Merge tag 'ntb-4.8' of git://github.com/jonmason/ntb

Pull NTB updates from Jon Mason:
 "NTB bug fixes for the ntb_tool and ntb_perf, and improvements to the
  ntb_perf and ntb_pingpong for increased debugability.

  Also, modification to the ntb_transport layer to increase/decrease
  the number of transport entries depending on the ring size"

* tag 'ntb-4.8' of git://github.com/jonmason/ntb:
  NTB: ntb_hw_intel: use local variable pdev
  NTB: ntb_hw_intel: show BAR size in debugfs info
  ntb_test: Add a selftest script for the NTB subsystem
  ntb_perf: clear link_is_up flag when the link goes down.
  ntb_pingpong: Add a debugfs file to get the ping count
  ntb_tool: Add link status and files to debugfs
  ntb_tool: Postpone memory window initialization for the user
  ntb_perf: Wait for link before running test
  ntb_perf: Return results by reading the run file
  ntb_perf: Improve thread handling to increase robustness
  ntb_perf: Schedule based on time not on performance
  ntb_transport: Check the number of spads the hardware supports
  ntb_tool: Add memory window debug support
  ntb_perf: Allow limiting the size of the memory windows
  NTB: allocate number transport entries depending on size of ring size
  ntb_tool: BUG: Ensure the buffer size is large enough to return all spads
  ntb_tool: Fix infinite loop bug when writing spad/peer_spad file

5 files changed, 749 insertions, 99 deletions

diff --git a/drivers/ntb/hw/intel/ntb_hw_intel.c b/drivers/ntb/hw/intel/ntb_hw_intel.c
index 40d04ef5da9e..0d5c29ae51de 100644
--- a/drivers/ntb/hw/intel/ntb_hw_intel.c
+++ b/drivers/ntb/hw/intel/ntb_hw_intel.c
@@ -551,13 +551,15 @@ static ssize_t ndev_debugfs_read(struct file *filp, char __user *ubuf,
 				 size_t count, loff_t *offp)
 {
 	struct intel_ntb_dev *ndev;
+	struct pci_dev *pdev;
 	void __iomem *mmio;
 	char *buf;
 	size_t buf_size;
 	ssize_t ret, off;
-	union { u64 v64; u32 v32; u16 v16; } u;
+	union { u64 v64; u32 v32; u16 v16; u8 v8; } u;
 
 	ndev = filp->private_data;
+	pdev = ndev_pdev(ndev);
 	mmio = ndev->self_mmio;
 
 	buf_size = min(count, 0x800ul);
@@ -632,6 +634,41 @@ static ssize_t ndev_debugfs_read(struct file *filp, char __user *ubuf,
 			 "Doorbell Bell -\t\t%#llx\n", u.v64);
 
 	off += scnprintf(buf + off, buf_size - off,
+			 "\nNTB Window Size:\n");
+
+	pci_read_config_byte(pdev, XEON_PBAR23SZ_OFFSET, &u.v8);
+	off += scnprintf(buf + off, buf_size - off,
+			 "PBAR23SZ %hhu\n", u.v8);
+	if (!ndev->bar4_split) {
+		pci_read_config_byte(pdev, XEON_PBAR45SZ_OFFSET, &u.v8);
+		off += scnprintf(buf + off, buf_size - off,
+				 "PBAR45SZ %hhu\n", u.v8);
+	} else {
+		pci_read_config_byte(pdev, XEON_PBAR4SZ_OFFSET, &u.v8);
+		off += scnprintf(buf + off, buf_size - off,
+				 "PBAR4SZ %hhu\n", u.v8);
+		pci_read_config_byte(pdev, XEON_PBAR5SZ_OFFSET, &u.v8);
+		off += scnprintf(buf + off, buf_size - off,
+				 "PBAR5SZ %hhu\n", u.v8);
+	}
+
+	pci_read_config_byte(pdev, XEON_SBAR23SZ_OFFSET, &u.v8);
+	off += scnprintf(buf + off, buf_size - off,
+			 "SBAR23SZ %hhu\n", u.v8);
+	if (!ndev->bar4_split) {
+		pci_read_config_byte(pdev, XEON_SBAR45SZ_OFFSET, &u.v8);
+		off += scnprintf(buf + off, buf_size - off,
+				 "SBAR45SZ %hhu\n", u.v8);
+	} else {
+		pci_read_config_byte(pdev, XEON_SBAR4SZ_OFFSET, &u.v8);
+		off += scnprintf(buf + off, buf_size - off,
+				 "SBAR4SZ %hhu\n", u.v8);
+		pci_read_config_byte(pdev, XEON_SBAR5SZ_OFFSET, &u.v8);
+		off += scnprintf(buf + off, buf_size - off,
+				 "SBAR5SZ %hhu\n", u.v8);
+	}
+
+	off += scnprintf(buf + off, buf_size - off,
 			 "\nNTB Incoming XLAT:\n");
 
 	u.v64 = ioread64(mmio + bar2_off(ndev->xlat_reg->bar2_xlat, 2));
@@ -669,7 +706,7 @@ static ssize_t ndev_debugfs_read(struct file *filp, char __user *ubuf,
 				 "LMT45 -\t\t\t%#018llx\n", u.v64);
 	}
 
-	if (pdev_is_xeon(ndev->ntb.pdev)) {
+	if (pdev_is_xeon(pdev)) {
 		if (ntb_topo_is_b2b(ndev->ntb.topo)) {
 			off += scnprintf(buf + off, buf_size - off,
 					 "\nNTB Outgoing B2B XLAT:\n");
@@ -750,22 +787,22 @@ static ssize_t ndev_debugfs_read(struct file *filp, char __user *ubuf,
 		off += scnprintf(buf + off, buf_size - off,
 				 "\nXEON NTB Hardware Errors:\n");
 
-		if (!pci_read_config_word(ndev->ntb.pdev,
+		if (!pci_read_config_word(pdev,
 					  XEON_DEVSTS_OFFSET, &u.v16))
 			off += scnprintf(buf + off, buf_size - off,
 					 "DEVSTS -\t\t%#06x\n", u.v16);
 
-		if (!pci_read_config_word(ndev->ntb.pdev,
+		if (!pci_read_config_word(pdev,
 					  XEON_LINK_STATUS_OFFSET, &u.v16))
 			off += scnprintf(buf + off, buf_size - off,
 					 "LNKSTS -\t\t%#06x\n", u.v16);
 
-		if (!pci_read_config_dword(ndev->ntb.pdev,
+		if (!pci_read_config_dword(pdev,
 					   XEON_UNCERRSTS_OFFSET, &u.v32))
 			off += scnprintf(buf + off, buf_size - off,
 					 "UNCERRSTS -\t\t%#06x\n", u.v32);
 
-		if (!pci_read_config_dword(ndev->ntb.pdev,
+		if (!pci_read_config_dword(pdev,
 					   XEON_CORERRSTS_OFFSET, &u.v32))
 			off += scnprintf(buf + off, buf_size - off,
 					 "CORERRSTS -\t\t%#06x\n", u.v32);
diff --git a/drivers/ntb/ntb_transport.c b/drivers/ntb/ntb_transport.c
index 2ef9d9130864..d5c5894f252e 100644
--- a/drivers/ntb/ntb_transport.c
+++ b/drivers/ntb/ntb_transport.c
@@ -153,6 +153,7 @@ struct ntb_transport_qp {
 	unsigned int rx_index;
 	unsigned int rx_max_entry;
 	unsigned int rx_max_frame;
+	unsigned int rx_alloc_entry;
 	dma_cookie_t last_cookie;
 	struct tasklet_struct rxc_db_work;
 
@@ -480,7 +481,9 @@ static ssize_t debugfs_read(struct file *filp, char __user *ubuf, size_t count,
 	out_offset += snprintf(buf + out_offset, out_count - out_offset,
 			       "rx_index - \t%u\n", qp->rx_index);
 	out_offset += snprintf(buf + out_offset, out_count - out_offset,
-			       "rx_max_entry - \t%u\n\n", qp->rx_max_entry);
+			       "rx_max_entry - \t%u\n", qp->rx_max_entry);
+	out_offset += snprintf(buf + out_offset, out_count - out_offset,
+			       "rx_alloc_entry - \t%u\n\n", qp->rx_alloc_entry);
 
 	out_offset += snprintf(buf + out_offset, out_count - out_offset,
 			       "tx_bytes - \t%llu\n", qp->tx_bytes);
@@ -597,9 +600,12 @@ static int ntb_transport_setup_qp_mw(struct ntb_transport_ctx *nt,
 {
 	struct ntb_transport_qp *qp = &nt->qp_vec[qp_num];
 	struct ntb_transport_mw *mw;
+	struct ntb_dev *ndev = nt->ndev;
+	struct ntb_queue_entry *entry;
 	unsigned int rx_size, num_qps_mw;
 	unsigned int mw_num, mw_count, qp_count;
 	unsigned int i;
+	int node;
 
 	mw_count = nt->mw_count;
 	qp_count = nt->qp_count;
@@ -626,6 +632,23 @@ static int ntb_transport_setup_qp_mw(struct ntb_transport_ctx *nt,
 	qp->rx_max_entry = rx_size / qp->rx_max_frame;
 	qp->rx_index = 0;
 
+	/*
+	 * Checking to see if we have more entries than the default.
+	 * We should add additional entries if that is the case so we
+	 * can be in sync with the transport frames.
+	 */
+	node = dev_to_node(&ndev->dev);
+	for (i = qp->rx_alloc_entry; i < qp->rx_max_entry; i++) {
+		entry = kzalloc_node(sizeof(*entry), GFP_ATOMIC, node);
+		if (!entry)
+			return -ENOMEM;
+
+		entry->qp = qp;
+		ntb_list_add(&qp->ntb_rx_q_lock, &entry->entry,
+			     &qp->rx_free_q);
+		qp->rx_alloc_entry++;
+	}
+
 	qp->remote_rx_info->entry = qp->rx_max_entry - 1;
 
 	/* setup the hdr offsets with 0's */
@@ -1037,6 +1060,13 @@ static int ntb_transport_probe(struct ntb_client *self, struct ntb_dev *ndev)
 	int node;
 	int rc, i;
 
+	mw_count = ntb_mw_count(ndev);
+	if (ntb_spad_count(ndev) < (NUM_MWS + 1 + mw_count * 2)) {
+		dev_err(&ndev->dev, "Not enough scratch pad registers for %s",
+			NTB_TRANSPORT_NAME);
+		return -EIO;
+	}
+
 	if (ntb_db_is_unsafe(ndev))
 		dev_dbg(&ndev->dev,
 			"doorbell is unsafe, proceed anyway...\n");
@@ -1052,8 +1082,6 @@ static int ntb_transport_probe(struct ntb_client *self, struct ntb_dev *ndev)
 
 	nt->ndev = ndev;
 
-	mw_count = ntb_mw_count(ndev);
-
 	nt->mw_count = mw_count;
 
 	nt->mw_vec = kzalloc_node(mw_count * sizeof(*nt->mw_vec),
@@ -1722,8 +1750,9 @@ ntb_transport_create_queue(void *data, struct device *client_dev,
 		ntb_list_add(&qp->ntb_rx_q_lock, &entry->entry,
 			     &qp->rx_free_q);
 	}
+	qp->rx_alloc_entry = NTB_QP_DEF_NUM_ENTRIES;
 
-	for (i = 0; i < NTB_QP_DEF_NUM_ENTRIES; i++) {
+	for (i = 0; i < qp->tx_max_entry; i++) {
 		entry = kzalloc_node(sizeof(*entry), GFP_ATOMIC, node);
 		if (!entry)
 			goto err2;
@@ -1744,6 +1773,7 @@ err2:
 	while ((entry = ntb_list_rm(&qp->ntb_tx_free_q_lock, &qp->tx_free_q)))
 		kfree(entry);
 err1:
+	qp->rx_alloc_entry = 0;
 	while ((entry = ntb_list_rm(&qp->ntb_rx_q_lock, &qp->rx_free_q)))
 		kfree(entry);
 	if (qp->tx_dma_chan)
diff --git a/drivers/ntb/test/ntb_perf.c b/drivers/ntb/test/ntb_perf.c
index 8dfce9c9aad0..6a50f20bf1cd 100644
--- a/drivers/ntb/test/ntb_perf.c
+++ b/drivers/ntb/test/ntb_perf.c
@@ -58,6 +58,7 @@
 #include <linux/delay.h>
 #include <linux/sizes.h>
 #include <linux/ntb.h>
+#include <linux/mutex.h>
 
 #define DRIVER_NAME		"ntb_perf"
 #define DRIVER_DESCRIPTION	"PCIe NTB Performance Measurement Tool"
@@ -83,6 +84,10 @@ MODULE_DESCRIPTION(DRIVER_DESCRIPTION);
 
 static struct dentry *perf_debugfs_dir;
 
+static unsigned long max_mw_size;
+module_param(max_mw_size, ulong, 0644);
+MODULE_PARM_DESC(max_mw_size, "Limit size of large memory windows");
+
 static unsigned int seg_order = 19; /* 512K */
 module_param(seg_order, uint, 0644);
 MODULE_PARM_DESC(seg_order, "size order [n^2] of buffer segment for testing");
@@ -117,6 +122,10 @@ struct pthr_ctx {
 	int			dma_prep_err;
 	int			src_idx;
 	void			*srcs[MAX_SRCS];
+	wait_queue_head_t       *wq;
+	int			status;
+	u64			copied;
+	u64			diff_us;
 };
 
 struct perf_ctx {
@@ -124,23 +133,23 @@ struct perf_ctx {
 	spinlock_t		db_lock;
 	struct perf_mw		mw;
 	bool			link_is_up;
-	struct work_struct	link_cleanup;
 	struct delayed_work	link_work;
+	wait_queue_head_t	link_wq;
 	struct dentry		*debugfs_node_dir;
 	struct dentry		*debugfs_run;
 	struct dentry		*debugfs_threads;
 	u8			perf_threads;
-	bool			run;
+	/* mutex ensures only one set of threads run at once */
+	struct mutex		run_mutex;
 	struct pthr_ctx		pthr_ctx[MAX_THREADS];
 	atomic_t		tsync;
+	atomic_t                tdone;
 };
 
 enum {
 	VERSION = 0,
 	MW_SZ_HIGH,
 	MW_SZ_LOW,
-	SPAD_MSG,
-	SPAD_ACK,
 	MAX_SPAD
 };
 
@@ -148,10 +157,16 @@ static void perf_link_event(void *ctx)
 {
 	struct perf_ctx *perf = ctx;
 
-	if (ntb_link_is_up(perf->ntb, NULL, NULL) == 1)
+	if (ntb_link_is_up(perf->ntb, NULL, NULL) == 1) {
 		schedule_delayed_work(&perf->link_work, 2*HZ);
-	else
-		schedule_work(&perf->link_cleanup);
+	} else {
+		dev_dbg(&perf->ntb->pdev->dev, "link down\n");
+
+		if (!perf->link_is_up)
+			cancel_delayed_work_sync(&perf->link_work);
+
+		perf->link_is_up = false;
+	}
 }
 
 static void perf_db_event(void *ctx, int vec)
@@ -271,6 +286,7 @@ static int perf_move_data(struct pthr_ctx *pctx, char __iomem *dst, char *src,
 	char __iomem *tmp = dst;
 	u64 perf, diff_us;
 	ktime_t kstart, kstop, kdiff;
+	unsigned long last_sleep = jiffies;
 
 	chunks = div64_u64(win_size, buf_size);
 	total_chunks = div64_u64(total, buf_size);
@@ -286,30 +302,40 @@ static int perf_move_data(struct pthr_ctx *pctx, char __iomem *dst, char *src,
 		} else
 			tmp += buf_size;
 
-		/* Probably should schedule every 4GB to prevent soft hang. */
-		if (((copied % SZ_4G) == 0) && !use_dma) {
+		/* Probably should schedule every 5s to prevent soft hang. */
+		if (unlikely((jiffies - last_sleep) > 5 * HZ)) {
+			last_sleep = jiffies;
 			set_current_state(TASK_INTERRUPTIBLE);
 			schedule_timeout(1);
 		}
+
+		if (unlikely(kthread_should_stop()))
+			break;
 	}
 
 	if (use_dma) {
-		pr_info("%s: All DMA descriptors submitted\n", current->comm);
-		while (atomic_read(&pctx->dma_sync) != 0)
+		pr_debug("%s: All DMA descriptors submitted\n", current->comm);
+		while (atomic_read(&pctx->dma_sync) != 0) {
+			if (kthread_should_stop())
+				break;
 			msleep(20);
+		}
 	}
 
 	kstop = ktime_get();
 	kdiff = ktime_sub(kstop, kstart);
 	diff_us = ktime_to_us(kdiff);
 
-	pr_info("%s: copied %llu bytes\n", current->comm, copied);
+	pr_debug("%s: copied %llu bytes\n", current->comm, copied);
 
-	pr_info("%s: lasted %llu usecs\n", current->comm, diff_us);
+	pr_debug("%s: lasted %llu usecs\n", current->comm, diff_us);
 
 	perf = div64_u64(copied, diff_us);
 
-	pr_info("%s: MBytes/s: %llu\n", current->comm, perf);
+	pr_debug("%s: MBytes/s: %llu\n", current->comm, perf);
+
+	pctx->copied = copied;
+	pctx->diff_us = diff_us;
 
 	return 0;
 }
@@ -331,7 +357,7 @@ static int ntb_perf_thread(void *data)
 	int rc, node, i;
 	struct dma_chan *dma_chan = NULL;
 
-	pr_info("kthread %s starting...\n", current->comm);
+	pr_debug("kthread %s starting...\n", current->comm);
 
 	node = dev_to_node(&pdev->dev);
 
@@ -389,7 +415,10 @@ static int ntb_perf_thread(void *data)
 		pctx->srcs[i] = NULL;
 	}
 
-	return 0;
+	atomic_inc(&perf->tdone);
+	wake_up(pctx->wq);
+	rc = 0;
+	goto done;
 
 err:
 	for (i = 0; i < MAX_SRCS; i++) {
@@ -402,6 +431,16 @@ err:
 		pctx->dma_chan = NULL;
 	}
 
+done:
+	/* Wait until we are told to stop */
+	for (;;) {
+		set_current_state(TASK_INTERRUPTIBLE);
+		if (kthread_should_stop())
+			break;
+		schedule();
+	}
+	__set_current_state(TASK_RUNNING);
+
 	return rc;
 }
 
@@ -472,6 +511,10 @@ static void perf_link_work(struct work_struct *work)
 	dev_dbg(&perf->ntb->pdev->dev, "%s called\n", __func__);
 
 	size = perf->mw.phys_size;
+
+	if (max_mw_size && size > max_mw_size)
+		size = max_mw_size;
+
 	ntb_peer_spad_write(ndev, MW_SZ_HIGH, upper_32_bits(size));
 	ntb_peer_spad_write(ndev, MW_SZ_LOW, lower_32_bits(size));
 	ntb_peer_spad_write(ndev, VERSION, PERF_VERSION);
@@ -496,6 +539,7 @@ static void perf_link_work(struct work_struct *work)
 		goto out1;
 
 	perf->link_is_up = true;
+	wake_up(&perf->link_wq);
 
 	return;
 
@@ -508,18 +552,6 @@ out:
 				      msecs_to_jiffies(PERF_LINK_DOWN_TIMEOUT));
 }
 
-static void perf_link_cleanup(struct work_struct *work)
-{
-	struct perf_ctx *perf = container_of(work,
-					     struct perf_ctx,
-					     link_cleanup);
-
-	dev_dbg(&perf->ntb->pdev->dev, "%s called\n", __func__);
-
-	if (!perf->link_is_up)
-		cancel_delayed_work_sync(&perf->link_work);
-}
-
 static int perf_setup_mw(struct ntb_dev *ntb, struct perf_ctx *perf)
 {
 	struct perf_mw *mw;
@@ -544,16 +576,44 @@ static ssize_t debugfs_run_read(struct file *filp, char __user *ubuf,
 {
 	struct perf_ctx *perf = filp->private_data;
 	char *buf;
-	ssize_t ret, out_offset;
+	ssize_t ret, out_off = 0;
+	struct pthr_ctx *pctx;
+	int i;
+	u64 rate;
 
 	if (!perf)
 		return 0;
 
-	buf = kmalloc(64, GFP_KERNEL);
+	buf = kmalloc(1024, GFP_KERNEL);
 	if (!buf)
 		return -ENOMEM;
-	out_offset = snprintf(buf, 64, "%d\n", perf->run);
-	ret = simple_read_from_buffer(ubuf, count, offp, buf, out_offset);
+
+	if (mutex_is_locked(&perf->run_mutex)) {
+		out_off = snprintf(buf, 64, "running\n");
+		goto read_from_buf;
+	}
+
+	for (i = 0; i < MAX_THREADS; i++) {
+		pctx = &perf->pthr_ctx[i];
+
+		if (pctx->status == -ENODATA)
+			break;
+
+		if (pctx->status) {
+			out_off += snprintf(buf + out_off, 1024 - out_off,
+					    "%d: error %d\n", i,
+					    pctx->status);
+			continue;
+		}
+
+		rate = div64_u64(pctx->copied, pctx->diff_us);
+		out_off += snprintf(buf + out_off, 1024 - out_off,
+			"%d: copied %llu bytes in %llu usecs, %llu MBytes/s\n",
+			i, pctx->copied, pctx->diff_us, rate);
+	}
+
+read_from_buf:
+	ret = simple_read_from_buffer(ubuf, count, offp, buf, out_off);
 	kfree(buf);
 
 	return ret;
@@ -564,80 +624,90 @@ static void threads_cleanup(struct perf_ctx *perf)
 	struct pthr_ctx *pctx;
 	int i;
 
-	perf->run = false;
 	for (i = 0; i < MAX_THREADS; i++) {
 		pctx = &perf->pthr_ctx[i];
 		if (pctx->thread) {
-			kthread_stop(pctx->thread);
+			pctx->status = kthread_stop(pctx->thread);
 			pctx->thread = NULL;
 		}
 	}
 }
 
+static void perf_clear_thread_status(struct perf_ctx *perf)
+{
+	int i;
+
+	for (i = 0; i < MAX_THREADS; i++)
+		perf->pthr_ctx[i].status = -ENODATA;
+}
+
 static ssize_t debugfs_run_write(struct file *filp, const char __user *ubuf,
 				 size_t count, loff_t *offp)
 {
 	struct perf_ctx *perf = filp->private_data;
 	int node, i;
+	DECLARE_WAIT_QUEUE_HEAD(wq);
 
-	if (!perf->link_is_up)
-		return 0;
+	if (wait_event_interruptible(perf->link_wq, perf->link_is_up))
+		return -ENOLINK;
 
 	if (perf->perf_threads == 0)
-		return 0;
+		return -EINVAL;
 
-	if (atomic_read(&perf->tsync) == 0)
-		perf->run = false;
+	if (!mutex_trylock(&perf->run_mutex))
+		return -EBUSY;
 
-	if (perf->run)
-		threads_cleanup(perf);
-	else {
-		perf->run = true;
+	perf_clear_thread_status(perf);
 
-		if (perf->perf_threads > MAX_THREADS) {
-			perf->perf_threads = MAX_THREADS;
-			pr_info("Reset total threads to: %u\n", MAX_THREADS);
-		}
+	if (perf->perf_threads > MAX_THREADS) {
+		perf->perf_threads = MAX_THREADS;
+		pr_info("Reset total threads to: %u\n", MAX_THREADS);
+	}
 
-		/* no greater than 1M */
-		if (seg_order > MAX_SEG_ORDER) {
-			seg_order = MAX_SEG_ORDER;
-			pr_info("Fix seg_order to %u\n", seg_order);
-		}
+	/* no greater than 1M */
+	if (seg_order > MAX_SEG_ORDER) {
+		seg_order = MAX_SEG_ORDER;
+		pr_info("Fix seg_order to %u\n", seg_order);
+	}
 
-		if (run_order < seg_order) {
-			run_order = seg_order;
-			pr_info("Fix run_order to %u\n", run_order);
-		}
+	if (run_order < seg_order) {
+		run_order = seg_order;
+		pr_info("Fix run_order to %u\n", run_order);
+	}
 
-		node = dev_to_node(&perf->ntb->pdev->dev);
-		/* launch kernel thread */
-		for (i = 0; i < perf->perf_threads; i++) {
-			struct pthr_ctx *pctx;
-
-			pctx = &perf->pthr_ctx[i];
-			atomic_set(&pctx->dma_sync, 0);
-			pctx->perf = perf;
-			pctx->thread =
-				kthread_create_on_node(ntb_perf_thread,
-						       (void *)pctx,
-						       node, "ntb_perf %d", i);
-			if (IS_ERR(pctx->thread)) {
-				pctx->thread = NULL;
-				goto err;
-			} else
-				wake_up_process(pctx->thread);
-
-			if (perf->run == false)
-				return -ENXIO;
-		}
+	node = dev_to_node(&perf->ntb->pdev->dev);
+	atomic_set(&perf->tdone, 0);
 
+	/* launch kernel thread */
+	for (i = 0; i < perf->perf_threads; i++) {
+		struct pthr_ctx *pctx;
+
+		pctx = &perf->pthr_ctx[i];
+		atomic_set(&pctx->dma_sync, 0);
+		pctx->perf = perf;
+		pctx->wq = &wq;
+		pctx->thread =
+			kthread_create_on_node(ntb_perf_thread,
+					       (void *)pctx,
+					       node, "ntb_perf %d", i);
+		if (IS_ERR(pctx->thread)) {
+			pctx->thread = NULL;
+			goto err;
+		} else {
+			wake_up_process(pctx->thread);
+		}
 	}
 
+	wait_event_interruptible(wq,
+		atomic_read(&perf->tdone) == perf->perf_threads);
+
+	threads_cleanup(perf);
+	mutex_unlock(&perf->run_mutex);
 	return count;
 
 err:
 	threads_cleanup(perf);
+	mutex_unlock(&perf->run_mutex);
 	return -ENXIO;
 }
 
@@ -688,6 +758,12 @@ static int perf_probe(struct ntb_client *client, struct ntb_dev *ntb)
 	int node;
 	int rc = 0;
 
+	if (ntb_spad_count(ntb) < MAX_SPAD) {
+		dev_err(&ntb->dev, "Not enough scratch pad registers for %s",
+			DRIVER_NAME);
+		return -EIO;
+	}
+
 	node = dev_to_node(&pdev->dev);
 
 	perf = kzalloc_node(sizeof(*perf), GFP_KERNEL, node);
@@ -699,11 +775,11 @@ static int perf_probe(struct ntb_client *client, struct ntb_dev *ntb)
 	perf->ntb = ntb;
 	perf->perf_threads = 1;
 	atomic_set(&perf->tsync, 0);
-	perf->run = false;
+	mutex_init(&perf->run_mutex);
 	spin_lock_init(&perf->db_lock);
 	perf_setup_mw(ntb, perf);
+	init_waitqueue_head(&perf->link_wq);
 	INIT_DELAYED_WORK(&perf->link_work, perf_link_work);
-	INIT_WORK(&perf->link_cleanup, perf_link_cleanup);
 
 	rc = ntb_set_ctx(ntb, perf, &perf_ops);
 	if (rc)
@@ -717,11 +793,12 @@ static int perf_probe(struct ntb_client *client, struct ntb_dev *ntb)
 	if (rc)
 		goto err_ctx;
 
+	perf_clear_thread_status(perf);
+
 	return 0;
 
 err_ctx:
 	cancel_delayed_work_sync(&perf->link_work);
-	cancel_work_sync(&perf->link_cleanup);
 	kfree(perf);
 err_perf:
 	return rc;
@@ -734,8 +811,9 @@ static void perf_remove(struct ntb_client *client, struct ntb_dev *ntb)
 
 	dev_dbg(&perf->ntb->dev, "%s called\n", __func__);
 
+	mutex_lock(&perf->run_mutex);
+
 	cancel_delayed_work_sync(&perf->link_work);
-	cancel_work_sync(&perf->link_cleanup);
 
 	ntb_clear_ctx(ntb);
 	ntb_link_disable(ntb);
diff --git a/drivers/ntb/test/ntb_pingpong.c b/drivers/ntb/test/ntb_pingpong.c
index fe1600566981..7d311799fca1 100644
--- a/drivers/ntb/test/ntb_pingpong.c
+++ b/drivers/ntb/test/ntb_pingpong.c
@@ -61,6 +61,7 @@
 #include <linux/pci.h>
 #include <linux/slab.h>
 #include <linux/spinlock.h>
+#include <linux/debugfs.h>
 
 #include <linux/ntb.h>
 
@@ -96,8 +97,13 @@ struct pp_ctx {
 	spinlock_t			db_lock;
 	struct timer_list		db_timer;
 	unsigned long			db_delay;
+	struct dentry			*debugfs_node_dir;
+	struct dentry			*debugfs_count;
+	atomic_t			count;
 };
 
+static struct dentry *pp_debugfs_dir;
+
 static void pp_ping(unsigned long ctx)
 {
 	struct pp_ctx *pp = (void *)ctx;
@@ -171,10 +177,32 @@ static void pp_db_event(void *ctx, int vec)
 		dev_dbg(&pp->ntb->dev,
 			"Pong vec %d bits %#llx\n",
 			vec, db_bits);
+		atomic_inc(&pp->count);
 	}
 	spin_unlock_irqrestore(&pp->db_lock, irqflags);
 }
 
+static int pp_debugfs_setup(struct pp_ctx *pp)
+{
+	struct pci_dev *pdev = pp->ntb->pdev;
+
+	if (!pp_debugfs_dir)
+		return -ENODEV;
+
+	pp->debugfs_node_dir = debugfs_create_dir(pci_name(pdev),
+						  pp_debugfs_dir);
+	if (!pp->debugfs_node_dir)
+		return -ENODEV;
+
+	pp->debugfs_count = debugfs_create_atomic_t("count", S_IRUSR | S_IWUSR,
+						    pp->debugfs_node_dir,
+						    &pp->count);
+	if (!pp->debugfs_count)
+		return -ENODEV;
+
+	return 0;
+}
+
 static const struct ntb_ctx_ops pp_ops = {
 	.link_event = pp_link_event,
 	.db_event = pp_db_event,
@@ -210,6 +238,7 @@ static int pp_probe(struct ntb_client *client,
 
 	pp->ntb = ntb;
 	pp->db_bits = 0;
+	atomic_set(&pp->count, 0);
 	spin_lock_init(&pp->db_lock);
 	setup_timer(&pp->db_timer, pp_ping, (unsigned long)pp);
 	pp->db_delay = msecs_to_jiffies(delay_ms);
@@ -218,6 +247,10 @@ static int pp_probe(struct ntb_client *client,
 	if (rc)
 		goto err_ctx;
 
+	rc = pp_debugfs_setup(pp);
+	if (rc)
+		goto err_ctx;
+
 	ntb_link_enable(ntb, NTB_SPEED_AUTO, NTB_WIDTH_AUTO);
 	ntb_link_event(ntb);
 
@@ -234,6 +267,8 @@ static void pp_remove(struct ntb_client *client,
 {
 	struct pp_ctx *pp = ntb->ctx;
 
+	debugfs_remove_recursive(pp->debugfs_node_dir);
+
 	ntb_clear_ctx(ntb);
 	del_timer_sync(&pp->db_timer);
 	ntb_link_disable(ntb);
@@ -247,4 +282,29 @@ static struct ntb_client pp_client = {
 		.remove = pp_remove,
 	},
 };
-module_ntb_client(pp_client);
+
+static int __init pp_init(void)
+{
+	int rc;
+
+	if (debugfs_initialized())
+		pp_debugfs_dir = debugfs_create_dir(KBUILD_MODNAME, NULL);
+
+	rc = ntb_register_client(&pp_client);
+	if (rc)
+		goto err_client;
+
+	return 0;
+
+err_client:
+	debugfs_remove_recursive(pp_debugfs_dir);
+	return rc;
+}
+module_init(pp_init);
+
+static void __exit pp_exit(void)
+{
+	ntb_unregister_client(&pp_client);
+	debugfs_remove_recursive(pp_debugfs_dir);
+}
+module_exit(pp_exit);
diff --git a/drivers/ntb/test/ntb_tool.c b/drivers/ntb/test/ntb_tool.c
index 6f5dc6ca673d..61bf2ef87e0e 100644
--- a/drivers/ntb/test/ntb_tool.c
+++ b/drivers/ntb/test/ntb_tool.c
@@ -59,6 +59,12 @@
  *
  * Eg: check if clearing the doorbell mask generates an interrupt.
  *
+ * # Check the link status
+ * root@self# cat $DBG_DIR/link
+ *
+ * # Block until the link is up
+ * root@self# echo Y > $DBG_DIR/link_event
+ *
  * # Set the doorbell mask
  * root@self# echo 's 1' > $DBG_DIR/mask
  *
@@ -79,6 +85,13 @@
  * root@self# cat $DBG_DIR/spad
  *
  * Observe that spad 0 and 1 have the values set by the peer.
+ *
+ * # Check the memory window translation info
+ * cat $DBG_DIR/peer_trans0
+ *
+ * # Setup a 16k memory window buffer
+ * echo 16384 > $DBG_DIR/peer_trans0
+ *
  */
 
 #include <linux/init.h>
@@ -89,6 +102,7 @@
 #include <linux/dma-mapping.h>
 #include <linux/pci.h>
 #include <linux/slab.h>
+#include <linux/uaccess.h>
 
 #include <linux/ntb.h>
 
@@ -105,11 +119,27 @@ MODULE_VERSION(DRIVER_VERSION);
 MODULE_AUTHOR(DRIVER_AUTHOR);
 MODULE_DESCRIPTION(DRIVER_DESCRIPTION);
 
+#define MAX_MWS 16
+
 static struct dentry *tool_dbgfs;
 
+struct tool_mw {
+	int idx;
+	struct tool_ctx *tc;
+	resource_size_t win_size;
+	resource_size_t size;
+	u8 __iomem *local;
+	u8 *peer;
+	dma_addr_t peer_dma;
+	struct dentry *peer_dbg_file;
+};
+
 struct tool_ctx {
 	struct ntb_dev *ntb;
 	struct dentry *dbgfs;
+	wait_queue_head_t link_wq;
+	int mw_count;
+	struct tool_mw mws[MAX_MWS];
 };
 
 #define SPAD_FNAME_SIZE 0x10
@@ -135,6 +165,8 @@ static void tool_link_event(void *ctx)
 
 	dev_dbg(&tc->ntb->dev, "link is %s speed %d width %d\n",
 		up ? "up" : "down", speed, width);
+
+	wake_up(&tc->link_wq);
 }
 
 static void tool_db_event(void *ctx, int vec)
@@ -239,7 +271,14 @@ static ssize_t tool_spadfn_read(struct tool_ctx *tc, char __user *ubuf,
 	if (!spad_read_fn)
 		return -EINVAL;
 
-	buf_size = min_t(size_t, size, 0x100);
+	spad_count = ntb_spad_count(tc->ntb);
+
+	/*
+	 * We multiply the number of spads by 15 to get the buffer size
+	 * this is from 3 for the %d, 10 for the largest hex value
+	 * (0x00000000) and 2 for the tab and line feed.
+	 */
+	buf_size = min_t(size_t, size, spad_count * 15);
 
 	buf = kmalloc(buf_size, GFP_KERNEL);
 	if (!buf)
@@ -247,7 +286,6 @@ static ssize_t tool_spadfn_read(struct tool_ctx *tc, char __user *ubuf,
 
 	pos = 0;
 
-	spad_count = ntb_spad_count(tc->ntb);
 	for (i = 0; i < spad_count; ++i) {
 		pos += scnprintf(buf + pos, buf_size - pos, "%d\t%#x\n",
 				 i, spad_read_fn(tc->ntb, i));
@@ -268,7 +306,7 @@ static ssize_t tool_spadfn_write(struct tool_ctx *tc,
 {
 	int spad_idx;
 	u32 spad_val;
-	char *buf;
+	char *buf, *buf_ptr;
 	int pos, n;
 	ssize_t rc;
 
@@ -288,14 +326,15 @@ static ssize_t tool_spadfn_write(struct tool_ctx *tc,
 	}
 
 	buf[size] = 0;
-
-	n = sscanf(buf, "%d %i%n", &spad_idx, &spad_val, &pos);
+	buf_ptr = buf;
+	n = sscanf(buf_ptr, "%d %i%n", &spad_idx, &spad_val, &pos);
 	while (n == 2) {
+		buf_ptr += pos;
 		rc = spad_write_fn(tc->ntb, spad_idx, spad_val);
 		if (rc)
 			break;
 
-		n = sscanf(buf + pos, "%d %i%n", &spad_idx, &spad_val, &pos);
+		n = sscanf(buf_ptr, "%d %i%n", &spad_idx, &spad_val, &pos);
 	}
 
 	if (n < 0)
@@ -442,8 +481,384 @@ static TOOL_FOPS_RDWR(tool_peer_spad_fops,
 		      tool_peer_spad_read,
 		      tool_peer_spad_write);
 
+static ssize_t tool_link_read(struct file *filep, char __user *ubuf,
+			      size_t size, loff_t *offp)
+{
+	struct tool_ctx *tc = filep->private_data;
+	char buf[3];
+
+	buf[0] = ntb_link_is_up(tc->ntb, NULL, NULL) ? 'Y' : 'N';
+	buf[1] = '\n';
+	buf[2] = '\0';
+
+	return simple_read_from_buffer(ubuf, size, offp, buf, 2);
+}
+
+static ssize_t tool_link_write(struct file *filep, const char __user *ubuf,
+			       size_t size, loff_t *offp)
+{
+	struct tool_ctx *tc = filep->private_data;
+	char buf[32];
+	size_t buf_size;
+	bool val;
+	int rc;
+
+	buf_size = min(size, (sizeof(buf) - 1));
+	if (copy_from_user(buf, ubuf, buf_size))
+		return -EFAULT;
+
+	buf[buf_size] = '\0';
+
+	rc = strtobool(buf, &val);
+	if (rc)
+		return rc;
+
+	if (val)
+		rc = ntb_link_enable(tc->ntb, NTB_SPEED_AUTO, NTB_WIDTH_AUTO);
+	else
+		rc = ntb_link_disable(tc->ntb);
+
+	if (rc)
+		return rc;
+
+	return size;
+}
+
+static TOOL_FOPS_RDWR(tool_link_fops,
+		      tool_link_read,
+		      tool_link_write);
+
+static ssize_t tool_link_event_write(struct file *filep,
+				     const char __user *ubuf,
+				     size_t size, loff_t *offp)
+{
+	struct tool_ctx *tc = filep->private_data;
+	char buf[32];
+	size_t buf_size;
+	bool val;
+	int rc;
+
+	buf_size = min(size, (sizeof(buf) - 1));
+	if (copy_from_user(buf, ubuf, buf_size))
+		return -EFAULT;
+
+	buf[buf_size] = '\0';
+
+	rc = strtobool(buf, &val);
+	if (rc)
+		return rc;
+
+	if (wait_event_interruptible(tc->link_wq,
+		ntb_link_is_up(tc->ntb, NULL, NULL) == val))
+		return -ERESTART;
+
+	return size;
+}
+
+static TOOL_FOPS_RDWR(tool_link_event_fops,
+		      NULL,
+		      tool_link_event_write);
+
+static ssize_t tool_mw_read(struct file *filep, char __user *ubuf,
+			    size_t size, loff_t *offp)
+{
+	struct tool_mw *mw = filep->private_data;
+	ssize_t rc;
+	loff_t pos = *offp;
+	void *buf;
+
+	if (mw->local == NULL)
+		return -EIO;
+	if (pos < 0)
+		return -EINVAL;
+	if (pos >= mw->win_size || !size)
+		return 0;
+	if (size > mw->win_size - pos)
+		size = mw->win_size - pos;
+
+	buf = kmalloc(size, GFP_KERNEL);
+	if (!buf)
+		return -ENOMEM;
+
+	memcpy_fromio(buf, mw->local + pos, size);
+	rc = copy_to_user(ubuf, buf, size);
+	if (rc == size) {
+		rc = -EFAULT;
+		goto err_free;
+	}
+
+	size -= rc;
+	*offp = pos + size;
+	rc = size;
+
+err_free:
+	kfree(buf);
+
+	return rc;
+}
+
+static ssize_t tool_mw_write(struct file *filep, const char __user *ubuf,
+			     size_t size, loff_t *offp)
+{
+	struct tool_mw *mw = filep->private_data;
+	ssize_t rc;
+	loff_t pos = *offp;
+	void *buf;
+
+	if (pos < 0)
+		return -EINVAL;
+	if (pos >= mw->win_size || !size)
+		return 0;
+	if (size > mw->win_size - pos)
+		size = mw->win_size - pos;
+
+	buf = kmalloc(size, GFP_KERNEL);
+	if (!buf)
+		return -ENOMEM;
+
+	rc = copy_from_user(buf, ubuf, size);
+	if (rc == size) {
+		rc = -EFAULT;
+		goto err_free;
+	}
+
+	size -= rc;
+	*offp = pos + size;
+	rc = size;
+
+	memcpy_toio(mw->local + pos, buf, size);
+
+err_free:
+	kfree(buf);
+
+	return rc;
+}
+
+static TOOL_FOPS_RDWR(tool_mw_fops,
+		      tool_mw_read,
+		      tool_mw_write);
+
+static ssize_t tool_peer_mw_read(struct file *filep, char __user *ubuf,
+				 size_t size, loff_t *offp)
+{
+	struct tool_mw *mw = filep->private_data;
+
+	if (!mw->peer)
+		return -ENXIO;
+
+	return simple_read_from_buffer(ubuf, size, offp, mw->peer, mw->size);
+}
+
+static ssize_t tool_peer_mw_write(struct file *filep, const char __user *ubuf,
+				  size_t size, loff_t *offp)
+{
+	struct tool_mw *mw = filep->private_data;
+
+	if (!mw->peer)
+		return -ENXIO;
+
+	return simple_write_to_buffer(mw->peer, mw->size, offp, ubuf, size);
+}
+
+static TOOL_FOPS_RDWR(tool_peer_mw_fops,
+		      tool_peer_mw_read,
+		      tool_peer_mw_write);
+
+static int tool_setup_mw(struct tool_ctx *tc, int idx, size_t req_size)
+{
+	int rc;
+	struct tool_mw *mw = &tc->mws[idx];
+	phys_addr_t base;
+	resource_size_t size, align, align_size;
+	char buf[16];
+
+	if (mw->peer)
+		return 0;
+
+	rc = ntb_mw_get_range(tc->ntb, idx, &base, &size, &align,
+			      &align_size);
+	if (rc)
+		return rc;
+
+	mw->size = min_t(resource_size_t, req_size, size);
+	mw->size = round_up(mw->size, align);
+	mw->size = round_up(mw->size, align_size);
+	mw->peer = dma_alloc_coherent(&tc->ntb->pdev->dev, mw->size,
+				      &mw->peer_dma, GFP_KERNEL);
+
+	if (!mw->peer)
+		return -ENOMEM;
+
+	rc = ntb_mw_set_trans(tc->ntb, idx, mw->peer_dma, mw->size);
+	if (rc)
+		goto err_free_dma;
+
+	snprintf(buf, sizeof(buf), "peer_mw%d", idx);
+	mw->peer_dbg_file = debugfs_create_file(buf, S_IRUSR | S_IWUSR,
+						mw->tc->dbgfs, mw,
+						&tool_peer_mw_fops);
+
+	return 0;
+
+err_free_dma:
+	dma_free_coherent(&tc->ntb->pdev->dev, mw->size,
+			  mw->peer,
+			  mw->peer_dma);
+	mw->peer = NULL;
+	mw->peer_dma = 0;
+	mw->size = 0;
+
+	return rc;
+}
+
+static void tool_free_mw(struct tool_ctx *tc, int idx)
+{
+	struct tool_mw *mw = &tc->mws[idx];
+
+	if (mw->peer) {
+		ntb_mw_clear_trans(tc->ntb, idx);
+		dma_free_coherent(&tc->ntb->pdev->dev, mw->size,
+				  mw->peer,
+				  mw->peer_dma);
+	}
+
+	mw->peer = NULL;
+	mw->peer_dma = 0;
+
+	debugfs_remove(mw->peer_dbg_file);
+
+	mw->peer_dbg_file = NULL;
+}
+
+static ssize_t tool_peer_mw_trans_read(struct file *filep,
+				       char __user *ubuf,
+				       size_t size, loff_t *offp)
+{
+	struct tool_mw *mw = filep->private_data;
+
+	char *buf;
+	size_t buf_size;
+	ssize_t ret, off = 0;
+
+	phys_addr_t base;
+	resource_size_t mw_size;
+	resource_size_t align;
+	resource_size_t align_size;
+
+	buf_size = min_t(size_t, size, 512);
+
+	buf = kmalloc(buf_size, GFP_KERNEL);
+	if (!buf)
+		return -ENOMEM;
+
+	ntb_mw_get_range(mw->tc->ntb, mw->idx,
+			 &base, &mw_size, &align, &align_size);
+
+	off += scnprintf(buf + off, buf_size - off,
+			 "Peer MW %d Information:\n", mw->idx);
+
+	off += scnprintf(buf + off, buf_size - off,
+			 "Physical Address      \t%pa[p]\n",
+			 &base);
+
+	off += scnprintf(buf + off, buf_size - off,
+			 "Window Size           \t%lld\n",
+			 (unsigned long long)mw_size);
+
+	off += scnprintf(buf + off, buf_size - off,
+			 "Alignment             \t%lld\n",
+			 (unsigned long long)align);
+
+	off += scnprintf(buf + off, buf_size - off,
+			 "Size Alignment        \t%lld\n",
+			 (unsigned long long)align_size);
+
+	off += scnprintf(buf + off, buf_size - off,
+			 "Ready                 \t%c\n",
+			 (mw->peer) ? 'Y' : 'N');
+
+	off += scnprintf(buf + off, buf_size - off,
+			 "Allocated Size       \t%zd\n",
+			 (mw->peer) ? (size_t)mw->size : 0);
+
+	ret = simple_read_from_buffer(ubuf, size, offp, buf, off);
+	kfree(buf);
+	return ret;
+}
+
+static ssize_t tool_peer_mw_trans_write(struct file *filep,
+					const char __user *ubuf,
+					size_t size, loff_t *offp)
+{
+	struct tool_mw *mw = filep->private_data;
+
+	char buf[32];
+	size_t buf_size;
+	unsigned long long val;
+	int rc;
+
+	buf_size = min(size, (sizeof(buf) - 1));
+	if (copy_from_user(buf, ubuf, buf_size))
+		return -EFAULT;
+
+	buf[buf_size] = '\0';
+
+	rc = kstrtoull(buf, 0, &val);
+	if (rc)
+		return rc;
+
+	tool_free_mw(mw->tc, mw->idx);
+	if (val)
+		rc = tool_setup_mw(mw->tc, mw->idx, val);
+
+	if (rc)
+		return rc;
+
+	return size;
+}
+
+static TOOL_FOPS_RDWR(tool_peer_mw_trans_fops,
+		      tool_peer_mw_trans_read,
+		      tool_peer_mw_trans_write);
+
+static int tool_init_mw(struct tool_ctx *tc, int idx)
+{
+	struct tool_mw *mw = &tc->mws[idx];
+	phys_addr_t base;
+	int rc;
+
+	rc = ntb_mw_get_range(tc->ntb, idx, &base, &mw->win_size,
+			      NULL, NULL);
+	if (rc)
+		return rc;
+
+	mw->tc = tc;
+	mw->idx = idx;
+	mw->local = ioremap_wc(base, mw->win_size);
+	if (!mw->local)
+		return -EFAULT;
+
+	return 0;
+}
+
+static void tool_free_mws(struct tool_ctx *tc)
+{
+	int i;
+
+	for (i = 0; i < tc->mw_count; i++) {
+		tool_free_mw(tc, i);
+
+		if (tc->mws[i].local)
+			iounmap(tc->mws[i].local);
+
+		tc->mws[i].local = NULL;
+	}
+}
+
 static void tool_setup_dbgfs(struct tool_ctx *tc)
 {
+	int i;
+
 	/* This modules is useless without dbgfs... */
 	if (!tool_dbgfs) {
 		tc->dbgfs = NULL;
@@ -472,12 +887,31 @@ static void tool_setup_dbgfs(struct tool_ctx *tc)
 
 	debugfs_create_file("peer_spad", S_IRUSR | S_IWUSR, tc->dbgfs,
 			    tc, &tool_peer_spad_fops);
+
+	debugfs_create_file("link", S_IRUSR | S_IWUSR, tc->dbgfs,
+			    tc, &tool_link_fops);
+
+	debugfs_create_file("link_event", S_IWUSR, tc->dbgfs,
+			    tc, &tool_link_event_fops);
+
+	for (i = 0; i < tc->mw_count; i++) {
+		char buf[30];
+
+		snprintf(buf, sizeof(buf), "mw%d", i);
+		debugfs_create_file(buf, S_IRUSR | S_IWUSR, tc->dbgfs,
+				    &tc->mws[i], &tool_mw_fops);
+
+		snprintf(buf, sizeof(buf), "peer_trans%d", i);
+		debugfs_create_file(buf, S_IRUSR | S_IWUSR, tc->dbgfs,
+				    &tc->mws[i], &tool_peer_mw_trans_fops);
+	}
 }
 
 static int tool_probe(struct ntb_client *self, struct ntb_dev *ntb)
 {
 	struct tool_ctx *tc;
 	int rc;
+	int i;
 
 	if (ntb_db_is_unsafe(ntb))
 		dev_dbg(&ntb->dev, "doorbell is unsafe\n");
@@ -485,13 +919,21 @@ static int tool_probe(struct ntb_client *self, struct ntb_dev *ntb)
 	if (ntb_spad_is_unsafe(ntb))
 		dev_dbg(&ntb->dev, "scratchpad is unsafe\n");
 
-	tc = kmalloc(sizeof(*tc), GFP_KERNEL);
+	tc = kzalloc(sizeof(*tc), GFP_KERNEL);
 	if (!tc) {
 		rc = -ENOMEM;
 		goto err_tc;
 	}
 
 	tc->ntb = ntb;
+	init_waitqueue_head(&tc->link_wq);
+
+	tc->mw_count = min(ntb_mw_count(tc->ntb), MAX_MWS);
+	for (i = 0; i < tc->mw_count; i++) {
+		rc = tool_init_mw(tc, i);
+		if (rc)
+			goto err_ctx;
+	}
 
 	tool_setup_dbgfs(tc);
 
@@ -505,6 +947,7 @@ static int tool_probe(struct ntb_client *self, struct ntb_dev *ntb)
 	return 0;
 
 err_ctx:
+	tool_free_mws(tc);
 	debugfs_remove_recursive(tc->dbgfs);
 	kfree(tc);
 err_tc:
@@ -515,6 +958,8 @@ static void tool_remove(struct ntb_client *self, struct ntb_dev *ntb)
 {
 	struct tool_ctx *tc = ntb->ctx;
 
+	tool_free_mws(tc);
+
 	ntb_clear_ctx(ntb);
 	ntb_link_disable(ntb);