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

Pull infiniband/rdma updates from Roland Dreier:

 - mostly cxgb4 fixes unblocked by the merge of some prerequisites via
   the net tree

 - drop deprecated MSI-X API use.

 - a couple other miscellaneous things.

* tag 'rdma-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/roland/infiniband:
  RDMA/cxgb4: Fix over-dereference when terminating
  RDMA/cxgb4: Use uninitialized_var()
  RDMA/cxgb4: Add missing debug stats
  RDMA/cxgb4: Initialize reserved fields in a FW work request
  RDMA/cxgb4: Use pr_warn_ratelimited
  RDMA/cxgb4: Max fastreg depth depends on DSGL support
  RDMA/cxgb4: SQ flush fix
  RDMA/cxgb4: rmb() after reading valid gen bit
  RDMA/cxgb4: Endpoint timeout fixes
  RDMA/cxgb4: Use the BAR2/WC path for kernel QPs and T5 devices
  IB/mlx5: Add block multicast loopback support
  IB/mthca: Use pci_enable_msix_exact() instead of pci_enable_msix()
  IB/qib: Use pci_enable_msix_range() instead of pci_enable_msix()

9 files changed, 224 insertions, 92 deletions

diff --git a/drivers/infiniband/hw/cxgb4/cm.c b/drivers/infiniband/hw/cxgb4/cm.c
index 02436d5d0da..185452abf32 100644
--- a/drivers/infiniband/hw/cxgb4/cm.c
+++ b/drivers/infiniband/hw/cxgb4/cm.c
@@ -173,12 +173,15 @@ static void start_ep_timer(struct c4iw_ep *ep)
 	add_timer(&ep->timer);
 }
 
-static void stop_ep_timer(struct c4iw_ep *ep)
+static int stop_ep_timer(struct c4iw_ep *ep)
 {
 	PDBG("%s ep %p stopping\n", __func__, ep);
 	del_timer_sync(&ep->timer);
-	if (!test_and_set_bit(TIMEOUT, &ep->com.flags))
+	if (!test_and_set_bit(TIMEOUT, &ep->com.flags)) {
 		c4iw_put_ep(&ep->com);
+		return 0;
+	}
+	return 1;
 }
 
 static int c4iw_l2t_send(struct c4iw_rdev *rdev, struct sk_buff *skb,
@@ -1165,12 +1168,11 @@ static void process_mpa_reply(struct c4iw_ep *ep, struct sk_buff *skb)
 	PDBG("%s ep %p tid %u\n", __func__, ep, ep->hwtid);
 
 	/*
-	 * Stop mpa timer.  If it expired, then the state has
-	 * changed and we bail since ep_timeout already aborted
-	 * the connection.
+	 * Stop mpa timer.  If it expired, then
+	 * we ignore the MPA reply.  process_timeout()
+	 * will abort the connection.
 	 */
-	stop_ep_timer(ep);
-	if (ep->com.state != MPA_REQ_SENT)
+	if (stop_ep_timer(ep))
 		return;
 
 	/*
@@ -1375,15 +1377,12 @@ static void process_mpa_request(struct c4iw_ep *ep, struct sk_buff *skb)
 
 	PDBG("%s ep %p tid %u\n", __func__, ep, ep->hwtid);
 
-	if (ep->com.state != MPA_REQ_WAIT)
-		return;
-
 	/*
 	 * If we get more than the supported amount of private data
 	 * then we must fail this connection.
 	 */
 	if (ep->mpa_pkt_len + skb->len > sizeof(ep->mpa_pkt)) {
-		stop_ep_timer(ep);
+		(void)stop_ep_timer(ep);
 		abort_connection(ep, skb, GFP_KERNEL);
 		return;
 	}
@@ -1413,13 +1412,13 @@ static void process_mpa_request(struct c4iw_ep *ep, struct sk_buff *skb)
 	if (mpa->revision > mpa_rev) {
 		printk(KERN_ERR MOD "%s MPA version mismatch. Local = %d,"
 		       " Received = %d\n", __func__, mpa_rev, mpa->revision);
-		stop_ep_timer(ep);
+		(void)stop_ep_timer(ep);
 		abort_connection(ep, skb, GFP_KERNEL);
 		return;
 	}
 
 	if (memcmp(mpa->key, MPA_KEY_REQ, sizeof(mpa->key))) {
-		stop_ep_timer(ep);
+		(void)stop_ep_timer(ep);
 		abort_connection(ep, skb, GFP_KERNEL);
 		return;
 	}
@@ -1430,7 +1429,7 @@ static void process_mpa_request(struct c4iw_ep *ep, struct sk_buff *skb)
 	 * Fail if there's too much private data.
 	 */
 	if (plen > MPA_MAX_PRIVATE_DATA) {
-		stop_ep_timer(ep);
+		(void)stop_ep_timer(ep);
 		abort_connection(ep, skb, GFP_KERNEL);
 		return;
 	}
@@ -1439,7 +1438,7 @@ static void process_mpa_request(struct c4iw_ep *ep, struct sk_buff *skb)
 	 * If plen does not account for pkt size
 	 */
 	if (ep->mpa_pkt_len > (sizeof(*mpa) + plen)) {
-		stop_ep_timer(ep);
+		(void)stop_ep_timer(ep);
 		abort_connection(ep, skb, GFP_KERNEL);
 		return;
 	}
@@ -1496,18 +1495,24 @@ static void process_mpa_request(struct c4iw_ep *ep, struct sk_buff *skb)
 	     ep->mpa_attr.xmit_marker_enabled, ep->mpa_attr.version,
 	     ep->mpa_attr.p2p_type);
 
-	__state_set(&ep->com, MPA_REQ_RCVD);
-	stop_ep_timer(ep);
-
-	/* drive upcall */
-	mutex_lock(&ep->parent_ep->com.mutex);
-	if (ep->parent_ep->com.state != DEAD) {
-		if (connect_request_upcall(ep))
+	/*
+	 * If the endpoint timer already expired, then we ignore
+	 * the start request.  process_timeout() will abort
+	 * the connection.
+	 */
+	if (!stop_ep_timer(ep)) {
+		__state_set(&ep->com, MPA_REQ_RCVD);
+
+		/* drive upcall */
+		mutex_lock(&ep->parent_ep->com.mutex);
+		if (ep->parent_ep->com.state != DEAD) {
+			if (connect_request_upcall(ep))
+				abort_connection(ep, skb, GFP_KERNEL);
+		} else {
 			abort_connection(ep, skb, GFP_KERNEL);
-	} else {
-		abort_connection(ep, skb, GFP_KERNEL);
+		}
+		mutex_unlock(&ep->parent_ep->com.mutex);
 	}
-	mutex_unlock(&ep->parent_ep->com.mutex);
 	return;
 }
 
@@ -2265,7 +2270,7 @@ static int peer_close(struct c4iw_dev *dev, struct sk_buff *skb)
 		disconnect = 0;
 		break;
 	case MORIBUND:
-		stop_ep_timer(ep);
+		(void)stop_ep_timer(ep);
 		if (ep->com.cm_id && ep->com.qp) {
 			attrs.next_state = C4IW_QP_STATE_IDLE;
 			c4iw_modify_qp(ep->com.qp->rhp, ep->com.qp,
@@ -2325,10 +2330,10 @@ static int peer_abort(struct c4iw_dev *dev, struct sk_buff *skb)
 	case CONNECTING:
 		break;
 	case MPA_REQ_WAIT:
-		stop_ep_timer(ep);
+		(void)stop_ep_timer(ep);
 		break;
 	case MPA_REQ_SENT:
-		stop_ep_timer(ep);
+		(void)stop_ep_timer(ep);
 		if (mpa_rev == 1 || (mpa_rev == 2 && ep->tried_with_mpa_v1))
 			connect_reply_upcall(ep, -ECONNRESET);
 		else {
@@ -2433,7 +2438,7 @@ static int close_con_rpl(struct c4iw_dev *dev, struct sk_buff *skb)
 		__state_set(&ep->com, MORIBUND);
 		break;
 	case MORIBUND:
-		stop_ep_timer(ep);
+		(void)stop_ep_timer(ep);
 		if ((ep->com.cm_id) && (ep->com.qp)) {
 			attrs.next_state = C4IW_QP_STATE_IDLE;
 			c4iw_modify_qp(ep->com.qp->rhp,
@@ -3028,7 +3033,7 @@ int c4iw_ep_disconnect(struct c4iw_ep *ep, int abrupt, gfp_t gfp)
 		if (!test_and_set_bit(CLOSE_SENT, &ep->com.flags)) {
 			close = 1;
 			if (abrupt) {
-				stop_ep_timer(ep);
+				(void)stop_ep_timer(ep);
 				ep->com.state = ABORTING;
 			} else
 				ep->com.state = MORIBUND;
@@ -3462,6 +3467,16 @@ static void process_timeout(struct c4iw_ep *ep)
 		__state_set(&ep->com, ABORTING);
 		close_complete_upcall(ep, -ETIMEDOUT);
 		break;
+	case ABORTING:
+	case DEAD:
+
+		/*
+		 * These states are expected if the ep timed out at the same
+		 * time as another thread was calling stop_ep_timer().
+		 * So we silently do nothing for these states.
+		 */
+		abort = 0;
+		break;
 	default:
 		WARN(1, "%s unexpected state ep %p tid %u state %u\n",
 			__func__, ep, ep->hwtid, ep->com.state);
@@ -3483,6 +3498,8 @@ static void process_timedout_eps(void)
 
 		tmp = timeout_list.next;
 		list_del(tmp);
+		tmp->next = NULL;
+		tmp->prev = NULL;
 		spin_unlock_irq(&timeout_lock);
 		ep = list_entry(tmp, struct c4iw_ep, entry);
 		process_timeout(ep);
@@ -3499,6 +3516,7 @@ static void process_work(struct work_struct *work)
 	unsigned int opcode;
 	int ret;
 
+	process_timedout_eps();
 	while ((skb = skb_dequeue(&rxq))) {
 		rpl = cplhdr(skb);
 		dev = *((struct c4iw_dev **) (skb->cb + sizeof(void *)));
@@ -3508,8 +3526,8 @@ static void process_work(struct work_struct *work)
 		ret = work_handlers[opcode](dev, skb);
 		if (!ret)
 			kfree_skb(skb);
+		process_timedout_eps();
 	}
-	process_timedout_eps();
 }
 
 static DECLARE_WORK(skb_work, process_work);
@@ -3521,8 +3539,13 @@ static void ep_timeout(unsigned long arg)
 
 	spin_lock(&timeout_lock);
 	if (!test_and_set_bit(TIMEOUT, &ep->com.flags)) {
-		list_add_tail(&ep->entry, &timeout_list);
-		kickit = 1;
+		/*
+		 * Only insert if it is not already on the list.
+		 */
+		if (!ep->entry.next) {
+			list_add_tail(&ep->entry, &timeout_list);
+			kickit = 1;
+		}
 	}
 	spin_unlock(&timeout_lock);
 	if (kickit)
diff --git a/drivers/infiniband/hw/cxgb4/cq.c b/drivers/infiniband/hw/cxgb4/cq.c
index ce468e54242..cfaa56ada18 100644
--- a/drivers/infiniband/hw/cxgb4/cq.c
+++ b/drivers/infiniband/hw/cxgb4/cq.c
@@ -235,27 +235,21 @@ int c4iw_flush_sq(struct c4iw_qp *qhp)
 	struct t4_cq *cq = &chp->cq;
 	int idx;
 	struct t4_swsqe *swsqe;
-	int error = (qhp->attr.state != C4IW_QP_STATE_CLOSING &&
-			qhp->attr.state != C4IW_QP_STATE_IDLE);
 
 	if (wq->sq.flush_cidx == -1)
 		wq->sq.flush_cidx = wq->sq.cidx;
 	idx = wq->sq.flush_cidx;
 	BUG_ON(idx >= wq->sq.size);
 	while (idx != wq->sq.pidx) {
-		if (error) {
-			swsqe = &wq->sq.sw_sq[idx];
-			BUG_ON(swsqe->flushed);
-			swsqe->flushed = 1;
-			insert_sq_cqe(wq, cq, swsqe);
-			if (wq->sq.oldest_read == swsqe) {
-				BUG_ON(swsqe->opcode != FW_RI_READ_REQ);
-				advance_oldest_read(wq);
-			}
-			flushed++;
-		} else {
-			t4_sq_consume(wq);
+		swsqe = &wq->sq.sw_sq[idx];
+		BUG_ON(swsqe->flushed);
+		swsqe->flushed = 1;
+		insert_sq_cqe(wq, cq, swsqe);
+		if (wq->sq.oldest_read == swsqe) {
+			BUG_ON(swsqe->opcode != FW_RI_READ_REQ);
+			advance_oldest_read(wq);
 		}
+		flushed++;
 		if (++idx == wq->sq.size)
 			idx = 0;
 	}
@@ -678,7 +672,7 @@ skip_cqe:
 static int c4iw_poll_cq_one(struct c4iw_cq *chp, struct ib_wc *wc)
 {
 	struct c4iw_qp *qhp = NULL;
-	struct t4_cqe cqe = {0, 0}, *rd_cqe;
+	struct t4_cqe uninitialized_var(cqe), *rd_cqe;
 	struct t4_wq *wq;
 	u32 credit = 0;
 	u8 cqe_flushed;
diff --git a/drivers/infiniband/hw/cxgb4/device.c b/drivers/infiniband/hw/cxgb4/device.c
index 9489a388376..f4fa50a609e 100644
--- a/drivers/infiniband/hw/cxgb4/device.c
+++ b/drivers/infiniband/hw/cxgb4/device.c
@@ -682,7 +682,10 @@ static void c4iw_dealloc(struct uld_ctx *ctx)
 	idr_destroy(&ctx->dev->hwtid_idr);
 	idr_destroy(&ctx->dev->stid_idr);
 	idr_destroy(&ctx->dev->atid_idr);
-	iounmap(ctx->dev->rdev.oc_mw_kva);
+	if (ctx->dev->rdev.bar2_kva)
+		iounmap(ctx->dev->rdev.bar2_kva);
+	if (ctx->dev->rdev.oc_mw_kva)
+		iounmap(ctx->dev->rdev.oc_mw_kva);
 	ib_dealloc_device(&ctx->dev->ibdev);
 	ctx->dev = NULL;
 }
@@ -722,11 +725,31 @@ static struct c4iw_dev *c4iw_alloc(const struct cxgb4_lld_info *infop)
 	}
 	devp->rdev.lldi = *infop;
 
-	devp->rdev.oc_mw_pa = pci_resource_start(devp->rdev.lldi.pdev, 2) +
-		(pci_resource_len(devp->rdev.lldi.pdev, 2) -
-		 roundup_pow_of_two(devp->rdev.lldi.vr->ocq.size));
-	devp->rdev.oc_mw_kva = ioremap_wc(devp->rdev.oc_mw_pa,
-					       devp->rdev.lldi.vr->ocq.size);
+	/*
+	 * For T5 devices, we map all of BAR2 with WC.
+	 * For T4 devices with onchip qp mem, we map only that part
+	 * of BAR2 with WC.
+	 */
+	devp->rdev.bar2_pa = pci_resource_start(devp->rdev.lldi.pdev, 2);
+	if (is_t5(devp->rdev.lldi.adapter_type)) {
+		devp->rdev.bar2_kva = ioremap_wc(devp->rdev.bar2_pa,
+			pci_resource_len(devp->rdev.lldi.pdev, 2));
+		if (!devp->rdev.bar2_kva) {
+			pr_err(MOD "Unable to ioremap BAR2\n");
+			return ERR_PTR(-EINVAL);
+		}
+	} else if (ocqp_supported(infop)) {
+		devp->rdev.oc_mw_pa =
+			pci_resource_start(devp->rdev.lldi.pdev, 2) +
+			pci_resource_len(devp->rdev.lldi.pdev, 2) -
+			roundup_pow_of_two(devp->rdev.lldi.vr->ocq.size);
+		devp->rdev.oc_mw_kva = ioremap_wc(devp->rdev.oc_mw_pa,
+			devp->rdev.lldi.vr->ocq.size);
+		if (!devp->rdev.oc_mw_kva) {
+			pr_err(MOD "Unable to ioremap onchip mem\n");
+			return ERR_PTR(-EINVAL);
+		}
+	}
 
 	PDBG(KERN_INFO MOD "ocq memory: "
 	       "hw_start 0x%x size %u mw_pa 0x%lx mw_kva %p\n",
@@ -1003,9 +1026,11 @@ static int enable_qp_db(int id, void *p, void *data)
 static void resume_rc_qp(struct c4iw_qp *qp)
 {
 	spin_lock(&qp->lock);
-	t4_ring_sq_db(&qp->wq, qp->wq.sq.wq_pidx_inc);
+	t4_ring_sq_db(&qp->wq, qp->wq.sq.wq_pidx_inc,
+		      is_t5(qp->rhp->rdev.lldi.adapter_type), NULL);
 	qp->wq.sq.wq_pidx_inc = 0;
-	t4_ring_rq_db(&qp->wq, qp->wq.rq.wq_pidx_inc);
+	t4_ring_rq_db(&qp->wq, qp->wq.rq.wq_pidx_inc,
+		      is_t5(qp->rhp->rdev.lldi.adapter_type), NULL);
 	qp->wq.rq.wq_pidx_inc = 0;
 	spin_unlock(&qp->lock);
 }
diff --git a/drivers/infiniband/hw/cxgb4/iw_cxgb4.h b/drivers/infiniband/hw/cxgb4/iw_cxgb4.h
index e872203c542..7b8c5806a09 100644
--- a/drivers/infiniband/hw/cxgb4/iw_cxgb4.h
+++ b/drivers/infiniband/hw/cxgb4/iw_cxgb4.h
@@ -149,6 +149,8 @@ struct c4iw_rdev {
 	struct gen_pool *ocqp_pool;
 	u32 flags;
 	struct cxgb4_lld_info lldi;
+	unsigned long bar2_pa;
+	void __iomem *bar2_kva;
 	unsigned long oc_mw_pa;
 	void __iomem *oc_mw_kva;
 	struct c4iw_stats stats;
diff --git a/drivers/infiniband/hw/cxgb4/mem.c b/drivers/infiniband/hw/cxgb4/mem.c
index f9ca072a99e..ec7a2988a70 100644
--- a/drivers/infiniband/hw/cxgb4/mem.c
+++ b/drivers/infiniband/hw/cxgb4/mem.c
@@ -259,8 +259,12 @@ static int write_tpt_entry(struct c4iw_rdev *rdev, u32 reset_tpt_entry,
 
 	if ((!reset_tpt_entry) && (*stag == T4_STAG_UNSET)) {
 		stag_idx = c4iw_get_resource(&rdev->resource.tpt_table);
-		if (!stag_idx)
+		if (!stag_idx) {
+			mutex_lock(&rdev->stats.lock);
+			rdev->stats.stag.fail++;
+			mutex_unlock(&rdev->stats.lock);
 			return -ENOMEM;
+		}
 		mutex_lock(&rdev->stats.lock);
 		rdev->stats.stag.cur += 32;
 		if (rdev->stats.stag.cur > rdev->stats.stag.max)
diff --git a/drivers/infiniband/hw/cxgb4/provider.c b/drivers/infiniband/hw/cxgb4/provider.c
index 79429256023..a94a3e12c34 100644
--- a/drivers/infiniband/hw/cxgb4/provider.c
+++ b/drivers/infiniband/hw/cxgb4/provider.c
@@ -328,7 +328,7 @@ static int c4iw_query_device(struct ib_device *ibdev,
 	props->max_mr = c4iw_num_stags(&dev->rdev);
 	props->max_pd = T4_MAX_NUM_PD;
 	props->local_ca_ack_delay = 0;
-	props->max_fast_reg_page_list_len = T4_MAX_FR_DEPTH;
+	props->max_fast_reg_page_list_len = t4_max_fr_depth(use_dsgl);
 
 	return 0;
 }
diff --git a/drivers/infiniband/hw/cxgb4/qp.c b/drivers/infiniband/hw/cxgb4/qp.c
index cb76eb5eee1..7b5114cb486 100644
--- a/drivers/infiniband/hw/cxgb4/qp.c
+++ b/drivers/infiniband/hw/cxgb4/qp.c
@@ -212,13 +212,23 @@ static int create_qp(struct c4iw_rdev *rdev, struct t4_wq *wq,
 
 	wq->db = rdev->lldi.db_reg;
 	wq->gts = rdev->lldi.gts_reg;
-	if (user) {
-		wq->sq.udb = (u64)pci_resource_start(rdev->lldi.pdev, 2) +
-					(wq->sq.qid << rdev->qpshift);
-		wq->sq.udb &= PAGE_MASK;
-		wq->rq.udb = (u64)pci_resource_start(rdev->lldi.pdev, 2) +
-					(wq->rq.qid << rdev->qpshift);
-		wq->rq.udb &= PAGE_MASK;
+	if (user || is_t5(rdev->lldi.adapter_type)) {
+		u32 off;
+
+		off = (wq->sq.qid << rdev->qpshift) & PAGE_MASK;
+		if (user) {
+			wq->sq.udb = (u64 __iomem *)(rdev->bar2_pa + off);
+		} else {
+			off += 128 * (wq->sq.qid & rdev->qpmask) + 8;
+			wq->sq.udb = (u64 __iomem *)(rdev->bar2_kva + off);
+		}
+		off = (wq->rq.qid << rdev->qpshift) & PAGE_MASK;
+		if (user) {
+			wq->rq.udb = (u64 __iomem *)(rdev->bar2_pa + off);
+		} else {
+			off += 128 * (wq->rq.qid & rdev->qpmask) + 8;
+			wq->rq.udb = (u64 __iomem *)(rdev->bar2_kva + off);
+		}
 	}
 	wq->rdev = rdev;
 	wq->rq.msn = 1;
@@ -299,9 +309,10 @@ static int create_qp(struct c4iw_rdev *rdev, struct t4_wq *wq,
 	if (ret)
 		goto free_dma;
 
-	PDBG("%s sqid 0x%x rqid 0x%x kdb 0x%p squdb 0x%llx rqudb 0x%llx\n",
+	PDBG("%s sqid 0x%x rqid 0x%x kdb 0x%p squdb 0x%lx rqudb 0x%lx\n",
 	     __func__, wq->sq.qid, wq->rq.qid, wq->db,
-	     (unsigned long long)wq->sq.udb, (unsigned long long)wq->rq.udb);
+	     (__force unsigned long) wq->sq.udb,
+	     (__force unsigned long) wq->rq.udb);
 
 	return 0;
 free_dma:
@@ -425,6 +436,8 @@ static int build_rdma_send(struct t4_sq *sq, union t4_wr *wqe,
 	default:
 		return -EINVAL;
 	}
+	wqe->send.r3 = 0;
+	wqe->send.r4 = 0;
 
 	plen = 0;
 	if (wr->num_sge) {
@@ -555,7 +568,8 @@ static int build_fastreg(struct t4_sq *sq, union t4_wr *wqe,
 	int pbllen = roundup(wr->wr.fast_reg.page_list_len * sizeof(u64), 32);
 	int rem;
 
-	if (wr->wr.fast_reg.page_list_len > T4_MAX_FR_DEPTH)
+	if (wr->wr.fast_reg.page_list_len >
+	    t4_max_fr_depth(use_dsgl))
 		return -EINVAL;
 
 	wqe->fr.qpbinde_to_dcacpu = 0;
@@ -650,9 +664,10 @@ static int ring_kernel_sq_db(struct c4iw_qp *qhp, u16 inc)
 
 	spin_lock_irqsave(&qhp->rhp->lock, flags);
 	spin_lock(&qhp->lock);
-	if (qhp->rhp->db_state == NORMAL) {
-		t4_ring_sq_db(&qhp->wq, inc);
-	} else {
+	if (qhp->rhp->db_state == NORMAL)
+		t4_ring_sq_db(&qhp->wq, inc,
+			      is_t5(qhp->rhp->rdev.lldi.adapter_type), NULL);
+	else {
 		add_to_fc_list(&qhp->rhp->db_fc_list, &qhp->db_fc_entry);
 		qhp->wq.sq.wq_pidx_inc += inc;
 	}
@@ -667,9 +682,10 @@ static int ring_kernel_rq_db(struct c4iw_qp *qhp, u16 inc)
 
 	spin_lock_irqsave(&qhp->rhp->lock, flags);
 	spin_lock(&qhp->lock);
-	if (qhp->rhp->db_state == NORMAL) {
-		t4_ring_rq_db(&qhp->wq, inc);
-	} else {
+	if (qhp->rhp->db_state == NORMAL)
+		t4_ring_rq_db(&qhp->wq, inc,
+			      is_t5(qhp->rhp->rdev.lldi.adapter_type), NULL);
+	else {
 		add_to_fc_list(&qhp->rhp->db_fc_list, &qhp->db_fc_entry);
 		qhp->wq.rq.wq_pidx_inc += inc;
 	}
@@ -686,7 +702,7 @@ int c4iw_post_send(struct ib_qp *ibqp, struct ib_send_wr *wr,
 	enum fw_wr_opcodes fw_opcode = 0;
 	enum fw_ri_wr_flags fw_flags;
 	struct c4iw_qp *qhp;
-	union t4_wr *wqe;
+	union t4_wr *wqe = NULL;
 	u32 num_wrs;
 	struct t4_swsqe *swsqe;
 	unsigned long flag;
@@ -792,7 +808,8 @@ int c4iw_post_send(struct ib_qp *ibqp, struct ib_send_wr *wr,
 		idx += DIV_ROUND_UP(len16*16, T4_EQ_ENTRY_SIZE);
 	}
 	if (!qhp->rhp->rdev.status_page->db_off) {
-		t4_ring_sq_db(&qhp->wq, idx);
+		t4_ring_sq_db(&qhp->wq, idx,
+			      is_t5(qhp->rhp->rdev.lldi.adapter_type), wqe);
 		spin_unlock_irqrestore(&qhp->lock, flag);
 	} else {
 		spin_unlock_irqrestore(&qhp->lock, flag);
@@ -806,7 +823,7 @@ int c4iw_post_receive(struct ib_qp *ibqp, struct ib_recv_wr *wr,
 {
 	int err = 0;
 	struct c4iw_qp *qhp;
-	union t4_recv_wr *wqe;
+	union t4_recv_wr *wqe = NULL;
 	u32 num_wrs;
 	u8 len16 = 0;
 	unsigned long flag;
@@ -858,7 +875,8 @@ int c4iw_post_receive(struct ib_qp *ibqp, struct ib_recv_wr *wr,
 		num_wrs--;
 	}
 	if (!qhp->rhp->rdev.status_page->db_off) {
-		t4_ring_rq_db(&qhp->wq, idx);
+		t4_ring_rq_db(&qhp->wq, idx,
+			      is_t5(qhp->rhp->rdev.lldi.adapter_type), wqe);
 		spin_unlock_irqrestore(&qhp->lock, flag);
 	} else {
 		spin_unlock_irqrestore(&qhp->lock, flag);
@@ -1352,6 +1370,7 @@ int c4iw_modify_qp(struct c4iw_dev *rhp, struct c4iw_qp *qhp,
 		switch (attrs->next_state) {
 		case C4IW_QP_STATE_CLOSING:
 			BUG_ON(atomic_read(&qhp->ep->com.kref.refcount) < 2);
+			t4_set_wq_in_error(&qhp->wq);
 			set_state(qhp, C4IW_QP_STATE_CLOSING);
 			ep = qhp->ep;
 			if (!internal) {
@@ -1359,18 +1378,18 @@ int c4iw_modify_qp(struct c4iw_dev *rhp, struct c4iw_qp *qhp,
 				disconnect = 1;
 				c4iw_get_ep(&qhp->ep->com);
 			}
-			t4_set_wq_in_error(&qhp->wq);
 			ret = rdma_fini(rhp, qhp, ep);
 			if (ret)
 				goto err;
 			break;
 		case C4IW_QP_STATE_TERMINATE:
+			t4_set_wq_in_error(&qhp->wq);
 			set_state(qhp, C4IW_QP_STATE_TERMINATE);
 			qhp->attr.layer_etype = attrs->layer_etype;
 			qhp->attr.ecode = attrs->ecode;
-			t4_set_wq_in_error(&qhp->wq);
 			ep = qhp->ep;
 			disconnect = 1;
+			c4iw_get_ep(&qhp->ep->com);
 			if (!internal)
 				terminate = 1;
 			else {
@@ -1378,11 +1397,10 @@ int c4iw_modify_qp(struct c4iw_dev *rhp, struct c4iw_qp *qhp,
 				if (ret)
 					goto err;
 			}
-			c4iw_get_ep(&qhp->ep->com);
 			break;
 		case C4IW_QP_STATE_ERROR:
-			set_state(qhp, C4IW_QP_STATE_ERROR);
 			t4_set_wq_in_error(&qhp->wq);
+			set_state(qhp, C4IW_QP_STATE_ERROR);
 			if (!internal) {
 				abort = 1;
 				disconnect = 1;
@@ -1677,11 +1695,11 @@ struct ib_qp *c4iw_create_qp(struct ib_pd *pd, struct ib_qp_init_attr *attrs,
 		mm2->len = PAGE_ALIGN(qhp->wq.rq.memsize);
 		insert_mmap(ucontext, mm2);
 		mm3->key = uresp.sq_db_gts_key;
-		mm3->addr = qhp->wq.sq.udb;
+		mm3->addr = (__force unsigned long) qhp->wq.sq.udb;
 		mm3->len = PAGE_SIZE;
 		insert_mmap(ucontext, mm3);
 		mm4->key = uresp.rq_db_gts_key;
-		mm4->addr = qhp->wq.rq.udb;
+		mm4->addr = (__force unsigned long) qhp->wq.rq.udb;
 		mm4->len = PAGE_SIZE;
 		insert_mmap(ucontext, mm4);
 		if (mm5) {
diff --git a/drivers/infiniband/hw/cxgb4/resource.c b/drivers/infiniband/hw/cxgb4/resource.c
index cdef4d7fb6d..67df71a7012 100644
--- a/drivers/infiniband/hw/cxgb4/resource.c
+++ b/drivers/infiniband/hw/cxgb4/resource.c
@@ -179,8 +179,12 @@ u32 c4iw_get_qpid(struct c4iw_rdev *rdev, struct c4iw_dev_ucontext *uctx)
 		kfree(entry);
 	} else {
 		qid = c4iw_get_resource(&rdev->resource.qid_table);
-		if (!qid)
+		if (!qid) {
+			mutex_lock(&rdev->stats.lock);
+			rdev->stats.qid.fail++;
+			mutex_unlock(&rdev->stats.lock);
 			goto out;
+		}
 		mutex_lock(&rdev->stats.lock);
 		rdev->stats.qid.cur += rdev->qpmask + 1;
 		mutex_unlock(&rdev->stats.lock);
@@ -322,8 +326,8 @@ u32 c4iw_rqtpool_alloc(struct c4iw_rdev *rdev, int size)
 	unsigned long addr = gen_pool_alloc(rdev->rqt_pool, size << 6);
 	PDBG("%s addr 0x%x size %d\n", __func__, (u32)addr, size << 6);
 	if (!addr)
-		printk_ratelimited(KERN_WARNING MOD "%s: Out of RQT memory\n",
-		       pci_name(rdev->lldi.pdev));
+		pr_warn_ratelimited(MOD "%s: Out of RQT memory\n",
+				    pci_name(rdev->lldi.pdev));
 	mutex_lock(&rdev->stats.lock);
 	if (addr) {
 		rdev->stats.rqt.cur += roundup(size << 6, 1 << MIN_RQT_SHIFT);
diff --git a/drivers/infiniband/hw/cxgb4/t4.h b/drivers/infiniband/hw/cxgb4/t4.h
index eeca8b1e637..2178f319841 100644
--- a/drivers/infiniband/hw/cxgb4/t4.h
+++ b/drivers/infiniband/hw/cxgb4/t4.h
@@ -84,7 +84,14 @@ struct t4_status_page {
 			sizeof(struct fw_ri_isgl)) / sizeof(struct fw_ri_sge))
 #define T4_MAX_FR_IMMD ((T4_SQ_NUM_BYTES - sizeof(struct fw_ri_fr_nsmr_wr) - \
 			sizeof(struct fw_ri_immd)) & ~31UL)
-#define T4_MAX_FR_DEPTH (1024 / sizeof(u64))
+#define T4_MAX_FR_IMMD_DEPTH (T4_MAX_FR_IMMD / sizeof(u64))
+#define T4_MAX_FR_DSGL 1024
+#define T4_MAX_FR_DSGL_DEPTH (T4_MAX_FR_DSGL / sizeof(u64))
+
+static inline int t4_max_fr_depth(int use_dsgl)
+{
+	return use_dsgl ? T4_MAX_FR_DSGL_DEPTH : T4_MAX_FR_IMMD_DEPTH;
+}
 
 #define T4_RQ_NUM_SLOTS 2
 #define T4_RQ_NUM_BYTES (T4_EQ_ENTRY_SIZE * T4_RQ_NUM_SLOTS)
@@ -292,7 +299,7 @@ struct t4_sq {
 	unsigned long phys_addr;
 	struct t4_swsqe *sw_sq;
 	struct t4_swsqe *oldest_read;
-	u64 udb;
+	u64 __iomem *udb;
 	size_t memsize;
 	u32 qid;
 	u16 in_use;
@@ -314,7 +321,7 @@ struct t4_rq {
 	dma_addr_t dma_addr;
 	DEFINE_DMA_UNMAP_ADDR(mapping);
 	struct t4_swrqe *sw_rq;
-	u64 udb;
+	u64 __iomem *udb;
 	size_t memsize;
 	u32 qid;
 	u32 msn;
@@ -435,15 +442,67 @@ static inline u16 t4_sq_wq_size(struct t4_wq *wq)
 		return wq->sq.size * T4_SQ_NUM_SLOTS;
 }
 
-static inline void t4_ring_sq_db(struct t4_wq *wq, u16 inc)
+/* This function copies 64 byte coalesced work request to memory
+ * mapped BAR2 space. For coalesced WRs, the SGE fetches data
+ * from the FIFO instead of from Host.
+ */
+static inline void pio_copy(u64 __iomem *dst, u64 *src)
+{
+	int count = 8;
+
+	while (count) {
+		writeq(*src, dst);
+		src++;
+		dst++;
+		count--;
+	}
+}
+
+static inline void t4_ring_sq_db(struct t4_wq *wq, u16 inc, u8 t5,
+				 union t4_wr *wqe)
 {
+
+	/* Flush host queue memory writes. */
 	wmb();
+	if (t5) {
+		if (inc == 1 && wqe) {
+			PDBG("%s: WC wq->sq.pidx = %d\n",
+			     __func__, wq->sq.pidx);
+			pio_copy(wq->sq.udb + 7, (void *)wqe);
+		} else {
+			PDBG("%s: DB wq->sq.pidx = %d\n",
+			     __func__, wq->sq.pidx);
+			writel(PIDX_T5(inc), wq->sq.udb);
+		}
+
+		/* Flush user doorbell area writes. */
+		wmb();
+		return;
+	}
 	writel(QID(wq->sq.qid) | PIDX(inc), wq->db);
 }
 
-static inline void t4_ring_rq_db(struct t4_wq *wq, u16 inc)
+static inline void t4_ring_rq_db(struct t4_wq *wq, u16 inc, u8 t5,
+				 union t4_recv_wr *wqe)
 {
+
+	/* Flush host queue memory writes. */
 	wmb();
+	if (t5) {
+		if (inc == 1 && wqe) {
+			PDBG("%s: WC wq->rq.pidx = %d\n",
+			     __func__, wq->rq.pidx);
+			pio_copy(wq->rq.udb + 7, (void *)wqe);
+		} else {
+			PDBG("%s: DB wq->rq.pidx = %d\n",
+			     __func__, wq->rq.pidx);
+			writel(PIDX_T5(inc), wq->rq.udb);
+		}
+
+		/* Flush user doorbell area writes. */
+		wmb();
+		return;
+	}
 	writel(QID(wq->rq.qid) | PIDX(inc), wq->db);
 }
 
@@ -568,6 +627,9 @@ static inline int t4_next_hw_cqe(struct t4_cq *cq, struct t4_cqe **cqe)
 		printk(KERN_ERR MOD "cq overflow cqid %u\n", cq->cqid);
 		BUG_ON(1);
 	} else if (t4_valid_cqe(cq, &cq->queue[cq->cidx])) {
+
+		/* Ensure CQE is flushed to memory */
+		rmb();
 		*cqe = &cq->queue[cq->cidx];
 		ret = 0;
 	} else