Merge branches 'cxgb3', 'ehca', 'ipath', 'ipoib', 'misc', 'mlx4', 'mthca' and 'nes' into for-linus

MAINTAINERS

@@ -439,7 +439,7 @@ F:	drivers/hwmon/ams/
 AMSO1100 RNIC DRIVER
 M:	Tom Tucker <tom@opengridcomputing.com>
 M:	Steve Wise <swise@opengridcomputing.com>
-L:	general@lists.openfabrics.org
+L:	linux-rdma@vger.kernel.org
 S:	Maintained
 F:	drivers/infiniband/hw/amso1100/
 
@@ -1494,7 +1494,7 @@ F:	drivers/net/cxgb3/
 
 CXGB3 IWARP RNIC DRIVER (IW_CXGB3)
 M:	Steve Wise <swise@chelsio.com>
-L:	general@lists.openfabrics.org
+L:	linux-rdma@vger.kernel.org
 W:	http://www.openfabrics.org
 S:	Supported
 F:	drivers/infiniband/hw/cxgb3/
@@ -1868,7 +1868,7 @@ F:	fs/efs/
 EHCA (IBM GX bus InfiniBand adapter) DRIVER
 M:	Hoang-Nam Nguyen <hnguyen@de.ibm.com>
 M:	Christoph Raisch <raisch@de.ibm.com>
-L:	general@lists.openfabrics.org
+L:	linux-rdma@vger.kernel.org
 S:	Supported
 F:	drivers/infiniband/hw/ehca/
 
@@ -2552,7 +2552,7 @@ INFINIBAND SUBSYSTEM
 M:	Roland Dreier <rolandd@cisco.com>
 M:	Sean Hefty <sean.hefty@intel.com>
 M:	Hal Rosenstock <hal.rosenstock@gmail.com>
-L:	general@lists.openfabrics.org (moderated for non-subscribers)
+L:	linux-rdma@vger.kernel.org
 W:	http://www.openib.org/
 T:	git git://git.kernel.org/pub/scm/linux/kernel/git/roland/infiniband.git
 S:	Supported
@@ -2729,7 +2729,7 @@ F:	drivers/net/ipg.c
 
 IPATH DRIVER
 M:	Ralph Campbell <infinipath@qlogic.com>
-L:	general@lists.openfabrics.org
+L:	linux-rdma@vger.kernel.org
 T:	git git://git.qlogic.com/ipath-linux-2.6
 S:	Supported
 F:	drivers/infiniband/hw/ipath/
@@ -3485,7 +3485,7 @@ F:	drivers/scsi/NCR_D700.*
 NETEFFECT IWARP RNIC DRIVER (IW_NES)
 M:	Faisal Latif <faisal.latif@intel.com>
 M:	Chien Tung <chien.tin.tung@intel.com>
-L:	general@lists.openfabrics.org
+L:	linux-rdma@vger.kernel.org
 W:	http://www.neteffect.com
 S:	Supported
 F:	drivers/infiniband/hw/nes/

drivers/infiniband/core/iwcm.c

@@ -362,6 +362,7 @@ static void destroy_cm_id(struct iw_cm_id *cm_id)
 		 * In either case, must tell the provider to reject.
 		 */
 		cm_id_priv->state = IW_CM_STATE_DESTROYING;
+		cm_id->device->iwcm->reject(cm_id, NULL, 0);
 		break;
 	case IW_CM_STATE_CONN_SENT:
 	case IW_CM_STATE_DESTROYING:

drivers/infiniband/core/mad.c

@@ -51,8 +51,7 @@ static struct list_head ib_mad_port_list;
 static u32 ib_mad_client_id = 0;
 
 /* Port list lock */
-static spinlock_t ib_mad_port_list_lock;
-
+static DEFINE_SPINLOCK(ib_mad_port_list_lock);
 
 /* Forward declarations */
 static int method_in_use(struct ib_mad_mgmt_method_table **method,
@@ -2984,8 +2983,6 @@ static int __init ib_mad_init_module(void)
 {
 	int ret;
 
-	spin_lock_init(&ib_mad_port_list_lock);
-
 	ib_mad_cache = kmem_cache_create("ib_mad",
 					 sizeof(struct ib_mad_private),
 					 0,
@@ -3021,4 +3018,3 @@ static void __exit ib_mad_cleanup_module(void)
 
 module_init(ib_mad_init_module);
 module_exit(ib_mad_cleanup_module);
-

drivers/infiniband/core/multicast.c

@@ -106,6 +106,8 @@ struct mcast_group {
 	struct ib_sa_query	*query;
 	int			query_id;
 	u16			pkey_index;
+	u8			leave_state;
+	int			retries;
 };
 
 struct mcast_member {
@@ -350,6 +352,7 @@ static int send_leave(struct mcast_group *group, u8 leave_state)
 
 	rec = group->rec;
 	rec.join_state = leave_state;
+	group->leave_state = leave_state;
 
 	ret = ib_sa_mcmember_rec_query(&sa_client, port->dev->device,
 				       port->port_num, IB_SA_METHOD_DELETE, &rec,
@@ -542,7 +545,11 @@ static void leave_handler(int status, struct ib_sa_mcmember_rec *rec,
 {
 	struct mcast_group *group = context;
 
-	mcast_work_handler(&group->work);
+	if (status && group->retries > 0 &&
+	    !send_leave(group, group->leave_state))
+		group->retries--;
+	else
+		mcast_work_handler(&group->work);
 }
 
 static struct mcast_group *acquire_group(struct mcast_port *port,
@@ -565,6 +572,7 @@ static struct mcast_group *acquire_group(struct mcast_port *port,
 	if (!group)
 		return NULL;
 
+	group->retries = 3;
 	group->port = port;
 	group->rec.mgid = *mgid;
 	group->pkey_index = MCAST_INVALID_PKEY_INDEX;

drivers/infiniband/core/sa_query.c

@@ -109,10 +109,10 @@ static struct ib_client sa_client = {
 	.remove = ib_sa_remove_one
 };
 
-static spinlock_t idr_lock;
+static DEFINE_SPINLOCK(idr_lock);
 static DEFINE_IDR(query_idr);
 
-static spinlock_t tid_lock;
+static DEFINE_SPINLOCK(tid_lock);
 static u32 tid;
 
 #define PATH_REC_FIELD(field) \
@@ -1077,9 +1077,6 @@ static int __init ib_sa_init(void)
 {
 	int ret;
 
-	spin_lock_init(&idr_lock);
-	spin_lock_init(&tid_lock);
-
 	get_random_bytes(&tid, sizeof tid);
 
 	ret = ib_register_client(&sa_client);

drivers/infiniband/core/uverbs_main.c

@@ -73,7 +73,7 @@ DEFINE_IDR(ib_uverbs_cq_idr);
 DEFINE_IDR(ib_uverbs_qp_idr);
 DEFINE_IDR(ib_uverbs_srq_idr);
 
-static spinlock_t map_lock;
+static DEFINE_SPINLOCK(map_lock);
 static struct ib_uverbs_device *dev_table[IB_UVERBS_MAX_DEVICES];
 static DECLARE_BITMAP(dev_map, IB_UVERBS_MAX_DEVICES);
 
@@ -584,14 +584,16 @@ static ssize_t ib_uverbs_write(struct file *filp, const char __user *buf,
 
 	if (hdr.command < 0				||
 	    hdr.command >= ARRAY_SIZE(uverbs_cmd_table) ||
-	    !uverbs_cmd_table[hdr.command]		||
-	    !(file->device->ib_dev->uverbs_cmd_mask & (1ull << hdr.command)))
+	    !uverbs_cmd_table[hdr.command])
 		return -EINVAL;
 
 	if (!file->ucontext &&
 	    hdr.command != IB_USER_VERBS_CMD_GET_CONTEXT)
 		return -EINVAL;
 
+	if (!(file->device->ib_dev->uverbs_cmd_mask & (1ull << hdr.command)))
+		return -ENOSYS;
+
 	return uverbs_cmd_table[hdr.command](file, buf + sizeof hdr,
 					     hdr.in_words * 4, hdr.out_words * 4);
 }
@@ -836,8 +838,6 @@ static int __init ib_uverbs_init(void)
 {
 	int ret;
 
-	spin_lock_init(&map_lock);
-
 	ret = register_chrdev_region(IB_UVERBS_BASE_DEV, IB_UVERBS_MAX_DEVICES,
 				     "infiniband_verbs");
 	if (ret) {

drivers/infiniband/hw/amso1100/c2.c

@@ -86,11 +86,7 @@ MODULE_DEVICE_TABLE(pci, c2_pci_table);
 
 static void c2_print_macaddr(struct net_device *netdev)
 {
-	pr_debug("%s: MAC %02X:%02X:%02X:%02X:%02X:%02X, "
-		"IRQ %u\n", netdev->name,
-		netdev->dev_addr[0], netdev->dev_addr[1], netdev->dev_addr[2],
-		netdev->dev_addr[3], netdev->dev_addr[4], netdev->dev_addr[5],
-		netdev->irq);
+	pr_debug("%s: MAC %pM, IRQ %u\n", netdev->name, netdev->dev_addr, netdev->irq);
 }
 
 static void c2_set_rxbufsize(struct c2_port *c2_port)

drivers/infiniband/hw/amso1100/c2_provider.c

@@ -780,11 +780,11 @@ int c2_register_device(struct c2_dev *dev)
 	/* Register pseudo network device */
 	dev->pseudo_netdev = c2_pseudo_netdev_init(dev);
 	if (!dev->pseudo_netdev)
-		goto out3;
+		goto out;
 
 	ret = register_netdev(dev->pseudo_netdev);
 	if (ret)
-		goto out2;
+		goto out_free_netdev;
 
 	pr_debug("%s:%u\n", __func__, __LINE__);
 	strlcpy(dev->ibdev.name, "amso%d", IB_DEVICE_NAME_MAX);
@@ -851,6 +851,10 @@ int c2_register_device(struct c2_dev *dev)
 	dev->ibdev.post_recv = c2_post_receive;
 
 	dev->ibdev.iwcm = kmalloc(sizeof(*dev->ibdev.iwcm), GFP_KERNEL);
+	if (dev->ibdev.iwcm == NULL) {
+		ret = -ENOMEM;
+		goto out_unregister_netdev;
+	}
 	dev->ibdev.iwcm->add_ref = c2_add_ref;
 	dev->ibdev.iwcm->rem_ref = c2_rem_ref;
 	dev->ibdev.iwcm->get_qp = c2_get_qp;
@@ -862,23 +866,25 @@ int c2_register_device(struct c2_dev *dev)
 
 	ret = ib_register_device(&dev->ibdev);
 	if (ret)
-		goto out1;
+		goto out_free_iwcm;
 
 	for (i = 0; i < ARRAY_SIZE(c2_dev_attributes); ++i) {
 		ret = device_create_file(&dev->ibdev.dev,
 					       c2_dev_attributes[i]);
 		if (ret)
-			goto out0;
+			goto out_unregister_ibdev;
 	}
-	goto out3;
+	goto out;
 
-out0:
+out_unregister_ibdev:
 	ib_unregister_device(&dev->ibdev);
-out1:
+out_free_iwcm:
+	kfree(dev->ibdev.iwcm);
+out_unregister_netdev:
 	unregister_netdev(dev->pseudo_netdev);
-out2:
+out_free_netdev:
 	free_netdev(dev->pseudo_netdev);
-out3:
+out:
 	pr_debug("%s:%u ret=%d\n", __func__, __LINE__, ret);
 	return ret;
 }

drivers/infiniband/hw/cxgb3/iwch.c

@@ -105,11 +105,9 @@ static void rnic_init(struct iwch_dev *rnicp)
 static void open_rnic_dev(struct t3cdev *tdev)
 {
 	struct iwch_dev *rnicp;
-	static int vers_printed;
 
 	PDBG("%s t3cdev %p\n", __func__,  tdev);
-	if (!vers_printed++)
-		printk(KERN_INFO MOD "Chelsio T3 RDMA Driver - version %s\n",
+	printk_once(KERN_INFO MOD "Chelsio T3 RDMA Driver - version %s\n",
 		       DRV_VERSION);
 	rnicp = (struct iwch_dev *)ib_alloc_device(sizeof(*rnicp));
 	if (!rnicp) {

drivers/infiniband/hw/ehca/ehca_main.c

@@ -52,7 +52,7 @@
 #include "ehca_tools.h"
 #include "hcp_if.h"
 
-#define HCAD_VERSION "0028"
+#define HCAD_VERSION "0029"
 
 MODULE_LICENSE("Dual BSD/GPL");
 MODULE_AUTHOR("Christoph Raisch <raisch@de.ibm.com>");
@@ -64,7 +64,7 @@ static int ehca_hw_level      = 0;
 static int ehca_poll_all_eqs  = 1;
 
 int ehca_debug_level   = 0;
-int ehca_nr_ports      = 2;
+int ehca_nr_ports      = -1;
 int ehca_use_hp_mr     = 0;
 int ehca_port_act_time = 30;
 int ehca_static_rate   = -1;
@@ -95,8 +95,8 @@ MODULE_PARM_DESC(hw_level,
 		 "Hardware level (0: autosensing (default), "
 		 "0x10..0x14: eHCA, 0x20..0x23: eHCA2)");
 MODULE_PARM_DESC(nr_ports,
-		 "number of connected ports (-1: autodetect, 1: port one only, "
-		 "2: two ports (default)");
+		 "number of connected ports (-1: autodetect (default), "
+		 "1: port one only, 2: two ports)");
 MODULE_PARM_DESC(use_hp_mr,
 		 "Use high performance MRs (default: no)");
 MODULE_PARM_DESC(port_act_time,

drivers/infiniband/hw/ehca/ehca_reqs.c

@@ -786,7 +786,11 @@ static inline int ehca_poll_cq_one(struct ib_cq *cq, struct ib_wc *wc)
 	wc->slid = cqe->rlid;
 	wc->dlid_path_bits = cqe->dlid;
 	wc->src_qp = cqe->remote_qp_number;
-	wc->wc_flags = cqe->w_completion_flags;
+	/*
+	 * HW has "Immed data present" and "GRH present" in bits 6 and 5.
+	 * SW defines those in bits 1 and 0, so we can just shift and mask.
+	 */
+	wc->wc_flags = (cqe->w_completion_flags >> 5) & 3;
 	wc->ex.imm_data = cpu_to_be32(cqe->immediate_data);
 	wc->sl = cqe->service_level;
 

drivers/infiniband/hw/ehca/ehca_sqp.c

@@ -125,14 +125,30 @@ struct ib_perf {
 	u8 data[192];
 } __attribute__ ((packed));
 
+/* TC/SL/FL packed into 32 bits, as in ClassPortInfo */
+struct tcslfl {
+	u32 tc:8;
+	u32 sl:4;
+	u32 fl:20;
+} __attribute__ ((packed));
+
+/* IP Version/TC/FL packed into 32 bits, as in GRH */
+struct vertcfl {
+	u32 ver:4;
+	u32 tc:8;
+	u32 fl:20;
+} __attribute__ ((packed));
 
 static int ehca_process_perf(struct ib_device *ibdev, u8 port_num,
+			     struct ib_wc *in_wc, struct ib_grh *in_grh,
 			     struct ib_mad *in_mad, struct ib_mad *out_mad)
 {
 	struct ib_perf *in_perf = (struct ib_perf *)in_mad;
 	struct ib_perf *out_perf = (struct ib_perf *)out_mad;
 	struct ib_class_port_info *poi =
 		(struct ib_class_port_info *)out_perf->data;
+	struct tcslfl *tcslfl =
+		(struct tcslfl *)&poi->redirect_tcslfl;
 	struct ehca_shca *shca =
 		container_of(ibdev, struct ehca_shca, ib_device);
 	struct ehca_sport *sport = &shca->sport[port_num - 1];
@@ -158,10 +174,29 @@ static int ehca_process_perf(struct ib_device *ibdev, u8 port_num,
 		poi->base_version = 1;
 		poi->class_version = 1;
 		poi->resp_time_value = 18;
-		poi->redirect_lid = sport->saved_attr.lid;
-		poi->redirect_qp = sport->pma_qp_nr;
+
+		/* copy local routing information from WC where applicable */
+		tcslfl->sl         = in_wc->sl;
+		poi->redirect_lid  =
+			sport->saved_attr.lid | in_wc->dlid_path_bits;
+		poi->redirect_qp   = sport->pma_qp_nr;
 		poi->redirect_qkey = IB_QP1_QKEY;
-		poi->redirect_pkey = IB_DEFAULT_PKEY_FULL;
+
+		ehca_query_pkey(ibdev, port_num, in_wc->pkey_index,
+				&poi->redirect_pkey);
+
+		/* if request was globally routed, copy route info */
+		if (in_grh) {
+			struct vertcfl *vertcfl =
+				(struct vertcfl *)&in_grh->version_tclass_flow;
+			memcpy(poi->redirect_gid, in_grh->dgid.raw,
+			       sizeof(poi->redirect_gid));
+			tcslfl->tc        = vertcfl->tc;
+			tcslfl->fl        = vertcfl->fl;
+		} else
+			/* else only fill in default GID */
+			ehca_query_gid(ibdev, port_num, 0,
+				       (union ib_gid *)&poi->redirect_gid);
 
 		ehca_dbg(ibdev, "ehca_pma_lid=%x ehca_pma_qp=%x",
 			 sport->saved_attr.lid, sport->pma_qp_nr);
@@ -183,8 +218,7 @@ static int ehca_process_perf(struct ib_device *ibdev, u8 port_num,
 
 int ehca_process_mad(struct ib_device *ibdev, int mad_flags, u8 port_num,
 		     struct ib_wc *in_wc, struct ib_grh *in_grh,
-		     struct ib_mad *in_mad,
-		     struct ib_mad *out_mad)
+		     struct ib_mad *in_mad, struct ib_mad *out_mad)
 {
 	int ret;
 
@@ -196,7 +230,8 @@ int ehca_process_mad(struct ib_device *ibdev, int mad_flags, u8 port_num,
 		return IB_MAD_RESULT_SUCCESS;
 
 	ehca_dbg(ibdev, "port_num=%x src_qp=%x", port_num, in_wc->src_qp);
-	ret = ehca_process_perf(ibdev, port_num, in_mad, out_mad);
+	ret = ehca_process_perf(ibdev, port_num, in_wc, in_grh,
+				in_mad, out_mad);
 
 	return ret;
 }

drivers/infiniband/hw/ipath/ipath_file_ops.c

@@ -1616,7 +1616,7 @@ static int try_alloc_port(struct ipath_devdata *dd, int port,
 		pd->port_cnt = 1;
 		port_fp(fp) = pd;
 		pd->port_pid = get_pid(task_pid(current));
-		strncpy(pd->port_comm, current->comm, sizeof(pd->port_comm));
+		strlcpy(pd->port_comm, current->comm, sizeof(pd->port_comm));
 		ipath_stats.sps_ports++;
 		ret = 0;
 	} else

drivers/infiniband/hw/ipath/ipath_mad.c

@@ -60,7 +60,7 @@ static int recv_subn_get_nodedescription(struct ib_smp *smp,
 	if (smp->attr_mod)
 		smp->status |= IB_SMP_INVALID_FIELD;
 
-	strncpy(smp->data, ibdev->node_desc, sizeof(smp->data));
+	memcpy(smp->data, ibdev->node_desc, sizeof(smp->data));
 
 	return reply(smp);
 }

drivers/infiniband/hw/mlx4/main.c

@@ -342,6 +342,9 @@ static struct ib_ucontext *mlx4_ib_alloc_ucontext(struct ib_device *ibdev,
 	struct mlx4_ib_alloc_ucontext_resp resp;
 	int err;
 
+	if (!dev->ib_active)
+		return ERR_PTR(-EAGAIN);
+
 	resp.qp_tab_size      = dev->dev->caps.num_qps;
 	resp.bf_reg_size      = dev->dev->caps.bf_reg_size;
 	resp.bf_regs_per_page = dev->dev->caps.bf_regs_per_page;
@@ -540,15 +543,11 @@ static struct device_attribute *mlx4_class_attributes[] = {
 
 static void *mlx4_ib_add(struct mlx4_dev *dev)
 {
-	static int mlx4_ib_version_printed;
 	struct mlx4_ib_dev *ibdev;
 	int num_ports = 0;
 	int i;
 
-	if (!mlx4_ib_version_printed) {
-		printk(KERN_INFO "%s", mlx4_ib_version);
-		++mlx4_ib_version_printed;
-	}
+	printk_once(KERN_INFO "%s", mlx4_ib_version);
 
 	mlx4_foreach_port(i, dev, MLX4_PORT_TYPE_IB)
 		num_ports++;
@@ -673,6 +672,8 @@ static void *mlx4_ib_add(struct mlx4_dev *dev)
 			goto err_reg;
 	}
 
+	ibdev->ib_active = true;
+
 	return ibdev;
 
 err_reg:
@@ -729,6 +730,7 @@ static void mlx4_ib_event(struct mlx4_dev *dev, void *ibdev_ptr,
 		break;
 
 	case MLX4_DEV_EVENT_CATASTROPHIC_ERROR:
+		ibdev->ib_active = false;
 		ibev.event = IB_EVENT_DEVICE_FATAL;
 		break;
 

drivers/infiniband/hw/mlx4/mlx4_ib.h

@@ -175,6 +175,7 @@ struct mlx4_ib_dev {
 	spinlock_t		sm_lock;
 
 	struct mutex		cap_mask_mutex;
+	bool			ib_active;
 };
 
 static inline struct mlx4_ib_dev *to_mdev(struct ib_device *ibdev)

drivers/infiniband/hw/mlx4/qp.c

@@ -615,10 +615,12 @@ static enum mlx4_qp_state to_mlx4_state(enum ib_qp_state state)
 }
 
 static void mlx4_ib_lock_cqs(struct mlx4_ib_cq *send_cq, struct mlx4_ib_cq *recv_cq)
+	__acquires(&send_cq->lock) __acquires(&recv_cq->lock)
 {
-	if (send_cq == recv_cq)
+	if (send_cq == recv_cq) {
 		spin_lock_irq(&send_cq->lock);
-	else if (send_cq->mcq.cqn < recv_cq->mcq.cqn) {
+		__acquire(&recv_cq->lock);
+	} else if (send_cq->mcq.cqn < recv_cq->mcq.cqn) {
 		spin_lock_irq(&send_cq->lock);
 		spin_lock_nested(&recv_cq->lock, SINGLE_DEPTH_NESTING);
 	} else {
@@ -628,10 +630,12 @@ static void mlx4_ib_lock_cqs(struct mlx4_ib_cq *send_cq, struct mlx4_ib_cq *recv
 }
 
 static void mlx4_ib_unlock_cqs(struct mlx4_ib_cq *send_cq, struct mlx4_ib_cq *recv_cq)
+	__releases(&send_cq->lock) __releases(&recv_cq->lock)
 {
-	if (send_cq == recv_cq)
+	if (send_cq == recv_cq) {
+		__release(&recv_cq->lock);
 		spin_unlock_irq(&send_cq->lock);
-	else if (send_cq->mcq.cqn < recv_cq->mcq.cqn) {
+	} else if (send_cq->mcq.cqn < recv_cq->mcq.cqn) {
 		spin_unlock(&recv_cq->lock);
 		spin_unlock_irq(&send_cq->lock);
 	} else {

drivers/infiniband/hw/mthca/mthca_catas.c

@@ -88,6 +88,7 @@ static void handle_catas(struct mthca_dev *dev)
 	event.device = &dev->ib_dev;
 	event.event  = IB_EVENT_DEVICE_FATAL;
 	event.element.port_num = 0;
+	dev->active = false;
 
 	ib_dispatch_event(&event);
 

drivers/infiniband/hw/mthca/mthca_config_reg.h

@@ -34,8 +34,6 @@
 #ifndef MTHCA_CONFIG_REG_H
 #define MTHCA_CONFIG_REG_H
 
-#include <asm/page.h>
-
 #define MTHCA_HCR_BASE         0x80680
 #define MTHCA_HCR_SIZE         0x0001c
 #define MTHCA_ECR_BASE         0x80700

drivers/infiniband/hw/mthca/mthca_dev.h

@@ -357,6 +357,7 @@ struct mthca_dev {
 	struct ib_ah         *sm_ah[MTHCA_MAX_PORTS];
 	spinlock_t            sm_lock;
 	u8                    rate[MTHCA_MAX_PORTS];
+	bool		      active;
 };
 
 #ifdef CONFIG_INFINIBAND_MTHCA_DEBUG

drivers/infiniband/hw/mthca/mthca_eq.c

@@ -829,27 +829,34 @@ int mthca_init_eq_table(struct mthca_dev *dev)
 
 	if (dev->mthca_flags & MTHCA_FLAG_MSI_X) {
 		static const char *eq_name[] = {
-			[MTHCA_EQ_COMP]  = DRV_NAME " (comp)",
-			[MTHCA_EQ_ASYNC] = DRV_NAME " (async)",
-			[MTHCA_EQ_CMD]   = DRV_NAME " (cmd)"
+			[MTHCA_EQ_COMP]  = DRV_NAME "-comp",
+			[MTHCA_EQ_ASYNC] = DRV_NAME "-async",
+			[MTHCA_EQ_CMD]   = DRV_NAME "-cmd"
 		};
 
 		for (i = 0; i < MTHCA_NUM_EQ; ++i) {
+			snprintf(dev->eq_table.eq[i].irq_name,
+				 IB_DEVICE_NAME_MAX,
+				 "%s@pci:%s", eq_name[i],
+				 pci_name(dev->pdev));
 			err = request_irq(dev->eq_table.eq[i].msi_x_vector,
 					  mthca_is_memfree(dev) ?
 					  mthca_arbel_msi_x_interrupt :
 					  mthca_tavor_msi_x_interrupt,
-					  0, eq_name[i], dev->eq_table.eq + i);
+					  0, dev->eq_table.eq[i].irq_name,
+					  dev->eq_table.eq + i);
 			if (err)
 				goto err_out_cmd;
 			dev->eq_table.eq[i].have_irq = 1;
 		}
 	} else {
+		snprintf(dev->eq_table.eq[0].irq_name, IB_DEVICE_NAME_MAX,
+			 DRV_NAME "@pci:%s", pci_name(dev->pdev));
 		err = request_irq(dev->pdev->irq,
 				  mthca_is_memfree(dev) ?
 				  mthca_arbel_interrupt :
 				  mthca_tavor_interrupt,
-				  IRQF_SHARED, DRV_NAME, dev);
+				  IRQF_SHARED, dev->eq_table.eq[0].irq_name, dev);
 		if (err)
 			goto err_out_cmd;
 		dev->eq_table.have_irq = 1;

drivers/infiniband/hw/mthca/mthca_main.c

@@ -1116,6 +1116,8 @@ static int __mthca_init_one(struct pci_dev *pdev, int hca_type)
 	pci_set_drvdata(pdev, mdev);
 	mdev->hca_type = hca_type;
 
+	mdev->active = true;
+
 	return 0;
 
 err_unregister:
@@ -1215,15 +1217,11 @@ int __mthca_restart_one(struct pci_dev *pdev)
 static int __devinit mthca_init_one(struct pci_dev *pdev,
 				    const struct pci_device_id *id)
 {
-	static int mthca_version_printed = 0;
 	int ret;
 
 	mutex_lock(&mthca_device_mutex);
 
-	if (!mthca_version_printed) {
-		printk(KERN_INFO "%s", mthca_version);
-		++mthca_version_printed;
-	}
+	printk_once(KERN_INFO "%s", mthca_version);
 
 	if (id->driver_data >= ARRAY_SIZE(mthca_hca_table)) {
 		printk(KERN_ERR PFX "%s has invalid driver data %lx\n",

drivers/infiniband/hw/mthca/mthca_provider.c

@@ -334,6 +334,9 @@ static struct ib_ucontext *mthca_alloc_ucontext(struct ib_device *ibdev,
 	struct mthca_ucontext           *context;
 	int                              err;
 
+	if (!(to_mdev(ibdev)->active))
+		return ERR_PTR(-EAGAIN);
+
 	memset(&uresp, 0, sizeof uresp);
 
 	uresp.qp_tab_size = to_mdev(ibdev)->limits.num_qps;

drivers/infiniband/hw/mthca/mthca_provider.h

@@ -113,6 +113,7 @@ struct mthca_eq {
 	int                    nent;
 	struct mthca_buf_list *page_list;
 	struct mthca_mr        mr;
+	char		       irq_name[IB_DEVICE_NAME_MAX];
 };
 
 struct mthca_av;

drivers/infiniband/hw/mthca/mthca_qp.c

@@ -1319,10 +1319,12 @@ int mthca_alloc_qp(struct mthca_dev *dev,
 }
 
 static void mthca_lock_cqs(struct mthca_cq *send_cq, struct mthca_cq *recv_cq)
+	__acquires(&send_cq->lock) __acquires(&recv_cq->lock)
 {
-	if (send_cq == recv_cq)
+	if (send_cq == recv_cq) {
 		spin_lock_irq(&send_cq->lock);
-	else if (send_cq->cqn < recv_cq->cqn) {
+		__acquire(&recv_cq->lock);
+	} else if (send_cq->cqn < recv_cq->cqn) {
 		spin_lock_irq(&send_cq->lock);
 		spin_lock_nested(&recv_cq->lock, SINGLE_DEPTH_NESTING);
 	} else {
@@ -1332,10 +1334,12 @@ static void mthca_lock_cqs(struct mthca_cq *send_cq, struct mthca_cq *recv_cq)
 }
 
 static void mthca_unlock_cqs(struct mthca_cq *send_cq, struct mthca_cq *recv_cq)
+	__releases(&send_cq->lock) __releases(&recv_cq->lock)
 {
-	if (send_cq == recv_cq)
+	if (send_cq == recv_cq) {
+		__release(&recv_cq->lock);
 		spin_unlock_irq(&send_cq->lock);
-	else if (send_cq->cqn < recv_cq->cqn) {
+	} else if (send_cq->cqn < recv_cq->cqn) {
 		spin_unlock(&recv_cq->lock);
 		spin_unlock_irq(&send_cq->lock);
 	} else {

drivers/infiniband/hw/mthca/mthca_reset.c

@@ -30,7 +30,6 @@
  * SOFTWARE.
  */
 
-#include <linux/init.h>
 #include <linux/errno.h>
 #include <linux/pci.h>
 #include <linux/delay.h>

drivers/infiniband/hw/nes/nes.h

@@ -523,7 +523,7 @@ int nes_cm_disconn(struct nes_qp *);
 void nes_cm_disconn_worker(void *);
 
 /* nes_verbs.c */
-int nes_hw_modify_qp(struct nes_device *, struct nes_qp *, u32, u32);
+int nes_hw_modify_qp(struct nes_device *, struct nes_qp *, u32, u32, u32);
 int nes_modify_qp(struct ib_qp *, struct ib_qp_attr *, int, struct ib_udata *);
 struct nes_ib_device *nes_init_ofa_device(struct net_device *);
 void nes_destroy_ofa_device(struct nes_ib_device *);

drivers/infiniband/hw/nes/nes_cm.c

@@ -2450,19 +2450,16 @@ static int nes_cm_init_tsa_conn(struct nes_qp *nesqp, struct nes_cm_node *cm_nod
  */
 int nes_cm_disconn(struct nes_qp *nesqp)
 {
-	unsigned long flags;
-
-	spin_lock_irqsave(&nesqp->lock, flags);
-	if (nesqp->disconn_pending == 0) {
-		nesqp->disconn_pending++;
-		spin_unlock_irqrestore(&nesqp->lock, flags);
-		/* init our disconnect work element, to */
-		INIT_WORK(&nesqp->disconn_work, nes_disconnect_worker);
+	struct disconn_work *work;
 
-		queue_work(g_cm_core->disconn_wq, &nesqp->disconn_work);
-	} else
-		spin_unlock_irqrestore(&nesqp->lock, flags);
+	work = kzalloc(sizeof *work, GFP_ATOMIC);
+	if (!work)
+		return -ENOMEM; /* Timer will clean up */
 
+	nes_add_ref(&nesqp->ibqp);
+	work->nesqp = nesqp;
+	INIT_WORK(&work->work, nes_disconnect_worker);
+	queue_work(g_cm_core->disconn_wq, &work->work);
 	return 0;
 }
 
@@ -2472,11 +2469,14 @@ int nes_cm_disconn(struct nes_qp *nesqp)
  */
 static void nes_disconnect_worker(struct work_struct *work)
 {
-	struct nes_qp *nesqp = container_of(work, struct nes_qp, disconn_work);
+	struct disconn_work *dwork = container_of(work, struct disconn_work, work);
+	struct nes_qp *nesqp = dwork->nesqp;
 
+	kfree(dwork);
 	nes_debug(NES_DBG_CM, "processing AEQE id 0x%04X for QP%u.\n",
 			nesqp->last_aeq, nesqp->hwqp.qp_id);
 	nes_cm_disconn_true(nesqp);
+	nes_rem_ref(&nesqp->ibqp);
 }
 
 
@@ -2493,7 +2493,12 @@ static int nes_cm_disconn_true(struct nes_qp *nesqp)
 	u16 last_ae;
 	u8 original_hw_tcp_state;
 	u8 original_ibqp_state;
-	u8 issued_disconnect_reset = 0;
+	enum iw_cm_event_type disconn_status = IW_CM_EVENT_STATUS_OK;
+	int issue_disconn = 0;
+	int issue_close = 0;
+	int issue_flush = 0;
+	u32 flush_q = NES_CQP_FLUSH_RQ;
+	struct ib_event ibevent;
 
 	if (!nesqp) {
 		nes_debug(NES_DBG_CM, "disconnect_worker nesqp is NULL\n");
@@ -2517,24 +2522,55 @@ static int nes_cm_disconn_true(struct nes_qp *nesqp)
 	original_ibqp_state   = nesqp->ibqp_state;
 	last_ae = nesqp->last_aeq;
 
+	if (nesqp->term_flags) {
+		issue_disconn = 1;
+		issue_close = 1;
+		nesqp->cm_id = NULL;
+		if (nesqp->flush_issued == 0) {
+			nesqp->flush_issued = 1;
+			issue_flush = 1;
+		}
+	} else if ((original_hw_tcp_state == NES_AEQE_TCP_STATE_CLOSE_WAIT) ||
+			((original_ibqp_state == IB_QPS_RTS) &&
+			(last_ae == NES_AEQE_AEID_LLP_CONNECTION_RESET))) {
+		issue_disconn = 1;
+		if (last_ae == NES_AEQE_AEID_LLP_CONNECTION_RESET)
+			disconn_status = IW_CM_EVENT_STATUS_RESET;
+	}
+
+	if (((original_hw_tcp_state == NES_AEQE_TCP_STATE_CLOSED) ||
+		 (original_hw_tcp_state == NES_AEQE_TCP_STATE_TIME_WAIT) ||
+		 (last_ae == NES_AEQE_AEID_RDMAP_ROE_BAD_LLP_CLOSE) ||
+		 (last_ae == NES_AEQE_AEID_LLP_CONNECTION_RESET))) {
+		issue_close = 1;
+		nesqp->cm_id = NULL;
+		if (nesqp->flush_issued == 0) {
+			nesqp->flush_issued = 1;
+			issue_flush = 1;
+		}
+	}
+
+	spin_unlock_irqrestore(&nesqp->lock, flags);
 
-	nes_debug(NES_DBG_CM, "set ibqp_state=%u\n", nesqp->ibqp_state);
+	if ((issue_flush) && (nesqp->destroyed == 0)) {
+		/* Flush the queue(s) */
+		if (nesqp->hw_iwarp_state >= NES_AEQE_IWARP_STATE_TERMINATE)
+			flush_q |= NES_CQP_FLUSH_SQ;
+		flush_wqes(nesvnic->nesdev, nesqp, flush_q, 1);
 
-	if ((nesqp->cm_id) && (cm_id->event_handler)) {
-		if ((original_hw_tcp_state == NES_AEQE_TCP_STATE_CLOSE_WAIT) ||
-				((original_ibqp_state == IB_QPS_RTS) &&
-				(last_ae == NES_AEQE_AEID_LLP_CONNECTION_RESET))) {
+		if (nesqp->term_flags) {
+			ibevent.device = nesqp->ibqp.device;
+			ibevent.event = nesqp->terminate_eventtype;
+			ibevent.element.qp = &nesqp->ibqp;
+			nesqp->ibqp.event_handler(&ibevent, nesqp->ibqp.qp_context);
+		}
+	}
+
+	if ((cm_id) && (cm_id->event_handler)) {
+		if (issue_disconn) {
 			atomic_inc(&cm_disconnects);
 			cm_event.event = IW_CM_EVENT_DISCONNECT;
-			if (last_ae == NES_AEQE_AEID_LLP_CONNECTION_RESET) {
-				cm_event.status = IW_CM_EVENT_STATUS_RESET;
-				nes_debug(NES_DBG_CM, "Generating a CM "
-					"Disconnect Event (status reset) for "
-					"QP%u, cm_id = %p. \n",
-					nesqp->hwqp.qp_id, cm_id);
-			} else
-				cm_event.status = IW_CM_EVENT_STATUS_OK;
-
+			cm_event.status = disconn_status;
 			cm_event.local_addr = cm_id->local_addr;
 			cm_event.remote_addr = cm_id->remote_addr;
 			cm_event.private_data = NULL;
@@ -2547,29 +2583,14 @@ static int nes_cm_disconn_true(struct nes_qp *nesqp)
 				nesqp->hwqp.sq_tail, cm_id,
 				atomic_read(&nesqp->refcount));
 
-			spin_unlock_irqrestore(&nesqp->lock, flags);
 			ret = cm_id->event_handler(cm_id, &cm_event);
 			if (ret)
 				nes_debug(NES_DBG_CM, "OFA CM event_handler "
 					"returned, ret=%d\n", ret);
-			spin_lock_irqsave(&nesqp->lock, flags);
 		}
 
-		nesqp->disconn_pending = 0;
-		/* There might have been another AE while the lock was released */
-		original_hw_tcp_state = nesqp->hw_tcp_state;
-		original_ibqp_state   = nesqp->ibqp_state;
-		last_ae = nesqp->last_aeq;
-
-		if ((issued_disconnect_reset == 0) && (nesqp->cm_id) &&
-				((original_hw_tcp_state == NES_AEQE_TCP_STATE_CLOSED) ||
-				 (original_hw_tcp_state == NES_AEQE_TCP_STATE_TIME_WAIT) ||
-				 (last_ae == NES_AEQE_AEID_RDMAP_ROE_BAD_LLP_CLOSE) ||
-				 (last_ae == NES_AEQE_AEID_LLP_CONNECTION_RESET))) {
+		if (issue_close) {
 			atomic_inc(&cm_closes);
-			nesqp->cm_id = NULL;
-			nesqp->in_disconnect = 0;
-			spin_unlock_irqrestore(&nesqp->lock, flags);
 			nes_disconnect(nesqp, 1);
 
 			cm_id->provider_data = nesqp;
@@ -2588,28 +2609,7 @@ static int nes_cm_disconn_true(struct nes_qp *nesqp)
 			}
 
 			cm_id->rem_ref(cm_id);
-
-			spin_lock_irqsave(&nesqp->lock, flags);
-			if (nesqp->flush_issued == 0) {
-				nesqp->flush_issued = 1;
-				spin_unlock_irqrestore(&nesqp->lock, flags);
-				flush_wqes(nesvnic->nesdev, nesqp,
-					NES_CQP_FLUSH_RQ, 1);
-			} else
-				spin_unlock_irqrestore(&nesqp->lock, flags);
-		} else {
-			cm_id = nesqp->cm_id;
-			spin_unlock_irqrestore(&nesqp->lock, flags);
-			/* check to see if the inbound reset beat the outbound reset */
-			if ((!cm_id) && (last_ae==NES_AEQE_AEID_RESET_SENT)) {
-				nes_debug(NES_DBG_CM, "QP%u: Decing refcount "
-					"due to inbound reset beating the "
-					"outbound reset.\n", nesqp->hwqp.qp_id);
-			}
 		}
-	} else {
-		nesqp->disconn_pending = 0;
-		spin_unlock_irqrestore(&nesqp->lock, flags);
 	}
 
 	return 0;

drivers/infiniband/hw/nes/nes_cm.h

@@ -410,8 +410,6 @@ struct nes_cm_ops {
 int schedule_nes_timer(struct nes_cm_node *, struct sk_buff *,
 		enum nes_timer_type, int, int);
 
-int nes_cm_disconn(struct nes_qp *);
-
 int nes_accept(struct iw_cm_id *, struct iw_cm_conn_param *);
 int nes_reject(struct iw_cm_id *, const void *, u8);
 int nes_connect(struct iw_cm_id *, struct iw_cm_conn_param *);

drivers/infiniband/hw/nes/nes_hw.h

@@ -241,6 +241,7 @@ enum nes_cqp_stag_wqeword_idx {
 };
 
 #define NES_CQP_OP_IWARP_STATE_SHIFT 28
+#define NES_CQP_OP_TERMLEN_SHIFT     28
 
 enum nes_cqp_qp_bits {
 	NES_CQP_QP_ARP_VALID = (1<<8),
@@ -265,12 +266,16 @@ enum nes_cqp_qp_bits {
 	NES_CQP_QP_IWARP_STATE_TERMINATE = (5<<NES_CQP_OP_IWARP_STATE_SHIFT),
 	NES_CQP_QP_IWARP_STATE_ERROR = (6<<NES_CQP_OP_IWARP_STATE_SHIFT),
 	NES_CQP_QP_IWARP_STATE_MASK = (7<<NES_CQP_OP_IWARP_STATE_SHIFT),
+	NES_CQP_QP_TERM_DONT_SEND_FIN = (1<<24),
+	NES_CQP_QP_TERM_DONT_SEND_TERM_MSG = (1<<25),
 	NES_CQP_QP_RESET = (1<<31),
 };
 
 enum nes_cqp_qp_wqe_word_idx {
 	NES_CQP_QP_WQE_CONTEXT_LOW_IDX = 6,
 	NES_CQP_QP_WQE_CONTEXT_HIGH_IDX = 7,
+	NES_CQP_QP_WQE_FLUSH_SQ_CODE = 8,
+	NES_CQP_QP_WQE_FLUSH_RQ_CODE = 9,
 	NES_CQP_QP_WQE_NEW_MSS_IDX = 15,
 };
 
@@ -361,6 +366,7 @@ enum nes_cqp_arp_bits {
 enum nes_cqp_flush_bits {
 	NES_CQP_FLUSH_SQ = (1<<30),
 	NES_CQP_FLUSH_RQ = (1<<31),
+	NES_CQP_FLUSH_MAJ_MIN = (1<<28),
 };
 
 enum nes_cqe_opcode_bits {
@@ -633,11 +639,14 @@ enum nes_aeqe_bits {
 	NES_AEQE_INBOUND_RDMA = (1<<19),
 	NES_AEQE_IWARP_STATE_MASK = (7<<20),
 	NES_AEQE_TCP_STATE_MASK = (0xf<<24),
+	NES_AEQE_Q2_DATA_WRITTEN = (0x3<<28),
 	NES_AEQE_VALID = (1<<31),
 };
 
 #define NES_AEQE_IWARP_STATE_SHIFT	20
 #define NES_AEQE_TCP_STATE_SHIFT	24
+#define NES_AEQE_Q2_DATA_ETHERNET       (1<<28)
+#define NES_AEQE_Q2_DATA_MPA            (1<<29)
 
 enum nes_aeqe_iwarp_state {
 	NES_AEQE_IWARP_STATE_NON_EXISTANT = 0,
@@ -751,6 +760,15 @@ enum nes_iwarp_sq_wqe_bits {
 	NES_IWARP_SQ_OP_NOP = 12,
 };
 
+enum nes_iwarp_cqe_major_code {
+	NES_IWARP_CQE_MAJOR_FLUSH = 1,
+	NES_IWARP_CQE_MAJOR_DRV = 0x8000
+};
+
+enum nes_iwarp_cqe_minor_code {
+	NES_IWARP_CQE_MINOR_FLUSH = 1
+};
+
 #define NES_EEPROM_READ_REQUEST (1<<16)
 #define NES_MAC_ADDR_VALID      (1<<20)
 
@@ -1119,6 +1137,7 @@ struct nes_adapter {
 	u8            netdev_max;	/* from host nic address count in EEPROM */
 	u8            port_count;
 	u8            virtwq;
+	u8            send_term_ok;
 	u8            et_use_adaptive_rx_coalesce;
 	u8            adapter_fcn_count;
 	u8 pft_mcast_map[NES_PFT_SIZE];
@@ -1217,6 +1236,90 @@ struct nes_ib_device {
 	u32 num_pd;
 };
 
+enum nes_hdrct_flags {
+	DDP_LEN_FLAG                    = 0x80,
+	DDP_HDR_FLAG                    = 0x40,
+	RDMA_HDR_FLAG                   = 0x20
+};
+
+enum nes_term_layers {
+	LAYER_RDMA			= 0,
+	LAYER_DDP			= 1,
+	LAYER_MPA			= 2
+};
+
+enum nes_term_error_types {
+	RDMAP_CATASTROPHIC		= 0,
+	RDMAP_REMOTE_PROT		= 1,
+	RDMAP_REMOTE_OP			= 2,
+	DDP_CATASTROPHIC		= 0,
+	DDP_TAGGED_BUFFER		= 1,
+	DDP_UNTAGGED_BUFFER		= 2,
+	DDP_LLP				= 3
+};
+
+enum nes_term_rdma_errors {
+	RDMAP_INV_STAG			= 0x00,
+	RDMAP_INV_BOUNDS		= 0x01,
+	RDMAP_ACCESS			= 0x02,
+	RDMAP_UNASSOC_STAG		= 0x03,
+	RDMAP_TO_WRAP			= 0x04,
+	RDMAP_INV_RDMAP_VER		= 0x05,
+	RDMAP_UNEXPECTED_OP		= 0x06,
+	RDMAP_CATASTROPHIC_LOCAL	= 0x07,
+	RDMAP_CATASTROPHIC_GLOBAL	= 0x08,
+	RDMAP_CANT_INV_STAG		= 0x09,
+	RDMAP_UNSPECIFIED		= 0xff
+};
+
+enum nes_term_ddp_errors {
+	DDP_CATASTROPHIC_LOCAL		= 0x00,
+	DDP_TAGGED_INV_STAG		= 0x00,
+	DDP_TAGGED_BOUNDS		= 0x01,
+	DDP_TAGGED_UNASSOC_STAG		= 0x02,
+	DDP_TAGGED_TO_WRAP		= 0x03,
+	DDP_TAGGED_INV_DDP_VER		= 0x04,
+	DDP_UNTAGGED_INV_QN		= 0x01,
+	DDP_UNTAGGED_INV_MSN_NO_BUF	= 0x02,
+	DDP_UNTAGGED_INV_MSN_RANGE	= 0x03,
+	DDP_UNTAGGED_INV_MO		= 0x04,
+	DDP_UNTAGGED_INV_TOO_LONG	= 0x05,
+	DDP_UNTAGGED_INV_DDP_VER	= 0x06
+};
+
+enum nes_term_mpa_errors {
+	MPA_CLOSED			= 0x01,
+	MPA_CRC				= 0x02,
+	MPA_MARKER			= 0x03,
+	MPA_REQ_RSP			= 0x04,
+};
+
+struct nes_terminate_hdr {
+	u8 layer_etype;
+	u8 error_code;
+	u8 hdrct;
+	u8 rsvd;
+};
+
+/* Used to determine how to fill in terminate error codes */
+#define IWARP_OPCODE_WRITE		0
+#define IWARP_OPCODE_READREQ		1
+#define IWARP_OPCODE_READRSP		2
+#define IWARP_OPCODE_SEND		3
+#define IWARP_OPCODE_SEND_INV		4
+#define IWARP_OPCODE_SEND_SE		5
+#define IWARP_OPCODE_SEND_SE_INV	6
+#define IWARP_OPCODE_TERM		7
+
+/* These values are used only during terminate processing */
+#define TERM_DDP_LEN_TAGGED	14
+#define TERM_DDP_LEN_UNTAGGED	18
+#define TERM_RDMA_LEN		28
+#define RDMA_OPCODE_MASK	0x0f
+#define RDMA_READ_REQ_OPCODE	1
+#define BAD_FRAME_OFFSET	64
+#define CQE_MAJOR_DRV		0x8000
+
 #define nes_vlan_rx vlan_hwaccel_receive_skb
 #define nes_netif_rx netif_receive_skb
 

drivers/infiniband/hw/nes/nes_utils.c

@@ -183,6 +183,9 @@ int nes_read_eeprom_values(struct nes_device *nesdev, struct nes_adapter *nesada
 		} else if (((major_ver == 2) && (minor_ver > 21)) || ((major_ver > 2) && (major_ver != 255))) {
 			nesadapter->virtwq = 1;
 		}
+		if (((major_ver == 3) && (minor_ver >= 16)) || (major_ver > 3))
+			nesadapter->send_term_ok = 1;
+
 		nesadapter->firmware_version = (((u32)(u8)(eeprom_data>>8))  <<  16) +
 				(u32)((u8)eeprom_data);
 
@@ -548,7 +551,7 @@ struct nes_cqp_request *nes_get_cqp_request(struct nes_device *nesdev)
 		spin_unlock_irqrestore(&nesdev->cqp.lock, flags);
 	}
 	if (cqp_request == NULL) {
-		cqp_request = kzalloc(sizeof(struct nes_cqp_request), GFP_KERNEL);
+		cqp_request = kzalloc(sizeof(struct nes_cqp_request), GFP_ATOMIC);
 		if (cqp_request) {
 			cqp_request->dynamic = 1;
 			INIT_LIST_HEAD(&cqp_request->list);

drivers/infiniband/hw/nes/nes_verbs.c

@@ -667,15 +667,32 @@ static int nes_query_device(struct ib_device *ibdev, struct ib_device_attr *prop
  */
 static int nes_query_port(struct ib_device *ibdev, u8 port, struct ib_port_attr *props)
 {
+	struct nes_vnic *nesvnic = to_nesvnic(ibdev);
+	struct net_device *netdev = nesvnic->netdev;
+
 	memset(props, 0, sizeof(*props));
 
-	props->max_mtu = IB_MTU_2048;
-	props->active_mtu = IB_MTU_2048;
+	props->max_mtu = IB_MTU_4096;
+
+	if (netdev->mtu  >= 4096)
+		props->active_mtu = IB_MTU_4096;
+	else if (netdev->mtu  >= 2048)
+		props->active_mtu = IB_MTU_2048;
+	else if (netdev->mtu  >= 1024)
+		props->active_mtu = IB_MTU_1024;
+	else if (netdev->mtu  >= 512)
+		props->active_mtu = IB_MTU_512;
+	else
+		props->active_mtu = IB_MTU_256;
+
 	props->lid = 1;
 	props->lmc = 0;
 	props->sm_lid = 0;
 	props->sm_sl = 0;
-	props->state = IB_PORT_ACTIVE;
+	if (nesvnic->linkup)
+		props->state = IB_PORT_ACTIVE;
+	else
+		props->state = IB_PORT_DOWN;
 	props->phys_state = 0;
 	props->port_cap_flags = IB_PORT_CM_SUP | IB_PORT_REINIT_SUP |
 			IB_PORT_VENDOR_CLASS_SUP | IB_PORT_BOOT_MGMT_SUP;
@@ -1505,13 +1522,46 @@ static struct ib_qp *nes_create_qp(struct ib_pd *ibpd,
 }
 
 
+/**
+ * nes_clean_cq
+ */
+static void nes_clean_cq(struct nes_qp *nesqp, struct nes_cq *nescq)
+{
+	u32 cq_head;
+	u32 lo;
+	u32 hi;
+	u64 u64temp;
+	unsigned long flags = 0;
+
+	spin_lock_irqsave(&nescq->lock, flags);
+
+	cq_head = nescq->hw_cq.cq_head;
+	while (le32_to_cpu(nescq->hw_cq.cq_vbase[cq_head].cqe_words[NES_CQE_OPCODE_IDX]) & NES_CQE_VALID) {
+		rmb();
+		lo = le32_to_cpu(nescq->hw_cq.cq_vbase[cq_head].cqe_words[NES_CQE_COMP_COMP_CTX_LOW_IDX]);
+		hi = le32_to_cpu(nescq->hw_cq.cq_vbase[cq_head].cqe_words[NES_CQE_COMP_COMP_CTX_HIGH_IDX]);
+		u64temp = (((u64)hi) << 32) | ((u64)lo);
+		u64temp &= ~(NES_SW_CONTEXT_ALIGN-1);
+		if (u64temp == (u64)(unsigned long)nesqp) {
+			/* Zero the context value so cqe will be ignored */
+			nescq->hw_cq.cq_vbase[cq_head].cqe_words[NES_CQE_COMP_COMP_CTX_LOW_IDX] = 0;
+			nescq->hw_cq.cq_vbase[cq_head].cqe_words[NES_CQE_COMP_COMP_CTX_HIGH_IDX] = 0;
+		}
+
+		if (++cq_head >= nescq->hw_cq.cq_size)
+			cq_head = 0;
+	}
+
+	spin_unlock_irqrestore(&nescq->lock, flags);
+}
+
+
 /**
  * nes_destroy_qp
  */
 static int nes_destroy_qp(struct ib_qp *ibqp)
 {
 	struct nes_qp *nesqp = to_nesqp(ibqp);
-	/* struct nes_vnic *nesvnic = to_nesvnic(ibqp->device); */
 	struct nes_ucontext *nes_ucontext;
 	struct ib_qp_attr attr;
 	struct iw_cm_id *cm_id;
@@ -1548,7 +1598,6 @@ static int nes_destroy_qp(struct ib_qp *ibqp)
 			nes_debug(NES_DBG_QP, "OFA CM event_handler returned, ret=%d\n", ret);
 	}
 
-
 	if (nesqp->user_mode) {
 		if ((ibqp->uobject)&&(ibqp->uobject->context)) {
 			nes_ucontext = to_nesucontext(ibqp->uobject->context);
@@ -1560,6 +1609,13 @@ static int nes_destroy_qp(struct ib_qp *ibqp)
 		}
 		if (nesqp->pbl_pbase)
 			kunmap(nesqp->page);
+	} else {
+		/* Clean any pending completions from the cq(s) */
+		if (nesqp->nesscq)
+			nes_clean_cq(nesqp, nesqp->nesscq);
+
+		if ((nesqp->nesrcq) && (nesqp->nesrcq != nesqp->nesscq))
+			nes_clean_cq(nesqp, nesqp->nesrcq);
 	}
 
 	nes_rem_ref(&nesqp->ibqp);
@@ -2884,7 +2940,7 @@ static int nes_query_qp(struct ib_qp *ibqp, struct ib_qp_attr *attr,
  * nes_hw_modify_qp
  */
 int nes_hw_modify_qp(struct nes_device *nesdev, struct nes_qp *nesqp,
-		u32 next_iwarp_state, u32 wait_completion)
+		u32 next_iwarp_state, u32 termlen, u32 wait_completion)
 {
 	struct nes_hw_cqp_wqe *cqp_wqe;
 	/* struct iw_cm_id *cm_id = nesqp->cm_id; */
@@ -2916,6 +2972,13 @@ int nes_hw_modify_qp(struct nes_device *nesdev, struct nes_qp *nesqp,
 	set_wqe_32bit_value(cqp_wqe->wqe_words, NES_CQP_WQE_ID_IDX, nesqp->hwqp.qp_id);
 	set_wqe_64bit_value(cqp_wqe->wqe_words, NES_CQP_QP_WQE_CONTEXT_LOW_IDX, (u64)nesqp->nesqp_context_pbase);
 
+	/* If sending a terminate message, fill in the length (in words) */
+	if (((next_iwarp_state & NES_CQP_QP_IWARP_STATE_MASK) == NES_CQP_QP_IWARP_STATE_TERMINATE) &&
+	    !(next_iwarp_state & NES_CQP_QP_TERM_DONT_SEND_TERM_MSG)) {
+		termlen = ((termlen + 3) >> 2) << NES_CQP_OP_TERMLEN_SHIFT;
+		set_wqe_32bit_value(cqp_wqe->wqe_words, NES_CQP_QP_WQE_NEW_MSS_IDX, termlen);
+	}
+
 	atomic_set(&cqp_request->refcount, 2);
 	nes_post_cqp_request(nesdev, cqp_request);
 
@@ -3086,6 +3149,9 @@ int nes_modify_qp(struct ib_qp *ibqp, struct ib_qp_attr *attr,
 				}
 				nes_debug(NES_DBG_MOD_QP, "QP%u: new state = error\n",
 						nesqp->hwqp.qp_id);
+				if (nesqp->term_flags)
+					del_timer(&nesqp->terminate_timer);
+
 				next_iwarp_state = NES_CQP_QP_IWARP_STATE_ERROR;
 				/* next_iwarp_state = (NES_CQP_QP_IWARP_STATE_TERMINATE | 0x02000000); */
 					if (nesqp->hte_added) {
@@ -3163,7 +3229,7 @@ int nes_modify_qp(struct ib_qp *ibqp, struct ib_qp_attr *attr,
 
 	if (issue_modify_qp) {
 		nes_debug(NES_DBG_MOD_QP, "call nes_hw_modify_qp\n");
-		ret = nes_hw_modify_qp(nesdev, nesqp, next_iwarp_state, 1);
+		ret = nes_hw_modify_qp(nesdev, nesqp, next_iwarp_state, 0, 1);
 		if (ret)
 			nes_debug(NES_DBG_MOD_QP, "nes_hw_modify_qp (next_iwarp_state = 0x%08X)"
 					" failed for QP%u.\n",
@@ -3328,6 +3394,12 @@ static int nes_post_send(struct ib_qp *ibqp, struct ib_send_wr *ib_wr,
 	head = nesqp->hwqp.sq_head;
 
 	while (ib_wr) {
+		/* Check for QP error */
+		if (nesqp->term_flags) {
+			err = -EINVAL;
+			break;
+		}
+
 		/* Check for SQ overflow */
 		if (((head + (2 * qsize) - nesqp->hwqp.sq_tail) % qsize) == (qsize - 1)) {
 			err = -EINVAL;
@@ -3484,6 +3556,12 @@ static int nes_post_recv(struct ib_qp *ibqp, struct ib_recv_wr *ib_wr,
 	head = nesqp->hwqp.rq_head;
 
 	while (ib_wr) {
+		/* Check for QP error */
+		if (nesqp->term_flags) {
+			err = -EINVAL;
+			break;
+		}
+
 		if (ib_wr->num_sge > nesdev->nesadapter->max_sge) {
 			err = -EINVAL;
 			break;
@@ -3547,7 +3625,6 @@ static int nes_poll_cq(struct ib_cq *ibcq, int num_entries, struct ib_wc *entry)
 {
 	u64 u64temp;
 	u64 wrid;
-	/* u64 u64temp; */
 	unsigned long flags = 0;
 	struct nes_vnic *nesvnic = to_nesvnic(ibcq->device);
 	struct nes_device *nesdev = nesvnic->nesdev;
@@ -3555,12 +3632,13 @@ static int nes_poll_cq(struct ib_cq *ibcq, int num_entries, struct ib_wc *entry)
 	struct nes_qp *nesqp;
 	struct nes_hw_cqe cqe;
 	u32 head;
-	u32 wq_tail;
+	u32 wq_tail = 0;
 	u32 cq_size;
 	u32 cqe_count = 0;
 	u32 wqe_index;
 	u32 u32temp;
-	/* u32 counter; */
+	u32 move_cq_head = 1;
+	u32 err_code;
 
 	nes_debug(NES_DBG_CQ, "\n");
 
@@ -3570,29 +3648,40 @@ static int nes_poll_cq(struct ib_cq *ibcq, int num_entries, struct ib_wc *entry)
 	cq_size = nescq->hw_cq.cq_size;
 
 	while (cqe_count < num_entries) {
-		if (le32_to_cpu(nescq->hw_cq.cq_vbase[head].cqe_words[NES_CQE_OPCODE_IDX]) &
-				NES_CQE_VALID) {
-			/*
-			 * Make sure we read CQ entry contents *after*
-			 * we've checked the valid bit.
-			 */
-			rmb();
-
-			cqe = nescq->hw_cq.cq_vbase[head];
-			nescq->hw_cq.cq_vbase[head].cqe_words[NES_CQE_OPCODE_IDX] = 0;
-			u32temp = le32_to_cpu(cqe.cqe_words[NES_CQE_COMP_COMP_CTX_LOW_IDX]);
-			wqe_index = u32temp &
-					(nesdev->nesadapter->max_qp_wr - 1);
-			u32temp &= ~(NES_SW_CONTEXT_ALIGN-1);
-			/* parse CQE, get completion context from WQE (either rq or sq */
-			u64temp = (((u64)(le32_to_cpu(cqe.cqe_words[NES_CQE_COMP_COMP_CTX_HIGH_IDX])))<<32) |
-					((u64)u32temp);
-			nesqp = *((struct nes_qp **)&u64temp);
+		if ((le32_to_cpu(nescq->hw_cq.cq_vbase[head].cqe_words[NES_CQE_OPCODE_IDX]) &
+				NES_CQE_VALID) == 0)
+			break;
+
+		/*
+		 * Make sure we read CQ entry contents *after*
+		 * we've checked the valid bit.
+		 */
+		rmb();
+
+		cqe = nescq->hw_cq.cq_vbase[head];
+		u32temp = le32_to_cpu(cqe.cqe_words[NES_CQE_COMP_COMP_CTX_LOW_IDX]);
+		wqe_index = u32temp & (nesdev->nesadapter->max_qp_wr - 1);
+		u32temp &= ~(NES_SW_CONTEXT_ALIGN-1);
+		/* parse CQE, get completion context from WQE (either rq or sq) */
+		u64temp = (((u64)(le32_to_cpu(cqe.cqe_words[NES_CQE_COMP_COMP_CTX_HIGH_IDX])))<<32) |
+				((u64)u32temp);
+
+		if (u64temp) {
+			nesqp = (struct nes_qp *)(unsigned long)u64temp;
 			memset(entry, 0, sizeof *entry);
 			if (cqe.cqe_words[NES_CQE_ERROR_CODE_IDX] == 0) {
 				entry->status = IB_WC_SUCCESS;
 			} else {
-				entry->status = IB_WC_WR_FLUSH_ERR;
+				err_code = le32_to_cpu(cqe.cqe_words[NES_CQE_ERROR_CODE_IDX]);
+				if (NES_IWARP_CQE_MAJOR_DRV == (err_code >> 16)) {
+					entry->status = err_code & 0x0000ffff;
+
+					/* The rest of the cqe's will be marked as flushed */
+					nescq->hw_cq.cq_vbase[head].cqe_words[NES_CQE_ERROR_CODE_IDX] =
+						cpu_to_le32((NES_IWARP_CQE_MAJOR_FLUSH << 16) |
+							    NES_IWARP_CQE_MINOR_FLUSH);
+				} else
+					entry->status = IB_WC_WR_FLUSH_ERR;
 			}
 
 			entry->qp = &nesqp->ibqp;
@@ -3601,20 +3690,18 @@ static int nes_poll_cq(struct ib_cq *ibcq, int num_entries, struct ib_wc *entry)
 			if (le32_to_cpu(cqe.cqe_words[NES_CQE_OPCODE_IDX]) & NES_CQE_SQ) {
 				if (nesqp->skip_lsmm) {
 					nesqp->skip_lsmm = 0;
-					wq_tail = nesqp->hwqp.sq_tail++;
+					nesqp->hwqp.sq_tail++;
 				}
 
 				/* Working on a SQ Completion*/
-				wq_tail = wqe_index;
-				nesqp->hwqp.sq_tail = (wqe_index+1)&(nesqp->hwqp.sq_size - 1);
-				wrid = (((u64)(cpu_to_le32((u32)nesqp->hwqp.sq_vbase[wq_tail].
+				wrid = (((u64)(cpu_to_le32((u32)nesqp->hwqp.sq_vbase[wqe_index].
 						wqe_words[NES_IWARP_SQ_WQE_COMP_SCRATCH_HIGH_IDX]))) << 32) |
-						((u64)(cpu_to_le32((u32)nesqp->hwqp.sq_vbase[wq_tail].
+						((u64)(cpu_to_le32((u32)nesqp->hwqp.sq_vbase[wqe_index].
 						wqe_words[NES_IWARP_SQ_WQE_COMP_SCRATCH_LOW_IDX])));
-				entry->byte_len = le32_to_cpu(nesqp->hwqp.sq_vbase[wq_tail].
+				entry->byte_len = le32_to_cpu(nesqp->hwqp.sq_vbase[wqe_index].
 						wqe_words[NES_IWARP_SQ_WQE_TOTAL_PAYLOAD_IDX]);
 
-				switch (le32_to_cpu(nesqp->hwqp.sq_vbase[wq_tail].
+				switch (le32_to_cpu(nesqp->hwqp.sq_vbase[wqe_index].
 						wqe_words[NES_IWARP_SQ_WQE_MISC_IDX]) & 0x3f) {
 					case NES_IWARP_SQ_OP_RDMAW:
 						nes_debug(NES_DBG_CQ, "Operation = RDMA WRITE.\n");
@@ -3623,7 +3710,7 @@ static int nes_poll_cq(struct ib_cq *ibcq, int num_entries, struct ib_wc *entry)
 					case NES_IWARP_SQ_OP_RDMAR:
 						nes_debug(NES_DBG_CQ, "Operation = RDMA READ.\n");
 						entry->opcode = IB_WC_RDMA_READ;
-						entry->byte_len = le32_to_cpu(nesqp->hwqp.sq_vbase[wq_tail].
+						entry->byte_len = le32_to_cpu(nesqp->hwqp.sq_vbase[wqe_index].
 								wqe_words[NES_IWARP_SQ_WQE_RDMA_LENGTH_IDX]);
 						break;
 					case NES_IWARP_SQ_OP_SENDINV:
@@ -3634,33 +3721,54 @@ static int nes_poll_cq(struct ib_cq *ibcq, int num_entries, struct ib_wc *entry)
 						entry->opcode = IB_WC_SEND;
 						break;
 				}
+
+				nesqp->hwqp.sq_tail = (wqe_index+1)&(nesqp->hwqp.sq_size - 1);
+				if ((entry->status != IB_WC_SUCCESS) && (nesqp->hwqp.sq_tail != nesqp->hwqp.sq_head)) {
+					move_cq_head = 0;
+					wq_tail = nesqp->hwqp.sq_tail;
+				}
 			} else {
 				/* Working on a RQ Completion*/
-				wq_tail = wqe_index;
-					nesqp->hwqp.rq_tail = (wqe_index+1)&(nesqp->hwqp.rq_size - 1);
 				entry->byte_len = le32_to_cpu(cqe.cqe_words[NES_CQE_PAYLOAD_LENGTH_IDX]);
-				wrid = ((u64)(le32_to_cpu(nesqp->hwqp.rq_vbase[wq_tail].wqe_words[NES_IWARP_RQ_WQE_COMP_SCRATCH_LOW_IDX]))) |
-					((u64)(le32_to_cpu(nesqp->hwqp.rq_vbase[wq_tail].wqe_words[NES_IWARP_RQ_WQE_COMP_SCRATCH_HIGH_IDX]))<<32);
+				wrid = ((u64)(le32_to_cpu(nesqp->hwqp.rq_vbase[wqe_index].wqe_words[NES_IWARP_RQ_WQE_COMP_SCRATCH_LOW_IDX]))) |
+					((u64)(le32_to_cpu(nesqp->hwqp.rq_vbase[wqe_index].wqe_words[NES_IWARP_RQ_WQE_COMP_SCRATCH_HIGH_IDX]))<<32);
 					entry->opcode = IB_WC_RECV;
+
+				nesqp->hwqp.rq_tail = (wqe_index+1)&(nesqp->hwqp.rq_size - 1);
+				if ((entry->status != IB_WC_SUCCESS) && (nesqp->hwqp.rq_tail != nesqp->hwqp.rq_head)) {
+					move_cq_head = 0;
+					wq_tail = nesqp->hwqp.rq_tail;
+				}
 			}
+
 			entry->wr_id = wrid;
+			entry++;
+			cqe_count++;
+		}
 
+		if (move_cq_head) {
+			nescq->hw_cq.cq_vbase[head].cqe_words[NES_CQE_OPCODE_IDX] = 0;
 			if (++head >= cq_size)
 				head = 0;
-			cqe_count++;
 			nescq->polled_completions++;
+
 			if ((nescq->polled_completions > (cq_size / 2)) ||
 					(nescq->polled_completions == 255)) {
 				nes_debug(NES_DBG_CQ, "CQ%u Issuing CQE Allocate since more than half of cqes"
-						" are pending %u of %u.\n",
-						nescq->hw_cq.cq_number, nescq->polled_completions, cq_size);
+					" are pending %u of %u.\n",
+					nescq->hw_cq.cq_number, nescq->polled_completions, cq_size);
 				nes_write32(nesdev->regs+NES_CQE_ALLOC,
-						nescq->hw_cq.cq_number | (nescq->polled_completions << 16));
+					nescq->hw_cq.cq_number | (nescq->polled_completions << 16));
 				nescq->polled_completions = 0;
 			}
-			entry++;
-		} else
-			break;
+		} else {
+			/* Update the wqe index and set status to flush */
+			wqe_index = le32_to_cpu(cqe.cqe_words[NES_CQE_COMP_COMP_CTX_LOW_IDX]);
+			wqe_index = (wqe_index & (~(nesdev->nesadapter->max_qp_wr - 1))) | wq_tail;
+			nescq->hw_cq.cq_vbase[head].cqe_words[NES_CQE_COMP_COMP_CTX_LOW_IDX] =
+				cpu_to_le32(wqe_index);
+			move_cq_head = 1; /* ready for next pass */
+		}
 	}
 
 	if (nescq->polled_completions) {

drivers/infiniband/hw/nes/nes_verbs.h

@@ -40,6 +40,10 @@ struct nes_device;
 #define NES_MAX_USER_DB_REGIONS  4096
 #define NES_MAX_USER_WQ_REGIONS  4096
 
+#define NES_TERM_SENT            0x01
+#define NES_TERM_RCVD            0x02
+#define NES_TERM_DONE            0x04
+
 struct nes_ucontext {
 	struct ib_ucontext ibucontext;
 	struct nes_device  *nesdev;
@@ -119,6 +123,11 @@ struct nes_wq {
 	spinlock_t lock;
 };
 
+struct disconn_work {
+	struct work_struct    work;
+	struct nes_qp         *nesqp;
+};
+
 struct iw_cm_id;
 struct ietf_mpa_frame;
 
@@ -127,7 +136,6 @@ struct nes_qp {
 	void                  *allocated_buffer;
 	struct iw_cm_id       *cm_id;
 	struct workqueue_struct *wq;
-	struct work_struct    disconn_work;
 	struct nes_cq         *nesscq;
 	struct nes_cq         *nesrcq;
 	struct nes_pd         *nespd;
@@ -155,9 +163,13 @@ struct nes_qp {
 	void	              *pbl_vbase;
 	dma_addr_t            pbl_pbase;
 	struct page           *page;
+	struct timer_list     terminate_timer;
+	enum ib_event_type    terminate_eventtype;
 	wait_queue_head_t     kick_waitq;
 	u16                   in_disconnect;
 	u16                   private_data_len;
+	u16                   term_sq_flush_code;
+	u16                   term_rq_flush_code;
 	u8                    active_conn;
 	u8                    skip_lsmm;
 	u8                    user_mode;
@@ -165,7 +177,7 @@ struct nes_qp {
 	u8                    hw_iwarp_state;
 	u8                    flush_issued;
 	u8                    hw_tcp_state;
-	u8                    disconn_pending;
+	u8                    term_flags;
 	u8                    destroyed;
 };
 #endif			/* NES_VERBS_H */

drivers/infiniband/ulp/ipoib/ipoib_cm.c

@@ -31,7 +31,6 @@
  */
 
 #include <rdma/ib_cm.h>
-#include <rdma/ib_cache.h>
 #include <net/dst.h>
 #include <net/icmp.h>
 #include <linux/icmpv6.h>

drivers/infiniband/ulp/ipoib/ipoib_ib.c

@@ -36,7 +36,6 @@
 #include <linux/delay.h>
 #include <linux/dma-mapping.h>
 
-#include <rdma/ib_cache.h>
 #include <linux/ip.h>
 #include <linux/tcp.h>
 

drivers/infiniband/ulp/ipoib/ipoib_main.c

@@ -604,8 +604,11 @@ static void neigh_add_path(struct sk_buff *skb, struct net_device *dev)
 					   skb_queue_len(&neigh->queue));
 				goto err_drop;
 			}
-		} else
+		} else {
+			spin_unlock_irqrestore(&priv->lock, flags);
 			ipoib_send(dev, skb, path->ah, IPOIB_QPN(skb_dst(skb)->neighbour->ha));
+			return;
+		}
 	} else {
 		neigh->ah  = NULL;
 
@@ -688,7 +691,9 @@ static void unicast_arp_send(struct sk_buff *skb, struct net_device *dev,
 		ipoib_dbg(priv, "Send unicast ARP to %04x\n",
 			  be16_to_cpu(path->pathrec.dlid));
 
+		spin_unlock_irqrestore(&priv->lock, flags);
 		ipoib_send(dev, skb, path->ah, IPOIB_QPN(phdr->hwaddr));
+		return;
 	} else if ((path->query || !path_rec_start(dev, path)) &&
 		   skb_queue_len(&path->queue) < IPOIB_MAX_PATH_REC_QUEUE) {
 		/* put pseudoheader back on for next time */

drivers/infiniband/ulp/ipoib/ipoib_multicast.c

@@ -720,7 +720,9 @@ void ipoib_mcast_send(struct net_device *dev, void *mgid, struct sk_buff *skb)
 			}
 		}
 
+		spin_unlock_irqrestore(&priv->lock, flags);
 		ipoib_send(dev, skb, mcast->ah, IB_MULTICAST_QPN);
+		return;
 	}
 
 unlock:
@@ -758,6 +760,20 @@ void ipoib_mcast_dev_flush(struct net_device *dev)
 	}
 }
 
+static int ipoib_mcast_addr_is_valid(const u8 *addr, unsigned int addrlen,
+				     const u8 *broadcast)
+{
+	if (addrlen != INFINIBAND_ALEN)
+		return 0;
+	/* reserved QPN, prefix, scope */
+	if (memcmp(addr, broadcast, 6))
+		return 0;
+	/* signature lower, pkey */
+	if (memcmp(addr + 7, broadcast + 7, 3))
+		return 0;
+	return 1;
+}
+
 void ipoib_mcast_restart_task(struct work_struct *work)
 {
 	struct ipoib_dev_priv *priv =
@@ -791,6 +807,11 @@ void ipoib_mcast_restart_task(struct work_struct *work)
 	for (mclist = dev->mc_list; mclist; mclist = mclist->next) {
 		union ib_gid mgid;
 
+		if (!ipoib_mcast_addr_is_valid(mclist->dmi_addr,
+					       mclist->dmi_addrlen,
+					       dev->broadcast))
+			continue;
+
 		memcpy(mgid.raw, mclist->dmi_addr + 4, sizeof mgid);
 
 		mcast = __ipoib_mcast_find(dev, &mgid);

drivers/net/mlx4/cq.c

@@ -34,7 +34,6 @@
  * SOFTWARE.
  */
 
-#include <linux/init.h>
 #include <linux/hardirq.h>
 
 #include <linux/mlx4/cmd.h>

drivers/net/mlx4/eq.c

@@ -31,7 +31,6 @@
  * SOFTWARE.
  */
 
-#include <linux/init.h>
 #include <linux/interrupt.h>
 #include <linux/mm.h>
 #include <linux/dma-mapping.h>
@@ -41,6 +40,10 @@
 #include "mlx4.h"
 #include "fw.h"
 
+enum {
+	MLX4_IRQNAME_SIZE	= 64
+};
+
 enum {
 	MLX4_NUM_ASYNC_EQE	= 0x100,
 	MLX4_NUM_SPARE_EQE	= 0x80,
@@ -526,48 +529,6 @@ static void mlx4_unmap_clr_int(struct mlx4_dev *dev)
 	iounmap(priv->clr_base);
 }
 
-int mlx4_map_eq_icm(struct mlx4_dev *dev, u64 icm_virt)
-{
-	struct mlx4_priv *priv = mlx4_priv(dev);
-	int ret;
-
-	/*
-	 * We assume that mapping one page is enough for the whole EQ
-	 * context table.  This is fine with all current HCAs, because
-	 * we only use 32 EQs and each EQ uses 64 bytes of context
-	 * memory, or 1 KB total.
-	 */
-	priv->eq_table.icm_virt = icm_virt;
-	priv->eq_table.icm_page = alloc_page(GFP_HIGHUSER);
-	if (!priv->eq_table.icm_page)
-		return -ENOMEM;
-	priv->eq_table.icm_dma  = pci_map_page(dev->pdev, priv->eq_table.icm_page, 0,
-					       PAGE_SIZE, PCI_DMA_BIDIRECTIONAL);
-	if (pci_dma_mapping_error(dev->pdev, priv->eq_table.icm_dma)) {
-		__free_page(priv->eq_table.icm_page);
-		return -ENOMEM;
-	}
-
-	ret = mlx4_MAP_ICM_page(dev, priv->eq_table.icm_dma, icm_virt);
-	if (ret) {
-		pci_unmap_page(dev->pdev, priv->eq_table.icm_dma, PAGE_SIZE,
-			       PCI_DMA_BIDIRECTIONAL);
-		__free_page(priv->eq_table.icm_page);
-	}
-
-	return ret;
-}
-
-void mlx4_unmap_eq_icm(struct mlx4_dev *dev)
-{
-	struct mlx4_priv *priv = mlx4_priv(dev);
-
-	mlx4_UNMAP_ICM(dev, priv->eq_table.icm_virt, 1);
-	pci_unmap_page(dev->pdev, priv->eq_table.icm_dma, PAGE_SIZE,
-		       PCI_DMA_BIDIRECTIONAL);
-	__free_page(priv->eq_table.icm_page);
-}
-
 int mlx4_alloc_eq_table(struct mlx4_dev *dev)
 {
 	struct mlx4_priv *priv = mlx4_priv(dev);
@@ -615,7 +576,9 @@ int mlx4_init_eq_table(struct mlx4_dev *dev)
 	priv->eq_table.clr_int  = priv->clr_base +
 		(priv->eq_table.inta_pin < 32 ? 4 : 0);
 
-	priv->eq_table.irq_names = kmalloc(16 * dev->caps.num_comp_vectors, GFP_KERNEL);
+	priv->eq_table.irq_names =
+		kmalloc(MLX4_IRQNAME_SIZE * (dev->caps.num_comp_vectors + 1),
+			GFP_KERNEL);
 	if (!priv->eq_table.irq_names) {
 		err = -ENOMEM;
 		goto err_out_bitmap;
@@ -638,17 +601,25 @@ int mlx4_init_eq_table(struct mlx4_dev *dev)
 		goto err_out_comp;
 
 	if (dev->flags & MLX4_FLAG_MSI_X) {
-		static const char async_eq_name[] = "mlx4-async";
 		const char *eq_name;
 
 		for (i = 0; i < dev->caps.num_comp_vectors + 1; ++i) {
 			if (i < dev->caps.num_comp_vectors) {
-				snprintf(priv->eq_table.irq_names + i * 16, 16,
-					 "mlx4-comp-%d", i);
-				eq_name = priv->eq_table.irq_names + i * 16;
-			} else
-				eq_name = async_eq_name;
+				snprintf(priv->eq_table.irq_names +
+					 i * MLX4_IRQNAME_SIZE,
+					 MLX4_IRQNAME_SIZE,
+					 "mlx4-comp-%d@pci:%s", i,
+					 pci_name(dev->pdev));
+			} else {
+				snprintf(priv->eq_table.irq_names +
+					 i * MLX4_IRQNAME_SIZE,
+					 MLX4_IRQNAME_SIZE,
+					 "mlx4-async@pci:%s",
+					 pci_name(dev->pdev));
+			}
 
+			eq_name = priv->eq_table.irq_names +
+				  i * MLX4_IRQNAME_SIZE;
 			err = request_irq(priv->eq_table.eq[i].irq,
 					  mlx4_msi_x_interrupt, 0, eq_name,
 					  priv->eq_table.eq + i);
@@ -658,8 +629,12 @@ int mlx4_init_eq_table(struct mlx4_dev *dev)
 			priv->eq_table.eq[i].have_irq = 1;
 		}
 	} else {
+		snprintf(priv->eq_table.irq_names,
+			 MLX4_IRQNAME_SIZE,
+			 DRV_NAME "@pci:%s",
+			 pci_name(dev->pdev));
 		err = request_irq(dev->pdev->irq, mlx4_interrupt,
-				  IRQF_SHARED, DRV_NAME, dev);
+				  IRQF_SHARED, priv->eq_table.irq_names, dev);
 		if (err)
 			goto err_out_async;
 

drivers/net/mlx4/icm.c

@@ -31,7 +31,6 @@
  * SOFTWARE.
  */
 
-#include <linux/init.h>
 #include <linux/errno.h>
 #include <linux/mm.h>
 #include <linux/scatterlist.h>

drivers/net/mlx4/main.c

@@ -525,7 +525,10 @@ static int mlx4_init_icm(struct mlx4_dev *dev, struct mlx4_dev_cap *dev_cap,
 		goto err_unmap_aux;
 	}
 
-	err = mlx4_map_eq_icm(dev, init_hca->eqc_base);
+	err = mlx4_init_icm_table(dev, &priv->eq_table.table,
+				  init_hca->eqc_base, dev_cap->eqc_entry_sz,
+				  dev->caps.num_eqs, dev->caps.num_eqs,
+				  0, 0);
 	if (err) {
 		mlx4_err(dev, "Failed to map EQ context memory, aborting.\n");
 		goto err_unmap_cmpt;
@@ -668,7 +671,7 @@ static int mlx4_init_icm(struct mlx4_dev *dev, struct mlx4_dev_cap *dev_cap,
 	mlx4_cleanup_icm_table(dev, &priv->mr_table.mtt_table);
 
 err_unmap_eq:
-	mlx4_unmap_eq_icm(dev);
+	mlx4_cleanup_icm_table(dev, &priv->eq_table.table);
 
 err_unmap_cmpt:
 	mlx4_cleanup_icm_table(dev, &priv->eq_table.cmpt_table);
@@ -698,11 +701,11 @@ static void mlx4_free_icms(struct mlx4_dev *dev)
 	mlx4_cleanup_icm_table(dev, &priv->qp_table.qp_table);
 	mlx4_cleanup_icm_table(dev, &priv->mr_table.dmpt_table);
 	mlx4_cleanup_icm_table(dev, &priv->mr_table.mtt_table);
+	mlx4_cleanup_icm_table(dev, &priv->eq_table.table);
 	mlx4_cleanup_icm_table(dev, &priv->eq_table.cmpt_table);
 	mlx4_cleanup_icm_table(dev, &priv->cq_table.cmpt_table);
 	mlx4_cleanup_icm_table(dev, &priv->srq_table.cmpt_table);
 	mlx4_cleanup_icm_table(dev, &priv->qp_table.cmpt_table);
-	mlx4_unmap_eq_icm(dev);
 
 	mlx4_UNMAP_ICM_AUX(dev);
 	mlx4_free_icm(dev, priv->fw.aux_icm, 0);
@@ -786,7 +789,7 @@ static int mlx4_init_hca(struct mlx4_dev *dev)
 	return 0;
 
 err_close:
-	mlx4_close_hca(dev);
+	mlx4_CLOSE_HCA(dev, 0);
 
 err_free_icm:
 	mlx4_free_icms(dev);
@@ -1070,18 +1073,12 @@ static int __mlx4_init_one(struct pci_dev *pdev, const struct pci_device_id *id)
 		goto err_disable_pdev;
 	}
 
-	err = pci_request_region(pdev, 0, DRV_NAME);
+	err = pci_request_regions(pdev, DRV_NAME);
 	if (err) {
-		dev_err(&pdev->dev, "Cannot request control region, aborting.\n");
+		dev_err(&pdev->dev, "Couldn't get PCI resources, aborting\n");
 		goto err_disable_pdev;
 	}
 
-	err = pci_request_region(pdev, 2, DRV_NAME);
-	if (err) {
-		dev_err(&pdev->dev, "Cannot request UAR region, aborting.\n");
-		goto err_release_bar0;
-	}
-
 	pci_set_master(pdev);
 
 	err = pci_set_dma_mask(pdev, DMA_BIT_MASK(64));
@@ -1090,7 +1087,7 @@ static int __mlx4_init_one(struct pci_dev *pdev, const struct pci_device_id *id)
 		err = pci_set_dma_mask(pdev, DMA_BIT_MASK(32));
 		if (err) {
 			dev_err(&pdev->dev, "Can't set PCI DMA mask, aborting.\n");
-			goto err_release_bar2;
+			goto err_release_regions;
 		}
 	}
 	err = pci_set_consistent_dma_mask(pdev, DMA_BIT_MASK(64));
@@ -1101,7 +1098,7 @@ static int __mlx4_init_one(struct pci_dev *pdev, const struct pci_device_id *id)
 		if (err) {
 			dev_err(&pdev->dev, "Can't set consistent PCI DMA mask, "
 				"aborting.\n");
-			goto err_release_bar2;
+			goto err_release_regions;
 		}
 	}
 
@@ -1110,7 +1107,7 @@ static int __mlx4_init_one(struct pci_dev *pdev, const struct pci_device_id *id)
 		dev_err(&pdev->dev, "Device struct alloc failed, "
 			"aborting.\n");
 		err = -ENOMEM;
-		goto err_release_bar2;
+		goto err_release_regions;
 	}
 
 	dev       = &priv->dev;
@@ -1205,11 +1202,8 @@ static int __mlx4_init_one(struct pci_dev *pdev, const struct pci_device_id *id)
 err_free_dev:
 	kfree(priv);
 
-err_release_bar2:
-	pci_release_region(pdev, 2);
-
-err_release_bar0:
-	pci_release_region(pdev, 0);
+err_release_regions:
+	pci_release_regions(pdev);
 
 err_disable_pdev:
 	pci_disable_device(pdev);
@@ -1265,8 +1259,7 @@ static void mlx4_remove_one(struct pci_dev *pdev)
 			pci_disable_msix(pdev);
 
 		kfree(priv);
-		pci_release_region(pdev, 2);
-		pci_release_region(pdev, 0);
+		pci_release_regions(pdev);
 		pci_disable_device(pdev);
 		pci_set_drvdata(pdev, NULL);
 	}

drivers/net/mlx4/mcg.c

@@ -31,7 +31,6 @@
  * SOFTWARE.
  */
 
-#include <linux/init.h>
 #include <linux/string.h>
 #include <linux/slab.h>
 

drivers/net/mlx4/mlx4.h

@@ -205,9 +205,7 @@ struct mlx4_eq_table {
 	void __iomem	      **uar_map;
 	u32			clr_mask;
 	struct mlx4_eq	       *eq;
-	u64			icm_virt;
-	struct page	       *icm_page;
-	dma_addr_t		icm_dma;
+	struct mlx4_icm_table	table;
 	struct mlx4_icm_table	cmpt_table;
 	int			have_irq;
 	u8			inta_pin;
@@ -373,9 +371,6 @@ u64 mlx4_make_profile(struct mlx4_dev *dev,
 		      struct mlx4_dev_cap *dev_cap,
 		      struct mlx4_init_hca_param *init_hca);
 
-int mlx4_map_eq_icm(struct mlx4_dev *dev, u64 icm_virt);
-void mlx4_unmap_eq_icm(struct mlx4_dev *dev);
-
 int mlx4_cmd_init(struct mlx4_dev *dev);
 void mlx4_cmd_cleanup(struct mlx4_dev *dev);
 void mlx4_cmd_event(struct mlx4_dev *dev, u16 token, u8 status, u64 out_param);

drivers/net/mlx4/mr.c

@@ -32,7 +32,6 @@
  * SOFTWARE.
  */
 
-#include <linux/init.h>
 #include <linux/errno.h>
 
 #include <linux/mlx4/cmd.h>

drivers/net/mlx4/pd.c

@@ -31,7 +31,6 @@
  * SOFTWARE.
  */
 
-#include <linux/init.h>
 #include <linux/errno.h>
 
 #include <asm/page.h>

drivers/net/mlx4/profile.c

@@ -32,8 +32,6 @@
  * SOFTWARE.
  */
 
-#include <linux/init.h>
-
 #include "mlx4.h"
 #include "fw.h"
 

drivers/net/mlx4/qp.c

@@ -33,8 +33,6 @@
  * SOFTWARE.
  */
 
-#include <linux/init.h>
-
 #include <linux/mlx4/cmd.h>
 #include <linux/mlx4/qp.h>
 

drivers/net/mlx4/reset.c

@@ -31,7 +31,6 @@
  * SOFTWARE.
  */
 
-#include <linux/init.h>
 #include <linux/errno.h>
 #include <linux/pci.h>
 #include <linux/delay.h>

drivers/net/mlx4/srq.c

@@ -31,8 +31,6 @@
  * SOFTWARE.
  */
 
-#include <linux/init.h>
-
 #include <linux/mlx4/cmd.h>
 
 #include "mlx4.h"