20 files changed, 166 insertions, 168 deletions

diff --git a/drivers/infiniband/hw/amso1100/c2_ae.c b/drivers/infiniband/hw/amso1100/c2_ae.c
index 32d34e88d5cf..706cf97cbe8f 100644
--- a/drivers/infiniband/hw/amso1100/c2_ae.c
+++ b/drivers/infiniband/hw/amso1100/c2_ae.c
@@ -311,6 +311,7 @@ void c2_ae_event(struct c2_dev *c2dev, u32 mq_index)
 		if (cq->ibcq.event_handler)
 			cq->ibcq.event_handler(&ib_event,
 					       cq->ibcq.cq_context);
+		break;
 	}
 
 	default:
diff --git a/drivers/infiniband/hw/cxgb3/iwch_cm.c b/drivers/infiniband/hw/cxgb3/iwch_cm.c
index aaf88ef9409c..3e094cd6a0e3 100644
--- a/drivers/infiniband/hw/cxgb3/iwch_cm.c
+++ b/drivers/infiniband/hw/cxgb3/iwch_cm.c
@@ -128,9 +128,8 @@ static void stop_ep_timer(struct iwch_ep *ep)
 {
 	PDBG("%s ep %p\n", __func__, ep);
 	if (!timer_pending(&ep->timer)) {
-		printk(KERN_ERR "%s timer stopped when its not running!  ep %p state %u\n",
+		WARN(1, "%s timer stopped when its not running!  ep %p state %u\n",
 			__func__, ep, ep->com.state);
-		WARN_ON(1);
 		return;
 	}
 	del_timer_sync(&ep->timer);
@@ -1756,9 +1755,8 @@ static void ep_timeout(unsigned long arg)
 		__state_set(&ep->com, ABORTING);
 		break;
 	default:
-		printk(KERN_ERR "%s unexpected state ep %p state %u\n",
+		WARN(1, "%s unexpected state ep %p state %u\n",
 			__func__, ep, ep->com.state);
-		WARN_ON(1);
 		abort = 0;
 	}
 	spin_unlock_irqrestore(&ep->com.lock, flags);
diff --git a/drivers/infiniband/hw/cxgb4/cm.c b/drivers/infiniband/hw/cxgb4/cm.c
index 6cfd4d8fd0bd..5de86968379d 100644
--- a/drivers/infiniband/hw/cxgb4/cm.c
+++ b/drivers/infiniband/hw/cxgb4/cm.c
@@ -151,9 +151,8 @@ static void stop_ep_timer(struct c4iw_ep *ep)
 {
 	PDBG("%s ep %p\n", __func__, ep);
 	if (!timer_pending(&ep->timer)) {
-		printk(KERN_ERR "%s timer stopped when its not running! "
+		WARN(1, "%s timer stopped when its not running! "
 		       "ep %p state %u\n", __func__, ep, ep->com.state);
-		WARN_ON(1);
 		return;
 	}
 	del_timer_sync(&ep->timer);
@@ -2551,9 +2550,8 @@ static void process_timeout(struct c4iw_ep *ep)
 		__state_set(&ep->com, ABORTING);
 		break;
 	default:
-		printk(KERN_ERR "%s unexpected state ep %p tid %u state %u\n",
+		WARN(1, "%s unexpected state ep %p tid %u state %u\n",
 			__func__, ep, ep->hwtid, ep->com.state);
-		WARN_ON(1);
 		abort = 0;
 	}
 	mutex_unlock(&ep->com.mutex);
diff --git a/drivers/infiniband/hw/cxgb4/mem.c b/drivers/infiniband/hw/cxgb4/mem.c
index 57e07c61ace2..afd81790ab3c 100644
--- a/drivers/infiniband/hw/cxgb4/mem.c
+++ b/drivers/infiniband/hw/cxgb4/mem.c
@@ -468,7 +468,7 @@ struct ib_mr *c4iw_register_phys_mem(struct ib_pd *pd,
 	ret = alloc_pbl(mhp, npages);
 	if (ret) {
 		kfree(page_list);
-		goto err_pbl;
+		goto err;
 	}
 
 	ret = write_pbl(&mhp->rhp->rdev, page_list, mhp->attr.pbl_addr,
diff --git a/drivers/infiniband/hw/ehca/hcp_if.c b/drivers/infiniband/hw/ehca/hcp_if.c
index 2d41d04fd959..89517ffb4389 100644
--- a/drivers/infiniband/hw/ehca/hcp_if.c
+++ b/drivers/infiniband/hw/ehca/hcp_if.c
@@ -90,26 +90,6 @@
 
 static DEFINE_SPINLOCK(hcall_lock);
 
-static u32 get_longbusy_msecs(int longbusy_rc)
-{
-	switch (longbusy_rc) {
-	case H_LONG_BUSY_ORDER_1_MSEC:
-		return 1;
-	case H_LONG_BUSY_ORDER_10_MSEC:
-		return 10;
-	case H_LONG_BUSY_ORDER_100_MSEC:
-		return 100;
-	case H_LONG_BUSY_ORDER_1_SEC:
-		return 1000;
-	case H_LONG_BUSY_ORDER_10_SEC:
-		return 10000;
-	case H_LONG_BUSY_ORDER_100_SEC:
-		return 100000;
-	default:
-		return 1;
-	}
-}
-
 static long ehca_plpar_hcall_norets(unsigned long opcode,
 				    unsigned long arg1,
 				    unsigned long arg2,
diff --git a/drivers/infiniband/hw/ipath/ipath_init_chip.c b/drivers/infiniband/hw/ipath/ipath_init_chip.c
index 49b09c697c7c..be2a60e142b0 100644
--- a/drivers/infiniband/hw/ipath/ipath_init_chip.c
+++ b/drivers/infiniband/hw/ipath/ipath_init_chip.c
@@ -719,16 +719,6 @@ int ipath_init_chip(struct ipath_devdata *dd, int reinit)
 		goto done;
 
 	/*
-	 * we ignore most issues after reporting them, but have to specially
-	 * handle hardware-disabled chips.
-	 */
-	if (ret == 2) {
-		/* unique error, known to ipath_init_one */
-		ret = -EPERM;
-		goto done;
-	}
-
-	/*
 	 * We could bump this to allow for full rcvegrcnt + rcvtidcnt,
 	 * but then it no longer nicely fits power of two, and since
 	 * we now use routines that backend onto __get_free_pages, the
diff --git a/drivers/infiniband/hw/mlx4/alias_GUID.c b/drivers/infiniband/hw/mlx4/alias_GUID.c
index d2fb38d43571..2f215b93db6b 100644
--- a/drivers/infiniband/hw/mlx4/alias_GUID.c
+++ b/drivers/infiniband/hw/mlx4/alias_GUID.c
@@ -107,7 +107,7 @@ static __be64 get_cached_alias_guid(struct mlx4_ib_dev *dev, int port, int index
 {
 	if (index >= NUM_ALIAS_GUID_PER_PORT) {
 		pr_err("%s: ERROR: asked for index:%d\n", __func__, index);
-		return  (__force __be64) ((u64) 0xFFFFFFFFFFFFFFFFUL);
+		return (__force __be64) -1;
 	}
 	return *(__be64 *)&dev->sriov.demux[port - 1].guid_cache[index];
 }
diff --git a/drivers/infiniband/hw/mlx4/cm.c b/drivers/infiniband/hw/mlx4/cm.c
index 80079e5a2e30..dbc99d41605c 100644
--- a/drivers/infiniband/hw/mlx4/cm.c
+++ b/drivers/infiniband/hw/mlx4/cm.c
@@ -268,15 +268,15 @@ static void schedule_delayed(struct ib_device *ibdev, struct id_map_entry *id)
 	struct mlx4_ib_sriov *sriov = &to_mdev(ibdev)->sriov;
 	unsigned long flags;
 
-	spin_lock_irqsave(&sriov->going_down_lock, flags);
 	spin_lock(&sriov->id_map_lock);
+	spin_lock_irqsave(&sriov->going_down_lock, flags);
 	/*make sure that there is no schedule inside the scheduled work.*/
 	if (!sriov->is_going_down) {
 		id->scheduled_delete = 1;
 		schedule_delayed_work(&id->timeout, CM_CLEANUP_CACHE_TIMEOUT);
 	}
-	spin_unlock(&sriov->id_map_lock);
 	spin_unlock_irqrestore(&sriov->going_down_lock, flags);
+	spin_unlock(&sriov->id_map_lock);
 }
 
 int mlx4_ib_multiplex_cm_handler(struct ib_device *ibdev, int port, int slave_id,
diff --git a/drivers/infiniband/hw/mlx4/cq.c b/drivers/infiniband/hw/mlx4/cq.c
index c9eb6a6815ce..ae67df35dd4d 100644
--- a/drivers/infiniband/hw/mlx4/cq.c
+++ b/drivers/infiniband/hw/mlx4/cq.c
@@ -66,7 +66,7 @@ static void mlx4_ib_cq_event(struct mlx4_cq *cq, enum mlx4_event type)
 
 static void *get_cqe_from_buf(struct mlx4_ib_cq_buf *buf, int n)
 {
-	return mlx4_buf_offset(&buf->buf, n * sizeof (struct mlx4_cqe));
+	return mlx4_buf_offset(&buf->buf, n * buf->entry_size);
 }
 
 static void *get_cqe(struct mlx4_ib_cq *cq, int n)
@@ -77,8 +77,9 @@ static void *get_cqe(struct mlx4_ib_cq *cq, int n)
 static void *get_sw_cqe(struct mlx4_ib_cq *cq, int n)
 {
 	struct mlx4_cqe *cqe = get_cqe(cq, n & cq->ibcq.cqe);
+	struct mlx4_cqe *tcqe = ((cq->buf.entry_size == 64) ? (cqe + 1) : cqe);
 
-	return (!!(cqe->owner_sr_opcode & MLX4_CQE_OWNER_MASK) ^
+	return (!!(tcqe->owner_sr_opcode & MLX4_CQE_OWNER_MASK) ^
 		!!(n & (cq->ibcq.cqe + 1))) ? NULL : cqe;
 }
 
@@ -99,12 +100,13 @@ static int mlx4_ib_alloc_cq_buf(struct mlx4_ib_dev *dev, struct mlx4_ib_cq_buf *
 {
 	int err;
 
-	err = mlx4_buf_alloc(dev->dev, nent * sizeof(struct mlx4_cqe),
+	err = mlx4_buf_alloc(dev->dev, nent * dev->dev->caps.cqe_size,
 			     PAGE_SIZE * 2, &buf->buf);
 
 	if (err)
 		goto out;
 
+	buf->entry_size = dev->dev->caps.cqe_size;
 	err = mlx4_mtt_init(dev->dev, buf->buf.npages, buf->buf.page_shift,
 				    &buf->mtt);
 	if (err)
@@ -120,8 +122,7 @@ err_mtt:
 	mlx4_mtt_cleanup(dev->dev, &buf->mtt);
 
 err_buf:
-	mlx4_buf_free(dev->dev, nent * sizeof(struct mlx4_cqe),
-			      &buf->buf);
+	mlx4_buf_free(dev->dev, nent * buf->entry_size, &buf->buf);
 
 out:
 	return err;
@@ -129,7 +130,7 @@ out:
 
 static void mlx4_ib_free_cq_buf(struct mlx4_ib_dev *dev, struct mlx4_ib_cq_buf *buf, int cqe)
 {
-	mlx4_buf_free(dev->dev, (cqe + 1) * sizeof(struct mlx4_cqe), &buf->buf);
+	mlx4_buf_free(dev->dev, (cqe + 1) * buf->entry_size, &buf->buf);
 }
 
 static int mlx4_ib_get_cq_umem(struct mlx4_ib_dev *dev, struct ib_ucontext *context,
@@ -137,8 +138,9 @@ static int mlx4_ib_get_cq_umem(struct mlx4_ib_dev *dev, struct ib_ucontext *cont
 			       u64 buf_addr, int cqe)
 {
 	int err;
+	int cqe_size = dev->dev->caps.cqe_size;
 
-	*umem = ib_umem_get(context, buf_addr, cqe * sizeof (struct mlx4_cqe),
+	*umem = ib_umem_get(context, buf_addr, cqe * cqe_size,
 			    IB_ACCESS_LOCAL_WRITE, 1);
 	if (IS_ERR(*umem))
 		return PTR_ERR(*umem);
@@ -331,16 +333,23 @@ static void mlx4_ib_cq_resize_copy_cqes(struct mlx4_ib_cq *cq)
 {
 	struct mlx4_cqe *cqe, *new_cqe;
 	int i;
+	int cqe_size = cq->buf.entry_size;
+	int cqe_inc = cqe_size == 64 ? 1 : 0;
 
 	i = cq->mcq.cons_index;
 	cqe = get_cqe(cq, i & cq->ibcq.cqe);
+	cqe += cqe_inc;
+
 	while ((cqe->owner_sr_opcode & MLX4_CQE_OPCODE_MASK) != MLX4_CQE_OPCODE_RESIZE) {
 		new_cqe = get_cqe_from_buf(&cq->resize_buf->buf,
 					   (i + 1) & cq->resize_buf->cqe);
-		memcpy(new_cqe, get_cqe(cq, i & cq->ibcq.cqe), sizeof(struct mlx4_cqe));
+		memcpy(new_cqe, get_cqe(cq, i & cq->ibcq.cqe), cqe_size);
+		new_cqe += cqe_inc;
+
 		new_cqe->owner_sr_opcode = (cqe->owner_sr_opcode & ~MLX4_CQE_OWNER_MASK) |
 			(((i + 1) & (cq->resize_buf->cqe + 1)) ? MLX4_CQE_OWNER_MASK : 0);
 		cqe = get_cqe(cq, ++i & cq->ibcq.cqe);
+		cqe += cqe_inc;
 	}
 	++cq->mcq.cons_index;
 }
@@ -438,6 +447,7 @@ err_buf:
 
 out:
 	mutex_unlock(&cq->resize_mutex);
+
 	return err;
 }
 
@@ -586,6 +596,9 @@ repoll:
 	if (!cqe)
 		return -EAGAIN;
 
+	if (cq->buf.entry_size == 64)
+		cqe++;
+
 	++cq->mcq.cons_index;
 
 	/*
@@ -807,6 +820,7 @@ void __mlx4_ib_cq_clean(struct mlx4_ib_cq *cq, u32 qpn, struct mlx4_ib_srq *srq)
 	int nfreed = 0;
 	struct mlx4_cqe *cqe, *dest;
 	u8 owner_bit;
+	int cqe_inc = cq->buf.entry_size == 64 ? 1 : 0;
 
 	/*
 	 * First we need to find the current producer index, so we
@@ -825,12 +839,16 @@ void __mlx4_ib_cq_clean(struct mlx4_ib_cq *cq, u32 qpn, struct mlx4_ib_srq *srq)
 	 */
 	while ((int) --prod_index - (int) cq->mcq.cons_index >= 0) {
 		cqe = get_cqe(cq, prod_index & cq->ibcq.cqe);
+		cqe += cqe_inc;
+
 		if ((be32_to_cpu(cqe->vlan_my_qpn) & MLX4_CQE_QPN_MASK) == qpn) {
 			if (srq && !(cqe->owner_sr_opcode & MLX4_CQE_IS_SEND_MASK))
 				mlx4_ib_free_srq_wqe(srq, be16_to_cpu(cqe->wqe_index));
 			++nfreed;
 		} else if (nfreed) {
 			dest = get_cqe(cq, (prod_index + nfreed) & cq->ibcq.cqe);
+			dest += cqe_inc;
+
 			owner_bit = dest->owner_sr_opcode & MLX4_CQE_OWNER_MASK;
 			memcpy(dest, cqe, sizeof *cqe);
 			dest->owner_sr_opcode = owner_bit |
diff --git a/drivers/infiniband/hw/mlx4/mad.c b/drivers/infiniband/hw/mlx4/mad.c
index 21a794152d15..0a903c129f0a 100644
--- a/drivers/infiniband/hw/mlx4/mad.c
+++ b/drivers/infiniband/hw/mlx4/mad.c
@@ -409,38 +409,45 @@ int mlx4_ib_find_real_gid(struct ib_device *ibdev, u8 port, __be64 guid)
 }
 
 
-static int get_pkey_phys_indices(struct mlx4_ib_dev *ibdev, u8 port, u8 ph_pkey_ix,
-				 u8 *full_pk_ix, u8 *partial_pk_ix,
-				 int *is_full_member)
+static int find_slave_port_pkey_ix(struct mlx4_ib_dev *dev, int slave,
+				   u8 port, u16 pkey, u16 *ix)
 {
-	u16 search_pkey;
-	int fm;
-	int err = 0;
-	u16 pk;
+	int i, ret;
+	u8 unassigned_pkey_ix, pkey_ix, partial_ix = 0xFF;
+	u16 slot_pkey;
 
-	err = ib_get_cached_pkey(&ibdev->ib_dev, port, ph_pkey_ix, &search_pkey);
-	if (err)
-		return err;
+	if (slave == mlx4_master_func_num(dev->dev))
+		return ib_find_cached_pkey(&dev->ib_dev, port, pkey, ix);
 
-	fm = (search_pkey & 0x8000) ? 1 : 0;
-	if (fm) {
-		*full_pk_ix = ph_pkey_ix;
-		search_pkey &= 0x7FFF;
-	} else {
-		*partial_pk_ix = ph_pkey_ix;
-		search_pkey |= 0x8000;
-	}
+	unassigned_pkey_ix = dev->dev->phys_caps.pkey_phys_table_len[port] - 1;
 
-	if (ib_find_exact_cached_pkey(&ibdev->ib_dev, port, search_pkey, &pk))
-		pk = 0xFFFF;
+	for (i = 0; i < dev->dev->caps.pkey_table_len[port]; i++) {
+		if (dev->pkeys.virt2phys_pkey[slave][port - 1][i] == unassigned_pkey_ix)
+			continue;
 
-	if (fm)
-		*partial_pk_ix = (pk & 0xFF);
-	else
-		*full_pk_ix = (pk & 0xFF);
+		pkey_ix = dev->pkeys.virt2phys_pkey[slave][port - 1][i];
 
-	*is_full_member = fm;
-	return err;
+		ret = ib_get_cached_pkey(&dev->ib_dev, port, pkey_ix, &slot_pkey);
+		if (ret)
+			continue;
+		if ((slot_pkey & 0x7FFF) == (pkey & 0x7FFF)) {
+			if (slot_pkey & 0x8000) {
+				*ix = (u16) pkey_ix;
+				return 0;
+			} else {
+				/* take first partial pkey index found */
+				if (partial_ix == 0xFF)
+					partial_ix = pkey_ix;
+			}
+		}
+	}
+
+	if (partial_ix < 0xFF) {
+		*ix = (u16) partial_ix;
+		return 0;
+	}
+
+	return -EINVAL;
 }
 
 int mlx4_ib_send_to_slave(struct mlx4_ib_dev *dev, int slave, u8 port,
@@ -458,10 +465,8 @@ int mlx4_ib_send_to_slave(struct mlx4_ib_dev *dev, int slave, u8 port,
 	unsigned tun_tx_ix = 0;
 	int dqpn;
 	int ret = 0;
-	int i;
-	int is_full_member = 0;
 	u16 tun_pkey_ix;
-	u8 ph_pkey_ix, full_pk_ix = 0, partial_pk_ix = 0;
+	u16 cached_pkey;
 
 	if (dest_qpt > IB_QPT_GSI)
 		return -EINVAL;
@@ -481,27 +486,17 @@ int mlx4_ib_send_to_slave(struct mlx4_ib_dev *dev, int slave, u8 port,
 	else
 		tun_qp = &tun_ctx->qp[1];
 
-	/* compute pkey index for slave */
-	/* get physical pkey -- virtualized Dom0 pkey to phys*/
+	/* compute P_Key index to put in tunnel header for slave */
 	if (dest_qpt) {
-		ph_pkey_ix =
-			dev->pkeys.virt2phys_pkey[mlx4_master_func_num(dev->dev)][port - 1][wc->pkey_index];
-
-		/* now, translate this to the slave pkey index */
-		ret = get_pkey_phys_indices(dev, port, ph_pkey_ix, &full_pk_ix,
-					    &partial_pk_ix, &is_full_member);
+		u16 pkey_ix;
+		ret = ib_get_cached_pkey(&dev->ib_dev, port, wc->pkey_index, &cached_pkey);
 		if (ret)
 			return -EINVAL;
 
-		for (i = 0; i < dev->dev->caps.pkey_table_len[port]; i++) {
-			if ((dev->pkeys.virt2phys_pkey[slave][port - 1][i] == full_pk_ix) ||
-			    (is_full_member &&
-			     (dev->pkeys.virt2phys_pkey[slave][port - 1][i] == partial_pk_ix)))
-				break;
-		}
-		if (i == dev->dev->caps.pkey_table_len[port])
+		ret = find_slave_port_pkey_ix(dev, slave, port, cached_pkey, &pkey_ix);
+		if (ret)
 			return -EINVAL;
-		tun_pkey_ix = i;
+		tun_pkey_ix = pkey_ix;
 	} else
 		tun_pkey_ix = dev->pkeys.virt2phys_pkey[slave][port - 1][0];
 
diff --git a/drivers/infiniband/hw/mlx4/main.c b/drivers/infiniband/hw/mlx4/main.c
index 718ec6b2bad2..e7d81c0d1ac5 100644
--- a/drivers/infiniband/hw/mlx4/main.c
+++ b/drivers/infiniband/hw/mlx4/main.c
@@ -563,15 +563,24 @@ static struct ib_ucontext *mlx4_ib_alloc_ucontext(struct ib_device *ibdev,
 {
 	struct mlx4_ib_dev *dev = to_mdev(ibdev);
 	struct mlx4_ib_ucontext *context;
+	struct mlx4_ib_alloc_ucontext_resp_v3 resp_v3;
 	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;
+	if (ibdev->uverbs_abi_ver == MLX4_IB_UVERBS_NO_DEV_CAPS_ABI_VERSION) {
+		resp_v3.qp_tab_size      = dev->dev->caps.num_qps;
+		resp_v3.bf_reg_size      = dev->dev->caps.bf_reg_size;
+		resp_v3.bf_regs_per_page = dev->dev->caps.bf_regs_per_page;
+	} else {
+		resp.dev_caps	      = dev->dev->caps.userspace_caps;
+		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;
+		resp.cqe_size	      = dev->dev->caps.cqe_size;
+	}
 
 	context = kmalloc(sizeof *context, GFP_KERNEL);
 	if (!context)
@@ -586,7 +595,11 @@ static struct ib_ucontext *mlx4_ib_alloc_ucontext(struct ib_device *ibdev,
 	INIT_LIST_HEAD(&context->db_page_list);
 	mutex_init(&context->db_page_mutex);
 
-	err = ib_copy_to_udata(udata, &resp, sizeof resp);
+	if (ibdev->uverbs_abi_ver == MLX4_IB_UVERBS_NO_DEV_CAPS_ABI_VERSION)
+		err = ib_copy_to_udata(udata, &resp_v3, sizeof(resp_v3));
+	else
+		err = ib_copy_to_udata(udata, &resp, sizeof(resp));
+
 	if (err) {
 		mlx4_uar_free(to_mdev(ibdev)->dev, &context->uar);
 		kfree(context);
@@ -1342,7 +1355,11 @@ static void *mlx4_ib_add(struct mlx4_dev *dev)
 	ibdev->ib_dev.num_comp_vectors	= dev->caps.num_comp_vectors;
 	ibdev->ib_dev.dma_device	= &dev->pdev->dev;
 
-	ibdev->ib_dev.uverbs_abi_ver	= MLX4_IB_UVERBS_ABI_VERSION;
+	if (dev->caps.userspace_caps)
+		ibdev->ib_dev.uverbs_abi_ver = MLX4_IB_UVERBS_ABI_VERSION;
+	else
+		ibdev->ib_dev.uverbs_abi_ver = MLX4_IB_UVERBS_NO_DEV_CAPS_ABI_VERSION;
+
 	ibdev->ib_dev.uverbs_cmd_mask	=
 		(1ull << IB_USER_VERBS_CMD_GET_CONTEXT)		|
 		(1ull << IB_USER_VERBS_CMD_QUERY_DEVICE)	|
diff --git a/drivers/infiniband/hw/mlx4/mcg.c b/drivers/infiniband/hw/mlx4/mcg.c
index 3c3b54c3fdd9..25b2cdff00f8 100644
--- a/drivers/infiniband/hw/mlx4/mcg.c
+++ b/drivers/infiniband/hw/mlx4/mcg.c
@@ -233,7 +233,8 @@ static int send_mad_to_slave(int slave, struct mlx4_ib_demux_ctx *ctx,
 
 	ib_query_ah(dev->sm_ah[ctx->port - 1], &ah_attr);
 
-	wc.pkey_index = 0;
+	if (ib_find_cached_pkey(&dev->ib_dev, ctx->port, IB_DEFAULT_PKEY_FULL, &wc.pkey_index))
+		return -EINVAL;
 	wc.sl = 0;
 	wc.dlid_path_bits = 0;
 	wc.port_num = ctx->port;
@@ -1074,10 +1075,6 @@ static void _mlx4_ib_mcg_port_cleanup(struct mlx4_ib_demux_ctx *ctx, int destroy
 	unsigned long end;
 	int count;
 
-	if (ctx->flushing)
-		return;
-
-	ctx->flushing = 1;
 	for (i = 0; i < MAX_VFS; ++i)
 		clean_vf_mcast(ctx, i);
 
@@ -1107,9 +1104,6 @@ static void _mlx4_ib_mcg_port_cleanup(struct mlx4_ib_demux_ctx *ctx, int destroy
 		force_clean_group(group);
 	}
 	mutex_unlock(&ctx->mcg_table_lock);
-
-	if (!destroy_wq)
-		ctx->flushing = 0;
 }
 
 struct clean_work {
@@ -1123,6 +1117,7 @@ static void mcg_clean_task(struct work_struct *work)
 	struct clean_work *cw = container_of(work, struct clean_work, work);
 
 	_mlx4_ib_mcg_port_cleanup(cw->ctx, cw->destroy_wq);
+	cw->ctx->flushing = 0;
 	kfree(cw);
 }
 
@@ -1130,13 +1125,20 @@ void mlx4_ib_mcg_port_cleanup(struct mlx4_ib_demux_ctx *ctx, int destroy_wq)
 {
 	struct clean_work *work;
 
+	if (ctx->flushing)
+		return;
+
+	ctx->flushing = 1;
+
 	if (destroy_wq) {
 		_mlx4_ib_mcg_port_cleanup(ctx, destroy_wq);
+		ctx->flushing = 0;
 		return;
 	}
 
 	work = kmalloc(sizeof *work, GFP_KERNEL);
 	if (!work) {
+		ctx->flushing = 0;
 		mcg_warn("failed allocating work for cleanup\n");
 		return;
 	}
diff --git a/drivers/infiniband/hw/mlx4/mlx4_ib.h b/drivers/infiniband/hw/mlx4/mlx4_ib.h
index e04cbc9a54a5..dcd845bc30f0 100644
--- a/drivers/infiniband/hw/mlx4/mlx4_ib.h
+++ b/drivers/infiniband/hw/mlx4/mlx4_ib.h
@@ -90,6 +90,7 @@ struct mlx4_ib_xrcd {
 struct mlx4_ib_cq_buf {
 	struct mlx4_buf		buf;
 	struct mlx4_mtt		mtt;
+	int			entry_size;
 };
 
 struct mlx4_ib_cq_resize {
diff --git a/drivers/infiniband/hw/mlx4/user.h b/drivers/infiniband/hw/mlx4/user.h
index 13beedeeef9f..07e6769ef43b 100644
--- a/drivers/infiniband/hw/mlx4/user.h
+++ b/drivers/infiniband/hw/mlx4/user.h
@@ -40,7 +40,9 @@
  * Increment this value if any changes that break userspace ABI
  * compatibility are made.
  */
-#define MLX4_IB_UVERBS_ABI_VERSION	3
+
+#define MLX4_IB_UVERBS_NO_DEV_CAPS_ABI_VERSION	3
+#define MLX4_IB_UVERBS_ABI_VERSION		4
 
 /*
  * Make sure that all structs defined in this file remain laid out so
@@ -50,10 +52,18 @@
  * instead.
  */
 
+struct mlx4_ib_alloc_ucontext_resp_v3 {
+	__u32	qp_tab_size;
+	__u16	bf_reg_size;
+	__u16	bf_regs_per_page;
+};
+
 struct mlx4_ib_alloc_ucontext_resp {
+	__u32	dev_caps;
 	__u32	qp_tab_size;
 	__u16	bf_reg_size;
 	__u16	bf_regs_per_page;
+	__u32	cqe_size;
 };
 
 struct mlx4_ib_alloc_pd_resp {
diff --git a/drivers/infiniband/hw/nes/nes.h b/drivers/infiniband/hw/nes/nes.h
index 5cac29e6bc1c..33cc58941a3e 100644
--- a/drivers/infiniband/hw/nes/nes.h
+++ b/drivers/infiniband/hw/nes/nes.h
@@ -532,6 +532,7 @@ void nes_iwarp_ce_handler(struct nes_device *, struct nes_hw_cq *);
 int nes_destroy_cqp(struct nes_device *);
 int nes_nic_cm_xmit(struct sk_buff *, struct net_device *);
 void nes_recheck_link_status(struct work_struct *work);
+void nes_terminate_timeout(unsigned long context);
 
 /* nes_nic.c */
 struct net_device *nes_netdev_init(struct nes_device *, void __iomem *);
diff --git a/drivers/infiniband/hw/nes/nes_cm.c b/drivers/infiniband/hw/nes/nes_cm.c
index cfaacaf6bf5f..22ea67eea5dc 100644
--- a/drivers/infiniband/hw/nes/nes_cm.c
+++ b/drivers/infiniband/hw/nes/nes_cm.c
@@ -629,11 +629,9 @@ static void build_rdma0_msg(struct nes_cm_node *cm_node, struct nes_qp **nesqp_a
 
 	case SEND_RDMA_READ_ZERO:
 	default:
-		if (cm_node->send_rdma0_op != SEND_RDMA_READ_ZERO) {
-			printk(KERN_ERR "%s[%u]: Unsupported RDMA0 len operation=%u\n",
-				 __func__, __LINE__, cm_node->send_rdma0_op);
-			WARN_ON(1);
-		}
+		if (cm_node->send_rdma0_op != SEND_RDMA_READ_ZERO)
+			WARN(1, "Unsupported RDMA0 len operation=%u\n",
+			     cm_node->send_rdma0_op);
 		nes_debug(NES_DBG_CM, "Sending first rdma operation.\n");
 		wqe->wqe_words[NES_IWARP_SQ_WQE_MISC_IDX] =
 			cpu_to_le32(NES_IWARP_SQ_OP_RDMAR);
@@ -671,7 +669,6 @@ int schedule_nes_timer(struct nes_cm_node *cm_node, struct sk_buff *skb,
 	struct nes_cm_core *cm_core = cm_node->cm_core;
 	struct nes_timer_entry *new_send;
 	int ret = 0;
-	u32 was_timer_set;
 
 	new_send = kzalloc(sizeof(*new_send), GFP_ATOMIC);
 	if (!new_send)
@@ -723,12 +720,8 @@ int schedule_nes_timer(struct nes_cm_node *cm_node, struct sk_buff *skb,
 		}
 	}
 
-	was_timer_set = timer_pending(&cm_core->tcp_timer);
-
-	if (!was_timer_set) {
-		cm_core->tcp_timer.expires = new_send->timetosend;
-		add_timer(&cm_core->tcp_timer);
-	}
+	if (!timer_pending(&cm_core->tcp_timer))
+		mod_timer(&cm_core->tcp_timer, new_send->timetosend);
 
 	return ret;
 }
@@ -946,10 +939,8 @@ static void nes_cm_timer_tick(unsigned long pass)
 	}
 
 	if (settimer) {
-		if (!timer_pending(&cm_core->tcp_timer)) {
-			cm_core->tcp_timer.expires = nexttimeout;
-			add_timer(&cm_core->tcp_timer);
-		}
+		if (!timer_pending(&cm_core->tcp_timer))
+			mod_timer(&cm_core->tcp_timer, nexttimeout);
 	}
 }
 
@@ -1314,8 +1305,6 @@ static int mini_cm_del_listen(struct nes_cm_core *cm_core,
 static inline int mini_cm_accelerated(struct nes_cm_core *cm_core,
 				      struct nes_cm_node *cm_node)
 {
-	u32 was_timer_set;
-
 	cm_node->accelerated = 1;
 
 	if (cm_node->accept_pend) {
@@ -1325,11 +1314,8 @@ static inline int mini_cm_accelerated(struct nes_cm_core *cm_core,
 		BUG_ON(atomic_read(&cm_node->listener->pend_accepts_cnt) < 0);
 	}
 
-	was_timer_set = timer_pending(&cm_core->tcp_timer);
-	if (!was_timer_set) {
-		cm_core->tcp_timer.expires = jiffies + NES_SHORT_TIME;
-		add_timer(&cm_core->tcp_timer);
-	}
+	if (!timer_pending(&cm_core->tcp_timer))
+		mod_timer(&cm_core->tcp_timer, (jiffies + NES_SHORT_TIME));
 
 	return 0;
 }
diff --git a/drivers/infiniband/hw/nes/nes_hw.c b/drivers/infiniband/hw/nes/nes_hw.c
index fe7965ee4096..67647e264611 100644
--- a/drivers/infiniband/hw/nes/nes_hw.c
+++ b/drivers/infiniband/hw/nes/nes_hw.c
@@ -75,7 +75,6 @@ static void nes_process_iwarp_aeqe(struct nes_device *nesdev,
 static void process_critical_error(struct nes_device *nesdev);
 static void nes_process_mac_intr(struct nes_device *nesdev, u32 mac_number);
 static unsigned int nes_reset_adapter_ne020(struct nes_device *nesdev, u8 *OneG_Mode);
-static void nes_terminate_timeout(unsigned long context);
 static void nes_terminate_start_timer(struct nes_qp *nesqp);
 
 #ifdef CONFIG_INFINIBAND_NES_DEBUG
@@ -3520,7 +3519,7 @@ static void nes_terminate_received(struct nes_device *nesdev,
 }
 
 /* Timeout routine in case terminate fails to complete */
-static void nes_terminate_timeout(unsigned long context)
+void nes_terminate_timeout(unsigned long context)
 {
 	struct nes_qp *nesqp = (struct nes_qp *)(unsigned long)context;
 
@@ -3530,11 +3529,7 @@ static void nes_terminate_timeout(unsigned long context)
 /* Set a timer in case hw cannot complete the terminate sequence */
 static void nes_terminate_start_timer(struct nes_qp *nesqp)
 {
-	init_timer(&nesqp->terminate_timer);
-	nesqp->terminate_timer.function = nes_terminate_timeout;
-	nesqp->terminate_timer.expires = jiffies + HZ;
-	nesqp->terminate_timer.data = (unsigned long)nesqp;
-	add_timer(&nesqp->terminate_timer);
+	mod_timer(&nesqp->terminate_timer, (jiffies + HZ));
 }
 
 /**
diff --git a/drivers/infiniband/hw/nes/nes_mgt.c b/drivers/infiniband/hw/nes/nes_mgt.c
index 3ba7be369452..416645259b0f 100644
--- a/drivers/infiniband/hw/nes/nes_mgt.c
+++ b/drivers/infiniband/hw/nes/nes_mgt.c
@@ -210,6 +210,9 @@ static struct sk_buff *nes_get_next_skb(struct nes_device *nesdev, struct nes_qp
 	}
 
 	while (1) {
+		if (skb_queue_empty(&nesqp->pau_list))
+			goto out;
+
 		seq = nes_get_seq(skb, ack, wnd, fin_rcvd, rst_rcvd);
 		if (seq == nextseq) {
 			if (skb->len || processacks)
@@ -218,14 +221,13 @@ static struct sk_buff *nes_get_next_skb(struct nes_device *nesdev, struct nes_qp
 			goto out;
 		}
 
-		if (skb->next == (struct sk_buff *)&nesqp->pau_list)
-			goto out;
-
 		old_skb = skb;
 		skb = skb->next;
 		skb_unlink(old_skb, &nesqp->pau_list);
 		nes_mgt_free_skb(nesdev, old_skb, PCI_DMA_TODEVICE);
 		nes_rem_ref_cm_node(nesqp->cm_node);
+		if (skb == (struct sk_buff *)&nesqp->pau_list)
+			goto out;
 	}
 	return skb;
 
@@ -245,7 +247,6 @@ static int get_fpdu_info(struct nes_device *nesdev, struct nes_qp *nesqp,
 	struct nes_rskb_cb *cb;
 	struct pau_fpdu_info *fpdu_info = NULL;
 	struct pau_fpdu_frag frags[MAX_FPDU_FRAGS];
-	unsigned long flags;
 	u32 fpdu_len = 0;
 	u32 tmp_len;
 	int frag_cnt = 0;
@@ -260,12 +261,10 @@ static int get_fpdu_info(struct nes_device *nesdev, struct nes_qp *nesqp,
 
 	*pau_fpdu_info = NULL;
 
-	spin_lock_irqsave(&nesqp->pau_lock, flags);
 	skb = nes_get_next_skb(nesdev, nesqp, NULL, nesqp->pau_rcv_nxt, &ack, &wnd, &fin_rcvd, &rst_rcvd);
-	if (!skb) {
-		spin_unlock_irqrestore(&nesqp->pau_lock, flags);
+	if (!skb)
 		goto out;
-	}
+
 	cb = (struct nes_rskb_cb *)&skb->cb[0];
 	if (skb->len) {
 		fpdu_len = be16_to_cpu(*(__be16 *) skb->data) + MPA_FRAMING;
@@ -290,10 +289,9 @@ static int get_fpdu_info(struct nes_device *nesdev, struct nes_qp *nesqp,
 
 			skb = nes_get_next_skb(nesdev, nesqp, skb,
 					       nesqp->pau_rcv_nxt + frag_tot, &ack, &wnd, &fin_rcvd, &rst_rcvd);
-			if (!skb) {
-				spin_unlock_irqrestore(&nesqp->pau_lock, flags);
+			if (!skb)
 				goto out;
-			} else if (rst_rcvd) {
+			if (rst_rcvd) {
 				/* rst received in the middle of fpdu */
 				for (; i >= 0; i--) {
 					skb_unlink(frags[i].skb, &nesqp->pau_list);
@@ -320,8 +318,6 @@ static int get_fpdu_info(struct nes_device *nesdev, struct nes_qp *nesqp,
 		frag_cnt = 1;
 	}
 
-	spin_unlock_irqrestore(&nesqp->pau_lock, flags);
-
 	/* Found one */
 	fpdu_info = kzalloc(sizeof(*fpdu_info), GFP_ATOMIC);
 	if (fpdu_info == NULL) {
@@ -383,9 +379,8 @@ static int get_fpdu_info(struct nes_device *nesdev, struct nes_qp *nesqp,
 
 		if (frags[i].skb->len == 0) {
 			/* Pull skb off the list - it will be freed in the callback */
-			spin_lock_irqsave(&nesqp->pau_lock, flags);
-			skb_unlink(frags[i].skb, &nesqp->pau_list);
-			spin_unlock_irqrestore(&nesqp->pau_lock, flags);
+			if (!skb_queue_empty(&nesqp->pau_list))
+				skb_unlink(frags[i].skb, &nesqp->pau_list);
 		} else {
 			/* Last skb still has data so update the seq */
 			iph = (struct iphdr *)(cb->data_start + ETH_HLEN);
@@ -414,14 +409,18 @@ static int forward_fpdus(struct nes_vnic *nesvnic, struct nes_qp *nesqp)
 	struct pau_fpdu_info *fpdu_info;
 	struct nes_hw_cqp_wqe *cqp_wqe;
 	struct nes_cqp_request *cqp_request;
+	unsigned long flags;
 	u64 u64tmp;
 	u32 u32tmp;
 	int rc;
 
 	while (1) {
+		spin_lock_irqsave(&nesqp->pau_lock, flags);
 		rc = get_fpdu_info(nesdev, nesqp, &fpdu_info);
-		if (fpdu_info == NULL)
+		if (rc || (fpdu_info == NULL)) {
+			spin_unlock_irqrestore(&nesqp->pau_lock, flags);
 			return rc;
+		}
 
 		cqp_request = fpdu_info->cqp_request;
 		cqp_wqe = &cqp_request->cqp_wqe;
@@ -447,7 +446,7 @@ static int forward_fpdus(struct nes_vnic *nesvnic, struct nes_qp *nesqp)
 		set_wqe_32bit_value(cqp_wqe->wqe_words, NES_NIC_SQ_WQE_FRAG0_LOW_IDX,
 				    lower_32_bits(u64tmp));
 		set_wqe_32bit_value(cqp_wqe->wqe_words, NES_NIC_SQ_WQE_FRAG0_HIGH_IDX,
-				    upper_32_bits(u64tmp >> 32));
+				    upper_32_bits(u64tmp));
 
 		set_wqe_32bit_value(cqp_wqe->wqe_words, NES_NIC_SQ_WQE_FRAG1_LOW_IDX,
 				    lower_32_bits(fpdu_info->frags[0].physaddr));
@@ -475,6 +474,7 @@ static int forward_fpdus(struct nes_vnic *nesvnic, struct nes_qp *nesqp)
 
 		atomic_set(&cqp_request->refcount, 1);
 		nes_post_cqp_request(nesdev, cqp_request);
+		spin_unlock_irqrestore(&nesqp->pau_lock, flags);
 	}
 
 	return 0;
@@ -649,11 +649,9 @@ static void nes_chg_qh_handler(struct nes_device *nesdev, struct nes_cqp_request
 	nesqp = qh_chg->nesqp;
 
 	/* Should we handle the bad completion */
-	if (cqp_request->major_code) {
-		printk(KERN_ERR PFX "Invalid cqp_request major_code=0x%x\n",
+	if (cqp_request->major_code)
+		WARN(1, PFX "Invalid cqp_request major_code=0x%x\n",
 		       cqp_request->major_code);
-		WARN_ON(1);
-	}
 
 	switch (nesqp->pau_state) {
 	case PAU_DEL_QH:
diff --git a/drivers/infiniband/hw/nes/nes_nic.c b/drivers/infiniband/hw/nes/nes_nic.c
index 0564be757d82..9542e1644a5c 100644
--- a/drivers/infiniband/hw/nes/nes_nic.c
+++ b/drivers/infiniband/hw/nes/nes_nic.c
@@ -944,12 +944,13 @@ static void nes_netdev_set_multicast_list(struct net_device *netdev)
 					  addr,
 					  perfect_filter_register_address+(mc_index * 8),
 					  mc_nic_index);
-				macaddr_high  = ((u16) addr[0]) << 8;
-				macaddr_high += (u16) addr[1];
-				macaddr_low   = ((u32) addr[2]) << 24;
-				macaddr_low  += ((u32) addr[3]) << 16;
-				macaddr_low  += ((u32) addr[4]) << 8;
-				macaddr_low  += (u32) addr[5];
+				macaddr_high  = ((u8) addr[0]) << 8;
+				macaddr_high += (u8) addr[1];
+				macaddr_low   = ((u8) addr[2]) << 24;
+				macaddr_low  += ((u8) addr[3]) << 16;
+				macaddr_low  += ((u8) addr[4]) << 8;
+				macaddr_low  += (u8) addr[5];
+
 				nes_write_indexed(nesdev,
 						perfect_filter_register_address+(mc_index * 8),
 						macaddr_low);
diff --git a/drivers/infiniband/hw/nes/nes_verbs.c b/drivers/infiniband/hw/nes/nes_verbs.c
index cd0ecb215cca..07e4fbad987a 100644
--- a/drivers/infiniband/hw/nes/nes_verbs.c
+++ b/drivers/infiniband/hw/nes/nes_verbs.c
@@ -1404,6 +1404,9 @@ static struct ib_qp *nes_create_qp(struct ib_pd *ibpd,
 	}
 
 	nesqp->sig_all = (init_attr->sq_sig_type == IB_SIGNAL_ALL_WR);
+	init_timer(&nesqp->terminate_timer);
+	nesqp->terminate_timer.function = nes_terminate_timeout;
+	nesqp->terminate_timer.data = (unsigned long)nesqp;
 
 	/* update the QP table */
 	nesdev->nesadapter->qp_table[nesqp->hwqp.qp_id-NES_FIRST_QPN] = nesqp;
@@ -1413,7 +1416,6 @@ static struct ib_qp *nes_create_qp(struct ib_pd *ibpd,
 	return &nesqp->ibqp;
 }
 
-
 /**
  * nes_clean_cq
  */
@@ -2559,6 +2561,11 @@ static struct ib_mr *nes_reg_user_mr(struct ib_pd *pd, u64 start, u64 length,
 			return ibmr;
 		case IWNES_MEMREG_TYPE_QP:
 		case IWNES_MEMREG_TYPE_CQ:
+			if (!region->length) {
+				nes_debug(NES_DBG_MR, "Unable to register zero length region for CQ\n");
+				ib_umem_release(region);
+				return ERR_PTR(-EINVAL);
+			}
 			nespbl = kzalloc(sizeof(*nespbl), GFP_KERNEL);
 			if (!nespbl) {
 				nes_debug(NES_DBG_MR, "Unable to allocate PBL\n");