mlx5_ib_alloc_ucontext(struct ib_device *ibdev,
 	int i;
 	size_t reqlen;
 
+	gen = &dev->mdev->caps.gen;
 	if (!dev->ib_active)
 		return ERR_PTR(-EAGAIN);
 
@@ -512,14 +519,14 @@ static struct ib_ucontext *mlx5_ib_alloc_ucontext(struct ib_device *ibdev,
 
 	num_uars = req.total_num_uuars / MLX5_NON_FP_BF_REGS_PER_PAGE;
 	gross_uuars = num_uars * MLX5_BF_REGS_PER_PAGE;
-	resp.qp_tab_size      = 1 << dev->mdev->caps.log_max_qp;
-	resp.bf_reg_size      = dev->mdev->caps.bf_reg_size;
+	resp.qp_tab_size      = 1 << gen->log_max_qp;
+	resp.bf_reg_size      = gen->bf_reg_size;
 	resp.cache_line_size  = L1_CACHE_BYTES;
-	resp.max_sq_desc_sz = dev->mdev->caps.max_sq_desc_sz;
-	resp.max_rq_desc_sz = dev->mdev->caps.max_rq_desc_sz;
-	resp.max_send_wqebb = dev->mdev->caps.max_wqes;
-	resp.max_recv_wr = dev->mdev->caps.max_wqes;
-	resp.max_srq_recv_wr = dev->mdev->caps.max_srq_wqes;
+	resp.max_sq_desc_sz = gen->max_sq_desc_sz;
+	resp.max_rq_desc_sz = gen->max_rq_desc_sz;
+	resp.max_send_wqebb = gen->max_wqes;
+	resp.max_recv_wr = gen->max_wqes;
+	resp.max_srq_recv_wr = gen->max_srq_wqes;
 
 	context = kzalloc(sizeof(*context), GFP_KERNEL);
 	if (!context)
@@ -565,7 +572,7 @@ static struct ib_ucontext *mlx5_ib_alloc_ucontext(struct ib_device *ibdev,
 	mutex_init(&context->db_page_mutex);
 
 	resp.tot_uuars = req.total_num_uuars;
-	resp.num_ports = dev->mdev->caps.num_ports;
+	resp.num_ports = gen->num_ports;
 	err = ib_copy_to_udata(udata, &resp,
 			       sizeof(resp) - sizeof(resp.reserved));
 	if (err)
@@ -650,13 +657,13 @@ static int mlx5_ib_mmap(struct ib_ucontext *ibcontext, struct vm_area_struct *vm
 			return -EINVAL;
 
 		idx = get_index(vma->vm_pgoff);
+		if (idx >= uuari->num_uars)
+			return -EINVAL;
+
 		pfn = uar_index2pfn(dev, uuari->uars[idx].index);
 		mlx5_ib_dbg(dev, "uar idx 0x%lx, pfn 0x%llx\n", idx,
 			    (unsigned long long)pfn);
 
-		if (idx >= uuari->num_uars)
-			return -EINVAL;
-
 		vma->vm_page_prot = pgprot_writecombine(vma->vm_page_prot);
 		if (io_remap_pfn_range(vma, vma->vm_start, pfn,
 				       PAGE_SIZE, vma->vm_page_prot))
@@ -967,9 +974,11 @@ static void mlx5_ib_event(struct mlx5_core_dev *dev, void *context,
 
 static void get_ext_port_caps(struct mlx5_ib_dev *dev)
 {
+	struct mlx5_general_caps *gen;
 	int port;
 
-	for (port = 1; port <= dev->mdev->caps.num_ports; port++)
+	gen = &dev->mdev->caps.gen;
+	for (port = 1; port <= gen->num_ports; port++)
 		mlx5_query_ext_port_caps(dev, port);
 }
 
@@ -977,9 +986,11 @@ static int get_port_caps(struct mlx5_ib_dev *dev)
 {
 	struct ib_device_attr *dprops = NULL;
 	struct ib_port_attr *pprops = NULL;
+	struct mlx5_general_caps *gen;
 	int err = 0;
 	int port;
 
+	gen = &dev->mdev->caps.gen;
 	pprops = kmalloc(sizeof(*pprops), GFP_KERNEL);
 	if (!pprops)
 		goto out;
@@ -994,14 +1005,14 @@ static int get_port_caps(struct mlx5_ib_dev *dev)
 		goto out;
 	}
 
-	for (port = 1; port <= dev->mdev->caps.num_ports; port++) {
+	for (port = 1; port <= gen->num_ports; port++) {
 		err = mlx5_ib_query_port(&dev->ib_dev, port, pprops);
 		if (err) {
 			mlx5_ib_warn(dev, "query_port %d failed %d\n", port, err);
 			break;
 		}
-		dev->mdev->caps.port[port - 1].pkey_table_len = dprops->max_pkeys;
-		dev->mdev->caps.port[port - 1].gid_table_len = pprops->gid_tbl_len;
+		gen->port[port - 1].pkey_table_len = dprops->max_pkeys;
+		gen->port[port - 1].gid_table_len = pprops->gid_tbl_len;
 		mlx5_ib_dbg(dev, "pkey_table_len %d, gid_table_len %d\n",
 			    dprops->max_pkeys, pprops->gid_tbl_len);
 	}
@@ -1279,8 +1290,8 @@ static void *mlx5_ib_add(struct mlx5_core_dev *mdev)
 	strlcpy(dev->ib_dev.name, "mlx5_%d", IB_DEVICE_NAME_MAX);
 	dev->ib_dev.owner		= THIS_MODULE;
 	dev->ib_dev.node_type		= RDMA_NODE_IB_CA;
-	dev->ib_dev.local_dma_lkey	= mdev->caps.reserved_lkey;
-	dev->num_ports		= mdev->caps.num_ports;
+	dev->ib_dev.local_dma_lkey	= mdev->caps.gen.reserved_lkey;
+	dev->num_ports		= mdev->caps.gen.num_ports;
 	dev->ib_dev.phys_port_cnt     = dev->num_ports;
 	dev->ib_dev.num_comp_vectors	= dev->num_comp_vectors;
 	dev->ib_dev.dma_device	= &mdev->pdev->dev;
@@ -1355,7 +1366,7 @@ static void *mlx5_ib_add(struct mlx5_core_dev *mdev)
 	dev->ib_dev.free_fast_reg_page_list  = mlx5_ib_free_fast_reg_page_list;
 	dev->ib_dev.check_mr_status	= mlx5_ib_check_mr_status;
 
-	if (mdev->caps.flags & MLX5_DEV_CAP_FLAG_XRC) {
+	if (mdev->caps.gen.flags & MLX5_DEV_CAP_FLAG_XRC) {
 		dev->ib_dev.alloc_xrcd = mlx5_ib_alloc_xrcd;
 		dev->ib_dev.dealloc_xrcd = mlx5_ib_dealloc_xrcd;
 		dev->ib_dev.uverbs_cmd_mask |=
@@ -1414,8 +1425,8 @@ err_dealloc:
 static void mlx5_ib_remove(struct mlx5_core_dev *mdev, void *context)
 {
 	struct mlx5_ib_dev *dev = context;
-	destroy_umrc_res(dev);
 	ib_unregister_device(&dev->ib_dev);
+	destroy_umrc_res(dev);
 	destroy_dev_resources(&dev->devr);
 	free_comp_eqs(dev);
 	ib_dealloc_device(&dev->ib_dev);
diff --git a/drivers/infiniband/hw/mlx5/mem.c b/drivers/infiniband/hw/mlx5/mem.c
index a3e81444c825..dae07eae9507 100644
--- a/drivers/infiniband/hw/mlx5/mem.c
+++ b/drivers/infiniband/hw/mlx5/mem.c
@@ -55,16 +55,17 @@ void mlx5_ib_cont_pages(struct ib_umem *umem, u64 addr, int *count, int *shift,
 	u64 pfn;
 	struct scatterlist *sg;
 	int entry;
+	unsigned long page_shift = ilog2(umem->page_size);
 
-	addr = addr >> PAGE_SHIFT;
+	addr = addr >> page_shift;
 	tmp = (unsigned long)addr;
 	m = find_first_bit(&tmp, sizeof(tmp));
 	skip = 1 << m;
 	mask = skip - 1;
 	i = 0;
 	for_each_sg(umem->sg_head.sgl, sg, umem->nmap, entry) {
-		len = sg_dma_len(sg) >> PAGE_SHIFT;
-		pfn = sg_dma_address(sg) >> PAGE_SHIFT;
+		len = sg_dma_len(sg) >> page_shift;
+		pfn = sg_dma_address(sg) >> page_shift;
 		for (k = 0; k < len; k++) {
 			if (!(i & mask)) {
 				tmp = (unsigned long)pfn;
@@ -103,14 +104,15 @@ void mlx5_ib_cont_pages(struct ib_umem *umem, u64 addr, int *count, int *shift,
 
 		*ncont = 0;
 	}
-	*shift = PAGE_SHIFT + m;
+	*shift = page_shift + m;
 	*count = i;
 }
 
 void mlx5_ib_populate_pas(struct mlx5_ib_dev *dev, struct ib_umem *umem,
 			  int page_shift, __be64 *pas, int umr)
 {
-	int shift = page_shift - PAGE_SHIFT;
+	unsigned long umem_page_shift = ilog2(umem->page_size);
+	int shift = page_shift - umem_page_shift;
 	int mask = (1 << shift) - 1;
 	int i, k;
 	u64 cur = 0;
@@ -121,11 +123,11 @@ void mlx5_ib_populate_pas(struct mlx5_ib_dev *dev, struct ib_umem *umem,
 
 	i = 0;
 	for_each_sg(umem->sg_head.sgl, sg, umem->nmap, entry) {
-		len = sg_dma_len(sg) >> PAGE_SHIFT;
+		len = sg_dma_len(sg) >> umem_page_shift;
 		base = sg_dma_address(sg);
 		for (k = 0; k < len; k++) {
 			if (!(i & mask)) {
-				cur = base + (k << PAGE_SHIFT);
+				cur = base + (k << umem_page_shift);
 				if (umr)
 					cur |= 3;
 
@@ -134,7 +136,7 @@ void mlx5_ib_populate_pas(struct mlx5_ib_dev *dev, struct ib_umem *umem,
 					    i >> shift, be64_to_cpu(pas[i >> shift]));
 			}  else
 				mlx5_ib_dbg(dev, "=====> 0x%llx\n",
-					    base + (k << PAGE_SHIFT));
+					    base + (k << umem_page_shift));
 			i++;
 		}
 	}
diff --git a/drivers/infiniband/hw/mlx5/mr.c b/drivers/infiniband/hw/mlx5/mr.c
index 80b3c63eab5d..8ee7cb46e059 100644
--- a/drivers/infiniband/hw/mlx5/mr.c
+++ b/drivers/infiniband/hw/mlx5/mr.c
@@ -881,12 +881,12 @@ struct ib_mr *mlx5_ib_reg_user_mr(struct ib_pd *pd, u64 start, u64 length,
 	int order;
 	int err;
 
-	mlx5_ib_dbg(dev, "start 0x%llx, virt_addr 0x%llx, length 0x%llx\n",
-		    start, virt_addr, length);
+	mlx5_ib_dbg(dev, "start 0x%llx, virt_addr 0x%llx, length 0x%llx, access_flags 0x%x\n",
+		    start, virt_addr, length, access_flags);
 	umem = ib_umem_get(pd->uobject->context, start, length, access_flags,
 			   0);
 	if (IS_ERR(umem)) {
-		mlx5_ib_dbg(dev, "umem get failed\n");
+		mlx5_ib_dbg(dev, "umem get failed (%ld)\n", PTR_ERR(umem));
 		return (void *)umem;
 	}
 
diff --git a/drivers/infiniband/hw/mlx5/qp.c b/drivers/infiniband/hw/mlx5/qp.c
index 7efe6e3f3542..e261a53f9a02 100644
--- a/drivers/infiniband/hw/mlx5/qp.c
+++ b/drivers/infiniband/hw/mlx5/qp.c
@@ -158,11 +158,13 @@ static void mlx5_ib_qp_event(struct mlx5_core_qp *qp, int type)
 static int set_rq_size(struct mlx5_ib_dev *dev, struct ib_qp_cap *cap,
 		       int has_rq, struct mlx5_ib_qp *qp, struct mlx5_ib_create_qp *ucmd)
 {
+	struct mlx5_general_caps *gen;
 	int wqe_size;
 	int wq_size;
 
+	gen = &dev->mdev->caps.gen;
 	/* Sanity check RQ size before proceeding */
-	if (cap->max_recv_wr  > dev->mdev->caps.max_wqes)
+	if (cap->max_recv_wr  > gen->max_wqes)
 		return -EINVAL;
 
 	if (!has_rq) {
@@ -182,10 +184,10 @@ static int set_rq_size(struct mlx5_ib_dev *dev, struct ib_qp_cap *cap,
 			wq_size = roundup_pow_of_two(cap->max_recv_wr) * wqe_size;
 			wq_size = max_t(int, wq_size, MLX5_SEND_WQE_BB);
 			qp->rq.wqe_cnt = wq_size / wqe_size;
-			if (wqe_size > dev->mdev->caps.max_rq_desc_sz) {
+			if (wqe_size > gen->max_rq_desc_sz) {
 				mlx5_ib_dbg(dev, "wqe_size %d, max %d\n",
 					    wqe_size,
-					    dev->mdev->caps.max_rq_desc_sz);
+					    gen->max_rq_desc_sz);
 				return -EINVAL;
 			}
 			qp->rq.wqe_shift = ilog2(wqe_size);
@@ -266,9 +268,11 @@ static int calc_send_wqe(struct ib_qp_init_attr *attr)
 static int calc_sq_size(struct mlx5_ib_dev *dev, struct ib_qp_init_attr *attr,
 			struct mlx5_ib_qp *qp)
 {
+	struct mlx5_general_caps *gen;
 	int wqe_size;
 	int wq_size;
 
+	gen = &dev->mdev->caps.gen;
 	if (!attr->cap.max_send_wr)
 		return 0;
 
@@ -277,9 +281,9 @@ static int calc_sq_size(struct mlx5_ib_dev *dev, struct ib_qp_init_attr *attr,
 	if (wqe_size < 0)
 		return wqe_size;
 
-	if (wqe_size > dev->mdev->caps.max_sq_desc_sz) {
+	if (wqe_size > gen->max_sq_desc_sz) {
 		mlx5_ib_dbg(dev, "wqe_size(%d) > max_sq_desc_sz(%d)\n",
-			    wqe_size, dev->mdev->caps.max_sq_desc_sz);
+			    wqe_size, gen->max_sq_desc_sz);
 		return -EINVAL;
 	}
 
@@ -292,9 +296,9 @@ static int calc_sq_size(struct mlx5_ib_dev *dev, struct ib_qp_init_attr *attr,
 
 	wq_size = roundup_pow_of_two(attr->cap.max_send_wr * wqe_size);
 	qp->sq.wqe_cnt = wq_size / MLX5_SEND_WQE_BB;
-	if (qp->sq.wqe_cnt > dev->mdev->caps.max_wqes) {
+	if (qp->sq.wqe_cnt > gen->max_wqes) {
 		mlx5_ib_dbg(dev, "wqe count(%d) exceeds limits(%d)\n",
-			    qp->sq.wqe_cnt, dev->mdev->caps.max_wqes);
+			    qp->sq.wqe_cnt, gen->max_wqes);
 		return -ENOMEM;
 	}
 	qp->sq.wqe_shift = ilog2(MLX5_SEND_WQE_BB);
@@ -309,11 +313,13 @@ static int set_user_buf_size(struct mlx5_ib_dev *dev,
 			    struct mlx5_ib_qp *qp,
 			    struct mlx5_ib_create_qp *ucmd)
 {
+	struct mlx5_general_caps *gen;
 	int desc_sz = 1 << qp->sq.wqe_shift;
 
-	if (desc_sz > dev->mdev->caps.max_sq_desc_sz) {
+	gen = &dev->mdev->caps.gen;
+	if (desc_sz > gen->max_sq_desc_sz) {
 		mlx5_ib_warn(dev, "desc_sz %d, max_sq_desc_sz %d\n",
-			     desc_sz, dev->mdev->caps.max_sq_desc_sz);
+			     desc_sz, gen->max_sq_desc_sz);
 		return -EINVAL;
 	}
 
@@ -325,9 +331,9 @@ static int set_user_buf_size(struct mlx5_ib_dev *dev,
 
 	qp->sq.wqe_cnt = ucmd->sq_wqe_count;
 
-	if (qp->sq.wqe_cnt > dev->mdev->caps.max_wqes) {
+	if (qp->sq.wqe_cnt > gen->max_wqes) {
 		mlx5_ib_warn(dev, "wqe_cnt %d, max_wqes %d\n",
-			     qp->sq.wqe_cnt, dev->mdev->caps.max_wqes);
+			     qp->sq.wqe_cnt, gen->max_wqes);
 		return -EINVAL;
 	}
 
@@ -803,16 +809,18 @@ static int create_qp_common(struct mlx5_ib_dev *dev, struct ib_pd *pd,
 	struct mlx5_ib_resources *devr = &dev->devr;
 	struct mlx5_ib_create_qp_resp resp;
 	struct mlx5_create_qp_mbox_in *in;
+	struct mlx5_general_caps *gen;
 	struct mlx5_ib_create_qp ucmd;
 	int inlen = sizeof(*in);
 	int err;
 
+	gen = &dev->mdev->caps.gen;
 	mutex_init(&qp->mutex);
 	spin_lock_init(&qp->sq.lock);
 	spin_lock_init(&qp->rq.lock);
 
 	if (init_attr->create_flags & IB_QP_CREATE_BLOCK_MULTICAST_LOOPBACK) {
-		if (!(dev->mdev->caps.flags & MLX5_DEV_CAP_FLAG_BLOCK_MCAST)) {
+		if (!(gen->flags & MLX5_DEV_CAP_FLAG_BLOCK_MCAST)) {
 			mlx5_ib_dbg(dev, "block multicast loopback isn't supported\n");
 			return -EINVAL;
 		} else {
@@ -851,9 +859,9 @@ static int create_qp_common(struct mlx5_ib_dev *dev, struct ib_pd *pd,
 				mlx5_ib_dbg(dev, "invalid rq params\n");
 				return -EINVAL;
 			}
-			if (ucmd.sq_wqe_count > dev->mdev->caps.max_wqes) {
+			if (ucmd.sq_wqe_count > gen->max_wqes) {
 				mlx5_ib_dbg(dev, "requested sq_wqe_count (%d) > max allowed (%d)\n",
-					    ucmd.sq_wqe_count, dev->mdev->caps.max_wqes);
+					    ucmd.sq_wqe_count, gen->max_wqes);
 				return -EINVAL;
 			}
 			err = create_user_qp(dev, pd, qp, udata, &in, &resp, &inlen);
@@ -1144,6 +1152,7 @@ struct ib_qp *mlx5_ib_create_qp(struct ib_pd *pd,
 				struct ib_qp_init_attr *init_attr,
 				struct ib_udata *udata)
 {
+	struct mlx5_general_caps *gen;
 	struct mlx5_ib_dev *dev;
 	struct mlx5_ib_qp *qp;
 	u16 xrcdn = 0;
@@ -1161,11 +1170,12 @@ struct ib_qp *mlx5_ib_create_qp(struct ib_pd *pd,
 		}
 		dev = to_mdev(to_mxrcd(init_attr->xrcd)->ibxrcd.device);
 	}
+	gen = &dev->mdev->caps.gen;
 
 	switch (init_attr->qp_type) {
 	case IB_QPT_XRC_TGT:
 	case IB_QPT_XRC_INI:
-		if (!(dev->mdev->caps.flags & MLX5_DEV_CAP_FLAG_XRC)) {
+		if (!(gen->flags & MLX5_DEV_CAP_FLAG_XRC)) {
 			mlx5_ib_dbg(dev, "XRC not supported\n");
 			return ERR_PTR(-ENOSYS);
 		}
@@ -1272,6 +1282,9 @@ enum {
 
 static int ib_rate_to_mlx5(struct mlx5_ib_dev *dev, u8 rate)
 {
+	struct mlx5_general_caps *gen;
+
+	gen = &dev->mdev->caps.gen;
 	if (rate == IB_RATE_PORT_CURRENT) {
 		return 0;
 	} else if (rate < IB_RATE_2_5_GBPS || rate > IB_RATE_300_GBPS) {
@@ -1279,7 +1292,7 @@ static int ib_rate_to_mlx5(struct mlx5_ib_dev *dev, u8 rate)
 	} else {
 		while (rate != IB_RATE_2_5_GBPS &&
 		       !(1 << (rate + MLX5_STAT_RATE_OFFSET) &
-			 dev->mdev->caps.stat_rate_support))
+			 gen->stat_rate_support))
 			--rate;
 	}
 
@@ -1290,8 +1303,10 @@ static int mlx5_set_path(struct mlx5_ib_dev *dev, const struct ib_ah_attr *ah,
 			 struct mlx5_qp_path *path, u8 port, int attr_mask,
 			 u32 path_flags, const struct ib_qp_attr *attr)
 {
+	struct mlx5_general_caps *gen;
 	int err;
 
+	gen = &dev->mdev->caps.gen;
 	path->fl = (path_flags & MLX5_PATH_FLAG_FL) ? 0x80 : 0;
 	path->free_ar = (path_flags & MLX5_PATH_FLAG_FREE_AR) ? 0x80 : 0;
 
@@ -1302,6 +1317,11 @@ static int mlx5_set_path(struct mlx5_ib_dev *dev, const struct ib_ah_attr *ah,
 	path->rlid	= cpu_to_be16(ah->dlid);
 
 	if (ah->ah_flags & IB_AH_GRH) {
+		if (ah->grh.sgid_index >= gen->port[port - 1].gid_table_len) {
+			pr_err(KERN_ERR "sgid_index (%u) too large. max is %d\n",
+			       ah->grh.sgid_index, gen->port[port - 1].gid_table_len);
+			return -EINVAL;
+		}
 		path->grh_mlid |= 1 << 7;
 		path->mgid_index = ah->grh.sgid_index;
 		path->hop_limit  = ah->grh.hop_limit;
@@ -1317,22 +1337,6 @@ static int mlx5_set_path(struct mlx5_ib_dev *dev, const struct ib_ah_attr *ah,
 	path->static_rate = err;
 	path->port = port;
 
-	if (ah->ah_flags & IB_AH_GRH) {
-		if (ah->grh.sgid_index >= dev->mdev->caps.port[port - 1].gid_table_len) {
-			pr_err(KERN_ERR "sgid_index (%u) too large. max is %d\n",
-			       ah->grh.sgid_index, dev->mdev->caps.port[port - 1].gid_table_len);
-			return -EINVAL;
-		}
-
-		path->grh_mlid |= 1 << 7;
-		path->mgid_index = ah->grh.sgid_index;
-		path->hop_limit  = ah->grh.hop_limit;
-		path->tclass_flowlabel =
-			cpu_to_be32((ah->grh.traffic_class << 20) |
-				    (ah->grh.flow_label));
-		memcpy(path->rgid, ah->grh.dgid.raw, 16);
-	}
-
 	if (attr_mask & IB_QP_TIMEOUT)
 		path->ackto_lt = attr->timeout << 3;
 
@@ -1492,6 +1496,7 @@ static int __mlx5_ib_modify_qp(struct ib_qp *ibqp,
 	struct mlx5_ib_qp *qp = to_mqp(ibqp);
 	struct mlx5_ib_cq *send_cq, *recv_cq;
 	struct mlx5_qp_context *context;
+	struct mlx5_general_caps *gen;
 	struct mlx5_modify_qp_mbox_in *in;
 	struct mlx5_ib_pd *pd;
 	enum mlx5_qp_state mlx5_cur, mlx5_new;
@@ -1500,6 +1505,7 @@ static int __mlx5_ib_modify_qp(struct ib_qp *ibqp,
 	int mlx5_st;
 	int err;
 
+	gen = &dev->mdev->caps.gen;
 	in = kzalloc(sizeof(*in), GFP_KERNEL);
 	if (!in)
 		return -ENOMEM;
@@ -1539,7 +1545,7 @@ static int __mlx5_ib_modify_qp(struct ib_qp *ibqp,
 			err = -EINVAL;
 			goto out;
 		}
-		context->mtu_msgmax = (attr->path_mtu << 5) | dev->mdev->caps.log_max_msg;
+		context->mtu_msgmax = (attr->path_mtu << 5) | gen->log_max_msg;
 	}
 
 	if (attr_mask & IB_QP_DEST_QPN)
@@ -1685,9 +1691,11 @@ int mlx5_ib_modify_qp(struct ib_qp *ibqp, struct ib_qp_attr *attr,
 	struct mlx5_ib_dev *dev = to_mdev(ibqp->device);
 	struct mlx5_ib_qp *qp = to_mqp(ibqp);
 	enum ib_qp_state cur_state, new_state;
+	struct mlx5_general_caps *gen;
 	int err = -EINVAL;
 	int port;
 
+	gen = &dev->mdev->caps.gen;
 	mutex_lock(&qp->mutex);
 
 	cur_state = attr_mask & IB_QP_CUR_STATE ? attr->cur_qp_state : qp->state;
@@ -1699,21 +1707,21 @@ int mlx5_ib_modify_qp(struct ib_qp *ibqp, struct ib_qp_attr *attr,
 		goto out;
 
 	if ((attr_mask & IB_QP_PORT) &&
-	    (attr->port_num == 0 || attr->port_num > dev->mdev->caps.num_ports))
+	    (attr->port_num == 0 || attr->port_num > gen->num_ports))
 		goto out;
 
 	if (attr_mask & IB_QP_PKEY_INDEX) {
 		port = attr_mask & IB_QP_PORT ? attr->port_num : qp->port;
-		if (attr->pkey_index >= dev->mdev->caps.port[port - 1].pkey_table_len)
+		if (attr->pkey_index >= gen->port[port - 1].pkey_table_len)
 			goto out;
 	}
 
 	if (attr_mask & IB_QP_MAX_QP_RD_ATOMIC &&
-	    attr->max_rd_atomic > dev->mdev->caps.max_ra_res_qp)
+	    attr->max_rd_atomic > (1 << gen->log_max_ra_res_qp))
 		goto out;
 
 	if (attr_mask & IB_QP_MAX_DEST_RD_ATOMIC &&
-	    attr->max_dest_rd_atomic > dev->mdev->caps.max_ra_req_qp)
+	    attr->max_dest_rd_atomic > (1 << gen->log_max_ra_req_qp))
 		goto out;
 
 	if (cur_state == new_state && cur_state == IB_QPS_RESET) {
@@ -2020,56 +2028,31 @@ static u8 bs_selector(int block_size)
 	}
 }
 
-static int format_selector(struct ib_sig_attrs *attr,
-			   struct ib_sig_domain *domain,
-			   int *selector)
+static void mlx5_fill_inl_bsf(struct ib_sig_domain *domain,
+			      struct mlx5_bsf_inl *inl)
 {
+	/* Valid inline section and allow BSF refresh */
+	inl->vld_refresh = cpu_to_be16(MLX5_BSF_INL_VALID |
+				       MLX5_BSF_REFRESH_DIF);
+	inl->dif_apptag = cpu_to_be16(domain->sig.dif.app_tag);
+	inl->dif_reftag = cpu_to_be32(domain->sig.dif.ref_tag);
+	/* repeating block */
+	inl->rp_inv_seed = MLX5_BSF_REPEAT_BLOCK;
+	inl->sig_type = domain->sig.dif.bg_type == IB_T10DIF_CRC ?
+			MLX5_DIF_CRC : MLX5_DIF_IPCS;
 
-#define FORMAT_DIF_NONE		0
-#define FORMAT_DIF_CRC_INC	8
-#define FORMAT_DIF_CRC_NO_INC	12
-#define FORMAT_DIF_CSUM_INC	13
-#define FORMAT_DIF_CSUM_NO_INC	14
+	if (domain->sig.dif.ref_remap)
+		inl->dif_inc_ref_guard_check |= MLX5_BSF_INC_REFTAG;
 
-	switch (domain->sig.dif.type) {
-	case IB_T10DIF_NONE:
-		/* No DIF */
-		*selector = FORMAT_DIF_NONE;
-		break;
-	case IB_T10DIF_TYPE1: /* Fall through */
-	case IB_T10DIF_TYPE2:
-		switch (domain->sig.dif.bg_type) {
-		case IB_T10DIF_CRC:
-			*selector = FORMAT_DIF_CRC_INC;
-			break;
-		case IB_T10DIF_CSUM:
-			*selector = FORMAT_DIF_CSUM_INC;
-			break;
-		default:
-			return 1;
-		}
-		break;
-	case IB_T10DIF_TYPE3:
-		switch (domain->sig.dif.bg_type) {
-		case IB_T10DIF_CRC:
-			*selector = domain->sig.dif.type3_inc_reftag ?
-					   FORMAT_DIF_CRC_INC :
-					   FORMAT_DIF_CRC_NO_INC;
-			break;
-		case IB_T10DIF_CSUM:
-			*selector = domain->sig.dif.type3_inc_reftag ?
-					   FORMAT_DIF_CSUM_INC :
-					   FORMAT_DIF_CSUM_NO_INC;
-			break;
-		default:
-			return 1;
-		}
-		break;
-	default:
-		return 1;
+	if (domain->sig.dif.app_escape) {
+		if (domain->sig.dif.ref_escape)
+			inl->dif_inc_ref_guard_check |= MLX5_BSF_APPREF_ESCAPE;
+		else
+			inl->dif_inc_ref_guard_check |= MLX5_BSF_APPTAG_ESCAPE;
 	}
 
-	return 0;
+	inl->dif_app_bitmask_check =
+		cpu_to_be16(domain->sig.dif.apptag_check_mask);
 }
 
 static int mlx5_set_bsf(struct ib_mr *sig_mr,
@@ -2080,45 +2063,49 @@ static int mlx5_set_bsf(struct ib_mr *sig_mr,
 	struct mlx5_bsf_basic *basic = &bsf->basic;
 	struct ib_sig_domain *mem = &sig_attrs->mem;
 	struct ib_sig_domain *wire = &sig_attrs->wire;
-	int ret, selector;
 
 	memset(bsf, 0, sizeof(*bsf));
+
+	/* Basic + Extended + Inline */
+	basic->bsf_size_sbs = 1 << 7;
+	/* Input domain check byte mask */
+	basic->check_byte_mask = sig_attrs->check_mask;
+	basic->raw_data_size = cpu_to_be32(data_size);
+
+	/* Memory domain */
 	switch (sig_attrs->mem.sig_type) {
+	case IB_SIG_TYPE_NONE:
+		break;
 	case IB_SIG_TYPE_T10_DIF:
-		if (sig_attrs->wire.sig_type != IB_SIG_TYPE_T10_DIF)
-			return -EINVAL;
+		basic->mem.bs_selector = bs_selector(mem->sig.dif.pi_interval);
+		basic->m_bfs_psv = cpu_to_be32(msig->psv_memory.psv_idx);
+		mlx5_fill_inl_bsf(mem, &bsf->m_inl);
+		break;
+	default:
+		return -EINVAL;
+	}
 
-		/* Input domain check byte mask */
-		basic->check_byte_mask = sig_attrs->check_mask;
+	/* Wire domain */
+	switch (sig_attrs->wire.sig_type) {
+	case IB_SIG_TYPE_NONE:
+		break;
+	case IB_SIG_TYPE_T10_DIF:
 		if (mem->sig.dif.pi_interval == wire->sig.dif.pi_interval &&
-		    mem->sig.dif.type == wire->sig.dif.type) {
+		    mem->sig_type == wire->sig_type) {
 			/* Same block structure */
-			basic->bsf_size_sbs = 1 << 4;
+			basic->bsf_size_sbs |= 1 << 4;
 			if (mem->sig.dif.bg_type == wire->sig.dif.bg_type)
-				basic->wire.copy_byte_mask |= 0xc0;
+				basic->wire.copy_byte_mask |= MLX5_CPY_GRD_MASK;
 			if (mem->sig.dif.app_tag == wire->sig.dif.app_tag)
-				basic->wire.copy_byte_mask |= 0x30;
+				basic->wire.copy_byte_mask |= MLX5_CPY_APP_MASK;
 			if (mem->sig.dif.ref_tag == wire->sig.dif.ref_tag)
-				basic->wire.copy_byte_mask |= 0x0f;
+				basic->wire.copy_byte_mask |= MLX5_CPY_REF_MASK;
 		} else
 			basic->wire.bs_selector = bs_selector(wire->sig.dif.pi_interval);
 
-		basic->mem.bs_selector = bs_selector(mem->sig.dif.pi_interval);
-		basic->raw_data_size = cpu_to_be32(data_size);
-
-		ret = format_selector(sig_attrs, mem, &selector);
-		if (ret)
-			return -EINVAL;
-		basic->m_bfs_psv = cpu_to_be32(selector << 24 |
-					       msig->psv_memory.psv_idx);
-
-		ret = format_selector(sig_attrs, wire, &selector);
-		if (ret)
-			return -EINVAL;
-		basic->w_bfs_psv = cpu_to_be32(selector << 24 |
-					       msig->psv_wire.psv_idx);
+		basic->w_bfs_psv = cpu_to_be32(msig->psv_wire.psv_idx);
+		mlx5_fill_inl_bsf(wire, &bsf->w_inl);
 		break;
-
 	default:
 		return -EINVAL;
 	}
@@ -2317,20 +2304,21 @@ static int set_psv_wr(struct ib_sig_domain *domain,
 	memset(psv_seg, 0, sizeof(*psv_seg));
 	psv_seg->psv_num = cpu_to_be32(psv_idx);
 	switch (domain->sig_type) {
+	case IB_SIG_TYPE_NONE:
+		break;
 	case IB_SIG_TYPE_T10_DIF:
 		psv_seg->transient_sig = cpu_to_be32(domain->sig.dif.bg << 16 |
 						     domain->sig.dif.app_tag);
 		psv_seg->ref_tag = cpu_to_be32(domain->sig.dif.ref_tag);
-
-		*seg += sizeof(*psv_seg);
-		*size += sizeof(*psv_seg) / 16;
 		break;
-
 	default:
 		pr_err("Bad signature type given.\n");
 		return 1;
 	}
 
+	*seg += sizeof(*psv_seg);
+	*size += sizeof(*psv_seg) / 16;
+
 	return 0;
 }
 
@@ -2501,7 +2489,7 @@ int mlx5_ib_post_send(struct ib_qp *ibqp, struct ib_send_wr *wr,
 	spin_lock_irqsave(&qp->sq.lock, flags);
 
 	for (nreq = 0; wr; nreq++, wr = wr->next) {
-		if (unlikely(wr->opcode >= sizeof(mlx5_ib_opcode) / sizeof(mlx5_ib_opcode[0]))) {
+		if (unlikely(wr->opcode >= ARRAY_SIZE(mlx5_ib_opcode))) {
 			mlx5_ib_warn(dev, "\n");
 			err = -EINVAL;
 			*bad_wr = wr;
@@ -2893,7 +2881,8 @@ static void to_ib_ah_attr(struct mlx5_ib_dev *ibdev, struct ib_ah_attr *ib_ah_at
 	memset(ib_ah_attr, 0, sizeof(*ib_ah_attr));
 	ib_ah_attr->port_num	  = path->port;
 
-	if (ib_ah_attr->port_num == 0 || ib_ah_attr->port_num > dev->caps.num_ports)
+	if (ib_ah_attr->port_num == 0 ||
+	    ib_ah_attr->port_num > dev->caps.gen.num_ports)
 		return;
 
 	ib_ah_attr->sl = path->sl & 0xf;
@@ -3011,10 +3000,12 @@ struct ib_xrcd *mlx5_ib_alloc_xrcd(struct ib_device *ibdev,
 					  struct ib_udata *udata)
 {
 	struct mlx5_ib_dev *dev = to_mdev(ibdev);
+	struct mlx5_general_caps *gen;
 	struct mlx5_ib_xrcd *xrcd;
 	int err;
 
-	if (!(dev->mdev->caps.flags & MLX5_DEV_CAP_FLAG_XRC))
+	gen = &dev->mdev->caps.gen;
+	if (!(gen->flags & MLX5_DEV_CAP_FLAG_XRC))
 		return ERR_PTR(-ENOSYS);
 
 	xrcd = kmalloc(sizeof(*xrcd), GFP_KERNEL);
diff --git a/drivers/infiniband/hw/mlx5/srq.c b/drivers/infiniband/hw/mlx5/srq.c
index 70bd131ba646..97cc1baaa8e3 100644
--- a/drivers/infiniband/hw/mlx5/srq.c
+++ b/drivers/infiniband/hw/mlx5/srq.c
@@ -238,6 +238,7 @@ struct ib_srq *mlx5_ib_create_srq(struct ib_pd *pd,
 				  struct ib_udata *udata)
 {
 	struct mlx5_ib_dev *dev = to_mdev(pd->device);
+	struct mlx5_general_caps *gen;
 	struct mlx5_ib_srq *srq;
 	int desc_size;
 	int buf_size;
@@ -247,11 +248,12 @@ struct ib_srq *mlx5_ib_create_srq(struct ib_pd *pd,
 	int is_xrc;
 	u32 flgs, xrcdn;
 
+	gen = &dev->mdev->caps.gen;
 	/* Sanity check SRQ size before proceeding */
-	if (init_attr->attr.max_wr >= dev->mdev->caps.max_srq_wqes) {
+	if (init_attr->attr.max_wr >= gen->max_srq_wqes) {
 		mlx5_ib_dbg(dev, "max_wr %d, cap %d\n",
 			    init_attr->attr.max_wr,
-			    dev->mdev->caps.max_srq_wqes);
+			    gen->max_srq_wqes);
 		return ERR_PTR(-EINVAL);
 	}
 
diff --git a/drivers/infiniband/hw/mthca/mthca_mad.c b/drivers/infiniband/hw/mthca/mthca_mad.c
index b6f7f457fc55..8881fa376e06 100644
--- a/drivers/infiniband/hw/mthca/mthca_mad.c
+++ b/drivers/infiniband/hw/mthca/mthca_mad.c
@@ -294,7 +294,7 @@ int mthca_create_agents(struct mthca_dev *dev)
 			agent = ib_register_mad_agent(&dev->ib_dev, p + 1,
 						      q ? IB_QPT_GSI : IB_QPT_SMI,
 						      NULL, 0, send_handler,
-						      NULL, NULL);
+						      NULL, NULL, 0);
 			if (IS_ERR(agent)) {
 				ret = PTR_ERR(agent);
 				goto err;
diff --git a/drivers/infiniband/hw/nes/nes_hw.c b/drivers/infiniband/hw/nes/nes_hw.c
index 90200245c5eb..02120d340d50 100644
--- a/drivers/infiniband/hw/nes/nes_hw.c
+++ b/drivers/infiniband/hw/nes/nes_hw.c
@@ -1003,13 +1003,13 @@ int nes_init_cqp(struct nes_device *nesdev)
 			(sizeof(struct nes_hw_aeqe) * nesadapter->max_qp) +
 			sizeof(struct nes_hw_cqp_qp_context);
 
-	nesdev->cqp_vbase = pci_alloc_consistent(nesdev->pcidev, nesdev->cqp_mem_size,
-			&nesdev->cqp_pbase);
+	nesdev->cqp_vbase = pci_zalloc_consistent(nesdev->pcidev,
+						  nesdev->cqp_mem_size,
+						  &nesdev->cqp_pbase);
 	if (!nesdev->cqp_vbase) {
 		nes_debug(NES_DBG_INIT, "Unable to allocate memory for host descriptor rings\n");
 		return -ENOMEM;
 	}
-	memset(nesdev->cqp_vbase, 0, nesdev->cqp_mem_size);
 
 	/* Allocate a twice the number of CQP requests as the SQ size */
 	nesdev->nes_cqp_requests = kzalloc(sizeof(struct nes_cqp_request) *
@@ -1691,13 +1691,13 @@ int nes_init_nic_qp(struct nes_device *nesdev, struct net_device *netdev)
 			(NES_NIC_WQ_SIZE * 2 * sizeof(struct nes_hw_nic_cqe)) +
 			sizeof(struct nes_hw_nic_qp_context);
 
-	nesvnic->nic_vbase = pci_alloc_consistent(nesdev->pcidev, nesvnic->nic_mem_size,
-			&nesvnic->nic_pbase);
+	nesvnic->nic_vbase = pci_zalloc_consistent(nesdev->pcidev,
+						   nesvnic->nic_mem_size,
+						   &nesvnic->nic_pbase);
 	if (!nesvnic->nic_vbase) {
 		nes_debug(NES_DBG_INIT, "Unable to allocate memory for NIC host descriptor rings\n");
 		return -ENOMEM;
 	}
-	memset(nesvnic->nic_vbase, 0, nesvnic->nic_mem_size);
 	nes_debug(NES_DBG_INIT, "Allocated NIC QP structures at %p (phys = %016lX), size = %u.\n",
 			nesvnic->nic_vbase, (unsigned long)nesvnic->nic_pbase, nesvnic->nic_mem_size);
 
diff --git a/drivers/infiniband/hw/nes/nes_verbs.c b/drivers/infiniband/hw/nes/nes_verbs.c
index 218dd3574285..fef067c959fc 100644
--- a/drivers/infiniband/hw/nes/nes_verbs.c
+++ b/drivers/infiniband/hw/nes/nes_verbs.c
@@ -1616,8 +1616,8 @@ static struct ib_cq *nes_create_cq(struct ib_device *ibdev, int entries,
 				entries, nescq->cq_mem_size, nescq->hw_cq.cq_number);
 
 		/* allocate the physical buffer space */
-		mem = pci_alloc_consistent(nesdev->pcidev, nescq->cq_mem_size,
-				&nescq->hw_cq.cq_pbase);
+		mem = pci_zalloc_consistent(nesdev->pcidev, nescq->cq_mem_size,
+					    &nescq->hw_cq.cq_pbase);
 		if (!mem) {
 			printk(KERN_ERR PFX "Unable to allocate pci memory for cq\n");
 			nes_free_resource(nesadapter, nesadapter->allocated_cqs, cq_num);
@@ -1625,7 +1625,6 @@ static struct ib_cq *nes_create_cq(struct ib_device *ibdev, int entries,
 			return ERR_PTR(-ENOMEM);
 		}
 
-		memset(mem, 0, nescq->cq_mem_size);
 		nescq->hw_cq.cq_vbase = mem;
 		nescq->hw_cq.cq_head = 0;
 		nes_debug(NES_DBG_CQ, "CQ%u virtual address @ %p, phys = 0x%08X\n",
diff --git a/drivers/infiniband/hw/ocrdma/ocrdma.h b/drivers/infiniband/hw/ocrdma/ocrdma.h
index 19011dbb930f..b43456ae124b 100644
--- a/drivers/infiniband/hw/ocrdma/ocrdma.h
+++ b/drivers/infiniband/hw/ocrdma/ocrdma.h
@@ -40,7 +40,7 @@
 #include <be_roce.h>
 #include "ocrdma_sli.h"
 
-#define OCRDMA_ROCE_DRV_VERSION "10.2.145.0u"
+#define OCRDMA_ROCE_DRV_VERSION "10.2.287.0u"
 
 #define OCRDMA_ROCE_DRV_DESC "Emulex OneConnect RoCE Driver"
 #define OCRDMA_NODE_DESC "Emulex OneConnect RoCE HCA"
@@ -137,6 +137,7 @@ struct mqe_ctx {
 	u16 cqe_status;
 	u16 ext_status;
 	bool cmd_done;
+	bool fw_error_state;
 };
 
 struct ocrdma_hw_mr {
@@ -235,7 +236,10 @@ struct ocrdma_dev {
 	struct list_head entry;
 	struct rcu_head rcu;
 	int id;
-	u64 stag_arr[OCRDMA_MAX_STAG];
+	u64 *stag_arr;
+	u8 sl; /* service level */
+	bool pfc_state;
+	atomic_t update_sl;
 	u16 pvid;
 	u32 asic_id;
 
@@ -518,4 +522,22 @@ static inline u8 ocrdma_get_asic_type(struct ocrdma_dev *dev)
 				OCRDMA_SLI_ASIC_GEN_NUM_SHIFT;
 }
 
+static inline u8 ocrdma_get_pfc_prio(u8 *pfc, u8 prio)
+{
+	return *(pfc + prio);
+}
+
+static inline u8 ocrdma_get_app_prio(u8 *app_prio, u8 prio)
+{
+	return *(app_prio + prio);
+}
+
+static inline u8 ocrdma_is_enabled_and_synced(u32 state)
+{	/* May also be used to interpret TC-state, QCN-state
+	 * Appl-state and Logical-link-state in future.
+	 */
+	return (state & OCRDMA_STATE_FLAG_ENABLED) &&
+		(state & OCRDMA_STATE_FLAG_SYNC);
+}
+
 #endif
diff --git a/drivers/infiniband/hw/ocrdma/ocrdma_ah.c b/drivers/infiniband/hw/ocrdma/ocrdma_ah.c
index d4cc01f10c01..ac02ce4e8040 100644
--- a/drivers/infiniband/hw/ocrdma/ocrdma_ah.c
+++ b/drivers/infiniband/hw/ocrdma/ocrdma_ah.c
@@ -35,8 +35,10 @@
 #include "ocrdma_ah.h"
 #include "ocrdma_hw.h"
 
+#define OCRDMA_VID_PCP_SHIFT	0xD
+
 static inline int set_av_attr(struct ocrdma_dev *dev, struct ocrdma_ah *ah,
-				struct ib_ah_attr *attr, int pdid)
+			struct ib_ah_attr *attr, union ib_gid *sgid, int pdid)
 {
 	int status = 0;
 	u16 vlan_tag; bool vlan_enabled = false;
@@ -47,15 +49,14 @@ static inline int set_av_attr(struct ocrdma_dev *dev, struct ocrdma_ah *ah,
 	memset(&eth, 0, sizeof(eth));
 	memset(&grh, 0, sizeof(grh));
 
-	ah->sgid_index = attr->grh.sgid_index;
-
+	/* VLAN */
 	vlan_tag = attr->vlan_id;
 	if (!vlan_tag || (vlan_tag > 0xFFF))
 		vlan_tag = dev->pvid;
 	if (vlan_tag && (vlan_tag < 0x1000)) {
 		eth.eth_type = cpu_to_be16(0x8100);
 		eth.roce_eth_type = cpu_to_be16(OCRDMA_ROCE_ETH_TYPE);
-		vlan_tag |= (attr->sl & 7) << 13;
+		vlan_tag |= (dev->sl & 0x07) << OCRDMA_VID_PCP_SHIFT;
 		eth.vlan_tag = cpu_to_be16(vlan_tag);
 		eth_sz = sizeof(struct ocrdma_eth_vlan);
 		vlan_enabled = true;
@@ -63,15 +64,14 @@ static inline int set_av_attr(struct ocrdma_dev *dev, struct ocrdma_ah *ah,
 		eth.eth_type = cpu_to_be16(OCRDMA_ROCE_ETH_TYPE);
 		eth_sz = sizeof(struct ocrdma_eth_basic);
 	}
+	/* MAC */
 	memcpy(&eth.smac[0], &dev->nic_info.mac_addr[0], ETH_ALEN);
-	memcpy(&eth.dmac[0], attr->dmac, ETH_ALEN);
 	status = ocrdma_resolve_dmac(dev, attr, &eth.dmac[0]);
 	if (status)
 		return status;
-	status = ocrdma_query_gid(&dev->ibdev, 1, attr->grh.sgid_index,
-			(union ib_gid *)&grh.sgid[0]);
-	if (status)
-		return status;
+	ah->sgid_index = attr->grh.sgid_index;
+	memcpy(&grh.sgid[0], sgid->raw, sizeof(union ib_gid));
+	memcpy(&grh.dgid[0], attr->grh.dgid.raw, sizeof(attr->grh.dgid.raw));
 
 	grh.tclass_flow = cpu_to_be32((6 << 28) |
 			(attr->grh.traffic_class << 24) |
@@ -79,8 +79,7 @@ static inline int set_av_attr(struct ocrdma_dev *dev, struct ocrdma_ah *ah,
 	/* 0x1b is next header value in GRH */
 	grh.pdid_hoplimit = cpu_to_be32((pdid << 16) |
 			(0x1b << 8) | attr->grh.hop_limit);
-
-	memcpy(&grh.dgid[0], attr->grh.dgid.raw, sizeof(attr->grh.dgid.raw));
+	/* Eth HDR */
 	memcpy(&ah->av->eth_hdr, &eth, eth_sz);
 	memcpy((u8 *)ah->av + eth_sz, &grh, sizeof(struct ocrdma_grh));
 	if (vlan_enabled)
@@ -96,10 +95,14 @@ struct ib_ah *ocrdma_create_ah(struct ib_pd *ibpd, struct ib_ah_attr *attr)
 	struct ocrdma_ah *ah;
 	struct ocrdma_pd *pd = get_ocrdma_pd(ibpd);
 	struct ocrdma_dev *dev = get_ocrdma_dev(ibpd->device);
+	union ib_gid sgid;
+	u8 zmac[ETH_ALEN];
 
 	if (!(attr->ah_flags & IB_AH_GRH))
 		return ERR_PTR(-EINVAL);
 
+	if (atomic_cmpxchg(&dev->update_sl, 1, 0))
+		ocrdma_init_service_level(dev);
 	ah = kzalloc(sizeof(*ah), GFP_ATOMIC);
 	if (!ah)
 		return ERR_PTR(-ENOMEM);
@@ -107,7 +110,27 @@ struct ib_ah *ocrdma_create_ah(struct ib_pd *ibpd, struct ib_ah_attr *attr)
 	status = ocrdma_alloc_av(dev, ah);
 	if (status)
 		goto av_err;
-	status = set_av_attr(dev, ah, attr, pd->id);
+
+	status = ocrdma_query_gid(&dev->ibdev, 1, attr->grh.sgid_index, &sgid);
+	if (status) {
+		pr_err("%s(): Failed to query sgid, status = %d\n",
+		      __func__, status);
+		goto av_conf_err;
+	}
+
+	memset(&zmac, 0, ETH_ALEN);
+	if (pd->uctx &&
+	    memcmp(attr->dmac, &zmac, ETH_ALEN)) {
+		status = rdma_addr_find_dmac_by_grh(&sgid, &attr->grh.dgid,
+                                        attr->dmac, &attr->vlan_id);
+		if (status) {
+			pr_err("%s(): Failed to resolve dmac from gid." 
+				"status = %d\n", __func__, status);
+			goto av_conf_err;
+		}
+	}
+
+	status = set_av_attr(dev, ah, attr, &sgid, pd->id);
 	if (status)
 		goto av_conf_err;
 
@@ -141,7 +164,7 @@ int ocrdma_query_ah(struct ib_ah *ibah, struct ib_ah_attr *attr)
 	struct ocrdma_av *av = ah->av;
 	struct ocrdma_grh *grh;
 	attr->ah_flags |= IB_AH_GRH;
-	if (ah->av->valid & Bit(1)) {
+	if (ah->av->valid & OCRDMA_AV_VALID) {
 		grh = (struct ocrdma_grh *)((u8 *)ah->av +
 				sizeof(struct ocrdma_eth_vlan));
 		attr->sl = be16_to_cpu(av->eth_hdr.vlan_tag) >> 13;
diff --git a/drivers/infiniband/hw/ocrdma/ocrdma_hw.c b/drivers/infiniband/hw/ocrdma/ocrdma_hw.c
index 3bbf2010a821..638bff1ffc6c 100644
--- a/drivers/infiniband/hw/ocrdma/ocrdma_hw.c
+++ b/drivers/infiniband/hw/ocrdma/ocrdma_hw.c
@@ -348,11 +348,6 @@ static void *ocrdma_init_emb_mqe(u8 opcode, u32 cmd_len)
 	return mqe;
 }
 
-static void *ocrdma_alloc_mqe(void)
-{
-	return kzalloc(sizeof(struct ocrdma_mqe), GFP_KERNEL);
-}
-
 static void ocrdma_free_q(struct ocrdma_dev *dev, struct ocrdma_queue_info *q)
 {
 	dma_free_coherent(&dev->nic_info.pdev->dev, q->size, q->va, q->dma);
@@ -525,7 +520,7 @@ static int ocrdma_mbx_mq_cq_create(struct ocrdma_dev *dev,
 
 	cmd->ev_cnt_flags = OCRDMA_CREATE_CQ_DEF_FLAGS;
 	cmd->eqn = eq->id;
-	cmd->cqe_count = cq->size / sizeof(struct ocrdma_mcqe);
+	cmd->pdid_cqecnt = cq->size / sizeof(struct ocrdma_mcqe);
 
 	ocrdma_build_q_pages(&cmd->pa[0], cq->size / OCRDMA_MIN_Q_PAGE_SIZE,
 			     cq->dma, PAGE_SIZE_4K);
@@ -566,8 +561,8 @@ static int ocrdma_mbx_create_mq(struct ocrdma_dev *dev,
 	cmd->cqid_pages |= (cq->id << OCRDMA_CREATE_MQ_CQ_ID_SHIFT);
 	cmd->async_cqid_valid = OCRDMA_CREATE_MQ_ASYNC_CQ_VALID;
 
-	cmd->async_event_bitmap = Bit(OCRDMA_ASYNC_GRP5_EVE_CODE);
-	cmd->async_event_bitmap |= Bit(OCRDMA_ASYNC_RDMA_EVE_CODE);
+	cmd->async_event_bitmap = BIT(OCRDMA_ASYNC_GRP5_EVE_CODE);
+	cmd->async_event_bitmap |= BIT(OCRDMA_ASYNC_RDMA_EVE_CODE);
 
 	cmd->async_cqid_ringsize = cq->id;
 	cmd->async_cqid_ringsize |= (ocrdma_encoded_q_len(mq->len) <<
@@ -661,7 +656,7 @@ static void ocrdma_dispatch_ibevent(struct ocrdma_dev *dev,
 {
 	struct ocrdma_qp *qp = NULL;
 	struct ocrdma_cq *cq = NULL;
-	struct ib_event ib_evt = { 0 };
+	struct ib_event ib_evt;
 	int cq_event = 0;
 	int qp_event = 1;
 	int srq_event = 0;
@@ -674,6 +669,8 @@ static void ocrdma_dispatch_ibevent(struct ocrdma_dev *dev,
 	if (cqe->cqvalid_cqid & OCRDMA_AE_MCQE_CQVALID)
 		cq = dev->cq_tbl[cqe->cqvalid_cqid & OCRDMA_AE_MCQE_CQID_MASK];
 
+	memset(&ib_evt, 0, sizeof(ib_evt));
+
 	ib_evt.device = &dev->ibdev;
 
 	switch (type) {
@@ -771,6 +768,10 @@ static void ocrdma_process_grp5_aync(struct ocrdma_dev *dev,
 					OCRDMA_AE_PVID_MCQE_TAG_MASK) >>
 					OCRDMA_AE_PVID_MCQE_TAG_SHIFT);
 		break;
+
+	case OCRDMA_ASYNC_EVENT_COS_VALUE:
+		atomic_set(&dev->update_sl, 1);
+		break;
 	default:
 		/* Not interested evts. */
 		break;
@@ -962,8 +963,12 @@ static int ocrdma_wait_mqe_cmpl(struct ocrdma_dev *dev)
 				    msecs_to_jiffies(30000));
 	if (status)
 		return 0;
-	else
+	else {
+		dev->mqe_ctx.fw_error_state = true;
+		pr_err("%s(%d) mailbox timeout: fw not responding\n",
+		       __func__, dev->id);
 		return -1;
+	}
 }
 
 /* issue a mailbox command on the MQ */
@@ -975,6 +980,8 @@ static int ocrdma_mbx_cmd(struct ocrdma_dev *dev, struct ocrdma_mqe *mqe)
 	struct ocrdma_mbx_rsp *rsp = NULL;
 
 	mutex_lock(&dev->mqe_ctx.lock);
+	if (dev->mqe_ctx.fw_error_state)
+		goto mbx_err;
 	ocrdma_post_mqe(dev, mqe);
 	status = ocrdma_wait_mqe_cmpl(dev);
 	if (status)
@@ -1078,7 +1085,8 @@ static void ocrdma_get_attr(struct ocrdma_dev *dev,
 	    OCRDMA_MBX_QUERY_CFG_CA_ACK_DELAY_SHIFT;
 	attr->max_mw = rsp->max_mw;
 	attr->max_mr = rsp->max_mr;
-	attr->max_mr_size = ~0ull;
+	attr->max_mr_size = ((u64)rsp->max_mr_size_hi << 32) |
+			      rsp->max_mr_size_lo;
 	attr->max_fmr = 0;
 	attr->max_pages_per_frmr = rsp->max_pages_per_frmr;
 	attr->max_num_mr_pbl = rsp->max_num_mr_pbl;
@@ -1176,10 +1184,10 @@ int ocrdma_mbx_rdma_stats(struct ocrdma_dev *dev, bool reset)
 {
 	struct ocrdma_rdma_stats_req *req = dev->stats_mem.va;
 	struct ocrdma_mqe *mqe = &dev->stats_mem.mqe;
-	struct ocrdma_rdma_stats_resp *old_stats = NULL;
+	struct ocrdma_rdma_stats_resp *old_stats;
 	int status;
 
-	old_stats = kzalloc(sizeof(*old_stats), GFP_KERNEL);
+	old_stats = kmalloc(sizeof(*old_stats), GFP_KERNEL);
 	if (old_stats == NULL)
 		return -ENOMEM;
 
@@ -1222,10 +1230,9 @@ static int ocrdma_mbx_get_ctrl_attribs(struct ocrdma_dev *dev)
 	struct ocrdma_get_ctrl_attribs_rsp *ctrl_attr_rsp;
 	struct mgmt_hba_attribs *hba_attribs;
 
-	mqe = ocrdma_alloc_mqe();
+	mqe = kzalloc(sizeof(struct ocrdma_mqe), GFP_KERNEL);
 	if (!mqe)
 		return status;
-	memset(mqe, 0, sizeof(*mqe));
 
 	dma.size = sizeof(struct ocrdma_get_ctrl_attribs_rsp);
 	dma.va	 = dma_alloc_coherent(&dev->nic_info.pdev->dev,
@@ -1252,7 +1259,9 @@ static int ocrdma_mbx_get_ctrl_attribs(struct ocrdma_dev *dev)
 		ctrl_attr_rsp = (struct ocrdma_get_ctrl_attribs_rsp *)dma.va;
 		hba_attribs = &ctrl_attr_rsp->ctrl_attribs.hba_attribs;
 
-		dev->hba_port_num = hba_attribs->phy_port;
+		dev->hba_port_num = (hba_attribs->ptpnum_maxdoms_hbast_cv &
+					OCRDMA_HBA_ATTRB_PTNUM_MASK)
+					>> OCRDMA_HBA_ATTRB_PTNUM_SHIFT;
 		strncpy(dev->model_number,
 			hba_attribs->controller_model_number, 31);
 	}
@@ -1302,7 +1311,8 @@ int ocrdma_mbx_get_link_speed(struct ocrdma_dev *dev, u8 *lnk_speed)
 		goto mbx_err;
 
 	rsp = (struct ocrdma_get_link_speed_rsp *)cmd;
-	*lnk_speed = rsp->phys_port_speed;
+	*lnk_speed = (rsp->pflt_pps_ld_pnum & OCRDMA_PHY_PS_MASK)
+			>> OCRDMA_PHY_PS_SHIFT;
 
 mbx_err:
 	kfree(cmd);
@@ -1328,11 +1338,16 @@ static int ocrdma_mbx_get_phy_info(struct ocrdma_dev *dev)
 		goto mbx_err;
 
 	rsp = (struct ocrdma_get_phy_info_rsp *)cmd;
-	dev->phy.phy_type = le16_to_cpu(rsp->phy_type);
+	dev->phy.phy_type =
+			(rsp->ityp_ptyp & OCRDMA_PHY_TYPE_MASK);
+	dev->phy.interface_type =
+			(rsp->ityp_ptyp & OCRDMA_IF_TYPE_MASK)
+				>> OCRDMA_IF_TYPE_SHIFT;
 	dev->phy.auto_speeds_supported  =
-			le16_to_cpu(rsp->auto_speeds_supported);
+			(rsp->fspeed_aspeed & OCRDMA_ASPEED_SUPP_MASK);
 	dev->phy.fixed_speeds_supported =
-			le16_to_cpu(rsp->fixed_speeds_supported);
+			(rsp->fspeed_aspeed & OCRDMA_FSPEED_SUPP_MASK)
+				>> OCRDMA_FSPEED_SUPP_SHIFT;
 mbx_err:
 	kfree(cmd);
 	return status;
@@ -1457,8 +1472,8 @@ static int ocrdma_mbx_create_ah_tbl(struct ocrdma_dev *dev)
 
 	pbes = (struct ocrdma_pbe *)dev->av_tbl.pbl.va;
 	for (i = 0; i < dev->av_tbl.size / OCRDMA_MIN_Q_PAGE_SIZE; i++) {
-		pbes[i].pa_lo = (u32) (pa & 0xffffffff);
-		pbes[i].pa_hi = (u32) upper_32_bits(pa);
+		pbes[i].pa_lo = (u32)cpu_to_le32(pa & 0xffffffff);
+		pbes[i].pa_hi = (u32)cpu_to_le32(upper_32_bits(pa));
 		pa += PAGE_SIZE;
 	}
 	cmd->tbl_addr[0].lo = (u32)(dev->av_tbl.pbl.pa & 0xFFFFFFFF);
@@ -1501,6 +1516,7 @@ static void ocrdma_mbx_delete_ah_tbl(struct ocrdma_dev *dev)
 	ocrdma_mbx_cmd(dev, (struct ocrdma_mqe *)cmd);
 	dma_free_coherent(&pdev->dev, dev->av_tbl.size, dev->av_tbl.va,
 			  dev->av_tbl.pa);
+	dev->av_tbl.va = NULL;
 	dma_free_coherent(&pdev->dev, PAGE_SIZE, dev->av_tbl.pbl.va,
 			  dev->av_tbl.pbl.pa);
 	kfree(cmd);
@@ -1624,14 +1640,16 @@ int ocrdma_mbx_create_cq(struct ocrdma_dev *dev, struct ocrdma_cq *cq,
 			cmd->cmd.pgsz_pgcnt |= OCRDMA_CREATE_CQ_DPP <<
 				OCRDMA_CREATE_CQ_TYPE_SHIFT;
 		cq->phase_change = false;
-		cmd->cmd.cqe_count = (cq->len / cqe_size);
+		cmd->cmd.pdid_cqecnt = (cq->len / cqe_size);
 	} else {
-		cmd->cmd.cqe_count = (cq->len / cqe_size) - 1;
+		cmd->cmd.pdid_cqecnt = (cq->len / cqe_size) - 1;
 		cmd->cmd.ev_cnt_flags |= OCRDMA_CREATE_CQ_FLAGS_AUTO_VALID;
 		cq->phase_change = true;
 	}
 
-	cmd->cmd.pd_id = pd_id; /* valid only for v3 */
+	/* pd_id valid only for v3 */
+	cmd->cmd.pdid_cqecnt |= (pd_id <<
+		OCRDMA_CREATE_CQ_CMD_PDID_SHIFT);
 	ocrdma_build_q_pages(&cmd->cmd.pa[0], hw_pages, cq->pa, page_size);
 	status = ocrdma_mbx_cmd(dev, (struct ocrdma_mqe *)cmd);
 	if (status)
@@ -2206,7 +2224,8 @@ int ocrdma_mbx_create_qp(struct ocrdma_qp *qp, struct ib_qp_init_attr *attrs,
 				OCRDMA_CREATE_QP_REQ_RQ_CQID_MASK;
 	qp->rq_cq = cq;
 
-	if (pd->dpp_enabled && pd->num_dpp_qp) {
+	if (pd->dpp_enabled && attrs->cap.max_inline_data && pd->num_dpp_qp &&
+	    (attrs->cap.max_inline_data <= dev->attr.max_inline_data)) {
 		ocrdma_set_create_qp_dpp_cmd(cmd, pd, qp, enable_dpp_cq,
 					     dpp_cq_id);
 	}
@@ -2254,7 +2273,8 @@ mbx_err:
 
 static int ocrdma_set_av_params(struct ocrdma_qp *qp,
 				struct ocrdma_modify_qp *cmd,
-				struct ib_qp_attr *attrs)
+				struct ib_qp_attr *attrs,
+				int attr_mask)
 {
 	int status;
 	struct ib_ah_attr *ah_attr = &attrs->ah_attr;
@@ -2264,6 +2284,8 @@ static int ocrdma_set_av_params(struct ocrdma_qp *qp,
 
 	if ((ah_attr->ah_flags & IB_AH_GRH) == 0)
 		return -EINVAL;
+	if (atomic_cmpxchg(&qp->dev->update_sl, 1, 0))
+		ocrdma_init_service_level(qp->dev);
 	cmd->params.tclass_sq_psn |=
 	    (ah_attr->grh.traffic_class << OCRDMA_QP_PARAMS_TCLASS_SHIFT);
 	cmd->params.rnt_rc_sl_fl |=
@@ -2292,11 +2314,13 @@ static int ocrdma_set_av_params(struct ocrdma_qp *qp,
 	ocrdma_cpu_to_le32(&cmd->params.dgid[0], sizeof(cmd->params.dgid));
 	ocrdma_cpu_to_le32(&cmd->params.sgid[0], sizeof(cmd->params.sgid));
 	cmd->params.vlan_dmac_b4_to_b5 = mac_addr[4] | (mac_addr[5] << 8);
-	vlan_id = ah_attr->vlan_id;
-	if (vlan_id && (vlan_id < 0x1000)) {
+	if (attr_mask & IB_QP_VID) {
+		vlan_id = attrs->vlan_id;
 		cmd->params.vlan_dmac_b4_to_b5 |=
 		    vlan_id << OCRDMA_QP_PARAMS_VLAN_SHIFT;
 		cmd->flags |= OCRDMA_QP_PARA_VLAN_EN_VALID;
+		cmd->params.rnt_rc_sl_fl |=
+			(qp->dev->sl & 0x07) << OCRDMA_QP_PARAMS_SL_SHIFT;
 	}
 	return 0;
 }
@@ -2318,7 +2342,7 @@ static int ocrdma_set_qp_params(struct ocrdma_qp *qp,
 		cmd->flags |= OCRDMA_QP_PARA_QKEY_VALID;
 	}
 	if (attr_mask & IB_QP_AV) {
-		status = ocrdma_set_av_params(qp, cmd, attrs);
+		status = ocrdma_set_av_params(qp, cmd, attrs, attr_mask);
 		if (status)
 			return status;
 	} else if (qp->qp_type == IB_QPT_GSI || qp->qp_type == IB_QPT_UD) {
@@ -2604,6 +2628,168 @@ int ocrdma_mbx_destroy_srq(struct ocrdma_dev *dev, struct ocrdma_srq *srq)
 	return status;
 }
 
+static int ocrdma_mbx_get_dcbx_config(struct ocrdma_dev *dev, u32 ptype,
+				      struct ocrdma_dcbx_cfg *dcbxcfg)
+{
+	int status = 0;
+	dma_addr_t pa;
+	struct ocrdma_mqe cmd;
+
+	struct ocrdma_get_dcbx_cfg_req *req = NULL;
+	struct ocrdma_get_dcbx_cfg_rsp *rsp = NULL;
+	struct pci_dev *pdev = dev->nic_info.pdev;
+	struct ocrdma_mqe_sge *mqe_sge = cmd.u.nonemb_req.sge;
+
+	memset(&cmd, 0, sizeof(struct ocrdma_mqe));
+	cmd.hdr.pyld_len = max_t (u32, sizeof(struct ocrdma_get_dcbx_cfg_rsp),
+					sizeof(struct ocrdma_get_dcbx_cfg_req));
+	req = dma_alloc_coherent(&pdev->dev, cmd.hdr.pyld_len, &pa, GFP_KERNEL);
+	if (!req) {
+		status = -ENOMEM;
+		goto mem_err;
+	}
+
+	cmd.hdr.spcl_sge_cnt_emb |= (1 << OCRDMA_MQE_HDR_SGE_CNT_SHIFT) &
+					OCRDMA_MQE_HDR_SGE_CNT_MASK;
+	mqe_sge->pa_lo = (u32) (pa & 0xFFFFFFFFUL);
+	mqe_sge->pa_hi = (u32) upper_32_bits(pa);
+	mqe_sge->len = cmd.hdr.pyld_len;
+
+	memset(req, 0, sizeof(struct ocrdma_get_dcbx_cfg_req));
+	ocrdma_init_mch(&req->hdr, OCRDMA_CMD_GET_DCBX_CONFIG,
+			OCRDMA_SUBSYS_DCBX, cmd.hdr.pyld_len);
+	req->param_type = ptype;
+
+	status = ocrdma_mbx_cmd(dev, &cmd);
+	if (status)
+		goto mbx_err;
+
+	rsp = (struct ocrdma_get_dcbx_cfg_rsp *)req;
+	ocrdma_le32_to_cpu(rsp, sizeof(struct ocrdma_get_dcbx_cfg_rsp));
+	memcpy(dcbxcfg, &rsp->cfg, sizeof(struct ocrdma_dcbx_cfg));
+
+mbx_err:
+	dma_free_coherent(&pdev->dev, cmd.hdr.pyld_len, req, pa);
+mem_err:
+	return status;
+}
+
+#define OCRDMA_MAX_SERVICE_LEVEL_INDEX	0x08
+#define OCRDMA_DEFAULT_SERVICE_LEVEL	0x05
+
+static int ocrdma_parse_dcbxcfg_rsp(struct ocrdma_dev *dev, int ptype,
+				    struct ocrdma_dcbx_cfg *dcbxcfg,
+				    u8 *srvc_lvl)
+{
+	int status = -EINVAL, indx, slindx;
+	int ventry_cnt;
+	struct ocrdma_app_parameter *app_param;
+	u8 valid, proto_sel;
+	u8 app_prio, pfc_prio;
+	u16 proto;
+
+	if (!(dcbxcfg->tcv_aev_opv_st & OCRDMA_DCBX_STATE_MASK)) {
+		pr_info("%s ocrdma%d DCBX is disabled\n",
+			dev_name(&dev->nic_info.pdev->dev), dev->id);
+		goto out;
+	}
+
+	if (!ocrdma_is_enabled_and_synced(dcbxcfg->pfc_state)) {
+		pr_info("%s ocrdma%d priority flow control(%s) is %s%s\n",
+			dev_name(&dev->nic_info.pdev->dev), dev->id,
+			(ptype > 0 ? "operational" : "admin"),
+			(dcbxcfg->pfc_state & OCRDMA_STATE_FLAG_ENABLED) ?
+			"enabled" : "disabled",
+			(dcbxcfg->pfc_state & OCRDMA_STATE_FLAG_SYNC) ?
+			"" : ", not sync'ed");
+		goto out;
+	} else {
+		pr_info("%s ocrdma%d priority flow control is enabled and sync'ed\n",
+			dev_name(&dev->nic_info.pdev->dev), dev->id);
+	}
+
+	ventry_cnt = (dcbxcfg->tcv_aev_opv_st >>
+				OCRDMA_DCBX_APP_ENTRY_SHIFT)
+				& OCRDMA_DCBX_STATE_MASK;
+
+	for (indx = 0; indx < ventry_cnt; indx++) {
+		app_param = &dcbxcfg->app_param[indx];
+		valid = (app_param->valid_proto_app >>
+				OCRDMA_APP_PARAM_VALID_SHIFT)
+				& OCRDMA_APP_PARAM_VALID_MASK;
+		proto_sel = (app_param->valid_proto_app
+				>>  OCRDMA_APP_PARAM_PROTO_SEL_SHIFT)
+				& OCRDMA_APP_PARAM_PROTO_SEL_MASK;
+		proto = app_param->valid_proto_app &
+				OCRDMA_APP_PARAM_APP_PROTO_MASK;
+
+		if (
+			valid && proto == OCRDMA_APP_PROTO_ROCE &&
+			proto_sel == OCRDMA_PROTO_SELECT_L2) {
+			for (slindx = 0; slindx <
+				OCRDMA_MAX_SERVICE_LEVEL_INDEX; slindx++) {
+				app_prio = ocrdma_get_app_prio(
+						(u8 *)app_param->app_prio,
+						slindx);
+				pfc_prio = ocrdma_get_pfc_prio(
+						(u8 *)dcbxcfg->pfc_prio,
+						slindx);
+
+				if (app_prio && pfc_prio) {
+					*srvc_lvl = slindx;
+					status = 0;
+					goto out;
+				}
+			}
+			if (slindx == OCRDMA_MAX_SERVICE_LEVEL_INDEX) {
+				pr_info("%s ocrdma%d application priority not set for 0x%x protocol\n",
+					dev_name(&dev->nic_info.pdev->dev),
+					dev->id, proto);
+			}
+		}
+	}
+
+out:
+	return status;
+}
+
+void ocrdma_init_service_level(struct ocrdma_dev *dev)
+{
+	int status = 0, indx;
+	struct ocrdma_dcbx_cfg dcbxcfg;
+	u8 srvc_lvl = OCRDMA_DEFAULT_SERVICE_LEVEL;
+	int ptype = OCRDMA_PARAMETER_TYPE_OPER;
+
+	for (indx = 0; indx < 2; indx++) {
+		status = ocrdma_mbx_get_dcbx_config(dev, ptype, &dcbxcfg);
+		if (status) {
+			pr_err("%s(): status=%d\n", __func__, status);
+			ptype = OCRDMA_PARAMETER_TYPE_ADMIN;
+			continue;
+		}
+
+		status = ocrdma_parse_dcbxcfg_rsp(dev, ptype,
+						  &dcbxcfg, &srvc_lvl);
+		if (status) {
+			ptype = OCRDMA_PARAMETER_TYPE_ADMIN;
+			continue;
+		}
+
+		break;
+	}
+
+	if (status)
+		pr_info("%s ocrdma%d service level default\n",
+			dev_name(&dev->nic_info.pdev->dev), dev->id);
+	else
+		pr_info("%s ocrdma%d service level %d\n",
+			dev_name(&dev->nic_info.pdev->dev), dev->id,
+			srvc_lvl);
+
+	dev->pfc_state = ocrdma_is_enabled_and_synced(dcbxcfg.pfc_state);
+	dev->sl = srvc_lvl;
+}
+
 int ocrdma_alloc_av(struct ocrdma_dev *dev, struct ocrdma_ah *ah)
 {
 	int i;
@@ -2709,13 +2895,15 @@ int ocrdma_init_hw(struct ocrdma_dev *dev)
 		goto conf_err;
 	status = ocrdma_mbx_get_phy_info(dev);
 	if (status)
-		goto conf_err;
+		goto info_attrb_err;
 	status = ocrdma_mbx_get_ctrl_attribs(dev);
 	if (status)
-		goto conf_err;
+		goto info_attrb_err;
 
 	return 0;
 
+info_attrb_err:
+	ocrdma_mbx_delete_ah_tbl(dev);
 conf_err:
 	ocrdma_destroy_mq(dev);
 mq_err:
diff --git a/drivers/infiniband/hw/ocrdma/ocrdma_hw.h b/drivers/infiniband/hw/ocrdma/ocrdma_hw.h
index e513f7293142..6eed8f191322 100644
--- a/drivers/infiniband/hw/ocrdma/ocrdma_hw.h
+++ b/drivers/infiniband/hw/ocrdma/ocrdma_hw.h
@@ -135,4 +135,6 @@ int ocrdma_get_irq(struct ocrdma_dev *dev, struct ocrdma_eq *eq);
 
 int ocrdma_mbx_rdma_stats(struct ocrdma_dev *, bool reset);
 char *port_speed_string(struct ocrdma_dev *dev);
+void ocrdma_init_service_level(struct ocrdma_dev *);
+
 #endif				/* __OCRDMA_HW_H__ */
diff --git a/drivers/infiniband/hw/ocrdma/ocrdma_main.c b/drivers/infiniband/hw/ocrdma/ocrdma_main.c
index 7c504e079744..b0b2257b8e04 100644
--- a/drivers/infiniband/hw/ocrdma/ocrdma_main.c
+++ b/drivers/infiniband/hw/ocrdma/ocrdma_main.c
@@ -324,6 +324,11 @@ static int ocrdma_alloc_resources(struct ocrdma_dev *dev)
 		if (!dev->qp_tbl)
 			goto alloc_err;
 	}
+
+	dev->stag_arr = kzalloc(sizeof(u64) * OCRDMA_MAX_STAG, GFP_KERNEL);
+	if (dev->stag_arr == NULL)
+		goto alloc_err;
+
 	spin_lock_init(&dev->av_tbl.lock);
 	spin_lock_init(&dev->flush_q_lock);
 	return 0;
@@ -334,6 +339,7 @@ alloc_err:
 
 static void ocrdma_free_resources(struct ocrdma_dev *dev)
 {
+	kfree(dev->stag_arr);
 	kfree(dev->qp_tbl);
 	kfree(dev->cq_tbl);
 	kfree(dev->sgid_tbl);
@@ -353,15 +359,25 @@ static ssize_t show_fw_ver(struct device *device, struct device_attribute *attr,
 {
 	struct ocrdma_dev *dev = dev_get_drvdata(device);
 
-	return scnprintf(buf, PAGE_SIZE, "%s", &dev->attr.fw_ver[0]);
+	return scnprintf(buf, PAGE_SIZE, "%s\n", &dev->attr.fw_ver[0]);
+}
+
+static ssize_t show_hca_type(struct device *device,
+			     struct device_attribute *attr, char *buf)
+{
+	struct ocrdma_dev *dev = dev_get_drvdata(device);
+
+	return scnprintf(buf, PAGE_SIZE, "%s\n", &dev->model_number[0]);
 }
 
 static DEVICE_ATTR(hw_rev, S_IRUGO, show_rev, NULL);
 static DEVICE_ATTR(fw_ver, S_IRUGO, show_fw_ver, NULL);
+static DEVICE_ATTR(hca_type, S_IRUGO, show_hca_type, NULL);
 
 static struct device_attribute *ocrdma_attributes[] = {
 	&dev_attr_hw_rev,
-	&dev_attr_fw_ver
+	&dev_attr_fw_ver,
+	&dev_attr_hca_type
 };
 
 static void ocrdma_remove_sysfiles(struct ocrdma_dev *dev)
@@ -372,6 +388,68 @@ static void ocrdma_remove_sysfiles(struct ocrdma_dev *dev)
 		device_remove_file(&dev->ibdev.dev, ocrdma_attributes[i]);
 }
 
+static void ocrdma_add_default_sgid(struct ocrdma_dev *dev)
+{
+	/* GID Index 0 - Invariant manufacturer-assigned EUI-64 */
+	union ib_gid *sgid = &dev->sgid_tbl[0];
+
+	sgid->global.subnet_prefix = cpu_to_be64(0xfe80000000000000LL);
+	ocrdma_get_guid(dev, &sgid->raw[8]);
+}
+
+static void ocrdma_init_ipv4_gids(struct ocrdma_dev *dev,
+				  struct net_device *net)
+{
+	struct in_device *in_dev;
+	union ib_gid gid;
+	in_dev = in_dev_get(net);
+	if (in_dev) {
+		for_ifa(in_dev) {
+			ipv6_addr_set_v4mapped(ifa->ifa_address,
+					       (struct in6_addr *)&gid);
+			ocrdma_add_sgid(dev, &gid);
+		}
+		endfor_ifa(in_dev);
+		in_dev_put(in_dev);
+	}
+}
+
+static void ocrdma_init_ipv6_gids(struct ocrdma_dev *dev,
+				  struct net_device *net)
+{
+#if IS_ENABLED(CONFIG_IPV6)
+	struct inet6_dev *in6_dev;
+	union ib_gid  *pgid;
+	struct inet6_ifaddr *ifp;
+	in6_dev = in6_dev_get(net);
+	if (in6_dev) {
+		read_lock_bh(&in6_dev->lock);
+		list_for_each_entry(ifp, &in6_dev->addr_list, if_list) {
+			pgid = (union ib_gid *)&ifp->addr;
+			ocrdma_add_sgid(dev, pgid);
+		}
+		read_unlock_bh(&in6_dev->lock);
+		in6_dev_put(in6_dev);
+	}
+#endif
+}
+
+static void ocrdma_init_gid_table(struct ocrdma_dev *dev)
+{
+	struct  net_device *net_dev;
+
+	for_each_netdev(&init_net, net_dev) {
+		struct net_device *real_dev = rdma_vlan_dev_real_dev(net_dev) ?
+				rdma_vlan_dev_real_dev(net_dev) : net_dev;
+
+		if (real_dev == dev->nic_info.netdev) {
+			ocrdma_add_default_sgid(dev);
+			ocrdma_init_ipv4_gids(dev, net_dev);
+			ocrdma_init_ipv6_gids(dev, net_dev);
+		}
+	}
+}
+
 static struct ocrdma_dev *ocrdma_add(struct be_dev_info *dev_info)
 {
 	int status = 0, i;
@@ -399,6 +477,8 @@ static struct ocrdma_dev *ocrdma_add(struct be_dev_info *dev_info)
 	if (status)
 		goto alloc_err;
 
+	ocrdma_init_service_level(dev);
+	ocrdma_init_gid_table(dev);
 	status = ocrdma_register_device(dev);
 	if (status)
 		goto alloc_err;
@@ -508,6 +588,12 @@ static int ocrdma_close(struct ocrdma_dev *dev)
 	return 0;
 }
 
+static void ocrdma_shutdown(struct ocrdma_dev *dev)
+{
+	ocrdma_close(dev);
+	ocrdma_remove(dev);
+}
+
 /* event handling via NIC driver ensures that all the NIC specific
  * initialization done before RoCE driver notifies
  * event to stack.
@@ -521,6 +607,9 @@ static void ocrdma_event_handler(struct ocrdma_dev *dev, u32 event)
 	case BE_DEV_DOWN:
 		ocrdma_close(dev);
 		break;
+	case BE_DEV_SHUTDOWN:
+		ocrdma_shutdown(dev);
+		break;
 	}
 }
 
@@ -567,8 +656,10 @@ static int __init ocrdma_init_module(void)
 	return 0;
 
 err_be_reg:
+#if IS_ENABLED(CONFIG_IPV6)
 	ocrdma_unregister_inet6addr_notifier();
 err_notifier6:
+#endif
 	ocrdma_unregister_inetaddr_notifier();
 	return status;
 }
diff --git a/drivers/infiniband/hw/ocrdma/ocrdma_sli.h b/drivers/infiniband/hw/ocrdma/ocrdma_sli.h
index 96c9ee602ba4..4e036480c1a8 100644
--- a/drivers/infiniband/hw/ocrdma/ocrdma_sli.h
+++ b/drivers/infiniband/hw/ocrdma/ocrdma_sli.h
@@ -28,8 +28,6 @@
 #ifndef __OCRDMA_SLI_H__
 #define __OCRDMA_SLI_H__
 
-#define Bit(_b) (1 << (_b))
-
 enum {
 	OCRDMA_ASIC_GEN_SKH_R = 0x04,
 	OCRDMA_ASIC_GEN_LANCER = 0x0B
@@ -44,35 +42,39 @@ enum {
 #define OCRDMA_SUBSYS_ROCE 10
 enum {
 	OCRDMA_CMD_QUERY_CONFIG = 1,
-	OCRDMA_CMD_ALLOC_PD,
-	OCRDMA_CMD_DEALLOC_PD,
-
-	OCRDMA_CMD_CREATE_AH_TBL,
-	OCRDMA_CMD_DELETE_AH_TBL,
-
-	OCRDMA_CMD_CREATE_QP,
-	OCRDMA_CMD_QUERY_QP,
-	OCRDMA_CMD_MODIFY_QP,
-	OCRDMA_CMD_DELETE_QP,
-
-	OCRDMA_CMD_RSVD1,
-	OCRDMA_CMD_ALLOC_LKEY,
-	OCRDMA_CMD_DEALLOC_LKEY,
-	OCRDMA_CMD_REGISTER_NSMR,
-	OCRDMA_CMD_REREGISTER_NSMR,
-	OCRDMA_CMD_REGISTER_NSMR_CONT,
-	OCRDMA_CMD_QUERY_NSMR,
-	OCRDMA_CMD_ALLOC_MW,
-	OCRDMA_CMD_QUERY_MW,
-
-	OCRDMA_CMD_CREATE_SRQ,
-	OCRDMA_CMD_QUERY_SRQ,
-	OCRDMA_CMD_MODIFY_SRQ,
-	OCRDMA_CMD_DELETE_SRQ,
-
-	OCRDMA_CMD_ATTACH_MCAST,
-	OCRDMA_CMD_DETACH_MCAST,
-	OCRDMA_CMD_GET_RDMA_STATS,
+	OCRDMA_CMD_ALLOC_PD = 2,
+	OCRDMA_CMD_DEALLOC_PD = 3,
+
+	OCRDMA_CMD_CREATE_AH_TBL = 4,
+	OCRDMA_CMD_DELETE_AH_TBL = 5,
+
+	OCRDMA_CMD_CREATE_QP = 6,
+	OCRDMA_CMD_QUERY_QP = 7,
+	OCRDMA_CMD_MODIFY_QP = 8 ,
+	OCRDMA_CMD_DELETE_QP = 9,
+
+	OCRDMA_CMD_RSVD1 = 10,
+	OCRDMA_CMD_ALLOC_LKEY = 11,
+	OCRDMA_CMD_DEALLOC_LKEY = 12,
+	OCRDMA_CMD_REGISTER_NSMR = 13,
+	OCRDMA_CMD_REREGISTER_NSMR = 14,
+	OCRDMA_CMD_REGISTER_NSMR_CONT = 15,
+	OCRDMA_CMD_QUERY_NSMR = 16,
+	OCRDMA_CMD_ALLOC_MW = 17,
+	OCRDMA_CMD_QUERY_MW = 18,
+
+	OCRDMA_CMD_CREATE_SRQ = 19,
+	OCRDMA_CMD_QUERY_SRQ = 20,
+	OCRDMA_CMD_MODIFY_SRQ = 21,
+	OCRDMA_CMD_DELETE_SRQ = 22,
+
+	OCRDMA_CMD_ATTACH_MCAST = 23,
+	OCRDMA_CMD_DETACH_MCAST = 24,
+
+	OCRDMA_CMD_CREATE_RBQ = 25,
+	OCRDMA_CMD_DESTROY_RBQ = 26,
+
+	OCRDMA_CMD_GET_RDMA_STATS = 27,
 
 	OCRDMA_CMD_MAX
 };
@@ -99,11 +101,11 @@ enum {
 	QTYPE_MCCQ	= 3
 };
 
-#define OCRDMA_MAX_SGID (8)
+#define OCRDMA_MAX_SGID		8
 
 #define OCRDMA_MAX_QP    2048
 #define OCRDMA_MAX_CQ    2048
-#define OCRDMA_MAX_STAG  8192
+#define OCRDMA_MAX_STAG 16384
 
 enum {
 	OCRDMA_DB_RQ_OFFSET		= 0xE0,
@@ -124,33 +126,33 @@ enum {
 #define OCRDMA_DB_CQ_RING_ID_EXT_MASK  0x0C00	/* bits 10-11 of qid at 12-11 */
 /* qid #2 msbits at 12-11 */
 #define OCRDMA_DB_CQ_RING_ID_EXT_MASK_SHIFT  0x1
-#define OCRDMA_DB_CQ_NUM_POPPED_SHIFT       (16)	/* bits 16 - 28 */
+#define OCRDMA_DB_CQ_NUM_POPPED_SHIFT	16	/* bits 16 - 28 */
 /* Rearm bit */
-#define OCRDMA_DB_CQ_REARM_SHIFT        (29)	/* bit 29 */
+#define OCRDMA_DB_CQ_REARM_SHIFT	29	/* bit 29 */
 /* solicited bit */
-#define OCRDMA_DB_CQ_SOLICIT_SHIFT   (31)	/* bit 31 */
+#define OCRDMA_DB_CQ_SOLICIT_SHIFT	31	/* bit 31 */
 
 #define OCRDMA_EQ_ID_MASK		0x1FF	/* bits 0 - 8 */
 #define OCRDMA_EQ_ID_EXT_MASK		0x3e00	/* bits 9-13 */
-#define OCRDMA_EQ_ID_EXT_MASK_SHIFT	(2)	/* qid bits 9-13 at 11-15 */
+#define OCRDMA_EQ_ID_EXT_MASK_SHIFT	2	/* qid bits 9-13 at 11-15 */
 
 /* Clear the interrupt for this eq */
-#define OCRDMA_EQ_CLR_SHIFT			(9)	/* bit 9 */
+#define OCRDMA_EQ_CLR_SHIFT		9	/* bit 9 */
 /* Must be 1 */
-#define OCRDMA_EQ_TYPE_SHIFT		(10)	/* bit 10 */
+#define OCRDMA_EQ_TYPE_SHIFT		10	/* bit 10 */
 /* Number of event entries processed */
-#define OCRDMA_NUM_EQE_SHIFT		(16)	/* bits 16 - 28 */
+#define OCRDMA_NUM_EQE_SHIFT		16	/* bits 16 - 28 */
 /* Rearm bit */
-#define OCRDMA_REARM_SHIFT		(29)	/* bit 29 */
+#define OCRDMA_REARM_SHIFT		29	/* bit 29 */
 
 #define OCRDMA_MQ_ID_MASK		0x7FF	/* bits 0 - 10 */
 /* Number of entries posted */
-#define OCRDMA_MQ_NUM_MQE_SHIFT	(16)	/* bits 16 - 29 */
+#define OCRDMA_MQ_NUM_MQE_SHIFT	16	/* bits 16 - 29 */
 
-#define OCRDMA_MIN_HPAGE_SIZE (4096)
+#define OCRDMA_MIN_HPAGE_SIZE	4096
 
-#define OCRDMA_MIN_Q_PAGE_SIZE (4096)
-#define OCRDMA_MAX_Q_PAGES     (8)
+#define OCRDMA_MIN_Q_PAGE_SIZE	4096
+#define OCRDMA_MAX_Q_PAGES	8
 
 #define OCRDMA_SLI_ASIC_ID_OFFSET	0x9C
 #define OCRDMA_SLI_ASIC_REV_MASK	0x000000FF
@@ -166,14 +168,14 @@ enum {
 # 6: 256K Bytes
 # 7: 512K Bytes
 */
-#define OCRDMA_MAX_Q_PAGE_SIZE_CNT (8)
+#define OCRDMA_MAX_Q_PAGE_SIZE_CNT	8
 #define OCRDMA_Q_PAGE_BASE_SIZE (OCRDMA_MIN_Q_PAGE_SIZE * OCRDMA_MAX_Q_PAGES)
 
-#define MAX_OCRDMA_QP_PAGES      (8)
+#define MAX_OCRDMA_QP_PAGES		8
 #define OCRDMA_MAX_WQE_MEM_SIZE (MAX_OCRDMA_QP_PAGES * OCRDMA_MIN_HQ_PAGE_SIZE)
 
-#define OCRDMA_CREATE_CQ_MAX_PAGES (4)
-#define OCRDMA_DPP_CQE_SIZE (4)
+#define OCRDMA_CREATE_CQ_MAX_PAGES	4
+#define OCRDMA_DPP_CQE_SIZE		4
 
 #define OCRDMA_GEN2_MAX_CQE 1024
 #define OCRDMA_GEN2_CQ_PAGE_SIZE 4096
@@ -234,7 +236,7 @@ struct ocrdma_mqe_sge {
 
 enum {
 	OCRDMA_MQE_HDR_EMB_SHIFT	= 0,
-	OCRDMA_MQE_HDR_EMB_MASK		= Bit(0),
+	OCRDMA_MQE_HDR_EMB_MASK		= BIT(0),
 	OCRDMA_MQE_HDR_SGE_CNT_SHIFT	= 3,
 	OCRDMA_MQE_HDR_SGE_CNT_MASK	= 0x1F << OCRDMA_MQE_HDR_SGE_CNT_SHIFT,
 	OCRDMA_MQE_HDR_SPECIAL_SHIFT	= 24,
@@ -288,7 +290,7 @@ struct ocrdma_pa {
 	u32 hi;
 };
 
-#define MAX_OCRDMA_EQ_PAGES (8)
+#define MAX_OCRDMA_EQ_PAGES	8
 struct ocrdma_create_eq_req {
 	struct ocrdma_mbx_hdr req;
 	u32 num_pages;
@@ -300,7 +302,7 @@ struct ocrdma_create_eq_req {
 };
 
 enum {
-	OCRDMA_CREATE_EQ_VALID	= Bit(29),
+	OCRDMA_CREATE_EQ_VALID	= BIT(29),
 	OCRDMA_CREATE_EQ_CNT_SHIFT	= 26,
 	OCRDMA_CREATE_CQ_DELAY_SHIFT	= 13,
 };
@@ -310,7 +312,7 @@ struct ocrdma_create_eq_rsp {
 	u32 vector_eqid;
 };
 
-#define OCRDMA_EQ_MINOR_OTHER (0x1)
+#define OCRDMA_EQ_MINOR_OTHER	0x1
 
 enum {
 	OCRDMA_MCQE_STATUS_SHIFT	= 0,
@@ -318,13 +320,13 @@ enum {
 	OCRDMA_MCQE_ESTATUS_SHIFT	= 16,
 	OCRDMA_MCQE_ESTATUS_MASK	= 0xFFFF << OCRDMA_MCQE_ESTATUS_SHIFT,
 	OCRDMA_MCQE_CONS_SHIFT		= 27,
-	OCRDMA_MCQE_CONS_MASK		= Bit(27),
+	OCRDMA_MCQE_CONS_MASK		= BIT(27),
 	OCRDMA_MCQE_CMPL_SHIFT		= 28,
-	OCRDMA_MCQE_CMPL_MASK		= Bit(28),
+	OCRDMA_MCQE_CMPL_MASK		= BIT(28),
 	OCRDMA_MCQE_AE_SHIFT		= 30,
-	OCRDMA_MCQE_AE_MASK		= Bit(30),
+	OCRDMA_MCQE_AE_MASK		= BIT(30),
 	OCRDMA_MCQE_VALID_SHIFT		= 31,
-	OCRDMA_MCQE_VALID_MASK		= Bit(31)
+	OCRDMA_MCQE_VALID_MASK		= BIT(31)
 };
 
 struct ocrdma_mcqe {
@@ -335,13 +337,13 @@ struct ocrdma_mcqe {
 };
 
 enum {
-	OCRDMA_AE_MCQE_QPVALID		= Bit(31),
+	OCRDMA_AE_MCQE_QPVALID		= BIT(31),
 	OCRDMA_AE_MCQE_QPID_MASK	= 0xFFFF,
 
-	OCRDMA_AE_MCQE_CQVALID		= Bit(31),
+	OCRDMA_AE_MCQE_CQVALID		= BIT(31),
 	OCRDMA_AE_MCQE_CQID_MASK	= 0xFFFF,
-	OCRDMA_AE_MCQE_VALID		= Bit(31),
-	OCRDMA_AE_MCQE_AE		= Bit(30),
+	OCRDMA_AE_MCQE_VALID		= BIT(31),
+	OCRDMA_AE_MCQE_AE		= BIT(30),
 	OCRDMA_AE_MCQE_EVENT_TYPE_SHIFT	= 16,
 	OCRDMA_AE_MCQE_EVENT_TYPE_MASK	=
 					0xFF << OCRDMA_AE_MCQE_EVENT_TYPE_SHIFT,
@@ -382,9 +384,9 @@ enum {
 	OCRDMA_AE_MPA_MCQE_EVENT_TYPE_MASK	= 0xFF <<
 					OCRDMA_AE_MPA_MCQE_EVENT_TYPE_SHIFT,
 	OCRDMA_AE_MPA_MCQE_EVENT_AE_SHIFT	= 30,
-	OCRDMA_AE_MPA_MCQE_EVENT_AE_MASK	= Bit(30),
+	OCRDMA_AE_MPA_MCQE_EVENT_AE_MASK	= BIT(30),
 	OCRDMA_AE_MPA_MCQE_EVENT_VALID_SHIFT	= 31,
-	OCRDMA_AE_MPA_MCQE_EVENT_VALID_MASK	= Bit(31)
+	OCRDMA_AE_MPA_MCQE_EVENT_VALID_MASK	= BIT(31)
 };
 
 struct ocrdma_ae_mpa_mcqe {
@@ -408,9 +410,9 @@ enum {
 	OCRDMA_AE_QP_MCQE_EVENT_TYPE_MASK	= 0xFF <<
 				OCRDMA_AE_QP_MCQE_EVENT_TYPE_SHIFT,
 	OCRDMA_AE_QP_MCQE_EVENT_AE_SHIFT	= 30,
-	OCRDMA_AE_QP_MCQE_EVENT_AE_MASK		= Bit(30),
+	OCRDMA_AE_QP_MCQE_EVENT_AE_MASK		= BIT(30),
 	OCRDMA_AE_QP_MCQE_EVENT_VALID_SHIFT	= 31,
-	OCRDMA_AE_QP_MCQE_EVENT_VALID_MASK	= Bit(31)
+	OCRDMA_AE_QP_MCQE_EVENT_VALID_MASK	= BIT(31)
 };
 
 struct ocrdma_ae_qp_mcqe {
@@ -422,7 +424,12 @@ struct ocrdma_ae_qp_mcqe {
 
 #define OCRDMA_ASYNC_RDMA_EVE_CODE 0x14
 #define OCRDMA_ASYNC_GRP5_EVE_CODE 0x5
-#define OCRDMA_ASYNC_EVENT_PVID_STATE 0x3
+
+enum ocrdma_async_grp5_events {
+	OCRDMA_ASYNC_EVENT_QOS_VALUE	= 0x01,
+	OCRDMA_ASYNC_EVENT_COS_VALUE	= 0x02,
+	OCRDMA_ASYNC_EVENT_PVID_STATE	= 0x03
+};
 
 enum OCRDMA_ASYNC_EVENT_TYPE {
 	OCRDMA_CQ_ERROR			= 0x00,
@@ -440,9 +447,9 @@ enum OCRDMA_ASYNC_EVENT_TYPE {
 /* mailbox command request and responses */
 enum {
 	OCRDMA_MBX_QUERY_CFG_CQ_OVERFLOW_SHIFT		= 2,
-	OCRDMA_MBX_QUERY_CFG_CQ_OVERFLOW_MASK		= Bit(2),
+	OCRDMA_MBX_QUERY_CFG_CQ_OVERFLOW_MASK		= BIT(2),
 	OCRDMA_MBX_QUERY_CFG_SRQ_SUPPORTED_SHIFT	= 3,
-	OCRDMA_MBX_QUERY_CFG_SRQ_SUPPORTED_MASK		= Bit(3),
+	OCRDMA_MBX_QUERY_CFG_SRQ_SUPPORTED_MASK		= BIT(3),
 	OCRDMA_MBX_QUERY_CFG_MAX_QP_SHIFT		= 8,
 	OCRDMA_MBX_QUERY_CFG_MAX_QP_MASK		= 0xFFFFFF <<
 				OCRDMA_MBX_QUERY_CFG_MAX_QP_SHIFT,
@@ -525,8 +532,8 @@ struct ocrdma_mbx_query_config {
 	u32 max_ird_ord_per_qp;
 	u32 max_shared_ird_ord;
 	u32 max_mr;
-	u32 max_mr_size_lo;
 	u32 max_mr_size_hi;
+	u32 max_mr_size_lo;
 	u32 max_num_mr_pbl;
 	u32 max_mw;
 	u32 max_fmr;
@@ -580,17 +587,26 @@ enum {
 	OCRDMA_FN_MODE_RDMA	= 0x4
 };
 
+enum {
+	OCRDMA_IF_TYPE_MASK		= 0xFFFF0000,
+	OCRDMA_IF_TYPE_SHIFT		= 0x10,
+	OCRDMA_PHY_TYPE_MASK		= 0x0000FFFF,
+	OCRDMA_FUTURE_DETAILS_MASK	= 0xFFFF0000,
+	OCRDMA_FUTURE_DETAILS_SHIFT	= 0x10,
+	OCRDMA_EX_PHY_DETAILS_MASK	= 0x0000FFFF,
+	OCRDMA_FSPEED_SUPP_MASK		= 0xFFFF0000,
+	OCRDMA_FSPEED_SUPP_SHIFT	= 0x10,
+	OCRDMA_ASPEED_SUPP_MASK		= 0x0000FFFF
+};
+
 struct ocrdma_get_phy_info_rsp {
 	struct ocrdma_mqe_hdr hdr;
 	struct ocrdma_mbx_rsp rsp;
 
-	u16 phy_type;
-	u16 interface_type;
+	u32 ityp_ptyp;
 	u32 misc_params;
-	u16 ext_phy_details;
-	u16 rsvd;
-	u16 auto_speeds_supported;
-	u16 fixed_speeds_supported;
+	u32 ftrdtl_exphydtl;
+	u32 fspeed_aspeed;
 	u32 future_use[2];
 };
 
@@ -603,19 +619,34 @@ enum {
 	OCRDMA_PHY_SPEED_40GBPS = 0x20
 };
 
+enum {
+	OCRDMA_PORT_NUM_MASK	= 0x3F,
+	OCRDMA_PT_MASK		= 0xC0,
+	OCRDMA_PT_SHIFT		= 0x6,
+	OCRDMA_LINK_DUP_MASK	= 0x0000FF00,
+	OCRDMA_LINK_DUP_SHIFT	= 0x8,
+	OCRDMA_PHY_PS_MASK	= 0x00FF0000,
+	OCRDMA_PHY_PS_SHIFT	= 0x10,
+	OCRDMA_PHY_PFLT_MASK	= 0xFF000000,
+	OCRDMA_PHY_PFLT_SHIFT	= 0x18,
+	OCRDMA_QOS_LNKSP_MASK	= 0xFFFF0000,
+	OCRDMA_QOS_LNKSP_SHIFT	= 0x10,
+	OCRDMA_LLST_MASK	= 0xFF,
+	OCRDMA_PLFC_MASK	= 0x00000400,
+	OCRDMA_PLFC_SHIFT	= 0x8,
+	OCRDMA_PLRFC_MASK	= 0x00000200,
+	OCRDMA_PLRFC_SHIFT	= 0x8,
+	OCRDMA_PLTFC_MASK	= 0x00000100,
+	OCRDMA_PLTFC_SHIFT	= 0x8
+};
 
 struct ocrdma_get_link_speed_rsp {
 	struct ocrdma_mqe_hdr hdr;
 	struct ocrdma_mbx_rsp rsp;
 
-	u8 pt_port_num;
-	u8 link_duplex;
-	u8 phys_port_speed;
-	u8 phys_port_fault;
-	u16 rsvd1;
-	u16 qos_lnk_speed;
-	u8 logical_lnk_status;
-	u8 rsvd2[3];
+	u32 pflt_pps_ld_pnum;
+	u32 qos_lsp;
+	u32 res_lls;
 };
 
 enum {
@@ -639,9 +670,9 @@ enum {
 	OCRDMA_CREATE_CQ_PAGE_SIZE_MASK		= 0xFF,
 
 	OCRDMA_CREATE_CQ_COALESCWM_SHIFT	= 12,
-	OCRDMA_CREATE_CQ_COALESCWM_MASK		= Bit(13) | Bit(12),
-	OCRDMA_CREATE_CQ_FLAGS_NODELAY		= Bit(14),
-	OCRDMA_CREATE_CQ_FLAGS_AUTO_VALID	= Bit(15),
+	OCRDMA_CREATE_CQ_COALESCWM_MASK		= BIT(13) | BIT(12),
+	OCRDMA_CREATE_CQ_FLAGS_NODELAY		= BIT(14),
+	OCRDMA_CREATE_CQ_FLAGS_AUTO_VALID	= BIT(15),
 
 	OCRDMA_CREATE_CQ_EQ_ID_MASK		= 0xFFFF,
 	OCRDMA_CREATE_CQ_CQE_COUNT_MASK		= 0xFFFF
@@ -654,8 +685,8 @@ enum {
 	OCRDMA_CREATE_CQ_EQID_SHIFT		= 22,
 
 	OCRDMA_CREATE_CQ_CNT_SHIFT		= 27,
-	OCRDMA_CREATE_CQ_FLAGS_VALID		= Bit(29),
-	OCRDMA_CREATE_CQ_FLAGS_EVENTABLE	= Bit(31),
+	OCRDMA_CREATE_CQ_FLAGS_VALID		= BIT(29),
+	OCRDMA_CREATE_CQ_FLAGS_EVENTABLE	= BIT(31),
 	OCRDMA_CREATE_CQ_DEF_FLAGS		= OCRDMA_CREATE_CQ_FLAGS_VALID |
 					OCRDMA_CREATE_CQ_FLAGS_EVENTABLE |
 					OCRDMA_CREATE_CQ_FLAGS_NODELAY
@@ -666,8 +697,7 @@ struct ocrdma_create_cq_cmd {
 	u32 pgsz_pgcnt;
 	u32 ev_cnt_flags;
 	u32 eqn;
-	u16 cqe_count;
-	u16 pd_id;
+	u32 pdid_cqecnt;
 	u32 rsvd6;
 	struct ocrdma_pa pa[OCRDMA_CREATE_CQ_MAX_PAGES];
 };
@@ -678,6 +708,10 @@ struct ocrdma_create_cq {
 };
 
 enum {
+	OCRDMA_CREATE_CQ_CMD_PDID_SHIFT	= 0x10
+};
+
+enum {
 	OCRDMA_CREATE_CQ_RSP_CQ_ID_MASK	= 0xFFFF
 };
 
@@ -695,8 +729,8 @@ enum {
 	OCRDMA_CREATE_MQ_V0_CQ_ID_SHIFT		= 22,
 	OCRDMA_CREATE_MQ_CQ_ID_SHIFT		= 16,
 	OCRDMA_CREATE_MQ_RING_SIZE_SHIFT	= 16,
-	OCRDMA_CREATE_MQ_VALID			= Bit(31),
-	OCRDMA_CREATE_MQ_ASYNC_CQ_VALID		= Bit(0)
+	OCRDMA_CREATE_MQ_VALID			= BIT(31),
+	OCRDMA_CREATE_MQ_ASYNC_CQ_VALID		= BIT(0)
 };
 
 struct ocrdma_create_mq_req {
@@ -747,7 +781,7 @@ enum {
 	OCRDMA_CREATE_QP_REQ_SQ_PAGE_SIZE_SHIFT	= 16,
 	OCRDMA_CREATE_QP_REQ_RQ_PAGE_SIZE_SHIFT	= 19,
 	OCRDMA_CREATE_QP_REQ_QPT_SHIFT		= 29,
-	OCRDMA_CREATE_QP_REQ_QPT_MASK		= Bit(31) | Bit(30) | Bit(29),
+	OCRDMA_CREATE_QP_REQ_QPT_MASK		= BIT(31) | BIT(30) | BIT(29),
 
 	OCRDMA_CREATE_QP_REQ_MAX_RQE_SHIFT	= 0,
 	OCRDMA_CREATE_QP_REQ_MAX_RQE_MASK	= 0xFFFF,
@@ -762,23 +796,23 @@ enum {
 					OCRDMA_CREATE_QP_REQ_MAX_SGE_SEND_SHIFT,
 
 	OCRDMA_CREATE_QP_REQ_FMR_EN_SHIFT		= 0,
-	OCRDMA_CREATE_QP_REQ_FMR_EN_MASK		= Bit(0),
+	OCRDMA_CREATE_QP_REQ_FMR_EN_MASK		= BIT(0),
 	OCRDMA_CREATE_QP_REQ_ZERO_LKEYEN_SHIFT		= 1,
-	OCRDMA_CREATE_QP_REQ_ZERO_LKEYEN_MASK		= Bit(1),
+	OCRDMA_CREATE_QP_REQ_ZERO_LKEYEN_MASK		= BIT(1),
 	OCRDMA_CREATE_QP_REQ_BIND_MEMWIN_SHIFT		= 2,
-	OCRDMA_CREATE_QP_REQ_BIND_MEMWIN_MASK		= Bit(2),
+	OCRDMA_CREATE_QP_REQ_BIND_MEMWIN_MASK		= BIT(2),
 	OCRDMA_CREATE_QP_REQ_INB_WREN_SHIFT		= 3,
-	OCRDMA_CREATE_QP_REQ_INB_WREN_MASK		= Bit(3),
+	OCRDMA_CREATE_QP_REQ_INB_WREN_MASK		= BIT(3),
 	OCRDMA_CREATE_QP_REQ_INB_RDEN_SHIFT		= 4,
-	OCRDMA_CREATE_QP_REQ_INB_RDEN_MASK		= Bit(4),
+	OCRDMA_CREATE_QP_REQ_INB_RDEN_MASK		= BIT(4),
 	OCRDMA_CREATE_QP_REQ_USE_SRQ_SHIFT		= 5,
-	OCRDMA_CREATE_QP_REQ_USE_SRQ_MASK		= Bit(5),
+	OCRDMA_CREATE_QP_REQ_USE_SRQ_MASK		= BIT(5),
 	OCRDMA_CREATE_QP_REQ_ENABLE_RPIR_SHIFT		= 6,
-	OCRDMA_CREATE_QP_REQ_ENABLE_RPIR_MASK		= Bit(6),
+	OCRDMA_CREATE_QP_REQ_ENABLE_RPIR_MASK		= BIT(6),
 	OCRDMA_CREATE_QP_REQ_ENABLE_DPP_SHIFT		= 7,
-	OCRDMA_CREATE_QP_REQ_ENABLE_DPP_MASK		= Bit(7),
+	OCRDMA_CREATE_QP_REQ_ENABLE_DPP_MASK		= BIT(7),
 	OCRDMA_CREATE_QP_REQ_ENABLE_DPP_CQ_SHIFT	= 8,
-	OCRDMA_CREATE_QP_REQ_ENABLE_DPP_CQ_MASK		= Bit(8),
+	OCRDMA_CREATE_QP_REQ_ENABLE_DPP_CQ_MASK		= BIT(8),
 	OCRDMA_CREATE_QP_REQ_MAX_SGE_RECV_SHIFT		= 16,
 	OCRDMA_CREATE_QP_REQ_MAX_SGE_RECV_MASK		= 0xFFFF <<
 				OCRDMA_CREATE_QP_REQ_MAX_SGE_RECV_SHIFT,
@@ -891,7 +925,7 @@ enum {
 	OCRDMA_CREATE_QP_RSP_SQ_ID_MASK			= 0xFFFF <<
 				OCRDMA_CREATE_QP_RSP_SQ_ID_SHIFT,
 
-	OCRDMA_CREATE_QP_RSP_DPP_ENABLED_MASK		= Bit(0),
+	OCRDMA_CREATE_QP_RSP_DPP_ENABLED_MASK		= BIT(0),
 	OCRDMA_CREATE_QP_RSP_DPP_PAGE_OFFSET_SHIFT	= 1,
 	OCRDMA_CREATE_QP_RSP_DPP_PAGE_OFFSET_MASK	= 0x7FFF <<
 				OCRDMA_CREATE_QP_RSP_DPP_PAGE_OFFSET_SHIFT,
@@ -928,38 +962,38 @@ enum {
 	OCRDMA_MODIFY_QP_ID_SHIFT	= 0,
 	OCRDMA_MODIFY_QP_ID_MASK	= 0xFFFF,
 
-	OCRDMA_QP_PARA_QPS_VALID	= Bit(0),
-	OCRDMA_QP_PARA_SQD_ASYNC_VALID	= Bit(1),
-	OCRDMA_QP_PARA_PKEY_VALID	= Bit(2),
-	OCRDMA_QP_PARA_QKEY_VALID	= Bit(3),
-	OCRDMA_QP_PARA_PMTU_VALID	= Bit(4),
-	OCRDMA_QP_PARA_ACK_TO_VALID	= Bit(5),
-	OCRDMA_QP_PARA_RETRY_CNT_VALID	= Bit(6),
-	OCRDMA_QP_PARA_RRC_VALID	= Bit(7),
-	OCRDMA_QP_PARA_RQPSN_VALID	= Bit(8),
-	OCRDMA_QP_PARA_MAX_IRD_VALID	= Bit(9),
-	OCRDMA_QP_PARA_MAX_ORD_VALID	= Bit(10),
-	OCRDMA_QP_PARA_RNT_VALID	= Bit(11),
-	OCRDMA_QP_PARA_SQPSN_VALID	= Bit(12),
-	OCRDMA_QP_PARA_DST_QPN_VALID	= Bit(13),
-	OCRDMA_QP_PARA_MAX_WQE_VALID	= Bit(14),
-	OCRDMA_QP_PARA_MAX_RQE_VALID	= Bit(15),
-	OCRDMA_QP_PARA_SGE_SEND_VALID	= Bit(16),
-	OCRDMA_QP_PARA_SGE_RECV_VALID	= Bit(17),
-	OCRDMA_QP_PARA_SGE_WR_VALID	= Bit(18),
-	OCRDMA_QP_PARA_INB_RDEN_VALID	= Bit(19),
-	OCRDMA_QP_PARA_INB_WREN_VALID	= Bit(20),
-	OCRDMA_QP_PARA_FLOW_LBL_VALID	= Bit(21),
-	OCRDMA_QP_PARA_BIND_EN_VALID	= Bit(22),
-	OCRDMA_QP_PARA_ZLKEY_EN_VALID	= Bit(23),
-	OCRDMA_QP_PARA_FMR_EN_VALID	= Bit(24),
-	OCRDMA_QP_PARA_INBAT_EN_VALID	= Bit(25),
-	OCRDMA_QP_PARA_VLAN_EN_VALID	= Bit(26),
-
-	OCRDMA_MODIFY_QP_FLAGS_RD	= Bit(0),
-	OCRDMA_MODIFY_QP_FLAGS_WR	= Bit(1),
-	OCRDMA_MODIFY_QP_FLAGS_SEND	= Bit(2),
-	OCRDMA_MODIFY_QP_FLAGS_ATOMIC	= Bit(3)
+	OCRDMA_QP_PARA_QPS_VALID	= BIT(0),
+	OCRDMA_QP_PARA_SQD_ASYNC_VALID	= BIT(1),
+	OCRDMA_QP_PARA_PKEY_VALID	= BIT(2),
+	OCRDMA_QP_PARA_QKEY_VALID	= BIT(3),
+	OCRDMA_QP_PARA_PMTU_VALID	= BIT(4),
+	OCRDMA_QP_PARA_ACK_TO_VALID	= BIT(5),
+	OCRDMA_QP_PARA_RETRY_CNT_VALID	= BIT(6),
+	OCRDMA_QP_PARA_RRC_VALID	= BIT(7),
+	OCRDMA_QP_PARA_RQPSN_VALID	= BIT(8),
+	OCRDMA_QP_PARA_MAX_IRD_VALID	= BIT(9),
+	OCRDMA_QP_PARA_MAX_ORD_VALID	= BIT(10),
+	OCRDMA_QP_PARA_RNT_VALID	= BIT(11),
+	OCRDMA_QP_PARA_SQPSN_VALID	= BIT(12),
+	OCRDMA_QP_PARA_DST_QPN_VALID	= BIT(13),
+	OCRDMA_QP_PARA_MAX_WQE_VALID	= BIT(14),
+	OCRDMA_QP_PARA_MAX_RQE_VALID	= BIT(15),
+	OCRDMA_QP_PARA_SGE_SEND_VALID	= BIT(16),
+	OCRDMA_QP_PARA_SGE_RECV_VALID	= BIT(17),
+	OCRDMA_QP_PARA_SGE_WR_VALID	= BIT(18),
+	OCRDMA_QP_PARA_INB_RDEN_VALID	= BIT(19),
+	OCRDMA_QP_PARA_INB_WREN_VALID	= BIT(20),
+	OCRDMA_QP_PARA_FLOW_LBL_VALID	= BIT(21),
+	OCRDMA_QP_PARA_BIND_EN_VALID	= BIT(22),
+	OCRDMA_QP_PARA_ZLKEY_EN_VALID	= BIT(23),
+	OCRDMA_QP_PARA_FMR_EN_VALID	= BIT(24),
+	OCRDMA_QP_PARA_INBAT_EN_VALID	= BIT(25),
+	OCRDMA_QP_PARA_VLAN_EN_VALID	= BIT(26),
+
+	OCRDMA_MODIFY_QP_FLAGS_RD	= BIT(0),
+	OCRDMA_MODIFY_QP_FLAGS_WR	= BIT(1),
+	OCRDMA_MODIFY_QP_FLAGS_SEND	= BIT(2),
+	OCRDMA_MODIFY_QP_FLAGS_ATOMIC	= BIT(3)
 };
 
 enum {
@@ -978,15 +1012,15 @@ enum {
 	OCRDMA_QP_PARAMS_MAX_SGE_SEND_MASK	= 0xFFFF <<
 					OCRDMA_QP_PARAMS_MAX_SGE_SEND_SHIFT,
 
-	OCRDMA_QP_PARAMS_FLAGS_FMR_EN		= Bit(0),
-	OCRDMA_QP_PARAMS_FLAGS_LKEY_0_EN	= Bit(1),
-	OCRDMA_QP_PARAMS_FLAGS_BIND_MW_EN	= Bit(2),
-	OCRDMA_QP_PARAMS_FLAGS_INBWR_EN		= Bit(3),
-	OCRDMA_QP_PARAMS_FLAGS_INBRD_EN		= Bit(4),
+	OCRDMA_QP_PARAMS_FLAGS_FMR_EN		= BIT(0),
+	OCRDMA_QP_PARAMS_FLAGS_LKEY_0_EN	= BIT(1),
+	OCRDMA_QP_PARAMS_FLAGS_BIND_MW_EN	= BIT(2),
+	OCRDMA_QP_PARAMS_FLAGS_INBWR_EN		= BIT(3),
+	OCRDMA_QP_PARAMS_FLAGS_INBRD_EN		= BIT(4),
 	OCRDMA_QP_PARAMS_STATE_SHIFT		= 5,
-	OCRDMA_QP_PARAMS_STATE_MASK		= Bit(5) | Bit(6) | Bit(7),
-	OCRDMA_QP_PARAMS_FLAGS_SQD_ASYNC	= Bit(8),
-	OCRDMA_QP_PARAMS_FLAGS_INB_ATEN		= Bit(9),
+	OCRDMA_QP_PARAMS_STATE_MASK		= BIT(5) | BIT(6) | BIT(7),
+	OCRDMA_QP_PARAMS_FLAGS_SQD_ASYNC	= BIT(8),
+	OCRDMA_QP_PARAMS_FLAGS_INB_ATEN		= BIT(9),
 	OCRDMA_QP_PARAMS_MAX_SGE_RECV_SHIFT	= 16,
 	OCRDMA_QP_PARAMS_MAX_SGE_RECV_MASK	= 0xFFFF <<
 					OCRDMA_QP_PARAMS_MAX_SGE_RECV_SHIFT,
@@ -1231,7 +1265,6 @@ struct ocrdma_destroy_srq {
 
 enum {
 	OCRDMA_ALLOC_PD_ENABLE_DPP	= BIT(16),
-	OCRDMA_PD_MAX_DPP_ENABLED_QP	= 8,
 	OCRDMA_DPP_PAGE_SIZE		= 4096
 };
 
@@ -1242,7 +1275,7 @@ struct ocrdma_alloc_pd {
 };
 
 enum {
-	OCRDMA_ALLOC_PD_RSP_DPP			= Bit(16),
+	OCRDMA_ALLOC_PD_RSP_DPP			= BIT(16),
 	OCRDMA_ALLOC_PD_RSP_DPP_PAGE_SHIFT	= 20,
 	OCRDMA_ALLOC_PD_RSP_PDID_MASK		= 0xFFFF,
 };
@@ -1274,18 +1307,18 @@ enum {
 	OCRDMA_ALLOC_LKEY_PD_ID_MASK		= 0xFFFF,
 
 	OCRDMA_ALLOC_LKEY_ADDR_CHECK_SHIFT	= 0,
-	OCRDMA_ALLOC_LKEY_ADDR_CHECK_MASK	= Bit(0),
+	OCRDMA_ALLOC_LKEY_ADDR_CHECK_MASK	= BIT(0),
 	OCRDMA_ALLOC_LKEY_FMR_SHIFT		= 1,
-	OCRDMA_ALLOC_LKEY_FMR_MASK		= Bit(1),
+	OCRDMA_ALLOC_LKEY_FMR_MASK		= BIT(1),
 	OCRDMA_ALLOC_LKEY_REMOTE_INV_SHIFT	= 2,
-	OCRDMA_ALLOC_LKEY_REMOTE_INV_MASK	= Bit(2),
+	OCRDMA_ALLOC_LKEY_REMOTE_INV_MASK	= BIT(2),
 	OCRDMA_ALLOC_LKEY_REMOTE_WR_SHIFT	= 3,
-	OCRDMA_ALLOC_LKEY_REMOTE_WR_MASK	= Bit(3),
+	OCRDMA_ALLOC_LKEY_REMOTE_WR_MASK	= BIT(3),
 	OCRDMA_ALLOC_LKEY_REMOTE_RD_SHIFT	= 4,
-	OCRDMA_ALLOC_LKEY_REMOTE_RD_MASK	= Bit(4),
+	OCRDMA_ALLOC_LKEY_REMOTE_RD_MASK	= BIT(4),
 	OCRDMA_ALLOC_LKEY_LOCAL_WR_SHIFT	= 5,
-	OCRDMA_ALLOC_LKEY_LOCAL_WR_MASK		= Bit(5),
-	OCRDMA_ALLOC_LKEY_REMOTE_ATOMIC_MASK	= Bit(6),
+	OCRDMA_ALLOC_LKEY_LOCAL_WR_MASK		= BIT(5),
+	OCRDMA_ALLOC_LKEY_REMOTE_ATOMIC_MASK	= BIT(6),
 	OCRDMA_ALLOC_LKEY_REMOTE_ATOMIC_SHIFT	= 6,
 	OCRDMA_ALLOC_LKEY_PBL_SIZE_SHIFT	= 16,
 	OCRDMA_ALLOC_LKEY_PBL_SIZE_MASK		= 0xFFFF <<
@@ -1344,21 +1377,21 @@ enum {
 	OCRDMA_REG_NSMR_HPAGE_SIZE_MASK		= 0xFF <<
 					OCRDMA_REG_NSMR_HPAGE_SIZE_SHIFT,
 	OCRDMA_REG_NSMR_BIND_MEMWIN_SHIFT	= 24,
-	OCRDMA_REG_NSMR_BIND_MEMWIN_MASK	= Bit(24),
+	OCRDMA_REG_NSMR_BIND_MEMWIN_MASK	= BIT(24),
 	OCRDMA_REG_NSMR_ZB_SHIFT		= 25,
-	OCRDMA_REG_NSMR_ZB_SHIFT_MASK		= Bit(25),
+	OCRDMA_REG_NSMR_ZB_SHIFT_MASK		= BIT(25),
 	OCRDMA_REG_NSMR_REMOTE_INV_SHIFT	= 26,
-	OCRDMA_REG_NSMR_REMOTE_INV_MASK		= Bit(26),
+	OCRDMA_REG_NSMR_REMOTE_INV_MASK		= BIT(26),
 	OCRDMA_REG_NSMR_REMOTE_WR_SHIFT		= 27,
-	OCRDMA_REG_NSMR_REMOTE_WR_MASK		= Bit(27),
+	OCRDMA_REG_NSMR_REMOTE_WR_MASK		= BIT(27),
 	OCRDMA_REG_NSMR_REMOTE_RD_SHIFT		= 28,
-	OCRDMA_REG_NSMR_REMOTE_RD_MASK		= Bit(28),
+	OCRDMA_REG_NSMR_REMOTE_RD_MASK		= BIT(28),
 	OCRDMA_REG_NSMR_LOCAL_WR_SHIFT		= 29,
-	OCRDMA_REG_NSMR_LOCAL_WR_MASK		= Bit(29),
+	OCRDMA_REG_NSMR_LOCAL_WR_MASK		= BIT(29),
 	OCRDMA_REG_NSMR_REMOTE_ATOMIC_SHIFT	= 30,
-	OCRDMA_REG_NSMR_REMOTE_ATOMIC_MASK	= Bit(30),
+	OCRDMA_REG_NSMR_REMOTE_ATOMIC_MASK	= BIT(30),
 	OCRDMA_REG_NSMR_LAST_SHIFT		= 31,
-	OCRDMA_REG_NSMR_LAST_MASK		= Bit(31)
+	OCRDMA_REG_NSMR_LAST_MASK		= BIT(31)
 };
 
 struct ocrdma_reg_nsmr {
@@ -1385,7 +1418,7 @@ enum {
 					OCRDMA_REG_NSMR_CONT_NUM_PBL_SHIFT,
 
 	OCRDMA_REG_NSMR_CONT_LAST_SHIFT		= 31,
-	OCRDMA_REG_NSMR_CONT_LAST_MASK		= Bit(31)
+	OCRDMA_REG_NSMR_CONT_LAST_MASK		= BIT(31)
 };
 
 struct ocrdma_reg_nsmr_cont {
@@ -1531,7 +1564,7 @@ struct ocrdma_delete_ah_tbl_rsp {
 
 enum {
 	OCRDMA_EQE_VALID_SHIFT		= 0,
-	OCRDMA_EQE_VALID_MASK		= Bit(0),
+	OCRDMA_EQE_VALID_MASK		= BIT(0),
 	OCRDMA_EQE_FOR_CQE_MASK		= 0xFFFE,
 	OCRDMA_EQE_RESOURCE_ID_SHIFT	= 16,
 	OCRDMA_EQE_RESOURCE_ID_MASK	= 0xFFFF <<
@@ -1589,11 +1622,11 @@ enum {
 	OCRDMA_CQE_UD_STATUS_MASK	= 0x7 << OCRDMA_CQE_UD_STATUS_SHIFT,
 	OCRDMA_CQE_STATUS_SHIFT		= 16,
 	OCRDMA_CQE_STATUS_MASK		= 0xFF << OCRDMA_CQE_STATUS_SHIFT,
-	OCRDMA_CQE_VALID		= Bit(31),
-	OCRDMA_CQE_INVALIDATE		= Bit(30),
-	OCRDMA_CQE_QTYPE		= Bit(29),
-	OCRDMA_CQE_IMM			= Bit(28),
-	OCRDMA_CQE_WRITE_IMM		= Bit(27),
+	OCRDMA_CQE_VALID		= BIT(31),
+	OCRDMA_CQE_INVALIDATE		= BIT(30),
+	OCRDMA_CQE_QTYPE		= BIT(29),
+	OCRDMA_CQE_IMM			= BIT(28),
+	OCRDMA_CQE_WRITE_IMM		= BIT(27),
 	OCRDMA_CQE_QTYPE_SQ		= 0,
 	OCRDMA_CQE_QTYPE_RQ		= 1,
 	OCRDMA_CQE_SRCQP_MASK		= 0xFFFFFF
@@ -1737,8 +1770,8 @@ struct ocrdma_grh {
 	u16	rsvd;
 } __packed;
 
-#define OCRDMA_AV_VALID		Bit(7)
-#define OCRDMA_AV_VLAN_VALID	Bit(1)
+#define OCRDMA_AV_VALID		BIT(7)
+#define OCRDMA_AV_VLAN_VALID	BIT(1)
 
 struct ocrdma_av {
 	struct ocrdma_eth_vlan eth_hdr;
@@ -1896,12 +1929,62 @@ struct ocrdma_rdma_stats_resp {
 	struct ocrdma_rx_dbg_stats	rx_dbg_stats;
 } __packed;
 
+enum {
+	OCRDMA_HBA_ATTRB_EPROM_VER_LO_MASK	= 0xFF,
+	OCRDMA_HBA_ATTRB_EPROM_VER_HI_MASK	= 0xFF00,
+	OCRDMA_HBA_ATTRB_EPROM_VER_HI_SHIFT	= 0x08,
+	OCRDMA_HBA_ATTRB_CDBLEN_MASK		= 0xFFFF,
+	OCRDMA_HBA_ATTRB_ASIC_REV_MASK		= 0xFF0000,
+	OCRDMA_HBA_ATTRB_ASIC_REV_SHIFT		= 0x10,
+	OCRDMA_HBA_ATTRB_GUID0_MASK		= 0xFF000000,
+	OCRDMA_HBA_ATTRB_GUID0_SHIFT		= 0x18,
+	OCRDMA_HBA_ATTRB_GUID13_MASK		= 0xFF,
+	OCRDMA_HBA_ATTRB_GUID14_MASK		= 0xFF00,
+	OCRDMA_HBA_ATTRB_GUID14_SHIFT		= 0x08,
+	OCRDMA_HBA_ATTRB_GUID15_MASK		= 0xFF0000,
+	OCRDMA_HBA_ATTRB_GUID15_SHIFT		= 0x10,
+	OCRDMA_HBA_ATTRB_PCNT_MASK		= 0xFF000000,
+	OCRDMA_HBA_ATTRB_PCNT_SHIFT		= 0x18,
+	OCRDMA_HBA_ATTRB_LDTOUT_MASK		= 0xFFFF,
+	OCRDMA_HBA_ATTRB_ISCSI_VER_MASK		= 0xFF0000,
+	OCRDMA_HBA_ATTRB_ISCSI_VER_SHIFT	= 0x10,
+	OCRDMA_HBA_ATTRB_MFUNC_DEV_MASK		= 0xFF000000,
+	OCRDMA_HBA_ATTRB_MFUNC_DEV_SHIFT	= 0x18,
+	OCRDMA_HBA_ATTRB_CV_MASK		= 0xFF,
+	OCRDMA_HBA_ATTRB_HBA_ST_MASK		= 0xFF00,
+	OCRDMA_HBA_ATTRB_HBA_ST_SHIFT		= 0x08,
+	OCRDMA_HBA_ATTRB_MAX_DOMS_MASK		= 0xFF0000,
+	OCRDMA_HBA_ATTRB_MAX_DOMS_SHIFT		= 0x10,
+	OCRDMA_HBA_ATTRB_PTNUM_MASK		= 0x3F000000,
+	OCRDMA_HBA_ATTRB_PTNUM_SHIFT		= 0x18,
+	OCRDMA_HBA_ATTRB_PT_MASK		= 0xC0000000,
+	OCRDMA_HBA_ATTRB_PT_SHIFT		= 0x1E,
+	OCRDMA_HBA_ATTRB_ISCSI_FET_MASK		= 0xFF,
+	OCRDMA_HBA_ATTRB_ASIC_GEN_MASK		= 0xFF00,
+	OCRDMA_HBA_ATTRB_ASIC_GEN_SHIFT		= 0x08,
+	OCRDMA_HBA_ATTRB_PCI_VID_MASK		= 0xFFFF,
+	OCRDMA_HBA_ATTRB_PCI_DID_MASK		= 0xFFFF0000,
+	OCRDMA_HBA_ATTRB_PCI_DID_SHIFT		= 0x10,
+	OCRDMA_HBA_ATTRB_PCI_SVID_MASK		= 0xFFFF,
+	OCRDMA_HBA_ATTRB_PCI_SSID_MASK		= 0xFFFF0000,
+	OCRDMA_HBA_ATTRB_PCI_SSID_SHIFT		= 0x10,
+	OCRDMA_HBA_ATTRB_PCI_BUSNUM_MASK	= 0xFF,
+	OCRDMA_HBA_ATTRB_PCI_DEVNUM_MASK	= 0xFF00,
+	OCRDMA_HBA_ATTRB_PCI_DEVNUM_SHIFT	= 0x08,
+	OCRDMA_HBA_ATTRB_PCI_FUNCNUM_MASK	= 0xFF0000,
+	OCRDMA_HBA_ATTRB_PCI_FUNCNUM_SHIFT	= 0x10,
+	OCRDMA_HBA_ATTRB_IF_TYPE_MASK		= 0xFF000000,
+	OCRDMA_HBA_ATTRB_IF_TYPE_SHIFT		= 0x18,
+	OCRDMA_HBA_ATTRB_NETFIL_MASK		=0xFF
+};
 
 struct mgmt_hba_attribs {
 	u8 flashrom_version_string[32];
 	u8 manufacturer_name[32];
 	u32 supported_modes;
-	u32 rsvd0[3];
+	u32 rsvd_eprom_verhi_verlo;
+	u32 mbx_ds_ver;
+	u32 epfw_ds_ver;
 	u8 ncsi_ver_string[12];
 	u32 default_extended_timeout;
 	u8 controller_model_number[32];
@@ -1914,34 +1997,26 @@ struct mgmt_hba_attribs {
 	u8 driver_version_string[32];
 	u8 fw_on_flash_version_string[32];
 	u32 functionalities_supported;
-	u16 max_cdblength;
-	u8 asic_revision;
-	u8 generational_guid[16];
-	u8 hba_port_count;
-	u16 default_link_down_timeout;
-	u8 iscsi_ver_min_max;
-	u8 multifunction_device;
-	u8 cache_valid;
-	u8 hba_status;
-	u8 max_domains_supported;
-	u8 phy_port;
+	u32 guid0_asicrev_cdblen;
+	u8 generational_guid[12];
+	u32 portcnt_guid15;
+	u32 mfuncdev_iscsi_ldtout;
+	u32 ptpnum_maxdoms_hbast_cv;
 	u32 firmware_post_status;
 	u32 hba_mtu[8];
-	u32 rsvd1[4];
+	u32 res_asicgen_iscsi_feaures;
+	u32 rsvd1[3];
 };
 
 struct mgmt_controller_attrib {
 	struct mgmt_hba_attribs hba_attribs;
-	u16 pci_vendor_id;
-	u16 pci_device_id;
-	u16 pci_sub_vendor_id;
-	u16 pci_sub_system_id;
-	u8 pci_bus_number;
-	u8 pci_device_number;
-	u8 pci_function_number;
-	u8 interface_type;
-	u64 unique_identifier;
-	u32 rsvd0[5];
+	u32 pci_did_vid;
+	u32 pci_ssid_svid;
+	u32 ityp_fnum_devnum_bnum;
+	u32 uid_hi;
+	u32 uid_lo;
+	u32 res_nnetfil;
+	u32 rsvd0[4];
 };
 
 struct ocrdma_get_ctrl_attribs_rsp {
@@ -1949,5 +2024,79 @@ struct ocrdma_get_ctrl_attribs_rsp {
 	struct mgmt_controller_attrib ctrl_attribs;
 };
 
+#define OCRDMA_SUBSYS_DCBX 0x10
+
+enum OCRDMA_DCBX_OPCODE {
+	OCRDMA_CMD_GET_DCBX_CONFIG = 0x01
+};
+
+enum OCRDMA_DCBX_PARAM_TYPE {
+	OCRDMA_PARAMETER_TYPE_ADMIN	= 0x00,
+	OCRDMA_PARAMETER_TYPE_OPER	= 0x01,
+	OCRDMA_PARAMETER_TYPE_PEER	= 0x02
+};
+
+enum OCRDMA_DCBX_APP_PROTO {
+	OCRDMA_APP_PROTO_ROCE	= 0x8915
+};
+
+enum OCRDMA_DCBX_PROTO {
+	OCRDMA_PROTO_SELECT_L2	= 0x00,
+	OCRDMA_PROTO_SELECT_L4	= 0x01
+};
+
+enum OCRDMA_DCBX_APP_PARAM {
+	OCRDMA_APP_PARAM_APP_PROTO_MASK = 0xFFFF,
+	OCRDMA_APP_PARAM_PROTO_SEL_MASK = 0xFF,
+	OCRDMA_APP_PARAM_PROTO_SEL_SHIFT = 0x10,
+	OCRDMA_APP_PARAM_VALID_MASK	= 0xFF,
+	OCRDMA_APP_PARAM_VALID_SHIFT	= 0x18
+};
+
+enum OCRDMA_DCBX_STATE_FLAGS {
+	OCRDMA_STATE_FLAG_ENABLED	= 0x01,
+	OCRDMA_STATE_FLAG_ADDVERTISED	= 0x02,
+	OCRDMA_STATE_FLAG_WILLING	= 0x04,
+	OCRDMA_STATE_FLAG_SYNC		= 0x08,
+	OCRDMA_STATE_FLAG_UNSUPPORTED	= 0x40000000,
+	OCRDMA_STATE_FLAG_NEG_FAILD	= 0x80000000
+};
+
+enum OCRDMA_TCV_AEV_OPV_ST {
+	OCRDMA_DCBX_TC_SUPPORT_MASK	= 0xFF,
+	OCRDMA_DCBX_TC_SUPPORT_SHIFT	= 0x18,
+	OCRDMA_DCBX_APP_ENTRY_SHIFT	= 0x10,
+	OCRDMA_DCBX_OP_PARAM_SHIFT	= 0x08,
+	OCRDMA_DCBX_STATE_MASK		= 0xFF
+};
+
+struct ocrdma_app_parameter {
+	u32 valid_proto_app;
+	u32 oui;
+	u32 app_prio[2];
+};
+
+struct ocrdma_dcbx_cfg {
+	u32 tcv_aev_opv_st;
+	u32 tc_state;
+	u32 pfc_state;
+	u32 qcn_state;
+	u32 appl_state;
+	u32 ll_state;
+	u32 tc_bw[2];
+	u32 tc_prio[8];
+	u32 pfc_prio[2];
+	struct ocrdma_app_parameter app_param[15];
+};
+
+struct ocrdma_get_dcbx_cfg_req {
+	struct ocrdma_mbx_hdr hdr;
+	u32 param_type;
+} __packed;
+
+struct ocrdma_get_dcbx_cfg_rsp {
+	struct ocrdma_mbx_rsp hdr;
+	struct ocrdma_dcbx_cfg cfg;
+} __packed;
 
 #endif				/* __OCRDMA_SLI_H__ */
diff --git a/drivers/infiniband/hw/ocrdma/ocrdma_verbs.c b/drivers/infiniband/hw/ocrdma/ocrdma_verbs.c
index 8bae8718fa53..4c68305ee781 100644
--- a/drivers/infiniband/hw/ocrdma/ocrdma_verbs.c
+++ b/drivers/infiniband/hw/ocrdma/ocrdma_verbs.c
@@ -69,11 +69,11 @@ int ocrdma_query_device(struct ib_device *ibdev, struct ib_device_attr *attr)
 	memcpy(&attr->fw_ver, &dev->attr.fw_ver[0],
 	       min(sizeof(dev->attr.fw_ver), sizeof(attr->fw_ver)));
 	ocrdma_get_guid(dev, (u8 *)&attr->sys_image_guid);
-	attr->max_mr_size = ~0ull;
+	attr->max_mr_size = dev->attr.max_mr_size;
 	attr->page_size_cap = 0xffff000;
 	attr->vendor_id = dev->nic_info.pdev->vendor;
 	attr->vendor_part_id = dev->nic_info.pdev->device;
-	attr->hw_ver = 0;
+	attr->hw_ver = dev->asic_id;
 	attr->max_qp = dev->attr.max_qp;
 	attr->max_ah = OCRDMA_MAX_AH;
 	attr->max_qp_wr = dev->attr.max_wqe;
@@ -101,7 +101,7 @@ int ocrdma_query_device(struct ib_device *ibdev, struct ib_device_attr *attr)
 	attr->max_srq_sge = dev->attr.max_srq_sge;
 	attr->max_srq_wr = dev->attr.max_rqe;
 	attr->local_ca_ack_delay = dev->attr.local_ca_ack_delay;
-	attr->max_fast_reg_page_list_len = 0;
+	attr->max_fast_reg_page_list_len = dev->attr.max_pages_per_frmr;
 	attr->max_pkeys = 1;
 	return 0;
 }
@@ -268,7 +268,8 @@ static struct ocrdma_pd *_ocrdma_alloc_pd(struct ocrdma_dev *dev,
 		pd->dpp_enabled =
 			ocrdma_get_asic_type(dev) == OCRDMA_ASIC_GEN_SKH_R;
 		pd->num_dpp_qp =
-			pd->dpp_enabled ? OCRDMA_PD_MAX_DPP_ENABLED_QP : 0;
+			pd->dpp_enabled ? (dev->nic_info.db_page_size /
+					   dev->attr.wqe_size) : 0;
 	}
 
 retry:
@@ -328,7 +329,10 @@ static int ocrdma_dealloc_ucontext_pd(struct ocrdma_ucontext *uctx)
 	struct ocrdma_pd *pd = uctx->cntxt_pd;
 	struct ocrdma_dev *dev = get_ocrdma_dev(pd->ibpd.device);
 
-	BUG_ON(uctx->pd_in_use);
+	if (uctx->pd_in_use) {
+		pr_err("%s(%d) Freeing in use pdid=0x%x.\n",
+		       __func__, dev->id, pd->id);
+	}
 	uctx->cntxt_pd = NULL;
 	status = _ocrdma_dealloc_pd(dev, pd);
 	return status;
@@ -384,7 +388,7 @@ struct ib_ucontext *ocrdma_alloc_ucontext(struct ib_device *ibdev,
 
 	memset(&resp, 0, sizeof(resp));
 	resp.ah_tbl_len = ctx->ah_tbl.len;
-	resp.ah_tbl_page = ctx->ah_tbl.pa;
+	resp.ah_tbl_page = virt_to_phys(ctx->ah_tbl.va);
 
 	status = ocrdma_add_mmap(ctx, resp.ah_tbl_page, resp.ah_tbl_len);
 	if (status)
@@ -843,6 +847,13 @@ int ocrdma_dereg_mr(struct ib_mr *ib_mr)
 	if (mr->umem)
 		ib_umem_release(mr->umem);
 	kfree(mr);
+
+	/* Don't stop cleanup, in case FW is unresponsive */
+	if (dev->mqe_ctx.fw_error_state) {
+		status = 0;
+		pr_err("%s(%d) fw not responding.\n",
+		       __func__, dev->id);
+	}
 	return status;
 }
 
@@ -859,7 +870,7 @@ static int ocrdma_copy_cq_uresp(struct ocrdma_dev *dev, struct ocrdma_cq *cq,
 	uresp.page_size = PAGE_ALIGN(cq->len);
 	uresp.num_pages = 1;
 	uresp.max_hw_cqe = cq->max_hw_cqe;
-	uresp.page_addr[0] = cq->pa;
+	uresp.page_addr[0] = virt_to_phys(cq->va);
 	uresp.db_page_addr =  ocrdma_get_db_addr(dev, uctx->cntxt_pd->id);
 	uresp.db_page_size = dev->nic_info.db_page_size;
 	uresp.phase_change = cq->phase_change ? 1 : 0;
@@ -1112,13 +1123,13 @@ static int ocrdma_copy_qp_uresp(struct ocrdma_qp *qp,
 	uresp.sq_dbid = qp->sq.dbid;
 	uresp.num_sq_pages = 1;
 	uresp.sq_page_size = PAGE_ALIGN(qp->sq.len);
-	uresp.sq_page_addr[0] = qp->sq.pa;
+	uresp.sq_page_addr[0] = virt_to_phys(qp->sq.va);
 	uresp.num_wqe_allocated = qp->sq.max_cnt;
 	if (!srq) {
 		uresp.rq_dbid = qp->rq.dbid;
 		uresp.num_rq_pages = 1;
 		uresp.rq_page_size = PAGE_ALIGN(qp->rq.len);
-		uresp.rq_page_addr[0] = qp->rq.pa;
+		uresp.rq_page_addr[0] = virt_to_phys(qp->rq.va);
 		uresp.num_rqe_allocated = qp->rq.max_cnt;
 	}
 	uresp.db_page_addr = usr_db;
@@ -1669,7 +1680,7 @@ static int ocrdma_copy_srq_uresp(struct ocrdma_dev *dev, struct ocrdma_srq *srq,
 	memset(&uresp, 0, sizeof(uresp));
 	uresp.rq_dbid = srq->rq.dbid;
 	uresp.num_rq_pages = 1;
-	uresp.rq_page_addr[0] = srq->rq.pa;
+	uresp.rq_page_addr[0] = virt_to_phys(srq->rq.va);
 	uresp.rq_page_size = srq->rq.len;
 	uresp.db_page_addr = dev->nic_info.unmapped_db +
 	    (srq->pd->id * dev->nic_info.db_page_size);
@@ -2054,6 +2065,13 @@ int ocrdma_post_send(struct ib_qp *ibqp, struct ib_send_wr *wr,
 	}
 
 	while (wr) {
+		if (qp->qp_type == IB_QPT_UD &&
+		    (wr->opcode != IB_WR_SEND &&
+		     wr->opcode != IB_WR_SEND_WITH_IMM)) {
+			*bad_wr = wr;
+			status = -EINVAL;
+			break;
+		}
 		if (ocrdma_hwq_free_cnt(&qp->sq) == 0 ||
 		    wr->num_sge > qp->sq.max_sges) {
 			*bad_wr = wr;
@@ -2488,6 +2506,11 @@ static bool ocrdma_poll_err_scqe(struct ocrdma_qp *qp,
 			*stop = true;
 			expand = false;
 		}
+	} else if (is_hw_sq_empty(qp)) {
+		/* Do nothing */
+		expand = false;
+		*polled = false;
+		*stop = false;
 	} else {
 		*polled = true;
 		expand = ocrdma_update_err_scqe(ibwc, cqe, qp, status);
@@ -2593,6 +2616,11 @@ static bool ocrdma_poll_err_rcqe(struct ocrdma_qp *qp, struct ocrdma_cqe *cqe,
 			*stop = true;
 			expand = false;
 		}
+	} else if (is_hw_rq_empty(qp)) {
+		/* Do nothing */
+		expand = false;
+		*polled = false;
+		*stop = false;
 	} else {
 		*polled = true;
 		expand = ocrdma_update_err_rcqe(ibwc, cqe, qp, status);
@@ -2818,11 +2846,9 @@ int ocrdma_arm_cq(struct ib_cq *ibcq, enum ib_cq_notify_flags cq_flags)
 	if (cq->first_arm) {
 		ocrdma_ring_cq_db(dev, cq_id, arm_needed, sol_needed, 0);
 		cq->first_arm = false;
-		goto skip_defer;
 	}
-	cq->deferred_arm = true;
 
-skip_defer:
+	cq->deferred_arm = true;
 	cq->deferred_sol = sol_needed;
 	spin_unlock_irqrestore(&cq->cq_lock, flags);
 
diff --git a/drivers/infiniband/hw/qib/qib_debugfs.c b/drivers/infiniband/hw/qib/qib_debugfs.c
index 799a0c3bffc4..6abd3ed3cd51 100644
--- a/drivers/infiniband/hw/qib/qib_debugfs.c
+++ b/drivers/infiniband/hw/qib/qib_debugfs.c
@@ -193,6 +193,7 @@ static void *_qp_stats_seq_start(struct seq_file *s, loff_t *pos)
 	struct qib_qp_iter *iter;
 	loff_t n = *pos;
 
+	rcu_read_lock();
 	iter = qib_qp_iter_init(s->private);
 	if (!iter)
 		return NULL;
@@ -224,7 +225,7 @@ static void *_qp_stats_seq_next(struct seq_file *s, void *iter_ptr,
 
 static void _qp_stats_seq_stop(struct seq_file *s, void *iter_ptr)
 {
-	/* nothing for now */
+	rcu_read_unlock();
 }
 
 static int _qp_stats_seq_show(struct seq_file *s, void *iter_ptr)
diff --git a/drivers/infiniband/hw/qib/qib_fs.c b/drivers/infiniband/hw/qib/qib_fs.c
index cab610ccd50e..81854586c081 100644
--- a/drivers/infiniband/hw/qib/qib_fs.c
+++ b/drivers/infiniband/hw/qib/qib_fs.c
@@ -89,7 +89,6 @@ static int create_file(const char *name, umode_t mode,
 {
 	int error;
 
-	*dentry = NULL;
 	mutex_lock(&parent->d_inode->i_mutex);
 	*dentry = lookup_one_len(name, parent, strlen(name));
 	if (!IS_ERR(*dentry))
diff --git a/drivers/infiniband/hw/qib/qib_init.c b/drivers/infiniband/hw/qib/qib_init.c
index 8d3c78ddc906..729da39c49ed 100644
--- a/drivers/infiniband/hw/qib/qib_init.c
+++ b/drivers/infiniband/hw/qib/qib_init.c
@@ -1222,7 +1222,7 @@ static int qib_init_one(struct pci_dev *, const struct pci_device_id *);
 #define DRIVER_LOAD_MSG "Intel " QIB_DRV_NAME " loaded: "
 #define PFX QIB_DRV_NAME ": "
 
-static DEFINE_PCI_DEVICE_TABLE(qib_pci_tbl) = {
+static const struct pci_device_id qib_pci_tbl[] = {
 	{ PCI_DEVICE(PCI_VENDOR_ID_PATHSCALE, PCI_DEVICE_ID_QLOGIC_IB_6120) },
 	{ PCI_DEVICE(PCI_VENDOR_ID_QLOGIC, PCI_DEVICE_ID_QLOGIC_IB_7220) },
 	{ PCI_DEVICE(PCI_VENDOR_ID_QLOGIC, PCI_DEVICE_ID_QLOGIC_IB_7322) },
diff --git a/drivers/infiniband/hw/qib/qib_mad.c b/drivers/infiniband/hw/qib/qib_mad.c
index 22c720e5740d..636be117b578 100644
--- a/drivers/infiniband/hw/qib/qib_mad.c
+++ b/drivers/infiniband/hw/qib/qib_mad.c
@@ -2476,7 +2476,7 @@ int qib_create_agents(struct qib_ibdev *dev)
 		ibp = &dd->pport[p].ibport_data;
 		agent = ib_register_mad_agent(&dev->ibdev, p + 1, IB_QPT_SMI,
 					      NULL, 0, send_handler,
-					      NULL, NULL);
+					      NULL, NULL, 0);
 		if (IS_ERR(agent)) {
 			ret = PTR_ERR(agent);
 			goto err;
diff --git a/drivers/infiniband/hw/qib/qib_qp.c b/drivers/infiniband/hw/qib/qib_qp.c
index 7fcc150d603c..6ddc0264aad2 100644
--- a/drivers/infiniband/hw/qib/qib_qp.c
+++ b/drivers/infiniband/hw/qib/qib_qp.c
@@ -1325,7 +1325,6 @@ int qib_qp_iter_next(struct qib_qp_iter *iter)
 	struct qib_qp *pqp = iter->qp;
 	struct qib_qp *qp;
 
-	rcu_read_lock();
 	for (; n < dev->qp_table_size; n++) {
 		if (pqp)
 			qp = rcu_dereference(pqp->next);
@@ -1333,18 +1332,11 @@ int qib_qp_iter_next(struct qib_qp_iter *iter)
 			qp = rcu_dereference(dev->qp_table[n]);
 		pqp = qp;
 		if (qp) {
-			if (iter->qp)
-				atomic_dec(&iter->qp->refcount);
-			atomic_inc(&qp->refcount);
-			rcu_read_unlock();
 			iter->qp = qp;
 			iter->n = n;
 			return 0;
 		}
 	}
-	rcu_read_unlock();
-	if (iter->qp)
-		atomic_dec(&iter->qp->refcount);
 	return ret;
 }
 
diff --git a/drivers/infiniband/hw/qib/qib_user_pages.c b/drivers/infiniband/hw/qib/qib_user_pages.c
index 2bc1d2b96298..74f90b2619f6 100644
--- a/drivers/infiniband/hw/qib/qib_user_pages.c
+++ b/drivers/infiniband/hw/qib/qib_user_pages.c
@@ -52,7 +52,7 @@ static void __qib_release_user_pages(struct page **p, size_t num_pages,
  * Call with current->mm->mmap_sem held.
  */
 static int __qib_get_user_pages(unsigned long start_page, size_t num_pages,
-				struct page **p, struct vm_area_struct **vma)
+				struct page **p)
 {
 	unsigned long lock_limit;
 	size_t got;
@@ -69,7 +69,7 @@ static int __qib_get_user_pages(unsigned long start_page, size_t num_pages,
 		ret = get_user_pages(current, current->mm,
 				     start_page + got * PAGE_SIZE,
 				     num_pages - got, 1, 1,
-				     p + got, vma);
+				     p + got, NULL);
 		if (ret < 0)
 			goto bail_release;
 	}
@@ -136,7 +136,7 @@ int qib_get_user_pages(unsigned long start_page, size_t num_pages,
 
 	down_write(&current->mm->mmap_sem);
 
-	ret = __qib_get_user_pages(start_page, num_pages, p, NULL);
+	ret = __qib_get_user_pages(start_page, num_pages, p);
 
 	up_write(&current->mm->mmap_sem);
 
diff --git a/drivers/infiniband/hw/usnic/usnic_ib_main.c b/drivers/infiniband/hw/usnic/usnic_ib_main.c
index fb6d026f92cd..0d0f98695d53 100644
--- a/drivers/infiniband/hw/usnic/usnic_ib_main.c
+++ b/drivers/infiniband/hw/usnic/usnic_ib_main.c
@@ -490,7 +490,7 @@ out:
 
 /* Start of PCI section */
 
-static DEFINE_PCI_DEVICE_TABLE(usnic_ib_pci_ids) = {
+static const struct pci_device_id usnic_ib_pci_ids[] = {
 	{PCI_DEVICE(PCI_VENDOR_ID_CISCO, PCI_DEVICE_ID_CISCO_VIC_USPACE_NIC)},
 	{0,}
 };
diff --git a/drivers/infiniband/hw/usnic/usnic_uiom.c b/drivers/infiniband/hw/usnic/usnic_uiom.c
index 801a1d6937e4..417de1f32960 100644
--- a/drivers/infiniband/hw/usnic/usnic_uiom.c
+++ b/drivers/infiniband/hw/usnic/usnic_uiom.c
@@ -507,7 +507,7 @@ int usnic_uiom_attach_dev_to_pd(struct usnic_uiom_pd *pd, struct device *dev)
 	if (err)
 		goto out_free_dev;
 
-	if (!iommu_domain_has_cap(pd->domain, IOMMU_CAP_CACHE_COHERENCY)) {
+	if (!iommu_capable(dev->bus, IOMMU_CAP_CACHE_COHERENCY)) {
 		usnic_err("IOMMU of %s does not support cache coherency\n",
 				dev_name(dev));
 		err = -EINVAL;
diff --git a/drivers/infiniband/ulp/ipoib/ipoib.h b/drivers/infiniband/ulp/ipoib/ipoib.h
index c639f90cfda4..d7562beb5423 100644
--- a/drivers/infiniband/ulp/ipoib/ipoib.h
+++ b/drivers/infiniband/ulp/ipoib/ipoib.h
@@ -86,7 +86,6 @@ enum {
 	IPOIB_FLAG_INITIALIZED	  = 1,
 	IPOIB_FLAG_ADMIN_UP	  = 2,
 	IPOIB_PKEY_ASSIGNED	  = 3,
-	IPOIB_PKEY_STOP		  = 4,
 	IPOIB_FLAG_SUBINTERFACE	  = 5,
 	IPOIB_MCAST_RUN		  = 6,
 	IPOIB_STOP_REAPER	  = 7,
@@ -132,6 +131,12 @@ struct ipoib_cb {
 	u8			hwaddr[INFINIBAND_ALEN];
 };
 
+static inline struct ipoib_cb *ipoib_skb_cb(const struct sk_buff *skb)
+{
+	BUILD_BUG_ON(sizeof(skb->cb) < sizeof(struct ipoib_cb));
+	return (struct ipoib_cb *)skb->cb;
+}
+
 /* Used for all multicast joins (broadcast, IPv4 mcast and IPv6 mcast) */
 struct ipoib_mcast {
 	struct ib_sa_mcmember_rec mcmember;
@@ -312,7 +317,6 @@ struct ipoib_dev_priv {
 	struct list_head multicast_list;
 	struct rb_root multicast_tree;
 
-	struct delayed_work pkey_poll_task;
 	struct delayed_work mcast_task;
 	struct work_struct carrier_on_task;
 	struct work_struct flush_light;
@@ -473,10 +477,11 @@ 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(struct net_device *dev);
+int ipoib_ib_dev_open(struct net_device *dev, int flush);
 int ipoib_ib_dev_up(struct net_device *dev);
 int ipoib_ib_dev_down(struct net_device *dev, int flush);
 int ipoib_ib_dev_stop(struct net_device *dev, int flush);
+void ipoib_pkey_dev_check_presence(struct net_device *dev);
 
 int ipoib_dev_init(struct net_device *dev, struct ib_device *ca, int port);
 void ipoib_dev_cleanup(struct net_device *dev);
@@ -532,8 +537,7 @@ int  ipoib_set_mode(struct net_device *dev, const char *buf);
 
 void ipoib_setup(struct net_device *dev);
 
-void ipoib_pkey_poll(struct work_struct *work);
-int ipoib_pkey_dev_delay_open(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);
diff --git a/drivers/infiniband/ulp/ipoib/ipoib_fs.c b/drivers/infiniband/ulp/ipoib/ipoib_fs.c
index 50061854616e..6bd5740e2691 100644
--- a/drivers/infiniband/ulp/ipoib/ipoib_fs.c
+++ b/drivers/infiniband/ulp/ipoib/ipoib_fs.c
@@ -281,10 +281,8 @@ void ipoib_delete_debug_files(struct net_device *dev)
 {
 	struct ipoib_dev_priv *priv = netdev_priv(dev);
 
-	if (priv->mcg_dentry)
-		debugfs_remove(priv->mcg_dentry);
-	if (priv->path_dentry)
-		debugfs_remove(priv->path_dentry);
+	debugfs_remove(priv->mcg_dentry);
+	debugfs_remove(priv->path_dentry);
 }
 
 int ipoib_register_debugfs(void)
diff --git a/drivers/infiniband/ulp/ipoib/ipoib_ib.c b/drivers/infiniband/ulp/ipoib/ipoib_ib.c
index 6a7003ddb0be..72626c348174 100644
--- a/drivers/infiniband/ulp/ipoib/ipoib_ib.c
+++ b/drivers/infiniband/ulp/ipoib/ipoib_ib.c
@@ -664,17 +664,18 @@ static void ipoib_ib_tx_timer_func(unsigned long ctx)
 	drain_tx_cq((struct net_device *)ctx);
 }
 
-int ipoib_ib_dev_open(struct net_device *dev)
+int ipoib_ib_dev_open(struct net_device *dev, int flush)
 {
 	struct ipoib_dev_priv *priv = netdev_priv(dev);
 	int ret;
 
-	if (ib_find_pkey(priv->ca, priv->port, priv->pkey, &priv->pkey_index)) {
-		ipoib_warn(priv, "P_Key 0x%04x not found\n", priv->pkey);
-		clear_bit(IPOIB_PKEY_ASSIGNED, &priv->flags);
+	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;
 	}
-	set_bit(IPOIB_PKEY_ASSIGNED, &priv->flags);
 
 	ret = ipoib_init_qp(dev);
 	if (ret) {
@@ -705,16 +706,17 @@ int ipoib_ib_dev_open(struct net_device *dev)
 dev_stop:
 	if (!test_and_set_bit(IPOIB_FLAG_INITIALIZED, &priv->flags))
 		napi_enable(&priv->napi);
-	ipoib_ib_dev_stop(dev, 1);
+	ipoib_ib_dev_stop(dev, flush);
 	return -1;
 }
 
-static void ipoib_pkey_dev_check_presence(struct net_device *dev)
+void ipoib_pkey_dev_check_presence(struct net_device *dev)
 {
 	struct ipoib_dev_priv *priv = netdev_priv(dev);
-	u16 pkey_index = 0;
 
-	if (ib_find_pkey(priv->ca, priv->port, priv->pkey, &pkey_index))
+	if (!(priv->pkey & 0x7fff) ||
+	    ib_find_pkey(priv->ca, priv->port, priv->pkey,
+			 &priv->pkey_index))
 		clear_bit(IPOIB_PKEY_ASSIGNED, &priv->flags);
 	else
 		set_bit(IPOIB_PKEY_ASSIGNED, &priv->flags);
@@ -745,14 +747,6 @@ int ipoib_ib_dev_down(struct net_device *dev, int flush)
 	clear_bit(IPOIB_FLAG_OPER_UP, &priv->flags);
 	netif_carrier_off(dev);
 
-	/* Shutdown the P_Key thread if still active */
-	if (!test_bit(IPOIB_PKEY_ASSIGNED, &priv->flags)) {
-		mutex_lock(&pkey_mutex);
-		set_bit(IPOIB_PKEY_STOP, &priv->flags);
-		cancel_delayed_work_sync(&priv->pkey_poll_task);
-		mutex_unlock(&pkey_mutex);
-	}
-
 	ipoib_mcast_stop_thread(dev, flush);
 	ipoib_mcast_dev_flush(dev);
 
@@ -924,7 +918,7 @@ int ipoib_ib_dev_init(struct net_device *dev, struct ib_device *ca, int port)
 		    (unsigned long) dev);
 
 	if (dev->flags & IFF_UP) {
-		if (ipoib_ib_dev_open(dev)) {
+		if (ipoib_ib_dev_open(dev, 1)) {
 			ipoib_transport_dev_cleanup(dev);
 			return -ENODEV;
 		}
@@ -966,13 +960,27 @@ static inline int update_parent_pkey(struct ipoib_dev_priv *priv)
 
 	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;
+}
 
 static void __ipoib_ib_dev_flush(struct ipoib_dev_priv *priv,
 				enum ipoib_flush_level level)
 {
 	struct ipoib_dev_priv *cpriv;
 	struct net_device *dev = priv->dev;
-	u16 new_index;
 	int result;
 
 	down_read(&priv->vlan_rwsem);
@@ -986,16 +994,20 @@ static void __ipoib_ib_dev_flush(struct ipoib_dev_priv *priv,
 
 	up_read(&priv->vlan_rwsem);
 
-	if (!test_bit(IPOIB_FLAG_INITIALIZED, &priv->flags)) {
-		/* for non-child devices must check/update the pkey value here */
-		if (level == IPOIB_FLUSH_HEAVY &&
-		    !test_bit(IPOIB_FLAG_SUBINTERFACE, &priv->flags))
-			update_parent_pkey(priv);
+	if (!test_bit(IPOIB_FLAG_INITIALIZED, &priv->flags) &&
+	    level != IPOIB_FLUSH_HEAVY) {
 		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);
+		}
 		ipoib_dbg(priv, "Not flushing - IPOIB_FLAG_ADMIN_UP not set.\n");
 		return;
 	}
@@ -1005,20 +1017,13 @@ static void __ipoib_ib_dev_flush(struct ipoib_dev_priv *priv,
 		 * (parent) devices should always takes what present in pkey index 0
 		 */
 		if (test_bit(IPOIB_FLAG_SUBINTERFACE, &priv->flags)) {
-			if (ib_find_pkey(priv->ca, priv->port, priv->pkey, &new_index)) {
-				clear_bit(IPOIB_PKEY_ASSIGNED, &priv->flags);
-				ipoib_ib_dev_down(dev, 0);
-				ipoib_ib_dev_stop(dev, 0);
-				if (ipoib_pkey_dev_delay_open(dev))
-					return;
-			}
-			/* restart QP only if P_Key index is changed */
-			if (test_and_set_bit(IPOIB_PKEY_ASSIGNED, &priv->flags) &&
-			    new_index == priv->pkey_index) {
+			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;
 			}
-			priv->pkey_index = new_index;
+
 		} else {
 			result = update_parent_pkey(priv);
 			/* restart QP only if P_Key value changed */
@@ -1038,8 +1043,12 @@ static void __ipoib_ib_dev_flush(struct ipoib_dev_priv *priv,
 		ipoib_ib_dev_down(dev, 0);
 
 	if (level == IPOIB_FLUSH_HEAVY) {
-		ipoib_ib_dev_stop(dev, 0);
-		ipoib_ib_dev_open(dev);
+		if (test_bit(IPOIB_FLAG_INITIALIZED, &priv->flags))
+			ipoib_ib_dev_stop(dev, 0);
+		if (ipoib_ib_dev_open(dev, 0) != 0)
+			return;
+		if (netif_queue_stopped(dev))
+			netif_start_queue(dev);
 	}
 
 	/*
@@ -1094,54 +1103,4 @@ void ipoib_ib_dev_cleanup(struct net_device *dev)
 	ipoib_transport_dev_cleanup(dev);
 }
 
-/*
- * Delayed P_Key Assigment Interim Support
- *
- * The following is initial implementation of delayed P_Key assigment
- * mechanism. It is using the same approach implemented for the multicast
- * group join. The single goal of this implementation is to quickly address
- * Bug #2507. This implementation will probably be removed when the P_Key
- * change async notification is available.
- */
-
-void ipoib_pkey_poll(struct work_struct *work)
-{
-	struct ipoib_dev_priv *priv =
-		container_of(work, struct ipoib_dev_priv, pkey_poll_task.work);
-	struct net_device *dev = priv->dev;
-
-	ipoib_pkey_dev_check_presence(dev);
-
-	if (test_bit(IPOIB_PKEY_ASSIGNED, &priv->flags))
-		ipoib_open(dev);
-	else {
-		mutex_lock(&pkey_mutex);
-		if (!test_bit(IPOIB_PKEY_STOP, &priv->flags))
-			queue_delayed_work(ipoib_workqueue,
-					   &priv->pkey_poll_task,
-					   HZ);
-		mutex_unlock(&pkey_mutex);
-	}
-}
-
-int ipoib_pkey_dev_delay_open(struct net_device *dev)
-{
-	struct ipoib_dev_priv *priv = netdev_priv(dev);
-
-	/* Look for the interface pkey value in the IB Port P_Key table and */
-	/* set the interface pkey assigment flag                            */
-	ipoib_pkey_dev_check_presence(dev);
 
-	/* P_Key value not assigned yet - start polling */
-	if (!test_bit(IPOIB_PKEY_ASSIGNED, &priv->flags)) {
-		mutex_lock(&pkey_mutex);
-		clear_bit(IPOIB_PKEY_STOP, &priv->flags);
-		queue_delayed_work(ipoib_workqueue,
-				   &priv->pkey_poll_task,
-				   HZ);
-		mutex_unlock(&pkey_mutex);
-		return 1;
-	}
-
-	return 0;
-}
diff --git a/drivers/infiniband/ulp/ipoib/ipoib_main.c b/drivers/infiniband/ulp/ipoib/ipoib_main.c
index 4e675f4fecc9..58b5aa3b6f2d 100644
--- a/drivers/infiniband/ulp/ipoib/ipoib_main.c
+++ b/drivers/infiniband/ulp/ipoib/ipoib_main.c
@@ -108,11 +108,11 @@ int ipoib_open(struct net_device *dev)
 
 	set_bit(IPOIB_FLAG_ADMIN_UP, &priv->flags);
 
-	if (ipoib_pkey_dev_delay_open(dev))
-		return 0;
-
-	if (ipoib_ib_dev_open(dev))
+	if (ipoib_ib_dev_open(dev, 1)) {
+		if (!test_bit(IPOIB_PKEY_ASSIGNED, &priv->flags))
+			return 0;
 		goto err_disable;
+	}
 
 	if (ipoib_ib_dev_up(dev))
 		goto err_stop;
@@ -716,7 +716,7 @@ static int ipoib_start_xmit(struct sk_buff *skb, struct net_device *dev)
 {
 	struct ipoib_dev_priv *priv = netdev_priv(dev);
 	struct ipoib_neigh *neigh;
-	struct ipoib_cb *cb = (struct ipoib_cb *) skb->cb;
+	struct ipoib_cb *cb = ipoib_skb_cb(skb);
 	struct ipoib_header *header;
 	unsigned long flags;
 
@@ -813,7 +813,7 @@ static int ipoib_hard_header(struct sk_buff *skb,
 			     const void *daddr, const void *saddr, unsigned len)
 {
 	struct ipoib_header *header;
-	struct ipoib_cb *cb = (struct ipoib_cb *) skb->cb;
+	struct ipoib_cb *cb = ipoib_skb_cb(skb);
 
 	header = (struct ipoib_header *) skb_push(skb, sizeof *header);
 
@@ -1364,7 +1364,7 @@ void ipoib_setup(struct net_device *dev)
 	dev->tx_queue_len	 = ipoib_sendq_size * 2;
 	dev->features		 = (NETIF_F_VLAN_CHALLENGED	|
 				    NETIF_F_HIGHDMA);
-	dev->priv_flags		&= ~IFF_XMIT_DST_RELEASE;
+	netif_keep_dst(dev);
 
 	memcpy(dev->broadcast, ipv4_bcast_addr, INFINIBAND_ALEN);
 
@@ -1379,7 +1379,6 @@ void ipoib_setup(struct net_device *dev)
 	INIT_LIST_HEAD(&priv->dead_ahs);
 	INIT_LIST_HEAD(&priv->multicast_list);
 
-	INIT_DELAYED_WORK(&priv->pkey_poll_task, ipoib_pkey_poll);
 	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);
diff --git a/drivers/infiniband/ulp/ipoib/ipoib_multicast.c b/drivers/infiniband/ulp/ipoib/ipoib_multicast.c
index d4e005720d01..ffb83b5f7e80 100644
--- a/drivers/infiniband/ulp/ipoib/ipoib_multicast.c
+++ b/drivers/infiniband/ulp/ipoib/ipoib_multicast.c
@@ -529,21 +529,13 @@ void ipoib_mcast_join_task(struct work_struct *work)
 			  port_attr.state);
 		return;
 	}
+	priv->local_lid = port_attr.lid;
 
 	if (ib_query_gid(priv->ca, priv->port, 0, &priv->local_gid))
 		ipoib_warn(priv, "ib_query_gid() failed\n");
 	else
 		memcpy(priv->dev->dev_addr + 4, priv->local_gid.raw, sizeof (union ib_gid));
 
-	{
-		struct ib_port_attr attr;
-
-		if (!ib_query_port(priv->ca, priv->port, &attr))
-			priv->local_lid = attr.lid;
-		else
-			ipoib_warn(priv, "ib_query_port failed\n");
-	}
-
 	if (!priv->broadcast) {
 		struct ipoib_mcast *broadcast;
 
diff --git a/drivers/infiniband/ulp/iser/iscsi_iser.c b/drivers/infiniband/ulp/iser/iscsi_iser.c
index eb7973957a6e..f42ab14105ac 100644
--- a/drivers/infiniband/ulp/iser/iscsi_iser.c
+++ b/drivers/infiniband/ulp/iser/iscsi_iser.c
@@ -83,7 +83,7 @@ module_param_named(max_lun, iscsi_max_lun, uint, S_IRUGO);
 
 int iser_debug_level = 0;
 bool iser_pi_enable = false;
-int iser_pi_guard = 0;
+int iser_pi_guard = 1;
 
 MODULE_DESCRIPTION("iSER (iSCSI Extensions for RDMA) Datamover");
 MODULE_LICENSE("Dual BSD/GPL");
@@ -97,14 +97,24 @@ module_param_named(pi_enable, iser_pi_enable, bool, 0644);
 MODULE_PARM_DESC(pi_enable, "Enable T10-PI offload support (default:disabled)");
 
 module_param_named(pi_guard, iser_pi_guard, int, 0644);
-MODULE_PARM_DESC(pi_guard, "T10-PI guard_type, 0:CRC|1:IP_CSUM (default:CRC)");
+MODULE_PARM_DESC(pi_guard, "T10-PI guard_type, 0:CRC|1:IP_CSUM (default:IP_CSUM)");
 
 static struct workqueue_struct *release_wq;
 struct iser_global ig;
 
+/*
+ * iscsi_iser_recv() - Process a successfull recv completion
+ * @conn:         iscsi connection
+ * @hdr:          iscsi header
+ * @rx_data:      buffer containing receive data payload
+ * @rx_data_len:  length of rx_data
+ *
+ * Notes: In case of data length errors or iscsi PDU completion failures
+ *        this routine will signal iscsi layer of connection failure.
+ */
 void
-iscsi_iser_recv(struct iscsi_conn *conn,
-		struct iscsi_hdr *hdr, char *rx_data, int rx_data_len)
+iscsi_iser_recv(struct iscsi_conn *conn, struct iscsi_hdr *hdr,
+		char *rx_data, int rx_data_len)
 {
 	int rc = 0;
 	int datalen;
@@ -135,20 +145,30 @@ error:
 	iscsi_conn_failure(conn, rc);
 }
 
-static int iscsi_iser_pdu_alloc(struct iscsi_task *task, uint8_t opcode)
+/**
+ * iscsi_iser_pdu_alloc() - allocate an iscsi-iser PDU
+ * @task:     iscsi task
+ * @opcode:   iscsi command opcode
+ *
+ * Netes: This routine can't fail, just assign iscsi task
+ *        hdr and max hdr size.
+ */
+static int
+iscsi_iser_pdu_alloc(struct iscsi_task *task, uint8_t opcode)
 {
 	struct iscsi_iser_task *iser_task = task->dd_data;
 
 	task->hdr = (struct iscsi_hdr *)&iser_task->desc.iscsi_header;
 	task->hdr_max = sizeof(iser_task->desc.iscsi_header);
+
 	return 0;
 }
 
 int iser_initialize_task_headers(struct iscsi_task *task,
 						struct iser_tx_desc *tx_desc)
 {
-	struct iser_conn       *ib_conn   = task->conn->dd_data;
-	struct iser_device     *device    = ib_conn->device;
+	struct iser_conn       *iser_conn   = task->conn->dd_data;
+	struct iser_device *device = iser_conn->ib_conn.device;
 	struct iscsi_iser_task *iser_task = task->dd_data;
 	u64 dma_addr;
 
@@ -162,14 +182,18 @@ int iser_initialize_task_headers(struct iscsi_task *task,
 	tx_desc->tx_sg[0].length = ISER_HEADERS_LEN;
 	tx_desc->tx_sg[0].lkey   = device->mr->lkey;
 
-	iser_task->ib_conn = ib_conn;
+	iser_task->iser_conn = iser_conn;
 	return 0;
 }
+
 /**
- * iscsi_iser_task_init - Initialize task
+ * iscsi_iser_task_init() - Initialize iscsi-iser task
  * @task: iscsi task
  *
  * Initialize the task for the scsi command or mgmt command.
+ *
+ * Return: Returns zero on success or -ENOMEM when failing
+ *         to init task headers (dma mapping error).
  */
 static int
 iscsi_iser_task_init(struct iscsi_task *task)
@@ -191,7 +215,7 @@ iscsi_iser_task_init(struct iscsi_task *task)
 }
 
 /**
- * iscsi_iser_mtask_xmit - xmit management(immediate) task
+ * iscsi_iser_mtask_xmit() - xmit management (immediate) task
  * @conn: iscsi connection
  * @task: task management task
  *
@@ -249,6 +273,12 @@ iscsi_iser_task_xmit_unsol_data_exit:
 	return error;
 }
 
+/**
+ * iscsi_iser_task_xmit() - xmit iscsi-iser task
+ * @task: iscsi task
+ *
+ * Return: zero on success or escalates $error on failure.
+ */
 static int
 iscsi_iser_task_xmit(struct iscsi_task *task)
 {
@@ -286,12 +316,24 @@ iscsi_iser_task_xmit(struct iscsi_task *task)
 	return error;
 }
 
+/**
+ * iscsi_iser_cleanup_task() - cleanup an iscsi-iser task
+ * @task: iscsi task
+ *
+ * Notes: In case the RDMA device is already NULL (might have
+ *        been removed in DEVICE_REMOVAL CM event it will bail-out
+ *        without doing dma unmapping.
+ */
 static void iscsi_iser_cleanup_task(struct iscsi_task *task)
 {
 	struct iscsi_iser_task *iser_task = task->dd_data;
 	struct iser_tx_desc    *tx_desc   = &iser_task->desc;
-	struct iser_conn       *ib_conn	  = task->conn->dd_data;
-	struct iser_device     *device	  = ib_conn->device;
+	struct iser_conn       *iser_conn	  = task->conn->dd_data;
+	struct iser_device *device = iser_conn->ib_conn.device;
+
+	/* DEVICE_REMOVAL event might have already released the device */
+	if (!device)
+		return;
 
 	ib_dma_unmap_single(device->ib_device,
 		tx_desc->dma_addr, ISER_HEADERS_LEN, DMA_TO_DEVICE);
@@ -306,7 +348,20 @@ static void iscsi_iser_cleanup_task(struct iscsi_task *task)
 	}
 }
 
-static u8 iscsi_iser_check_protection(struct iscsi_task *task, sector_t *sector)
+/**
+ * iscsi_iser_check_protection() - check protection information status of task.
+ * @task:     iscsi task
+ * @sector:   error sector if exsists (output)
+ *
+ * Return: zero if no data-integrity errors have occured
+ *         0x1: data-integrity error occured in the guard-block
+ *         0x2: data-integrity error occured in the reference tag
+ *         0x3: data-integrity error occured in the application tag
+ *
+ *         In addition the error sector is marked.
+ */
+static u8
+iscsi_iser_check_protection(struct iscsi_task *task, sector_t *sector)
 {
 	struct iscsi_iser_task *iser_task = task->dd_data;
 
@@ -318,8 +373,17 @@ static u8 iscsi_iser_check_protection(struct iscsi_task *task, sector_t *sector)
 						 sector);
 }
 
+/**
+ * iscsi_iser_conn_create() - create a new iscsi-iser connection
+ * @cls_session: iscsi class connection
+ * @conn_idx:    connection index within the session (for MCS)
+ *
+ * Return: iscsi_cls_conn when iscsi_conn_setup succeeds or NULL
+ *         otherwise.
+ */
 static struct iscsi_cls_conn *
-iscsi_iser_conn_create(struct iscsi_cls_session *cls_session, uint32_t conn_idx)
+iscsi_iser_conn_create(struct iscsi_cls_session *cls_session,
+		       uint32_t conn_idx)
 {
 	struct iscsi_conn *conn;
 	struct iscsi_cls_conn *cls_conn;
@@ -338,14 +402,25 @@ iscsi_iser_conn_create(struct iscsi_cls_session *cls_session, uint32_t conn_idx)
 	return cls_conn;
 }
 
+/**
+ * iscsi_iser_conn_bind() - bind iscsi and iser connection structures
+ * @cls_session:     iscsi class session
+ * @cls_conn:        iscsi class connection
+ * @transport_eph:   transport end-point handle
+ * @is_leading:      indicate if this is the session leading connection (MCS)
+ *
+ * Return: zero on success, $error if iscsi_conn_bind fails and
+ *         -EINVAL in case end-point doesn't exsits anymore or iser connection
+ *         state is not UP (teardown already started).
+ */
 static int
 iscsi_iser_conn_bind(struct iscsi_cls_session *cls_session,
-		     struct iscsi_cls_conn *cls_conn, uint64_t transport_eph,
+		     struct iscsi_cls_conn *cls_conn,
+		     uint64_t transport_eph,
 		     int is_leading)
 {
 	struct iscsi_conn *conn = cls_conn->dd_data;
-	struct iscsi_session *session;
-	struct iser_conn *ib_conn;
+	struct iser_conn *iser_conn;
 	struct iscsi_endpoint *ep;
 	int error;
 
@@ -361,56 +436,100 @@ iscsi_iser_conn_bind(struct iscsi_cls_session *cls_session,
 			 (unsigned long long)transport_eph);
 		return -EINVAL;
 	}
-	ib_conn = ep->dd_data;
+	iser_conn = ep->dd_data;
+
+	mutex_lock(&iser_conn->state_mutex);
+	if (iser_conn->state != ISER_CONN_UP) {
+		error = -EINVAL;
+		iser_err("iser_conn %p state is %d, teardown started\n",
+			 iser_conn, iser_conn->state);
+		goto out;
+	}
 
-	session = conn->session;
-	if (iser_alloc_rx_descriptors(ib_conn, session))
-		return -ENOMEM;
+	error = iser_alloc_rx_descriptors(iser_conn, conn->session);
+	if (error)
+		goto out;
 
 	/* binds the iSER connection retrieved from the previously
 	 * connected ep_handle to the iSCSI layer connection. exchanges
 	 * connection pointers */
-	iser_info("binding iscsi conn %p to ib_conn %p\n", conn, ib_conn);
+	iser_info("binding iscsi conn %p to iser_conn %p\n", conn, iser_conn);
 
-	conn->dd_data = ib_conn;
-	ib_conn->iscsi_conn = conn;
+	conn->dd_data = iser_conn;
+	iser_conn->iscsi_conn = conn;
 
-	return 0;
+out:
+	mutex_unlock(&iser_conn->state_mutex);
+	return error;
 }
 
+/**
+ * iscsi_iser_conn_start() - start iscsi-iser connection
+ * @cls_conn: iscsi class connection
+ *
+ * Notes: Here iser intialize (or re-initialize) stop_completion as
+ *        from this point iscsi must call conn_stop in session/connection
+ *        teardown so iser transport must wait for it.
+ */
 static int
 iscsi_iser_conn_start(struct iscsi_cls_conn *cls_conn)
 {
 	struct iscsi_conn *iscsi_conn;
-	struct iser_conn *ib_conn;
+	struct iser_conn *iser_conn;
 
 	iscsi_conn = cls_conn->dd_data;
-	ib_conn = iscsi_conn->dd_data;
-	reinit_completion(&ib_conn->stop_completion);
+	iser_conn = iscsi_conn->dd_data;
+	reinit_completion(&iser_conn->stop_completion);
 
 	return iscsi_conn_start(cls_conn);
 }
 
+/**
+ * iscsi_iser_conn_stop() - stop iscsi-iser connection
+ * @cls_conn:  iscsi class connection
+ * @flag:      indicate if recover or terminate (passed as is)
+ *
+ * Notes: Calling iscsi_conn_stop might theoretically race with
+ *        DEVICE_REMOVAL event and dereference a previously freed RDMA device
+ *        handle, so we call it under iser the state lock to protect against
+ *        this kind of race.
+ */
 static void
 iscsi_iser_conn_stop(struct iscsi_cls_conn *cls_conn, int flag)
 {
 	struct iscsi_conn *conn = cls_conn->dd_data;
-	struct iser_conn *ib_conn = conn->dd_data;
+	struct iser_conn *iser_conn = conn->dd_data;
 
-	iser_dbg("stopping iscsi_conn: %p, ib_conn: %p\n", conn, ib_conn);
-	iscsi_conn_stop(cls_conn, flag);
+	iser_info("stopping iscsi_conn: %p, iser_conn: %p\n", conn, iser_conn);
 
 	/*
 	 * Userspace may have goofed up and not bound the connection or
 	 * might have only partially setup the connection.
 	 */
-	if (ib_conn) {
+	if (iser_conn) {
+		mutex_lock(&iser_conn->state_mutex);
+		iscsi_conn_stop(cls_conn, flag);
+		iser_conn_terminate(iser_conn);
+
+		/* unbind */
+		iser_conn->iscsi_conn = NULL;
 		conn->dd_data = NULL;
-		complete(&ib_conn->stop_completion);
+
+		complete(&iser_conn->stop_completion);
+		mutex_unlock(&iser_conn->state_mutex);
+	} else {
+		iscsi_conn_stop(cls_conn, flag);
 	}
 }
 
-static void iscsi_iser_session_destroy(struct iscsi_cls_session *cls_session)
+/**
+ * iscsi_iser_session_destroy() - destroy iscsi-iser session
+ * @cls_session: iscsi class session
+ *
+ * Removes and free iscsi host.
+ */
+static void
+iscsi_iser_session_destroy(struct iscsi_cls_session *cls_session)
 {
 	struct Scsi_Host *shost = iscsi_session_to_shost(cls_session);
 
@@ -430,6 +549,16 @@ iser_dif_prot_caps(int prot_caps)
 						      SHOST_DIX_TYPE3_PROTECTION : 0);
 }
 
+/**
+ * iscsi_iser_session_create() - create an iscsi-iser session
+ * @ep:             iscsi end-point handle
+ * @cmds_max:       maximum commands in this session
+ * @qdepth:         session command queue depth
+ * @initial_cmdsn:  initiator command sequnce number
+ *
+ * Allocates and adds a scsi host, expose DIF supprot if
+ * exists, and sets up an iscsi session.
+ */
 static struct iscsi_cls_session *
 iscsi_iser_session_create(struct iscsi_endpoint *ep,
 			  uint16_t cmds_max, uint16_t qdepth,
@@ -438,7 +567,8 @@ iscsi_iser_session_create(struct iscsi_endpoint *ep,
 	struct iscsi_cls_session *cls_session;
 	struct iscsi_session *session;
 	struct Scsi_Host *shost;
-	struct iser_conn *ib_conn = NULL;
+	struct iser_conn *iser_conn = NULL;
+	struct ib_conn *ib_conn;
 
 	shost = iscsi_host_alloc(&iscsi_iser_sht, 0, 0);
 	if (!shost)
@@ -455,7 +585,8 @@ iscsi_iser_session_create(struct iscsi_endpoint *ep,
 	 * the leading conn's ep so this will be NULL;
 	 */
 	if (ep) {
-		ib_conn = ep->dd_data;
+		iser_conn = ep->dd_data;
+		ib_conn = &iser_conn->ib_conn;
 		if (ib_conn->pi_support) {
 			u32 sig_caps = ib_conn->device->dev_attr.sig_prot_cap;
 
@@ -467,8 +598,8 @@ iscsi_iser_session_create(struct iscsi_endpoint *ep,
 		}
 	}
 
-	if (iscsi_host_add(shost,
-			   ep ? ib_conn->device->ib_device->dma_device : NULL))
+	if (iscsi_host_add(shost, ep ?
+			   ib_conn->device->ib_device->dma_device : NULL))
 		goto free_host;
 
 	if (cmds_max > ISER_DEF_XMIT_CMDS_MAX) {
@@ -540,6 +671,13 @@ iscsi_iser_set_param(struct iscsi_cls_conn *cls_conn,
 	return 0;
 }
 
+/**
+ * iscsi_iser_set_param() - set class connection parameter
+ * @cls_conn:    iscsi class connection
+ * @stats:       iscsi stats to output
+ *
+ * Output connection statistics.
+ */
 static void
 iscsi_iser_conn_get_stats(struct iscsi_cls_conn *cls_conn, struct iscsi_stats *stats)
 {
@@ -568,18 +706,18 @@ iscsi_iser_conn_get_stats(struct iscsi_cls_conn *cls_conn, struct iscsi_stats *s
 static int iscsi_iser_get_ep_param(struct iscsi_endpoint *ep,
 				   enum iscsi_param param, char *buf)
 {
-	struct iser_conn *ib_conn = ep->dd_data;
+	struct iser_conn *iser_conn = ep->dd_data;
 	int len;
 
 	switch (param) {
 	case ISCSI_PARAM_CONN_PORT:
 	case ISCSI_PARAM_CONN_ADDRESS:
-		if (!ib_conn || !ib_conn->cma_id)
+		if (!iser_conn || !iser_conn->ib_conn.cma_id)
 			return -ENOTCONN;
 
 		return iscsi_conn_get_addr_param((struct sockaddr_storage *)
-					&ib_conn->cma_id->route.addr.dst_addr,
-					param, buf);
+				&iser_conn->ib_conn.cma_id->route.addr.dst_addr,
+				param, buf);
 		break;
 	default:
 		return -ENOSYS;
@@ -588,48 +726,85 @@ static int iscsi_iser_get_ep_param(struct iscsi_endpoint *ep,
 	return len;
 }
 
+/**
+ * iscsi_iser_ep_connect() - Initiate iSER connection establishment
+ * @shost:          scsi_host
+ * @dst_addr:       destination address
+ * @non-blocking:   indicate if routine can block
+ *
+ * Allocate an iscsi endpoint, an iser_conn structure and bind them.
+ * After that start RDMA connection establishment via rdma_cm. We
+ * don't allocate iser_conn embedded in iscsi_endpoint since in teardown
+ * the endpoint will be destroyed at ep_disconnect while iser_conn will
+ * cleanup its resources asynchronuously.
+ *
+ * Return: iscsi_endpoint created by iscsi layer or ERR_PTR(error)
+ *         if fails.
+ */
 static struct iscsi_endpoint *
 iscsi_iser_ep_connect(struct Scsi_Host *shost, struct sockaddr *dst_addr,
 		      int non_blocking)
 {
 	int err;
-	struct iser_conn *ib_conn;
+	struct iser_conn *iser_conn;
 	struct iscsi_endpoint *ep;
 
-	ep = iscsi_create_endpoint(sizeof(*ib_conn));
+	ep = iscsi_create_endpoint(0);
 	if (!ep)
 		return ERR_PTR(-ENOMEM);
 
-	ib_conn = ep->dd_data;
-	ib_conn->ep = ep;
-	iser_conn_init(ib_conn);
+	iser_conn = kzalloc(sizeof(*iser_conn), GFP_KERNEL);
+	if (!iser_conn) {
+		err = -ENOMEM;
+		goto failure;
+	}
+
+	ep->dd_data = iser_conn;
+	iser_conn->ep = ep;
+	iser_conn_init(iser_conn);
 
-	err = iser_connect(ib_conn, NULL, (struct sockaddr_in *)dst_addr,
-			   non_blocking);
+	err = iser_connect(iser_conn, NULL, dst_addr, non_blocking);
 	if (err)
-		return ERR_PTR(err);
+		goto failure;
 
 	return ep;
+failure:
+	iscsi_destroy_endpoint(ep);
+	return ERR_PTR(err);
 }
 
+/**
+ * iscsi_iser_ep_poll() - poll for iser connection establishment to complete
+ * @ep:            iscsi endpoint (created at ep_connect)
+ * @timeout_ms:    polling timeout allowed in ms.
+ *
+ * This routine boils down to waiting for up_completion signaling
+ * that cma_id got CONNECTED event.
+ *
+ * Return: 1 if succeeded in connection establishment, 0 if timeout expired
+ *         (libiscsi will retry will kick in) or -1 if interrupted by signal
+ *         or more likely iser connection state transitioned to TEMINATING or
+ *         DOWN during the wait period.
+ */
 static int
 iscsi_iser_ep_poll(struct iscsi_endpoint *ep, int timeout_ms)
 {
-	struct iser_conn *ib_conn;
+	struct iser_conn *iser_conn;
 	int rc;
 
-	ib_conn = ep->dd_data;
-	rc = wait_event_interruptible_timeout(ib_conn->wait,
-			     ib_conn->state == ISER_CONN_UP,
-			     msecs_to_jiffies(timeout_ms));
-
+	iser_conn = ep->dd_data;
+	rc = wait_for_completion_interruptible_timeout(&iser_conn->up_completion,
+						       msecs_to_jiffies(timeout_ms));
 	/* if conn establishment failed, return error code to iscsi */
-	if (!rc &&
-	    (ib_conn->state == ISER_CONN_TERMINATING ||
-	     ib_conn->state == ISER_CONN_DOWN))
-		rc = -1;
+	if (rc == 0) {
+		mutex_lock(&iser_conn->state_mutex);
+		if (iser_conn->state == ISER_CONN_TERMINATING ||
+		    iser_conn->state == ISER_CONN_DOWN)
+			rc = -1;
+		mutex_unlock(&iser_conn->state_mutex);
+	}
 
-	iser_info("ib conn %p rc = %d\n", ib_conn, rc);
+	iser_info("ib conn %p rc = %d\n", iser_conn, rc);
 
 	if (rc > 0)
 		return 1; /* success, this is the equivalent of POLLOUT */
@@ -639,26 +814,43 @@ iscsi_iser_ep_poll(struct iscsi_endpoint *ep, int timeout_ms)
 		return rc; /* signal */
 }
 
+/**
+ * iscsi_iser_ep_disconnect() - Initiate connection teardown process
+ * @ep:    iscsi endpoint handle
+ *
+ * This routine is not blocked by iser and RDMA termination process
+ * completion as we queue a deffered work for iser/RDMA destruction
+ * and cleanup or actually call it immediately in case we didn't pass
+ * iscsi conn bind/start stage, thus it is safe.
+ */
 static void
 iscsi_iser_ep_disconnect(struct iscsi_endpoint *ep)
 {
-	struct iser_conn *ib_conn;
+	struct iser_conn *iser_conn;
+
+	iser_conn = ep->dd_data;
+	iser_info("ep %p iser conn %p state %d\n",
+		  ep, iser_conn, iser_conn->state);
 
-	ib_conn = ep->dd_data;
-	iser_info("ep %p ib conn %p state %d\n", ep, ib_conn, ib_conn->state);
-	iser_conn_terminate(ib_conn);
+	mutex_lock(&iser_conn->state_mutex);
+	iser_conn_terminate(iser_conn);
 
 	/*
-	 * if iser_conn and iscsi_conn are bound, we must wait iscsi_conn_stop
-	 * call and ISER_CONN_DOWN state before freeing the iser resources.
-	 * otherwise we are safe to free resources immediately.
+	 * if iser_conn and iscsi_conn are bound, we must wait for
+	 * iscsi_conn_stop and flush errors completion before freeing
+	 * the iser resources. Otherwise we are safe to free resources
+	 * immediately.
 	 */
-	if (ib_conn->iscsi_conn) {
-		INIT_WORK(&ib_conn->release_work, iser_release_work);
-		queue_work(release_wq, &ib_conn->release_work);
+	if (iser_conn->iscsi_conn) {
+		INIT_WORK(&iser_conn->release_work, iser_release_work);
+		queue_work(release_wq, &iser_conn->release_work);
+		mutex_unlock(&iser_conn->state_mutex);
 	} else {
-		iser_conn_release(ib_conn);
+		iser_conn->state = ISER_CONN_DOWN;
+		mutex_unlock(&iser_conn->state_mutex);
+		iser_conn_release(iser_conn);
 	}
+	iscsi_destroy_endpoint(ep);
 }
 
 static umode_t iser_attr_is_visible(int param_type, int param)
@@ -819,7 +1011,7 @@ register_transport_failure:
 
 static void __exit iser_exit(void)
 {
-	struct iser_conn *ib_conn, *n;
+	struct iser_conn *iser_conn, *n;
 	int connlist_empty;
 
 	iser_dbg("Removing iSER datamover...\n");
@@ -832,8 +1024,9 @@ static void __exit iser_exit(void)
 	if (!connlist_empty) {
 		iser_err("Error cleanup stage completed but we still have iser "
 			 "connections, destroying them anyway.\n");
-		list_for_each_entry_safe(ib_conn, n, &ig.connlist, conn_list) {
-			iser_conn_release(ib_conn);
+		list_for_each_entry_safe(iser_conn, n, &ig.connlist,
+					 conn_list) {
+			iser_conn_release(iser_conn);
 		}
 	}
 
diff --git a/drivers/infiniband/ulp/iser/iscsi_iser.h b/drivers/infiniband/ulp/iser/iscsi_iser.h
index 97cd385bf7f7..cd4174ca9a76 100644
--- a/drivers/infiniband/ulp/iser/iscsi_iser.h
+++ b/drivers/infiniband/ulp/iser/iscsi_iser.h
@@ -69,39 +69,38 @@
 
 #define DRV_NAME	"iser"
 #define PFX		DRV_NAME ": "
-#define DRV_VER		"1.4"
+#define DRV_VER		"1.4.8"
 
-#define iser_dbg(fmt, arg...)				\
-	do {						\
-		if (iser_debug_level > 2)		\
-			printk(KERN_DEBUG PFX "%s:" fmt,\
-				__func__ , ## arg);	\
+#define iser_dbg(fmt, arg...)				 \
+	do {						 \
+		if (iser_debug_level > 2)		 \
+			printk(KERN_DEBUG PFX "%s: " fmt,\
+				__func__ , ## arg);	 \
 	} while (0)
 
 #define iser_warn(fmt, arg...)				\
 	do {						\
 		if (iser_debug_level > 0)		\
-			pr_warn(PFX "%s:" fmt,          \
+			pr_warn(PFX "%s: " fmt,		\
 				__func__ , ## arg);	\
 	} while (0)
 
 #define iser_info(fmt, arg...)				\
 	do {						\
 		if (iser_debug_level > 1)		\
-			pr_info(PFX "%s:" fmt,          \
+			pr_info(PFX "%s: " fmt,		\
 				__func__ , ## arg);	\
 	} while (0)
 
 #define iser_err(fmt, arg...)				\
 	do {						\
-		printk(KERN_ERR PFX "%s:" fmt,          \
+		printk(KERN_ERR PFX "%s: " fmt,		\
 		       __func__ , ## arg);		\
 	} while (0)
 
 #define SHIFT_4K	12
 #define SIZE_4K	(1ULL << SHIFT_4K)
 #define MASK_4K	(~(SIZE_4K-1))
-
 					/* support up to 512KB in one RDMA */
 #define ISCSI_ISER_SG_TABLESIZE         (0x80000 >> SHIFT_4K)
 #define ISER_DEF_XMIT_CMDS_DEFAULT		512
@@ -145,18 +144,32 @@
 					ISER_MAX_TX_MISC_PDUS         + \
 					ISER_MAX_RX_MISC_PDUS)
 
+#define ISER_WC_BATCH_COUNT   16
+#define ISER_SIGNAL_CMD_COUNT 32
+
 #define ISER_VER			0x10
 #define ISER_WSV			0x08
 #define ISER_RSV			0x04
 
 #define ISER_FASTREG_LI_WRID		0xffffffffffffffffULL
+#define ISER_BEACON_WRID		0xfffffffffffffffeULL
 
+/**
+ * struct iser_hdr - iSER header
+ *
+ * @flags:        flags support (zbva, remote_inv)
+ * @rsvd:         reserved
+ * @write_stag:   write rkey
+ * @write_va:     write virtual address
+ * @reaf_stag:    read rkey
+ * @read_va:      read virtual address
+ */
 struct iser_hdr {
 	u8      flags;
 	u8      rsvd[3];
-	__be32  write_stag; /* write rkey */
+	__be32  write_stag;
 	__be64  write_va;
-	__be32  read_stag;  /* read rkey */
+	__be32  read_stag;
 	__be64  read_va;
 } __attribute__((packed));
 
@@ -179,7 +192,7 @@ struct iser_cm_hdr {
 /* Length of an object name string */
 #define ISER_OBJECT_NAME_SIZE		    64
 
-enum iser_ib_conn_state {
+enum iser_conn_state {
 	ISER_CONN_INIT,		   /* descriptor allocd, no conn          */
 	ISER_CONN_PENDING,	   /* in the process of being established */
 	ISER_CONN_UP,		   /* up and running                      */
@@ -200,23 +213,42 @@ enum iser_data_dir {
 	ISER_DIRS_NUM
 };
 
+/**
+ * struct iser_data_buf - iSER data buffer
+ *
+ * @buf:          pointer to the sg list
+ * @size:         num entries of this sg
+ * @data_len:     total beffer byte len
+ * @dma_nents:    returned by dma_map_sg
+ * @copy_buf:     allocated copy buf for SGs unaligned
+ *                for rdma which are copied
+ * @sg_single:    SG-ified clone of a non SG SC or
+ *                unaligned SG
+ */
 struct iser_data_buf {
-	void               *buf;      /* pointer to the sg list               */
-	unsigned int       size;      /* num entries of this sg               */
-	unsigned long      data_len;  /* total data len                       */
-	unsigned int       dma_nents; /* returned by dma_map_sg               */
-	char       	   *copy_buf; /* allocated copy buf for SGs unaligned *
-	                               * for rdma which are copied            */
-	struct scatterlist sg_single; /* SG-ified clone of a non SG SC or     *
-				       * unaligned SG                         */
+	void               *buf;
+	unsigned int       size;
+	unsigned long      data_len;
+	unsigned int       dma_nents;
+	char               *copy_buf;
+	struct scatterlist sg_single;
   };
 
 /* fwd declarations */
 struct iser_device;
-struct iser_cq_desc;
 struct iscsi_iser_task;
 struct iscsi_endpoint;
 
+/**
+ * struct iser_mem_reg - iSER memory registration info
+ *
+ * @lkey:         MR local key
+ * @rkey:         MR remote key
+ * @va:           MR start address (buffer va)
+ * @len:          MR length
+ * @mem_h:        pointer to registration context (FMR/Fastreg)
+ * @is_mr:        indicates weather we registered the buffer
+ */
 struct iser_mem_reg {
 	u32  lkey;
 	u32  rkey;
@@ -226,11 +258,20 @@ struct iser_mem_reg {
 	int  is_mr;
 };
 
+/**
+ * struct iser_regd_buf - iSER buffer registration desc
+ *
+ * @reg:          memory registration info
+ * @virt_addr:    virtual address of buffer
+ * @device:       reference to iser device
+ * @direction:    dma direction (for dma_unmap)
+ * @data_size:    data buffer size in bytes
+ */
 struct iser_regd_buf {
-	struct iser_mem_reg     reg;        /* memory registration info        */
+	struct iser_mem_reg     reg;
 	void                    *virt_addr;
-	struct iser_device      *device;    /* device->device for dma_unmap    */
-	enum dma_data_direction direction;  /* direction for dma_unmap	       */
+	struct iser_device      *device;
+	enum dma_data_direction direction;
 	unsigned int            data_size;
 };
 
@@ -240,19 +281,39 @@ enum iser_desc_type {
 	ISCSI_TX_DATAOUT
 };
 
+/**
+ * struct iser_tx_desc - iSER TX descriptor (for send wr_id)
+ *
+ * @iser_header:   iser header
+ * @iscsi_header:  iscsi header
+ * @type:          command/control/dataout
+ * @dam_addr:      header buffer dma_address
+ * @tx_sg:         sg[0] points to iser/iscsi headers
+ *                 sg[1] optionally points to either of immediate data
+ *                 unsolicited data-out or control
+ * @num_sge:       number sges used on this TX task
+ */
 struct iser_tx_desc {
 	struct iser_hdr              iser_header;
 	struct iscsi_hdr             iscsi_header;
 	enum   iser_desc_type        type;
 	u64		             dma_addr;
-	/* sg[0] points to iser/iscsi headers, sg[1] optionally points to either
-	of immediate data, unsolicited data-out or control (login,text) */
 	struct ib_sge		     tx_sg[2];
 	int                          num_sge;
 };
 
 #define ISER_RX_PAD_SIZE	(256 - (ISER_RX_PAYLOAD_SIZE + \
 					sizeof(u64) + sizeof(struct ib_sge)))
+/**
+ * struct iser_rx_desc - iSER RX descriptor (for recv wr_id)
+ *
+ * @iser_header:   iser header
+ * @iscsi_header:  iscsi header
+ * @data:          received data segment
+ * @dma_addr:      receive buffer dma address
+ * @rx_sg:         ib_sge of receive buffer
+ * @pad:           for sense data TODO: Modify to maximum sense length supported
+ */
 struct iser_rx_desc {
 	struct iser_hdr              iser_header;
 	struct iscsi_hdr             iscsi_header;
@@ -265,25 +326,59 @@ struct iser_rx_desc {
 #define ISER_MAX_CQ 4
 
 struct iser_conn;
+struct ib_conn;
 struct iscsi_iser_task;
 
+/**
+ * struct iser_comp - iSER completion context
+ *
+ * @device:     pointer to device handle
+ * @cq:         completion queue
+ * @wcs:        work completion array
+ * @tasklet:    Tasklet handle
+ * @active_qps: Number of active QPs attached
+ *              to completion context
+ */
+struct iser_comp {
+	struct iser_device      *device;
+	struct ib_cq		*cq;
+	struct ib_wc		 wcs[ISER_WC_BATCH_COUNT];
+	struct tasklet_struct	 tasklet;
+	int                      active_qps;
+};
+
+/**
+ * struct iser_device - iSER device handle
+ *
+ * @ib_device:     RDMA device
+ * @pd:            Protection Domain for this device
+ * @dev_attr:      Device attributes container
+ * @mr:            Global DMA memory region
+ * @event_handler: IB events handle routine
+ * @ig_list:	   entry in devices list
+ * @refcount:      Reference counter, dominated by open iser connections
+ * @comps_used:    Number of completion contexts used, Min between online
+ *                 cpus and device max completion vectors
+ * @comps:         Dinamically allocated array of completion handlers
+ * Memory registration pool Function pointers (FMR or Fastreg):
+ *     @iser_alloc_rdma_reg_res: Allocation of memory regions pool
+ *     @iser_free_rdma_reg_res:  Free of memory regions pool
+ *     @iser_reg_rdma_mem:       Memory registration routine
+ *     @iser_unreg_rdma_mem:     Memory deregistration routine
+ */
 struct iser_device {
 	struct ib_device             *ib_device;
 	struct ib_pd	             *pd;
 	struct ib_device_attr	     dev_attr;
-	struct ib_cq	             *rx_cq[ISER_MAX_CQ];
-	struct ib_cq	             *tx_cq[ISER_MAX_CQ];
 	struct ib_mr	             *mr;
-	struct tasklet_struct	     cq_tasklet[ISER_MAX_CQ];
 	struct ib_event_handler      event_handler;
-	struct list_head             ig_list; /* entry in ig devices list */
+	struct list_head             ig_list;
 	int                          refcount;
-	int                          cq_active_qps[ISER_MAX_CQ];
-	int			     cqs_used;
-	struct iser_cq_desc	     *cq_desc;
-	int                          (*iser_alloc_rdma_reg_res)(struct iser_conn *ib_conn,
+	int			     comps_used;
+	struct iser_comp	     comps[ISER_MAX_CQ];
+	int                          (*iser_alloc_rdma_reg_res)(struct ib_conn *ib_conn,
 								unsigned cmds_max);
-	void                         (*iser_free_rdma_reg_res)(struct iser_conn *ib_conn);
+	void                         (*iser_free_rdma_reg_res)(struct ib_conn *ib_conn);
 	int                          (*iser_reg_rdma_mem)(struct iscsi_iser_task *iser_task,
 							  enum iser_data_dir cmd_dir);
 	void                         (*iser_unreg_rdma_mem)(struct iscsi_iser_task *iser_task,
@@ -301,76 +396,160 @@ enum iser_reg_indicator {
 	ISER_FASTREG_PROTECTED	= 1 << 3,
 };
 
+/**
+ * struct iser_pi_context - Protection information context
+ *
+ * @prot_mr:        protection memory region
+ * @prot_frpl:      protection fastreg page list
+ * @sig_mr:         signature feature enabled memory region
+ */
 struct iser_pi_context {
 	struct ib_mr                   *prot_mr;
 	struct ib_fast_reg_page_list   *prot_frpl;
 	struct ib_mr                   *sig_mr;
 };
 
+/**
+ * struct fast_reg_descriptor - Fast registration descriptor
+ *
+ * @list:           entry in connection fastreg pool
+ * @data_mr:        data memory region
+ * @data_frpl:      data fastreg page list
+ * @pi_ctx:         protection information context
+ * @reg_indicators: fast registration indicators
+ */
 struct fast_reg_descriptor {
 	struct list_head		  list;
-	/* For fast registration - FRWR */
 	struct ib_mr			 *data_mr;
 	struct ib_fast_reg_page_list     *data_frpl;
 	struct iser_pi_context		 *pi_ctx;
-	/* registration indicators container */
 	u8				  reg_indicators;
 };
 
+/**
+ * struct ib_conn - Infiniband related objects
+ *
+ * @cma_id:              rdma_cm connection maneger handle
+ * @qp:                  Connection Queue-pair
+ * @post_recv_buf_count: post receive counter
+ * @rx_wr:               receive work request for batch posts
+ * @device:              reference to iser device
+ * @comp:                iser completion context
+ * @pi_support:          Indicate device T10-PI support
+ * @beacon:              beacon send wr to signal all flush errors were drained
+ * @flush_comp:          completes when all connection completions consumed
+ * @lock:                protects fmr/fastreg pool
+ * @union.fmr:
+ *     @pool:            FMR pool for fast registrations
+ *     @page_vec:        page vector to hold mapped commands pages
+ *                       used for registration
+ * @union.fastreg:
+ *     @pool:            Fast registration descriptors pool for fast
+ *                       registrations
+ *     @pool_size:       Size of pool
+ */
+struct ib_conn {
+	struct rdma_cm_id           *cma_id;
+	struct ib_qp	            *qp;
+	int                          post_recv_buf_count;
+	struct ib_recv_wr	     rx_wr[ISER_MIN_POSTED_RX];
+	struct iser_device          *device;
+	struct iser_comp	    *comp;
+	bool			     pi_support;
+	struct ib_send_wr	     beacon;
+	struct completion	     flush_comp;
+	spinlock_t		     lock;
+	union {
+		struct {
+			struct ib_fmr_pool      *pool;
+			struct iser_page_vec	*page_vec;
+		} fmr;
+		struct {
+			struct list_head	 pool;
+			int			 pool_size;
+		} fastreg;
+	};
+};
+
+/**
+ * struct iser_conn - iSER connection context
+ *
+ * @ib_conn:          connection RDMA resources
+ * @iscsi_conn:       link to matching iscsi connection
+ * @ep:               transport handle
+ * @state:            connection logical state
+ * @qp_max_recv_dtos: maximum number of data outs, corresponds
+ *                    to max number of post recvs
+ * @qp_max_recv_dtos_mask: (qp_max_recv_dtos - 1)
+ * @min_posted_rx:    (qp_max_recv_dtos >> 2)
+ * @name:             connection peer portal
+ * @release_work:     deffered work for release job
+ * @state_mutex:      protects iser onnection state
+ * @stop_completion:  conn_stop completion
+ * @ib_completion:    RDMA cleanup completion
+ * @up_completion:    connection establishment completed
+ *                    (state is ISER_CONN_UP)
+ * @conn_list:        entry in ig conn list
+ * @login_buf:        login data buffer (stores login parameters)
+ * @login_req_buf:    login request buffer
+ * @login_req_dma:    login request buffer dma address
+ * @login_resp_buf:   login response buffer
+ * @login_resp_dma:   login response buffer dma address
+ * @rx_desc_head:     head of rx_descs cyclic buffer
+ * @rx_descs:         rx buffers array (cyclic buffer)
+ * @num_rx_descs:     number of rx descriptors
+ */
 struct iser_conn {
+	struct ib_conn		     ib_conn;
 	struct iscsi_conn	     *iscsi_conn;
 	struct iscsi_endpoint	     *ep;
-	enum iser_ib_conn_state	     state;	    /* rdma connection state   */
-	atomic_t		     refcount;
-	spinlock_t		     lock;	    /* used for state changes  */
-	struct iser_device           *device;       /* device context          */
-	struct rdma_cm_id            *cma_id;       /* CMA ID		       */
-	struct ib_qp	             *qp;           /* QP 		       */
-	wait_queue_head_t	     wait;          /* waitq for conn/disconn  */
-	unsigned		     qp_max_recv_dtos; /* num of rx buffers */
-	unsigned		     qp_max_recv_dtos_mask; /* above minus 1 */
-	unsigned		     min_posted_rx; /* qp_max_recv_dtos >> 2 */
-	int                          post_recv_buf_count; /* posted rx count  */
-	atomic_t                     post_send_buf_count; /* posted tx count   */
+	enum iser_conn_state	     state;
+	unsigned		     qp_max_recv_dtos;
+	unsigned		     qp_max_recv_dtos_mask;
+	unsigned		     min_posted_rx;
 	char 			     name[ISER_OBJECT_NAME_SIZE];
 	struct work_struct	     release_work;
+	struct mutex		     state_mutex;
 	struct completion	     stop_completion;
-	struct list_head	     conn_list;       /* entry in ig conn list */
+	struct completion	     ib_completion;
+	struct completion	     up_completion;
+	struct list_head	     conn_list;
 
 	char  			     *login_buf;
 	char			     *login_req_buf, *login_resp_buf;
 	u64			     login_req_dma, login_resp_dma;
 	unsigned int 		     rx_desc_head;
 	struct iser_rx_desc	     *rx_descs;
-	struct ib_recv_wr	     rx_wr[ISER_MIN_POSTED_RX];
-	bool			     pi_support;
-
-	/* Connection memory registration pool */
-	union {
-		struct {
-			struct ib_fmr_pool      *pool;	   /* pool of IB FMRs         */
-			struct iser_page_vec	*page_vec; /* represents SG to fmr maps*
-							    * maps serialized as tx is*/
-		} fmr;
-		struct {
-			struct list_head	pool;
-			int			pool_size;
-		} fastreg;
-	};
+	u32                          num_rx_descs;
 };
 
+/**
+ * struct iscsi_iser_task - iser task context
+ *
+ * @desc:     TX descriptor
+ * @iser_conn:        link to iser connection
+ * @status:           current task status
+ * @sc:               link to scsi command
+ * @command_sent:     indicate if command was sent
+ * @dir:              iser data direction
+ * @rdma_regd:        task rdma registration desc
+ * @data:             iser data buffer desc
+ * @data_copy:        iser data copy buffer desc (bounce buffer)
+ * @prot:             iser protection buffer desc
+ * @prot_copy:        iser protection copy buffer desc (bounce buffer)
+ */
 struct iscsi_iser_task {
 	struct iser_tx_desc          desc;
-	struct iser_conn	     *ib_conn;
+	struct iser_conn	     *iser_conn;
 	enum iser_task_status 	     status;
 	struct scsi_cmnd	     *sc;
-	int                          command_sent;  /* set if command  sent  */
-	int                          dir[ISER_DIRS_NUM];      /* set if dir use*/
-	struct iser_regd_buf         rdma_regd[ISER_DIRS_NUM];/* regd rdma buf */
-	struct iser_data_buf         data[ISER_DIRS_NUM];     /* orig. data des*/
-	struct iser_data_buf         data_copy[ISER_DIRS_NUM];/* contig. copy  */
-	struct iser_data_buf         prot[ISER_DIRS_NUM];     /* prot desc     */
-	struct iser_data_buf         prot_copy[ISER_DIRS_NUM];/* prot copy     */
+	int                          command_sent;
+	int                          dir[ISER_DIRS_NUM];
+	struct iser_regd_buf         rdma_regd[ISER_DIRS_NUM];
+	struct iser_data_buf         data[ISER_DIRS_NUM];
+	struct iser_data_buf         data_copy[ISER_DIRS_NUM];
+	struct iser_data_buf         prot[ISER_DIRS_NUM];
+	struct iser_data_buf         prot_copy[ISER_DIRS_NUM];
 };
 
 struct iser_page_vec {
@@ -380,17 +559,20 @@ struct iser_page_vec {
 	int data_size;
 };
 
-struct iser_cq_desc {
-	struct iser_device           *device;
-	int                          cq_index;
-};
-
+/**
+ * struct iser_global: iSER global context
+ *
+ * @device_list_mutex:    protects device_list
+ * @device_list:          iser devices global list
+ * @connlist_mutex:       protects connlist
+ * @connlist:             iser connections global list
+ * @desc_cache:           kmem cache for tx dataout
+ */
 struct iser_global {
-	struct mutex      device_list_mutex;/*                   */
-	struct list_head  device_list;	     /* all iSER devices */
+	struct mutex      device_list_mutex;
+	struct list_head  device_list;
 	struct mutex      connlist_mutex;
-	struct list_head  connlist;		/* all iSER IB connections */
-
+	struct list_head  connlist;
 	struct kmem_cache *desc_cache;
 };
 
@@ -399,9 +581,6 @@ extern int iser_debug_level;
 extern bool iser_pi_enable;
 extern int iser_pi_guard;
 
-/* allocate connection resources needed for rdma functionality */
-int iser_conn_set_full_featured_mode(struct iscsi_conn *conn);
-
 int iser_send_control(struct iscsi_conn *conn,
 		      struct iscsi_task *task);
 
@@ -413,29 +592,30 @@ int iser_send_data_out(struct iscsi_conn *conn,
 		       struct iscsi_data *hdr);
 
 void iscsi_iser_recv(struct iscsi_conn *conn,
-		     struct iscsi_hdr       *hdr,
-		     char                   *rx_data,
-		     int                    rx_data_len);
+		     struct iscsi_hdr *hdr,
+		     char *rx_data,
+		     int rx_data_len);
 
-void iser_conn_init(struct iser_conn *ib_conn);
+void iser_conn_init(struct iser_conn *iser_conn);
 
-void iser_conn_release(struct iser_conn *ib_conn);
+void iser_conn_release(struct iser_conn *iser_conn);
 
-void iser_conn_terminate(struct iser_conn *ib_conn);
+int iser_conn_terminate(struct iser_conn *iser_conn);
 
 void iser_release_work(struct work_struct *work);
 
 void iser_rcv_completion(struct iser_rx_desc *desc,
-			 unsigned long    dto_xfer_len,
-			struct iser_conn *ib_conn);
+			 unsigned long dto_xfer_len,
+			 struct ib_conn *ib_conn);
 
-void iser_snd_completion(struct iser_tx_desc *desc, struct iser_conn *ib_conn);
+void iser_snd_completion(struct iser_tx_desc *desc,
+			 struct ib_conn *ib_conn);
 
 void iser_task_rdma_init(struct iscsi_iser_task *task);
 
 void iser_task_rdma_finalize(struct iscsi_iser_task *task);
 
-void iser_free_rx_descriptors(struct iser_conn *ib_conn);
+void iser_free_rx_descriptors(struct iser_conn *iser_conn);
 
 void iser_finalize_rdma_unaligned_sg(struct iscsi_iser_task *iser_task,
 				     struct iser_data_buf *mem,
@@ -447,38 +627,40 @@ int  iser_reg_rdma_mem_fmr(struct iscsi_iser_task *task,
 int  iser_reg_rdma_mem_fastreg(struct iscsi_iser_task *task,
 			       enum iser_data_dir cmd_dir);
 
-int  iser_connect(struct iser_conn   *ib_conn,
-		  struct sockaddr_in *src_addr,
-		  struct sockaddr_in *dst_addr,
-		  int                non_blocking);
+int  iser_connect(struct iser_conn *iser_conn,
+		  struct sockaddr *src_addr,
+		  struct sockaddr *dst_addr,
+		  int non_blocking);
 
-int  iser_reg_page_vec(struct iser_conn     *ib_conn,
+int  iser_reg_page_vec(struct ib_conn *ib_conn,
 		       struct iser_page_vec *page_vec,
-		       struct iser_mem_reg  *mem_reg);
+		       struct iser_mem_reg *mem_reg);
 
 void iser_unreg_mem_fmr(struct iscsi_iser_task *iser_task,
 			enum iser_data_dir cmd_dir);
 void iser_unreg_mem_fastreg(struct iscsi_iser_task *iser_task,
 			    enum iser_data_dir cmd_dir);
 
-int  iser_post_recvl(struct iser_conn *ib_conn);
-int  iser_post_recvm(struct iser_conn *ib_conn, int count);
-int  iser_post_send(struct iser_conn *ib_conn, struct iser_tx_desc *tx_desc);
+int  iser_post_recvl(struct iser_conn *iser_conn);
+int  iser_post_recvm(struct iser_conn *iser_conn, int count);
+int  iser_post_send(struct ib_conn *ib_conn, struct iser_tx_desc *tx_desc,
+		    bool signal);
 
 int iser_dma_map_task_data(struct iscsi_iser_task *iser_task,
-			    struct iser_data_buf       *data,
-			    enum   iser_data_dir       iser_dir,
-			    enum   dma_data_direction  dma_dir);
+			   struct iser_data_buf *data,
+			   enum iser_data_dir iser_dir,
+			   enum dma_data_direction dma_dir);
 
 void iser_dma_unmap_task_data(struct iscsi_iser_task *iser_task,
 			      struct iser_data_buf *data);
 int  iser_initialize_task_headers(struct iscsi_task *task,
 			struct iser_tx_desc *tx_desc);
-int iser_alloc_rx_descriptors(struct iser_conn *ib_conn, struct iscsi_session *session);
-int iser_create_fmr_pool(struct iser_conn *ib_conn, unsigned cmds_max);
-void iser_free_fmr_pool(struct iser_conn *ib_conn);
-int iser_create_fastreg_pool(struct iser_conn *ib_conn, unsigned cmds_max);
-void iser_free_fastreg_pool(struct iser_conn *ib_conn);
+int iser_alloc_rx_descriptors(struct iser_conn *iser_conn,
+			      struct iscsi_session *session);
+int iser_create_fmr_pool(struct ib_conn *ib_conn, unsigned cmds_max);
+void iser_free_fmr_pool(struct ib_conn *ib_conn);
+int iser_create_fastreg_pool(struct ib_conn *ib_conn, unsigned cmds_max);
+void iser_free_fastreg_pool(struct ib_conn *ib_conn);
 u8 iser_check_task_pi_status(struct iscsi_iser_task *iser_task,
 			     enum iser_data_dir cmd_dir, sector_t *sector);
 #endif
diff --git a/drivers/infiniband/ulp/iser/iser_initiator.c b/drivers/infiniband/ulp/iser/iser_initiator.c
index 8d44a4060634..5a489ea63732 100644
--- a/drivers/infiniband/ulp/iser/iser_initiator.c
+++ b/drivers/infiniband/ulp/iser/iser_initiator.c
@@ -49,7 +49,7 @@ static int iser_prepare_read_cmd(struct iscsi_task *task)
 
 {
 	struct iscsi_iser_task *iser_task = task->dd_data;
-	struct iser_device  *device = iser_task->ib_conn->device;
+	struct iser_device  *device = iser_task->iser_conn->ib_conn.device;
 	struct iser_regd_buf *regd_buf;
 	int err;
 	struct iser_hdr *hdr = &iser_task->desc.iser_header;
@@ -103,7 +103,7 @@ iser_prepare_write_cmd(struct iscsi_task *task,
 		       unsigned int edtl)
 {
 	struct iscsi_iser_task *iser_task = task->dd_data;
-	struct iser_device  *device = iser_task->ib_conn->device;
+	struct iser_device  *device = iser_task->iser_conn->ib_conn.device;
 	struct iser_regd_buf *regd_buf;
 	int err;
 	struct iser_hdr *hdr = &iser_task->desc.iser_header;
@@ -160,10 +160,10 @@ iser_prepare_write_cmd(struct iscsi_task *task,
 }
 
 /* creates a new tx descriptor and adds header regd buffer */
-static void iser_create_send_desc(struct iser_conn	*ib_conn,
+static void iser_create_send_desc(struct iser_conn	*iser_conn,
 				  struct iser_tx_desc	*tx_desc)
 {
-	struct iser_device *device = ib_conn->device;
+	struct iser_device *device = iser_conn->ib_conn.device;
 
 	ib_dma_sync_single_for_cpu(device->ib_device,
 		tx_desc->dma_addr, ISER_HEADERS_LEN, DMA_TO_DEVICE);
@@ -179,103 +179,108 @@ static void iser_create_send_desc(struct iser_conn	*ib_conn,
 	}
 }
 
-static void iser_free_login_buf(struct iser_conn *ib_conn)
+static void iser_free_login_buf(struct iser_conn *iser_conn)
 {
-	if (!ib_conn->login_buf)
+	struct iser_device *device = iser_conn->ib_conn.device;
+
+	if (!iser_conn->login_buf)
 		return;
 
-	if (ib_conn->login_req_dma)
-		ib_dma_unmap_single(ib_conn->device->ib_device,
-				    ib_conn->login_req_dma,
+	if (iser_conn->login_req_dma)
+		ib_dma_unmap_single(device->ib_device,
+				    iser_conn->login_req_dma,
 				    ISCSI_DEF_MAX_RECV_SEG_LEN, DMA_TO_DEVICE);
 
-	if (ib_conn->login_resp_dma)
-		ib_dma_unmap_single(ib_conn->device->ib_device,
-				    ib_conn->login_resp_dma,
+	if (iser_conn->login_resp_dma)
+		ib_dma_unmap_single(device->ib_device,
+				    iser_conn->login_resp_dma,
 				    ISER_RX_LOGIN_SIZE, DMA_FROM_DEVICE);
 
-	kfree(ib_conn->login_buf);
+	kfree(iser_conn->login_buf);
 
 	/* make sure we never redo any unmapping */
-	ib_conn->login_req_dma = 0;
-	ib_conn->login_resp_dma = 0;
-	ib_conn->login_buf = NULL;
+	iser_conn->login_req_dma = 0;
+	iser_conn->login_resp_dma = 0;
+	iser_conn->login_buf = NULL;
 }
 
-static int iser_alloc_login_buf(struct iser_conn *ib_conn)
+static int iser_alloc_login_buf(struct iser_conn *iser_conn)
 {
-	struct iser_device	*device;
+	struct iser_device *device = iser_conn->ib_conn.device;
 	int			req_err, resp_err;
 
-	BUG_ON(ib_conn->device == NULL);
+	BUG_ON(device == NULL);
 
-	device = ib_conn->device;
-
-	ib_conn->login_buf = kmalloc(ISCSI_DEF_MAX_RECV_SEG_LEN +
+	iser_conn->login_buf = kmalloc(ISCSI_DEF_MAX_RECV_SEG_LEN +
 				     ISER_RX_LOGIN_SIZE, GFP_KERNEL);
-	if (!ib_conn->login_buf)
+	if (!iser_conn->login_buf)
 		goto out_err;
 
-	ib_conn->login_req_buf  = ib_conn->login_buf;
-	ib_conn->login_resp_buf = ib_conn->login_buf +
+	iser_conn->login_req_buf  = iser_conn->login_buf;
+	iser_conn->login_resp_buf = iser_conn->login_buf +
 						ISCSI_DEF_MAX_RECV_SEG_LEN;
 
-	ib_conn->login_req_dma = ib_dma_map_single(ib_conn->device->ib_device,
-				(void *)ib_conn->login_req_buf,
-				ISCSI_DEF_MAX_RECV_SEG_LEN, DMA_TO_DEVICE);
+	iser_conn->login_req_dma = ib_dma_map_single(device->ib_device,
+						     iser_conn->login_req_buf,
+						     ISCSI_DEF_MAX_RECV_SEG_LEN,
+						     DMA_TO_DEVICE);
 
-	ib_conn->login_resp_dma = ib_dma_map_single(ib_conn->device->ib_device,
-				(void *)ib_conn->login_resp_buf,
-				ISER_RX_LOGIN_SIZE, DMA_FROM_DEVICE);
+	iser_conn->login_resp_dma = ib_dma_map_single(device->ib_device,
+						      iser_conn->login_resp_buf,
+						      ISER_RX_LOGIN_SIZE,
+						      DMA_FROM_DEVICE);
 
 	req_err  = ib_dma_mapping_error(device->ib_device,
-					ib_conn->login_req_dma);
+					iser_conn->login_req_dma);
 	resp_err = ib_dma_mapping_error(device->ib_device,
-					ib_conn->login_resp_dma);
+					iser_conn->login_resp_dma);
 
 	if (req_err || resp_err) {
 		if (req_err)
-			ib_conn->login_req_dma = 0;
+			iser_conn->login_req_dma = 0;
 		if (resp_err)
-			ib_conn->login_resp_dma = 0;
+			iser_conn->login_resp_dma = 0;
 		goto free_login_buf;
 	}
 	return 0;
 
 free_login_buf:
-	iser_free_login_buf(ib_conn);
+	iser_free_login_buf(iser_conn);
 
 out_err:
 	iser_err("unable to alloc or map login buf\n");
 	return -ENOMEM;
 }
 
-int iser_alloc_rx_descriptors(struct iser_conn *ib_conn, struct iscsi_session *session)
+int iser_alloc_rx_descriptors(struct iser_conn *iser_conn,
+			      struct iscsi_session *session)
 {
 	int i, j;
 	u64 dma_addr;
 	struct iser_rx_desc *rx_desc;
 	struct ib_sge       *rx_sg;
-	struct iser_device  *device = ib_conn->device;
+	struct ib_conn *ib_conn = &iser_conn->ib_conn;
+	struct iser_device *device = ib_conn->device;
 
-	ib_conn->qp_max_recv_dtos = session->cmds_max;
-	ib_conn->qp_max_recv_dtos_mask = session->cmds_max - 1; /* cmds_max is 2^N */
-	ib_conn->min_posted_rx = ib_conn->qp_max_recv_dtos >> 2;
+	iser_conn->qp_max_recv_dtos = session->cmds_max;
+	iser_conn->qp_max_recv_dtos_mask = session->cmds_max - 1; /* cmds_max is 2^N */
+	iser_conn->min_posted_rx = iser_conn->qp_max_recv_dtos >> 2;
 
 	if (device->iser_alloc_rdma_reg_res(ib_conn, session->scsi_cmds_max))
 		goto create_rdma_reg_res_failed;
 
-	if (iser_alloc_login_buf(ib_conn))
+	if (iser_alloc_login_buf(iser_conn))
 		goto alloc_login_buf_fail;
 
-	ib_conn->rx_descs = kmalloc(session->cmds_max *
+	iser_conn->num_rx_descs = session->cmds_max;
+	iser_conn->rx_descs = kmalloc(iser_conn->num_rx_descs *
 				sizeof(struct iser_rx_desc), GFP_KERNEL);
-	if (!ib_conn->rx_descs)
+	if (!iser_conn->rx_descs)
 		goto rx_desc_alloc_fail;
 
-	rx_desc = ib_conn->rx_descs;
+	rx_desc = iser_conn->rx_descs;
 
-	for (i = 0; i < ib_conn->qp_max_recv_dtos; i++, rx_desc++)  {
+	for (i = 0; i < iser_conn->qp_max_recv_dtos; i++, rx_desc++)  {
 		dma_addr = ib_dma_map_single(device->ib_device, (void *)rx_desc,
 					ISER_RX_PAYLOAD_SIZE, DMA_FROM_DEVICE);
 		if (ib_dma_mapping_error(device->ib_device, dma_addr))
@@ -289,18 +294,18 @@ int iser_alloc_rx_descriptors(struct iser_conn *ib_conn, struct iscsi_session *s
 		rx_sg->lkey   = device->mr->lkey;
 	}
 
-	ib_conn->rx_desc_head = 0;
+	iser_conn->rx_desc_head = 0;
 	return 0;
 
 rx_desc_dma_map_failed:
-	rx_desc = ib_conn->rx_descs;
+	rx_desc = iser_conn->rx_descs;
 	for (j = 0; j < i; j++, rx_desc++)
 		ib_dma_unmap_single(device->ib_device, rx_desc->dma_addr,
 				    ISER_RX_PAYLOAD_SIZE, DMA_FROM_DEVICE);
-	kfree(ib_conn->rx_descs);
-	ib_conn->rx_descs = NULL;
+	kfree(iser_conn->rx_descs);
+	iser_conn->rx_descs = NULL;
 rx_desc_alloc_fail:
-	iser_free_login_buf(ib_conn);
+	iser_free_login_buf(iser_conn);
 alloc_login_buf_fail:
 	device->iser_free_rdma_reg_res(ib_conn);
 create_rdma_reg_res_failed:
@@ -308,33 +313,35 @@ create_rdma_reg_res_failed:
 	return -ENOMEM;
 }
 
-void iser_free_rx_descriptors(struct iser_conn *ib_conn)
+void iser_free_rx_descriptors(struct iser_conn *iser_conn)
 {
 	int i;
 	struct iser_rx_desc *rx_desc;
+	struct ib_conn *ib_conn = &iser_conn->ib_conn;
 	struct iser_device *device = ib_conn->device;
 
-	if (!ib_conn->rx_descs)
+	if (!iser_conn->rx_descs)
 		goto free_login_buf;
 
 	if (device->iser_free_rdma_reg_res)
 		device->iser_free_rdma_reg_res(ib_conn);
 
-	rx_desc = ib_conn->rx_descs;
-	for (i = 0; i < ib_conn->qp_max_recv_dtos; i++, rx_desc++)
+	rx_desc = iser_conn->rx_descs;
+	for (i = 0; i < iser_conn->qp_max_recv_dtos; i++, rx_desc++)
 		ib_dma_unmap_single(device->ib_device, rx_desc->dma_addr,
 				    ISER_RX_PAYLOAD_SIZE, DMA_FROM_DEVICE);
-	kfree(ib_conn->rx_descs);
+	kfree(iser_conn->rx_descs);
 	/* make sure we never redo any unmapping */
-	ib_conn->rx_descs = NULL;
+	iser_conn->rx_descs = NULL;
 
 free_login_buf:
-	iser_free_login_buf(ib_conn);
+	iser_free_login_buf(iser_conn);
 }
 
 static int iser_post_rx_bufs(struct iscsi_conn *conn, struct iscsi_hdr *req)
 {
-	struct iser_conn *ib_conn = conn->dd_data;
+	struct iser_conn *iser_conn = conn->dd_data;
+	struct ib_conn *ib_conn = &iser_conn->ib_conn;
 	struct iscsi_session *session = conn->session;
 
 	iser_dbg("req op %x flags %x\n", req->opcode, req->flags);
@@ -343,34 +350,37 @@ static int iser_post_rx_bufs(struct iscsi_conn *conn, struct iscsi_hdr *req)
 		return 0;
 
 	/*
-	 * Check that there is one posted recv buffer (for the last login
-	 * response) and no posted send buffers left - they must have been
-	 * consumed during previous login phases.
+	 * Check that there is one posted recv buffer
+	 * (for the last login response).
 	 */
 	WARN_ON(ib_conn->post_recv_buf_count != 1);
-	WARN_ON(atomic_read(&ib_conn->post_send_buf_count) != 0);
 
 	if (session->discovery_sess) {
 		iser_info("Discovery session, re-using login RX buffer\n");
 		return 0;
 	} else
 		iser_info("Normal session, posting batch of RX %d buffers\n",
-			  ib_conn->min_posted_rx);
+			  iser_conn->min_posted_rx);
 
 	/* Initial post receive buffers */
-	if (iser_post_recvm(ib_conn, ib_conn->min_posted_rx))
+	if (iser_post_recvm(iser_conn, iser_conn->min_posted_rx))
 		return -ENOMEM;
 
 	return 0;
 }
 
+static inline bool iser_signal_comp(int sig_count)
+{
+	return ((sig_count % ISER_SIGNAL_CMD_COUNT) == 0);
+}
+
 /**
  * iser_send_command - send command PDU
  */
 int iser_send_command(struct iscsi_conn *conn,
 		      struct iscsi_task *task)
 {
-	struct iser_conn *ib_conn = conn->dd_data;
+	struct iser_conn *iser_conn = conn->dd_data;
 	struct iscsi_iser_task *iser_task = task->dd_data;
 	unsigned long edtl;
 	int err;
@@ -378,12 +388,13 @@ int iser_send_command(struct iscsi_conn *conn,
 	struct iscsi_scsi_req *hdr = (struct iscsi_scsi_req *)task->hdr;
 	struct scsi_cmnd *sc  =  task->sc;
 	struct iser_tx_desc *tx_desc = &iser_task->desc;
+	static unsigned sig_count;
 
 	edtl = ntohl(hdr->data_length);
 
 	/* build the tx desc regd header and add it to the tx desc dto */
 	tx_desc->type = ISCSI_TX_SCSI_COMMAND;
-	iser_create_send_desc(ib_conn, tx_desc);
+	iser_create_send_desc(iser_conn, tx_desc);
 
 	if (hdr->flags & ISCSI_FLAG_CMD_READ) {
 		data_buf = &iser_task->data[ISER_DIR_IN];
@@ -423,7 +434,8 @@ int iser_send_command(struct iscsi_conn *conn,
 
 	iser_task->status = ISER_TASK_STATUS_STARTED;
 
-	err = iser_post_send(ib_conn, tx_desc);
+	err = iser_post_send(&iser_conn->ib_conn, tx_desc,
+			     iser_signal_comp(++sig_count));
 	if (!err)
 		return 0;
 
@@ -439,7 +451,7 @@ int iser_send_data_out(struct iscsi_conn *conn,
 		       struct iscsi_task *task,
 		       struct iscsi_data *hdr)
 {
-	struct iser_conn *ib_conn = conn->dd_data;
+	struct iser_conn *iser_conn = conn->dd_data;
 	struct iscsi_iser_task *iser_task = task->dd_data;
 	struct iser_tx_desc *tx_desc = NULL;
 	struct iser_regd_buf *regd_buf;
@@ -488,7 +500,7 @@ int iser_send_data_out(struct iscsi_conn *conn,
 		 itt, buf_offset, data_seg_len);
 
 
-	err = iser_post_send(ib_conn, tx_desc);
+	err = iser_post_send(&iser_conn->ib_conn, tx_desc, true);
 	if (!err)
 		return 0;
 
@@ -501,7 +513,7 @@ send_data_out_error:
 int iser_send_control(struct iscsi_conn *conn,
 		      struct iscsi_task *task)
 {
-	struct iser_conn *ib_conn = conn->dd_data;
+	struct iser_conn *iser_conn = conn->dd_data;
 	struct iscsi_iser_task *iser_task = task->dd_data;
 	struct iser_tx_desc *mdesc = &iser_task->desc;
 	unsigned long data_seg_len;
@@ -510,9 +522,9 @@ int iser_send_control(struct iscsi_conn *conn,
 
 	/* build the tx desc regd header and add it to the tx desc dto */
 	mdesc->type = ISCSI_TX_CONTROL;
-	iser_create_send_desc(ib_conn, mdesc);
+	iser_create_send_desc(iser_conn, mdesc);
 
-	device = ib_conn->device;
+	device = iser_conn->ib_conn.device;
 
 	data_seg_len = ntoh24(task->hdr->dlength);
 
@@ -524,16 +536,16 @@ int iser_send_control(struct iscsi_conn *conn,
 		}
 
 		ib_dma_sync_single_for_cpu(device->ib_device,
-			ib_conn->login_req_dma, task->data_count,
+			iser_conn->login_req_dma, task->data_count,
 			DMA_TO_DEVICE);
 
-		memcpy(ib_conn->login_req_buf, task->data, task->data_count);
+		memcpy(iser_conn->login_req_buf, task->data, task->data_count);
 
 		ib_dma_sync_single_for_device(device->ib_device,
-			ib_conn->login_req_dma, task->data_count,
+			iser_conn->login_req_dma, task->data_count,
 			DMA_TO_DEVICE);
 
-		tx_dsg->addr    = ib_conn->login_req_dma;
+		tx_dsg->addr    = iser_conn->login_req_dma;
 		tx_dsg->length  = task->data_count;
 		tx_dsg->lkey    = device->mr->lkey;
 		mdesc->num_sge = 2;
@@ -542,7 +554,7 @@ int iser_send_control(struct iscsi_conn *conn,
 	if (task == conn->login_task) {
 		iser_dbg("op %x dsl %lx, posting login rx buffer\n",
 			 task->hdr->opcode, data_seg_len);
-		err = iser_post_recvl(ib_conn);
+		err = iser_post_recvl(iser_conn);
 		if (err)
 			goto send_control_error;
 		err = iser_post_rx_bufs(conn, task->hdr);
@@ -550,7 +562,7 @@ int iser_send_control(struct iscsi_conn *conn,
 			goto send_control_error;
 	}
 
-	err = iser_post_send(ib_conn, mdesc);
+	err = iser_post_send(&iser_conn->ib_conn, mdesc, true);
 	if (!err)
 		return 0;
 
@@ -564,15 +576,17 @@ send_control_error:
  */
 void iser_rcv_completion(struct iser_rx_desc *rx_desc,
 			 unsigned long rx_xfer_len,
-			 struct iser_conn *ib_conn)
+			 struct ib_conn *ib_conn)
 {
+	struct iser_conn *iser_conn = container_of(ib_conn, struct iser_conn,
+						   ib_conn);
 	struct iscsi_hdr *hdr;
 	u64 rx_dma;
 	int rx_buflen, outstanding, count, err;
 
 	/* differentiate between login to all other PDUs */
-	if ((char *)rx_desc == ib_conn->login_resp_buf) {
-		rx_dma = ib_conn->login_resp_dma;
+	if ((char *)rx_desc == iser_conn->login_resp_buf) {
+		rx_dma = iser_conn->login_resp_dma;
 		rx_buflen = ISER_RX_LOGIN_SIZE;
 	} else {
 		rx_dma = rx_desc->dma_addr;
@@ -580,14 +594,14 @@ void iser_rcv_completion(struct iser_rx_desc *rx_desc,
 	}
 
 	ib_dma_sync_single_for_cpu(ib_conn->device->ib_device, rx_dma,
-			rx_buflen, DMA_FROM_DEVICE);
+				   rx_buflen, DMA_FROM_DEVICE);
 
 	hdr = &rx_desc->iscsi_header;
 
 	iser_dbg("op 0x%x itt 0x%x dlen %d\n", hdr->opcode,
 			hdr->itt, (int)(rx_xfer_len - ISER_HEADERS_LEN));
 
-	iscsi_iser_recv(ib_conn->iscsi_conn, hdr, rx_desc->data,
+	iscsi_iser_recv(iser_conn->iscsi_conn, hdr, rx_desc->data,
 			rx_xfer_len - ISER_HEADERS_LEN);
 
 	ib_dma_sync_single_for_device(ib_conn->device->ib_device, rx_dma,
@@ -599,21 +613,21 @@ void iser_rcv_completion(struct iser_rx_desc *rx_desc,
 	 * for the posted rx bufs refcount to become zero handles everything   */
 	ib_conn->post_recv_buf_count--;
 
-	if (rx_dma == ib_conn->login_resp_dma)
+	if (rx_dma == iser_conn->login_resp_dma)
 		return;
 
 	outstanding = ib_conn->post_recv_buf_count;
-	if (outstanding + ib_conn->min_posted_rx <= ib_conn->qp_max_recv_dtos) {
-		count = min(ib_conn->qp_max_recv_dtos - outstanding,
-						ib_conn->min_posted_rx);
-		err = iser_post_recvm(ib_conn, count);
+	if (outstanding + iser_conn->min_posted_rx <= iser_conn->qp_max_recv_dtos) {
+		count = min(iser_conn->qp_max_recv_dtos - outstanding,
+			    iser_conn->min_posted_rx);
+		err = iser_post_recvm(iser_conn, count);
 		if (err)
 			iser_err("posting %d rx bufs err %d\n", count, err);
 	}
 }
 
 void iser_snd_completion(struct iser_tx_desc *tx_desc,
-			struct iser_conn *ib_conn)
+			struct ib_conn *ib_conn)
 {
 	struct iscsi_task *task;
 	struct iser_device *device = ib_conn->device;
@@ -625,8 +639,6 @@ void iser_snd_completion(struct iser_tx_desc *tx_desc,
 		tx_desc = NULL;
 	}
 
-	atomic_dec(&ib_conn->post_send_buf_count);
-
 	if (tx_desc && tx_desc->type == ISCSI_TX_CONTROL) {
 		/* this arithmetic is legal by libiscsi dd_data allocation */
 		task = (void *) ((long)(void *)tx_desc -
@@ -658,7 +670,7 @@ void iser_task_rdma_init(struct iscsi_iser_task *iser_task)
 
 void iser_task_rdma_finalize(struct iscsi_iser_task *iser_task)
 {
-	struct iser_device *device = iser_task->ib_conn->device;
+	struct iser_device *device = iser_task->iser_conn->ib_conn.device;
 	int is_rdma_data_aligned = 1;
 	int is_rdma_prot_aligned = 1;
 	int prot_count = scsi_prot_sg_count(iser_task->sc);
diff --git a/drivers/infiniband/ulp/iser/iser_memory.c b/drivers/infiniband/ulp/iser/iser_memory.c
index 47acd3ad3a17..6c5ce357fba6 100644
--- a/drivers/infiniband/ulp/iser/iser_memory.c
+++ b/drivers/infiniband/ulp/iser/iser_memory.c
@@ -49,7 +49,7 @@ static int iser_start_rdma_unaligned_sg(struct iscsi_iser_task *iser_task,
 					struct iser_data_buf *data_copy,
 					enum iser_data_dir cmd_dir)
 {
-	struct ib_device *dev = iser_task->ib_conn->device->ib_device;
+	struct ib_device *dev = iser_task->iser_conn->ib_conn.device->ib_device;
 	struct scatterlist *sgl = (struct scatterlist *)data->buf;
 	struct scatterlist *sg;
 	char *mem = NULL;
@@ -116,7 +116,7 @@ void iser_finalize_rdma_unaligned_sg(struct iscsi_iser_task *iser_task,
 	struct ib_device *dev;
 	unsigned long  cmd_data_len;
 
-	dev = iser_task->ib_conn->device->ib_device;
+	dev = iser_task->iser_conn->ib_conn.device->ib_device;
 
 	ib_dma_unmap_sg(dev, &data_copy->sg_single, 1,
 			(cmd_dir == ISER_DIR_OUT) ?
@@ -322,7 +322,7 @@ int iser_dma_map_task_data(struct iscsi_iser_task *iser_task,
 	struct ib_device *dev;
 
 	iser_task->dir[iser_dir] = 1;
-	dev = iser_task->ib_conn->device->ib_device;
+	dev = iser_task->iser_conn->ib_conn.device->ib_device;
 
 	data->dma_nents = ib_dma_map_sg(dev, data->buf, data->size, dma_dir);
 	if (data->dma_nents == 0) {
@@ -337,7 +337,7 @@ void iser_dma_unmap_task_data(struct iscsi_iser_task *iser_task,
 {
 	struct ib_device *dev;
 
-	dev = iser_task->ib_conn->device->ib_device;
+	dev = iser_task->iser_conn->ib_conn.device->ib_device;
 	ib_dma_unmap_sg(dev, data->buf, data->size, DMA_FROM_DEVICE);
 }
 
@@ -348,7 +348,7 @@ static int fall_to_bounce_buf(struct iscsi_iser_task *iser_task,
 			      enum iser_data_dir cmd_dir,
 			      int aligned_len)
 {
-	struct iscsi_conn    *iscsi_conn = iser_task->ib_conn->iscsi_conn;
+	struct iscsi_conn    *iscsi_conn = iser_task->iser_conn->iscsi_conn;
 
 	iscsi_conn->fmr_unalign_cnt++;
 	iser_warn("rdma alignment violation (%d/%d aligned) or FMR not supported\n",
@@ -377,7 +377,7 @@ static int fall_to_bounce_buf(struct iscsi_iser_task *iser_task,
 int iser_reg_rdma_mem_fmr(struct iscsi_iser_task *iser_task,
 			  enum iser_data_dir cmd_dir)
 {
-	struct iser_conn     *ib_conn = iser_task->ib_conn;
+	struct ib_conn *ib_conn = &iser_task->iser_conn->ib_conn;
 	struct iser_device   *device = ib_conn->device;
 	struct ib_device     *ibdev = device->ib_device;
 	struct iser_data_buf *mem = &iser_task->data[cmd_dir];
@@ -432,7 +432,7 @@ int iser_reg_rdma_mem_fmr(struct iscsi_iser_task *iser_task,
 				 ib_conn->fmr.page_vec->offset);
 			for (i = 0; i < ib_conn->fmr.page_vec->length; i++)
 				iser_err("page_vec[%d] = 0x%llx\n", i,
-					 (unsigned long long) ib_conn->fmr.page_vec->pages[i]);
+					 (unsigned long long)ib_conn->fmr.page_vec->pages[i]);
 		}
 		if (err)
 			return err;
@@ -440,77 +440,74 @@ int iser_reg_rdma_mem_fmr(struct iscsi_iser_task *iser_task,
 	return 0;
 }
 
-static inline enum ib_t10_dif_type
-scsi2ib_prot_type(unsigned char prot_type)
+static inline void
+iser_set_dif_domain(struct scsi_cmnd *sc, struct ib_sig_attrs *sig_attrs,
+		    struct ib_sig_domain *domain)
 {
-	switch (prot_type) {
-	case SCSI_PROT_DIF_TYPE0:
-		return IB_T10DIF_NONE;
-	case SCSI_PROT_DIF_TYPE1:
-		return IB_T10DIF_TYPE1;
-	case SCSI_PROT_DIF_TYPE2:
-		return IB_T10DIF_TYPE2;
-	case SCSI_PROT_DIF_TYPE3:
-		return IB_T10DIF_TYPE3;
-	default:
-		return IB_T10DIF_NONE;
-	}
-}
-
+	domain->sig_type = IB_SIG_TYPE_T10_DIF;
+	domain->sig.dif.pi_interval = sc->device->sector_size;
+	domain->sig.dif.ref_tag = scsi_get_lba(sc) & 0xffffffff;
+	/*
+	 * At the moment we hard code those, but in the future
+	 * we will take them from sc.
+	 */
+	domain->sig.dif.apptag_check_mask = 0xffff;
+	domain->sig.dif.app_escape = true;
+	domain->sig.dif.ref_escape = true;
+	if (scsi_get_prot_type(sc) == SCSI_PROT_DIF_TYPE1 ||
+	    scsi_get_prot_type(sc) == SCSI_PROT_DIF_TYPE2)
+		domain->sig.dif.ref_remap = true;
+};
 
 static int
 iser_set_sig_attrs(struct scsi_cmnd *sc, struct ib_sig_attrs *sig_attrs)
 {
-	unsigned char scsi_ptype = scsi_get_prot_type(sc);
-
-	sig_attrs->mem.sig_type = IB_SIG_TYPE_T10_DIF;
-	sig_attrs->wire.sig_type = IB_SIG_TYPE_T10_DIF;
-	sig_attrs->mem.sig.dif.pi_interval = sc->device->sector_size;
-	sig_attrs->wire.sig.dif.pi_interval = sc->device->sector_size;
-
 	switch (scsi_get_prot_op(sc)) {
 	case SCSI_PROT_WRITE_INSERT:
 	case SCSI_PROT_READ_STRIP:
-		sig_attrs->mem.sig.dif.type = IB_T10DIF_NONE;
-		sig_attrs->wire.sig.dif.type = scsi2ib_prot_type(scsi_ptype);
+		sig_attrs->mem.sig_type = IB_SIG_TYPE_NONE;
+		iser_set_dif_domain(sc, sig_attrs, &sig_attrs->wire);
 		sig_attrs->wire.sig.dif.bg_type = IB_T10DIF_CRC;
-		sig_attrs->wire.sig.dif.ref_tag = scsi_get_lba(sc) &
-						  0xffffffff;
 		break;
 	case SCSI_PROT_READ_INSERT:
 	case SCSI_PROT_WRITE_STRIP:
-		sig_attrs->mem.sig.dif.type = scsi2ib_prot_type(scsi_ptype);
-		sig_attrs->mem.sig.dif.bg_type = IB_T10DIF_CRC;
-		sig_attrs->mem.sig.dif.ref_tag = scsi_get_lba(sc) &
-						 0xffffffff;
-		sig_attrs->wire.sig.dif.type = IB_T10DIF_NONE;
+		sig_attrs->wire.sig_type = IB_SIG_TYPE_NONE;
+		iser_set_dif_domain(sc, sig_attrs, &sig_attrs->mem);
+		/*
+		 * At the moment we use this modparam to tell what is
+		 * the memory bg_type, in the future we will take it
+		 * from sc.
+		 */
+		sig_attrs->mem.sig.dif.bg_type = iser_pi_guard ? IB_T10DIF_CSUM :
+						 IB_T10DIF_CRC;
 		break;
 	case SCSI_PROT_READ_PASS:
 	case SCSI_PROT_WRITE_PASS:
-		sig_attrs->mem.sig.dif.type = scsi2ib_prot_type(scsi_ptype);
-		sig_attrs->mem.sig.dif.bg_type = IB_T10DIF_CRC;
-		sig_attrs->mem.sig.dif.ref_tag = scsi_get_lba(sc) &
-						 0xffffffff;
-		sig_attrs->wire.sig.dif.type = scsi2ib_prot_type(scsi_ptype);
+		iser_set_dif_domain(sc, sig_attrs, &sig_attrs->wire);
 		sig_attrs->wire.sig.dif.bg_type = IB_T10DIF_CRC;
-		sig_attrs->wire.sig.dif.ref_tag = scsi_get_lba(sc) &
-						  0xffffffff;
+		iser_set_dif_domain(sc, sig_attrs, &sig_attrs->mem);
+		/*
+		 * At the moment we use this modparam to tell what is
+		 * the memory bg_type, in the future we will take it
+		 * from sc.
+		 */
+		sig_attrs->mem.sig.dif.bg_type = iser_pi_guard ? IB_T10DIF_CSUM :
+						 IB_T10DIF_CRC;
 		break;
 	default:
 		iser_err("Unsupported PI operation %d\n",
 			 scsi_get_prot_op(sc));
 		return -EINVAL;
 	}
+
 	return 0;
 }
 
-
 static int
 iser_set_prot_checks(struct scsi_cmnd *sc, u8 *mask)
 {
 	switch (scsi_get_prot_type(sc)) {
 	case SCSI_PROT_DIF_TYPE0:
-		*mask = 0x0;
 		break;
 	case SCSI_PROT_DIF_TYPE1:
 	case SCSI_PROT_DIF_TYPE2:
@@ -533,7 +530,7 @@ iser_reg_sig_mr(struct iscsi_iser_task *iser_task,
 		struct fast_reg_descriptor *desc, struct ib_sge *data_sge,
 		struct ib_sge *prot_sge, struct ib_sge *sig_sge)
 {
-	struct iser_conn *ib_conn = iser_task->ib_conn;
+	struct ib_conn *ib_conn = &iser_task->iser_conn->ib_conn;
 	struct iser_pi_context *pi_ctx = desc->pi_ctx;
 	struct ib_send_wr sig_wr, inv_wr;
 	struct ib_send_wr *bad_wr, *wr = NULL;
@@ -609,7 +606,7 @@ static int iser_fast_reg_mr(struct iscsi_iser_task *iser_task,
 			    struct ib_sge *sge)
 {
 	struct fast_reg_descriptor *desc = regd_buf->reg.mem_h;
-	struct iser_conn *ib_conn = iser_task->ib_conn;
+	struct ib_conn *ib_conn = &iser_task->iser_conn->ib_conn;
 	struct iser_device *device = ib_conn->device;
 	struct ib_device *ibdev = device->ib_device;
 	struct ib_mr *mr;
@@ -700,7 +697,7 @@ static int iser_fast_reg_mr(struct iscsi_iser_task *iser_task,
 int iser_reg_rdma_mem_fastreg(struct iscsi_iser_task *iser_task,
 			      enum iser_data_dir cmd_dir)
 {
-	struct iser_conn *ib_conn = iser_task->ib_conn;
+	struct ib_conn *ib_conn = &iser_task->iser_conn->ib_conn;
 	struct iser_device *device = ib_conn->device;
 	struct ib_device *ibdev = device->ib_device;
 	struct iser_data_buf *mem = &iser_task->data[cmd_dir];
diff --git a/drivers/infiniband/ulp/iser/iser_verbs.c b/drivers/infiniband/ulp/iser/iser_verbs.c
index ea01075f9f9b..67225bb82bb5 100644
--- a/drivers/infiniband/ulp/iser/iser_verbs.c
+++ b/drivers/infiniband/ulp/iser/iser_verbs.c
@@ -39,8 +39,12 @@
 #include "iscsi_iser.h"
 
 #define ISCSI_ISER_MAX_CONN	8
-#define ISER_MAX_RX_CQ_LEN	(ISER_QP_MAX_RECV_DTOS * ISCSI_ISER_MAX_CONN)
-#define ISER_MAX_TX_CQ_LEN	(ISER_QP_MAX_REQ_DTOS  * ISCSI_ISER_MAX_CONN)
+#define ISER_MAX_RX_LEN		(ISER_QP_MAX_RECV_DTOS * ISCSI_ISER_MAX_CONN)
+#define ISER_MAX_TX_LEN		(ISER_QP_MAX_REQ_DTOS  * ISCSI_ISER_MAX_CONN)
+#define ISER_MAX_CQ_LEN		(ISER_MAX_RX_LEN + ISER_MAX_TX_LEN + \
+				 ISCSI_ISER_MAX_CONN)
+
+static int iser_cq_poll_limit = 512;
 
 static void iser_cq_tasklet_fn(unsigned long data);
 static void iser_cq_callback(struct ib_cq *cq, void *cq_context);
@@ -71,9 +75,8 @@ static void iser_event_handler(struct ib_event_handler *handler,
  */
 static int iser_create_device_ib_res(struct iser_device *device)
 {
-	struct iser_cq_desc *cq_desc;
 	struct ib_device_attr *dev_attr = &device->dev_attr;
-	int ret, i, j;
+	int ret, i;
 
 	ret = ib_query_device(device->ib_device, dev_attr);
 	if (ret) {
@@ -101,47 +104,35 @@ static int iser_create_device_ib_res(struct iser_device *device)
 		return -1;
 	}
 
-	device->cqs_used = min(ISER_MAX_CQ, device->ib_device->num_comp_vectors);
+	device->comps_used = min(ISER_MAX_CQ,
+				 device->ib_device->num_comp_vectors);
 	iser_info("using %d CQs, device %s supports %d vectors\n",
-		  device->cqs_used, device->ib_device->name,
+		  device->comps_used, device->ib_device->name,
 		  device->ib_device->num_comp_vectors);
 
-	device->cq_desc = kmalloc(sizeof(struct iser_cq_desc) * device->cqs_used,
-				  GFP_KERNEL);
-	if (device->cq_desc == NULL)
-		goto cq_desc_err;
-	cq_desc = device->cq_desc;
-
 	device->pd = ib_alloc_pd(device->ib_device);
 	if (IS_ERR(device->pd))
 		goto pd_err;
 
-	for (i = 0; i < device->cqs_used; i++) {
-		cq_desc[i].device   = device;
-		cq_desc[i].cq_index = i;
-
-		device->rx_cq[i] = ib_create_cq(device->ib_device,
-					  iser_cq_callback,
-					  iser_cq_event_callback,
-					  (void *)&cq_desc[i],
-					  ISER_MAX_RX_CQ_LEN, i);
-		if (IS_ERR(device->rx_cq[i]))
-			goto cq_err;
-
-		device->tx_cq[i] = ib_create_cq(device->ib_device,
-					  NULL, iser_cq_event_callback,
-					  (void *)&cq_desc[i],
-					  ISER_MAX_TX_CQ_LEN, i);
-
-		if (IS_ERR(device->tx_cq[i]))
+	for (i = 0; i < device->comps_used; i++) {
+		struct iser_comp *comp = &device->comps[i];
+
+		comp->device = device;
+		comp->cq = ib_create_cq(device->ib_device,
+					iser_cq_callback,
+					iser_cq_event_callback,
+					(void *)comp,
+					ISER_MAX_CQ_LEN, i);
+		if (IS_ERR(comp->cq)) {
+			comp->cq = NULL;
 			goto cq_err;
+		}
 
-		if (ib_req_notify_cq(device->rx_cq[i], IB_CQ_NEXT_COMP))
+		if (ib_req_notify_cq(comp->cq, IB_CQ_NEXT_COMP))
 			goto cq_err;
 
-		tasklet_init(&device->cq_tasklet[i],
-			     iser_cq_tasklet_fn,
-			(unsigned long)&cq_desc[i]);
+		tasklet_init(&comp->tasklet, iser_cq_tasklet_fn,
+			     (unsigned long)comp);
 	}
 
 	device->mr = ib_get_dma_mr(device->pd, IB_ACCESS_LOCAL_WRITE |
@@ -160,19 +151,17 @@ static int iser_create_device_ib_res(struct iser_device *device)
 handler_err:
 	ib_dereg_mr(device->mr);
 dma_mr_err:
-	for (j = 0; j < device->cqs_used; j++)
-		tasklet_kill(&device->cq_tasklet[j]);
+	for (i = 0; i < device->comps_used; i++)
+		tasklet_kill(&device->comps[i].tasklet);
 cq_err:
-	for (j = 0; j < i; j++) {
-		if (device->tx_cq[j])
-			ib_destroy_cq(device->tx_cq[j]);
-		if (device->rx_cq[j])
-			ib_destroy_cq(device->rx_cq[j]);
+	for (i = 0; i < device->comps_used; i++) {
+		struct iser_comp *comp = &device->comps[i];
+
+		if (comp->cq)
+			ib_destroy_cq(comp->cq);
 	}
 	ib_dealloc_pd(device->pd);
 pd_err:
-	kfree(device->cq_desc);
-cq_desc_err:
 	iser_err("failed to allocate an IB resource\n");
 	return -1;
 }
@@ -186,20 +175,18 @@ static void iser_free_device_ib_res(struct iser_device *device)
 	int i;
 	BUG_ON(device->mr == NULL);
 
-	for (i = 0; i < device->cqs_used; i++) {
-		tasklet_kill(&device->cq_tasklet[i]);
-		(void)ib_destroy_cq(device->tx_cq[i]);
-		(void)ib_destroy_cq(device->rx_cq[i]);
-		device->tx_cq[i] = NULL;
-		device->rx_cq[i] = NULL;
+	for (i = 0; i < device->comps_used; i++) {
+		struct iser_comp *comp = &device->comps[i];
+
+		tasklet_kill(&comp->tasklet);
+		ib_destroy_cq(comp->cq);
+		comp->cq = NULL;
 	}
 
 	(void)ib_unregister_event_handler(&device->event_handler);
 	(void)ib_dereg_mr(device->mr);
 	(void)ib_dealloc_pd(device->pd);
 
-	kfree(device->cq_desc);
-
 	device->mr = NULL;
 	device->pd = NULL;
 }
@@ -209,7 +196,7 @@ static void iser_free_device_ib_res(struct iser_device *device)
  *
  * returns 0 on success, or errno code on failure
  */
-int iser_create_fmr_pool(struct iser_conn *ib_conn, unsigned cmds_max)
+int iser_create_fmr_pool(struct ib_conn *ib_conn, unsigned cmds_max)
 {
 	struct iser_device *device = ib_conn->device;
 	struct ib_fmr_pool_param params;
@@ -259,7 +246,7 @@ int iser_create_fmr_pool(struct iser_conn *ib_conn, unsigned cmds_max)
 /**
  * iser_free_fmr_pool - releases the FMR pool and page vec
  */
-void iser_free_fmr_pool(struct iser_conn *ib_conn)
+void iser_free_fmr_pool(struct ib_conn *ib_conn)
 {
 	iser_info("freeing conn %p fmr pool %p\n",
 		  ib_conn, ib_conn->fmr.pool);
@@ -363,10 +350,10 @@ fast_reg_mr_failure:
  * for fast registration work requests.
  * returns 0 on success, or errno code on failure
  */
-int iser_create_fastreg_pool(struct iser_conn *ib_conn, unsigned cmds_max)
+int iser_create_fastreg_pool(struct ib_conn *ib_conn, unsigned cmds_max)
 {
-	struct iser_device	*device = ib_conn->device;
-	struct fast_reg_descriptor	*desc;
+	struct iser_device *device = ib_conn->device;
+	struct fast_reg_descriptor *desc;
 	int i, ret;
 
 	INIT_LIST_HEAD(&ib_conn->fastreg.pool);
@@ -402,7 +389,7 @@ err:
 /**
  * iser_free_fastreg_pool - releases the pool of fast_reg descriptors
  */
-void iser_free_fastreg_pool(struct iser_conn *ib_conn)
+void iser_free_fastreg_pool(struct ib_conn *ib_conn)
 {
 	struct fast_reg_descriptor *desc, *tmp;
 	int i = 0;
@@ -436,7 +423,7 @@ void iser_free_fastreg_pool(struct iser_conn *ib_conn)
  *
  * returns 0 on success, -1 on failure
  */
-static int iser_create_ib_conn_res(struct iser_conn *ib_conn)
+static int iser_create_ib_conn_res(struct ib_conn *ib_conn)
 {
 	struct iser_device	*device;
 	struct ib_qp_init_attr	init_attr;
@@ -451,28 +438,30 @@ static int iser_create_ib_conn_res(struct iser_conn *ib_conn)
 
 	mutex_lock(&ig.connlist_mutex);
 	/* select the CQ with the minimal number of usages */
-	for (index = 0; index < device->cqs_used; index++)
-		if (device->cq_active_qps[index] <
-		    device->cq_active_qps[min_index])
+	for (index = 0; index < device->comps_used; index++) {
+		if (device->comps[index].active_qps <
+		    device->comps[min_index].active_qps)
 			min_index = index;
-	device->cq_active_qps[min_index]++;
+	}
+	ib_conn->comp = &device->comps[min_index];
+	ib_conn->comp->active_qps++;
 	mutex_unlock(&ig.connlist_mutex);
 	iser_info("cq index %d used for ib_conn %p\n", min_index, ib_conn);
 
 	init_attr.event_handler = iser_qp_event_callback;
 	init_attr.qp_context	= (void *)ib_conn;
-	init_attr.send_cq	= device->tx_cq[min_index];
-	init_attr.recv_cq	= device->rx_cq[min_index];
+	init_attr.send_cq	= ib_conn->comp->cq;
+	init_attr.recv_cq	= ib_conn->comp->cq;
 	init_attr.cap.max_recv_wr  = ISER_QP_MAX_RECV_DTOS;
 	init_attr.cap.max_send_sge = 2;
 	init_attr.cap.max_recv_sge = 1;
 	init_attr.sq_sig_type	= IB_SIGNAL_REQ_WR;
 	init_attr.qp_type	= IB_QPT_RC;
 	if (ib_conn->pi_support) {
-		init_attr.cap.max_send_wr = ISER_QP_SIG_MAX_REQ_DTOS;
+		init_attr.cap.max_send_wr = ISER_QP_SIG_MAX_REQ_DTOS + 1;
 		init_attr.create_flags |= IB_QP_CREATE_SIGNATURE_EN;
 	} else {
-		init_attr.cap.max_send_wr  = ISER_QP_MAX_REQ_DTOS;
+		init_attr.cap.max_send_wr  = ISER_QP_MAX_REQ_DTOS + 1;
 	}
 
 	ret = rdma_create_qp(ib_conn->cma_id, device->pd, &init_attr);
@@ -491,33 +480,6 @@ out_err:
 }
 
 /**
- * releases the QP objects, returns 0 on success,
- * -1 on failure
- */
-static int iser_free_ib_conn_res(struct iser_conn *ib_conn)
-{
-	int cq_index;
-	BUG_ON(ib_conn == NULL);
-
-	iser_info("freeing conn %p cma_id %p qp %p\n",
-		  ib_conn, ib_conn->cma_id,
-		  ib_conn->qp);
-
-	/* qp is created only once both addr & route are resolved */
-
-	if (ib_conn->qp != NULL) {
-		cq_index = ((struct iser_cq_desc *)ib_conn->qp->recv_cq->cq_context)->cq_index;
-		ib_conn->device->cq_active_qps[cq_index]--;
-
-		rdma_destroy_qp(ib_conn->cma_id);
-	}
-
-	ib_conn->qp	  = NULL;
-
-	return 0;
-}
-
-/**
  * based on the resolved device node GUID see if there already allocated
  * device for this device. If there's no such, create one.
  */
@@ -568,96 +530,174 @@ static void iser_device_try_release(struct iser_device *device)
 	mutex_unlock(&ig.device_list_mutex);
 }
 
-static int iser_conn_state_comp_exch(struct iser_conn *ib_conn,
-				     enum iser_ib_conn_state comp,
-				     enum iser_ib_conn_state exch)
+/**
+ * Called with state mutex held
+ **/
+static int iser_conn_state_comp_exch(struct iser_conn *iser_conn,
+				     enum iser_conn_state comp,
+				     enum iser_conn_state exch)
 {
 	int ret;
 
-	spin_lock_bh(&ib_conn->lock);
-	if ((ret = (ib_conn->state == comp)))
-		ib_conn->state = exch;
-	spin_unlock_bh(&ib_conn->lock);
+	ret = (iser_conn->state == comp);
+	if (ret)
+		iser_conn->state = exch;
+
 	return ret;
 }
 
 void iser_release_work(struct work_struct *work)
 {
-	struct iser_conn *ib_conn;
+	struct iser_conn *iser_conn;
 
-	ib_conn = container_of(work, struct iser_conn, release_work);
+	iser_conn = container_of(work, struct iser_conn, release_work);
 
-	/* wait for .conn_stop callback */
-	wait_for_completion(&ib_conn->stop_completion);
+	/* Wait for conn_stop to complete */
+	wait_for_completion(&iser_conn->stop_completion);
+	/* Wait for IB resouces cleanup to complete */
+	wait_for_completion(&iser_conn->ib_completion);
 
-	/* wait for the qp`s post send and post receive buffers to empty */
-	wait_event_interruptible(ib_conn->wait,
-				 ib_conn->state == ISER_CONN_DOWN);
+	mutex_lock(&iser_conn->state_mutex);
+	iser_conn->state = ISER_CONN_DOWN;
+	mutex_unlock(&iser_conn->state_mutex);
 
-	iser_conn_release(ib_conn);
+	iser_conn_release(iser_conn);
 }
 
 /**
- * Frees all conn objects and deallocs conn descriptor
+ * iser_free_ib_conn_res - release IB related resources
+ * @iser_conn: iser connection struct
+ * @destroy_device: indicator if we need to try to release
+ *     the iser device (only iscsi shutdown and DEVICE_REMOVAL
+ *     will use this.
+ *
+ * This routine is called with the iser state mutex held
+ * so the cm_id removal is out of here. It is Safe to
+ * be invoked multiple times.
  */
-void iser_conn_release(struct iser_conn *ib_conn)
+static void iser_free_ib_conn_res(struct iser_conn *iser_conn,
+				  bool destroy_device)
 {
-	struct iser_device  *device = ib_conn->device;
+	struct ib_conn *ib_conn = &iser_conn->ib_conn;
+	struct iser_device *device = ib_conn->device;
+
+	iser_info("freeing conn %p cma_id %p qp %p\n",
+		  iser_conn, ib_conn->cma_id, ib_conn->qp);
+
+	iser_free_rx_descriptors(iser_conn);
+
+	if (ib_conn->qp != NULL) {
+		ib_conn->comp->active_qps--;
+		rdma_destroy_qp(ib_conn->cma_id);
+		ib_conn->qp = NULL;
+	}
+
+	if (destroy_device && device != NULL) {
+		iser_device_try_release(device);
+		ib_conn->device = NULL;
+	}
+}
 
-	BUG_ON(ib_conn->state == ISER_CONN_UP);
+/**
+ * Frees all conn objects and deallocs conn descriptor
+ */
+void iser_conn_release(struct iser_conn *iser_conn)
+{
+	struct ib_conn *ib_conn = &iser_conn->ib_conn;
 
 	mutex_lock(&ig.connlist_mutex);
-	list_del(&ib_conn->conn_list);
+	list_del(&iser_conn->conn_list);
 	mutex_unlock(&ig.connlist_mutex);
-	iser_free_rx_descriptors(ib_conn);
-	iser_free_ib_conn_res(ib_conn);
-	ib_conn->device = NULL;
-	/* on EVENT_ADDR_ERROR there's no device yet for this conn */
-	if (device != NULL)
-		iser_device_try_release(device);
-	/* if cma handler context, the caller actually destroy the id */
+
+	mutex_lock(&iser_conn->state_mutex);
+	if (iser_conn->state != ISER_CONN_DOWN)
+		iser_warn("iser conn %p state %d, expected state down.\n",
+			  iser_conn, iser_conn->state);
+	/*
+	 * In case we never got to bind stage, we still need to
+	 * release IB resources (which is safe to call more than once).
+	 */
+	iser_free_ib_conn_res(iser_conn, true);
+	mutex_unlock(&iser_conn->state_mutex);
+
 	if (ib_conn->cma_id != NULL) {
 		rdma_destroy_id(ib_conn->cma_id);
 		ib_conn->cma_id = NULL;
 	}
-	iscsi_destroy_endpoint(ib_conn->ep);
+
+	kfree(iser_conn);
 }
 
 /**
  * triggers start of the disconnect procedures and wait for them to be done
+ * Called with state mutex held
  */
-void iser_conn_terminate(struct iser_conn *ib_conn)
+int iser_conn_terminate(struct iser_conn *iser_conn)
 {
+	struct ib_conn *ib_conn = &iser_conn->ib_conn;
+	struct ib_send_wr *bad_wr;
 	int err = 0;
 
-	/* change the ib conn state only if the conn is UP, however always call
-	 * rdma_disconnect since this is the only way to cause the CMA to change
-	 * the QP state to ERROR
+	/* terminate the iser conn only if the conn state is UP */
+	if (!iser_conn_state_comp_exch(iser_conn, ISER_CONN_UP,
+				       ISER_CONN_TERMINATING))
+		return 0;
+
+	iser_info("iser_conn %p state %d\n", iser_conn, iser_conn->state);
+
+	/* suspend queuing of new iscsi commands */
+	if (iser_conn->iscsi_conn)
+		iscsi_suspend_queue(iser_conn->iscsi_conn);
+
+	/*
+	 * In case we didn't already clean up the cma_id (peer initiated
+	 * a disconnection), we need to Cause the CMA to change the QP
+	 * state to ERROR.
 	 */
+	if (ib_conn->cma_id) {
+		err = rdma_disconnect(ib_conn->cma_id);
+		if (err)
+			iser_err("Failed to disconnect, conn: 0x%p err %d\n",
+				 iser_conn, err);
+
+		/* post an indication that all flush errors were consumed */
+		err = ib_post_send(ib_conn->qp, &ib_conn->beacon, &bad_wr);
+		if (err)
+			iser_err("conn %p failed to post beacon", ib_conn);
+
+		wait_for_completion(&ib_conn->flush_comp);
+	}
 
-	iser_conn_state_comp_exch(ib_conn, ISER_CONN_UP, ISER_CONN_TERMINATING);
-	err = rdma_disconnect(ib_conn->cma_id);
-	if (err)
-		iser_err("Failed to disconnect, conn: 0x%p err %d\n",
-			 ib_conn,err);
+	return 1;
 }
 
+/**
+ * Called with state mutex held
+ **/
 static void iser_connect_error(struct rdma_cm_id *cma_id)
 {
-	struct iser_conn *ib_conn;
-
-	ib_conn = (struct iser_conn *)cma_id->context;
+	struct iser_conn *iser_conn;
 
-	ib_conn->state = ISER_CONN_DOWN;
-	wake_up_interruptible(&ib_conn->wait);
+	iser_conn = (struct iser_conn *)cma_id->context;
+	iser_conn->state = ISER_CONN_DOWN;
 }
 
+/**
+ * Called with state mutex held
+ **/
 static void iser_addr_handler(struct rdma_cm_id *cma_id)
 {
 	struct iser_device *device;
-	struct iser_conn   *ib_conn;
+	struct iser_conn   *iser_conn;
+	struct ib_conn   *ib_conn;
 	int    ret;
 
+	iser_conn = (struct iser_conn *)cma_id->context;
+	if (iser_conn->state != ISER_CONN_PENDING)
+		/* bailout */
+		return;
+
+	ib_conn = &iser_conn->ib_conn;
 	device = iser_device_find_by_ib_device(cma_id);
 	if (!device) {
 		iser_err("device lookup/creation failed\n");
@@ -665,7 +705,6 @@ static void iser_addr_handler(struct rdma_cm_id *cma_id)
 		return;
 	}
 
-	ib_conn = (struct iser_conn *)cma_id->context;
 	ib_conn->device = device;
 
 	/* connection T10-PI support */
@@ -689,18 +728,28 @@ static void iser_addr_handler(struct rdma_cm_id *cma_id)
 	}
 }
 
+/**
+ * Called with state mutex held
+ **/
 static void iser_route_handler(struct rdma_cm_id *cma_id)
 {
 	struct rdma_conn_param conn_param;
 	int    ret;
 	struct iser_cm_hdr req_hdr;
+	struct iser_conn *iser_conn = (struct iser_conn *)cma_id->context;
+	struct ib_conn *ib_conn = &iser_conn->ib_conn;
+	struct iser_device *device = ib_conn->device;
 
-	ret = iser_create_ib_conn_res((struct iser_conn *)cma_id->context);
+	if (iser_conn->state != ISER_CONN_PENDING)
+		/* bailout */
+		return;
+
+	ret = iser_create_ib_conn_res(ib_conn);
 	if (ret)
 		goto failure;
 
 	memset(&conn_param, 0, sizeof conn_param);
-	conn_param.responder_resources = 4;
+	conn_param.responder_resources = device->dev_attr.max_qp_rd_atom;
 	conn_param.initiator_depth     = 1;
 	conn_param.retry_count	       = 7;
 	conn_param.rnr_retry_count     = 6;
@@ -724,47 +773,60 @@ failure:
 
 static void iser_connected_handler(struct rdma_cm_id *cma_id)
 {
-	struct iser_conn *ib_conn;
+	struct iser_conn *iser_conn;
 	struct ib_qp_attr attr;
 	struct ib_qp_init_attr init_attr;
 
+	iser_conn = (struct iser_conn *)cma_id->context;
+	if (iser_conn->state != ISER_CONN_PENDING)
+		/* bailout */
+		return;
+
 	(void)ib_query_qp(cma_id->qp, &attr, ~0, &init_attr);
 	iser_info("remote qpn:%x my qpn:%x\n", attr.dest_qp_num, cma_id->qp->qp_num);
 
-	ib_conn = (struct iser_conn *)cma_id->context;
-	if (iser_conn_state_comp_exch(ib_conn, ISER_CONN_PENDING, ISER_CONN_UP))
-		wake_up_interruptible(&ib_conn->wait);
+	iser_conn->state = ISER_CONN_UP;
+	complete(&iser_conn->up_completion);
 }
 
 static void iser_disconnected_handler(struct rdma_cm_id *cma_id)
 {
-	struct iser_conn *ib_conn;
+	struct iser_conn *iser_conn = (struct iser_conn *)cma_id->context;
 
-	ib_conn = (struct iser_conn *)cma_id->context;
-
-	/* getting here when the state is UP means that the conn is being *
-	 * terminated asynchronously from the iSCSI layer's perspective.  */
-	if (iser_conn_state_comp_exch(ib_conn, ISER_CONN_UP,
-					ISER_CONN_TERMINATING)){
-		if (ib_conn->iscsi_conn)
-			iscsi_conn_failure(ib_conn->iscsi_conn, ISCSI_ERR_CONN_FAILED);
+	if (iser_conn_terminate(iser_conn)) {
+		if (iser_conn->iscsi_conn)
+			iscsi_conn_failure(iser_conn->iscsi_conn,
+					   ISCSI_ERR_CONN_FAILED);
 		else
 			iser_err("iscsi_iser connection isn't bound\n");
 	}
-
-	/* Complete the termination process if no posts are pending */
-	if (ib_conn->post_recv_buf_count == 0 &&
-	    (atomic_read(&ib_conn->post_send_buf_count) == 0)) {
-		ib_conn->state = ISER_CONN_DOWN;
-		wake_up_interruptible(&ib_conn->wait);
-	}
 }
 
+static void iser_cleanup_handler(struct rdma_cm_id *cma_id,
+				 bool destroy_device)
+{
+	struct iser_conn *iser_conn = (struct iser_conn *)cma_id->context;
+
+	/*
+	 * We are not guaranteed that we visited disconnected_handler
+	 * by now, call it here to be safe that we handle CM drep
+	 * and flush errors.
+	 */
+	iser_disconnected_handler(cma_id);
+	iser_free_ib_conn_res(iser_conn, destroy_device);
+	complete(&iser_conn->ib_completion);
+};
+
 static int iser_cma_handler(struct rdma_cm_id *cma_id, struct rdma_cm_event *event)
 {
+	struct iser_conn *iser_conn;
+	int ret = 0;
+
+	iser_conn = (struct iser_conn *)cma_id->context;
 	iser_info("event %d status %d conn %p id %p\n",
 		  event->event, event->status, cma_id->context, cma_id);
 
+	mutex_lock(&iser_conn->state_mutex);
 	switch (event->event) {
 	case RDMA_CM_EVENT_ADDR_RESOLVED:
 		iser_addr_handler(cma_id);
@@ -783,89 +845,107 @@ static int iser_cma_handler(struct rdma_cm_id *cma_id, struct rdma_cm_event *eve
 		iser_connect_error(cma_id);
 		break;
 	case RDMA_CM_EVENT_DISCONNECTED:
-	case RDMA_CM_EVENT_DEVICE_REMOVAL:
 	case RDMA_CM_EVENT_ADDR_CHANGE:
 		iser_disconnected_handler(cma_id);
 		break;
+	case RDMA_CM_EVENT_DEVICE_REMOVAL:
+		/*
+		 * we *must* destroy the device as we cannot rely
+		 * on iscsid to be around to initiate error handling.
+		 * also implicitly destroy the cma_id.
+		 */
+		iser_cleanup_handler(cma_id, true);
+		iser_conn->ib_conn.cma_id = NULL;
+		ret = 1;
+		break;
+	case RDMA_CM_EVENT_TIMEWAIT_EXIT:
+		iser_cleanup_handler(cma_id, false);
+		break;
 	default:
 		iser_err("Unexpected RDMA CM event (%d)\n", event->event);
 		break;
 	}
-	return 0;
+	mutex_unlock(&iser_conn->state_mutex);
+
+	return ret;
 }
 
-void iser_conn_init(struct iser_conn *ib_conn)
+void iser_conn_init(struct iser_conn *iser_conn)
 {
-	ib_conn->state = ISER_CONN_INIT;
-	init_waitqueue_head(&ib_conn->wait);
-	ib_conn->post_recv_buf_count = 0;
-	atomic_set(&ib_conn->post_send_buf_count, 0);
-	init_completion(&ib_conn->stop_completion);
-	INIT_LIST_HEAD(&ib_conn->conn_list);
-	spin_lock_init(&ib_conn->lock);
+	iser_conn->state = ISER_CONN_INIT;
+	iser_conn->ib_conn.post_recv_buf_count = 0;
+	init_completion(&iser_conn->ib_conn.flush_comp);
+	init_completion(&iser_conn->stop_completion);
+	init_completion(&iser_conn->ib_completion);
+	init_completion(&iser_conn->up_completion);
+	INIT_LIST_HEAD(&iser_conn->conn_list);
+	spin_lock_init(&iser_conn->ib_conn.lock);
+	mutex_init(&iser_conn->state_mutex);
 }
 
  /**
  * starts the process of connecting to the target
  * sleeps until the connection is established or rejected
  */
-int iser_connect(struct iser_conn   *ib_conn,
-		 struct sockaddr_in *src_addr,
-		 struct sockaddr_in *dst_addr,
+int iser_connect(struct iser_conn   *iser_conn,
+		 struct sockaddr    *src_addr,
+		 struct sockaddr    *dst_addr,
 		 int                 non_blocking)
 {
-	struct sockaddr *src, *dst;
+	struct ib_conn *ib_conn = &iser_conn->ib_conn;
 	int err = 0;
 
-	sprintf(ib_conn->name, "%pI4:%d",
-		&dst_addr->sin_addr.s_addr, dst_addr->sin_port);
+	mutex_lock(&iser_conn->state_mutex);
+
+	sprintf(iser_conn->name, "%pISp", dst_addr);
+
+	iser_info("connecting to: %s\n", iser_conn->name);
 
 	/* the device is known only --after-- address resolution */
 	ib_conn->device = NULL;
 
-	iser_info("connecting to: %pI4, port 0x%x\n",
-		  &dst_addr->sin_addr, dst_addr->sin_port);
+	iser_conn->state = ISER_CONN_PENDING;
 
-	ib_conn->state = ISER_CONN_PENDING;
+	ib_conn->beacon.wr_id = ISER_BEACON_WRID;
+	ib_conn->beacon.opcode = IB_WR_SEND;
 
 	ib_conn->cma_id = rdma_create_id(iser_cma_handler,
-					     (void *)ib_conn,
-					     RDMA_PS_TCP, IB_QPT_RC);
+					 (void *)iser_conn,
+					 RDMA_PS_TCP, IB_QPT_RC);
 	if (IS_ERR(ib_conn->cma_id)) {
 		err = PTR_ERR(ib_conn->cma_id);
 		iser_err("rdma_create_id failed: %d\n", err);
 		goto id_failure;
 	}
 
-	src = (struct sockaddr *)src_addr;
-	dst = (struct sockaddr *)dst_addr;
-	err = rdma_resolve_addr(ib_conn->cma_id, src, dst, 1000);
+	err = rdma_resolve_addr(ib_conn->cma_id, src_addr, dst_addr, 1000);
 	if (err) {
 		iser_err("rdma_resolve_addr failed: %d\n", err);
 		goto addr_failure;
 	}
 
 	if (!non_blocking) {
-		wait_event_interruptible(ib_conn->wait,
-					 (ib_conn->state != ISER_CONN_PENDING));
+		wait_for_completion_interruptible(&iser_conn->up_completion);
 
-		if (ib_conn->state != ISER_CONN_UP) {
+		if (iser_conn->state != ISER_CONN_UP) {
 			err =  -EIO;
 			goto connect_failure;
 		}
 	}
+	mutex_unlock(&iser_conn->state_mutex);
 
 	mutex_lock(&ig.connlist_mutex);
-	list_add(&ib_conn->conn_list, &ig.connlist);
+	list_add(&iser_conn->conn_list, &ig.connlist);
 	mutex_unlock(&ig.connlist_mutex);
 	return 0;
 
 id_failure:
 	ib_conn->cma_id = NULL;
 addr_failure:
-	ib_conn->state = ISER_CONN_DOWN;
+	iser_conn->state = ISER_CONN_DOWN;
 connect_failure:
-	iser_conn_release(ib_conn);
+	mutex_unlock(&iser_conn->state_mutex);
+	iser_conn_release(iser_conn);
 	return err;
 }
 
@@ -874,7 +954,7 @@ connect_failure:
  *
  * returns: 0 on success, errno code on failure
  */
-int iser_reg_page_vec(struct iser_conn     *ib_conn,
+int iser_reg_page_vec(struct ib_conn *ib_conn,
 		      struct iser_page_vec *page_vec,
 		      struct iser_mem_reg  *mem_reg)
 {
@@ -944,7 +1024,8 @@ void iser_unreg_mem_fastreg(struct iscsi_iser_task *iser_task,
 			    enum iser_data_dir cmd_dir)
 {
 	struct iser_mem_reg *reg = &iser_task->rdma_regd[cmd_dir].reg;
-	struct iser_conn *ib_conn = iser_task->ib_conn;
+	struct iser_conn *iser_conn = iser_task->iser_conn;
+	struct ib_conn *ib_conn = &iser_conn->ib_conn;
 	struct fast_reg_descriptor *desc = reg->mem_h;
 
 	if (!reg->is_mr)
@@ -957,17 +1038,18 @@ void iser_unreg_mem_fastreg(struct iscsi_iser_task *iser_task,
 	spin_unlock_bh(&ib_conn->lock);
 }
 
-int iser_post_recvl(struct iser_conn *ib_conn)
+int iser_post_recvl(struct iser_conn *iser_conn)
 {
 	struct ib_recv_wr rx_wr, *rx_wr_failed;
+	struct ib_conn *ib_conn = &iser_conn->ib_conn;
 	struct ib_sge	  sge;
 	int ib_ret;
 
-	sge.addr   = ib_conn->login_resp_dma;
+	sge.addr   = iser_conn->login_resp_dma;
 	sge.length = ISER_RX_LOGIN_SIZE;
 	sge.lkey   = ib_conn->device->mr->lkey;
 
-	rx_wr.wr_id   = (unsigned long)ib_conn->login_resp_buf;
+	rx_wr.wr_id   = (unsigned long)iser_conn->login_resp_buf;
 	rx_wr.sg_list = &sge;
 	rx_wr.num_sge = 1;
 	rx_wr.next    = NULL;
@@ -981,20 +1063,21 @@ int iser_post_recvl(struct iser_conn *ib_conn)
 	return ib_ret;
 }
 
-int iser_post_recvm(struct iser_conn *ib_conn, int count)
+int iser_post_recvm(struct iser_conn *iser_conn, int count)
 {
 	struct ib_recv_wr *rx_wr, *rx_wr_failed;
 	int i, ib_ret;
-	unsigned int my_rx_head = ib_conn->rx_desc_head;
+	struct ib_conn *ib_conn = &iser_conn->ib_conn;
+	unsigned int my_rx_head = iser_conn->rx_desc_head;
 	struct iser_rx_desc *rx_desc;
 
 	for (rx_wr = ib_conn->rx_wr, i = 0; i < count; i++, rx_wr++) {
-		rx_desc		= &ib_conn->rx_descs[my_rx_head];
+		rx_desc		= &iser_conn->rx_descs[my_rx_head];
 		rx_wr->wr_id	= (unsigned long)rx_desc;
 		rx_wr->sg_list	= &rx_desc->rx_sg;
 		rx_wr->num_sge	= 1;
 		rx_wr->next	= rx_wr + 1;
-		my_rx_head = (my_rx_head + 1) & ib_conn->qp_max_recv_dtos_mask;
+		my_rx_head = (my_rx_head + 1) & iser_conn->qp_max_recv_dtos_mask;
 	}
 
 	rx_wr--;
@@ -1006,7 +1089,7 @@ int iser_post_recvm(struct iser_conn *ib_conn, int count)
 		iser_err("ib_post_recv failed ret=%d\n", ib_ret);
 		ib_conn->post_recv_buf_count -= count;
 	} else
-		ib_conn->rx_desc_head = my_rx_head;
+		iser_conn->rx_desc_head = my_rx_head;
 	return ib_ret;
 }
 
@@ -1016,138 +1099,166 @@ int iser_post_recvm(struct iser_conn *ib_conn, int count)
  *
  * returns 0 on success, -1 on failure
  */
-int iser_post_send(struct iser_conn *ib_conn, struct iser_tx_desc *tx_desc)
+int iser_post_send(struct ib_conn *ib_conn, struct iser_tx_desc *tx_desc,
+		   bool signal)
 {
 	int		  ib_ret;
 	struct ib_send_wr send_wr, *send_wr_failed;
 
 	ib_dma_sync_single_for_device(ib_conn->device->ib_device,
-		tx_desc->dma_addr, ISER_HEADERS_LEN, DMA_TO_DEVICE);
+				      tx_desc->dma_addr, ISER_HEADERS_LEN,
+				      DMA_TO_DEVICE);
 
 	send_wr.next	   = NULL;
 	send_wr.wr_id	   = (unsigned long)tx_desc;
 	send_wr.sg_list	   = tx_desc->tx_sg;
 	send_wr.num_sge	   = tx_desc->num_sge;
 	send_wr.opcode	   = IB_WR_SEND;
-	send_wr.send_flags = IB_SEND_SIGNALED;
-
-	atomic_inc(&ib_conn->post_send_buf_count);
+	send_wr.send_flags = signal ? IB_SEND_SIGNALED : 0;
 
 	ib_ret = ib_post_send(ib_conn->qp, &send_wr, &send_wr_failed);
-	if (ib_ret) {
+	if (ib_ret)
 		iser_err("ib_post_send failed, ret:%d\n", ib_ret);
-		atomic_dec(&ib_conn->post_send_buf_count);
-	}
+
 	return ib_ret;
 }
 
-static void iser_handle_comp_error(struct iser_tx_desc *desc,
-				struct iser_conn *ib_conn)
+/**
+ * is_iser_tx_desc - Indicate if the completion wr_id
+ *     is a TX descriptor or not.
+ * @iser_conn: iser connection
+ * @wr_id: completion WR identifier
+ *
+ * Since we cannot rely on wc opcode in FLUSH errors
+ * we must work around it by checking if the wr_id address
+ * falls in the iser connection rx_descs buffer. If so
+ * it is an RX descriptor, otherwize it is a TX.
+ */
+static inline bool
+is_iser_tx_desc(struct iser_conn *iser_conn, void *wr_id)
+{
+	void *start = iser_conn->rx_descs;
+	int len = iser_conn->num_rx_descs * sizeof(*iser_conn->rx_descs);
+
+	if (wr_id >= start && wr_id < start + len)
+		return false;
+
+	return true;
+}
+
+/**
+ * iser_handle_comp_error() - Handle error completion
+ * @ib_conn:   connection RDMA resources
+ * @wc:        work completion
+ *
+ * Notes: We may handle a FLUSH error completion and in this case
+ *        we only cleanup in case TX type was DATAOUT. For non-FLUSH
+ *        error completion we should also notify iscsi layer that
+ *        connection is failed (in case we passed bind stage).
+ */
+static void
+iser_handle_comp_error(struct ib_conn *ib_conn,
+		       struct ib_wc *wc)
 {
-	if (desc && desc->type == ISCSI_TX_DATAOUT)
-		kmem_cache_free(ig.desc_cache, desc);
-
-	if (ib_conn->post_recv_buf_count == 0 &&
-	    atomic_read(&ib_conn->post_send_buf_count) == 0) {
-		/* getting here when the state is UP means that the conn is *
-		 * being terminated asynchronously from the iSCSI layer's   *
-		 * perspective.                                             */
-		if (iser_conn_state_comp_exch(ib_conn, ISER_CONN_UP,
-		    ISER_CONN_TERMINATING))
-			iscsi_conn_failure(ib_conn->iscsi_conn,
+	struct iser_conn *iser_conn = container_of(ib_conn, struct iser_conn,
+						   ib_conn);
+
+	if (wc->status != IB_WC_WR_FLUSH_ERR)
+		if (iser_conn->iscsi_conn)
+			iscsi_conn_failure(iser_conn->iscsi_conn,
 					   ISCSI_ERR_CONN_FAILED);
 
-		/* no more non completed posts to the QP, complete the
-		 * termination process w.o worrying on disconnect event */
-		ib_conn->state = ISER_CONN_DOWN;
-		wake_up_interruptible(&ib_conn->wait);
+	if (is_iser_tx_desc(iser_conn, (void *)wc->wr_id)) {
+		struct iser_tx_desc *desc = (struct iser_tx_desc *)wc->wr_id;
+
+		if (desc->type == ISCSI_TX_DATAOUT)
+			kmem_cache_free(ig.desc_cache, desc);
+	} else {
+		ib_conn->post_recv_buf_count--;
 	}
 }
 
-static int iser_drain_tx_cq(struct iser_device  *device, int cq_index)
+/**
+ * iser_handle_wc - handle a single work completion
+ * @wc: work completion
+ *
+ * Soft-IRQ context, work completion can be either
+ * SEND or RECV, and can turn out successful or
+ * with error (or flush error).
+ */
+static void iser_handle_wc(struct ib_wc *wc)
 {
-	struct ib_cq  *cq = device->tx_cq[cq_index];
-	struct ib_wc  wc;
+	struct ib_conn *ib_conn;
 	struct iser_tx_desc *tx_desc;
-	struct iser_conn *ib_conn;
-	int completed_tx = 0;
-
-	while (ib_poll_cq(cq, 1, &wc) == 1) {
-		tx_desc	= (struct iser_tx_desc *) (unsigned long) wc.wr_id;
-		ib_conn = wc.qp->qp_context;
-		if (wc.status == IB_WC_SUCCESS) {
-			if (wc.opcode == IB_WC_SEND)
-				iser_snd_completion(tx_desc, ib_conn);
-			else
-				iser_err("expected opcode %d got %d\n",
-					IB_WC_SEND, wc.opcode);
+	struct iser_rx_desc *rx_desc;
+
+	ib_conn = wc->qp->qp_context;
+	if (wc->status == IB_WC_SUCCESS) {
+		if (wc->opcode == IB_WC_RECV) {
+			rx_desc = (struct iser_rx_desc *)wc->wr_id;
+			iser_rcv_completion(rx_desc, wc->byte_len,
+					    ib_conn);
+		} else
+		if (wc->opcode == IB_WC_SEND) {
+			tx_desc = (struct iser_tx_desc *)wc->wr_id;
+			iser_snd_completion(tx_desc, ib_conn);
 		} else {
-			iser_err("tx id %llx status %d vend_err %x\n",
-				 wc.wr_id, wc.status, wc.vendor_err);
-			if (wc.wr_id != ISER_FASTREG_LI_WRID) {
-				atomic_dec(&ib_conn->post_send_buf_count);
-				iser_handle_comp_error(tx_desc, ib_conn);
-			}
+			iser_err("Unknown wc opcode %d\n", wc->opcode);
 		}
-		completed_tx++;
+	} else {
+		if (wc->status != IB_WC_WR_FLUSH_ERR)
+			iser_err("wr id %llx status %d vend_err %x\n",
+				 wc->wr_id, wc->status, wc->vendor_err);
+		else
+			iser_dbg("flush error: wr id %llx\n", wc->wr_id);
+
+		if (wc->wr_id != ISER_FASTREG_LI_WRID &&
+		    wc->wr_id != ISER_BEACON_WRID)
+			iser_handle_comp_error(ib_conn, wc);
+
+		/* complete in case all flush errors were consumed */
+		if (wc->wr_id == ISER_BEACON_WRID)
+			complete(&ib_conn->flush_comp);
 	}
-	return completed_tx;
 }
 
-
+/**
+ * iser_cq_tasklet_fn - iSER completion polling loop
+ * @data: iSER completion context
+ *
+ * Soft-IRQ context, polling connection CQ until
+ * either CQ was empty or we exausted polling budget
+ */
 static void iser_cq_tasklet_fn(unsigned long data)
 {
-	struct iser_cq_desc *cq_desc = (struct iser_cq_desc *)data;
-	struct iser_device  *device = cq_desc->device;
-	int cq_index = cq_desc->cq_index;
-	struct ib_cq	     *cq = device->rx_cq[cq_index];
-	 struct ib_wc	     wc;
-	 struct iser_rx_desc *desc;
-	 unsigned long	     xfer_len;
-	struct iser_conn *ib_conn;
-	int completed_tx, completed_rx = 0;
-
-	/* First do tx drain, so in a case where we have rx flushes and a successful
-	 * tx completion we will still go through completion error handling.
-	 */
-	completed_tx = iser_drain_tx_cq(device, cq_index);
-
-	while (ib_poll_cq(cq, 1, &wc) == 1) {
-		desc	 = (struct iser_rx_desc *) (unsigned long) wc.wr_id;
-		BUG_ON(desc == NULL);
-		ib_conn = wc.qp->qp_context;
-		if (wc.status == IB_WC_SUCCESS) {
-			if (wc.opcode == IB_WC_RECV) {
-				xfer_len = (unsigned long)wc.byte_len;
-				iser_rcv_completion(desc, xfer_len, ib_conn);
-			} else
-				iser_err("expected opcode %d got %d\n",
-					IB_WC_RECV, wc.opcode);
-		} else {
-			if (wc.status != IB_WC_WR_FLUSH_ERR)
-				iser_err("rx id %llx status %d vend_err %x\n",
-					wc.wr_id, wc.status, wc.vendor_err);
-			ib_conn->post_recv_buf_count--;
-			iser_handle_comp_error(NULL, ib_conn);
-		}
-		completed_rx++;
-		if (!(completed_rx & 63))
-			completed_tx += iser_drain_tx_cq(device, cq_index);
+	struct iser_comp *comp = (struct iser_comp *)data;
+	struct ib_cq *cq = comp->cq;
+	struct ib_wc *const wcs = comp->wcs;
+	int i, n, completed = 0;
+
+	while ((n = ib_poll_cq(cq, ARRAY_SIZE(comp->wcs), wcs)) > 0) {
+		for (i = 0; i < n; i++)
+			iser_handle_wc(&wcs[i]);
+
+		completed += n;
+		if (completed >= iser_cq_poll_limit)
+			break;
 	}
-	/* #warning "it is assumed here that arming CQ only once its empty" *
-	 * " would not cause interrupts to be missed"                       */
+
+	/*
+	 * It is assumed here that arming CQ only once its empty
+	 * would not cause interrupts to be missed.
+	 */
 	ib_req_notify_cq(cq, IB_CQ_NEXT_COMP);
 
-	iser_dbg("got %d rx %d tx completions\n", completed_rx, completed_tx);
+	iser_dbg("got %d completions\n", completed);
 }
 
 static void iser_cq_callback(struct ib_cq *cq, void *cq_context)
 {
-	struct iser_cq_desc *cq_desc = (struct iser_cq_desc *)cq_context;
-	struct iser_device  *device = cq_desc->device;
-	int cq_index = cq_desc->cq_index;
+	struct iser_comp *comp = cq_context;
 
-	tasklet_schedule(&device->cq_tasklet[cq_index]);
+	tasklet_schedule(&comp->tasklet);
 }
 
 u8 iser_check_task_pi_status(struct iscsi_iser_task *iser_task,
diff --git a/drivers/infiniband/ulp/isert/ib_isert.c b/drivers/infiniband/ulp/isert/ib_isert.c
index d4c7928a0f36..3effa931fce2 100644
--- a/drivers/infiniband/ulp/isert/ib_isert.c
+++ b/drivers/infiniband/ulp/isert/ib_isert.c
@@ -586,17 +586,12 @@ isert_connect_request(struct rdma_cm_id *cma_id, struct rdma_cm_event *event)
 	init_completion(&isert_conn->conn_wait);
 	init_completion(&isert_conn->conn_wait_comp_err);
 	kref_init(&isert_conn->conn_kref);
-	kref_get(&isert_conn->conn_kref);
 	mutex_init(&isert_conn->conn_mutex);
 	spin_lock_init(&isert_conn->conn_lock);
 	INIT_LIST_HEAD(&isert_conn->conn_fr_pool);
 
 	cma_id->context = isert_conn;
 	isert_conn->conn_cm_id = cma_id;
-	isert_conn->responder_resources = event->param.conn.responder_resources;
-	isert_conn->initiator_depth = event->param.conn.initiator_depth;
-	pr_debug("Using responder_resources: %u initiator_depth: %u\n",
-		 isert_conn->responder_resources, isert_conn->initiator_depth);
 
 	isert_conn->login_buf = kzalloc(ISCSI_DEF_MAX_RECV_SEG_LEN +
 					ISER_RX_LOGIN_SIZE, GFP_KERNEL);
@@ -643,6 +638,12 @@ isert_connect_request(struct rdma_cm_id *cma_id, struct rdma_cm_event *event)
 		goto out_rsp_dma_map;
 	}
 
+	/* Set max inflight RDMA READ requests */
+	isert_conn->initiator_depth = min_t(u8,
+				event->param.conn.initiator_depth,
+				device->dev_attr.max_qp_init_rd_atom);
+	pr_debug("Using initiator_depth: %u\n", isert_conn->initiator_depth);
+
 	isert_conn->conn_device = device;
 	isert_conn->conn_pd = ib_alloc_pd(isert_conn->conn_device->ib_device);
 	if (IS_ERR(isert_conn->conn_pd)) {
@@ -746,7 +747,9 @@ isert_connect_release(struct isert_conn *isert_conn)
 static void
 isert_connected_handler(struct rdma_cm_id *cma_id)
 {
-	return;
+	struct isert_conn *isert_conn = cma_id->context;
+
+	kref_get(&isert_conn->conn_kref);
 }
 
 static void
@@ -798,7 +801,6 @@ isert_disconnect_work(struct work_struct *work)
 
 wake_up:
 	complete(&isert_conn->conn_wait);
-	isert_put_conn(isert_conn);
 }
 
 static void
@@ -2183,7 +2185,7 @@ isert_put_response(struct iscsi_conn *conn, struct iscsi_cmd *cmd)
 		isert_cmd->tx_desc.num_sge = 2;
 	}
 
-	isert_init_send_wr(isert_conn, isert_cmd, send_wr, true);
+	isert_init_send_wr(isert_conn, isert_cmd, send_wr, false);
 
 	pr_debug("Posting SCSI Response IB_WR_SEND >>>>>>>>>>>>>>>>>>>>>>\n");
 
@@ -2607,58 +2609,45 @@ isert_fast_reg_mr(struct isert_conn *isert_conn,
 	return ret;
 }
 
-static inline enum ib_t10_dif_type
-se2ib_prot_type(enum target_prot_type prot_type)
+static inline void
+isert_set_dif_domain(struct se_cmd *se_cmd, struct ib_sig_attrs *sig_attrs,
+		     struct ib_sig_domain *domain)
 {
-	switch (prot_type) {
-	case TARGET_DIF_TYPE0_PROT:
-		return IB_T10DIF_NONE;
-	case TARGET_DIF_TYPE1_PROT:
-		return IB_T10DIF_TYPE1;
-	case TARGET_DIF_TYPE2_PROT:
-		return IB_T10DIF_TYPE2;
-	case TARGET_DIF_TYPE3_PROT:
-		return IB_T10DIF_TYPE3;
-	default:
-		return IB_T10DIF_NONE;
-	}
-}
+	domain->sig_type = IB_SIG_TYPE_T10_DIF;
+	domain->sig.dif.bg_type = IB_T10DIF_CRC;
+	domain->sig.dif.pi_interval = se_cmd->se_dev->dev_attrib.block_size;
+	domain->sig.dif.ref_tag = se_cmd->reftag_seed;
+	/*
+	 * At the moment we hard code those, but if in the future
+	 * the target core would like to use it, we will take it
+	 * from se_cmd.
+	 */
+	domain->sig.dif.apptag_check_mask = 0xffff;
+	domain->sig.dif.app_escape = true;
+	domain->sig.dif.ref_escape = true;
+	if (se_cmd->prot_type == TARGET_DIF_TYPE1_PROT ||
+	    se_cmd->prot_type == TARGET_DIF_TYPE2_PROT)
+		domain->sig.dif.ref_remap = true;
+};
 
 static int
 isert_set_sig_attrs(struct se_cmd *se_cmd, struct ib_sig_attrs *sig_attrs)
 {
-	enum ib_t10_dif_type ib_prot_type = se2ib_prot_type(se_cmd->prot_type);
-
-	sig_attrs->mem.sig_type = IB_SIG_TYPE_T10_DIF;
-	sig_attrs->wire.sig_type = IB_SIG_TYPE_T10_DIF;
-	sig_attrs->mem.sig.dif.pi_interval =
-				se_cmd->se_dev->dev_attrib.block_size;
-	sig_attrs->wire.sig.dif.pi_interval =
-				se_cmd->se_dev->dev_attrib.block_size;
-
 	switch (se_cmd->prot_op) {
 	case TARGET_PROT_DIN_INSERT:
 	case TARGET_PROT_DOUT_STRIP:
-		sig_attrs->mem.sig.dif.type = IB_T10DIF_NONE;
-		sig_attrs->wire.sig.dif.type = ib_prot_type;
-		sig_attrs->wire.sig.dif.bg_type = IB_T10DIF_CRC;
-		sig_attrs->wire.sig.dif.ref_tag = se_cmd->reftag_seed;
+		sig_attrs->mem.sig_type = IB_SIG_TYPE_NONE;
+		isert_set_dif_domain(se_cmd, sig_attrs, &sig_attrs->wire);
 		break;
 	case TARGET_PROT_DOUT_INSERT:
 	case TARGET_PROT_DIN_STRIP:
-		sig_attrs->mem.sig.dif.type = ib_prot_type;
-		sig_attrs->mem.sig.dif.bg_type = IB_T10DIF_CRC;
-		sig_attrs->mem.sig.dif.ref_tag = se_cmd->reftag_seed;
-		sig_attrs->wire.sig.dif.type = IB_T10DIF_NONE;
+		sig_attrs->wire.sig_type = IB_SIG_TYPE_NONE;
+		isert_set_dif_domain(se_cmd, sig_attrs, &sig_attrs->mem);
 		break;
 	case TARGET_PROT_DIN_PASS:
 	case TARGET_PROT_DOUT_PASS:
-		sig_attrs->mem.sig.dif.type = ib_prot_type;
-		sig_attrs->mem.sig.dif.bg_type = IB_T10DIF_CRC;
-		sig_attrs->mem.sig.dif.ref_tag = se_cmd->reftag_seed;
-		sig_attrs->wire.sig.dif.type = ib_prot_type;
-		sig_attrs->wire.sig.dif.bg_type = IB_T10DIF_CRC;
-		sig_attrs->wire.sig.dif.ref_tag = se_cmd->reftag_seed;
+		isert_set_dif_domain(se_cmd, sig_attrs, &sig_attrs->wire);
+		isert_set_dif_domain(se_cmd, sig_attrs, &sig_attrs->mem);
 		break;
 	default:
 		pr_err("Unsupported PI operation %d\n", se_cmd->prot_op);
@@ -2882,7 +2871,7 @@ isert_put_datain(struct iscsi_conn *conn, struct iscsi_cmd *cmd)
 				     &isert_cmd->tx_desc.iscsi_header);
 		isert_init_tx_hdrs(isert_conn, &isert_cmd->tx_desc);
 		isert_init_send_wr(isert_conn, isert_cmd,
-				   &isert_cmd->tx_desc.send_wr, true);
+				   &isert_cmd->tx_desc.send_wr, false);
 		isert_cmd->rdma_wr.s_send_wr.next = &isert_cmd->tx_desc.send_wr;
 		wr->send_wr_num += 1;
 	}
@@ -3067,7 +3056,6 @@ isert_rdma_accept(struct isert_conn *isert_conn)
 	int ret;
 
 	memset(&cp, 0, sizeof(struct rdma_conn_param));
-	cp.responder_resources = isert_conn->responder_resources;
 	cp.initiator_depth = isert_conn->initiator_depth;
 	cp.retry_count = 7;
 	cp.rnr_retry_count = 7;
@@ -3152,7 +3140,7 @@ isert_accept_np(struct iscsi_np *np, struct iscsi_conn *conn)
 
 accept_wait:
 	ret = down_interruptible(&isert_np->np_sem);
-	if (max_accept > 5)
+	if (ret || max_accept > 5)
 		return -ENODEV;
 
 	spin_lock_bh(&np->np_thread_lock);
@@ -3215,7 +3203,7 @@ static void isert_wait_conn(struct iscsi_conn *conn)
 	pr_debug("isert_wait_conn: Starting \n");
 
 	mutex_lock(&isert_conn->conn_mutex);
-	if (isert_conn->conn_cm_id) {
+	if (isert_conn->conn_cm_id && !isert_conn->disconnect) {
 		pr_debug("Calling rdma_disconnect from isert_wait_conn\n");
 		rdma_disconnect(isert_conn->conn_cm_id);
 	}
@@ -3234,6 +3222,7 @@ static void isert_wait_conn(struct iscsi_conn *conn)
 	wait_for_completion(&isert_conn->conn_wait_comp_err);
 
 	wait_for_completion(&isert_conn->conn_wait);
+	isert_put_conn(isert_conn);
 }
 
 static void isert_free_conn(struct iscsi_conn *conn)
diff --git a/drivers/infiniband/ulp/srp/ib_srp.c b/drivers/infiniband/ulp/srp/ib_srp.c
index e3c2c5b4297f..62d2a18e1b41 100644
--- a/drivers/infiniband/ulp/srp/ib_srp.c
+++ b/drivers/infiniband/ulp/srp/ib_srp.c
@@ -130,6 +130,7 @@ static void srp_send_completion(struct ib_cq *cq, void *target_ptr);
 static int srp_cm_handler(struct ib_cm_id *cm_id, struct ib_cm_event *event);
 
 static struct scsi_transport_template *ib_srp_transport_template;
+static struct workqueue_struct *srp_remove_wq;
 
 static struct ib_client srp_client = {
 	.name   = "srp",
@@ -731,7 +732,7 @@ static bool srp_queue_remove_work(struct srp_target_port *target)
 	spin_unlock_irq(&target->lock);
 
 	if (changed)
-		queue_work(system_long_wq, &target->remove_work);
+		queue_work(srp_remove_wq, &target->remove_work);
 
 	return changed;
 }
@@ -1643,10 +1644,14 @@ static void srp_process_rsp(struct srp_target_port *target, struct srp_rsp *rsp)
 				     SCSI_SENSE_BUFFERSIZE));
 		}
 
-		if (rsp->flags & (SRP_RSP_FLAG_DOOVER | SRP_RSP_FLAG_DOUNDER))
-			scsi_set_resid(scmnd, be32_to_cpu(rsp->data_out_res_cnt));
-		else if (rsp->flags & (SRP_RSP_FLAG_DIOVER | SRP_RSP_FLAG_DIUNDER))
+		if (unlikely(rsp->flags & SRP_RSP_FLAG_DIUNDER))
 			scsi_set_resid(scmnd, be32_to_cpu(rsp->data_in_res_cnt));
+		else if (unlikely(rsp->flags & SRP_RSP_FLAG_DIOVER))
+			scsi_set_resid(scmnd, -be32_to_cpu(rsp->data_in_res_cnt));
+		else if (unlikely(rsp->flags & SRP_RSP_FLAG_DOUNDER))
+			scsi_set_resid(scmnd, be32_to_cpu(rsp->data_out_res_cnt));
+		else if (unlikely(rsp->flags & SRP_RSP_FLAG_DOOVER))
+			scsi_set_resid(scmnd, -be32_to_cpu(rsp->data_out_res_cnt));
 
 		srp_free_req(target, req, scmnd,
 			     be32_to_cpu(rsp->req_lim_delta));
@@ -3261,9 +3266,10 @@ static void srp_remove_one(struct ib_device *device)
 		spin_unlock(&host->target_lock);
 
 		/*
-		 * Wait for target port removal tasks.
+		 * Wait for tl_err and target port removal tasks.
 		 */
 		flush_workqueue(system_long_wq);
+		flush_workqueue(srp_remove_wq);
 
 		kfree(host);
 	}
@@ -3313,16 +3319,22 @@ static int __init srp_init_module(void)
 		indirect_sg_entries = cmd_sg_entries;
 	}
 
+	srp_remove_wq = create_workqueue("srp_remove");
+	if (!srp_remove_wq) {
+		ret = -ENOMEM;
+		goto out;
+	}
+
+	ret = -ENOMEM;
 	ib_srp_transport_template =
 		srp_attach_transport(&ib_srp_transport_functions);
 	if (!ib_srp_transport_template)
-		return -ENOMEM;
+		goto destroy_wq;
 
 	ret = class_register(&srp_class);
 	if (ret) {
 		pr_err("couldn't register class infiniband_srp\n");
-		srp_release_transport(ib_srp_transport_template);
-		return ret;
+		goto release_tr;
 	}
 
 	ib_sa_register_client(&srp_sa_client);
@@ -3330,13 +3342,22 @@ static int __init srp_init_module(void)
 	ret = ib_register_client(&srp_client);
 	if (ret) {
 		pr_err("couldn't register IB client\n");
-		srp_release_transport(ib_srp_transport_template);
-		ib_sa_unregister_client(&srp_sa_client);
-		class_unregister(&srp_class);
-		return ret;
+		goto unreg_sa;
 	}
 
-	return 0;
+out:
+	return ret;
+
+unreg_sa:
+	ib_sa_unregister_client(&srp_sa_client);
+	class_unregister(&srp_class);
+
+release_tr:
+	srp_release_transport(ib_srp_transport_template);
+
+destroy_wq:
+	destroy_workqueue(srp_remove_wq);
+	goto out;
 }
 
 static void __exit srp_cleanup_module(void)
@@ -3345,6 +3366,7 @@ static void __exit srp_cleanup_module(void)
 	ib_sa_unregister_client(&srp_sa_client);
 	class_unregister(&srp_class);
 	srp_release_transport(ib_srp_transport_template);
+	destroy_workqueue(srp_remove_wq);
 }
 
 module_init(srp_init_module);
diff --git a/drivers/infiniband/ulp/srpt/ib_srpt.c b/drivers/infiniband/ulp/srpt/ib_srpt.c
index fe09f2788b15..7206547c13ce 100644
--- a/drivers/infiniband/ulp/srpt/ib_srpt.c
+++ b/drivers/infiniband/ulp/srpt/ib_srpt.c
@@ -198,6 +198,7 @@ static void srpt_event_handler(struct ib_event_handler *handler,
 	case IB_EVENT_PKEY_CHANGE:
 	case IB_EVENT_SM_CHANGE:
 	case IB_EVENT_CLIENT_REREGISTER:
+	case IB_EVENT_GID_CHANGE:
 		/* Refresh port data asynchronously. */
 		if (event->element.port_num <= sdev->device->phys_port_cnt) {
 			sport = &sdev->port[event->element.port_num - 1];
@@ -563,7 +564,7 @@ static int srpt_refresh_port(struct srpt_port *sport)
 							 &reg_req, 0,
 							 srpt_mad_send_handler,
 							 srpt_mad_recv_handler,
-							 sport);
+							 sport, 0);
 		if (IS_ERR(sport->mad_agent)) {
 			ret = PTR_ERR(sport->mad_agent);
 			sport->mad_agent = NULL;
@@ -3573,7 +3574,7 @@ static int srpt_parse_i_port_id(u8 i_port_id[16], const char *name)
 	int ret, rc;
 
 	p = name;
-	if (strnicmp(p, "0x", 2) == 0)
+	if (strncasecmp(p, "0x", 2) == 0)
 		p += 2;
 	ret = -EINVAL;
 	len = strlen(p);