diff --git a/drivers/infiniband/hw/qib/qib.h b/drivers/infiniband/hw/qib/qib.h
index 7e62f4137148..cbe577151457 100644
--- a/drivers/infiniband/hw/qib/qib.h
+++ b/drivers/infiniband/hw/qib/qib.h
@@ -1,8 +1,8 @@
 #ifndef _QIB_KERNEL_H
 #define _QIB_KERNEL_H
 /*
- * Copyright (c) 2006, 2007, 2008, 2009, 2010 QLogic Corporation.
- * All rights reserved.
+ * Copyright (c) 2012 Intel Corporation.  All rights reserved.
+ * Copyright (c) 2006 - 2012 QLogic Corporation. All rights reserved.
  * Copyright (c) 2003, 2004, 2005, 2006 PathScale, Inc. All rights reserved.
  *
  * This software is available to you under a choice of one of two
@@ -544,6 +544,7 @@ struct qib_pportdata {
 
 	/* read mostly */
 	struct qib_sdma_desc *sdma_descq;
+	struct workqueue_struct *qib_wq;
 	struct qib_sdma_state sdma_state;
 	dma_addr_t       sdma_descq_phys;
 	volatile __le64 *sdma_head_dma; /* DMA'ed by chip */
@@ -1267,6 +1268,11 @@ int qib_sdma_verbs_send(struct qib_pportdata *, struct qib_sge_state *,
 /* ppd->sdma_lock should be locked before calling this. */
 int qib_sdma_make_progress(struct qib_pportdata *dd);
 
+static inline int qib_sdma_empty(const struct qib_pportdata *ppd)
+{
+	return ppd->sdma_descq_added == ppd->sdma_descq_removed;
+}
+
 /* must be called under qib_sdma_lock */
 static inline u16 qib_sdma_descq_freecnt(const struct qib_pportdata *ppd)
 {
diff --git a/drivers/infiniband/hw/qib/qib_init.c b/drivers/infiniband/hw/qib/qib_init.c
index dc14e100a7f1..306e65e99e99 100644
--- a/drivers/infiniband/hw/qib/qib_init.c
+++ b/drivers/infiniband/hw/qib/qib_init.c
@@ -1,6 +1,6 @@
 /*
- * Copyright (c) 2006, 2007, 2008, 2009, 2010 QLogic Corporation.
- * All rights reserved.
+ * Copyright (c) 2012 Intel Corporation.  All rights reserved.
+ * Copyright (c) 2006 - 2012 QLogic Corporation. All rights reserved.
  * Copyright (c) 2003, 2004, 2005, 2006 PathScale, Inc. All rights reserved.
  *
  * This software is available to you under a choice of one of two
@@ -210,6 +210,8 @@ void qib_init_pportdata(struct qib_pportdata *ppd, struct qib_devdata *dd,
 	init_timer(&ppd->symerr_clear_timer);
 	ppd->symerr_clear_timer.function = qib_clear_symerror_on_linkup;
 	ppd->symerr_clear_timer.data = (unsigned long)ppd;
+
+	ppd->qib_wq = NULL;
 }
 
 static int init_pioavailregs(struct qib_devdata *dd)
@@ -482,6 +484,42 @@ static void init_piobuf_state(struct qib_devdata *dd)
 	dd->f_initvl15_bufs(dd);
 }
 
+/**
+ * qib_create_workqueues - create per port workqueues
+ * @dd: the qlogic_ib device
+ */
+static int qib_create_workqueues(struct qib_devdata *dd)
+{
+	int pidx;
+	struct qib_pportdata *ppd;
+
+	for (pidx = 0; pidx < dd->num_pports; ++pidx) {
+		ppd = dd->pport + pidx;
+		if (!ppd->qib_wq) {
+			char wq_name[8]; /* 3 + 2 + 1 + 1 + 1 */
+			snprintf(wq_name, sizeof(wq_name), "qib%d_%d",
+				dd->unit, pidx);
+			ppd->qib_wq =
+				create_singlethread_workqueue(wq_name);
+			if (!ppd->qib_wq)
+				goto wq_error;
+		}
+	}
+	return 0;
+wq_error:
+	pr_err(
+	 QIB_DRV_NAME ": create_singlethread_workqueue failed for port %d\n",
+	 pidx + 1);
+	for (pidx = 0; pidx < dd->num_pports; ++pidx) {
+		ppd = dd->pport + pidx;
+		if (ppd->qib_wq) {
+			destroy_workqueue(ppd->qib_wq);
+			ppd->qib_wq = NULL;
+		}
+	}
+	return -ENOMEM;
+}
+
 /**
  * qib_init - do the actual initialization sequence on the chip
  * @dd: the qlogic_ib device
@@ -764,6 +802,11 @@ static void qib_shutdown_device(struct qib_devdata *dd)
 		 * We can't count on interrupts since we are stopping.
 		 */
 		dd->f_quiet_serdes(ppd);
+
+		if (ppd->qib_wq) {
+			destroy_workqueue(ppd->qib_wq);
+			ppd->qib_wq = NULL;
+		}
 	}
 
 	qib_update_eeprom_log(dd);
@@ -1249,6 +1292,10 @@ static int __devinit qib_init_one(struct pci_dev *pdev,
 	if (ret)
 		goto bail; /* error already printed */
 
+	ret = qib_create_workqueues(dd);
+	if (ret)
+		goto bail;
+
 	/* do the generic initialization */
 	initfail = qib_init(dd, 0);
 
diff --git a/drivers/infiniband/hw/qib/qib_verbs.c b/drivers/infiniband/hw/qib/qib_verbs.c
index 03ace0650a8f..fc9b205c2412 100644
--- a/drivers/infiniband/hw/qib/qib_verbs.c
+++ b/drivers/infiniband/hw/qib/qib_verbs.c
@@ -333,7 +333,8 @@ static void qib_copy_from_sge(void *data, struct qib_sge_state *ss, u32 length)
  * @qp: the QP to post on
  * @wr: the work request to send
  */
-static int qib_post_one_send(struct qib_qp *qp, struct ib_send_wr *wr)
+static int qib_post_one_send(struct qib_qp *qp, struct ib_send_wr *wr,
+	int *scheduled)
 {
 	struct qib_swqe *wqe;
 	u32 next;
@@ -440,6 +441,12 @@ static int qib_post_one_send(struct qib_qp *qp, struct ib_send_wr *wr)
 bail_inval:
 	ret = -EINVAL;
 bail:
+	if (!ret && !wr->next &&
+	 !qib_sdma_empty(
+	   dd_from_ibdev(qp->ibqp.device)->pport + qp->port_num - 1)) {
+		qib_schedule_send(qp);
+		*scheduled = 1;
+	}
 	spin_unlock_irqrestore(&qp->s_lock, flags);
 	return ret;
 }
@@ -457,9 +464,10 @@ static int qib_post_send(struct ib_qp *ibqp, struct ib_send_wr *wr,
 {
 	struct qib_qp *qp = to_iqp(ibqp);
 	int err = 0;
+	int scheduled = 0;
 
 	for (; wr; wr = wr->next) {
-		err = qib_post_one_send(qp, wr);
+		err = qib_post_one_send(qp, wr, &scheduled);
 		if (err) {
 			*bad_wr = wr;
 			goto bail;
@@ -467,7 +475,8 @@ static int qib_post_send(struct ib_qp *ibqp, struct ib_send_wr *wr,
 	}
 
 	/* Try to do the send work in the caller's context. */
-	qib_do_send(&qp->s_work);
+	if (!scheduled)
+		qib_do_send(&qp->s_work);
 
 bail:
 	return err;
@@ -2308,3 +2317,17 @@ void qib_unregister_ib_device(struct qib_devdata *dd)
 		   get_order(lk_tab_size));
 	kfree(dev->qp_table);
 }
+
+/*
+ * This must be called with s_lock held.
+ */
+void qib_schedule_send(struct qib_qp *qp)
+{
+	if (qib_send_ok(qp)) {
+		struct qib_ibport *ibp =
+			to_iport(qp->ibqp.device, qp->port_num);
+		struct qib_pportdata *ppd = ppd_from_ibp(ibp);
+
+		queue_work(ppd->qib_wq, &qp->s_work);
+	}
+}
diff --git a/drivers/infiniband/hw/qib/qib_verbs.h b/drivers/infiniband/hw/qib/qib_verbs.h
index 61fad05328ca..aff8b2c17886 100644
--- a/drivers/infiniband/hw/qib/qib_verbs.h
+++ b/drivers/infiniband/hw/qib/qib_verbs.h
@@ -727,6 +727,7 @@ struct qib_ibport {
 	struct qib_opcode_stats opstats[128];
 };
 
+
 struct qib_ibdev {
 	struct ib_device ibdev;
 	struct list_head pending_mmaps;
@@ -836,11 +837,7 @@ extern struct workqueue_struct *qib_cq_wq;
 /*
  * This must be called with s_lock held.
  */
-static inline void qib_schedule_send(struct qib_qp *qp)
-{
-	if (qib_send_ok(qp))
-		queue_work(ib_wq, &qp->s_work);
-}
+void qib_schedule_send(struct qib_qp *qp);
 
 static inline int qib_pkey_ok(u16 pkey1, u16 pkey2)
 {