Merge branch 'for-linus' of master.kernel.org:/pub/scm/linux/kernel/git/roland/infiniband

25 files changed, 178 insertions, 124 deletions

diff --git a/drivers/infiniband/core/agent.c b/drivers/infiniband/core/agent.c
index 0c3c6952faa..7545775d38e 100644
--- a/drivers/infiniband/core/agent.c
+++ b/drivers/infiniband/core/agent.c
@@ -155,13 +155,12 @@ int ib_agent_port_open(struct ib_device *device, int port_num)
 	int ret;
 
 	/* Create new device info */
-	port_priv = kmalloc(sizeof *port_priv, GFP_KERNEL);
+	port_priv = kzalloc(sizeof *port_priv, GFP_KERNEL);
 	if (!port_priv) {
 		printk(KERN_ERR SPFX "No memory for ib_agent_port_private\n");
 		ret = -ENOMEM;
 		goto error1;
 	}
-	memset(port_priv, 0, sizeof *port_priv);
 
 	/* Obtain send only MAD agent for SMI QP */
 	port_priv->agent[0] = ib_register_mad_agent(device, port_num,
diff --git a/drivers/infiniband/core/cm.c b/drivers/infiniband/core/cm.c
index 580c3a2bb10..02110e00d14 100644
--- a/drivers/infiniband/core/cm.c
+++ b/drivers/infiniband/core/cm.c
@@ -544,11 +544,10 @@ struct ib_cm_id *ib_create_cm_id(struct ib_device *device,
 	struct cm_id_private *cm_id_priv;
 	int ret;
 
-	cm_id_priv = kmalloc(sizeof *cm_id_priv, GFP_KERNEL);
+	cm_id_priv = kzalloc(sizeof *cm_id_priv, GFP_KERNEL);
 	if (!cm_id_priv)
 		return ERR_PTR(-ENOMEM);
 
-	memset(cm_id_priv, 0, sizeof *cm_id_priv);
 	cm_id_priv->id.state = IB_CM_IDLE;
 	cm_id_priv->id.device = device;
 	cm_id_priv->id.cm_handler = cm_handler;
@@ -621,10 +620,9 @@ static struct cm_timewait_info * cm_create_timewait_info(__be32 local_id)
 {
 	struct cm_timewait_info *timewait_info;
 
-	timewait_info = kmalloc(sizeof *timewait_info, GFP_KERNEL);
+	timewait_info = kzalloc(sizeof *timewait_info, GFP_KERNEL);
 	if (!timewait_info)
 		return ERR_PTR(-ENOMEM);
-	memset(timewait_info, 0, sizeof *timewait_info);
 
 	timewait_info->work.local_id = local_id;
 	INIT_WORK(&timewait_info->work.work, cm_work_handler,
diff --git a/drivers/infiniband/core/device.c b/drivers/infiniband/core/device.c
index 5a6e4497640..e169e798354 100644
--- a/drivers/infiniband/core/device.c
+++ b/drivers/infiniband/core/device.c
@@ -161,17 +161,9 @@ static int alloc_name(char *name)
  */
 struct ib_device *ib_alloc_device(size_t size)
 {
-	void *dev;
-
 	BUG_ON(size < sizeof (struct ib_device));
 
-	dev = kmalloc(size, GFP_KERNEL);
-	if (!dev)
-		return NULL;
-
-	memset(dev, 0, size);
-
-	return dev;
+	return kzalloc(size, GFP_KERNEL);
 }
 EXPORT_SYMBOL(ib_alloc_device);
 
diff --git a/drivers/infiniband/core/mad.c b/drivers/infiniband/core/mad.c
index 88f9f8c9eac..3d8175e5f05 100644
--- a/drivers/infiniband/core/mad.c
+++ b/drivers/infiniband/core/mad.c
@@ -255,12 +255,11 @@ struct ib_mad_agent *ib_register_mad_agent(struct ib_device *device,
 	}
 
 	/* Allocate structures */
-	mad_agent_priv = kmalloc(sizeof *mad_agent_priv, GFP_KERNEL);
+	mad_agent_priv = kzalloc(sizeof *mad_agent_priv, GFP_KERNEL);
 	if (!mad_agent_priv) {
 		ret = ERR_PTR(-ENOMEM);
 		goto error1;
 	}
-	memset(mad_agent_priv, 0, sizeof *mad_agent_priv);
 
 	mad_agent_priv->agent.mr = ib_get_dma_mr(port_priv->qp_info[qpn].qp->pd,
 						 IB_ACCESS_LOCAL_WRITE);
@@ -448,14 +447,13 @@ struct ib_mad_agent *ib_register_mad_snoop(struct ib_device *device,
 		goto error1;
 	}
 	/* Allocate structures */
-	mad_snoop_priv = kmalloc(sizeof *mad_snoop_priv, GFP_KERNEL);
+	mad_snoop_priv = kzalloc(sizeof *mad_snoop_priv, GFP_KERNEL);
 	if (!mad_snoop_priv) {
 		ret = ERR_PTR(-ENOMEM);
 		goto error1;
 	}
 
 	/* Now, fill in the various structures */
-	memset(mad_snoop_priv, 0, sizeof *mad_snoop_priv);
 	mad_snoop_priv->qp_info = &port_priv->qp_info[qpn];
 	mad_snoop_priv->agent.device = device;
 	mad_snoop_priv->agent.recv_handler = recv_handler;
@@ -794,10 +792,9 @@ struct ib_mad_send_buf * ib_create_send_mad(struct ib_mad_agent *mad_agent,
 	    (!rmpp_active && buf_size > sizeof(struct ib_mad)))
 		return ERR_PTR(-EINVAL);
 
-	buf = kmalloc(sizeof *mad_send_wr + buf_size, gfp_mask);
+	buf = kzalloc(sizeof *mad_send_wr + buf_size, gfp_mask);
 	if (!buf)
 		return ERR_PTR(-ENOMEM);
-	memset(buf, 0, sizeof *mad_send_wr + buf_size);
 
 	mad_send_wr = buf + buf_size;
 	mad_send_wr->send_buf.mad = buf;
@@ -1039,14 +1036,12 @@ static int method_in_use(struct ib_mad_mgmt_method_table **method,
 static int allocate_method_table(struct ib_mad_mgmt_method_table **method)
 {
 	/* Allocate management method table */
-	*method = kmalloc(sizeof **method, GFP_ATOMIC);
+	*method = kzalloc(sizeof **method, GFP_ATOMIC);
 	if (!*method) {
 		printk(KERN_ERR PFX "No memory for "
 		       "ib_mad_mgmt_method_table\n");
 		return -ENOMEM;
 	}
-	/* Clear management method table */
-	memset(*method, 0, sizeof **method);
 
 	return 0;
 }
@@ -1137,15 +1132,14 @@ static int add_nonoui_reg_req(struct ib_mad_reg_req *mad_reg_req,
 	class = &port_priv->version[mad_reg_req->mgmt_class_version].class;
 	if (!*class) {
 		/* Allocate management class table for "new" class version */
-		*class = kmalloc(sizeof **class, GFP_ATOMIC);
+		*class = kzalloc(sizeof **class, GFP_ATOMIC);
 		if (!*class) {
 			printk(KERN_ERR PFX "No memory for "
 			       "ib_mad_mgmt_class_table\n");
 			ret = -ENOMEM;
 			goto error1;
 		}
-		/* Clear management class table */
-		memset(*class, 0, sizeof(**class));
+
 		/* Allocate method table for this management class */
 		method = &(*class)->method_table[mgmt_class];
 		if ((ret = allocate_method_table(method)))
@@ -1209,25 +1203,24 @@ static int add_oui_reg_req(struct ib_mad_reg_req *mad_reg_req,
 				mad_reg_req->mgmt_class_version].vendor;
 	if (!*vendor_table) {
 		/* Allocate mgmt vendor class table for "new" class version */
-		vendor = kmalloc(sizeof *vendor, GFP_ATOMIC);
+		vendor = kzalloc(sizeof *vendor, GFP_ATOMIC);
 		if (!vendor) {
 			printk(KERN_ERR PFX "No memory for "
 			       "ib_mad_mgmt_vendor_class_table\n");
 			goto error1;
 		}
-		/* Clear management vendor class table */
-		memset(vendor, 0, sizeof(*vendor));
+
 		*vendor_table = vendor;
 	}
 	if (!(*vendor_table)->vendor_class[vclass]) {
 		/* Allocate table for this management vendor class */
-		vendor_class = kmalloc(sizeof *vendor_class, GFP_ATOMIC);
+		vendor_class = kzalloc(sizeof *vendor_class, GFP_ATOMIC);
 		if (!vendor_class) {
 			printk(KERN_ERR PFX "No memory for "
 			       "ib_mad_mgmt_vendor_class\n");
 			goto error2;
 		}
-		memset(vendor_class, 0, sizeof(*vendor_class));
+
 		(*vendor_table)->vendor_class[vclass] = vendor_class;
 	}
 	for (i = 0; i < MAX_MGMT_OUI; i++) {
@@ -2524,12 +2517,12 @@ static int ib_mad_port_open(struct ib_device *device,
 	char name[sizeof "ib_mad123"];
 
 	/* Create new device info */
-	port_priv = kmalloc(sizeof *port_priv, GFP_KERNEL);
+	port_priv = kzalloc(sizeof *port_priv, GFP_KERNEL);
 	if (!port_priv) {
 		printk(KERN_ERR PFX "No memory for ib_mad_port_private\n");
 		return -ENOMEM;
 	}
-	memset(port_priv, 0, sizeof *port_priv);
+
 	port_priv->device = device;
 	port_priv->port_num = port_num;
 	spin_lock_init(&port_priv->reg_lock);
diff --git a/drivers/infiniband/core/sysfs.c b/drivers/infiniband/core/sysfs.c
index 7ce7a6c782f..b8120650e71 100644
--- a/drivers/infiniband/core/sysfs.c
+++ b/drivers/infiniband/core/sysfs.c
@@ -307,14 +307,13 @@ static ssize_t show_pma_counter(struct ib_port *p, struct port_attribute *attr,
 	if (!p->ibdev->process_mad)
 		return sprintf(buf, "N/A (no PMA)\n");
 
-	in_mad  = kmalloc(sizeof *in_mad, GFP_KERNEL);
+	in_mad  = kzalloc(sizeof *in_mad, GFP_KERNEL);
 	out_mad = kmalloc(sizeof *in_mad, GFP_KERNEL);
 	if (!in_mad || !out_mad) {
 		ret = -ENOMEM;
 		goto out;
 	}
 
-	memset(in_mad, 0, sizeof *in_mad);
 	in_mad->mad_hdr.base_version  = 1;
 	in_mad->mad_hdr.mgmt_class    = IB_MGMT_CLASS_PERF_MGMT;
 	in_mad->mad_hdr.class_version = 1;
@@ -508,10 +507,9 @@ static int add_port(struct ib_device *device, int port_num)
 	if (ret)
 		return ret;
 
-	p = kmalloc(sizeof *p, GFP_KERNEL);
+	p = kzalloc(sizeof *p, GFP_KERNEL);
 	if (!p)
 		return -ENOMEM;
-	memset(p, 0, sizeof *p);
 
 	p->ibdev      = device;
 	p->port_num   = port_num;
diff --git a/drivers/infiniband/core/ucm.c b/drivers/infiniband/core/ucm.c
index 28477565ecb..6e15787d1de 100644
--- a/drivers/infiniband/core/ucm.c
+++ b/drivers/infiniband/core/ucm.c
@@ -172,11 +172,10 @@ static struct ib_ucm_context *ib_ucm_ctx_alloc(struct ib_ucm_file *file)
 	struct ib_ucm_context *ctx;
 	int result;
 
-	ctx = kmalloc(sizeof(*ctx), GFP_KERNEL);
+	ctx = kzalloc(sizeof *ctx, GFP_KERNEL);
 	if (!ctx)
 		return NULL;
 
-	memset(ctx, 0, sizeof *ctx);
 	atomic_set(&ctx->ref, 1);
 	init_waitqueue_head(&ctx->wait);
 	ctx->file = file;
@@ -386,11 +385,10 @@ static int ib_ucm_event_handler(struct ib_cm_id *cm_id,
 
 	ctx = cm_id->context;
 
-	uevent = kmalloc(sizeof(*uevent), GFP_KERNEL);
+	uevent = kzalloc(sizeof *uevent, GFP_KERNEL);
 	if (!uevent)
 		goto err1;
 
-	memset(uevent, 0, sizeof(*uevent));
 	uevent->ctx = ctx;
 	uevent->cm_id = cm_id;
 	uevent->resp.uid = ctx->uid;
@@ -1345,11 +1343,10 @@ static void ib_ucm_add_one(struct ib_device *device)
 	if (!device->alloc_ucontext)
 		return;
 
-	ucm_dev = kmalloc(sizeof *ucm_dev, GFP_KERNEL);
+	ucm_dev = kzalloc(sizeof *ucm_dev, GFP_KERNEL);
 	if (!ucm_dev)
 		return;
 
-	memset(ucm_dev, 0, sizeof *ucm_dev);
 	ucm_dev->ib_dev = device;
 
 	ucm_dev->devnum = find_first_zero_bit(dev_map, IB_UCM_MAX_DEVICES);
diff --git a/drivers/infiniband/core/user_mad.c b/drivers/infiniband/core/user_mad.c
index 97128e25f78..aed5ca23fb2 100644
--- a/drivers/infiniband/core/user_mad.c
+++ b/drivers/infiniband/core/user_mad.c
@@ -94,6 +94,9 @@ struct ib_umad_port {
 	struct class_device   *sm_class_dev;
 	struct semaphore       sm_sem;
 
+	struct rw_semaphore    mutex;
+	struct list_head       file_list;
+
 	struct ib_device      *ib_dev;
 	struct ib_umad_device *umad_dev;
 	int                    dev_num;
@@ -108,10 +111,10 @@ struct ib_umad_device {
 
 struct ib_umad_file {
 	struct ib_umad_port *port;
-	spinlock_t           recv_lock;
 	struct list_head     recv_list;
+	struct list_head     port_list;
+	spinlock_t           recv_lock;
 	wait_queue_head_t    recv_wait;
-	struct rw_semaphore  agent_mutex;
 	struct ib_mad_agent *agent[IB_UMAD_MAX_AGENTS];
 	struct ib_mr        *mr[IB_UMAD_MAX_AGENTS];
 };
@@ -148,7 +151,7 @@ static int queue_packet(struct ib_umad_file *file,
 {
 	int ret = 1;
 
-	down_read(&file->agent_mutex);
+	down_read(&file->port->mutex);
 	for (packet->mad.hdr.id = 0;
 	     packet->mad.hdr.id < IB_UMAD_MAX_AGENTS;
 	     packet->mad.hdr.id++)
@@ -161,7 +164,7 @@ static int queue_packet(struct ib_umad_file *file,
 			break;
 		}
 
-	up_read(&file->agent_mutex);
+	up_read(&file->port->mutex);
 
 	return ret;
 }
@@ -322,7 +325,7 @@ static ssize_t ib_umad_write(struct file *filp, const char __user *buf,
 		goto err;
 	}
 
-	down_read(&file->agent_mutex);
+	down_read(&file->port->mutex);
 
 	agent = file->agent[packet->mad.hdr.id];
 	if (!agent) {
@@ -419,7 +422,7 @@ static ssize_t ib_umad_write(struct file *filp, const char __user *buf,
 	if (ret)
 		goto err_msg;
 
-	up_read(&file->agent_mutex);
+	up_read(&file->port->mutex);
 
 	return count;
 
@@ -430,7 +433,7 @@ err_ah:
 	ib_destroy_ah(ah);
 
 err_up:
-	up_read(&file->agent_mutex);
+	up_read(&file->port->mutex);
 
 err:
 	kfree(packet);
@@ -460,7 +463,12 @@ static int ib_umad_reg_agent(struct ib_umad_file *file, unsigned long arg)
 	int agent_id;
 	int ret;
 
-	down_write(&file->agent_mutex);
+	down_write(&file->port->mutex);
+
+	if (!file->port->ib_dev) {
+		ret = -EPIPE;
+		goto out;
+	}
 
 	if (copy_from_user(&ureq, (void __user *) arg, sizeof ureq)) {
 		ret = -EFAULT;
@@ -522,7 +530,7 @@ err:
 	ib_unregister_mad_agent(agent);
 
 out:
-	up_write(&file->agent_mutex);
+	up_write(&file->port->mutex);
 	return ret;
 }
 
@@ -531,7 +539,7 @@ static int ib_umad_unreg_agent(struct ib_umad_file *file, unsigned long arg)
 	u32 id;
 	int ret = 0;
 
-	down_write(&file->agent_mutex);
+	down_write(&file->port->mutex);
 
 	if (get_user(id, (u32 __user *) arg)) {
 		ret = -EFAULT;
@@ -548,7 +556,7 @@ static int ib_umad_unreg_agent(struct ib_umad_file *file, unsigned long arg)
 	file->agent[id] = NULL;
 
 out:
-	up_write(&file->agent_mutex);
+	up_write(&file->port->mutex);
 	return ret;
 }
 
@@ -569,6 +577,7 @@ static int ib_umad_open(struct inode *inode, struct file *filp)
 {
 	struct ib_umad_port *port;
 	struct ib_umad_file *file;
+	int ret = 0;
 
 	spin_lock(&port_lock);
 	port = umad_port[iminor(inode) - IB_UMAD_MINOR_BASE];
@@ -579,21 +588,32 @@ static int ib_umad_open(struct inode *inode, struct file *filp)
 	if (!port)
 		return -ENXIO;
 
+	down_write(&port->mutex);
+
+	if (!port->ib_dev) {
+		ret = -ENXIO;
+		goto out;
+	}
+
 	file = kzalloc(sizeof *file, GFP_KERNEL);
 	if (!file) {
 		kref_put(&port->umad_dev->ref, ib_umad_release_dev);
-		return -ENOMEM;
+		ret = -ENOMEM;
+		goto out;
 	}
 
 	spin_lock_init(&file->recv_lock);
-	init_rwsem(&file->agent_mutex);
 	INIT_LIST_HEAD(&file->recv_list);
 	init_waitqueue_head(&file->recv_wait);
 
 	file->port = port;
 	filp->private_data = file;
 
-	return 0;
+	list_add_tail(&file->port_list, &port->file_list);
+
+out:
+	up_write(&port->mutex);
+	return ret;
 }
 
 static int ib_umad_close(struct inode *inode, struct file *filp)
@@ -603,6 +623,7 @@ static int ib_umad_close(struct inode *inode, struct file *filp)
 	struct ib_umad_packet *packet, *tmp;
 	int i;
 
+	down_write(&file->port->mutex);
 	for (i = 0; i < IB_UMAD_MAX_AGENTS; ++i)
 		if (file->agent[i]) {
 			ib_dereg_mr(file->mr[i]);
@@ -612,6 +633,9 @@ static int ib_umad_close(struct inode *inode, struct file *filp)
 	list_for_each_entry_safe(packet, tmp, &file->recv_list, list)
 		kfree(packet);
 
+	list_del(&file->port_list);
+	up_write(&file->port->mutex);
+
 	kfree(file);
 
 	kref_put(&dev->ref, ib_umad_release_dev);
@@ -680,9 +704,13 @@ static int ib_umad_sm_close(struct inode *inode, struct file *filp)
 	struct ib_port_modify props = {
 		.clr_port_cap_mask = IB_PORT_SM
 	};
-	int ret;
+	int ret = 0;
+
+	down_write(&port->mutex);
+	if (port->ib_dev)
+		ret = ib_modify_port(port->ib_dev, port->port_num, 0, &props);
+	up_write(&port->mutex);
 
-	ret = ib_modify_port(port->ib_dev, port->port_num, 0, &props);
 	up(&port->sm_sem);
 
 	kref_put(&port->umad_dev->ref, ib_umad_release_dev);
@@ -745,6 +773,8 @@ static int ib_umad_init_port(struct ib_device *device, int port_num,
 	port->ib_dev   = device;
 	port->port_num = port_num;
 	init_MUTEX(&port->sm_sem);
+	init_rwsem(&port->mutex);
+	INIT_LIST_HEAD(&port->file_list);
 
 	port->dev = cdev_alloc();
 	if (!port->dev)
@@ -813,6 +843,9 @@ err_cdev:
 
 static void ib_umad_kill_port(struct ib_umad_port *port)
 {
+	struct ib_umad_file *file;
+	int id;
+
 	class_set_devdata(port->class_dev,    NULL);
 	class_set_devdata(port->sm_class_dev, NULL);
 
@@ -826,6 +859,21 @@ static void ib_umad_kill_port(struct ib_umad_port *port)
 	umad_port[port->dev_num] = NULL;
 	spin_unlock(&port_lock);
 
+	down_write(&port->mutex);
+
+	port->ib_dev = NULL;
+
+	list_for_each_entry(file, &port->file_list, port_list)
+		for (id = 0; id < IB_UMAD_MAX_AGENTS; ++id) {
+			if (!file->agent[id])
+				continue;
+			ib_dereg_mr(file->mr[id]);
+			ib_unregister_mad_agent(file->agent[id]);
+			file->agent[id] = NULL;
+		}
+
+	up_write(&port->mutex);
+
 	clear_bit(port->dev_num, dev_map);
 }
 
diff --git a/drivers/infiniband/core/uverbs.h b/drivers/infiniband/core/uverbs.h
index 031cdf3c066..ecb83012786 100644
--- a/drivers/infiniband/core/uverbs.h
+++ b/drivers/infiniband/core/uverbs.h
@@ -113,6 +113,7 @@ struct ib_uevent_object {
 
 struct ib_ucq_object {
 	struct ib_uobject	uobject;
+	struct ib_uverbs_file  *uverbs_file;
 	struct list_head	comp_list;
 	struct list_head	async_list;
 	u32			comp_events_reported;
diff --git a/drivers/infiniband/core/uverbs_cmd.c b/drivers/infiniband/core/uverbs_cmd.c
index 8c89abc8c76..63a74151c60 100644
--- a/drivers/infiniband/core/uverbs_cmd.c
+++ b/drivers/infiniband/core/uverbs_cmd.c
@@ -602,6 +602,7 @@ ssize_t ib_uverbs_create_cq(struct ib_uverbs_file *file,
 
 	uobj->uobject.user_handle   = cmd.user_handle;
 	uobj->uobject.context       = file->ucontext;
+	uobj->uverbs_file	    = file;
 	uobj->comp_events_reported  = 0;
 	uobj->async_events_reported = 0;
 	INIT_LIST_HEAD(&uobj->comp_list);
diff --git a/drivers/infiniband/core/uverbs_main.c b/drivers/infiniband/core/uverbs_main.c
index 0eb38f479b3..de6581d7cb8 100644
--- a/drivers/infiniband/core/uverbs_main.c
+++ b/drivers/infiniband/core/uverbs_main.c
@@ -442,13 +442,10 @@ static void ib_uverbs_async_handler(struct ib_uverbs_file *file,
 
 void ib_uverbs_cq_event_handler(struct ib_event *event, void *context_ptr)
 {
-	struct ib_uverbs_event_file *ev_file = context_ptr;
-	struct ib_ucq_object *uobj;
+	struct ib_ucq_object *uobj = container_of(event->element.cq->uobject,
+						  struct ib_ucq_object, uobject);
 
-	uobj = container_of(event->element.cq->uobject,
-			    struct ib_ucq_object, uobject);
-
-	ib_uverbs_async_handler(ev_file->uverbs_file, uobj->uobject.user_handle,
+	ib_uverbs_async_handler(uobj->uverbs_file, uobj->uobject.user_handle,
 				event->event, &uobj->async_list,
 				&uobj->async_events_reported);
 				
@@ -728,12 +725,10 @@ static void ib_uverbs_add_one(struct ib_device *device)
 	if (!device->alloc_ucontext)
 		return;
 
-	uverbs_dev = kmalloc(sizeof *uverbs_dev, GFP_KERNEL);
+	uverbs_dev = kzalloc(sizeof *uverbs_dev, GFP_KERNEL);
 	if (!uverbs_dev)
 		return;
 
-	memset(uverbs_dev, 0, sizeof *uverbs_dev);
-
 	kref_init(&uverbs_dev->ref);
 
 	spin_lock(&map_lock);
diff --git a/drivers/infiniband/hw/mthca/mthca_cq.c b/drivers/infiniband/hw/mthca/mthca_cq.c
index 8600b6c3e0c..f98e2355582 100644
--- a/drivers/infiniband/hw/mthca/mthca_cq.c
+++ b/drivers/infiniband/hw/mthca/mthca_cq.c
@@ -208,7 +208,7 @@ static inline void update_cons_index(struct mthca_dev *dev, struct mthca_cq *cq,
 	}
 }
 
-void mthca_cq_event(struct mthca_dev *dev, u32 cqn)
+void mthca_cq_completion(struct mthca_dev *dev, u32 cqn)
 {
 	struct mthca_cq *cq;
 
@@ -224,6 +224,35 @@ void mthca_cq_event(struct mthca_dev *dev, u32 cqn)
 	cq->ibcq.comp_handler(&cq->ibcq, cq->ibcq.cq_context);
 }
 
+void mthca_cq_event(struct mthca_dev *dev, u32 cqn,
+		    enum ib_event_type event_type)
+{
+	struct mthca_cq *cq;
+	struct ib_event event;
+
+	spin_lock(&dev->cq_table.lock);
+
+	cq = mthca_array_get(&dev->cq_table.cq, cqn & (dev->limits.num_cqs - 1));
+
+	if (cq)
+		atomic_inc(&cq->refcount);
+	spin_unlock(&dev->cq_table.lock);
+
+	if (!cq) {
+		mthca_warn(dev, "Async event for bogus CQ %08x\n", cqn);
+		return;
+	}
+
+	event.device      = &dev->ib_dev;
+	event.event       = event_type;
+	event.element.cq  = &cq->ibcq;
+	if (cq->ibcq.event_handler)
+		cq->ibcq.event_handler(&event, cq->ibcq.cq_context);
+
+	if (atomic_dec_and_test(&cq->refcount))
+		wake_up(&cq->wait);
+}
+
 void mthca_cq_clean(struct mthca_dev *dev, u32 cqn, u32 qpn,
 		    struct mthca_srq *srq)
 {
diff --git a/drivers/infiniband/hw/mthca/mthca_dev.h b/drivers/infiniband/hw/mthca/mthca_dev.h
index 7e68bd4a378..e7e5d3b4f00 100644
--- a/drivers/infiniband/hw/mthca/mthca_dev.h
+++ b/drivers/infiniband/hw/mthca/mthca_dev.h
@@ -460,7 +460,9 @@ int mthca_init_cq(struct mthca_dev *dev, int nent,
 		  struct mthca_cq *cq);
 void mthca_free_cq(struct mthca_dev *dev,
 		   struct mthca_cq *cq);
-void mthca_cq_event(struct mthca_dev *dev, u32 cqn);
+void mthca_cq_completion(struct mthca_dev *dev, u32 cqn);
+void mthca_cq_event(struct mthca_dev *dev, u32 cqn,
+		    enum ib_event_type event_type);
 void mthca_cq_clean(struct mthca_dev *dev, u32 cqn, u32 qpn,
 		    struct mthca_srq *srq);
 
diff --git a/drivers/infiniband/hw/mthca/mthca_eq.c b/drivers/infiniband/hw/mthca/mthca_eq.c
index e5a047a6dbe..34d68e5a72d 100644
--- a/drivers/infiniband/hw/mthca/mthca_eq.c
+++ b/drivers/infiniband/hw/mthca/mthca_eq.c
@@ -292,7 +292,7 @@ static int mthca_eq_int(struct mthca_dev *dev, struct mthca_eq *eq)
 		case MTHCA_EVENT_TYPE_COMP:
 			disarm_cqn = be32_to_cpu(eqe->event.comp.cqn) & 0xffffff;
 			disarm_cq(dev, eq->eqn, disarm_cqn);
-			mthca_cq_event(dev, disarm_cqn);
+			mthca_cq_completion(dev, disarm_cqn);
 			break;
 
 		case MTHCA_EVENT_TYPE_PATH_MIG:
@@ -364,6 +364,8 @@ static int mthca_eq_int(struct mthca_dev *dev, struct mthca_eq *eq)
 				   eqe->event.cq_err.syndrome == 1 ?
 				   "overrun" : "access violation",
 				   be32_to_cpu(eqe->event.cq_err.cqn) & 0xffffff);
+			mthca_cq_event(dev, be32_to_cpu(eqe->event.cq_err.cqn),
+				       IB_EVENT_CQ_ERR);
 			break;
 
 		case MTHCA_EVENT_TYPE_EQ_OVERFLOW:
diff --git a/drivers/infiniband/hw/mthca/mthca_main.c b/drivers/infiniband/hw/mthca/mthca_main.c
index 883d1e5a79b..45c6328e780 100644
--- a/drivers/infiniband/hw/mthca/mthca_main.c
+++ b/drivers/infiniband/hw/mthca/mthca_main.c
@@ -1057,7 +1057,7 @@ static int __devinit mthca_init_one(struct pci_dev *pdev,
 		goto err_cmd;
 
 	if (mdev->fw_ver < mthca_hca_table[id->driver_data].latest_fw) {
-		mthca_warn(mdev, "HCA FW version %x.%x.%x is old (%x.%x.%x is current).\n",
+		mthca_warn(mdev, "HCA FW version %d.%d.%d is old (%d.%d.%d is current).\n",
 			   (int) (mdev->fw_ver >> 32), (int) (mdev->fw_ver >> 16) & 0xffff,
 			   (int) (mdev->fw_ver & 0xffff),
 			   (int) (mthca_hca_table[id->driver_data].latest_fw >> 32),
diff --git a/drivers/infiniband/hw/mthca/mthca_mr.c b/drivers/infiniband/hw/mthca/mthca_mr.c
index 1f97a44477f..e995e2aa016 100644
--- a/drivers/infiniband/hw/mthca/mthca_mr.c
+++ b/drivers/infiniband/hw/mthca/mthca_mr.c
@@ -140,13 +140,11 @@ static int __devinit mthca_buddy_init(struct mthca_buddy *buddy, int max_order)
 	buddy->max_order = max_order;
 	spin_lock_init(&buddy->lock);
 
-	buddy->bits = kmalloc((buddy->max_order + 1) * sizeof (long *),
+	buddy->bits = kzalloc((buddy->max_order + 1) * sizeof (long *),
 			      GFP_KERNEL);
 	if (!buddy->bits)
 		goto err_out;
 
-	memset(buddy->bits, 0, (buddy->max_order + 1) * sizeof (long *));
-
 	for (i = 0; i <= buddy->max_order; ++i) {
 		s = BITS_TO_LONGS(1 << (buddy->max_order - i));
 		buddy->bits[i] = kmalloc(s * sizeof (long), GFP_KERNEL);
diff --git a/drivers/infiniband/hw/mthca/mthca_profile.c b/drivers/infiniband/hw/mthca/mthca_profile.c
index bd133868207..08a909371b0 100644
--- a/drivers/infiniband/hw/mthca/mthca_profile.c
+++ b/drivers/infiniband/hw/mthca/mthca_profile.c
@@ -82,12 +82,10 @@ u64 mthca_make_profile(struct mthca_dev *dev,
 	struct mthca_resource tmp;
 	int i, j;
 
-	profile = kmalloc(MTHCA_RES_NUM * sizeof *profile, GFP_KERNEL);
+	profile = kzalloc(MTHCA_RES_NUM * sizeof *profile, GFP_KERNEL);
 	if (!profile)
 		return -ENOMEM;
 
-	memset(profile, 0, MTHCA_RES_NUM * sizeof *profile);
-
 	profile[MTHCA_RES_QP].size   = dev_lim->qpc_entry_sz;
 	profile[MTHCA_RES_EEC].size  = dev_lim->eec_entry_sz;
 	profile[MTHCA_RES_SRQ].size  = dev_lim->srq_entry_sz;
diff --git a/drivers/infiniband/hw/mthca/mthca_provider.c b/drivers/infiniband/hw/mthca/mthca_provider.c
index 1b9477edbd7..6b016666826 100644
--- a/drivers/infiniband/hw/mthca/mthca_provider.c
+++ b/drivers/infiniband/hw/mthca/mthca_provider.c
@@ -1028,7 +1028,7 @@ static ssize_t show_rev(struct class_device *cdev, char *buf)
 static ssize_t show_fw_ver(struct class_device *cdev, char *buf)
 {
 	struct mthca_dev *dev = container_of(cdev, struct mthca_dev, ib_dev.class_dev);
-	return sprintf(buf, "%x.%x.%x\n", (int) (dev->fw_ver >> 32),
+	return sprintf(buf, "%d.%d.%d\n", (int) (dev->fw_ver >> 32),
 		       (int) (dev->fw_ver >> 16) & 0xffff,
 		       (int) dev->fw_ver & 0xffff);
 }
diff --git a/drivers/infiniband/hw/mthca/mthca_qp.c b/drivers/infiniband/hw/mthca/mthca_qp.c
index 7c9afde5ace..8852ea477c2 100644
--- a/drivers/infiniband/hw/mthca/mthca_qp.c
+++ b/drivers/infiniband/hw/mthca/mthca_qp.c
@@ -584,6 +584,13 @@ int mthca_modify_qp(struct ib_qp *ibqp, struct ib_qp_attr *attr, int attr_mask)
 		return -EINVAL;
 	}
 
+	if ((attr_mask & IB_QP_PKEY_INDEX) && 
+	     attr->pkey_index >= dev->limits.pkey_table_len) {
+		mthca_dbg(dev, "PKey index (%u) too large. max is %d\n",
+			  attr->pkey_index,dev->limits.pkey_table_len-1); 
+		return -EINVAL;
+	}
+
 	mailbox = mthca_alloc_mailbox(dev, GFP_KERNEL);
 	if (IS_ERR(mailbox))
 		return PTR_ERR(mailbox);
diff --git a/drivers/infiniband/hw/mthca/mthca_srq.c b/drivers/infiniband/hw/mthca/mthca_srq.c
index 64f70aa1b3c..292f55be8cb 100644
--- a/drivers/infiniband/hw/mthca/mthca_srq.c
+++ b/drivers/infiniband/hw/mthca/mthca_srq.c
@@ -75,15 +75,16 @@ static void *get_wqe(struct mthca_srq *srq, int n)
 
 /*
  * Return a pointer to the location within a WQE that we're using as a
- * link when the WQE is in the free list.  We use an offset of 4
- * because in the Tavor case, posting a WQE may overwrite the first
- * four bytes of the previous WQE.  The offset avoids corrupting our
- * free list if the WQE has already completed and been put on the free
- * list when we post the next WQE.
+ * link when the WQE is in the free list.  We use the imm field
+ * because in the Tavor case, posting a WQE may overwrite the next
+ * segment of the previous WQE, but a receive WQE will never touch the
+ * imm field.  This avoids corrupting our free list if the previous
+ * WQE has already completed and been put on the free list when we
+ * post the next WQE.
  */
 static inline int *wqe_to_link(void *wqe)
 {
-	return (int *) (wqe + 4);
+	return (int *) (wqe + offsetof(struct mthca_next_seg, imm));
 }
 
 static void mthca_tavor_init_srq_context(struct mthca_dev *dev,
diff --git a/drivers/infiniband/ulp/ipoib/ipoib.h b/drivers/infiniband/ulp/ipoib/ipoib.h
index c994a916a58..0095acc0fbb 100644
--- a/drivers/infiniband/ulp/ipoib/ipoib.h
+++ b/drivers/infiniband/ulp/ipoib/ipoib.h
@@ -235,6 +235,7 @@ static inline void ipoib_put_ah(struct ipoib_ah *ah)
 	kref_put(&ah->ref, ipoib_free_ah);
 }
 
+int ipoib_open(struct net_device *dev);
 int ipoib_add_pkey_attr(struct net_device *dev);
 
 void ipoib_send(struct net_device *dev, struct sk_buff *skb,
@@ -267,6 +268,7 @@ int ipoib_mcast_stop_thread(struct net_device *dev, int flush);
 void ipoib_mcast_dev_down(struct net_device *dev);
 void ipoib_mcast_dev_flush(struct net_device *dev);
 
+#ifdef CONFIG_INFINIBAND_IPOIB_DEBUG
 struct ipoib_mcast_iter *ipoib_mcast_iter_init(struct net_device *dev);
 void ipoib_mcast_iter_free(struct ipoib_mcast_iter *iter);
 int ipoib_mcast_iter_next(struct ipoib_mcast_iter *iter);
@@ -276,6 +278,7 @@ void ipoib_mcast_iter_read(struct ipoib_mcast_iter *iter,
 				  unsigned int *queuelen,
 				  unsigned int *complete,
 				  unsigned int *send_only);
+#endif
 
 int ipoib_mcast_attach(struct net_device *dev, u16 mlid,
 		       union ib_gid *mgid);
diff --git a/drivers/infiniband/ulp/ipoib/ipoib_ib.c b/drivers/infiniband/ulp/ipoib/ipoib_ib.c
index 192fef884e2..54ef2fea530 100644
--- a/drivers/infiniband/ulp/ipoib/ipoib_ib.c
+++ b/drivers/infiniband/ulp/ipoib/ipoib_ib.c
@@ -486,15 +486,16 @@ int ipoib_ib_dev_stop(struct net_device *dev)
 {
 	struct ipoib_dev_priv *priv = netdev_priv(dev);
 	struct ib_qp_attr qp_attr;
-	int attr_mask;
 	unsigned long begin;
 	struct ipoib_tx_buf *tx_req;
 	int i;
 
-	/* Kill the existing QP and allocate a new one */
+	/*
+	 * Move our QP to the error state and then reinitialize in
+	 * when all work requests have completed or have been flushed.
+	 */
 	qp_attr.qp_state = IB_QPS_ERR;
-	attr_mask        = IB_QP_STATE;
-	if (ib_modify_qp(priv->qp, &qp_attr, attr_mask))
+	if (ib_modify_qp(priv->qp, &qp_attr, IB_QP_STATE))
 		ipoib_warn(priv, "Failed to modify QP to ERROR state\n");
 
 	/* Wait for all sends and receives to complete */
@@ -541,8 +542,7 @@ int ipoib_ib_dev_stop(struct net_device *dev)
 
 timeout:
 	qp_attr.qp_state = IB_QPS_RESET;
-	attr_mask        = IB_QP_STATE;
-	if (ib_modify_qp(priv->qp, &qp_attr, attr_mask))
+	if (ib_modify_qp(priv->qp, &qp_attr, IB_QP_STATE))
 		ipoib_warn(priv, "Failed to modify QP to RESET state\n");
 
 	/* Wait for all AHs to be reaped */
@@ -636,7 +636,6 @@ void ipoib_ib_dev_cleanup(struct net_device *dev)
  * Bug #2507. This implementation will probably be removed when the P_Key
  * change async notification is available.
  */
-int ipoib_open(struct net_device *dev);
 
 static void ipoib_pkey_dev_check_presence(struct net_device *dev)
 {
diff --git a/drivers/infiniband/ulp/ipoib/ipoib_main.c b/drivers/infiniband/ulp/ipoib/ipoib_main.c
index cd4f42328db..ce0296273e7 100644
--- a/drivers/infiniband/ulp/ipoib/ipoib_main.c
+++ b/drivers/infiniband/ulp/ipoib/ipoib_main.c
@@ -356,18 +356,15 @@ static struct ipoib_path *path_rec_create(struct net_device *dev,
 	struct ipoib_dev_priv *priv = netdev_priv(dev);
 	struct ipoib_path *path;
 
-	path = kmalloc(sizeof *path, GFP_ATOMIC);
+	path = kzalloc(sizeof *path, GFP_ATOMIC);
 	if (!path)
 		return NULL;
 
-	path->dev          = dev;
-	path->pathrec.dlid = 0;
-	path->ah           = NULL;
+	path->dev = dev;
 
 	skb_queue_head_init(&path->queue);
 
 	INIT_LIST_HEAD(&path->neigh_list);
-	path->query = NULL;
 	init_completion(&path->done);
 
 	memcpy(path->pathrec.dgid.raw, gid->raw, sizeof (union ib_gid));
@@ -551,11 +548,8 @@ static int ipoib_start_xmit(struct sk_buff *skb, struct net_device *dev)
 	struct ipoib_neigh *neigh;
 	unsigned long flags;
 
-	local_irq_save(flags);
-	if (!spin_trylock(&priv->tx_lock)) {
-		local_irq_restore(flags);
+	if (!spin_trylock_irqsave(&priv->tx_lock, flags))
 		return NETDEV_TX_LOCKED;
-	}
 
 	/*
 	 * Check if our queue is stopped.  Since we have the LLTX bit
@@ -732,25 +726,21 @@ int ipoib_dev_init(struct net_device *dev, struct ib_device *ca, int port)
 
 	/* Allocate RX/TX "rings" to hold queued skbs */
 
-	priv->rx_ring =	kmalloc(IPOIB_RX_RING_SIZE * sizeof (struct ipoib_rx_buf),
+	priv->rx_ring =	kzalloc(IPOIB_RX_RING_SIZE * sizeof (struct ipoib_rx_buf),
 				GFP_KERNEL);
 	if (!priv->rx_ring) {
 		printk(KERN_WARNING "%s: failed to allocate RX ring (%d entries)\n",
 		       ca->name, IPOIB_RX_RING_SIZE);
 		goto out;
 	}
-	memset(priv->rx_ring, 0,
-	       IPOIB_RX_RING_SIZE * sizeof (struct ipoib_rx_buf));
 
-	priv->tx_ring = kmalloc(IPOIB_TX_RING_SIZE * sizeof (struct ipoib_tx_buf),
+	priv->tx_ring = kzalloc(IPOIB_TX_RING_SIZE * sizeof (struct ipoib_tx_buf),
 				GFP_KERNEL);
 	if (!priv->tx_ring) {
 		printk(KERN_WARNING "%s: failed to allocate TX ring (%d entries)\n",
 		       ca->name, IPOIB_TX_RING_SIZE);
 		goto out_rx_ring_cleanup;
 	}
-	memset(priv->tx_ring, 0,
-	       IPOIB_TX_RING_SIZE * sizeof (struct ipoib_tx_buf));
 
 	/* priv->tx_head & tx_tail are already 0 */
 
@@ -807,10 +797,6 @@ static void ipoib_setup(struct net_device *dev)
 
 	dev->watchdog_timeo 	 = HZ;
 
-	dev->rebuild_header 	 = NULL;
-	dev->set_mac_address 	 = NULL;
-	dev->header_cache_update = NULL;
-
 	dev->flags              |= IFF_BROADCAST | IFF_MULTICAST;
 
 	/*
diff --git a/drivers/infiniband/ulp/ipoib/ipoib_multicast.c b/drivers/infiniband/ulp/ipoib/ipoib_multicast.c
index 36ce29836bf..3ecf78a9493 100644
--- a/drivers/infiniband/ulp/ipoib/ipoib_multicast.c
+++ b/drivers/infiniband/ulp/ipoib/ipoib_multicast.c
@@ -135,12 +135,10 @@ static struct ipoib_mcast *ipoib_mcast_alloc(struct net_device *dev,
 {
 	struct ipoib_mcast *mcast;
 
-	mcast = kmalloc(sizeof (*mcast), can_sleep ? GFP_KERNEL : GFP_ATOMIC);
+	mcast = kzalloc(sizeof *mcast, can_sleep ? GFP_KERNEL : GFP_ATOMIC);
 	if (!mcast)
 		return NULL;
 
-	memset(mcast, 0, sizeof (*mcast));
-
 	init_completion(&mcast->done);
 
 	mcast->dev = dev;
@@ -919,6 +917,8 @@ void ipoib_mcast_restart_task(void *dev_ptr)
 		ipoib_mcast_start_thread(dev);
 }
 
+#ifdef CONFIG_INFINIBAND_IPOIB_DEBUG
+
 struct ipoib_mcast_iter *ipoib_mcast_iter_init(struct net_device *dev)
 {
 	struct ipoib_mcast_iter *iter;
@@ -991,3 +991,5 @@ void ipoib_mcast_iter_read(struct ipoib_mcast_iter *iter,
 	*complete  = iter->complete;
 	*send_only = iter->send_only;
 }
+
+#endif /* CONFIG_INFINIBAND_IPOIB_DEBUG */
diff --git a/drivers/infiniband/ulp/ipoib/ipoib_verbs.c b/drivers/infiniband/ulp/ipoib/ipoib_verbs.c
index b5902a7ec24..e829e10400e 100644
--- a/drivers/infiniband/ulp/ipoib/ipoib_verbs.c
+++ b/drivers/infiniband/ulp/ipoib/ipoib_verbs.c
@@ -41,7 +41,6 @@ int ipoib_mcast_attach(struct net_device *dev, u16 mlid, union ib_gid *mgid)
 {
 	struct ipoib_dev_priv *priv = netdev_priv(dev);
 	struct ib_qp_attr *qp_attr;
-	int attr_mask;
 	int ret;
 	u16 pkey_index;
 
@@ -59,8 +58,7 @@ int ipoib_mcast_attach(struct net_device *dev, u16 mlid, union ib_gid *mgid)
 
 	/* set correct QKey for QP */
 	qp_attr->qkey = priv->qkey;
-	attr_mask = IB_QP_QKEY;
-	ret = ib_modify_qp(priv->qp, qp_attr, attr_mask);
+	ret = ib_modify_qp(priv->qp, qp_attr, IB_QP_QKEY);
 	if (ret) {
 		ipoib_warn(priv, "failed to modify QP, ret = %d\n", ret);
 		goto out;
diff --git a/include/rdma/ib_user_cm.h b/include/rdma/ib_user_cm.h
index 3037588b846..19be116047f 100644
--- a/include/rdma/ib_user_cm.h
+++ b/include/rdma/ib_user_cm.h
@@ -38,7 +38,7 @@
 
 #include <linux/types.h>
 
-#define IB_USER_CM_ABI_VERSION 3
+#define IB_USER_CM_ABI_VERSION 4
 
 enum {
 	IB_USER_CM_CMD_CREATE_ID,
@@ -84,6 +84,7 @@ struct ib_ucm_create_id_resp {
 struct ib_ucm_destroy_id {
 	__u64 response;
 	__u32 id;
+	__u32 reserved;
 };
 
 struct ib_ucm_destroy_id_resp {
@@ -93,6 +94,7 @@ struct ib_ucm_destroy_id_resp {
 struct ib_ucm_attr_id {
 	__u64 response;
 	__u32 id;
+	__u32 reserved;
 };
 
 struct ib_ucm_attr_id_resp {
@@ -164,6 +166,7 @@ struct ib_ucm_listen {
 	__be64 service_id;
 	__be64 service_mask;
 	__u32 id;
+	__u32 reserved;
 };
 
 struct ib_ucm_establish {
@@ -219,7 +222,7 @@ struct ib_ucm_req {
 	__u8  rnr_retry_count;
 	__u8  max_cm_retries;
 	__u8  srq;
-	__u8  reserved[1];
+	__u8  reserved[5];
 };
 
 struct ib_ucm_rep {
@@ -236,6 +239,7 @@ struct ib_ucm_rep {
 	__u8  flow_control;
 	__u8  rnr_retry_count;
 	__u8  srq;
+	__u8  reserved[4];
 };
 
 struct ib_ucm_info {
@@ -245,7 +249,7 @@ struct ib_ucm_info {
 	__u64 data;
 	__u8  info_len;
 	__u8  data_len;
-	__u8  reserved[2];
+	__u8  reserved[6];
 };
 
 struct ib_ucm_mra {
@@ -273,6 +277,7 @@ struct ib_ucm_sidr_req {
 	__u16 pkey;
 	__u8  len;
 	__u8  max_cm_retries;
+	__u8  reserved[4];
 };
 
 struct ib_ucm_sidr_rep {
@@ -284,7 +289,7 @@ struct ib_ucm_sidr_rep {
 	__u64 data;
 	__u8  info_len;
 	__u8  data_len;
-	__u8  reserved[2];
+	__u8  reserved[6];
 };
 /*
  * event notification ABI structures.
@@ -295,7 +300,7 @@ struct ib_ucm_event_get {
 	__u64 info;
 	__u8  data_len;
 	__u8  info_len;
-	__u8  reserved[2];
+	__u8  reserved[6];
 };
 
 struct ib_ucm_req_event_resp {
@@ -315,6 +320,7 @@ struct ib_ucm_req_event_resp {
 	__u8  rnr_retry_count;
 	__u8  srq;
 	__u8  port;
+	__u8  reserved[7];
 };
 
 struct ib_ucm_rep_event_resp {
@@ -329,7 +335,7 @@ struct ib_ucm_rep_event_resp {
 	__u8  flow_control;
 	__u8  rnr_retry_count;
 	__u8  srq;
-	__u8  reserved[1];
+	__u8  reserved[5];
 };
 
 struct ib_ucm_rej_event_resp {
@@ -374,6 +380,7 @@ struct ib_ucm_event_resp {
 	__u32 id;
 	__u32 event;
 	__u32 present;
+	__u32 reserved;
 	union {
 		struct ib_ucm_req_event_resp req_resp;
 		struct ib_ucm_rep_event_resp rep_resp;