Merge branches 'amso1100', 'cma', 'cxgb3', 'ehca', 'ipath', 'ipoib', 'iser',...

Merge branches 'amso1100', 'cma', 'cxgb3', 'ehca', 'ipath', 'ipoib', 'iser', 'misc', 'mlx4' and 'nes' into for-next

Documentation/infiniband/ipoib.txt

@@ -36,11 +36,11 @@ Datagram vs Connected modes
   fabric with a 2K MTU, the IPoIB MTU will be 2048 - 4 = 2044 bytes.
 
   In connected mode, the IB RC (Reliable Connected) transport is used.
-  Connected mode is to takes advantage of the connected nature of the
-  IB transport and allows an MTU up to the maximal IP packet size of
-  64K, which reduces the number of IP packets needed for handling
-  large UDP datagrams, TCP segments, etc and increases the performance
-  for large messages.
+  Connected mode takes advantage of the connected nature of the IB
+  transport and allows an MTU up to the maximal IP packet size of 64K,
+  which reduces the number of IP packets needed for handling large UDP
+  datagrams, TCP segments, etc and increases the performance for large
+  messages.
 
   In connected mode, the interface's UD QP is still used for multicast
   and communication with peers that don't support connected mode. In

drivers/infiniband/core/cma.c

@@ -330,17 +330,7 @@ static int cma_acquire_dev(struct rdma_id_private *id_priv)
 	union ib_gid gid;
 	int ret = -ENODEV;
 
-	switch (rdma_node_get_transport(dev_addr->dev_type)) {
-	case RDMA_TRANSPORT_IB:
-		ib_addr_get_sgid(dev_addr, &gid);
-		break;
-	case RDMA_TRANSPORT_IWARP:
-		iw_addr_get_sgid(dev_addr, &gid);
-		break;
-	default:
-		return -ENODEV;
-	}
-
+	rdma_addr_get_sgid(dev_addr, &gid);
 	list_for_each_entry(cma_dev, &dev_list, list) {
 		ret = ib_find_cached_gid(cma_dev->device, &gid,
 					 &id_priv->id.port_num, NULL);
@@ -1032,11 +1022,17 @@ static struct rdma_id_private *cma_new_conn_id(struct rdma_cm_id *listen_id,
 	if (rt->num_paths == 2)
 		rt->path_rec[1] = *ib_event->param.req_rcvd.alternate_path;
 
-	ib_addr_set_dgid(&rt->addr.dev_addr, &rt->path_rec[0].dgid);
-	ret = rdma_translate_ip((struct sockaddr *) &id->route.addr.src_addr,
-				&id->route.addr.dev_addr);
+	if (cma_any_addr((struct sockaddr *) &rt->addr.src_addr)) {
+		rt->addr.dev_addr.dev_type = ARPHRD_INFINIBAND;
+		rdma_addr_set_sgid(&rt->addr.dev_addr, &rt->path_rec[0].sgid);
+		ib_addr_set_pkey(&rt->addr.dev_addr, rt->path_rec[0].pkey);
+	} else {
+		ret = rdma_translate_ip((struct sockaddr *) &rt->addr.src_addr,
+					&rt->addr.dev_addr);
 		if (ret)
 			goto destroy_id;
+	}
+	rdma_addr_set_dgid(&rt->addr.dev_addr, &rt->path_rec[0].dgid);
 
 	id_priv = container_of(id, struct rdma_id_private, id);
 	id_priv->state = CMA_CONNECT;
@@ -1071,10 +1067,12 @@ static struct rdma_id_private *cma_new_udp_id(struct rdma_cm_id *listen_id,
 	cma_save_net_info(&id->route.addr, &listen_id->route.addr,
 			  ip_ver, port, src, dst);
 
+	if (!cma_any_addr((struct sockaddr *) &id->route.addr.src_addr)) {
 		ret = rdma_translate_ip((struct sockaddr *) &id->route.addr.src_addr,
 					&id->route.addr.dev_addr);
 		if (ret)
 			goto err;
+	}
 
 	id_priv = container_of(id, struct rdma_id_private, id);
 	id_priv->state = CMA_CONNECT;
@@ -1474,15 +1472,6 @@ static void cma_listen_on_all(struct rdma_id_private *id_priv)
 	mutex_unlock(&lock);
 }
 
-static int cma_bind_any(struct rdma_cm_id *id, sa_family_t af)
-{
-	struct sockaddr_storage addr_in;
-
-	memset(&addr_in, 0, sizeof addr_in);
-	addr_in.ss_family = af;
-	return rdma_bind_addr(id, (struct sockaddr *) &addr_in);
-}
-
 int rdma_listen(struct rdma_cm_id *id, int backlog)
 {
 	struct rdma_id_private *id_priv;
@@ -1490,7 +1479,8 @@ int rdma_listen(struct rdma_cm_id *id, int backlog)
 
 	id_priv = container_of(id, struct rdma_id_private, id);
 	if (id_priv->state == CMA_IDLE) {
-		ret = cma_bind_any(id, AF_INET);
+		((struct sockaddr *) &id->route.addr.src_addr)->sa_family = AF_INET;
+		ret = rdma_bind_addr(id, (struct sockaddr *) &id->route.addr.src_addr);
 		if (ret)
 			return ret;
 	}
@@ -1565,8 +1555,8 @@ static int cma_query_ib_route(struct rdma_id_private *id_priv, int timeout_ms,
 	struct sockaddr_in6 *sin6;
 
 	memset(&path_rec, 0, sizeof path_rec);
-	ib_addr_get_sgid(&addr->dev_addr, &path_rec.sgid);
-	ib_addr_get_dgid(&addr->dev_addr, &path_rec.dgid);
+	rdma_addr_get_sgid(&addr->dev_addr, &path_rec.sgid);
+	rdma_addr_get_dgid(&addr->dev_addr, &path_rec.dgid);
 	path_rec.pkey = cpu_to_be16(ib_addr_get_pkey(&addr->dev_addr));
 	path_rec.numb_path = 1;
 	path_rec.reversible = 1;
@@ -1781,7 +1771,11 @@ static int cma_bind_loopback(struct rdma_id_private *id_priv)
 	if (ret)
 		goto out;
 
-	ib_addr_set_sgid(&id_priv->id.route.addr.dev_addr, &gid);
+	id_priv->id.route.addr.dev_addr.dev_type =
+		(rdma_node_get_transport(cma_dev->device->node_type) == RDMA_TRANSPORT_IB) ?
+		ARPHRD_INFINIBAND : ARPHRD_ETHER;
+
+	rdma_addr_set_sgid(&id_priv->id.route.addr.dev_addr, &gid);
 	ib_addr_set_pkey(&id_priv->id.route.addr.dev_addr, pkey);
 	id_priv->id.port_num = p;
 	cma_attach_to_dev(id_priv, cma_dev);
@@ -1839,7 +1833,7 @@ static void addr_handler(int status, struct sockaddr *src_addr,
 static int cma_resolve_loopback(struct rdma_id_private *id_priv)
 {
 	struct cma_work *work;
-	struct sockaddr_in *src_in, *dst_in;
+	struct sockaddr *src, *dst;
 	union ib_gid gid;
 	int ret;
 
@@ -1853,14 +1847,19 @@ static int cma_resolve_loopback(struct rdma_id_private *id_priv)
 			goto err;
 	}
 
-	ib_addr_get_sgid(&id_priv->id.route.addr.dev_addr, &gid);
-	ib_addr_set_dgid(&id_priv->id.route.addr.dev_addr, &gid);
+	rdma_addr_get_sgid(&id_priv->id.route.addr.dev_addr, &gid);
+	rdma_addr_set_dgid(&id_priv->id.route.addr.dev_addr, &gid);
 
-	if (cma_zero_addr((struct sockaddr *) &id_priv->id.route.addr.src_addr)) {
-		src_in = (struct sockaddr_in *)&id_priv->id.route.addr.src_addr;
-		dst_in = (struct sockaddr_in *)&id_priv->id.route.addr.dst_addr;
-		src_in->sin_family = dst_in->sin_family;
-		src_in->sin_addr.s_addr = dst_in->sin_addr.s_addr;
+	src = (struct sockaddr *) &id_priv->id.route.addr.src_addr;
+	if (cma_zero_addr(src)) {
+		dst = (struct sockaddr *) &id_priv->id.route.addr.dst_addr;
+		if ((src->sa_family = dst->sa_family) == AF_INET) {
+			((struct sockaddr_in *) src)->sin_addr.s_addr =
+				((struct sockaddr_in *) dst)->sin_addr.s_addr;
+		} else {
+			ipv6_addr_copy(&((struct sockaddr_in6 *) src)->sin6_addr,
+				       &((struct sockaddr_in6 *) dst)->sin6_addr);
+		}
 	}
 
 	work->id = id_priv;
@@ -1878,10 +1877,14 @@ static int cma_resolve_loopback(struct rdma_id_private *id_priv)
 static int cma_bind_addr(struct rdma_cm_id *id, struct sockaddr *src_addr,
 			 struct sockaddr *dst_addr)
 {
-	if (src_addr && src_addr->sa_family)
+	if (!src_addr || !src_addr->sa_family) {
+		src_addr = (struct sockaddr *) &id->route.addr.src_addr;
+		if ((src_addr->sa_family = dst_addr->sa_family) == AF_INET6) {
+			((struct sockaddr_in6 *) src_addr)->sin6_scope_id =
+				((struct sockaddr_in6 *) dst_addr)->sin6_scope_id;
+		}
+	}
 	return rdma_bind_addr(id, src_addr);
-	else
-		return cma_bind_any(id, dst_addr->sa_family);
 }
 
 int rdma_resolve_addr(struct rdma_cm_id *id, struct sockaddr *src_addr,
@@ -2077,6 +2080,25 @@ static int cma_get_port(struct rdma_id_private *id_priv)
 	return ret;
 }
 
+static int cma_check_linklocal(struct rdma_dev_addr *dev_addr,
+			       struct sockaddr *addr)
+{
+#if defined(CONFIG_IPv6) || defined(CONFIG_IPV6_MODULE)
+	struct sockaddr_in6 *sin6;
+
+	if (addr->sa_family != AF_INET6)
+		return 0;
+
+	sin6 = (struct sockaddr_in6 *) addr;
+	if ((ipv6_addr_type(&sin6->sin6_addr) & IPV6_ADDR_LINKLOCAL) &&
+	    !sin6->sin6_scope_id)
+			return -EINVAL;
+
+	dev_addr->bound_dev_if = sin6->sin6_scope_id;
+#endif
+	return 0;
+}
+
 int rdma_bind_addr(struct rdma_cm_id *id, struct sockaddr *addr)
 {
 	struct rdma_id_private *id_priv;
@@ -2089,7 +2111,13 @@ int rdma_bind_addr(struct rdma_cm_id *id, struct sockaddr *addr)
 	if (!cma_comp_exch(id_priv, CMA_IDLE, CMA_ADDR_BOUND))
 		return -EINVAL;
 
-	if (!cma_any_addr(addr)) {
+	ret = cma_check_linklocal(&id->route.addr.dev_addr, addr);
+	if (ret)
+		goto err1;
+
+	if (cma_loopback_addr(addr)) {
+		ret = cma_bind_loopback(id_priv);
+	} else if (!cma_zero_addr(addr)) {
 		ret = rdma_translate_ip(addr, &id->route.addr.dev_addr);
 		if (ret)
 			goto err1;
@@ -2108,7 +2136,7 @@ int rdma_bind_addr(struct rdma_cm_id *id, struct sockaddr *addr)
 
 	return 0;
 err2:
-	if (!cma_any_addr(addr)) {
+	if (id_priv->cma_dev) {
 		mutex_lock(&lock);
 		cma_detach_from_dev(id_priv);
 		mutex_unlock(&lock);
@@ -2687,10 +2715,15 @@ static void cma_set_mgid(struct rdma_id_private *id_priv,
 	if (cma_any_addr(addr)) {
 		memset(mgid, 0, sizeof *mgid);
 	} else if ((addr->sa_family == AF_INET6) &&
-		   ((be32_to_cpu(sin6->sin6_addr.s6_addr32[0]) & 0xFF10A01B) ==
+		   ((be32_to_cpu(sin6->sin6_addr.s6_addr32[0]) & 0xFFF0FFFF) ==
 								 0xFF10A01B)) {
 		/* IPv6 address is an SA assigned MGID. */
 		memcpy(mgid, &sin6->sin6_addr, sizeof *mgid);
+	} else if ((addr->sa_family == AF_INET6)) {
+		ipv6_ib_mc_map(&sin6->sin6_addr, dev_addr->broadcast, mc_map);
+		if (id_priv->id.ps == RDMA_PS_UDP)
+			mc_map[7] = 0x01;	/* Use RDMA CM signature */
+		*mgid = *(union ib_gid *) (mc_map + 4);
 	} else {
 		ip_ib_mc_map(sin->sin_addr.s_addr, dev_addr->broadcast, mc_map);
 		if (id_priv->id.ps == RDMA_PS_UDP)
@@ -2716,7 +2749,7 @@ static int cma_join_ib_multicast(struct rdma_id_private *id_priv,
 	cma_set_mgid(id_priv, (struct sockaddr *) &mc->addr, &rec.mgid);
 	if (id_priv->id.ps == RDMA_PS_UDP)
 		rec.qkey = cpu_to_be32(RDMA_UDP_QKEY);
-	ib_addr_get_sgid(dev_addr, &rec.port_gid);
+	rdma_addr_get_sgid(dev_addr, &rec.port_gid);
 	rec.pkey = cpu_to_be16(ib_addr_get_pkey(dev_addr));
 	rec.join_state = 1;
 
@@ -2815,7 +2848,7 @@ static int cma_netdev_change(struct net_device *ndev, struct rdma_id_private *id
 
 	dev_addr = &id_priv->id.route.addr.dev_addr;
 
-	if ((dev_addr->src_dev == ndev) &&
+	if ((dev_addr->bound_dev_if == ndev->ifindex) &&
 	    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);

drivers/infiniband/core/sa_query.c

@@ -604,6 +604,12 @@ static int send_mad(struct ib_sa_query *query, int timeout_ms, gfp_t gfp_mask)
 	return ret ? ret : id;
 }
 
+void ib_sa_unpack_path(void *attribute, struct ib_sa_path_rec *rec)
+{
+	ib_unpack(path_rec_table, ARRAY_SIZE(path_rec_table), attribute, rec);
+}
+EXPORT_SYMBOL(ib_sa_unpack_path);
+
 static void ib_sa_path_rec_callback(struct ib_sa_query *sa_query,
 				    int status,
 				    struct ib_sa_mad *mad)

drivers/infiniband/core/ucma.c

@@ -43,6 +43,7 @@
 #include <rdma/rdma_user_cm.h>
 #include <rdma/ib_marshall.h>
 #include <rdma/rdma_cm.h>
+#include <rdma/rdma_cm_ib.h>
 
 MODULE_AUTHOR("Sean Hefty");
 MODULE_DESCRIPTION("RDMA Userspace Connection Manager Access");
@@ -562,9 +563,9 @@ static void ucma_copy_ib_route(struct rdma_ucm_query_route_resp *resp,
 	switch (route->num_paths) {
 	case 0:
 		dev_addr = &route->addr.dev_addr;
-		ib_addr_get_dgid(dev_addr,
+		rdma_addr_get_dgid(dev_addr,
 				   (union ib_gid *) &resp->ib_route[0].dgid);
-		ib_addr_get_sgid(dev_addr,
+		rdma_addr_get_sgid(dev_addr,
 				   (union ib_gid *) &resp->ib_route[0].sgid);
 		resp->ib_route[0].pkey = cpu_to_be16(ib_addr_get_pkey(dev_addr));
 		break;
@@ -812,6 +813,51 @@ static int ucma_set_option_id(struct ucma_context *ctx, int optname,
 	return ret;
 }
 
+static int ucma_set_ib_path(struct ucma_context *ctx,
+			    struct ib_path_rec_data *path_data, size_t optlen)
+{
+	struct ib_sa_path_rec sa_path;
+	struct rdma_cm_event event;
+	int ret;
+
+	if (optlen % sizeof(*path_data))
+		return -EINVAL;
+
+	for (; optlen; optlen -= sizeof(*path_data), path_data++) {
+		if (path_data->flags == (IB_PATH_GMP | IB_PATH_PRIMARY |
+					 IB_PATH_BIDIRECTIONAL))
+			break;
+	}
+
+	if (!optlen)
+		return -EINVAL;
+
+	ib_sa_unpack_path(path_data->path_rec, &sa_path);
+	ret = rdma_set_ib_paths(ctx->cm_id, &sa_path, 1);
+	if (ret)
+		return ret;
+
+	memset(&event, 0, sizeof event);
+	event.event = RDMA_CM_EVENT_ROUTE_RESOLVED;
+	return ucma_event_handler(ctx->cm_id, &event);
+}
+
+static int ucma_set_option_ib(struct ucma_context *ctx, int optname,
+			      void *optval, size_t optlen)
+{
+	int ret;
+
+	switch (optname) {
+	case RDMA_OPTION_IB_PATH:
+		ret = ucma_set_ib_path(ctx, optval, optlen);
+		break;
+	default:
+		ret = -ENOSYS;
+	}
+
+	return ret;
+}
+
 static int ucma_set_option_level(struct ucma_context *ctx, int level,
 				 int optname, void *optval, size_t optlen)
 {
@@ -821,6 +867,9 @@ static int ucma_set_option_level(struct ucma_context *ctx, int level,
 	case RDMA_OPTION_ID:
 		ret = ucma_set_option_id(ctx, optname, optval, optlen);
 		break;
+	case RDMA_OPTION_IB:
+		ret = ucma_set_option_ib(ctx, optname, optval, optlen);
+		break;
 	default:
 		ret = -ENOSYS;
 	}

drivers/infiniband/core/uverbs_cmd.c

@@ -285,7 +285,7 @@ ssize_t ib_uverbs_get_context(struct ib_uverbs_file *file,
 
 	ucontext = ibdev->alloc_ucontext(ibdev, &udata);
 	if (IS_ERR(ucontext)) {
-		ret = PTR_ERR(file->ucontext);
+		ret = PTR_ERR(ucontext);
 		goto err;
 	}
 

drivers/infiniband/hw/cxgb3/iwch_qp.c

@@ -365,18 +365,19 @@ int iwch_post_send(struct ib_qp *ibqp, struct ib_send_wr *wr,
 	spin_lock_irqsave(&qhp->lock, flag);
 	if (qhp->attr.state > IWCH_QP_STATE_RTS) {
 		spin_unlock_irqrestore(&qhp->lock, flag);
-		return -EINVAL;
+		err = -EINVAL;
+		goto out;
 	}
 	num_wrs = Q_FREECNT(qhp->wq.sq_rptr, qhp->wq.sq_wptr,
 		  qhp->wq.sq_size_log2);
 	if (num_wrs <= 0) {
 		spin_unlock_irqrestore(&qhp->lock, flag);
-		return -ENOMEM;
+		err = -ENOMEM;
+		goto out;
 	}
 	while (wr) {
 		if (num_wrs == 0) {
 			err = -ENOMEM;
-			*bad_wr = wr;
 			break;
 		}
 		idx = Q_PTR2IDX(qhp->wq.wptr, qhp->wq.size_log2);
@@ -428,10 +429,8 @@ int iwch_post_send(struct ib_qp *ibqp, struct ib_send_wr *wr,
 			     wr->opcode);
 			err = -EINVAL;
 		}
-		if (err) {
-			*bad_wr = wr;
+		if (err)
 			break;
-		}
 		wqe->send.wrid.id0.hi = qhp->wq.sq_wptr;
 		sqp->wr_id = wr->wr_id;
 		sqp->opcode = wr2opcode(t3_wr_opcode);
@@ -454,6 +453,10 @@ int iwch_post_send(struct ib_qp *ibqp, struct ib_send_wr *wr,
 	}
 	spin_unlock_irqrestore(&qhp->lock, flag);
 	ring_doorbell(qhp->wq.doorbell, qhp->wq.qpid);
+
+out:
+	if (err)
+		*bad_wr = wr;
 	return err;
 }
 
@@ -471,18 +474,19 @@ int iwch_post_receive(struct ib_qp *ibqp, struct ib_recv_wr *wr,
 	spin_lock_irqsave(&qhp->lock, flag);
 	if (qhp->attr.state > IWCH_QP_STATE_RTS) {
 		spin_unlock_irqrestore(&qhp->lock, flag);
-		return -EINVAL;
+		err = -EINVAL;
+		goto out;
 	}
 	num_wrs = Q_FREECNT(qhp->wq.rq_rptr, qhp->wq.rq_wptr,
 			    qhp->wq.rq_size_log2) - 1;
 	if (!wr) {
 		spin_unlock_irqrestore(&qhp->lock, flag);
-		return -EINVAL;
+		err = -ENOMEM;
+		goto out;
 	}
 	while (wr) {
 		if (wr->num_sge > T3_MAX_SGE) {
 			err = -EINVAL;
-			*bad_wr = wr;
 			break;
 		}
 		idx = Q_PTR2IDX(qhp->wq.wptr, qhp->wq.size_log2);
@@ -494,10 +498,10 @@ int iwch_post_receive(struct ib_qp *ibqp, struct ib_recv_wr *wr,
 				err = build_zero_stag_recv(qhp, wqe, wr);
 		else
 			err = -ENOMEM;
-		if (err) {
-			*bad_wr = wr;
+
+		if (err)
 			break;
-		}
+
 		build_fw_riwrh((void *) wqe, T3_WR_RCV, T3_COMPLETION_FLAG,
 			       Q_GENBIT(qhp->wq.wptr, qhp->wq.size_log2),
 			       0, sizeof(struct t3_receive_wr) >> 3, T3_SOPEOP);
@@ -511,6 +515,10 @@ int iwch_post_receive(struct ib_qp *ibqp, struct ib_recv_wr *wr,
 	}
 	spin_unlock_irqrestore(&qhp->lock, flag);
 	ring_doorbell(qhp->wq.doorbell, qhp->wq.qpid);
+
+out:
+	if (err)
+		*bad_wr = wr;
 	return err;
 }
 

drivers/infiniband/hw/ehca/ehca_classes.h

@@ -375,6 +375,7 @@ extern rwlock_t ehca_qp_idr_lock;
 extern rwlock_t ehca_cq_idr_lock;
 extern struct idr ehca_qp_idr;
 extern struct idr ehca_cq_idr;
+extern spinlock_t shca_list_lock;
 
 extern int ehca_static_rate;
 extern int ehca_port_act_time;

drivers/infiniband/hw/ehca/ehca_eq.c

@@ -169,12 +169,15 @@ int ehca_destroy_eq(struct ehca_shca *shca, struct ehca_eq *eq)
 	unsigned long flags;
 	u64 h_ret;
 
-	spin_lock_irqsave(&eq->spinlock, flags);
 	ibmebus_free_irq(eq->ist, (void *)shca);
 
-	h_ret = hipz_h_destroy_eq(shca->ipz_hca_handle, eq);
+	spin_lock_irqsave(&shca_list_lock, flags);
+	eq->is_initialized = 0;
+	spin_unlock_irqrestore(&shca_list_lock, flags);
 
-	spin_unlock_irqrestore(&eq->spinlock, flags);
+	tasklet_kill(&eq->interrupt_task);
+
+	h_ret = hipz_h_destroy_eq(shca->ipz_hca_handle, eq);
 
 	if (h_ret != H_SUCCESS) {
 		ehca_err(&shca->ib_device, "Can't free EQ resources.");

drivers/infiniband/hw/ehca/ehca_main.c

@@ -123,7 +123,7 @@ DEFINE_IDR(ehca_qp_idr);
 DEFINE_IDR(ehca_cq_idr);
 
 static LIST_HEAD(shca_list); /* list of all registered ehcas */
-static DEFINE_SPINLOCK(shca_list_lock);
+DEFINE_SPINLOCK(shca_list_lock);
 
 static struct timer_list poll_eqs_timer;
 

drivers/infiniband/hw/ehca/ehca_reqs.c

@@ -400,7 +400,6 @@ static inline void map_ib_wc_status(u32 cqe_status,
 
 static inline int post_one_send(struct ehca_qp *my_qp,
 			 struct ib_send_wr *cur_send_wr,
-			 struct ib_send_wr **bad_send_wr,
 			 int hidden)
 {
 	struct ehca_wqe *wqe_p;
@@ -412,8 +411,6 @@ static inline int post_one_send(struct ehca_qp *my_qp,
 	wqe_p = ipz_qeit_get_inc(&my_qp->ipz_squeue);
 	if (unlikely(!wqe_p)) {
 		/* too many posted work requests: queue overflow */
-		if (bad_send_wr)
-			*bad_send_wr = cur_send_wr;
 		ehca_err(my_qp->ib_qp.device, "Too many posted WQEs "
 			 "qp_num=%x", my_qp->ib_qp.qp_num);
 		return -ENOMEM;
@@ -433,8 +430,6 @@ static inline int post_one_send(struct ehca_qp *my_qp,
 	 */
 	if (unlikely(ret)) {
 		my_qp->ipz_squeue.current_q_offset = start_offset;
-		if (bad_send_wr)
-			*bad_send_wr = cur_send_wr;
 		ehca_err(my_qp->ib_qp.device, "Could not write WQE "
 			 "qp_num=%x", my_qp->ib_qp.qp_num);
 		return -EINVAL;
@@ -448,7 +443,6 @@ int ehca_post_send(struct ib_qp *qp,
 		   struct ib_send_wr **bad_send_wr)
 {
 	struct ehca_qp *my_qp = container_of(qp, struct ehca_qp, ib_qp);
-	struct ib_send_wr *cur_send_wr;
 	int wqe_cnt = 0;
 	int ret = 0;
 	unsigned long flags;
@@ -457,7 +451,8 @@ int ehca_post_send(struct ib_qp *qp,
 	if (unlikely(my_qp->state < IB_QPS_RTS)) {
 		ehca_err(qp->device, "Invalid QP state  qp_state=%d qpn=%x",
 			 my_qp->state, qp->qp_num);
-		return -EINVAL;
+		ret = -EINVAL;
+		goto out;
 	}
 
 	/* LOCK the QUEUE */
@@ -476,24 +471,21 @@ int ehca_post_send(struct ib_qp *qp,
 		struct ib_send_wr circ_wr;
 		memset(&circ_wr, 0, sizeof(circ_wr));
 		circ_wr.opcode = IB_WR_RDMA_READ;
-		post_one_send(my_qp, &circ_wr, NULL, 1); /* ignore retcode */
+		post_one_send(my_qp, &circ_wr, 1); /* ignore retcode */
 		wqe_cnt++;
 		ehca_dbg(qp->device, "posted circ wr  qp_num=%x", qp->qp_num);
 		my_qp->message_count = my_qp->packet_count = 0;
 	}
 
 	/* loop processes list of send reqs */
-	for (cur_send_wr = send_wr; cur_send_wr != NULL;
-	     cur_send_wr = cur_send_wr->next) {
-		ret = post_one_send(my_qp, cur_send_wr, bad_send_wr, 0);
+	while (send_wr) {
+		ret = post_one_send(my_qp, send_wr, 0);
 		if (unlikely(ret)) {
-			/* if one or more WQEs were successful, don't fail */
-			if (wqe_cnt)
-				ret = 0;
 			goto post_send_exit0;
 		}
 		wqe_cnt++;
-	} /* eof for cur_send_wr */
+		send_wr = send_wr->next;
+	}
 
 post_send_exit0:
 	iosync(); /* serialize GAL register access */
@@ -503,6 +495,10 @@ int ehca_post_send(struct ib_qp *qp,
 			 my_qp, qp->qp_num, wqe_cnt, ret);
 	my_qp->message_count += wqe_cnt;
 	spin_unlock_irqrestore(&my_qp->spinlock_s, flags);
+
+out:
+	if (ret)
+		*bad_send_wr = send_wr;
 	return ret;
 }
 
@@ -511,7 +507,6 @@ static int internal_post_recv(struct ehca_qp *my_qp,
 			      struct ib_recv_wr *recv_wr,
 			      struct ib_recv_wr **bad_recv_wr)
 {
-	struct ib_recv_wr *cur_recv_wr;
 	struct ehca_wqe *wqe_p;
 	int wqe_cnt = 0;
 	int ret = 0;
@@ -522,27 +517,23 @@ static int internal_post_recv(struct ehca_qp *my_qp,
 	if (unlikely(!HAS_RQ(my_qp))) {
 		ehca_err(dev, "QP has no RQ  ehca_qp=%p qp_num=%x ext_type=%d",
 			 my_qp, my_qp->real_qp_num, my_qp->ext_type);
-		return -ENODEV;
+		ret = -ENODEV;
+		goto out;
 	}
 
 	/* LOCK the QUEUE */
 	spin_lock_irqsave(&my_qp->spinlock_r, flags);
 
-	/* loop processes list of send reqs */
-	for (cur_recv_wr = recv_wr; cur_recv_wr != NULL;
-	     cur_recv_wr = cur_recv_wr->next) {
+	/* loop processes list of recv reqs */
+	while (recv_wr) {
 		u64 start_offset = my_qp->ipz_rqueue.current_q_offset;
 		/* get pointer next to free WQE */
 		wqe_p = ipz_qeit_get_inc(&my_qp->ipz_rqueue);
 		if (unlikely(!wqe_p)) {
 			/* too many posted work requests: queue overflow */
-			if (bad_recv_wr)
-				*bad_recv_wr = cur_recv_wr;
-			if (wqe_cnt == 0) {
 			ret = -ENOMEM;
 			ehca_err(dev, "Too many posted WQEs "
 				"qp_num=%x", my_qp->real_qp_num);
-			}
 			goto post_recv_exit0;
 		}
 		/*
@@ -552,7 +543,7 @@ static int internal_post_recv(struct ehca_qp *my_qp,
 		rq_map_idx = start_offset / my_qp->ipz_rqueue.qe_size;
 
 		/* write a RECV WQE into the QUEUE */
-		ret = ehca_write_rwqe(&my_qp->ipz_rqueue, wqe_p, cur_recv_wr,
+		ret = ehca_write_rwqe(&my_qp->ipz_rqueue, wqe_p, recv_wr,
 				rq_map_idx);
 		/*
 		 * if something failed,
@@ -560,22 +551,20 @@ static int internal_post_recv(struct ehca_qp *my_qp,
 		 */
 		if (unlikely(ret)) {
 			my_qp->ipz_rqueue.current_q_offset = start_offset;
-			*bad_recv_wr = cur_recv_wr;
-			if (wqe_cnt == 0) {
 			ret = -EINVAL;
 			ehca_err(dev, "Could not write WQE "
 				"qp_num=%x", my_qp->real_qp_num);
-			}
 			goto post_recv_exit0;
 		}
 
 		qmap_entry = &my_qp->rq_map.map[rq_map_idx];
-		qmap_entry->app_wr_id = get_app_wr_id(cur_recv_wr->wr_id);
+		qmap_entry->app_wr_id = get_app_wr_id(recv_wr->wr_id);
 		qmap_entry->reported = 0;
 		qmap_entry->cqe_req = 1;
 
 		wqe_cnt++;
-	} /* eof for cur_recv_wr */
+		recv_wr = recv_wr->next;
+	} /* eof for recv_wr */
 
 post_recv_exit0:
 	iosync(); /* serialize GAL register access */
@@ -584,6 +573,11 @@ static int internal_post_recv(struct ehca_qp *my_qp,
 	    ehca_dbg(dev, "ehca_qp=%p qp_num=%x wqe_cnt=%d ret=%i",
 		     my_qp, my_qp->real_qp_num, wqe_cnt, ret);
 	spin_unlock_irqrestore(&my_qp->spinlock_r, flags);
+
+out:
+	if (ret)
+		*bad_recv_wr = recv_wr;
+
 	return ret;
 }
 
@@ -597,6 +591,7 @@ int ehca_post_recv(struct ib_qp *qp,
 	if (unlikely(my_qp->state == IB_QPS_RESET)) {
 		ehca_err(qp->device, "Invalid QP state  qp_state=%d qpn=%x",
 			 my_qp->state, qp->qp_num);
+		*bad_recv_wr = recv_wr;
 		return -EINVAL;
 	}
 

drivers/infiniband/hw/ipath/ipath_driver.c

@@ -39,6 +39,7 @@
 #include <linux/delay.h>
 #include <linux/netdevice.h>
 #include <linux/vmalloc.h>
+#include <linux/bitmap.h>
 
 #include "ipath_kernel.h"
 #include "ipath_verbs.h"
@@ -1697,7 +1698,7 @@ void ipath_chg_pioavailkernel(struct ipath_devdata *dd, unsigned start,
 			      unsigned len, int avail)
 {
 	unsigned long flags;
-	unsigned end, cnt = 0, next;
+	unsigned end, cnt = 0;
 
 	/* There are two bits per send buffer (busy and generation) */
 	start *= 2;
@@ -1748,12 +1749,7 @@ void ipath_chg_pioavailkernel(struct ipath_devdata *dd, unsigned start,
 
 	if (dd->ipath_pioupd_thresh) {
 		end = 2 * (dd->ipath_piobcnt2k + dd->ipath_piobcnt4k);
-		next = find_first_bit(dd->ipath_pioavailkernel, end);
-		while (next < end) {
-			cnt++;
-			next = find_next_bit(dd->ipath_pioavailkernel, end,
-					next + 1);
-		}
+		cnt = bitmap_weight(dd->ipath_pioavailkernel, end);
 	}
 	spin_unlock_irqrestore(&ipath_pioavail_lock, flags);
 

drivers/infiniband/hw/mlx4/main.c

@@ -103,7 +103,7 @@ static int mlx4_ib_query_device(struct ib_device *ibdev,
 		props->device_cap_flags |= IB_DEVICE_UD_AV_PORT_ENFORCE;
 	if (dev->dev->caps.flags & MLX4_DEV_CAP_FLAG_IPOIB_CSUM)
 		props->device_cap_flags |= IB_DEVICE_UD_IP_CSUM;
-	if (dev->dev->caps.max_gso_sz)
+	if (dev->dev->caps.max_gso_sz && dev->dev->caps.flags & MLX4_DEV_CAP_FLAG_BLH)
 		props->device_cap_flags |= IB_DEVICE_UD_TSO;
 	if (dev->dev->caps.bmme_flags & MLX4_BMME_FLAG_RESERVED_LKEY)
 		props->device_cap_flags |= IB_DEVICE_LOCAL_DMA_LKEY;

drivers/infiniband/hw/mlx4/qp.c

@@ -54,7 +54,8 @@ enum {
 	/*
 	 * Largest possible UD header: send with GRH and immediate data.
 	 */
-	MLX4_IB_UD_HEADER_SIZE		= 72
+	MLX4_IB_UD_HEADER_SIZE		= 72,
+	MLX4_IB_LSO_HEADER_SPARE	= 128,
 };
 
 struct mlx4_ib_sqp {
@@ -67,7 +68,8 @@ struct mlx4_ib_sqp {
 };
 
 enum {
-	MLX4_IB_MIN_SQ_STRIDE = 6
+	MLX4_IB_MIN_SQ_STRIDE	= 6,
+	MLX4_IB_CACHE_LINE_SIZE	= 64,
 };
 
 static const __be32 mlx4_ib_opcode[] = {
@@ -261,7 +263,7 @@ static int send_wqe_overhead(enum ib_qp_type type, u32 flags)
 	case IB_QPT_UD:
 		return sizeof (struct mlx4_wqe_ctrl_seg) +
 			sizeof (struct mlx4_wqe_datagram_seg) +
-			((flags & MLX4_IB_QP_LSO) ? 64 : 0);
+			((flags & MLX4_IB_QP_LSO) ? MLX4_IB_LSO_HEADER_SPARE : 0);
 	case IB_QPT_UC:
 		return sizeof (struct mlx4_wqe_ctrl_seg) +
 			sizeof (struct mlx4_wqe_raddr_seg);
@@ -897,7 +899,6 @@ static int __mlx4_ib_modify_qp(struct ib_qp *ibqp,
 
 	context->flags = cpu_to_be32((to_mlx4_state(new_state) << 28) |
 				     (to_mlx4_st(ibqp->qp_type) << 16));
-	context->flags     |= cpu_to_be32(1 << 8); /* DE? */
 
 	if (!(attr_mask & IB_QP_PATH_MIG_STATE))
 		context->flags |= cpu_to_be32(MLX4_QP_PM_MIGRATED << 11);
@@ -1467,16 +1468,12 @@ static void __set_data_seg(struct mlx4_wqe_data_seg *dseg, struct ib_sge *sg)
 
 static int build_lso_seg(struct mlx4_wqe_lso_seg *wqe, struct ib_send_wr *wr,
 			 struct mlx4_ib_qp *qp, unsigned *lso_seg_len,
-			 __be32 *lso_hdr_sz)
+			 __be32 *lso_hdr_sz, __be32 *blh)
 {
 	unsigned halign = ALIGN(sizeof *wqe + wr->wr.ud.hlen, 16);
 
-	/*
-	 * This is a temporary limitation and will be removed in
-	 * a forthcoming FW release:
-	 */
-	if (unlikely(halign > 64))
-		return -EINVAL;
+	if (unlikely(halign > MLX4_IB_CACHE_LINE_SIZE))
+		*blh = cpu_to_be32(1 << 6);
 
 	if (unlikely(!(qp->flags & MLX4_IB_QP_LSO) &&
 		     wr->num_sge > qp->sq.max_gs - (halign >> 4)))
@@ -1522,6 +1519,7 @@ int mlx4_ib_post_send(struct ib_qp *ibqp, struct ib_send_wr *wr,
 	__be32 dummy;
 	__be32 *lso_wqe;
 	__be32 uninitialized_var(lso_hdr_sz);
+	__be32 blh;
 	int i;
 
 	spin_lock_irqsave(&qp->sq.lock, flags);
@@ -1530,6 +1528,7 @@ int mlx4_ib_post_send(struct ib_qp *ibqp, struct ib_send_wr *wr,
 
 	for (nreq = 0; wr; ++nreq, wr = wr->next) {
 		lso_wqe = &dummy;
+		blh = 0;
 
 		if (mlx4_wq_overflow(&qp->sq, nreq, qp->ibqp.send_cq)) {
 			err = -ENOMEM;
@@ -1616,7 +1615,7 @@ int mlx4_ib_post_send(struct ib_qp *ibqp, struct ib_send_wr *wr,
 			size += sizeof (struct mlx4_wqe_datagram_seg) / 16;
 
 			if (wr->opcode == IB_WR_LSO) {
-				err = build_lso_seg(wqe, wr, qp, &seglen, &lso_hdr_sz);
+				err = build_lso_seg(wqe, wr, qp, &seglen, &lso_hdr_sz, &blh);
 				if (unlikely(err)) {
 					*bad_wr = wr;
 					goto out;
@@ -1687,7 +1686,7 @@ int mlx4_ib_post_send(struct ib_qp *ibqp, struct ib_send_wr *wr,
 		}
 
 		ctrl->owner_opcode = mlx4_ib_opcode[wr->opcode] |
-			(ind & qp->sq.wqe_cnt ? cpu_to_be32(1 << 31) : 0);
+			(ind & qp->sq.wqe_cnt ? cpu_to_be32(1 << 31) : 0) | blh;
 
 		stamp = ind + qp->sq_spare_wqes;
 		ind += DIV_ROUND_UP(size * 16, 1U << qp->sq.wqe_shift);

drivers/infiniband/hw/nes/Kconfig

@@ -4,14 +4,13 @@ config INFINIBAND_NES
 	select LIBCRC32C
 	select INET_LRO
 	---help---
-	  This is a low-level driver for NetEffect RDMA enabled
-	  Network Interface Cards (RNIC).
+	  This is the RDMA Network Interface Card (RNIC) driver for
+	  NetEffect Ethernet Cluster Server Adapters.
 
 config INFINIBAND_NES_DEBUG
 	bool "Verbose debugging output"
 	depends on INFINIBAND_NES
 	default n
 	---help---
-	  This option causes the NetEffect RNIC driver to produce debug
-	  messages.  Select this if you are developing the driver
-	  or trying to diagnose a problem.
+	  This option enables debug messages from the NetEffect RNIC
+	  driver.  Select this if you are diagnosing a problem.

drivers/infiniband/hw/nes/nes.c

 /*
- * Copyright (c) 2006 - 2009 Intel-NE, Inc.  All rights reserved.
+ * Copyright (c) 2006 - 2009 Intel Corporation.  All rights reserved.
  * Copyright (c) 2005 Open Grid Computing, Inc. All rights reserved.
  *
  * This software is available to you under a choice of one of two
@@ -521,7 +521,8 @@ static int __devinit nes_probe(struct pci_dev *pcidev, const struct pci_device_i
 	spin_lock_init(&nesdev->indexed_regs_lock);
 
 	/* Remap the PCI registers in adapter BAR0 to kernel VA space */
-	mmio_regs = ioremap_nocache(pci_resource_start(pcidev, BAR_0), sizeof(mmio_regs));
+	mmio_regs = ioremap_nocache(pci_resource_start(pcidev, BAR_0),
+				    pci_resource_len(pcidev, BAR_0));
 	if (mmio_regs == NULL) {
 		printk(KERN_ERR PFX "Unable to remap BAR0\n");
 		ret = -EIO;

drivers/infiniband/hw/nes/nes.h

 /*
- * Copyright (c) 2006 - 2009 Intel-NE, Inc.  All rights reserved.
+ * Copyright (c) 2006 - 2009 Intel Corporation.  All rights reserved.
  * Copyright (c) 2005 Open Grid Computing, Inc. All rights reserved.
  *
  * This software is available to you under a choice of one of two

drivers/infiniband/hw/nes/nes_cm.h

 /*
- * Copyright (c) 2006 - 2009 Intel-NE, Inc.  All rights reserved.
+ * Copyright (c) 2006 - 2009 Intel Corporation.  All rights reserved.
  *
  * This software is available to you under a choice of one of two
  * licenses.  You may choose to be licensed under the terms of the GNU
@@ -47,6 +47,8 @@
 #define IEFT_MPA_KEY_REP  "MPA ID Rep Frame"
 #define IETF_MPA_KEY_SIZE 16
 #define IETF_MPA_VERSION  1
+#define IETF_MAX_PRIV_DATA_LEN 512
+#define IETF_MPA_FRAME_SIZE     20
 
 enum ietf_mpa_flags {
 	IETF_MPA_FLAGS_MARKERS = 0x80,	/* receive Markers */
@@ -169,7 +171,7 @@ struct nes_timer_entry {
 
 #define NES_CM_DEF_SEQ2      0x18ed5740
 #define NES_CM_DEF_LOCAL_ID2 0xb807
-#define	MAX_CM_BUFFER	512
+#define	MAX_CM_BUFFER	(IETF_MPA_FRAME_SIZE + IETF_MAX_PRIV_DATA_LEN)
 
 
 typedef u32 nes_addr_t;
@@ -198,6 +200,7 @@ enum nes_cm_node_state {
 	NES_CM_STATE_TIME_WAIT,
 	NES_CM_STATE_LAST_ACK,
 	NES_CM_STATE_CLOSING,
+	NES_CM_STATE_LISTENER_DESTROYED,
 	NES_CM_STATE_CLOSED
 };
 

drivers/infiniband/hw/nes/nes_context.h

 /*
- * Copyright (c) 2006 - 2009 Intel-NE, Inc.  All rights reserved.
+ * Copyright (c) 2006 - 2009 Intel Corporation.  All rights reserved.
  *
  * This software is available to you under a choice of one of two
  * licenses.  You may choose to be licensed under the terms of the GNU

drivers/infiniband/hw/nes/nes_hw.c

 /*
- * Copyright (c) 2006 - 2009 Intel-NE, Inc.  All rights reserved.
+ * Copyright (c) 2006 - 2009 Intel Corporation.  All rights reserved.
  *
  * This software is available to you under a choice of one of two
  * licenses.  You may choose to be licensed under the terms of the GNU
@@ -424,8 +424,9 @@ struct nes_adapter *nes_init_adapter(struct nes_device *nesdev, u8 hw_rev) {
 
 	nesadapter->base_pd = 1;
 
-	nesadapter->device_cap_flags =
-		IB_DEVICE_LOCAL_DMA_LKEY | IB_DEVICE_MEM_WINDOW;
+	nesadapter->device_cap_flags = IB_DEVICE_LOCAL_DMA_LKEY |
+				       IB_DEVICE_MEM_WINDOW |
+				       IB_DEVICE_MEM_MGT_EXTENSIONS;
 
 	nesadapter->allocated_qps = (unsigned long *)&(((unsigned char *)nesadapter)
 			[(sizeof(struct nes_adapter)+(sizeof(unsigned long)-1))&(~(sizeof(unsigned long)-1))]);
@@ -436,11 +437,12 @@ struct nes_adapter *nes_init_adapter(struct nes_device *nesdev, u8 hw_rev) {
 	nesadapter->qp_table = (struct nes_qp **)(&nesadapter->allocated_arps[BITS_TO_LONGS(arp_table_size)]);
 
 
-	/* mark the usual suspect QPs and CQs as in use */
+	/* mark the usual suspect QPs, MR and CQs as in use */
 	for (u32temp = 0; u32temp < NES_FIRST_QPN; u32temp++) {
 		set_bit(u32temp, nesadapter->allocated_qps);
 		set_bit(u32temp, nesadapter->allocated_cqs);
 	}
+	set_bit(0, nesadapter->allocated_mrs);
 
 	for (u32temp = 0; u32temp < 20; u32temp++)
 		set_bit(u32temp, nesadapter->allocated_pds);
@@ -481,7 +483,7 @@ struct nes_adapter *nes_init_adapter(struct nes_device *nesdev, u8 hw_rev) {
 	nesadapter->max_irrq_wr = (u32temp >> 16) & 3;
 
 	nesadapter->max_sge = 4;
-	nesadapter->max_cqe = 32767;
+	nesadapter->max_cqe = 32766;
 
 	if (nes_read_eeprom_values(nesdev, nesadapter)) {
 		printk(KERN_ERR PFX "Unable to read EEPROM data.\n");
@@ -1355,6 +1357,8 @@ int nes_init_phy(struct nes_device *nesdev)
 	}
 	if ((phy_type == NES_PHY_TYPE_ARGUS) ||
 	    (phy_type == NES_PHY_TYPE_SFP_D)) {
+		u32 first_time = 1;
+
 		/* Check firmware heartbeat */
 		nes_read_10G_phy_reg(nesdev, phy_index, 0x3, 0xd7ee);
 		temp_phy_data = (u16)nes_read_indexed(nesdev, NES_IDX_MAC_MDIO_CONTROL);
@@ -1362,8 +1366,13 @@ int nes_init_phy(struct nes_device *nesdev)
 		nes_read_10G_phy_reg(nesdev, phy_index, 0x3, 0xd7ee);
 		temp_phy_data2 = (u16)nes_read_indexed(nesdev, NES_IDX_MAC_MDIO_CONTROL);
 
-		if (temp_phy_data != temp_phy_data2)
+		if (temp_phy_data != temp_phy_data2) {
+			nes_read_10G_phy_reg(nesdev, phy_index, 0x3, 0xd7fd);
+			temp_phy_data = (u16)nes_read_indexed(nesdev, NES_IDX_MAC_MDIO_CONTROL);
+			if ((temp_phy_data & 0xff) > 0x20)
 				return 0;
+			printk(PFX "Reinitializing PHY\n");
+		}
 
 		/* no heartbeat, configure the PHY */
 		nes_write_10G_phy_reg(nesdev, phy_index, 0x1, 0x0000, 0x8000);
@@ -1399,7 +1408,7 @@ int nes_init_phy(struct nes_device *nesdev)
 		temp_phy_data = (u16)nes_read_indexed(nesdev, NES_IDX_MAC_MDIO_CONTROL);
 		do {
 			if (counter++ > 150) {
-				nes_debug(NES_DBG_PHY, "No PHY heartbeat\n");
+				printk(PFX "No PHY heartbeat\n");
 				break;
 			}
 			mdelay(1);
@@ -1413,11 +1422,20 @@ int nes_init_phy(struct nes_device *nesdev)
 			nes_read_10G_phy_reg(nesdev, phy_index, 0x3, 0xd7fd);
 			temp_phy_data = (u16)nes_read_indexed(nesdev, NES_IDX_MAC_MDIO_CONTROL);
 			if (counter++ > 300) {
-				nes_debug(NES_DBG_PHY, "PHY did not track\n");
+				if (((temp_phy_data & 0xff) == 0x0) && first_time) {
+					first_time = 0;
+					counter = 0;
+					/* reset AMCC PHY and try again */
+					nes_write_10G_phy_reg(nesdev, phy_index, 0x3, 0xe854, 0x00c0);
+					nes_write_10G_phy_reg(nesdev, phy_index, 0x3, 0xe854, 0x0040);
+					continue;
+				} else {
+					printk(PFX "PHY did not track\n");
 					break;
 				}
+			}
 			mdelay(10);
-		} while (((temp_phy_data & 0xff) != 0x50) && ((temp_phy_data & 0xff) != 0x70));
+		} while ((temp_phy_data & 0xff) < 0x30);
 
 		/* setup signal integrity */
 		nes_write_10G_phy_reg(nesdev, phy_index, 0x1, 0xd003, 0x0000);

drivers/infiniband/hw/nes/nes_hw.h

 /*
-* Copyright (c) 2006 - 2009 Intel-NE, Inc.  All rights reserved.
+* Copyright (c) 2006 - 2009 Intel Corporation.  All rights reserved.
 *
 * This software is available to you under a choice of one of two
 * licenses.  You may choose to be licensed under the terms of the GNU
@@ -546,11 +546,23 @@ enum nes_iwarp_sq_fmr_wqe_word_idx {
 	NES_IWARP_SQ_FMR_WQE_PBL_LENGTH_IDX = 14,
 };
 
+enum nes_iwarp_sq_fmr_opcodes {
+	NES_IWARP_SQ_FMR_WQE_ZERO_BASED			= (1<<6),
+	NES_IWARP_SQ_FMR_WQE_PAGE_SIZE_4K		= (0<<7),
+	NES_IWARP_SQ_FMR_WQE_PAGE_SIZE_2M		= (1<<7),
+	NES_IWARP_SQ_FMR_WQE_RIGHTS_ENABLE_LOCAL_READ	= (1<<16),
+	NES_IWARP_SQ_FMR_WQE_RIGHTS_ENABLE_LOCAL_WRITE 	= (1<<17),
+	NES_IWARP_SQ_FMR_WQE_RIGHTS_ENABLE_REMOTE_READ 	= (1<<18),
+	NES_IWARP_SQ_FMR_WQE_RIGHTS_ENABLE_REMOTE_WRITE = (1<<19),
+	NES_IWARP_SQ_FMR_WQE_RIGHTS_ENABLE_WINDOW_BIND 	= (1<<20),
+};
+
+#define NES_IWARP_SQ_FMR_WQE_MR_LENGTH_HIGH_MASK	0xFF;
+
 enum nes_iwarp_sq_locinv_wqe_word_idx {
 	NES_IWARP_SQ_LOCINV_WQE_INV_STAG_IDX = 6,
 };
 
-
 enum nes_iwarp_rq_wqe_word_idx {
 	NES_IWARP_RQ_WQE_TOTAL_PAYLOAD_IDX = 1,
 	NES_IWARP_RQ_WQE_COMP_CTX_LOW_IDX = 2,
@@ -1153,6 +1165,19 @@ struct nes_pbl {
 	/* TODO: need to add list for two level tables */
 };
 
+#define NES_4K_PBL_CHUNK_SIZE	4096
+
+struct nes_fast_mr_wqe_pbl {
+	u64		*kva;
+	dma_addr_t	paddr;
+};
+
+struct nes_ib_fast_reg_page_list {
+	struct ib_fast_reg_page_list	ibfrpl;
+	struct nes_fast_mr_wqe_pbl 	nes_wqe_pbl;
+	u64 				pbl;
+};
+
 struct nes_listener {
 	struct work_struct      work;
 	struct workqueue_struct *wq;

drivers/infiniband/hw/nes/nes_nic.c

 /*
- * Copyright (c) 2006 - 2009 Intel-NE, Inc.  All rights reserved.
+ * Copyright (c) 2006 - 2009 Intel Corporation.  All rights reserved.
  *
  * This software is available to you under a choice of one of two
  * licenses.  You may choose to be licensed under the terms of the GNU

drivers/infiniband/hw/nes/nes_user.h

 /*
- * Copyright (c) 2006 - 2009 Intel-NE, Inc.  All rights reserved.
+ * Copyright (c) 2006 - 2009 Intel Corporation.  All rights reserved.
  * Copyright (c) 2005 Topspin Communications.  All rights reserved.
  * Copyright (c) 2005 Cisco Systems.  All rights reserved.
  * Copyright (c) 2005 Open Grid Computing, Inc. All rights reserved.
@@ -86,6 +86,7 @@ enum iwnes_memreg_type {
 	IWNES_MEMREG_TYPE_CQ = 0x0002,
 	IWNES_MEMREG_TYPE_MW = 0x0003,
 	IWNES_MEMREG_TYPE_FMR = 0x0004,
+	IWNES_MEMREG_TYPE_FMEM = 0x0005,
 };
 
 struct nes_mem_reg_req {

drivers/infiniband/hw/nes/nes_utils.c

 /*
- * Copyright (c) 2006 - 2009 Intel-NE, Inc.  All rights reserved.
+ * Copyright (c) 2006 - 2009 Intel Corporation.  All rights reserved.
  *
  * This software is available to you under a choice of one of two
  * licenses.  You may choose to be licensed under the terms of the GNU

drivers/infiniband/hw/nes/nes_verbs.h

 /*
- * Copyright (c) 2006 - 2009 Intel-NE, Inc.  All rights reserved.
+ * Copyright (c) 2006 - 2009 Intel Corporation.  All rights reserved.
  * Copyright (c) 2005 Open Grid Computing, Inc. All rights reserved.
  *
  * This software is available to you under a choice of one of two
@@ -135,19 +135,15 @@ struct nes_qp {
 	struct ib_qp          ibqp;
 	void                  *allocated_buffer;
 	struct iw_cm_id       *cm_id;
-	struct workqueue_struct *wq;
 	struct nes_cq         *nesscq;
 	struct nes_cq         *nesrcq;
 	struct nes_pd         *nespd;
 	void *cm_node; /* handle of the node this QP is associated with */
 	struct ietf_mpa_frame *ietf_frame;
 	dma_addr_t            ietf_frame_pbase;
-	wait_queue_head_t     state_waitq;
 	struct ib_mr          *lsmm_mr;
-	unsigned long         socket;
 	struct nes_hw_qp      hwqp;
 	struct work_struct    work;
-	struct work_struct    ae_work;
 	enum ib_qp_state      ibqp_state;
 	u32                   iwarp_state;
 	u32                   hte_index;
@@ -165,19 +161,20 @@ struct nes_qp {
 	struct page           *page;
 	struct timer_list     terminate_timer;
 	enum ib_event_type    terminate_eventtype;
-	wait_queue_head_t     kick_waitq;
-	u16                   in_disconnect;
+	u16                   active_conn:1;
+	u16                   skip_lsmm:1;
+	u16                   user_mode:1;
+	u16                   hte_added:1;
+	u16                   flush_issued:1;
+	u16                   destroyed:1;
+	u16                   sig_all:1;
+	u16                   rsvd:9;
 	u16                   private_data_len;
 	u16                   term_sq_flush_code;
 	u16                   term_rq_flush_code;
-	u8                    active_conn;
-	u8                    skip_lsmm;
-	u8                    user_mode;
-	u8                    hte_added;
 	u8                    hw_iwarp_state;
-	u8                    flush_issued;
 	u8                    hw_tcp_state;
 	u8                    term_flags;
-	u8                    destroyed;
+	u8                    sq_kmapped;
 };
 #endif			/* NES_VERBS_H */

drivers/infiniband/ulp/ipoib/ipoib_main.c

@@ -884,6 +884,7 @@ struct ipoib_neigh *ipoib_neigh_alloc(struct neighbour *neighbour,
 
 	neigh->neighbour = neighbour;
 	neigh->dev = dev;
+	memset(&neigh->dgid.raw, 0, sizeof (union ib_gid));
 	*to_ipoib_neigh(neighbour) = neigh;
 	skb_queue_head_init(&neigh->queue);
 	ipoib_cm_set(neigh, NULL);

drivers/infiniband/ulp/iser/iser_memory.c

@@ -209,6 +209,8 @@ void iser_finalize_rdma_unaligned_sg(struct iscsi_iser_task *iser_task,
 	mem_copy->copy_buf = NULL;
 }
 
+#define IS_4K_ALIGNED(addr)	((((unsigned long)addr) & ~MASK_4K) == 0)
+
 /**
  * iser_sg_to_page_vec - Translates scatterlist entries to physical addresses
  * and returns the length of resulting physical address array (may be less than
@@ -221,62 +223,52 @@ void iser_finalize_rdma_unaligned_sg(struct iscsi_iser_task *iser_task,
  * where --few fragments of the same page-- are present in the SG as
  * consecutive elements. Also, it handles one entry SG.
  */
+
 static int iser_sg_to_page_vec(struct iser_data_buf *data,
 			       struct iser_page_vec *page_vec,
 			       struct ib_device *ibdev)
 {
-	struct scatterlist *sgl = (struct scatterlist *)data->buf;
-	struct scatterlist *sg;
-	u64 first_addr, last_addr, page;
-	int end_aligned;
-	unsigned int cur_page = 0;
+	struct scatterlist *sg, *sgl = (struct scatterlist *)data->buf;
+	u64 start_addr, end_addr, page, chunk_start = 0;
 	unsigned long total_sz = 0;
-	int i;
+	unsigned int dma_len;
+	int i, new_chunk, cur_page, last_ent = data->dma_nents - 1;
 
 	/* compute the offset of first element */
 	page_vec->offset = (u64) sgl[0].offset & ~MASK_4K;
 
+	new_chunk = 1;
+	cur_page  = 0;
 	for_each_sg(sgl, sg, data->dma_nents, i) {
-		unsigned int dma_len = ib_sg_dma_len(ibdev, sg);
-
-		total_sz += dma_len;
-
-		first_addr = ib_sg_dma_address(ibdev, sg);
-		last_addr  = first_addr + dma_len;
-
-		end_aligned   = !(last_addr  & ~MASK_4K);
-
-		/* continue to collect page fragments till aligned or SG ends */
-		while (!end_aligned && (i + 1 < data->dma_nents)) {
-			sg = sg_next(sg);
-			i++;
+		start_addr = ib_sg_dma_address(ibdev, sg);
+		if (new_chunk)
+			chunk_start = start_addr;
 		dma_len = ib_sg_dma_len(ibdev, sg);
+		end_addr = start_addr + dma_len;
 		total_sz += dma_len;
-			last_addr = ib_sg_dma_address(ibdev, sg) + dma_len;
-			end_aligned = !(last_addr  & ~MASK_4K);
-		}
 
-		/* handle the 1st page in the 1st DMA element */
-		if (cur_page == 0) {
-			page = first_addr & MASK_4K;
-			page_vec->pages[cur_page] = page;
-			cur_page++;
+		/* collect page fragments until aligned or end of SG list */
+		if (!IS_4K_ALIGNED(end_addr) && i < last_ent) {
+			new_chunk = 0;
+			continue;
+		}
+		new_chunk = 1;
+
+		/* address of the first page in the contiguous chunk;
+		   masking relevant for the very first SG entry,
+		   which might be unaligned */
+		page = chunk_start & MASK_4K;
+		do {
+			page_vec->pages[cur_page++] = page;
 			page += SIZE_4K;
-		} else
-			page = first_addr;
-
-		for (; page < last_addr; page += SIZE_4K) {
-			page_vec->pages[cur_page] = page;
-			cur_page++;
+		} while (page < end_addr);
 	}
 
-	}
 	page_vec->data_size = total_sz;
 	iser_dbg("page_vec->data_size:%d cur_page %d\n", page_vec->data_size,cur_page);
 	return cur_page;
 }
 
-#define IS_4K_ALIGNED(addr)	((((unsigned long)addr) & ~MASK_4K) == 0)
 
 /**
  * iser_data_buf_aligned_len - Tries to determine the maximal correctly aligned
@@ -284,42 +276,40 @@ static int iser_sg_to_page_vec(struct iser_data_buf *data,
  * the number of entries which are aligned correctly. Supports the case where
  * consecutive SG elements are actually fragments of the same physcial page.
  */
-static unsigned int iser_data_buf_aligned_len(struct iser_data_buf *data,
+static int iser_data_buf_aligned_len(struct iser_data_buf *data,
 				      struct ib_device *ibdev)
 {
-	struct scatterlist *sgl, *sg;
-	u64 end_addr, next_addr;
-	int i, cnt;
-	unsigned int ret_len = 0;
+	struct scatterlist *sgl, *sg, *next_sg = NULL;
+	u64 start_addr, end_addr;
+	int i, ret_len, start_check = 0;
+
+	if (data->dma_nents == 1)
+		return 1;
 
 	sgl = (struct scatterlist *)data->buf;
+	start_addr  = ib_sg_dma_address(ibdev, sgl);
 
-	cnt = 0;
 	for_each_sg(sgl, sg, data->dma_nents, i) {
-		/* iser_dbg("Checking sg iobuf [%d]: phys=0x%08lX "
-		   "offset: %ld sz: %ld\n", i,
-		   (unsigned long)sg_phys(sg),
-		   (unsigned long)sg->offset,
-		   (unsigned long)sg->length); */
-		end_addr = ib_sg_dma_address(ibdev, sg) +
-			   ib_sg_dma_len(ibdev, sg);
-		/* iser_dbg("Checking sg iobuf end address "
-		       "0x%08lX\n", end_addr); */
-		if (i + 1 < data->dma_nents) {
-			next_addr = ib_sg_dma_address(ibdev, sg_next(sg));
-			/* are i, i+1 fragments of the same page? */
-			if (end_addr == next_addr) {
-				cnt++;
+		if (start_check && !IS_4K_ALIGNED(start_addr))
+			break;
+
+		next_sg = sg_next(sg);
+		if (!next_sg)
+			break;
+
+		end_addr    = start_addr + ib_sg_dma_len(ibdev, sg);
+		start_addr  = ib_sg_dma_address(ibdev, next_sg);
+
+		if (end_addr == start_addr) {
+			start_check = 0;
 			continue;
-			} else if (!IS_4K_ALIGNED(end_addr)) {
-				ret_len = cnt + 1;
+		} else
+			start_check = 1;
+
+		if (!IS_4K_ALIGNED(end_addr))
 			break;
 	}
-		}
-		cnt++;
-	}
-	if (i == data->dma_nents)
-		ret_len = cnt;	/* loop ended */
+	ret_len = (next_sg) ? i : i+1;
 	iser_dbg("Found %d aligned entries out of %d in sg:0x%p\n",
 		 ret_len, data->dma_nents, data);
 	return ret_len;

drivers/net/mlx4/fw.c

@@ -90,6 +90,7 @@ static void dump_dev_cap_flags(struct mlx4_dev *dev, u32 flags)
 		[ 9] = "Q_Key violation counter",
 		[10] = "VMM",
 		[12] = "DPDP",
+		[15] = "Big LSO headers",
 		[16] = "MW support",
 		[17] = "APM support",
 		[18] = "Atomic ops support",
@@ -235,7 +236,7 @@ int mlx4_QUERY_DEV_CAP(struct mlx4_dev *dev, struct mlx4_dev_cap *dev_cap)
 	MLX4_GET(field, outbox, QUERY_DEV_CAP_MAX_MPT_OFFSET);
 	dev_cap->max_mpts = 1 << (field & 0x3f);
 	MLX4_GET(field, outbox, QUERY_DEV_CAP_RSVD_EQ_OFFSET);
-	dev_cap->reserved_eqs = 1 << (field & 0xf);
+	dev_cap->reserved_eqs = field & 0xf;
 	MLX4_GET(field, outbox, QUERY_DEV_CAP_MAX_EQ_OFFSET);
 	dev_cap->max_eqs = 1 << (field & 0xf);
 	MLX4_GET(field, outbox, QUERY_DEV_CAP_RSVD_MTT_OFFSET);

include/linux/mlx4/device.h

@@ -61,6 +61,7 @@ enum {
 	MLX4_DEV_CAP_FLAG_BAD_PKEY_CNTR	= 1 <<  8,
 	MLX4_DEV_CAP_FLAG_BAD_QKEY_CNTR	= 1 <<  9,
 	MLX4_DEV_CAP_FLAG_DPDP		= 1 << 12,
+	MLX4_DEV_CAP_FLAG_BLH		= 1 << 15,
 	MLX4_DEV_CAP_FLAG_MEM_WINDOW	= 1 << 16,
 	MLX4_DEV_CAP_FLAG_APM		= 1 << 17,
 	MLX4_DEV_CAP_FLAG_ATOMIC	= 1 << 18,

include/rdma/ib_addr.h

@@ -36,6 +36,7 @@
 
 #include <linux/in.h>
 #include <linux/in6.h>
+#include <linux/if_arp.h>
 #include <linux/netdevice.h>
 #include <linux/socket.h>
 #include <rdma/ib_verbs.h>
@@ -60,8 +61,8 @@ struct rdma_dev_addr {
 	unsigned char src_dev_addr[MAX_ADDR_LEN];
 	unsigned char dst_dev_addr[MAX_ADDR_LEN];
 	unsigned char broadcast[MAX_ADDR_LEN];
-	enum rdma_node_type dev_type;
-	struct net_device *src_dev;
+	unsigned short dev_type;
+	int bound_dev_if;
 };
 
 /**
@@ -121,40 +122,29 @@ static inline void ib_addr_get_mgid(struct rdma_dev_addr *dev_addr,
 	memcpy(gid, dev_addr->broadcast + 4, sizeof *gid);
 }
 
-static inline void ib_addr_get_sgid(struct rdma_dev_addr *dev_addr,
-				    union ib_gid *gid)
+static inline int rdma_addr_gid_offset(struct rdma_dev_addr *dev_addr)
 {
-	memcpy(gid, dev_addr->src_dev_addr + 4, sizeof *gid);
+	return dev_addr->dev_type == ARPHRD_INFINIBAND ? 4 : 0;
 }
 
-static inline void ib_addr_set_sgid(struct rdma_dev_addr *dev_addr,
-				    union ib_gid *gid)
+static inline void rdma_addr_get_sgid(struct rdma_dev_addr *dev_addr, union ib_gid *gid)
 {
-	memcpy(dev_addr->src_dev_addr + 4, gid, sizeof *gid);
+	memcpy(gid, dev_addr->src_dev_addr + rdma_addr_gid_offset(dev_addr), sizeof *gid);
 }
 
-static inline void ib_addr_get_dgid(struct rdma_dev_addr *dev_addr,
-				    union ib_gid *gid)
+static inline void rdma_addr_set_sgid(struct rdma_dev_addr *dev_addr, union ib_gid *gid)
 {
-	memcpy(gid, dev_addr->dst_dev_addr + 4, sizeof *gid);
+	memcpy(dev_addr->src_dev_addr + rdma_addr_gid_offset(dev_addr), gid, sizeof *gid);
 }
 
-static inline void ib_addr_set_dgid(struct rdma_dev_addr *dev_addr,
-				    union ib_gid *gid)
+static inline void rdma_addr_get_dgid(struct rdma_dev_addr *dev_addr, union ib_gid *gid)
 {
-	memcpy(dev_addr->dst_dev_addr + 4, gid, sizeof *gid);
+	memcpy(gid, dev_addr->dst_dev_addr + rdma_addr_gid_offset(dev_addr), sizeof *gid);
 }
 
-static inline void iw_addr_get_sgid(struct rdma_dev_addr *dev_addr,
-				    union ib_gid *gid)
-{
-	memcpy(gid, dev_addr->src_dev_addr, sizeof *gid);
-}
-
-static inline void iw_addr_get_dgid(struct rdma_dev_addr *dev_addr,
-				    union ib_gid *gid)
+static inline void rdma_addr_set_dgid(struct rdma_dev_addr *dev_addr, union ib_gid *gid)
 {
-	memcpy(gid, dev_addr->dst_dev_addr, sizeof *gid);
+	memcpy(dev_addr->dst_dev_addr + rdma_addr_gid_offset(dev_addr), gid, sizeof *gid);
 }
 
 #endif /* IB_ADDR_H */

include/rdma/ib_sa.h

@@ -379,4 +379,10 @@ int ib_init_ah_from_path(struct ib_device *device, u8 port_num,
 			 struct ib_sa_path_rec *rec,
 			 struct ib_ah_attr *ah_attr);
 
+/**
+ * ib_sa_unpack_path - Convert a path record from MAD format to struct
+ * ib_sa_path_rec.
+ */
+void ib_sa_unpack_path(void *attribute, struct ib_sa_path_rec *rec);
+
 #endif /* IB_SA_H */

include/rdma/ib_user_sa.h

@@ -35,6 +35,22 @@
 
 #include <linux/types.h>
 
+enum {
+	IB_PATH_GMP		= 1,
+	IB_PATH_PRIMARY		= (1<<1),
+	IB_PATH_ALTERNATE	= (1<<2),
+	IB_PATH_OUTBOUND	= (1<<3),
+	IB_PATH_INBOUND		= (1<<4),
+	IB_PATH_INBOUND_REVERSE = (1<<5),
+	IB_PATH_BIDIRECTIONAL	= IB_PATH_OUTBOUND | IB_PATH_INBOUND_REVERSE
+};
+
+struct ib_path_rec_data {
+	__u32	flags;
+	__u32	reserved;
+	__u32	path_rec[16];
+};
+
 struct ib_user_path_rec {
 	__u8	dgid[16];
 	__u8	sgid[16];

include/rdma/ib_verbs.h

@@ -1425,6 +1425,11 @@ int ib_destroy_qp(struct ib_qp *qp);
  * @send_wr: A list of work requests to post on the send queue.
  * @bad_send_wr: On an immediate failure, this parameter will reference
  *   the work request that failed to be posted on the QP.
+ *
+ * While IBA Vol. 1 section 11.4.1.1 specifies that if an immediate
+ * error is returned, the QP state shall not be affected,
+ * ib_post_send() will return an immediate error after queueing any
+ * earlier work requests in the list.
  */
 static inline int ib_post_send(struct ib_qp *qp,
 			       struct ib_send_wr *send_wr,

include/rdma/rdma_user_cm.h

@@ -215,12 +215,14 @@ struct rdma_ucm_event_resp {
 
 /* Option levels */
 enum {
-	RDMA_OPTION_ID		= 0
+	RDMA_OPTION_ID		= 0,
+	RDMA_OPTION_IB		= 1
 };
 
 /* Option details */
 enum {
-	RDMA_OPTION_ID_TOS	= 0
+	RDMA_OPTION_ID_TOS	= 0,
+	RDMA_OPTION_IB_PATH	= 1
 };
 
 struct rdma_ucm_set_option {

net/rds/ib.c

@@ -182,8 +182,8 @@ static int rds_ib_conn_info_visitor(struct rds_connection *conn,
 		ic = conn->c_transport_data;
 		dev_addr = &ic->i_cm_id->route.addr.dev_addr;
 
-		ib_addr_get_sgid(dev_addr, (union ib_gid *) &iinfo->src_gid);
-		ib_addr_get_dgid(dev_addr, (union ib_gid *) &iinfo->dst_gid);
+		rdma_addr_get_sgid(dev_addr, (union ib_gid *) &iinfo->src_gid);
+		rdma_addr_get_dgid(dev_addr, (union ib_gid *) &iinfo->dst_gid);
 
 		rds_ibdev = ib_get_client_data(ic->i_cm_id->device, &rds_ib_client);
 		iinfo->max_send_wr = ic->i_send_ring.w_nr;

net/rds/iw.c

@@ -184,8 +184,8 @@ static int rds_iw_conn_info_visitor(struct rds_connection *conn,
 		ic = conn->c_transport_data;
 		dev_addr = &ic->i_cm_id->route.addr.dev_addr;
 
-		ib_addr_get_sgid(dev_addr, (union ib_gid *) &iinfo->src_gid);
-		ib_addr_get_dgid(dev_addr, (union ib_gid *) &iinfo->dst_gid);
+		rdma_addr_get_sgid(dev_addr, (union ib_gid *) &iinfo->src_gid);
+		rdma_addr_get_dgid(dev_addr, (union ib_gid *) &iinfo->dst_gid);
 
 		rds_iwdev = ib_get_client_data(ic->i_cm_id->device, &rds_iw_client);
 		iinfo->max_send_wr = ic->i_send_ring.w_nr;