Merge commit 'v2.6.27-rc1' into for-linus

6 files changed, 148 insertions, 50 deletions

diff --git a/drivers/infiniband/core/cm.c b/drivers/infiniband/core/cm.c
index 55738eead3b..922d35f4fc0 100644
--- a/drivers/infiniband/core/cm.c
+++ b/drivers/infiniband/core/cm.c
@@ -44,6 +44,7 @@
 #include <linux/spinlock.h>
 #include <linux/sysfs.h>
 #include <linux/workqueue.h>
+#include <linux/kdev_t.h>
 
 #include <rdma/ib_cache.h>
 #include <rdma/ib_cm.h>
@@ -162,8 +163,8 @@ struct cm_port {
 
 struct cm_device {
 	struct list_head list;
-	struct ib_device *device;
-	struct kobject dev_obj;
+	struct ib_device *ib_device;
+	struct device *device;
 	u8 ack_delay;
 	struct cm_port *port[0];
 };
@@ -339,7 +340,7 @@ static void cm_init_av_for_response(struct cm_port *port, struct ib_wc *wc,
 {
 	av->port = port;
 	av->pkey_index = wc->pkey_index;
-	ib_init_ah_from_wc(port->cm_dev->device, port->port_num, wc,
+	ib_init_ah_from_wc(port->cm_dev->ib_device, port->port_num, wc,
 			   grh, &av->ah_attr);
 }
 
@@ -353,7 +354,7 @@ static int cm_init_av_by_path(struct ib_sa_path_rec *path, struct cm_av *av)
 
 	read_lock_irqsave(&cm.device_lock, flags);
 	list_for_each_entry(cm_dev, &cm.device_list, list) {
-		if (!ib_find_cached_gid(cm_dev->device, &path->sgid,
+		if (!ib_find_cached_gid(cm_dev->ib_device, &path->sgid,
 					&p, NULL)) {
 			port = cm_dev->port[p-1];
 			break;
@@ -364,13 +365,13 @@ static int cm_init_av_by_path(struct ib_sa_path_rec *path, struct cm_av *av)
 	if (!port)
 		return -EINVAL;
 
-	ret = ib_find_cached_pkey(cm_dev->device, port->port_num,
+	ret = ib_find_cached_pkey(cm_dev->ib_device, port->port_num,
 				  be16_to_cpu(path->pkey), &av->pkey_index);
 	if (ret)
 		return ret;
 
 	av->port = port;
-	ib_init_ah_from_path(cm_dev->device, port->port_num, path,
+	ib_init_ah_from_path(cm_dev->ib_device, port->port_num, path,
 			     &av->ah_attr);
 	av->timeout = path->packet_life_time + 1;
 	return 0;
@@ -1515,7 +1516,7 @@ static int cm_req_handler(struct cm_work *work)
 
 	req_msg = (struct cm_req_msg *)work->mad_recv_wc->recv_buf.mad;
 
-	cm_id = ib_create_cm_id(work->port->cm_dev->device, NULL, NULL);
+	cm_id = ib_create_cm_id(work->port->cm_dev->ib_device, NULL, NULL);
 	if (IS_ERR(cm_id))
 		return PTR_ERR(cm_id);
 
@@ -1550,7 +1551,7 @@ static int cm_req_handler(struct cm_work *work)
 	cm_format_paths_from_req(req_msg, &work->path[0], &work->path[1]);
 	ret = cm_init_av_by_path(&work->path[0], &cm_id_priv->av);
 	if (ret) {
-		ib_get_cached_gid(work->port->cm_dev->device,
+		ib_get_cached_gid(work->port->cm_dev->ib_device,
 				  work->port->port_num, 0, &work->path[0].sgid);
 		ib_send_cm_rej(cm_id, IB_CM_REJ_INVALID_GID,
 			       &work->path[0].sgid, sizeof work->path[0].sgid,
@@ -2950,7 +2951,7 @@ static int cm_sidr_req_handler(struct cm_work *work)
 	struct cm_sidr_req_msg *sidr_req_msg;
 	struct ib_wc *wc;
 
-	cm_id = ib_create_cm_id(work->port->cm_dev->device, NULL, NULL);
+	cm_id = ib_create_cm_id(work->port->cm_dev->ib_device, NULL, NULL);
 	if (IS_ERR(cm_id))
 		return PTR_ERR(cm_id);
 	cm_id_priv = container_of(cm_id, struct cm_id_private, id);
@@ -3578,7 +3579,7 @@ static void cm_get_ack_delay(struct cm_device *cm_dev)
 {
 	struct ib_device_attr attr;
 
-	if (ib_query_device(cm_dev->device, &attr))
+	if (ib_query_device(cm_dev->ib_device, &attr))
 		cm_dev->ack_delay = 0; /* acks will rely on packet life time */
 	else
 		cm_dev->ack_delay = attr.local_ca_ack_delay;
@@ -3618,18 +3619,6 @@ static struct kobj_type cm_port_obj_type = {
 	.release = cm_release_port_obj
 };
 
-static void cm_release_dev_obj(struct kobject *obj)
-{
-	struct cm_device *cm_dev;
-
-	cm_dev = container_of(obj, struct cm_device, dev_obj);
-	kfree(cm_dev);
-}
-
-static struct kobj_type cm_dev_obj_type = {
-	.release = cm_release_dev_obj
-};
-
 struct class cm_class = {
 	.name    = "infiniband_cm",
 };
@@ -3640,7 +3629,7 @@ static int cm_create_port_fs(struct cm_port *port)
 	int i, ret;
 
 	ret = kobject_init_and_add(&port->port_obj, &cm_port_obj_type,
-				   &port->cm_dev->dev_obj,
+				   &port->cm_dev->device->kobj,
 				   "%d", port->port_num);
 	if (ret) {
 		kfree(port);
@@ -3676,7 +3665,7 @@ static void cm_remove_port_fs(struct cm_port *port)
 	kobject_put(&port->port_obj);
 }
 
-static void cm_add_one(struct ib_device *device)
+static void cm_add_one(struct ib_device *ib_device)
 {
 	struct cm_device *cm_dev;
 	struct cm_port *port;
@@ -3691,26 +3680,27 @@ static void cm_add_one(struct ib_device *device)
 	int ret;
 	u8 i;
 
-	if (rdma_node_get_transport(device->node_type) != RDMA_TRANSPORT_IB)
+	if (rdma_node_get_transport(ib_device->node_type) != RDMA_TRANSPORT_IB)
 		return;
 
 	cm_dev = kzalloc(sizeof(*cm_dev) + sizeof(*port) *
-			 device->phys_port_cnt, GFP_KERNEL);
+			 ib_device->phys_port_cnt, GFP_KERNEL);
 	if (!cm_dev)
 		return;
 
-	cm_dev->device = device;
+	cm_dev->ib_device = ib_device;
 	cm_get_ack_delay(cm_dev);
 
-	ret = kobject_init_and_add(&cm_dev->dev_obj, &cm_dev_obj_type,
-				   &cm_class.subsys.kobj, "%s", device->name);
-	if (ret) {
+	cm_dev->device = device_create_drvdata(&cm_class, &ib_device->dev,
+					       MKDEV(0, 0), NULL,
+					       "%s", ib_device->name);
+	if (!cm_dev->device) {
 		kfree(cm_dev);
 		return;
 	}
 
 	set_bit(IB_MGMT_METHOD_SEND, reg_req.method_mask);
-	for (i = 1; i <= device->phys_port_cnt; i++) {
+	for (i = 1; i <= ib_device->phys_port_cnt; i++) {
 		port = kzalloc(sizeof *port, GFP_KERNEL);
 		if (!port)
 			goto error1;
@@ -3723,7 +3713,7 @@ static void cm_add_one(struct ib_device *device)
 		if (ret)
 			goto error1;
 
-		port->mad_agent = ib_register_mad_agent(device, i,
+		port->mad_agent = ib_register_mad_agent(ib_device, i,
 							IB_QPT_GSI,
 							&reg_req,
 							0,
@@ -3733,11 +3723,11 @@ static void cm_add_one(struct ib_device *device)
 		if (IS_ERR(port->mad_agent))
 			goto error2;
 
-		ret = ib_modify_port(device, i, 0, &port_modify);
+		ret = ib_modify_port(ib_device, i, 0, &port_modify);
 		if (ret)
 			goto error3;
 	}
-	ib_set_client_data(device, &cm_client, cm_dev);
+	ib_set_client_data(ib_device, &cm_client, cm_dev);
 
 	write_lock_irqsave(&cm.device_lock, flags);
 	list_add_tail(&cm_dev->list, &cm.device_list);
@@ -3753,14 +3743,14 @@ error1:
 	port_modify.clr_port_cap_mask = IB_PORT_CM_SUP;
 	while (--i) {
 		port = cm_dev->port[i-1];
-		ib_modify_port(device, port->port_num, 0, &port_modify);
+		ib_modify_port(ib_device, port->port_num, 0, &port_modify);
 		ib_unregister_mad_agent(port->mad_agent);
 		cm_remove_port_fs(port);
 	}
-	kobject_put(&cm_dev->dev_obj);
+	device_unregister(cm_dev->device);
 }
 
-static void cm_remove_one(struct ib_device *device)
+static void cm_remove_one(struct ib_device *ib_device)
 {
 	struct cm_device *cm_dev;
 	struct cm_port *port;
@@ -3770,7 +3760,7 @@ static void cm_remove_one(struct ib_device *device)
 	unsigned long flags;
 	int i;
 
-	cm_dev = ib_get_client_data(device, &cm_client);
+	cm_dev = ib_get_client_data(ib_device, &cm_client);
 	if (!cm_dev)
 		return;
 
@@ -3778,14 +3768,14 @@ static void cm_remove_one(struct ib_device *device)
 	list_del(&cm_dev->list);
 	write_unlock_irqrestore(&cm.device_lock, flags);
 
-	for (i = 1; i <= device->phys_port_cnt; i++) {
+	for (i = 1; i <= ib_device->phys_port_cnt; i++) {
 		port = cm_dev->port[i-1];
-		ib_modify_port(device, port->port_num, 0, &port_modify);
+		ib_modify_port(ib_device, port->port_num, 0, &port_modify);
 		ib_unregister_mad_agent(port->mad_agent);
 		flush_workqueue(cm.wq);
 		cm_remove_port_fs(port);
 	}
-	kobject_put(&cm_dev->dev_obj);
+	device_unregister(cm_dev->device);
 }
 
 static int __init ib_cm_init(void)
diff --git a/drivers/infiniband/core/cma.c b/drivers/infiniband/core/cma.c
index ae11d5cc74d..e980ff3335d 100644
--- a/drivers/infiniband/core/cma.c
+++ b/drivers/infiniband/core/cma.c
@@ -168,6 +168,12 @@ struct cma_work {
 	struct rdma_cm_event	event;
 };
 
+struct cma_ndev_work {
+	struct work_struct	work;
+	struct rdma_id_private	*id;
+	struct rdma_cm_event	event;
+};
+
 union cma_ip_addr {
 	struct in6_addr ip6;
 	struct {
@@ -914,7 +920,10 @@ static int cma_ib_handler(struct ib_cm_id *cm_id, struct ib_cm_event *ib_event)
 	struct rdma_cm_event event;
 	int ret = 0;
 
-	if (cma_disable_callback(id_priv, CMA_CONNECT))
+	if ((ib_event->event != IB_CM_TIMEWAIT_EXIT &&
+		cma_disable_callback(id_priv, CMA_CONNECT)) ||
+	    (ib_event->event == IB_CM_TIMEWAIT_EXIT &&
+		cma_disable_callback(id_priv, CMA_DISCONNECT)))
 		return 0;
 
 	memset(&event, 0, sizeof event);
@@ -950,6 +959,8 @@ static int cma_ib_handler(struct ib_cm_id *cm_id, struct ib_cm_event *ib_event)
 		event.event = RDMA_CM_EVENT_DISCONNECTED;
 		break;
 	case IB_CM_TIMEWAIT_EXIT:
+		event.event = RDMA_CM_EVENT_TIMEWAIT_EXIT;
+		break;
 	case IB_CM_MRA_RECEIVED:
 		/* ignore event */
 		goto out;
@@ -1598,6 +1609,30 @@ out:
 	kfree(work);
 }
 
+static void cma_ndev_work_handler(struct work_struct *_work)
+{
+	struct cma_ndev_work *work = container_of(_work, struct cma_ndev_work, work);
+	struct rdma_id_private *id_priv = work->id;
+	int destroy = 0;
+
+	mutex_lock(&id_priv->handler_mutex);
+	if (id_priv->state == CMA_DESTROYING ||
+	    id_priv->state == CMA_DEVICE_REMOVAL)
+		goto out;
+
+	if (id_priv->id.event_handler(&id_priv->id, &work->event)) {
+		cma_exch(id_priv, CMA_DESTROYING);
+		destroy = 1;
+	}
+
+out:
+	mutex_unlock(&id_priv->handler_mutex);
+	cma_deref_id(id_priv);
+	if (destroy)
+		rdma_destroy_id(&id_priv->id);
+	kfree(work);
+}
+
 static int cma_resolve_ib_route(struct rdma_id_private *id_priv, int timeout_ms)
 {
 	struct rdma_route *route = &id_priv->id.route;
@@ -2723,6 +2758,65 @@ void rdma_leave_multicast(struct rdma_cm_id *id, struct sockaddr *addr)
 }
 EXPORT_SYMBOL(rdma_leave_multicast);
 
+static int cma_netdev_change(struct net_device *ndev, struct rdma_id_private *id_priv)
+{
+	struct rdma_dev_addr *dev_addr;
+	struct cma_ndev_work *work;
+
+	dev_addr = &id_priv->id.route.addr.dev_addr;
+
+	if ((dev_addr->src_dev == ndev) &&
+	    memcmp(dev_addr->src_dev_addr, ndev->dev_addr, ndev->addr_len)) {
+		printk(KERN_INFO "RDMA CM addr change for ndev %s used by id %p\n",
+		       ndev->name, &id_priv->id);
+		work = kzalloc(sizeof *work, GFP_KERNEL);
+		if (!work)
+			return -ENOMEM;
+
+		INIT_WORK(&work->work, cma_ndev_work_handler);
+		work->id = id_priv;
+		work->event.event = RDMA_CM_EVENT_ADDR_CHANGE;
+		atomic_inc(&id_priv->refcount);
+		queue_work(cma_wq, &work->work);
+	}
+
+	return 0;
+}
+
+static int cma_netdev_callback(struct notifier_block *self, unsigned long event,
+			       void *ctx)
+{
+	struct net_device *ndev = (struct net_device *)ctx;
+	struct cma_device *cma_dev;
+	struct rdma_id_private *id_priv;
+	int ret = NOTIFY_DONE;
+
+	if (dev_net(ndev) != &init_net)
+		return NOTIFY_DONE;
+
+	if (event != NETDEV_BONDING_FAILOVER)
+		return NOTIFY_DONE;
+
+	if (!(ndev->flags & IFF_MASTER) || !(ndev->priv_flags & IFF_BONDING))
+		return NOTIFY_DONE;
+
+	mutex_lock(&lock);
+	list_for_each_entry(cma_dev, &dev_list, list)
+		list_for_each_entry(id_priv, &cma_dev->id_list, list) {
+			ret = cma_netdev_change(ndev, id_priv);
+			if (ret)
+				goto out;
+		}
+
+out:
+	mutex_unlock(&lock);
+	return ret;
+}
+
+static struct notifier_block cma_nb = {
+	.notifier_call = cma_netdev_callback
+};
+
 static void cma_add_one(struct ib_device *device)
 {
 	struct cma_device *cma_dev;
@@ -2831,6 +2925,7 @@ static int cma_init(void)
 
 	ib_sa_register_client(&sa_client);
 	rdma_addr_register_client(&addr_client);
+	register_netdevice_notifier(&cma_nb);
 
 	ret = ib_register_client(&cma_client);
 	if (ret)
@@ -2838,6 +2933,7 @@ static int cma_init(void)
 	return 0;
 
 err:
+	unregister_netdevice_notifier(&cma_nb);
 	rdma_addr_unregister_client(&addr_client);
 	ib_sa_unregister_client(&sa_client);
 	destroy_workqueue(cma_wq);
@@ -2847,6 +2943,7 @@ err:
 static void cma_cleanup(void)
 {
 	ib_unregister_client(&cma_client);
+	unregister_netdevice_notifier(&cma_nb);
 	rdma_addr_unregister_client(&addr_client);
 	ib_sa_unregister_client(&sa_client);
 	destroy_workqueue(cma_wq);
diff --git a/drivers/infiniband/core/iwcm.c b/drivers/infiniband/core/iwcm.c
index 81c9195b512..8f9509e1ebf 100644
--- a/drivers/infiniband/core/iwcm.c
+++ b/drivers/infiniband/core/iwcm.c
@@ -942,8 +942,7 @@ static int iwcm_init_qp_init_attr(struct iwcm_id_private *cm_id_priv,
 	case IW_CM_STATE_CONN_RECV:
 	case IW_CM_STATE_ESTABLISHED:
 		*qp_attr_mask = IB_QP_STATE | IB_QP_ACCESS_FLAGS;
-		qp_attr->qp_access_flags = IB_ACCESS_LOCAL_WRITE |
-					   IB_ACCESS_REMOTE_WRITE|
+		qp_attr->qp_access_flags = IB_ACCESS_REMOTE_WRITE|
 					   IB_ACCESS_REMOTE_READ;
 		ret = 0;
 		break;
diff --git a/drivers/infiniband/core/sa_query.c b/drivers/infiniband/core/sa_query.c
index 1341de793e5..7863a50d56f 100644
--- a/drivers/infiniband/core/sa_query.c
+++ b/drivers/infiniband/core/sa_query.c
@@ -1064,7 +1064,8 @@ static void ib_sa_remove_one(struct ib_device *device)
 
 	for (i = 0; i <= sa_dev->end_port - sa_dev->start_port; ++i) {
 		ib_unregister_mad_agent(sa_dev->port[i].agent);
-		kref_put(&sa_dev->port[i].sm_ah->ref, free_sm_ah);
+		if (sa_dev->port[i].sm_ah)
+			kref_put(&sa_dev->port[i].sm_ah->ref, free_sm_ah);
 	}
 
 	kfree(sa_dev);
diff --git a/drivers/infiniband/core/ucm.c b/drivers/infiniband/core/ucm.c
index 9494005d1c9..e603736682b 100644
--- a/drivers/infiniband/core/ucm.c
+++ b/drivers/infiniband/core/ucm.c
@@ -43,7 +43,6 @@
 #include <linux/cdev.h>
 #include <linux/idr.h>
 #include <linux/mutex.h>
-#include <linux/smp_lock.h>
 
 #include <asm/uaccess.h>
 
@@ -1154,11 +1153,18 @@ static unsigned int ib_ucm_poll(struct file *filp,
 	return mask;
 }
 
+/*
+ * ib_ucm_open() does not need the BKL:
+ *
+ *  - no global state is referred to;
+ *  - there is no ioctl method to race against;
+ *  - no further module initialization is required for open to work
+ *    after the device is registered.
+ */
 static int ib_ucm_open(struct inode *inode, struct file *filp)
 {
 	struct ib_ucm_file *file;
 
-	cycle_kernel_lock();
 	file = kmalloc(sizeof(*file), GFP_KERNEL);
 	if (!file)
 		return -ENOMEM;
diff --git a/drivers/infiniband/core/ucma.c b/drivers/infiniband/core/ucma.c
index 195f97302fe..b41dd26bbfa 100644
--- a/drivers/infiniband/core/ucma.c
+++ b/drivers/infiniband/core/ucma.c
@@ -38,7 +38,6 @@
 #include <linux/in.h>
 #include <linux/in6.h>
 #include <linux/miscdevice.h>
-#include <linux/smp_lock.h>
 
 #include <rdma/rdma_user_cm.h>
 #include <rdma/ib_marshall.h>
@@ -1149,6 +1148,14 @@ static unsigned int ucma_poll(struct file *filp, struct poll_table_struct *wait)
 	return mask;
 }
 
+/*
+ * ucma_open() does not need the BKL:
+ *
+ *  - no global state is referred to;
+ *  - there is no ioctl method to race against;
+ *  - no further module initialization is required for open to work
+ *    after the device is registered.
+ */
 static int ucma_open(struct inode *inode, struct file *filp)
 {
 	struct ucma_file *file;
@@ -1157,7 +1164,6 @@ static int ucma_open(struct inode *inode, struct file *filp)
 	if (!file)
 		return -ENOMEM;
 
-	lock_kernel();
 	INIT_LIST_HEAD(&file->event_list);
 	INIT_LIST_HEAD(&file->ctx_list);
 	init_waitqueue_head(&file->poll_wait);
@@ -1165,7 +1171,6 @@ static int ucma_open(struct inode *inode, struct file *filp)
 
 	filp->private_data = file;
 	file->filp = filp;
-	unlock_kernel();
 	return 0;
 }