Adding upstream version 4.19.249.upstream/4.19.249

Signed-off-by: Daniel Baumann <daniel.baumann@progress-linux.org>

41 files changed, 42511 insertions, 0 deletions

diff --git a/drivers/net/ethernet/mellanox/mlx4/Kconfig b/drivers/net/ethernet/mellanox/mlx4/Kconfig
new file mode 100644
index 000000000..f200b8c42
--- /dev/null
+++ b/drivers/net/ethernet/mellanox/mlx4/Kconfig
@@ -0,0 +1,48 @@
+#
+# Mellanox driver configuration
+#
+
+config MLX4_EN
+	tristate "Mellanox Technologies 1/10/40Gbit Ethernet support"
+	depends on MAY_USE_DEVLINK
+	depends on PCI && NETDEVICES && ETHERNET && INET
+	select MLX4_CORE
+	imply PTP_1588_CLOCK
+	---help---
+	  This driver supports Mellanox Technologies ConnectX Ethernet
+	  devices.
+
+config MLX4_EN_DCB
+	bool "Data Center Bridging (DCB) Support"
+	default y
+	depends on MLX4_EN && DCB
+	---help---
+	  Say Y here if you want to use Data Center Bridging (DCB) in the
+	  driver.
+	  If set to N, will not be able to configure QoS and ratelimit attributes.
+	  This flag is depended on the kernel's DCB support.
+
+	  If unsure, set to Y
+
+config MLX4_CORE
+	tristate
+	depends on PCI
+	default n
+
+config MLX4_DEBUG
+	bool "Verbose debugging output" if (MLX4_CORE && EXPERT)
+	depends on MLX4_CORE
+	default y
+	---help---
+	  This option causes debugging code to be compiled into the
+	  mlx4_core driver.  The output can be turned on via the
+	  debug_level module parameter (which can also be set after
+	  the driver is loaded through sysfs).
+
+config MLX4_CORE_GEN2
+	bool "Support for old gen2 Mellanox PCI IDs" if (MLX4_CORE)
+	depends on MLX4_CORE
+	default y
+	---help---
+	  Say Y here if you want to use old gen2 Mellanox devices in the
+	  driver.
diff --git a/drivers/net/ethernet/mellanox/mlx4/Makefile b/drivers/net/ethernet/mellanox/mlx4/Makefile
new file mode 100644
index 000000000..3f400770f
--- /dev/null
+++ b/drivers/net/ethernet/mellanox/mlx4/Makefile
@@ -0,0 +1,12 @@
+# SPDX-License-Identifier: GPL-2.0
+obj-$(CONFIG_MLX4_CORE)		+= mlx4_core.o
+
+mlx4_core-y :=	alloc.o catas.o cmd.o cq.o eq.o fw.o fw_qos.o icm.o intf.o \
+		main.o mcg.o mr.o pd.o port.o profile.o qp.o reset.o sense.o \
+		srq.o resource_tracker.o crdump.o
+
+obj-$(CONFIG_MLX4_EN)               += mlx4_en.o
+
+mlx4_en-y := 	en_main.o en_tx.o en_rx.o en_ethtool.o en_port.o en_cq.o \
+		en_resources.o en_netdev.o en_selftest.o en_clock.o
+mlx4_en-$(CONFIG_MLX4_EN_DCB) += en_dcb_nl.o
diff --git a/drivers/net/ethernet/mellanox/mlx4/alloc.c b/drivers/net/ethernet/mellanox/mlx4/alloc.c
new file mode 100644
index 000000000..21788d4f9
--- /dev/null
+++ b/drivers/net/ethernet/mellanox/mlx4/alloc.c
@@ -0,0 +1,819 @@
+/*
+ * Copyright (c) 2006, 2007 Cisco Systems, Inc.  All rights reserved.
+ * Copyright (c) 2007, 2008 Mellanox Technologies. All rights reserved.
+ *
+ * This software is available to you under a choice of one of two
+ * licenses.  You may choose to be licensed under the terms of the GNU
+ * General Public License (GPL) Version 2, available from the file
+ * COPYING in the main directory of this source tree, or the
+ * OpenIB.org BSD license below:
+ *
+ *     Redistribution and use in source and binary forms, with or
+ *     without modification, are permitted provided that the following
+ *     conditions are met:
+ *
+ *      - Redistributions of source code must retain the above
+ *        copyright notice, this list of conditions and the following
+ *        disclaimer.
+ *
+ *      - Redistributions in binary form must reproduce the above
+ *        copyright notice, this list of conditions and the following
+ *        disclaimer in the documentation and/or other materials
+ *        provided with the distribution.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
+ * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
+ * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS
+ * BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN
+ * ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
+ * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+ * SOFTWARE.
+ */
+
+#include <linux/errno.h>
+#include <linux/slab.h>
+#include <linux/mm.h>
+#include <linux/export.h>
+#include <linux/bitmap.h>
+#include <linux/dma-mapping.h>
+#include <linux/vmalloc.h>
+
+#include "mlx4.h"
+
+u32 mlx4_bitmap_alloc(struct mlx4_bitmap *bitmap)
+{
+	u32 obj;
+
+	spin_lock(&bitmap->lock);
+
+	obj = find_next_zero_bit(bitmap->table, bitmap->max, bitmap->last);
+	if (obj >= bitmap->max) {
+		bitmap->top = (bitmap->top + bitmap->max + bitmap->reserved_top)
+				& bitmap->mask;
+		obj = find_first_zero_bit(bitmap->table, bitmap->max);
+	}
+
+	if (obj < bitmap->max) {
+		set_bit(obj, bitmap->table);
+		bitmap->last = (obj + 1);
+		if (bitmap->last == bitmap->max)
+			bitmap->last = 0;
+		obj |= bitmap->top;
+	} else
+		obj = -1;
+
+	if (obj != -1)
+		--bitmap->avail;
+
+	spin_unlock(&bitmap->lock);
+
+	return obj;
+}
+
+void mlx4_bitmap_free(struct mlx4_bitmap *bitmap, u32 obj, int use_rr)
+{
+	mlx4_bitmap_free_range(bitmap, obj, 1, use_rr);
+}
+
+static unsigned long find_aligned_range(unsigned long *bitmap,
+					u32 start, u32 nbits,
+					int len, int align, u32 skip_mask)
+{
+	unsigned long end, i;
+
+again:
+	start = ALIGN(start, align);
+
+	while ((start < nbits) && (test_bit(start, bitmap) ||
+				   (start & skip_mask)))
+		start += align;
+
+	if (start >= nbits)
+		return -1;
+
+	end = start+len;
+	if (end > nbits)
+		return -1;
+
+	for (i = start + 1; i < end; i++) {
+		if (test_bit(i, bitmap) || ((u32)i & skip_mask)) {
+			start = i + 1;
+			goto again;
+		}
+	}
+
+	return start;
+}
+
+u32 mlx4_bitmap_alloc_range(struct mlx4_bitmap *bitmap, int cnt,
+			    int align, u32 skip_mask)
+{
+	u32 obj;
+
+	if (likely(cnt == 1 && align == 1 && !skip_mask))
+		return mlx4_bitmap_alloc(bitmap);
+
+	spin_lock(&bitmap->lock);
+
+	obj = find_aligned_range(bitmap->table, bitmap->last,
+				 bitmap->max, cnt, align, skip_mask);
+	if (obj >= bitmap->max) {
+		bitmap->top = (bitmap->top + bitmap->max + bitmap->reserved_top)
+				& bitmap->mask;
+		obj = find_aligned_range(bitmap->table, 0, bitmap->max,
+					 cnt, align, skip_mask);
+	}
+
+	if (obj < bitmap->max) {
+		bitmap_set(bitmap->table, obj, cnt);
+		if (obj == bitmap->last) {
+			bitmap->last = (obj + cnt);
+			if (bitmap->last >= bitmap->max)
+				bitmap->last = 0;
+		}
+		obj |= bitmap->top;
+	} else
+		obj = -1;
+
+	if (obj != -1)
+		bitmap->avail -= cnt;
+
+	spin_unlock(&bitmap->lock);
+
+	return obj;
+}
+
+u32 mlx4_bitmap_avail(struct mlx4_bitmap *bitmap)
+{
+	return bitmap->avail;
+}
+
+static u32 mlx4_bitmap_masked_value(struct mlx4_bitmap *bitmap, u32 obj)
+{
+	return obj & (bitmap->max + bitmap->reserved_top - 1);
+}
+
+void mlx4_bitmap_free_range(struct mlx4_bitmap *bitmap, u32 obj, int cnt,
+			    int use_rr)
+{
+	obj &= bitmap->max + bitmap->reserved_top - 1;
+
+	spin_lock(&bitmap->lock);
+	if (!use_rr) {
+		bitmap->last = min(bitmap->last, obj);
+		bitmap->top = (bitmap->top + bitmap->max + bitmap->reserved_top)
+				& bitmap->mask;
+	}
+	bitmap_clear(bitmap->table, obj, cnt);
+	bitmap->avail += cnt;
+	spin_unlock(&bitmap->lock);
+}
+
+int mlx4_bitmap_init(struct mlx4_bitmap *bitmap, u32 num, u32 mask,
+		     u32 reserved_bot, u32 reserved_top)
+{
+	/* num must be a power of 2 */
+	if (num != roundup_pow_of_two(num))
+		return -EINVAL;
+
+	bitmap->last = 0;
+	bitmap->top  = 0;
+	bitmap->max  = num - reserved_top;
+	bitmap->mask = mask;
+	bitmap->reserved_top = reserved_top;
+	bitmap->avail = num - reserved_top - reserved_bot;
+	bitmap->effective_len = bitmap->avail;
+	spin_lock_init(&bitmap->lock);
+	bitmap->table = kcalloc(BITS_TO_LONGS(bitmap->max), sizeof(long),
+				GFP_KERNEL);
+	if (!bitmap->table)
+		return -ENOMEM;
+
+	bitmap_set(bitmap->table, 0, reserved_bot);
+
+	return 0;
+}
+
+void mlx4_bitmap_cleanup(struct mlx4_bitmap *bitmap)
+{
+	kfree(bitmap->table);
+}
+
+struct mlx4_zone_allocator {
+	struct list_head		entries;
+	struct list_head		prios;
+	u32				last_uid;
+	u32				mask;
+	/* protect the zone_allocator from concurrent accesses */
+	spinlock_t			lock;
+	enum mlx4_zone_alloc_flags	flags;
+};
+
+struct mlx4_zone_entry {
+	struct list_head		list;
+	struct list_head		prio_list;
+	u32				uid;
+	struct mlx4_zone_allocator	*allocator;
+	struct mlx4_bitmap		*bitmap;
+	int				use_rr;
+	int				priority;
+	int				offset;
+	enum mlx4_zone_flags		flags;
+};
+
+struct mlx4_zone_allocator *mlx4_zone_allocator_create(enum mlx4_zone_alloc_flags flags)
+{
+	struct mlx4_zone_allocator *zones = kmalloc(sizeof(*zones), GFP_KERNEL);
+
+	if (NULL == zones)
+		return NULL;
+
+	INIT_LIST_HEAD(&zones->entries);
+	INIT_LIST_HEAD(&zones->prios);
+	spin_lock_init(&zones->lock);
+	zones->last_uid = 0;
+	zones->mask = 0;
+	zones->flags = flags;
+
+	return zones;
+}
+
+int mlx4_zone_add_one(struct mlx4_zone_allocator *zone_alloc,
+		      struct mlx4_bitmap *bitmap,
+		      u32 flags,
+		      int priority,
+		      int offset,
+		      u32 *puid)
+{
+	u32 mask = mlx4_bitmap_masked_value(bitmap, (u32)-1);
+	struct mlx4_zone_entry *it;
+	struct mlx4_zone_entry *zone = kmalloc(sizeof(*zone), GFP_KERNEL);
+
+	if (NULL == zone)
+		return -ENOMEM;
+
+	zone->flags = flags;
+	zone->bitmap = bitmap;
+	zone->use_rr = (flags & MLX4_ZONE_USE_RR) ? MLX4_USE_RR : 0;
+	zone->priority = priority;
+	zone->offset = offset;
+
+	spin_lock(&zone_alloc->lock);
+
+	zone->uid = zone_alloc->last_uid++;
+	zone->allocator = zone_alloc;
+
+	if (zone_alloc->mask < mask)
+		zone_alloc->mask = mask;
+
+	list_for_each_entry(it, &zone_alloc->prios, prio_list)
+		if (it->priority >= priority)
+			break;
+
+	if (&it->prio_list == &zone_alloc->prios || it->priority > priority)
+		list_add_tail(&zone->prio_list, &it->prio_list);
+	list_add_tail(&zone->list, &it->list);
+
+	spin_unlock(&zone_alloc->lock);
+
+	*puid = zone->uid;
+
+	return 0;
+}
+
+/* Should be called under a lock */
+static void __mlx4_zone_remove_one_entry(struct mlx4_zone_entry *entry)
+{
+	struct mlx4_zone_allocator *zone_alloc = entry->allocator;
+
+	if (!list_empty(&entry->prio_list)) {
+		/* Check if we need to add an alternative node to the prio list */
+		if (!list_is_last(&entry->list, &zone_alloc->entries)) {
+			struct mlx4_zone_entry *next = list_first_entry(&entry->list,
+									typeof(*next),
+									list);
+
+			if (next->priority == entry->priority)
+				list_add_tail(&next->prio_list, &entry->prio_list);
+		}
+
+		list_del(&entry->prio_list);
+	}
+
+	list_del(&entry->list);
+
+	if (zone_alloc->flags & MLX4_ZONE_ALLOC_FLAGS_NO_OVERLAP) {
+		u32 mask = 0;
+		struct mlx4_zone_entry *it;
+
+		list_for_each_entry(it, &zone_alloc->prios, prio_list) {
+			u32 cur_mask = mlx4_bitmap_masked_value(it->bitmap, (u32)-1);
+
+			if (mask < cur_mask)
+				mask = cur_mask;
+		}
+		zone_alloc->mask = mask;
+	}
+}
+
+void mlx4_zone_allocator_destroy(struct mlx4_zone_allocator *zone_alloc)
+{
+	struct mlx4_zone_entry *zone, *tmp;
+
+	spin_lock(&zone_alloc->lock);
+
+	list_for_each_entry_safe(zone, tmp, &zone_alloc->entries, list) {
+		list_del(&zone->list);
+		list_del(&zone->prio_list);
+		kfree(zone);
+	}
+
+	spin_unlock(&zone_alloc->lock);
+	kfree(zone_alloc);
+}
+
+/* Should be called under a lock */
+static u32 __mlx4_alloc_from_zone(struct mlx4_zone_entry *zone, int count,
+				  int align, u32 skip_mask, u32 *puid)
+{
+	u32 uid = 0;
+	u32 res;
+	struct mlx4_zone_allocator *zone_alloc = zone->allocator;
+	struct mlx4_zone_entry *curr_node;
+
+	res = mlx4_bitmap_alloc_range(zone->bitmap, count,
+				      align, skip_mask);
+
+	if (res != (u32)-1) {
+		res += zone->offset;
+		uid = zone->uid;
+		goto out;
+	}
+
+	list_for_each_entry(curr_node, &zone_alloc->prios, prio_list) {
+		if (unlikely(curr_node->priority == zone->priority))
+			break;
+	}
+
+	if (zone->flags & MLX4_ZONE_ALLOW_ALLOC_FROM_LOWER_PRIO) {
+		struct mlx4_zone_entry *it = curr_node;
+
+		list_for_each_entry_continue_reverse(it, &zone_alloc->entries, list) {
+			res = mlx4_bitmap_alloc_range(it->bitmap, count,
+						      align, skip_mask);
+			if (res != (u32)-1) {
+				res += it->offset;
+				uid = it->uid;
+				goto out;
+			}
+		}
+	}
+
+	if (zone->flags & MLX4_ZONE_ALLOW_ALLOC_FROM_EQ_PRIO) {
+		struct mlx4_zone_entry *it = curr_node;
+
+		list_for_each_entry_from(it, &zone_alloc->entries, list) {
+			if (unlikely(it == zone))
+				continue;
+
+			if (unlikely(it->priority != curr_node->priority))
+				break;
+
+			res = mlx4_bitmap_alloc_range(it->bitmap, count,
+						      align, skip_mask);
+			if (res != (u32)-1) {
+				res += it->offset;
+				uid = it->uid;
+				goto out;
+			}
+		}
+	}
+
+	if (zone->flags & MLX4_ZONE_FALLBACK_TO_HIGHER_PRIO) {
+		if (list_is_last(&curr_node->prio_list, &zone_alloc->prios))
+			goto out;
+
+		curr_node = list_first_entry(&curr_node->prio_list,
+					     typeof(*curr_node),
+					     prio_list);
+
+		list_for_each_entry_from(curr_node, &zone_alloc->entries, list) {
+			res = mlx4_bitmap_alloc_range(curr_node->bitmap, count,
+						      align, skip_mask);
+			if (res != (u32)-1) {
+				res += curr_node->offset;
+				uid = curr_node->uid;
+				goto out;
+			}
+		}
+	}
+
+out:
+	if (NULL != puid && res != (u32)-1)
+		*puid = uid;
+	return res;
+}
+
+/* Should be called under a lock */
+static void __mlx4_free_from_zone(struct mlx4_zone_entry *zone, u32 obj,
+				  u32 count)
+{
+	mlx4_bitmap_free_range(zone->bitmap, obj - zone->offset, count, zone->use_rr);
+}
+
+/* Should be called under a lock */
+static struct mlx4_zone_entry *__mlx4_find_zone_by_uid(
+		struct mlx4_zone_allocator *zones, u32 uid)
+{
+	struct mlx4_zone_entry *zone;
+
+	list_for_each_entry(zone, &zones->entries, list) {
+		if (zone->uid == uid)
+			return zone;
+	}
+
+	return NULL;
+}
+
+struct mlx4_bitmap *mlx4_zone_get_bitmap(struct mlx4_zone_allocator *zones, u32 uid)
+{
+	struct mlx4_zone_entry *zone;
+	struct mlx4_bitmap *bitmap;
+
+	spin_lock(&zones->lock);
+
+	zone = __mlx4_find_zone_by_uid(zones, uid);
+
+	bitmap = zone == NULL ? NULL : zone->bitmap;
+
+	spin_unlock(&zones->lock);
+
+	return bitmap;
+}
+
+int mlx4_zone_remove_one(struct mlx4_zone_allocator *zones, u32 uid)
+{
+	struct mlx4_zone_entry *zone;
+	int res = 0;
+
+	spin_lock(&zones->lock);
+
+	zone = __mlx4_find_zone_by_uid(zones, uid);
+
+	if (NULL == zone) {
+		res = -1;
+		goto out;
+	}
+
+	__mlx4_zone_remove_one_entry(zone);
+
+out:
+	spin_unlock(&zones->lock);
+	kfree(zone);
+
+	return res;
+}
+
+/* Should be called under a lock */
+static struct mlx4_zone_entry *__mlx4_find_zone_by_uid_unique(
+		struct mlx4_zone_allocator *zones, u32 obj)
+{
+	struct mlx4_zone_entry *zone, *zone_candidate = NULL;
+	u32 dist = (u32)-1;
+
+	/* Search for the smallest zone that this obj could be
+	 * allocated from. This is done in order to handle
+	 * situations when small bitmaps are allocated from bigger
+	 * bitmaps (and the allocated space is marked as reserved in
+	 * the bigger bitmap.
+	 */
+	list_for_each_entry(zone, &zones->entries, list) {
+		if (obj >= zone->offset) {
+			u32 mobj = (obj - zone->offset) & zones->mask;
+
+			if (mobj < zone->bitmap->max) {
+				u32 curr_dist = zone->bitmap->effective_len;
+
+				if (curr_dist < dist) {
+					dist = curr_dist;
+					zone_candidate = zone;
+				}
+			}
+		}
+	}
+
+	return zone_candidate;
+}
+
+u32 mlx4_zone_alloc_entries(struct mlx4_zone_allocator *zones, u32 uid, int count,
+			    int align, u32 skip_mask, u32 *puid)
+{
+	struct mlx4_zone_entry *zone;
+	int res = -1;
+
+	spin_lock(&zones->lock);
+
+	zone = __mlx4_find_zone_by_uid(zones, uid);
+
+	if (NULL == zone)
+		goto out;
+
+	res = __mlx4_alloc_from_zone(zone, count, align, skip_mask, puid);
+
+out:
+	spin_unlock(&zones->lock);
+
+	return res;
+}
+
+u32 mlx4_zone_free_entries(struct mlx4_zone_allocator *zones, u32 uid, u32 obj, u32 count)
+{
+	struct mlx4_zone_entry *zone;
+	int res = 0;
+
+	spin_lock(&zones->lock);
+
+	zone = __mlx4_find_zone_by_uid(zones, uid);
+
+	if (NULL == zone) {
+		res = -1;
+		goto out;
+	}
+
+	__mlx4_free_from_zone(zone, obj, count);
+
+out:
+	spin_unlock(&zones->lock);
+
+	return res;
+}
+
+u32 mlx4_zone_free_entries_unique(struct mlx4_zone_allocator *zones, u32 obj, u32 count)
+{
+	struct mlx4_zone_entry *zone;
+	int res;
+
+	if (!(zones->flags & MLX4_ZONE_ALLOC_FLAGS_NO_OVERLAP))
+		return -EFAULT;
+
+	spin_lock(&zones->lock);
+
+	zone = __mlx4_find_zone_by_uid_unique(zones, obj);
+
+	if (NULL == zone) {
+		res = -1;
+		goto out;
+	}
+
+	__mlx4_free_from_zone(zone, obj, count);
+	res = 0;
+
+out:
+	spin_unlock(&zones->lock);
+
+	return res;
+}
+
+static int mlx4_buf_direct_alloc(struct mlx4_dev *dev, int size,
+				 struct mlx4_buf *buf)
+{
+	dma_addr_t t;
+
+	buf->nbufs        = 1;
+	buf->npages       = 1;
+	buf->page_shift   = get_order(size) + PAGE_SHIFT;
+	buf->direct.buf   =
+		dma_zalloc_coherent(&dev->persist->pdev->dev,
+				    size, &t, GFP_KERNEL);
+	if (!buf->direct.buf)
+		return -ENOMEM;
+
+	buf->direct.map = t;
+
+	while (t & ((1 << buf->page_shift) - 1)) {
+		--buf->page_shift;
+		buf->npages *= 2;
+	}
+
+	return 0;
+}
+
+/* Handling for queue buffers -- we allocate a bunch of memory and
+ * register it in a memory region at HCA virtual address 0. If the
+ *  requested size is > max_direct, we split the allocation into
+ *  multiple pages, so we don't require too much contiguous memory.
+ */
+int mlx4_buf_alloc(struct mlx4_dev *dev, int size, int max_direct,
+		   struct mlx4_buf *buf)
+{
+	if (size <= max_direct) {
+		return mlx4_buf_direct_alloc(dev, size, buf);
+	} else {
+		dma_addr_t t;
+		int i;
+
+		buf->direct.buf = NULL;
+		buf->nbufs	= (size + PAGE_SIZE - 1) / PAGE_SIZE;
+		buf->npages	= buf->nbufs;
+		buf->page_shift  = PAGE_SHIFT;
+		buf->page_list   = kcalloc(buf->nbufs, sizeof(*buf->page_list),
+					   GFP_KERNEL);
+		if (!buf->page_list)
+			return -ENOMEM;
+
+		for (i = 0; i < buf->nbufs; ++i) {
+			buf->page_list[i].buf =
+				dma_zalloc_coherent(&dev->persist->pdev->dev,
+						    PAGE_SIZE, &t, GFP_KERNEL);
+			if (!buf->page_list[i].buf)
+				goto err_free;
+
+			buf->page_list[i].map = t;
+		}
+	}
+
+	return 0;
+
+err_free:
+	mlx4_buf_free(dev, size, buf);
+
+	return -ENOMEM;
+}
+EXPORT_SYMBOL_GPL(mlx4_buf_alloc);
+
+void mlx4_buf_free(struct mlx4_dev *dev, int size, struct mlx4_buf *buf)
+{
+	if (buf->nbufs == 1) {
+		dma_free_coherent(&dev->persist->pdev->dev, size,
+				  buf->direct.buf, buf->direct.map);
+	} else {
+		int i;
+
+		for (i = 0; i < buf->nbufs; ++i)
+			if (buf->page_list[i].buf)
+				dma_free_coherent(&dev->persist->pdev->dev,
+						  PAGE_SIZE,
+						  buf->page_list[i].buf,
+						  buf->page_list[i].map);
+		kfree(buf->page_list);
+	}
+}
+EXPORT_SYMBOL_GPL(mlx4_buf_free);
+
+static struct mlx4_db_pgdir *mlx4_alloc_db_pgdir(struct device *dma_device)
+{
+	struct mlx4_db_pgdir *pgdir;
+
+	pgdir = kzalloc(sizeof(*pgdir), GFP_KERNEL);
+	if (!pgdir)
+		return NULL;
+
+	bitmap_fill(pgdir->order1, MLX4_DB_PER_PAGE / 2);
+	pgdir->bits[0] = pgdir->order0;
+	pgdir->bits[1] = pgdir->order1;
+	pgdir->db_page = dma_alloc_coherent(dma_device, PAGE_SIZE,
+					    &pgdir->db_dma, GFP_KERNEL);
+	if (!pgdir->db_page) {
+		kfree(pgdir);
+		return NULL;
+	}
+
+	return pgdir;
+}
+
+static int mlx4_alloc_db_from_pgdir(struct mlx4_db_pgdir *pgdir,
+				    struct mlx4_db *db, int order)
+{
+	int o;
+	int i;
+
+	for (o = order; o <= 1; ++o) {
+		i = find_first_bit(pgdir->bits[o], MLX4_DB_PER_PAGE >> o);
+		if (i < MLX4_DB_PER_PAGE >> o)
+			goto found;
+	}
+
+	return -ENOMEM;
+
+found:
+	clear_bit(i, pgdir->bits[o]);
+
+	i <<= o;
+
+	if (o > order)
+		set_bit(i ^ 1, pgdir->bits[order]);
+
+	db->u.pgdir = pgdir;
+	db->index   = i;
+	db->db      = pgdir->db_page + db->index;
+	db->dma     = pgdir->db_dma  + db->index * 4;
+	db->order   = order;
+
+	return 0;
+}
+
+int mlx4_db_alloc(struct mlx4_dev *dev, struct mlx4_db *db, int order)
+{
+	struct mlx4_priv *priv = mlx4_priv(dev);
+	struct mlx4_db_pgdir *pgdir;
+	int ret = 0;
+
+	mutex_lock(&priv->pgdir_mutex);
+
+	list_for_each_entry(pgdir, &priv->pgdir_list, list)
+		if (!mlx4_alloc_db_from_pgdir(pgdir, db, order))
+			goto out;
+
+	pgdir = mlx4_alloc_db_pgdir(&dev->persist->pdev->dev);
+	if (!pgdir) {
+		ret = -ENOMEM;
+		goto out;
+	}
+
+	list_add(&pgdir->list, &priv->pgdir_list);
+
+	/* This should never fail -- we just allocated an empty page: */
+	WARN_ON(mlx4_alloc_db_from_pgdir(pgdir, db, order));
+
+out:
+	mutex_unlock(&priv->pgdir_mutex);
+
+	return ret;
+}
+EXPORT_SYMBOL_GPL(mlx4_db_alloc);
+
+void mlx4_db_free(struct mlx4_dev *dev, struct mlx4_db *db)
+{
+	struct mlx4_priv *priv = mlx4_priv(dev);
+	int o;
+	int i;
+
+	mutex_lock(&priv->pgdir_mutex);
+
+	o = db->order;
+	i = db->index;
+
+	if (db->order == 0 && test_bit(i ^ 1, db->u.pgdir->order0)) {
+		clear_bit(i ^ 1, db->u.pgdir->order0);
+		++o;
+	}
+	i >>= o;
+	set_bit(i, db->u.pgdir->bits[o]);
+
+	if (bitmap_full(db->u.pgdir->order1, MLX4_DB_PER_PAGE / 2)) {
+		dma_free_coherent(&dev->persist->pdev->dev, PAGE_SIZE,
+				  db->u.pgdir->db_page, db->u.pgdir->db_dma);
+		list_del(&db->u.pgdir->list);
+		kfree(db->u.pgdir);
+	}
+
+	mutex_unlock(&priv->pgdir_mutex);
+}
+EXPORT_SYMBOL_GPL(mlx4_db_free);
+
+int mlx4_alloc_hwq_res(struct mlx4_dev *dev, struct mlx4_hwq_resources *wqres,
+		       int size)
+{
+	int err;
+
+	err = mlx4_db_alloc(dev, &wqres->db, 1);
+	if (err)
+		return err;
+
+	*wqres->db.db = 0;
+
+	err = mlx4_buf_direct_alloc(dev, size, &wqres->buf);
+	if (err)
+		goto err_db;
+
+	err = mlx4_mtt_init(dev, wqres->buf.npages, wqres->buf.page_shift,
+			    &wqres->mtt);
+	if (err)
+		goto err_buf;
+
+	err = mlx4_buf_write_mtt(dev, &wqres->mtt, &wqres->buf);
+	if (err)
+		goto err_mtt;
+
+	return 0;
+
+err_mtt:
+	mlx4_mtt_cleanup(dev, &wqres->mtt);
+err_buf:
+	mlx4_buf_free(dev, size, &wqres->buf);
+err_db:
+	mlx4_db_free(dev, &wqres->db);
+
+	return err;
+}
+EXPORT_SYMBOL_GPL(mlx4_alloc_hwq_res);
+
+void mlx4_free_hwq_res(struct mlx4_dev *dev, struct mlx4_hwq_resources *wqres,
+		       int size)
+{
+	mlx4_mtt_cleanup(dev, &wqres->mtt);
+	mlx4_buf_free(dev, size, &wqres->buf);
+	mlx4_db_free(dev, &wqres->db);
+}
+EXPORT_SYMBOL_GPL(mlx4_free_hwq_res);
diff --git a/drivers/net/ethernet/mellanox/mlx4/catas.c b/drivers/net/ethernet/mellanox/mlx4/catas.c
new file mode 100644
index 000000000..c81d15bf2
--- /dev/null
+++ b/drivers/net/ethernet/mellanox/mlx4/catas.c
@@ -0,0 +1,334 @@
+/*
+ * Copyright (c) 2007 Cisco Systems, Inc. All rights reserved.
+ * Copyright (c) 2007, 2008 Mellanox Technologies. All rights reserved.
+ *
+ * This software is available to you under a choice of one of two
+ * licenses.  You may choose to be licensed under the terms of the GNU
+ * General Public License (GPL) Version 2, available from the file
+ * COPYING in the main directory of this source tree, or the
+ * OpenIB.org BSD license below:
+ *
+ *     Redistribution and use in source and binary forms, with or
+ *     without modification, are permitted provided that the following
+ *     conditions are met:
+ *
+ *      - Redistributions of source code must retain the above
+ *        copyright notice, this list of conditions and the following
+ *        disclaimer.
+ *
+ *      - Redistributions in binary form must reproduce the above
+ *        copyright notice, this list of conditions and the following
+ *        disclaimer in the documentation and/or other materials
+ *        provided with the distribution.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
+ * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
+ * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS
+ * BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN
+ * ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
+ * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+ * SOFTWARE.
+ */
+
+#include <linux/workqueue.h>
+#include <linux/module.h>
+
+#include "mlx4.h"
+
+enum {
+	MLX4_CATAS_POLL_INTERVAL	= 5 * HZ,
+};
+
+
+
+int mlx4_internal_err_reset = 1;
+module_param_named(internal_err_reset, mlx4_internal_err_reset,  int, 0644);
+MODULE_PARM_DESC(internal_err_reset,
+		 "Reset device on internal errors if non-zero (default 1)");
+
+static int read_vendor_id(struct mlx4_dev *dev)
+{
+	u16 vendor_id = 0;
+	int ret;
+
+	ret = pci_read_config_word(dev->persist->pdev, 0, &vendor_id);
+	if (ret) {
+		mlx4_err(dev, "Failed to read vendor ID, ret=%d\n", ret);
+		return ret;
+	}
+
+	if (vendor_id == 0xffff) {
+		mlx4_err(dev, "PCI can't be accessed to read vendor id\n");
+		return -EINVAL;
+	}
+
+	return 0;
+}
+
+static int mlx4_reset_master(struct mlx4_dev *dev)
+{
+	int err = 0;
+
+	if (mlx4_is_master(dev))
+		mlx4_report_internal_err_comm_event(dev);
+
+	if (!pci_channel_offline(dev->persist->pdev)) {
+		err = read_vendor_id(dev);
+		/* If PCI can't be accessed to read vendor ID we assume that its
+		 * link was disabled and chip was already reset.
+		 */
+		if (err)
+			return 0;
+
+		err = mlx4_reset(dev);
+		if (err)
+			mlx4_err(dev, "Fail to reset HCA\n");
+	}
+
+	return err;
+}
+
+static int mlx4_reset_slave(struct mlx4_dev *dev)
+{
+#define COM_CHAN_RST_REQ_OFFSET 0x10
+#define COM_CHAN_RST_ACK_OFFSET 0x08
+
+	u32 comm_flags;
+	u32 rst_req;
+	u32 rst_ack;
+	unsigned long end;
+	struct mlx4_priv *priv = mlx4_priv(dev);
+
+	if (pci_channel_offline(dev->persist->pdev))
+		return 0;
+
+	comm_flags = swab32(readl((__iomem char *)priv->mfunc.comm +
+				  MLX4_COMM_CHAN_FLAGS));
+	if (comm_flags == 0xffffffff) {
+		mlx4_err(dev, "VF reset is not needed\n");
+		return 0;
+	}
+
+	if (!(dev->caps.vf_caps & MLX4_VF_CAP_FLAG_RESET)) {
+		mlx4_err(dev, "VF reset is not supported\n");
+		return -EOPNOTSUPP;
+	}
+
+	rst_req = (comm_flags & (u32)(1 << COM_CHAN_RST_REQ_OFFSET)) >>
+		COM_CHAN_RST_REQ_OFFSET;
+	rst_ack = (comm_flags & (u32)(1 << COM_CHAN_RST_ACK_OFFSET)) >>
+		COM_CHAN_RST_ACK_OFFSET;
+	if (rst_req != rst_ack) {
+		mlx4_err(dev, "Communication channel isn't sync, fail to send reset\n");
+		return -EIO;
+	}
+
+	rst_req ^= 1;
+	mlx4_warn(dev, "VF is sending reset request to Firmware\n");
+	comm_flags = rst_req << COM_CHAN_RST_REQ_OFFSET;
+	__raw_writel((__force u32)cpu_to_be32(comm_flags),
+		     (__iomem char *)priv->mfunc.comm + MLX4_COMM_CHAN_FLAGS);
+	/* Make sure that our comm channel write doesn't
+	 * get mixed in with writes from another CPU.
+	 */
+	mmiowb();
+
+	end = msecs_to_jiffies(MLX4_COMM_TIME) + jiffies;
+	while (time_before(jiffies, end)) {
+		comm_flags = swab32(readl((__iomem char *)priv->mfunc.comm +
+					  MLX4_COMM_CHAN_FLAGS));
+		rst_ack = (comm_flags & (u32)(1 << COM_CHAN_RST_ACK_OFFSET)) >>
+			COM_CHAN_RST_ACK_OFFSET;
+
+		/* Reading rst_req again since the communication channel can
+		 * be reset at any time by the PF and all its bits will be
+		 * set to zero.
+		 */
+		rst_req = (comm_flags & (u32)(1 << COM_CHAN_RST_REQ_OFFSET)) >>
+			COM_CHAN_RST_REQ_OFFSET;
+
+		if (rst_ack == rst_req) {
+			mlx4_warn(dev, "VF Reset succeed\n");
+			return 0;
+		}
+		cond_resched();
+	}
+	mlx4_err(dev, "Fail to send reset over the communication channel\n");
+	return -ETIMEDOUT;
+}
+
+int mlx4_comm_internal_err(u32 slave_read)
+{
+	return (u32)COMM_CHAN_EVENT_INTERNAL_ERR ==
+		(slave_read & (u32)COMM_CHAN_EVENT_INTERNAL_ERR) ? 1 : 0;
+}
+
+void mlx4_enter_error_state(struct mlx4_dev_persistent *persist)
+{
+	int err;
+	struct mlx4_dev *dev;
+
+	if (!mlx4_internal_err_reset)
+		return;
+
+	mutex_lock(&persist->device_state_mutex);
+	if (persist->state & MLX4_DEVICE_STATE_INTERNAL_ERROR)
+		goto out;
+
+	dev = persist->dev;
+	mlx4_err(dev, "device is going to be reset\n");
+	if (mlx4_is_slave(dev)) {
+		err = mlx4_reset_slave(dev);
+	} else {
+		mlx4_crdump_collect(dev);
+		err = mlx4_reset_master(dev);
+	}
+
+	if (!err) {
+		mlx4_err(dev, "device was reset successfully\n");
+	} else {
+		/* EEH could have disabled the PCI channel during reset. That's
+		 * recoverable and the PCI error flow will handle it.
+		 */
+		if (!pci_channel_offline(dev->persist->pdev))
+			BUG_ON(1);
+	}
+	dev->persist->state |= MLX4_DEVICE_STATE_INTERNAL_ERROR;
+	mutex_unlock(&persist->device_state_mutex);
+
+	/* At that step HW was already reset, now notify clients */
+	mlx4_dispatch_event(dev, MLX4_DEV_EVENT_CATASTROPHIC_ERROR, 0);
+	mlx4_cmd_wake_completions(dev);
+	return;
+
+out:
+	mutex_unlock(&persist->device_state_mutex);
+}
+
+static void mlx4_handle_error_state(struct mlx4_dev_persistent *persist)
+{
+	int err = 0;
+
+	mlx4_enter_error_state(persist);
+	mutex_lock(&persist->interface_state_mutex);
+	if (persist->interface_state & MLX4_INTERFACE_STATE_UP &&
+	    !(persist->interface_state & MLX4_INTERFACE_STATE_DELETION)) {
+		err = mlx4_restart_one(persist->pdev, false, NULL);
+		mlx4_info(persist->dev, "mlx4_restart_one was ended, ret=%d\n",
+			  err);
+	}
+	mutex_unlock(&persist->interface_state_mutex);
+}
+
+static void dump_err_buf(struct mlx4_dev *dev)
+{
+	struct mlx4_priv *priv = mlx4_priv(dev);
+
+	int i;
+
+	mlx4_err(dev, "Internal error detected:\n");
+	for (i = 0; i < priv->fw.catas_size; ++i)
+		mlx4_err(dev, "  buf[%02x]: %08x\n",
+			 i, swab32(readl(priv->catas_err.map + i)));
+}
+
+static void poll_catas(struct timer_list *t)
+{
+	struct mlx4_priv *priv = from_timer(priv, t, catas_err.timer);
+	struct mlx4_dev *dev = &priv->dev;
+	u32 slave_read;
+
+	if (mlx4_is_slave(dev)) {
+		slave_read = swab32(readl(&priv->mfunc.comm->slave_read));
+		if (mlx4_comm_internal_err(slave_read)) {
+			mlx4_warn(dev, "Internal error detected on the communication channel\n");
+			goto internal_err;
+		}
+	} else if (readl(priv->catas_err.map)) {
+		dump_err_buf(dev);
+		goto internal_err;
+	}
+
+	if (dev->persist->state & MLX4_DEVICE_STATE_INTERNAL_ERROR) {
+		mlx4_warn(dev, "Internal error mark was detected on device\n");
+		goto internal_err;
+	}
+
+	mod_timer(&priv->catas_err.timer,
+		  round_jiffies(jiffies + MLX4_CATAS_POLL_INTERVAL));
+	return;
+
+internal_err:
+	if (mlx4_internal_err_reset)
+		queue_work(dev->persist->catas_wq, &dev->persist->catas_work);
+}
+
+static void catas_reset(struct work_struct *work)
+{
+	struct mlx4_dev_persistent *persist =
+		container_of(work, struct mlx4_dev_persistent,
+			     catas_work);
+
+	mlx4_handle_error_state(persist);
+}
+
+void mlx4_start_catas_poll(struct mlx4_dev *dev)
+{
+	struct mlx4_priv *priv = mlx4_priv(dev);
+	phys_addr_t addr;
+
+	INIT_LIST_HEAD(&priv->catas_err.list);
+	timer_setup(&priv->catas_err.timer, poll_catas, 0);
+	priv->catas_err.map = NULL;
+
+	if (!mlx4_is_slave(dev)) {
+		addr = pci_resource_start(dev->persist->pdev,
+					  priv->fw.catas_bar) +
+					  priv->fw.catas_offset;
+
+		priv->catas_err.map = ioremap(addr, priv->fw.catas_size * 4);
+		if (!priv->catas_err.map) {
+			mlx4_warn(dev, "Failed to map internal error buffer at 0x%llx\n",
+				  (unsigned long long)addr);
+			return;
+		}
+	}
+
+	priv->catas_err.timer.expires  =
+		round_jiffies(jiffies + MLX4_CATAS_POLL_INTERVAL);
+	add_timer(&priv->catas_err.timer);
+}
+
+void mlx4_stop_catas_poll(struct mlx4_dev *dev)
+{
+	struct mlx4_priv *priv = mlx4_priv(dev);
+
+	del_timer_sync(&priv->catas_err.timer);
+
+	if (priv->catas_err.map) {
+		iounmap(priv->catas_err.map);
+		priv->catas_err.map = NULL;
+	}
+
+	if (dev->persist->interface_state & MLX4_INTERFACE_STATE_DELETION)
+		flush_workqueue(dev->persist->catas_wq);
+}
+
+int  mlx4_catas_init(struct mlx4_dev *dev)
+{
+	INIT_WORK(&dev->persist->catas_work, catas_reset);
+	dev->persist->catas_wq = create_singlethread_workqueue("mlx4_health");
+	if (!dev->persist->catas_wq)
+		return -ENOMEM;
+
+	return 0;
+}
+
+void mlx4_catas_end(struct mlx4_dev *dev)
+{
+	if (dev->persist->catas_wq) {
+		destroy_workqueue(dev->persist->catas_wq);
+		dev->persist->catas_wq = NULL;
+	}
+}
diff --git a/drivers/net/ethernet/mellanox/mlx4/cmd.c b/drivers/net/ethernet/mellanox/mlx4/cmd.c
new file mode 100644
index 000000000..857588e24
--- /dev/null
+++ b/drivers/net/ethernet/mellanox/mlx4/cmd.c
@@ -0,0 +1,3443 @@
+/*
+ * Copyright (c) 2004, 2005 Topspin Communications.  All rights reserved.
+ * Copyright (c) 2005, 2006, 2007, 2008 Mellanox Technologies. All rights reserved.
+ * Copyright (c) 2005, 2006, 2007 Cisco Systems, Inc.  All rights reserved.
+ *
+ * This software is available to you under a choice of one of two
+ * licenses.  You may choose to be licensed under the terms of the GNU
+ * General Public License (GPL) Version 2, available from the file
+ * COPYING in the main directory of this source tree, or the
+ * OpenIB.org BSD license below:
+ *
+ *     Redistribution and use in source and binary forms, with or
+ *     without modification, are permitted provided that the following
+ *     conditions are met:
+ *
+ *      - Redistributions of source code must retain the above
+ *        copyright notice, this list of conditions and the following
+ *        disclaimer.
+ *
+ *      - Redistributions in binary form must reproduce the above
+ *        copyright notice, this list of conditions and the following
+ *        disclaimer in the documentation and/or other materials
+ *        provided with the distribution.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
+ * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
+ * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS
+ * BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN
+ * ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
+ * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+ * SOFTWARE.
+ */
+
+#include <linux/sched.h>
+#include <linux/slab.h>
+#include <linux/export.h>
+#include <linux/pci.h>
+#include <linux/errno.h>
+
+#include <linux/mlx4/cmd.h>
+#include <linux/mlx4/device.h>
+#include <linux/semaphore.h>
+#include <rdma/ib_smi.h>
+#include <linux/delay.h>
+#include <linux/etherdevice.h>
+
+#include <asm/io.h>
+
+#include "mlx4.h"
+#include "fw.h"
+#include "fw_qos.h"
+#include "mlx4_stats.h"
+
+#define CMD_POLL_TOKEN 0xffff
+#define INBOX_MASK	0xffffffffffffff00ULL
+
+#define CMD_CHAN_VER 1
+#define CMD_CHAN_IF_REV 1
+
+enum {
+	/* command completed successfully: */
+	CMD_STAT_OK		= 0x00,
+	/* Internal error (such as a bus error) occurred while processing command: */
+	CMD_STAT_INTERNAL_ERR	= 0x01,
+	/* Operation/command not supported or opcode modifier not supported: */
+	CMD_STAT_BAD_OP		= 0x02,
+	/* Parameter not supported or parameter out of range: */
+	CMD_STAT_BAD_PARAM	= 0x03,
+	/* System not enabled or bad system state: */
+	CMD_STAT_BAD_SYS_STATE	= 0x04,
+	/* Attempt to access reserved or unallocaterd resource: */
+	CMD_STAT_BAD_RESOURCE	= 0x05,
+	/* Requested resource is currently executing a command, or is otherwise busy: */
+	CMD_STAT_RESOURCE_BUSY	= 0x06,
+	/* Required capability exceeds device limits: */
+	CMD_STAT_EXCEED_LIM	= 0x08,
+	/* Resource is not in the appropriate state or ownership: */
+	CMD_STAT_BAD_RES_STATE	= 0x09,
+	/* Index out of range: */
+	CMD_STAT_BAD_INDEX	= 0x0a,
+	/* FW image corrupted: */
+	CMD_STAT_BAD_NVMEM	= 0x0b,
+	/* Error in ICM mapping (e.g. not enough auxiliary ICM pages to execute command): */
+	CMD_STAT_ICM_ERROR	= 0x0c,
+	/* Attempt to modify a QP/EE which is not in the presumed state: */
+	CMD_STAT_BAD_QP_STATE   = 0x10,
+	/* Bad segment parameters (Address/Size): */
+	CMD_STAT_BAD_SEG_PARAM	= 0x20,
+	/* Memory Region has Memory Windows bound to: */
+	CMD_STAT_REG_BOUND	= 0x21,
+	/* HCA local attached memory not present: */
+	CMD_STAT_LAM_NOT_PRE	= 0x22,
+	/* Bad management packet (silently discarded): */
+	CMD_STAT_BAD_PKT	= 0x30,
+	/* More outstanding CQEs in CQ than new CQ size: */
+	CMD_STAT_BAD_SIZE	= 0x40,
+	/* Multi Function device support required: */
+	CMD_STAT_MULTI_FUNC_REQ	= 0x50,
+};
+
+enum {
+	HCR_IN_PARAM_OFFSET	= 0x00,
+	HCR_IN_MODIFIER_OFFSET	= 0x08,
+	HCR_OUT_PARAM_OFFSET	= 0x0c,
+	HCR_TOKEN_OFFSET	= 0x14,
+	HCR_STATUS_OFFSET	= 0x18,
+
+	HCR_OPMOD_SHIFT		= 12,
+	HCR_T_BIT		= 21,
+	HCR_E_BIT		= 22,
+	HCR_GO_BIT		= 23
+};
+
+enum {
+	GO_BIT_TIMEOUT_MSECS	= 10000
+};
+
+enum mlx4_vlan_transition {
+	MLX4_VLAN_TRANSITION_VST_VST = 0,
+	MLX4_VLAN_TRANSITION_VST_VGT = 1,
+	MLX4_VLAN_TRANSITION_VGT_VST = 2,
+	MLX4_VLAN_TRANSITION_VGT_VGT = 3,
+};
+
+
+struct mlx4_cmd_context {
+	struct completion	done;
+	int			result;
+	int			next;
+	u64			out_param;
+	u16			token;
+	u8			fw_status;
+};
+
+static int mlx4_master_process_vhcr(struct mlx4_dev *dev, int slave,
+				    struct mlx4_vhcr_cmd *in_vhcr);
+
+static int mlx4_status_to_errno(u8 status)
+{
+	static const int trans_table[] = {
+		[CMD_STAT_INTERNAL_ERR]	  = -EIO,
+		[CMD_STAT_BAD_OP]	  = -EPERM,
+		[CMD_STAT_BAD_PARAM]	  = -EINVAL,
+		[CMD_STAT_BAD_SYS_STATE]  = -ENXIO,
+		[CMD_STAT_BAD_RESOURCE]	  = -EBADF,
+		[CMD_STAT_RESOURCE_BUSY]  = -EBUSY,
+		[CMD_STAT_EXCEED_LIM]	  = -ENOMEM,
+		[CMD_STAT_BAD_RES_STATE]  = -EBADF,
+		[CMD_STAT_BAD_INDEX]	  = -EBADF,
+		[CMD_STAT_BAD_NVMEM]	  = -EFAULT,
+		[CMD_STAT_ICM_ERROR]	  = -ENFILE,
+		[CMD_STAT_BAD_QP_STATE]   = -EINVAL,
+		[CMD_STAT_BAD_SEG_PARAM]  = -EFAULT,
+		[CMD_STAT_REG_BOUND]	  = -EBUSY,
+		[CMD_STAT_LAM_NOT_PRE]	  = -EAGAIN,
+		[CMD_STAT_BAD_PKT]	  = -EINVAL,
+		[CMD_STAT_BAD_SIZE]	  = -ENOMEM,
+		[CMD_STAT_MULTI_FUNC_REQ] = -EACCES,
+	};
+
+	if (status >= ARRAY_SIZE(trans_table) ||
+	    (status != CMD_STAT_OK && trans_table[status] == 0))
+		return -EIO;
+
+	return trans_table[status];
+}
+
+static u8 mlx4_errno_to_status(int errno)
+{
+	switch (errno) {
+	case -EPERM:
+		return CMD_STAT_BAD_OP;
+	case -EINVAL:
+		return CMD_STAT_BAD_PARAM;
+	case -ENXIO:
+		return CMD_STAT_BAD_SYS_STATE;
+	case -EBUSY:
+		return CMD_STAT_RESOURCE_BUSY;
+	case -ENOMEM:
+		return CMD_STAT_EXCEED_LIM;
+	case -ENFILE:
+		return CMD_STAT_ICM_ERROR;
+	default:
+		return CMD_STAT_INTERNAL_ERR;
+	}
+}
+
+static int mlx4_internal_err_ret_value(struct mlx4_dev *dev, u16 op,
+				       u8 op_modifier)
+{
+	switch (op) {
+	case MLX4_CMD_UNMAP_ICM:
+	case MLX4_CMD_UNMAP_ICM_AUX:
+	case MLX4_CMD_UNMAP_FA:
+	case MLX4_CMD_2RST_QP:
+	case MLX4_CMD_HW2SW_EQ:
+	case MLX4_CMD_HW2SW_CQ:
+	case MLX4_CMD_HW2SW_SRQ:
+	case MLX4_CMD_HW2SW_MPT:
+	case MLX4_CMD_CLOSE_HCA:
+	case MLX4_QP_FLOW_STEERING_DETACH:
+	case MLX4_CMD_FREE_RES:
+	case MLX4_CMD_CLOSE_PORT:
+		return CMD_STAT_OK;
+
+	case MLX4_CMD_QP_ATTACH:
+		/* On Detach case return success */
+		if (op_modifier == 0)
+			return CMD_STAT_OK;
+		return mlx4_status_to_errno(CMD_STAT_INTERNAL_ERR);
+
+	default:
+		return mlx4_status_to_errno(CMD_STAT_INTERNAL_ERR);
+	}
+}
+
+static int mlx4_closing_cmd_fatal_error(u16 op, u8 fw_status)
+{
+	/* Any error during the closing commands below is considered fatal */
+	if (op == MLX4_CMD_CLOSE_HCA ||
+	    op == MLX4_CMD_HW2SW_EQ ||
+	    op == MLX4_CMD_HW2SW_CQ ||
+	    op == MLX4_CMD_2RST_QP ||
+	    op == MLX4_CMD_HW2SW_SRQ ||
+	    op == MLX4_CMD_SYNC_TPT ||
+	    op == MLX4_CMD_UNMAP_ICM ||
+	    op == MLX4_CMD_UNMAP_ICM_AUX ||
+	    op == MLX4_CMD_UNMAP_FA)
+		return 1;
+	/* Error on MLX4_CMD_HW2SW_MPT is fatal except when fw status equals
+	  * CMD_STAT_REG_BOUND.
+	  * This status indicates that memory region has memory windows bound to it
+	  * which may result from invalid user space usage and is not fatal.
+	  */
+	if (op == MLX4_CMD_HW2SW_MPT && fw_status != CMD_STAT_REG_BOUND)
+		return 1;
+	return 0;
+}
+
+static int mlx4_cmd_reset_flow(struct mlx4_dev *dev, u16 op, u8 op_modifier,
+			       int err)
+{
+	/* Only if reset flow is really active return code is based on
+	  * command, otherwise current error code is returned.
+	  */
+	if (mlx4_internal_err_reset) {
+		mlx4_enter_error_state(dev->persist);
+		err = mlx4_internal_err_ret_value(dev, op, op_modifier);
+	}
+
+	return err;
+}
+
+static int comm_pending(struct mlx4_dev *dev)
+{
+	struct mlx4_priv *priv = mlx4_priv(dev);
+	u32 status = readl(&priv->mfunc.comm->slave_read);
+
+	return (swab32(status) >> 31) != priv->cmd.comm_toggle;
+}
+
+static int mlx4_comm_cmd_post(struct mlx4_dev *dev, u8 cmd, u16 param)
+{
+	struct mlx4_priv *priv = mlx4_priv(dev);
+	u32 val;
+
+	/* To avoid writing to unknown addresses after the device state was
+	 * changed to internal error and the function was rest,
+	 * check the INTERNAL_ERROR flag which is updated under
+	 * device_state_mutex lock.
+	 */
+	mutex_lock(&dev->persist->device_state_mutex);
+
+	if (dev->persist->state & MLX4_DEVICE_STATE_INTERNAL_ERROR) {
+		mutex_unlock(&dev->persist->device_state_mutex);
+		return -EIO;
+	}
+
+	priv->cmd.comm_toggle ^= 1;
+	val = param | (cmd << 16) | (priv->cmd.comm_toggle << 31);
+	__raw_writel((__force u32) cpu_to_be32(val),
+		     &priv->mfunc.comm->slave_write);
+	mmiowb();
+	mutex_unlock(&dev->persist->device_state_mutex);
+	return 0;
+}
+
+static int mlx4_comm_cmd_poll(struct mlx4_dev *dev, u8 cmd, u16 param,
+		       unsigned long timeout)
+{
+	struct mlx4_priv *priv = mlx4_priv(dev);
+	unsigned long end;
+	int err = 0;
+	int ret_from_pending = 0;
+
+	/* First, verify that the master reports correct status */
+	if (comm_pending(dev)) {
+		mlx4_warn(dev, "Communication channel is not idle - my toggle is %d (cmd:0x%x)\n",
+			  priv->cmd.comm_toggle, cmd);
+		return -EAGAIN;
+	}
+
+	/* Write command */
+	down(&priv->cmd.poll_sem);
+	if (mlx4_comm_cmd_post(dev, cmd, param)) {
+		/* Only in case the device state is INTERNAL_ERROR,
+		 * mlx4_comm_cmd_post returns with an error
+		 */
+		err = mlx4_status_to_errno(CMD_STAT_INTERNAL_ERR);
+		goto out;
+	}
+
+	end = msecs_to_jiffies(timeout) + jiffies;
+	while (comm_pending(dev) && time_before(jiffies, end))
+		cond_resched();
+	ret_from_pending = comm_pending(dev);
+	if (ret_from_pending) {
+		/* check if the slave is trying to boot in the middle of
+		 * FLR process. The only non-zero result in the RESET command
+		 * is MLX4_DELAY_RESET_SLAVE*/
+		if ((MLX4_COMM_CMD_RESET == cmd)) {
+			err = MLX4_DELAY_RESET_SLAVE;
+			goto out;
+		} else {
+			mlx4_warn(dev, "Communication channel command 0x%x timed out\n",
+				  cmd);
+			err = mlx4_status_to_errno(CMD_STAT_INTERNAL_ERR);
+		}
+	}
+
+	if (err)
+		mlx4_enter_error_state(dev->persist);
+out:
+	up(&priv->cmd.poll_sem);
+	return err;
+}
+
+static int mlx4_comm_cmd_wait(struct mlx4_dev *dev, u8 vhcr_cmd,
+			      u16 param, u16 op, unsigned long timeout)
+{
+	struct mlx4_cmd *cmd = &mlx4_priv(dev)->cmd;
+	struct mlx4_cmd_context *context;
+	unsigned long end;
+	int err = 0;
+
+	down(&cmd->event_sem);
+
+	spin_lock(&cmd->context_lock);
+	BUG_ON(cmd->free_head < 0);
+	context = &cmd->context[cmd->free_head];
+	context->token += cmd->token_mask + 1;
+	cmd->free_head = context->next;
+	spin_unlock(&cmd->context_lock);
+
+	reinit_completion(&context->done);
+
+	if (mlx4_comm_cmd_post(dev, vhcr_cmd, param)) {
+		/* Only in case the device state is INTERNAL_ERROR,
+		 * mlx4_comm_cmd_post returns with an error
+		 */
+		err = mlx4_status_to_errno(CMD_STAT_INTERNAL_ERR);
+		goto out;
+	}
+
+	if (!wait_for_completion_timeout(&context->done,
+					 msecs_to_jiffies(timeout))) {
+		mlx4_warn(dev, "communication channel command 0x%x (op=0x%x) timed out\n",
+			  vhcr_cmd, op);
+		goto out_reset;
+	}
+
+	err = context->result;
+	if (err && context->fw_status != CMD_STAT_MULTI_FUNC_REQ) {
+		mlx4_err(dev, "command 0x%x failed: fw status = 0x%x\n",
+			 vhcr_cmd, context->fw_status);
+		if (mlx4_closing_cmd_fatal_error(op, context->fw_status))
+			goto out_reset;
+	}
+
+	/* wait for comm channel ready
+	 * this is necessary for prevention the race
+	 * when switching between event to polling mode
+	 * Skipping this section in case the device is in FATAL_ERROR state,
+	 * In this state, no commands are sent via the comm channel until
+	 * the device has returned from reset.
+	 */
+	if (!(dev->persist->state & MLX4_DEVICE_STATE_INTERNAL_ERROR)) {
+		end = msecs_to_jiffies(timeout) + jiffies;
+		while (comm_pending(dev) && time_before(jiffies, end))
+			cond_resched();
+	}
+	goto out;
+
+out_reset:
+	err = mlx4_status_to_errno(CMD_STAT_INTERNAL_ERR);
+	mlx4_enter_error_state(dev->persist);
+out:
+	spin_lock(&cmd->context_lock);
+	context->next = cmd->free_head;
+	cmd->free_head = context - cmd->context;
+	spin_unlock(&cmd->context_lock);
+
+	up(&cmd->event_sem);
+	return err;
+}
+
+int mlx4_comm_cmd(struct mlx4_dev *dev, u8 cmd, u16 param,
+		  u16 op, unsigned long timeout)
+{
+	if (dev->persist->state & MLX4_DEVICE_STATE_INTERNAL_ERROR)
+		return mlx4_status_to_errno(CMD_STAT_INTERNAL_ERR);
+
+	if (mlx4_priv(dev)->cmd.use_events)
+		return mlx4_comm_cmd_wait(dev, cmd, param, op, timeout);
+	return mlx4_comm_cmd_poll(dev, cmd, param, timeout);
+}
+
+static int cmd_pending(struct mlx4_dev *dev)
+{
+	u32 status;
+
+	if (pci_channel_offline(dev->persist->pdev))
+		return -EIO;
+
+	status = readl(mlx4_priv(dev)->cmd.hcr + HCR_STATUS_OFFSET);
+
+	return (status & swab32(1 << HCR_GO_BIT)) ||
+		(mlx4_priv(dev)->cmd.toggle ==
+		 !!(status & swab32(1 << HCR_T_BIT)));
+}
+
+static int mlx4_cmd_post(struct mlx4_dev *dev, u64 in_param, u64 out_param,
+			 u32 in_modifier, u8 op_modifier, u16 op, u16 token,
+			 int event)
+{
+	struct mlx4_cmd *cmd = &mlx4_priv(dev)->cmd;
+	u32 __iomem *hcr = cmd->hcr;
+	int ret = -EIO;
+	unsigned long end;
+
+	mutex_lock(&dev->persist->device_state_mutex);
+	/* To avoid writing to unknown addresses after the device state was
+	  * changed to internal error and the chip was reset,
+	  * check the INTERNAL_ERROR flag which is updated under
+	  * device_state_mutex lock.
+	  */
+	if (pci_channel_offline(dev->persist->pdev) ||
+	    (dev->persist->state & MLX4_DEVICE_STATE_INTERNAL_ERROR)) {
+		/*
+		 * Device is going through error recovery
+		 * and cannot accept commands.
+		 */
+		goto out;
+	}
+
+	end = jiffies;
+	if (event)
+		end += msecs_to_jiffies(GO_BIT_TIMEOUT_MSECS);
+
+	while (cmd_pending(dev)) {
+		if (pci_channel_offline(dev->persist->pdev)) {
+			/*
+			 * Device is going through error recovery
+			 * and cannot accept commands.
+			 */
+			goto out;
+		}
+
+		if (time_after_eq(jiffies, end)) {
+			mlx4_err(dev, "%s:cmd_pending failed\n", __func__);
+			goto out;
+		}
+		cond_resched();
+	}
+
+	/*
+	 * We use writel (instead of something like memcpy_toio)
+	 * because writes of less than 32 bits to the HCR don't work
+	 * (and some architectures such as ia64 implement memcpy_toio
+	 * in terms of writeb).
+	 */
+	__raw_writel((__force u32) cpu_to_be32(in_param >> 32),		  hcr + 0);
+	__raw_writel((__force u32) cpu_to_be32(in_param & 0xfffffffful),  hcr + 1);
+	__raw_writel((__force u32) cpu_to_be32(in_modifier),		  hcr + 2);
+	__raw_writel((__force u32) cpu_to_be32(out_param >> 32),	  hcr + 3);
+	__raw_writel((__force u32) cpu_to_be32(out_param & 0xfffffffful), hcr + 4);
+	__raw_writel((__force u32) cpu_to_be32(token << 16),		  hcr + 5);
+
+	/* __raw_writel may not order writes. */
+	wmb();
+
+	__raw_writel((__force u32) cpu_to_be32((1 << HCR_GO_BIT)		|
+					       (cmd->toggle << HCR_T_BIT)	|
+					       (event ? (1 << HCR_E_BIT) : 0)	|
+					       (op_modifier << HCR_OPMOD_SHIFT) |
+					       op), hcr + 6);
+
+	/*
+	 * Make sure that our HCR writes don't get mixed in with
+	 * writes from another CPU starting a FW command.
+	 */
+	mmiowb();
+
+	cmd->toggle = cmd->toggle ^ 1;
+
+	ret = 0;
+
+out:
+	if (ret)
+		mlx4_warn(dev, "Could not post command 0x%x: ret=%d, in_param=0x%llx, in_mod=0x%x, op_mod=0x%x\n",
+			  op, ret, in_param, in_modifier, op_modifier);
+	mutex_unlock(&dev->persist->device_state_mutex);
+
+	return ret;
+}
+
+static int mlx4_slave_cmd(struct mlx4_dev *dev, u64 in_param, u64 *out_param,
+			  int out_is_imm, u32 in_modifier, u8 op_modifier,
+			  u16 op, unsigned long timeout)
+{
+	struct mlx4_priv *priv = mlx4_priv(dev);
+	struct mlx4_vhcr_cmd *vhcr = priv->mfunc.vhcr;
+	int ret;
+
+	mutex_lock(&priv->cmd.slave_cmd_mutex);
+
+	vhcr->in_param = cpu_to_be64(in_param);
+	vhcr->out_param = out_param ? cpu_to_be64(*out_param) : 0;
+	vhcr->in_modifier = cpu_to_be32(in_modifier);
+	vhcr->opcode = cpu_to_be16((((u16) op_modifier) << 12) | (op & 0xfff));
+	vhcr->token = cpu_to_be16(CMD_POLL_TOKEN);
+	vhcr->status = 0;
+	vhcr->flags = !!(priv->cmd.use_events) << 6;
+
+	if (mlx4_is_master(dev)) {
+		ret = mlx4_master_process_vhcr(dev, dev->caps.function, vhcr);
+		if (!ret) {
+			if (out_is_imm) {
+				if (out_param)
+					*out_param =
+						be64_to_cpu(vhcr->out_param);
+				else {
+					mlx4_err(dev, "response expected while output mailbox is NULL for command 0x%x\n",
+						 op);
+					vhcr->status = CMD_STAT_BAD_PARAM;
+				}
+			}
+			ret = mlx4_status_to_errno(vhcr->status);
+		}
+		if (ret &&
+		    dev->persist->state & MLX4_DEVICE_STATE_INTERNAL_ERROR)
+			ret = mlx4_internal_err_ret_value(dev, op, op_modifier);
+	} else {
+		ret = mlx4_comm_cmd(dev, MLX4_COMM_CMD_VHCR_POST, 0, op,
+				    MLX4_COMM_TIME + timeout);
+		if (!ret) {
+			if (out_is_imm) {
+				if (out_param)
+					*out_param =
+						be64_to_cpu(vhcr->out_param);
+				else {
+					mlx4_err(dev, "response expected while output mailbox is NULL for command 0x%x\n",
+						 op);
+					vhcr->status = CMD_STAT_BAD_PARAM;
+				}
+			}
+			ret = mlx4_status_to_errno(vhcr->status);
+		} else {
+			if (dev->persist->state &
+			    MLX4_DEVICE_STATE_INTERNAL_ERROR)
+				ret = mlx4_internal_err_ret_value(dev, op,
+								  op_modifier);
+			else
+				mlx4_err(dev, "failed execution of VHCR_POST command opcode 0x%x\n", op);
+		}
+	}
+
+	mutex_unlock(&priv->cmd.slave_cmd_mutex);
+	return ret;
+}
+
+static int mlx4_cmd_poll(struct mlx4_dev *dev, u64 in_param, u64 *out_param,
+			 int out_is_imm, u32 in_modifier, u8 op_modifier,
+			 u16 op, unsigned long timeout)
+{
+	struct mlx4_priv *priv = mlx4_priv(dev);
+	void __iomem *hcr = priv->cmd.hcr;
+	int err = 0;
+	unsigned long end;
+	u32 stat;
+
+	down(&priv->cmd.poll_sem);
+
+	if (dev->persist->state & MLX4_DEVICE_STATE_INTERNAL_ERROR) {
+		/*
+		 * Device is going through error recovery
+		 * and cannot accept commands.
+		 */
+		err = mlx4_internal_err_ret_value(dev, op, op_modifier);
+		goto out;
+	}
+
+	if (out_is_imm && !out_param) {
+		mlx4_err(dev, "response expected while output mailbox is NULL for command 0x%x\n",
+			 op);
+		err = -EINVAL;
+		goto out;
+	}
+
+	err = mlx4_cmd_post(dev, in_param, out_param ? *out_param : 0,
+			    in_modifier, op_modifier, op, CMD_POLL_TOKEN, 0);
+	if (err)
+		goto out_reset;
+
+	end = msecs_to_jiffies(timeout) + jiffies;
+	while (cmd_pending(dev) && time_before(jiffies, end)) {
+		if (pci_channel_offline(dev->persist->pdev)) {
+			/*
+			 * Device is going through error recovery
+			 * and cannot accept commands.
+			 */
+			err = -EIO;
+			goto out_reset;
+		}
+
+		if (dev->persist->state & MLX4_DEVICE_STATE_INTERNAL_ERROR) {
+			err = mlx4_internal_err_ret_value(dev, op, op_modifier);
+			goto out;
+		}
+
+		cond_resched();
+	}
+
+	if (cmd_pending(dev)) {
+		mlx4_warn(dev, "command 0x%x timed out (go bit not cleared)\n",
+			  op);
+		err = -EIO;
+		goto out_reset;
+	}
+
+	if (out_is_imm)
+		*out_param =
+			(u64) be32_to_cpu((__force __be32)
+					  __raw_readl(hcr + HCR_OUT_PARAM_OFFSET)) << 32 |
+			(u64) be32_to_cpu((__force __be32)
+					  __raw_readl(hcr + HCR_OUT_PARAM_OFFSET + 4));
+	stat = be32_to_cpu((__force __be32)
+			   __raw_readl(hcr + HCR_STATUS_OFFSET)) >> 24;
+	err = mlx4_status_to_errno(stat);
+	if (err) {
+		mlx4_err(dev, "command 0x%x failed: fw status = 0x%x\n",
+			 op, stat);
+		if (mlx4_closing_cmd_fatal_error(op, stat))
+			goto out_reset;
+		goto out;
+	}
+
+out_reset:
+	if (err)
+		err = mlx4_cmd_reset_flow(dev, op, op_modifier, err);
+out:
+	up(&priv->cmd.poll_sem);
+	return err;
+}
+
+void mlx4_cmd_event(struct mlx4_dev *dev, u16 token, u8 status, u64 out_param)
+{
+	struct mlx4_priv *priv = mlx4_priv(dev);
+	struct mlx4_cmd_context *context =
+		&priv->cmd.context[token & priv->cmd.token_mask];
+
+	/* previously timed out command completing at long last */
+	if (token != context->token)
+		return;
+
+	context->fw_status = status;
+	context->result    = mlx4_status_to_errno(status);
+	context->out_param = out_param;
+
+	complete(&context->done);
+}
+
+static int mlx4_cmd_wait(struct mlx4_dev *dev, u64 in_param, u64 *out_param,
+			 int out_is_imm, u32 in_modifier, u8 op_modifier,
+			 u16 op, unsigned long timeout)
+{
+	struct mlx4_cmd *cmd = &mlx4_priv(dev)->cmd;
+	struct mlx4_cmd_context *context;
+	long ret_wait;
+	int err = 0;
+
+	down(&cmd->event_sem);
+
+	spin_lock(&cmd->context_lock);
+	BUG_ON(cmd->free_head < 0);
+	context = &cmd->context[cmd->free_head];
+	context->token += cmd->token_mask + 1;
+	cmd->free_head = context->next;
+	spin_unlock(&cmd->context_lock);
+
+	if (out_is_imm && !out_param) {
+		mlx4_err(dev, "response expected while output mailbox is NULL for command 0x%x\n",
+			 op);
+		err = -EINVAL;
+		goto out;
+	}
+
+	reinit_completion(&context->done);
+
+	err = mlx4_cmd_post(dev, in_param, out_param ? *out_param : 0,
+			    in_modifier, op_modifier, op, context->token, 1);
+	if (err)
+		goto out_reset;
+
+	if (op == MLX4_CMD_SENSE_PORT) {
+		ret_wait =
+			wait_for_completion_interruptible_timeout(&context->done,
+								  msecs_to_jiffies(timeout));
+		if (ret_wait < 0) {
+			context->fw_status = 0;
+			context->out_param = 0;
+			context->result = 0;
+		}
+	} else {
+		ret_wait = (long)wait_for_completion_timeout(&context->done,
+							     msecs_to_jiffies(timeout));
+	}
+	if (!ret_wait) {
+		mlx4_warn(dev, "command 0x%x timed out (go bit not cleared)\n",
+			  op);
+		if (op == MLX4_CMD_NOP) {
+			err = -EBUSY;
+			goto out;
+		} else {
+			err = -EIO;
+			goto out_reset;
+		}
+	}
+
+	err = context->result;
+	if (err) {
+		/* Since we do not want to have this error message always
+		 * displayed at driver start when there are ConnectX2 HCAs
+		 * on the host, we deprecate the error message for this
+		 * specific command/input_mod/opcode_mod/fw-status to be debug.
+		 */
+		if (op == MLX4_CMD_SET_PORT &&
+		    (in_modifier == 1 || in_modifier == 2) &&
+		    op_modifier == MLX4_SET_PORT_IB_OPCODE &&
+		    context->fw_status == CMD_STAT_BAD_SIZE)
+			mlx4_dbg(dev, "command 0x%x failed: fw status = 0x%x\n",
+				 op, context->fw_status);
+		else
+			mlx4_err(dev, "command 0x%x failed: fw status = 0x%x\n",
+				 op, context->fw_status);
+		if (dev->persist->state & MLX4_DEVICE_STATE_INTERNAL_ERROR)
+			err = mlx4_internal_err_ret_value(dev, op, op_modifier);
+		else if (mlx4_closing_cmd_fatal_error(op, context->fw_status))
+			goto out_reset;
+
+		goto out;
+	}
+
+	if (out_is_imm)
+		*out_param = context->out_param;
+
+out_reset:
+	if (err)
+		err = mlx4_cmd_reset_flow(dev, op, op_modifier, err);
+out:
+	spin_lock(&cmd->context_lock);
+	context->next = cmd->free_head;
+	cmd->free_head = context - cmd->context;
+	spin_unlock(&cmd->context_lock);
+
+	up(&cmd->event_sem);
+	return err;
+}
+
+int __mlx4_cmd(struct mlx4_dev *dev, u64 in_param, u64 *out_param,
+	       int out_is_imm, u32 in_modifier, u8 op_modifier,
+	       u16 op, unsigned long timeout, int native)
+{
+	if (pci_channel_offline(dev->persist->pdev))
+		return mlx4_cmd_reset_flow(dev, op, op_modifier, -EIO);
+
+	if (!mlx4_is_mfunc(dev) || (native && mlx4_is_master(dev))) {
+		int ret;
+
+		if (dev->persist->state & MLX4_DEVICE_STATE_INTERNAL_ERROR)
+			return mlx4_internal_err_ret_value(dev, op,
+							  op_modifier);
+		down_read(&mlx4_priv(dev)->cmd.switch_sem);
+		if (mlx4_priv(dev)->cmd.use_events)
+			ret = mlx4_cmd_wait(dev, in_param, out_param,
+					    out_is_imm, in_modifier,
+					    op_modifier, op, timeout);
+		else
+			ret = mlx4_cmd_poll(dev, in_param, out_param,
+					    out_is_imm, in_modifier,
+					    op_modifier, op, timeout);
+
+		up_read(&mlx4_priv(dev)->cmd.switch_sem);
+		return ret;
+	}
+	return mlx4_slave_cmd(dev, in_param, out_param, out_is_imm,
+			      in_modifier, op_modifier, op, timeout);
+}
+EXPORT_SYMBOL_GPL(__mlx4_cmd);
+
+
+int mlx4_ARM_COMM_CHANNEL(struct mlx4_dev *dev)
+{
+	return mlx4_cmd(dev, 0, 0, 0, MLX4_CMD_ARM_COMM_CHANNEL,
+			MLX4_CMD_TIME_CLASS_B, MLX4_CMD_NATIVE);
+}
+
+static int mlx4_ACCESS_MEM(struct mlx4_dev *dev, u64 master_addr,
+			   int slave, u64 slave_addr,
+			   int size, int is_read)
+{
+	u64 in_param;
+	u64 out_param;
+
+	if ((slave_addr & 0xfff) | (master_addr & 0xfff) |
+	    (slave & ~0x7f) | (size & 0xff)) {
+		mlx4_err(dev, "Bad access mem params - slave_addr:0x%llx master_addr:0x%llx slave_id:%d size:%d\n",
+			 slave_addr, master_addr, slave, size);
+		return -EINVAL;
+	}
+
+	if (is_read) {
+		in_param = (u64) slave | slave_addr;
+		out_param = (u64) dev->caps.function | master_addr;
+	} else {
+		in_param = (u64) dev->caps.function | master_addr;
+		out_param = (u64) slave | slave_addr;
+	}
+
+	return mlx4_cmd_imm(dev, in_param, &out_param, size, 0,
+			    MLX4_CMD_ACCESS_MEM,
+			    MLX4_CMD_TIME_CLASS_A, MLX4_CMD_NATIVE);
+}
+
+static int query_pkey_block(struct mlx4_dev *dev, u8 port, u16 index, u16 *pkey,
+			       struct mlx4_cmd_mailbox *inbox,
+			       struct mlx4_cmd_mailbox *outbox)
+{
+	struct ib_smp *in_mad = (struct ib_smp *)(inbox->buf);
+	struct ib_smp *out_mad = (struct ib_smp *)(outbox->buf);
+	int err;
+	int i;
+
+	if (index & 0x1f)
+		return -EINVAL;
+
+	in_mad->attr_mod = cpu_to_be32(index / 32);
+
+	err = mlx4_cmd_box(dev, inbox->dma, outbox->dma, port, 3,
+			   MLX4_CMD_MAD_IFC, MLX4_CMD_TIME_CLASS_C,
+			   MLX4_CMD_NATIVE);
+	if (err)
+		return err;
+
+	for (i = 0; i < 32; ++i)
+		pkey[i] = be16_to_cpu(((__be16 *) out_mad->data)[i]);
+
+	return err;
+}
+
+static int get_full_pkey_table(struct mlx4_dev *dev, u8 port, u16 *table,
+			       struct mlx4_cmd_mailbox *inbox,
+			       struct mlx4_cmd_mailbox *outbox)
+{
+	int i;
+	int err;
+
+	for (i = 0; i < dev->caps.pkey_table_len[port]; i += 32) {
+		err = query_pkey_block(dev, port, i, table + i, inbox, outbox);
+		if (err)
+			return err;
+	}
+
+	return 0;
+}
+#define PORT_CAPABILITY_LOCATION_IN_SMP 20
+#define PORT_STATE_OFFSET 32
+
+static enum ib_port_state vf_port_state(struct mlx4_dev *dev, int port, int vf)
+{
+	if (mlx4_get_slave_port_state(dev, vf, port) == SLAVE_PORT_UP)
+		return IB_PORT_ACTIVE;
+	else
+		return IB_PORT_DOWN;
+}
+
+static int mlx4_MAD_IFC_wrapper(struct mlx4_dev *dev, int slave,
+				struct mlx4_vhcr *vhcr,
+				struct mlx4_cmd_mailbox *inbox,
+				struct mlx4_cmd_mailbox *outbox,
+				struct mlx4_cmd_info *cmd)
+{
+	struct ib_smp *smp = inbox->buf;
+	u32 index;
+	u8 port, slave_port;
+	u8 opcode_modifier;
+	u16 *table;
+	int err;
+	int vidx, pidx;
+	int network_view;
+	struct mlx4_priv *priv = mlx4_priv(dev);
+	struct ib_smp *outsmp = outbox->buf;
+	__be16 *outtab = (__be16 *)(outsmp->data);
+	__be32 slave_cap_mask;
+	__be64 slave_node_guid;
+
+	slave_port = vhcr->in_modifier;
+	port = mlx4_slave_convert_port(dev, slave, slave_port);
+
+	/* network-view bit is for driver use only, and should not be passed to FW */
+	opcode_modifier = vhcr->op_modifier & ~0x8; /* clear netw view bit */
+	network_view = !!(vhcr->op_modifier & 0x8);
+
+	if (smp->base_version == 1 &&
+	    smp->mgmt_class == IB_MGMT_CLASS_SUBN_LID_ROUTED &&
+	    smp->class_version == 1) {
+		/* host view is paravirtualized */
+		if (!network_view && smp->method == IB_MGMT_METHOD_GET) {
+			if (smp->attr_id == IB_SMP_ATTR_PKEY_TABLE) {
+				index = be32_to_cpu(smp->attr_mod);
+				if (port < 1 || port > dev->caps.num_ports)
+					return -EINVAL;
+				table = kcalloc((dev->caps.pkey_table_len[port] / 32) + 1,
+						sizeof(*table) * 32, GFP_KERNEL);
+
+				if (!table)
+					return -ENOMEM;
+				/* need to get the full pkey table because the paravirtualized
+				 * pkeys may be scattered among several pkey blocks.
+				 */
+				err = get_full_pkey_table(dev, port, table, inbox, outbox);
+				if (!err) {
+					for (vidx = index * 32; vidx < (index + 1) * 32; ++vidx) {
+						pidx = priv->virt2phys_pkey[slave][port - 1][vidx];
+						outtab[vidx % 32] = cpu_to_be16(table[pidx]);
+					}
+				}
+				kfree(table);
+				return err;
+			}
+			if (smp->attr_id == IB_SMP_ATTR_PORT_INFO) {
+				/*get the slave specific caps:*/
+				/*do the command */
+				smp->attr_mod = cpu_to_be32(port);
+				err = mlx4_cmd_box(dev, inbox->dma, outbox->dma,
+					    port, opcode_modifier,
+					    vhcr->op, MLX4_CMD_TIME_CLASS_C, MLX4_CMD_NATIVE);
+				/* modify the response for slaves */
+				if (!err && slave != mlx4_master_func_num(dev)) {
+					u8 *state = outsmp->data + PORT_STATE_OFFSET;
+
+					*state = (*state & 0xf0) | vf_port_state(dev, port, slave);
+					slave_cap_mask = priv->mfunc.master.slave_state[slave].ib_cap_mask[port];
+					memcpy(outsmp->data + PORT_CAPABILITY_LOCATION_IN_SMP, &slave_cap_mask, 4);
+				}
+				return err;
+			}
+			if (smp->attr_id == IB_SMP_ATTR_GUID_INFO) {
+				__be64 guid = mlx4_get_admin_guid(dev, slave,
+								  port);
+
+				/* set the PF admin guid to the FW/HW burned
+				 * GUID, if it wasn't yet set
+				 */
+				if (slave == 0 && guid == 0) {
+					smp->attr_mod = 0;
+					err = mlx4_cmd_box(dev,
+							   inbox->dma,
+							   outbox->dma,
+							   vhcr->in_modifier,
+							   opcode_modifier,
+							   vhcr->op,
+							   MLX4_CMD_TIME_CLASS_C,
+							   MLX4_CMD_NATIVE);
+					if (err)
+						return err;
+					mlx4_set_admin_guid(dev,
+							    *(__be64 *)outsmp->
+							    data, slave, port);
+				} else {
+					memcpy(outsmp->data, &guid, 8);
+				}
+
+				/* clean all other gids */
+				memset(outsmp->data + 8, 0, 56);
+				return 0;
+			}
+			if (smp->attr_id == IB_SMP_ATTR_NODE_INFO) {
+				err = mlx4_cmd_box(dev, inbox->dma, outbox->dma,
+					     port, opcode_modifier,
+					     vhcr->op, MLX4_CMD_TIME_CLASS_C, MLX4_CMD_NATIVE);
+				if (!err) {
+					slave_node_guid =  mlx4_get_slave_node_guid(dev, slave);
+					memcpy(outsmp->data + 12, &slave_node_guid, 8);
+				}
+				return err;
+			}
+		}
+	}
+
+	/* Non-privileged VFs are only allowed "host" view LID-routed 'Get' MADs.
+	 * These are the MADs used by ib verbs (such as ib_query_gids).
+	 */
+	if (slave != mlx4_master_func_num(dev) &&
+	    !mlx4_vf_smi_enabled(dev, slave, port)) {
+		if (!(smp->mgmt_class == IB_MGMT_CLASS_SUBN_LID_ROUTED &&
+		      smp->method == IB_MGMT_METHOD_GET) || network_view) {
+			mlx4_err(dev, "Unprivileged slave %d is trying to execute a Subnet MGMT MAD, class 0x%x, method 0x%x, view=%s for attr 0x%x. Rejecting\n",
+				 slave, smp->mgmt_class, smp->method,
+				 network_view ? "Network" : "Host",
+				 be16_to_cpu(smp->attr_id));
+			return -EPERM;
+		}
+	}
+
+	return mlx4_cmd_box(dev, inbox->dma, outbox->dma,
+				    vhcr->in_modifier, opcode_modifier,
+				    vhcr->op, MLX4_CMD_TIME_CLASS_C, MLX4_CMD_NATIVE);
+}
+
+static int mlx4_CMD_EPERM_wrapper(struct mlx4_dev *dev, int slave,
+		     struct mlx4_vhcr *vhcr,
+		     struct mlx4_cmd_mailbox *inbox,
+		     struct mlx4_cmd_mailbox *outbox,
+		     struct mlx4_cmd_info *cmd)
+{
+	return -EPERM;
+}
+
+int mlx4_DMA_wrapper(struct mlx4_dev *dev, int slave,
+		     struct mlx4_vhcr *vhcr,
+		     struct mlx4_cmd_mailbox *inbox,
+		     struct mlx4_cmd_mailbox *outbox,
+		     struct mlx4_cmd_info *cmd)
+{
+	u64 in_param;
+	u64 out_param;
+	int err;
+
+	in_param = cmd->has_inbox ? (u64) inbox->dma : vhcr->in_param;
+	out_param = cmd->has_outbox ? (u64) outbox->dma : vhcr->out_param;
+	if (cmd->encode_slave_id) {
+		in_param &= 0xffffffffffffff00ll;
+		in_param |= slave;
+	}
+
+	err = __mlx4_cmd(dev, in_param, &out_param, cmd->out_is_imm,
+			 vhcr->in_modifier, vhcr->op_modifier, vhcr->op,
+			 MLX4_CMD_TIME_CLASS_A, MLX4_CMD_NATIVE);
+
+	if (cmd->out_is_imm)
+		vhcr->out_param = out_param;
+
+	return err;
+}
+
+static struct mlx4_cmd_info cmd_info[] = {
+	{
+		.opcode = MLX4_CMD_QUERY_FW,
+		.has_inbox = false,
+		.has_outbox = true,
+		.out_is_imm = false,
+		.encode_slave_id = false,
+		.verify = NULL,
+		.wrapper = mlx4_QUERY_FW_wrapper
+	},
+	{
+		.opcode = MLX4_CMD_QUERY_HCA,
+		.has_inbox = false,
+		.has_outbox = true,
+		.out_is_imm = false,
+		.encode_slave_id = false,
+		.verify = NULL,
+		.wrapper = NULL
+	},
+	{
+		.opcode = MLX4_CMD_QUERY_DEV_CAP,
+		.has_inbox = false,
+		.has_outbox = true,
+		.out_is_imm = false,
+		.encode_slave_id = false,
+		.verify = NULL,
+		.wrapper = mlx4_QUERY_DEV_CAP_wrapper
+	},
+	{
+		.opcode = MLX4_CMD_QUERY_FUNC_CAP,
+		.has_inbox = false,
+		.has_outbox = true,
+		.out_is_imm = false,
+		.encode_slave_id = false,
+		.verify = NULL,
+		.wrapper = mlx4_QUERY_FUNC_CAP_wrapper
+	},
+	{
+		.opcode = MLX4_CMD_QUERY_ADAPTER,
+		.has_inbox = false,
+		.has_outbox = true,
+		.out_is_imm = false,
+		.encode_slave_id = false,
+		.verify = NULL,
+		.wrapper = NULL
+	},
+	{
+		.opcode = MLX4_CMD_INIT_PORT,
+		.has_inbox = false,
+		.has_outbox = false,
+		.out_is_imm = false,
+		.encode_slave_id = false,
+		.verify = NULL,
+		.wrapper = mlx4_INIT_PORT_wrapper
+	},
+	{
+		.opcode = MLX4_CMD_CLOSE_PORT,
+		.has_inbox = false,
+		.has_outbox = false,
+		.out_is_imm  = false,
+		.encode_slave_id = false,
+		.verify = NULL,
+		.wrapper = mlx4_CLOSE_PORT_wrapper
+	},
+	{
+		.opcode = MLX4_CMD_QUERY_PORT,
+		.has_inbox = false,
+		.has_outbox = true,
+		.out_is_imm = false,
+		.encode_slave_id = false,
+		.verify = NULL,
+		.wrapper = mlx4_QUERY_PORT_wrapper
+	},
+	{
+		.opcode = MLX4_CMD_SET_PORT,
+		.has_inbox = true,
+		.has_outbox = false,
+		.out_is_imm = false,
+		.encode_slave_id = false,
+		.verify = NULL,
+		.wrapper = mlx4_SET_PORT_wrapper
+	},
+	{
+		.opcode = MLX4_CMD_MAP_EQ,
+		.has_inbox = false,
+		.has_outbox = false,
+		.out_is_imm = false,
+		.encode_slave_id = false,
+		.verify = NULL,
+		.wrapper = mlx4_MAP_EQ_wrapper
+	},
+	{
+		.opcode = MLX4_CMD_SW2HW_EQ,
+		.has_inbox = true,
+		.has_outbox = false,
+		.out_is_imm = false,
+		.encode_slave_id = true,
+		.verify = NULL,
+		.wrapper = mlx4_SW2HW_EQ_wrapper
+	},
+	{
+		.opcode = MLX4_CMD_HW_HEALTH_CHECK,
+		.has_inbox = false,
+		.has_outbox = false,
+		.out_is_imm = false,
+		.encode_slave_id = false,
+		.verify = NULL,
+		.wrapper = NULL
+	},
+	{
+		.opcode = MLX4_CMD_NOP,
+		.has_inbox = false,
+		.has_outbox = false,
+		.out_is_imm = false,
+		.encode_slave_id = false,
+		.verify = NULL,
+		.wrapper = NULL
+	},
+	{
+		.opcode = MLX4_CMD_CONFIG_DEV,
+		.has_inbox = false,
+		.has_outbox = true,
+		.out_is_imm = false,
+		.encode_slave_id = false,
+		.verify = NULL,
+		.wrapper = mlx4_CONFIG_DEV_wrapper
+	},
+	{
+		.opcode = MLX4_CMD_ALLOC_RES,
+		.has_inbox = false,
+		.has_outbox = false,
+		.out_is_imm = true,
+		.encode_slave_id = false,
+		.verify = NULL,
+		.wrapper = mlx4_ALLOC_RES_wrapper
+	},
+	{
+		.opcode = MLX4_CMD_FREE_RES,
+		.has_inbox = false,
+		.has_outbox = false,
+		.out_is_imm = false,
+		.encode_slave_id = false,
+		.verify = NULL,
+		.wrapper = mlx4_FREE_RES_wrapper
+	},
+	{
+		.opcode = MLX4_CMD_SW2HW_MPT,
+		.has_inbox = true,
+		.has_outbox = false,
+		.out_is_imm = false,
+		.encode_slave_id = true,
+		.verify = NULL,
+		.wrapper = mlx4_SW2HW_MPT_wrapper
+	},
+	{
+		.opcode = MLX4_CMD_QUERY_MPT,
+		.has_inbox = false,
+		.has_outbox = true,
+		.out_is_imm = false,
+		.encode_slave_id = false,
+		.verify = NULL,
+		.wrapper = mlx4_QUERY_MPT_wrapper
+	},
+	{
+		.opcode = MLX4_CMD_HW2SW_MPT,
+		.has_inbox = false,
+		.has_outbox = false,
+		.out_is_imm = false,
+		.encode_slave_id = false,
+		.verify = NULL,
+		.wrapper = mlx4_HW2SW_MPT_wrapper
+	},
+	{
+		.opcode = MLX4_CMD_READ_MTT,
+		.has_inbox = false,
+		.has_outbox = true,
+		.out_is_imm = false,
+		.encode_slave_id = false,
+		.verify = NULL,
+		.wrapper = NULL
+	},
+	{
+		.opcode = MLX4_CMD_WRITE_MTT,
+		.has_inbox = true,
+		.has_outbox = false,
+		.out_is_imm = false,
+		.encode_slave_id = false,
+		.verify = NULL,
+		.wrapper = mlx4_WRITE_MTT_wrapper
+	},
+	{
+		.opcode = MLX4_CMD_SYNC_TPT,
+		.has_inbox = true,
+		.has_outbox = false,
+		.out_is_imm = false,
+		.encode_slave_id = false,
+		.verify = NULL,
+		.wrapper = NULL
+	},
+	{
+		.opcode = MLX4_CMD_HW2SW_EQ,
+		.has_inbox = false,
+		.has_outbox = false,
+		.out_is_imm = false,
+		.encode_slave_id = true,
+		.verify = NULL,
+		.wrapper = mlx4_HW2SW_EQ_wrapper
+	},
+	{
+		.opcode = MLX4_CMD_QUERY_EQ,
+		.has_inbox = false,
+		.has_outbox = true,
+		.out_is_imm = false,
+		.encode_slave_id = true,
+		.verify = NULL,
+		.wrapper = mlx4_QUERY_EQ_wrapper
+	},
+	{
+		.opcode = MLX4_CMD_SW2HW_CQ,
+		.has_inbox = true,
+		.has_outbox = false,
+		.out_is_imm = false,
+		.encode_slave_id = true,
+		.verify = NULL,
+		.wrapper = mlx4_SW2HW_CQ_wrapper
+	},
+	{
+		.opcode = MLX4_CMD_HW2SW_CQ,
+		.has_inbox = false,
+		.has_outbox = false,
+		.out_is_imm = false,
+		.encode_slave_id = false,
+		.verify = NULL,
+		.wrapper = mlx4_HW2SW_CQ_wrapper
+	},
+	{
+		.opcode = MLX4_CMD_QUERY_CQ,
+		.has_inbox = false,
+		.has_outbox = true,
+		.out_is_imm = false,
+		.encode_slave_id = false,
+		.verify = NULL,
+		.wrapper = mlx4_QUERY_CQ_wrapper
+	},
+	{
+		.opcode = MLX4_CMD_MODIFY_CQ,
+		.has_inbox = true,
+		.has_outbox = false,
+		.out_is_imm = true,
+		.encode_slave_id = false,
+		.verify = NULL,
+		.wrapper = mlx4_MODIFY_CQ_wrapper
+	},
+	{
+		.opcode = MLX4_CMD_SW2HW_SRQ,
+		.has_inbox = true,
+		.has_outbox = false,
+		.out_is_imm = false,
+		.encode_slave_id = true,
+		.verify = NULL,
+		.wrapper = mlx4_SW2HW_SRQ_wrapper
+	},
+	{
+		.opcode = MLX4_CMD_HW2SW_SRQ,
+		.has_inbox = false,
+		.has_outbox = false,
+		.out_is_imm = false,
+		.encode_slave_id = false,
+		.verify = NULL,
+		.wrapper = mlx4_HW2SW_SRQ_wrapper
+	},
+	{
+		.opcode = MLX4_CMD_QUERY_SRQ,
+		.has_inbox = false,
+		.has_outbox = true,
+		.out_is_imm = false,
+		.encode_slave_id = false,
+		.verify = NULL,
+		.wrapper = mlx4_QUERY_SRQ_wrapper
+	},
+	{
+		.opcode = MLX4_CMD_ARM_SRQ,
+		.has_inbox = false,
+		.has_outbox = false,
+		.out_is_imm = false,
+		.encode_slave_id = false,
+		.verify = NULL,
+		.wrapper = mlx4_ARM_SRQ_wrapper
+	},
+	{
+		.opcode = MLX4_CMD_RST2INIT_QP,
+		.has_inbox = true,
+		.has_outbox = false,
+		.out_is_imm = false,
+		.encode_slave_id = true,
+		.verify = NULL,
+		.wrapper = mlx4_RST2INIT_QP_wrapper
+	},
+	{
+		.opcode = MLX4_CMD_INIT2INIT_QP,
+		.has_inbox = true,
+		.has_outbox = false,
+		.out_is_imm = false,
+		.encode_slave_id = false,
+		.verify = NULL,
+		.wrapper = mlx4_INIT2INIT_QP_wrapper
+	},
+	{
+		.opcode = MLX4_CMD_INIT2RTR_QP,
+		.has_inbox = true,
+		.has_outbox = false,
+		.out_is_imm = false,
+		.encode_slave_id = false,
+		.verify = NULL,
+		.wrapper = mlx4_INIT2RTR_QP_wrapper
+	},
+	{
+		.opcode = MLX4_CMD_RTR2RTS_QP,
+		.has_inbox = true,
+		.has_outbox = false,
+		.out_is_imm = false,
+		.encode_slave_id = false,
+		.verify = NULL,
+		.wrapper = mlx4_RTR2RTS_QP_wrapper
+	},
+	{
+		.opcode = MLX4_CMD_RTS2RTS_QP,
+		.has_inbox = true,
+		.has_outbox = false,
+		.out_is_imm = false,
+		.encode_slave_id = false,
+		.verify = NULL,
+		.wrapper = mlx4_RTS2RTS_QP_wrapper
+	},
+	{
+		.opcode = MLX4_CMD_SQERR2RTS_QP,
+		.has_inbox = true,
+		.has_outbox = false,
+		.out_is_imm = false,
+		.encode_slave_id = false,
+		.verify = NULL,
+		.wrapper = mlx4_SQERR2RTS_QP_wrapper
+	},
+	{
+		.opcode = MLX4_CMD_2ERR_QP,
+		.has_inbox = false,
+		.has_outbox = false,
+		.out_is_imm = false,
+		.encode_slave_id = false,
+		.verify = NULL,
+		.wrapper = mlx4_GEN_QP_wrapper
+	},
+	{
+		.opcode = MLX4_CMD_RTS2SQD_QP,
+		.has_inbox = false,
+		.has_outbox = false,
+		.out_is_imm = false,
+		.encode_slave_id = false,
+		.verify = NULL,
+		.wrapper = mlx4_GEN_QP_wrapper
+	},
+	{
+		.opcode = MLX4_CMD_SQD2SQD_QP,
+		.has_inbox = true,
+		.has_outbox = false,
+		.out_is_imm = false,
+		.encode_slave_id = false,
+		.verify = NULL,
+		.wrapper = mlx4_SQD2SQD_QP_wrapper
+	},
+	{
+		.opcode = MLX4_CMD_SQD2RTS_QP,
+		.has_inbox = true,
+		.has_outbox = false,
+		.out_is_imm = false,
+		.encode_slave_id = false,
+		.verify = NULL,
+		.wrapper = mlx4_SQD2RTS_QP_wrapper
+	},
+	{
+		.opcode = MLX4_CMD_2RST_QP,
+		.has_inbox = false,
+		.has_outbox = false,
+		.out_is_imm = false,
+		.encode_slave_id = false,
+		.verify = NULL,
+		.wrapper = mlx4_2RST_QP_wrapper
+	},
+	{
+		.opcode = MLX4_CMD_QUERY_QP,
+		.has_inbox = false,
+		.has_outbox = true,
+		.out_is_imm = false,
+		.encode_slave_id = false,
+		.verify = NULL,
+		.wrapper = mlx4_GEN_QP_wrapper
+	},
+	{
+		.opcode = MLX4_CMD_SUSPEND_QP,
+		.has_inbox = false,
+		.has_outbox = false,
+		.out_is_imm = false,
+		.encode_slave_id = false,
+		.verify = NULL,
+		.wrapper = mlx4_GEN_QP_wrapper
+	},
+	{
+		.opcode = MLX4_CMD_UNSUSPEND_QP,
+		.has_inbox = false,
+		.has_outbox = false,
+		.out_is_imm = false,
+		.encode_slave_id = false,
+		.verify = NULL,
+		.wrapper = mlx4_GEN_QP_wrapper
+	},
+	{
+		.opcode = MLX4_CMD_UPDATE_QP,
+		.has_inbox = true,
+		.has_outbox = false,
+		.out_is_imm = false,
+		.encode_slave_id = false,
+		.verify = NULL,
+		.wrapper = mlx4_UPDATE_QP_wrapper
+	},
+	{
+		.opcode = MLX4_CMD_GET_OP_REQ,
+		.has_inbox = false,
+		.has_outbox = false,
+		.out_is_imm = false,
+		.encode_slave_id = false,
+		.verify = NULL,
+		.wrapper = mlx4_CMD_EPERM_wrapper,
+	},
+	{
+		.opcode = MLX4_CMD_ALLOCATE_VPP,
+		.has_inbox = false,
+		.has_outbox = true,
+		.out_is_imm = false,
+		.encode_slave_id = false,
+		.verify = NULL,
+		.wrapper = mlx4_CMD_EPERM_wrapper,
+	},
+	{
+		.opcode = MLX4_CMD_SET_VPORT_QOS,
+		.has_inbox = false,
+		.has_outbox = true,
+		.out_is_imm = false,
+		.encode_slave_id = false,
+		.verify = NULL,
+		.wrapper = mlx4_CMD_EPERM_wrapper,
+	},
+	{
+		.opcode = MLX4_CMD_CONF_SPECIAL_QP,
+		.has_inbox = false,
+		.has_outbox = false,
+		.out_is_imm = false,
+		.encode_slave_id = false,
+		.verify = NULL, /* XXX verify: only demux can do this */
+		.wrapper = NULL
+	},
+	{
+		.opcode = MLX4_CMD_MAD_IFC,
+		.has_inbox = true,
+		.has_outbox = true,
+		.out_is_imm = false,
+		.encode_slave_id = false,
+		.verify = NULL,
+		.wrapper = mlx4_MAD_IFC_wrapper
+	},
+	{
+		.opcode = MLX4_CMD_MAD_DEMUX,
+		.has_inbox = false,
+		.has_outbox = false,
+		.out_is_imm = false,
+		.encode_slave_id = false,
+		.verify = NULL,
+		.wrapper = mlx4_CMD_EPERM_wrapper
+	},
+	{
+		.opcode = MLX4_CMD_QUERY_IF_STAT,
+		.has_inbox = false,
+		.has_outbox = true,
+		.out_is_imm = false,
+		.encode_slave_id = false,
+		.verify = NULL,
+		.wrapper = mlx4_QUERY_IF_STAT_wrapper
+	},
+	{
+		.opcode = MLX4_CMD_ACCESS_REG,
+		.has_inbox = true,
+		.has_outbox = true,
+		.out_is_imm = false,
+		.encode_slave_id = false,
+		.verify = NULL,
+		.wrapper = mlx4_ACCESS_REG_wrapper,
+	},
+	{
+		.opcode = MLX4_CMD_CONGESTION_CTRL_OPCODE,
+		.has_inbox = false,
+		.has_outbox = false,
+		.out_is_imm = false,
+		.encode_slave_id = false,
+		.verify = NULL,
+		.wrapper = mlx4_CMD_EPERM_wrapper,
+	},
+	/* Native multicast commands are not available for guests */
+	{
+		.opcode = MLX4_CMD_QP_ATTACH,
+		.has_inbox = true,
+		.has_outbox = false,
+		.out_is_imm = false,
+		.encode_slave_id = false,
+		.verify = NULL,
+		.wrapper = mlx4_QP_ATTACH_wrapper
+	},
+	{
+		.opcode = MLX4_CMD_PROMISC,
+		.has_inbox = false,
+		.has_outbox = false,
+		.out_is_imm = false,
+		.encode_slave_id = false,
+		.verify = NULL,
+		.wrapper = mlx4_PROMISC_wrapper
+	},
+	/* Ethernet specific commands */
+	{
+		.opcode = MLX4_CMD_SET_VLAN_FLTR,
+		.has_inbox = true,
+		.has_outbox = false,
+		.out_is_imm = false,
+		.encode_slave_id = false,
+		.verify = NULL,
+		.wrapper = mlx4_SET_VLAN_FLTR_wrapper
+	},
+	{
+		.opcode = MLX4_CMD_SET_MCAST_FLTR,
+		.has_inbox = false,
+		.has_outbox = false,
+		.out_is_imm = false,
+		.encode_slave_id = false,
+		.verify = NULL,
+		.wrapper = mlx4_SET_MCAST_FLTR_wrapper
+	},
+	{
+		.opcode = MLX4_CMD_DUMP_ETH_STATS,
+		.has_inbox = false,
+		.has_outbox = true,
+		.out_is_imm = false,
+		.encode_slave_id = false,
+		.verify = NULL,
+		.wrapper = mlx4_DUMP_ETH_STATS_wrapper
+	},
+	{
+		.opcode = MLX4_CMD_INFORM_FLR_DONE,
+		.has_inbox = false,
+		.has_outbox = false,
+		.out_is_imm = false,
+		.encode_slave_id = false,
+		.verify = NULL,
+		.wrapper = NULL
+	},
+	/* flow steering commands */
+	{
+		.opcode = MLX4_QP_FLOW_STEERING_ATTACH,
+		.has_inbox = true,
+		.has_outbox = false,
+		.out_is_imm = true,
+		.encode_slave_id = false,
+		.verify = NULL,
+		.wrapper = mlx4_QP_FLOW_STEERING_ATTACH_wrapper
+	},
+	{
+		.opcode = MLX4_QP_FLOW_STEERING_DETACH,
+		.has_inbox = false,
+		.has_outbox = false,
+		.out_is_imm = false,
+		.encode_slave_id = false,
+		.verify = NULL,
+		.wrapper = mlx4_QP_FLOW_STEERING_DETACH_wrapper
+	},
+	{
+		.opcode = MLX4_FLOW_STEERING_IB_UC_QP_RANGE,
+		.has_inbox = false,
+		.has_outbox = false,
+		.out_is_imm = false,
+		.encode_slave_id = false,
+		.verify = NULL,
+		.wrapper = mlx4_CMD_EPERM_wrapper
+	},
+	{
+		.opcode = MLX4_CMD_VIRT_PORT_MAP,
+		.has_inbox = false,
+		.has_outbox = false,
+		.out_is_imm = false,
+		.encode_slave_id = false,
+		.verify = NULL,
+		.wrapper = mlx4_CMD_EPERM_wrapper
+	},
+};
+
+static int mlx4_master_process_vhcr(struct mlx4_dev *dev, int slave,
+				    struct mlx4_vhcr_cmd *in_vhcr)
+{
+	struct mlx4_priv *priv = mlx4_priv(dev);
+	struct mlx4_cmd_info *cmd = NULL;
+	struct mlx4_vhcr_cmd *vhcr_cmd = in_vhcr ? in_vhcr : priv->mfunc.vhcr;
+	struct mlx4_vhcr *vhcr;
+	struct mlx4_cmd_mailbox *inbox = NULL;
+	struct mlx4_cmd_mailbox *outbox = NULL;
+	u64 in_param;
+	u64 out_param;
+	int ret = 0;
+	int i;
+	int err = 0;
+
+	/* Create sw representation of Virtual HCR */
+	vhcr = kzalloc(sizeof(struct mlx4_vhcr), GFP_KERNEL);
+	if (!vhcr)
+		return -ENOMEM;
+
+	/* DMA in the vHCR */
+	if (!in_vhcr) {
+		ret = mlx4_ACCESS_MEM(dev, priv->mfunc.vhcr_dma, slave,
+				      priv->mfunc.master.slave_state[slave].vhcr_dma,
+				      ALIGN(sizeof(struct mlx4_vhcr_cmd),
+					    MLX4_ACCESS_MEM_ALIGN), 1);
+		if (ret) {
+			if (!(dev->persist->state &
+			    MLX4_DEVICE_STATE_INTERNAL_ERROR))
+				mlx4_err(dev, "%s: Failed reading vhcr ret: 0x%x\n",
+					 __func__, ret);
+			kfree(vhcr);
+			return ret;
+		}
+	}
+
+	/* Fill SW VHCR fields */
+	vhcr->in_param = be64_to_cpu(vhcr_cmd->in_param);
+	vhcr->out_param = be64_to_cpu(vhcr_cmd->out_param);
+	vhcr->in_modifier = be32_to_cpu(vhcr_cmd->in_modifier);
+	vhcr->token = be16_to_cpu(vhcr_cmd->token);
+	vhcr->op = be16_to_cpu(vhcr_cmd->opcode) & 0xfff;
+	vhcr->op_modifier = (u8) (be16_to_cpu(vhcr_cmd->opcode) >> 12);
+	vhcr->e_bit = vhcr_cmd->flags & (1 << 6);
+
+	/* Lookup command */
+	for (i = 0; i < ARRAY_SIZE(cmd_info); ++i) {
+		if (vhcr->op == cmd_info[i].opcode) {
+			cmd = &cmd_info[i];
+			break;
+		}
+	}
+	if (!cmd) {
+		mlx4_err(dev, "Unknown command:0x%x accepted from slave:%d\n",
+			 vhcr->op, slave);
+		vhcr_cmd->status = CMD_STAT_BAD_PARAM;
+		goto out_status;
+	}
+
+	/* Read inbox */
+	if (cmd->has_inbox) {
+		vhcr->in_param &= INBOX_MASK;
+		inbox = mlx4_alloc_cmd_mailbox(dev);
+		if (IS_ERR(inbox)) {
+			vhcr_cmd->status = CMD_STAT_BAD_SIZE;
+			inbox = NULL;
+			goto out_status;
+		}
+
+		ret = mlx4_ACCESS_MEM(dev, inbox->dma, slave,
+				      vhcr->in_param,
+				      MLX4_MAILBOX_SIZE, 1);
+		if (ret) {
+			if (!(dev->persist->state &
+			    MLX4_DEVICE_STATE_INTERNAL_ERROR))
+				mlx4_err(dev, "%s: Failed reading inbox (cmd:0x%x)\n",
+					 __func__, cmd->opcode);
+			vhcr_cmd->status = CMD_STAT_INTERNAL_ERR;
+			goto out_status;
+		}
+	}
+
+	/* Apply permission and bound checks if applicable */
+	if (cmd->verify && cmd->verify(dev, slave, vhcr, inbox)) {
+		mlx4_warn(dev, "Command:0x%x from slave: %d failed protection checks for resource_id:%d\n",
+			  vhcr->op, slave, vhcr->in_modifier);
+		vhcr_cmd->status = CMD_STAT_BAD_OP;
+		goto out_status;
+	}
+
+	/* Allocate outbox */
+	if (cmd->has_outbox) {
+		outbox = mlx4_alloc_cmd_mailbox(dev);
+		if (IS_ERR(outbox)) {
+			vhcr_cmd->status = CMD_STAT_BAD_SIZE;
+			outbox = NULL;
+			goto out_status;
+		}
+	}
+
+	/* Execute the command! */
+	if (cmd->wrapper) {
+		err = cmd->wrapper(dev, slave, vhcr, inbox, outbox,
+				   cmd);
+		if (cmd->out_is_imm)
+			vhcr_cmd->out_param = cpu_to_be64(vhcr->out_param);
+	} else {
+		in_param = cmd->has_inbox ? (u64) inbox->dma :
+			vhcr->in_param;
+		out_param = cmd->has_outbox ? (u64) outbox->dma :
+			vhcr->out_param;
+		err = __mlx4_cmd(dev, in_param, &out_param,
+				 cmd->out_is_imm, vhcr->in_modifier,
+				 vhcr->op_modifier, vhcr->op,
+				 MLX4_CMD_TIME_CLASS_A,
+				 MLX4_CMD_NATIVE);
+
+		if (cmd->out_is_imm) {
+			vhcr->out_param = out_param;
+			vhcr_cmd->out_param = cpu_to_be64(vhcr->out_param);
+		}
+	}
+
+	if (err) {
+		if (!(dev->persist->state & MLX4_DEVICE_STATE_INTERNAL_ERROR)) {
+			if (vhcr->op == MLX4_CMD_ALLOC_RES &&
+			    (vhcr->in_modifier & 0xff) == RES_COUNTER &&
+			    err == -EDQUOT)
+				mlx4_dbg(dev,
+					 "Unable to allocate counter for slave %d (%d)\n",
+					 slave, err);
+			else
+				mlx4_warn(dev, "vhcr command:0x%x slave:%d failed with error:%d, status %d\n",
+					  vhcr->op, slave, vhcr->errno, err);
+		}
+		vhcr_cmd->status = mlx4_errno_to_status(err);
+		goto out_status;
+	}
+
+
+	/* Write outbox if command completed successfully */
+	if (cmd->has_outbox && !vhcr_cmd->status) {
+		ret = mlx4_ACCESS_MEM(dev, outbox->dma, slave,
+				      vhcr->out_param,
+				      MLX4_MAILBOX_SIZE, MLX4_CMD_WRAPPED);
+		if (ret) {
+			/* If we failed to write back the outbox after the
+			 *command was successfully executed, we must fail this
+			 * slave, as it is now in undefined state */
+			if (!(dev->persist->state &
+			    MLX4_DEVICE_STATE_INTERNAL_ERROR))
+				mlx4_err(dev, "%s:Failed writing outbox\n", __func__);
+			goto out;
+		}
+	}
+
+out_status:
+	/* DMA back vhcr result */
+	if (!in_vhcr) {
+		ret = mlx4_ACCESS_MEM(dev, priv->mfunc.vhcr_dma, slave,
+				      priv->mfunc.master.slave_state[slave].vhcr_dma,
+				      ALIGN(sizeof(struct mlx4_vhcr),
+					    MLX4_ACCESS_MEM_ALIGN),
+				      MLX4_CMD_WRAPPED);
+		if (ret)
+			mlx4_err(dev, "%s:Failed writing vhcr result\n",
+				 __func__);
+		else if (vhcr->e_bit &&
+			 mlx4_GEN_EQE(dev, slave, &priv->mfunc.master.cmd_eqe))
+				mlx4_warn(dev, "Failed to generate command completion eqe for slave %d\n",
+					  slave);
+	}
+
+out:
+	kfree(vhcr);
+	mlx4_free_cmd_mailbox(dev, inbox);
+	mlx4_free_cmd_mailbox(dev, outbox);
+	return ret;
+}
+
+static int mlx4_master_immediate_activate_vlan_qos(struct mlx4_priv *priv,
+					    int slave, int port)
+{
+	struct mlx4_vport_oper_state *vp_oper;
+	struct mlx4_vport_state *vp_admin;
+	struct mlx4_vf_immed_vlan_work *work;
+	struct mlx4_dev *dev = &(priv->dev);
+	int err;
+	int admin_vlan_ix = NO_INDX;
+
+	vp_oper = &priv->mfunc.master.vf_oper[slave].vport[port];
+	vp_admin = &priv->mfunc.master.vf_admin[slave].vport[port];
+
+	if (vp_oper->state.default_vlan == vp_admin->default_vlan &&
+	    vp_oper->state.default_qos == vp_admin->default_qos &&
+	    vp_oper->state.vlan_proto == vp_admin->vlan_proto &&
+	    vp_oper->state.link_state == vp_admin->link_state &&
+	    vp_oper->state.qos_vport == vp_admin->qos_vport)
+		return 0;
+
+	if (!(priv->mfunc.master.slave_state[slave].active &&
+	      dev->caps.flags2 & MLX4_DEV_CAP_FLAG2_UPDATE_QP)) {
+		/* even if the UPDATE_QP command isn't supported, we still want
+		 * to set this VF link according to the admin directive
+		 */
+		vp_oper->state.link_state = vp_admin->link_state;
+		return -1;
+	}
+
+	mlx4_dbg(dev, "updating immediately admin params slave %d port %d\n",
+		 slave, port);
+	mlx4_dbg(dev, "vlan %d QoS %d link down %d\n",
+		 vp_admin->default_vlan, vp_admin->default_qos,
+		 vp_admin->link_state);
+
+	work = kzalloc(sizeof(*work), GFP_KERNEL);
+	if (!work)
+		return -ENOMEM;
+
+	if (vp_oper->state.default_vlan != vp_admin->default_vlan) {
+		if (MLX4_VGT != vp_admin->default_vlan) {
+			err = __mlx4_register_vlan(&priv->dev, port,
+						   vp_admin->default_vlan,
+						   &admin_vlan_ix);
+			if (err) {
+				kfree(work);
+				mlx4_warn(&priv->dev,
+					  "No vlan resources slave %d, port %d\n",
+					  slave, port);
+				return err;
+			}
+		} else {
+			admin_vlan_ix = NO_INDX;
+		}
+		work->flags |= MLX4_VF_IMMED_VLAN_FLAG_VLAN;
+		mlx4_dbg(&priv->dev,
+			 "alloc vlan %d idx  %d slave %d port %d\n",
+			 (int)(vp_admin->default_vlan),
+			 admin_vlan_ix, slave, port);
+	}
+
+	/* save original vlan ix and vlan id */
+	work->orig_vlan_id = vp_oper->state.default_vlan;
+	work->orig_vlan_ix = vp_oper->vlan_idx;
+
+	/* handle new qos */
+	if (vp_oper->state.default_qos != vp_admin->default_qos)
+		work->flags |= MLX4_VF_IMMED_VLAN_FLAG_QOS;
+
+	if (work->flags & MLX4_VF_IMMED_VLAN_FLAG_VLAN)
+		vp_oper->vlan_idx = admin_vlan_ix;
+
+	vp_oper->state.default_vlan = vp_admin->default_vlan;
+	vp_oper->state.default_qos = vp_admin->default_qos;
+	vp_oper->state.vlan_proto = vp_admin->vlan_proto;
+	vp_oper->state.link_state = vp_admin->link_state;
+	vp_oper->state.qos_vport = vp_admin->qos_vport;
+
+	if (vp_admin->link_state == IFLA_VF_LINK_STATE_DISABLE)
+		work->flags |= MLX4_VF_IMMED_VLAN_FLAG_LINK_DISABLE;
+
+	/* iterate over QPs owned by this slave, using UPDATE_QP */
+	work->port = port;
+	work->slave = slave;
+	work->qos = vp_oper->state.default_qos;
+	work->qos_vport = vp_oper->state.qos_vport;
+	work->vlan_id = vp_oper->state.default_vlan;
+	work->vlan_ix = vp_oper->vlan_idx;
+	work->vlan_proto = vp_oper->state.vlan_proto;
+	work->priv = priv;
+	INIT_WORK(&work->work, mlx4_vf_immed_vlan_work_handler);
+	queue_work(priv->mfunc.master.comm_wq, &work->work);
+
+	return 0;
+}
+
+static void mlx4_set_default_port_qos(struct mlx4_dev *dev, int port)
+{
+	struct mlx4_qos_manager *port_qos_ctl;
+	struct mlx4_priv *priv = mlx4_priv(dev);
+
+	port_qos_ctl = &priv->mfunc.master.qos_ctl[port];
+	bitmap_zero(port_qos_ctl->priority_bm, MLX4_NUM_UP);
+
+	/* Enable only default prio at PF init routine */
+	set_bit(MLX4_DEFAULT_QOS_PRIO, port_qos_ctl->priority_bm);
+}
+
+static void mlx4_allocate_port_vpps(struct mlx4_dev *dev, int port)
+{
+	int i;
+	int err;
+	int num_vfs;
+	u16 available_vpp;
+	u8 vpp_param[MLX4_NUM_UP];
+	struct mlx4_qos_manager *port_qos;
+	struct mlx4_priv *priv = mlx4_priv(dev);
+
+	err = mlx4_ALLOCATE_VPP_get(dev, port, &available_vpp, vpp_param);
+	if (err) {
+		mlx4_info(dev, "Failed query available VPPs\n");
+		return;
+	}
+
+	port_qos = &priv->mfunc.master.qos_ctl[port];
+	num_vfs = (available_vpp /
+		   bitmap_weight(port_qos->priority_bm, MLX4_NUM_UP));
+
+	for (i = 0; i < MLX4_NUM_UP; i++) {
+		if (test_bit(i, port_qos->priority_bm))
+			vpp_param[i] = num_vfs;
+	}
+
+	err = mlx4_ALLOCATE_VPP_set(dev, port, vpp_param);
+	if (err) {
+		mlx4_info(dev, "Failed allocating VPPs\n");
+		return;
+	}
+
+	/* Query actual allocated VPP, just to make sure */
+	err = mlx4_ALLOCATE_VPP_get(dev, port, &available_vpp, vpp_param);
+	if (err) {
+		mlx4_info(dev, "Failed query available VPPs\n");
+		return;
+	}
+
+	port_qos->num_of_qos_vfs = num_vfs;
+	mlx4_dbg(dev, "Port %d Available VPPs %d\n", port, available_vpp);
+
+	for (i = 0; i < MLX4_NUM_UP; i++)
+		mlx4_dbg(dev, "Port %d UP %d Allocated %d VPPs\n", port, i,
+			 vpp_param[i]);
+}
+
+static int mlx4_master_activate_admin_state(struct mlx4_priv *priv, int slave)
+{
+	int port, err;
+	struct mlx4_vport_state *vp_admin;
+	struct mlx4_vport_oper_state *vp_oper;
+	struct mlx4_slave_state *slave_state =
+		&priv->mfunc.master.slave_state[slave];
+	struct mlx4_active_ports actv_ports = mlx4_get_active_ports(
+			&priv->dev, slave);
+	int min_port = find_first_bit(actv_ports.ports,
+				      priv->dev.caps.num_ports) + 1;
+	int max_port = min_port - 1 +
+		bitmap_weight(actv_ports.ports, priv->dev.caps.num_ports);
+
+	for (port = min_port; port <= max_port; port++) {
+		if (!test_bit(port - 1, actv_ports.ports))
+			continue;
+		priv->mfunc.master.vf_oper[slave].smi_enabled[port] =
+			priv->mfunc.master.vf_admin[slave].enable_smi[port];
+		vp_oper = &priv->mfunc.master.vf_oper[slave].vport[port];
+		vp_admin = &priv->mfunc.master.vf_admin[slave].vport[port];
+		if (vp_admin->vlan_proto != htons(ETH_P_8021AD) ||
+		    slave_state->vst_qinq_supported) {
+			vp_oper->state.vlan_proto   = vp_admin->vlan_proto;
+			vp_oper->state.default_vlan = vp_admin->default_vlan;
+			vp_oper->state.default_qos  = vp_admin->default_qos;
+		}
+		vp_oper->state.link_state = vp_admin->link_state;
+		vp_oper->state.mac        = vp_admin->mac;
+		vp_oper->state.spoofchk   = vp_admin->spoofchk;
+		vp_oper->state.tx_rate    = vp_admin->tx_rate;
+		vp_oper->state.qos_vport  = vp_admin->qos_vport;
+		vp_oper->state.guid       = vp_admin->guid;
+
+		if (MLX4_VGT != vp_admin->default_vlan) {
+			err = __mlx4_register_vlan(&priv->dev, port,
+						   vp_admin->default_vlan, &(vp_oper->vlan_idx));
+			if (err) {
+				vp_oper->vlan_idx = NO_INDX;
+				vp_oper->state.default_vlan = MLX4_VGT;
+				vp_oper->state.vlan_proto = htons(ETH_P_8021Q);
+				mlx4_warn(&priv->dev,
+					  "No vlan resources slave %d, port %d\n",
+					  slave, port);
+				return err;
+			}
+			mlx4_dbg(&priv->dev, "alloc vlan %d idx  %d slave %d port %d\n",
+				 (int)(vp_oper->state.default_vlan),
+				 vp_oper->vlan_idx, slave, port);
+		}
+		if (vp_admin->spoofchk) {
+			vp_oper->mac_idx = __mlx4_register_mac(&priv->dev,
+							       port,
+							       vp_admin->mac);
+			if (0 > vp_oper->mac_idx) {
+				err = vp_oper->mac_idx;
+				vp_oper->mac_idx = NO_INDX;
+				mlx4_warn(&priv->dev,
+					  "No mac resources slave %d, port %d\n",
+					  slave, port);
+				return err;
+			}
+			mlx4_dbg(&priv->dev, "alloc mac %llx idx  %d slave %d port %d\n",
+				 vp_oper->state.mac, vp_oper->mac_idx, slave, port);
+		}
+	}
+	return 0;
+}
+
+static void mlx4_master_deactivate_admin_state(struct mlx4_priv *priv, int slave)
+{
+	int port;
+	struct mlx4_vport_oper_state *vp_oper;
+	struct mlx4_active_ports actv_ports = mlx4_get_active_ports(
+			&priv->dev, slave);
+	int min_port = find_first_bit(actv_ports.ports,
+				      priv->dev.caps.num_ports) + 1;
+	int max_port = min_port - 1 +
+		bitmap_weight(actv_ports.ports, priv->dev.caps.num_ports);
+
+
+	for (port = min_port; port <= max_port; port++) {
+		if (!test_bit(port - 1, actv_ports.ports))
+			continue;
+		priv->mfunc.master.vf_oper[slave].smi_enabled[port] =
+			MLX4_VF_SMI_DISABLED;
+		vp_oper = &priv->mfunc.master.vf_oper[slave].vport[port];
+		if (NO_INDX != vp_oper->vlan_idx) {
+			__mlx4_unregister_vlan(&priv->dev,
+					       port, vp_oper->state.default_vlan);
+			vp_oper->vlan_idx = NO_INDX;
+		}
+		if (NO_INDX != vp_oper->mac_idx) {
+			__mlx4_unregister_mac(&priv->dev, port, vp_oper->state.mac);
+			vp_oper->mac_idx = NO_INDX;
+		}
+	}
+	return;
+}
+
+static void mlx4_master_do_cmd(struct mlx4_dev *dev, int slave, u8 cmd,
+			       u16 param, u8 toggle)
+{
+	struct mlx4_priv *priv = mlx4_priv(dev);
+	struct mlx4_slave_state *slave_state = priv->mfunc.master.slave_state;
+	u32 reply;
+	u8 is_going_down = 0;
+	int i;
+	unsigned long flags;
+
+	slave_state[slave].comm_toggle ^= 1;
+	reply = (u32) slave_state[slave].comm_toggle << 31;
+	if (toggle != slave_state[slave].comm_toggle) {
+		mlx4_warn(dev, "Incorrect toggle %d from slave %d. *** MASTER STATE COMPROMISED ***\n",
+			  toggle, slave);
+		goto reset_slave;
+	}
+	if (cmd == MLX4_COMM_CMD_RESET) {
+		mlx4_warn(dev, "Received reset from slave:%d\n", slave);
+		slave_state[slave].active = false;
+		slave_state[slave].old_vlan_api = false;
+		slave_state[slave].vst_qinq_supported = false;
+		mlx4_master_deactivate_admin_state(priv, slave);
+		for (i = 0; i < MLX4_EVENT_TYPES_NUM; ++i) {
+				slave_state[slave].event_eq[i].eqn = -1;
+				slave_state[slave].event_eq[i].token = 0;
+		}
+		/*check if we are in the middle of FLR process,
+		if so return "retry" status to the slave*/
+		if (MLX4_COMM_CMD_FLR == slave_state[slave].last_cmd)
+			goto inform_slave_state;
+
+		mlx4_dispatch_event(dev, MLX4_DEV_EVENT_SLAVE_SHUTDOWN, slave);
+
+		/* write the version in the event field */
+		reply |= mlx4_comm_get_version();
+
+		goto reset_slave;
+	}
+	/*command from slave in the middle of FLR*/
+	if (cmd != MLX4_COMM_CMD_RESET &&
+	    MLX4_COMM_CMD_FLR == slave_state[slave].last_cmd) {
+		mlx4_warn(dev, "slave:%d is Trying to run cmd(0x%x) in the middle of FLR\n",
+			  slave, cmd);
+		return;
+	}
+
+	switch (cmd) {
+	case MLX4_COMM_CMD_VHCR0:
+		if (slave_state[slave].last_cmd != MLX4_COMM_CMD_RESET)
+			goto reset_slave;
+		slave_state[slave].vhcr_dma = ((u64) param) << 48;
+		priv->mfunc.master.slave_state[slave].cookie = 0;
+		break;
+	case MLX4_COMM_CMD_VHCR1:
+		if (slave_state[slave].last_cmd != MLX4_COMM_CMD_VHCR0)
+			goto reset_slave;
+		slave_state[slave].vhcr_dma |= ((u64) param) << 32;
+		break;
+	case MLX4_COMM_CMD_VHCR2:
+		if (slave_state[slave].last_cmd != MLX4_COMM_CMD_VHCR1)
+			goto reset_slave;
+		slave_state[slave].vhcr_dma |= ((u64) param) << 16;
+		break;
+	case MLX4_COMM_CMD_VHCR_EN:
+		if (slave_state[slave].last_cmd != MLX4_COMM_CMD_VHCR2)
+			goto reset_slave;
+		slave_state[slave].vhcr_dma |= param;
+		if (mlx4_master_activate_admin_state(priv, slave))
+				goto reset_slave;
+		slave_state[slave].active = true;
+		mlx4_dispatch_event(dev, MLX4_DEV_EVENT_SLAVE_INIT, slave);
+		break;
+	case MLX4_COMM_CMD_VHCR_POST:
+		if ((slave_state[slave].last_cmd != MLX4_COMM_CMD_VHCR_EN) &&
+		    (slave_state[slave].last_cmd != MLX4_COMM_CMD_VHCR_POST)) {
+			mlx4_warn(dev, "slave:%d is out of sync, cmd=0x%x, last command=0x%x, reset is needed\n",
+				  slave, cmd, slave_state[slave].last_cmd);
+			goto reset_slave;
+		}
+
+		mutex_lock(&priv->cmd.slave_cmd_mutex);
+		if (mlx4_master_process_vhcr(dev, slave, NULL)) {
+			mlx4_err(dev, "Failed processing vhcr for slave:%d, resetting slave\n",
+				 slave);
+			mutex_unlock(&priv->cmd.slave_cmd_mutex);
+			goto reset_slave;
+		}
+		mutex_unlock(&priv->cmd.slave_cmd_mutex);
+		break;
+	default:
+		mlx4_warn(dev, "Bad comm cmd:%d from slave:%d\n", cmd, slave);
+		goto reset_slave;
+	}
+	spin_lock_irqsave(&priv->mfunc.master.slave_state_lock, flags);
+	if (!slave_state[slave].is_slave_going_down)
+		slave_state[slave].last_cmd = cmd;
+	else
+		is_going_down = 1;
+	spin_unlock_irqrestore(&priv->mfunc.master.slave_state_lock, flags);
+	if (is_going_down) {
+		mlx4_warn(dev, "Slave is going down aborting command(%d) executing from slave:%d\n",
+			  cmd, slave);
+		return;
+	}
+	__raw_writel((__force u32) cpu_to_be32(reply),
+		     &priv->mfunc.comm[slave].slave_read);
+	mmiowb();
+
+	return;
+
+reset_slave:
+	/* cleanup any slave resources */
+	if (dev->persist->interface_state & MLX4_INTERFACE_STATE_UP)
+		mlx4_delete_all_resources_for_slave(dev, slave);
+
+	if (cmd != MLX4_COMM_CMD_RESET) {
+		mlx4_warn(dev, "Turn on internal error to force reset, slave=%d, cmd=0x%x\n",
+			  slave, cmd);
+		/* Turn on internal error letting slave reset itself immeditaly,
+		 * otherwise it might take till timeout on command is passed
+		 */
+		reply |= ((u32)COMM_CHAN_EVENT_INTERNAL_ERR);
+	}
+
+	spin_lock_irqsave(&priv->mfunc.master.slave_state_lock, flags);
+	if (!slave_state[slave].is_slave_going_down)
+		slave_state[slave].last_cmd = MLX4_COMM_CMD_RESET;
+	spin_unlock_irqrestore(&priv->mfunc.master.slave_state_lock, flags);
+	/*with slave in the middle of flr, no need to clean resources again.*/
+inform_slave_state:
+	memset(&slave_state[slave].event_eq, 0,
+	       sizeof(struct mlx4_slave_event_eq_info));
+	__raw_writel((__force u32) cpu_to_be32(reply),
+		     &priv->mfunc.comm[slave].slave_read);
+	wmb();
+}
+
+/* master command processing */
+void mlx4_master_comm_channel(struct work_struct *work)
+{
+	struct mlx4_mfunc_master_ctx *master =
+		container_of(work,
+			     struct mlx4_mfunc_master_ctx,
+			     comm_work);
+	struct mlx4_mfunc *mfunc =
+		container_of(master, struct mlx4_mfunc, master);
+	struct mlx4_priv *priv =
+		container_of(mfunc, struct mlx4_priv, mfunc);
+	struct mlx4_dev *dev = &priv->dev;
+	__be32 *bit_vec;
+	u32 comm_cmd;
+	u32 vec;
+	int i, j, slave;
+	int toggle;
+	int served = 0;
+	int reported = 0;
+	u32 slt;
+
+	bit_vec = master->comm_arm_bit_vector;
+	for (i = 0; i < COMM_CHANNEL_BIT_ARRAY_SIZE; i++) {
+		vec = be32_to_cpu(bit_vec[i]);
+		for (j = 0; j < 32; j++) {
+			if (!(vec & (1 << j)))
+				continue;
+			++reported;
+			slave = (i * 32) + j;
+			comm_cmd = swab32(readl(
+					  &mfunc->comm[slave].slave_write));
+			slt = swab32(readl(&mfunc->comm[slave].slave_read))
+				     >> 31;
+			toggle = comm_cmd >> 31;
+			if (toggle != slt) {
+				if (master->slave_state[slave].comm_toggle
+				    != slt) {
+					pr_info("slave %d out of sync. read toggle %d, state toggle %d. Resynching.\n",
+						slave, slt,
+						master->slave_state[slave].comm_toggle);
+					master->slave_state[slave].comm_toggle =
+						slt;
+				}
+				mlx4_master_do_cmd(dev, slave,
+						   comm_cmd >> 16 & 0xff,
+						   comm_cmd & 0xffff, toggle);
+				++served;
+			}
+		}
+	}
+
+	if (reported && reported != served)
+		mlx4_warn(dev, "Got command event with bitmask from %d slaves but %d were served\n",
+			  reported, served);
+
+	if (mlx4_ARM_COMM_CHANNEL(dev))
+		mlx4_warn(dev, "Failed to arm comm channel events\n");
+}
+
+static int sync_toggles(struct mlx4_dev *dev)
+{
+	struct mlx4_priv *priv = mlx4_priv(dev);
+	u32 wr_toggle;
+	u32 rd_toggle;
+	unsigned long end;
+
+	wr_toggle = swab32(readl(&priv->mfunc.comm->slave_write));
+	if (wr_toggle == 0xffffffff)
+		end = jiffies + msecs_to_jiffies(30000);
+	else
+		end = jiffies + msecs_to_jiffies(5000);
+
+	while (time_before(jiffies, end)) {
+		rd_toggle = swab32(readl(&priv->mfunc.comm->slave_read));
+		if (wr_toggle == 0xffffffff || rd_toggle == 0xffffffff) {
+			/* PCI might be offline */
+
+			/* If device removal has been requested,
+			 * do not continue retrying.
+			 */
+			if (dev->persist->interface_state &
+			    MLX4_INTERFACE_STATE_NOWAIT) {
+				mlx4_warn(dev,
+					  "communication channel is offline\n");
+				return -EIO;
+			}
+
+			msleep(100);
+			wr_toggle = swab32(readl(&priv->mfunc.comm->
+					   slave_write));
+			continue;
+		}
+
+		if (rd_toggle >> 31 == wr_toggle >> 31) {
+			priv->cmd.comm_toggle = rd_toggle >> 31;
+			return 0;
+		}
+
+		cond_resched();
+	}
+
+	/*
+	 * we could reach here if for example the previous VM using this
+	 * function misbehaved and left the channel with unsynced state. We
+	 * should fix this here and give this VM a chance to use a properly
+	 * synced channel
+	 */
+	mlx4_warn(dev, "recovering from previously mis-behaved VM\n");
+	__raw_writel((__force u32) 0, &priv->mfunc.comm->slave_read);
+	__raw_writel((__force u32) 0, &priv->mfunc.comm->slave_write);
+	priv->cmd.comm_toggle = 0;
+
+	return 0;
+}
+
+int mlx4_multi_func_init(struct mlx4_dev *dev)
+{
+	struct mlx4_priv *priv = mlx4_priv(dev);
+	struct mlx4_slave_state *s_state;
+	int i, j, err, port;
+
+	if (mlx4_is_master(dev))
+		priv->mfunc.comm =
+		ioremap(pci_resource_start(dev->persist->pdev,
+					   priv->fw.comm_bar) +
+			priv->fw.comm_base, MLX4_COMM_PAGESIZE);
+	else
+		priv->mfunc.comm =
+		ioremap(pci_resource_start(dev->persist->pdev, 2) +
+			MLX4_SLAVE_COMM_BASE, MLX4_COMM_PAGESIZE);
+	if (!priv->mfunc.comm) {
+		mlx4_err(dev, "Couldn't map communication vector\n");
+		goto err_vhcr;
+	}
+
+	if (mlx4_is_master(dev)) {
+		struct mlx4_vf_oper_state *vf_oper;
+		struct mlx4_vf_admin_state *vf_admin;
+
+		priv->mfunc.master.slave_state =
+			kcalloc(dev->num_slaves,
+				sizeof(struct mlx4_slave_state),
+				GFP_KERNEL);
+		if (!priv->mfunc.master.slave_state)
+			goto err_comm;
+
+		priv->mfunc.master.vf_admin =
+			kcalloc(dev->num_slaves,
+				sizeof(struct mlx4_vf_admin_state),
+				GFP_KERNEL);
+		if (!priv->mfunc.master.vf_admin)
+			goto err_comm_admin;
+
+		priv->mfunc.master.vf_oper =
+			kcalloc(dev->num_slaves,
+				sizeof(struct mlx4_vf_oper_state),
+				GFP_KERNEL);
+		if (!priv->mfunc.master.vf_oper)
+			goto err_comm_oper;
+
+		for (i = 0; i < dev->num_slaves; ++i) {
+			vf_admin = &priv->mfunc.master.vf_admin[i];
+			vf_oper = &priv->mfunc.master.vf_oper[i];
+			s_state = &priv->mfunc.master.slave_state[i];
+			s_state->last_cmd = MLX4_COMM_CMD_RESET;
+			s_state->vst_qinq_supported = false;
+			mutex_init(&priv->mfunc.master.gen_eqe_mutex[i]);
+			for (j = 0; j < MLX4_EVENT_TYPES_NUM; ++j)
+				s_state->event_eq[j].eqn = -1;
+			__raw_writel((__force u32) 0,
+				     &priv->mfunc.comm[i].slave_write);
+			__raw_writel((__force u32) 0,
+				     &priv->mfunc.comm[i].slave_read);
+			mmiowb();
+			for (port = 1; port <= MLX4_MAX_PORTS; port++) {
+				struct mlx4_vport_state *admin_vport;
+				struct mlx4_vport_state *oper_vport;
+
+				s_state->vlan_filter[port] =
+					kzalloc(sizeof(struct mlx4_vlan_fltr),
+						GFP_KERNEL);
+				if (!s_state->vlan_filter[port]) {
+					if (--port)
+						kfree(s_state->vlan_filter[port]);
+					goto err_slaves;
+				}
+
+				admin_vport = &vf_admin->vport[port];
+				oper_vport = &vf_oper->vport[port].state;
+				INIT_LIST_HEAD(&s_state->mcast_filters[port]);
+				admin_vport->default_vlan = MLX4_VGT;
+				oper_vport->default_vlan = MLX4_VGT;
+				admin_vport->qos_vport =
+						MLX4_VPP_DEFAULT_VPORT;
+				oper_vport->qos_vport = MLX4_VPP_DEFAULT_VPORT;
+				admin_vport->vlan_proto = htons(ETH_P_8021Q);
+				oper_vport->vlan_proto = htons(ETH_P_8021Q);
+				vf_oper->vport[port].vlan_idx = NO_INDX;
+				vf_oper->vport[port].mac_idx = NO_INDX;
+				mlx4_set_random_admin_guid(dev, i, port);
+			}
+			spin_lock_init(&s_state->lock);
+		}
+
+		if (dev->caps.flags2 & MLX4_DEV_CAP_FLAG2_QOS_VPP) {
+			for (port = 1; port <= dev->caps.num_ports; port++) {
+				if (mlx4_is_eth(dev, port)) {
+					mlx4_set_default_port_qos(dev, port);
+					mlx4_allocate_port_vpps(dev, port);
+				}
+			}
+		}
+
+		memset(&priv->mfunc.master.cmd_eqe, 0, sizeof(struct mlx4_eqe));
+		priv->mfunc.master.cmd_eqe.type = MLX4_EVENT_TYPE_CMD;
+		INIT_WORK(&priv->mfunc.master.comm_work,
+			  mlx4_master_comm_channel);
+		INIT_WORK(&priv->mfunc.master.slave_event_work,
+			  mlx4_gen_slave_eqe);
+		INIT_WORK(&priv->mfunc.master.slave_flr_event_work,
+			  mlx4_master_handle_slave_flr);
+		spin_lock_init(&priv->mfunc.master.slave_state_lock);
+		spin_lock_init(&priv->mfunc.master.slave_eq.event_lock);
+		priv->mfunc.master.comm_wq =
+			create_singlethread_workqueue("mlx4_comm");
+		if (!priv->mfunc.master.comm_wq)
+			goto err_slaves;
+
+		if (mlx4_init_resource_tracker(dev))
+			goto err_thread;
+
+	} else {
+		err = sync_toggles(dev);
+		if (err) {
+			mlx4_err(dev, "Couldn't sync toggles\n");
+			goto err_comm;
+		}
+	}
+	return 0;
+
+err_thread:
+	flush_workqueue(priv->mfunc.master.comm_wq);
+	destroy_workqueue(priv->mfunc.master.comm_wq);
+err_slaves:
+	while (i--) {
+		for (port = 1; port <= MLX4_MAX_PORTS; port++)
+			kfree(priv->mfunc.master.slave_state[i].vlan_filter[port]);
+	}
+	kfree(priv->mfunc.master.vf_oper);
+err_comm_oper:
+	kfree(priv->mfunc.master.vf_admin);
+err_comm_admin:
+	kfree(priv->mfunc.master.slave_state);
+err_comm:
+	iounmap(priv->mfunc.comm);
+	priv->mfunc.comm = NULL;
+err_vhcr:
+	dma_free_coherent(&dev->persist->pdev->dev, PAGE_SIZE,
+			  priv->mfunc.vhcr,
+			  priv->mfunc.vhcr_dma);
+	priv->mfunc.vhcr = NULL;
+	return -ENOMEM;
+}
+
+int mlx4_cmd_init(struct mlx4_dev *dev)
+{
+	struct mlx4_priv *priv = mlx4_priv(dev);
+	int flags = 0;
+
+	if (!priv->cmd.initialized) {
+		init_rwsem(&priv->cmd.switch_sem);
+		mutex_init(&priv->cmd.slave_cmd_mutex);
+		sema_init(&priv->cmd.poll_sem, 1);
+		priv->cmd.use_events = 0;
+		priv->cmd.toggle     = 1;
+		priv->cmd.initialized = 1;
+		flags |= MLX4_CMD_CLEANUP_STRUCT;
+	}
+
+	if (!mlx4_is_slave(dev) && !priv->cmd.hcr) {
+		priv->cmd.hcr = ioremap(pci_resource_start(dev->persist->pdev,
+					0) + MLX4_HCR_BASE, MLX4_HCR_SIZE);
+		if (!priv->cmd.hcr) {
+			mlx4_err(dev, "Couldn't map command register\n");
+			goto err;
+		}
+		flags |= MLX4_CMD_CLEANUP_HCR;
+	}
+
+	if (mlx4_is_mfunc(dev) && !priv->mfunc.vhcr) {
+		priv->mfunc.vhcr = dma_alloc_coherent(&dev->persist->pdev->dev,
+						      PAGE_SIZE,
+						      &priv->mfunc.vhcr_dma,
+						      GFP_KERNEL);
+		if (!priv->mfunc.vhcr)
+			goto err;
+
+		flags |= MLX4_CMD_CLEANUP_VHCR;
+	}
+
+	if (!priv->cmd.pool) {
+		priv->cmd.pool = dma_pool_create("mlx4_cmd",
+						 &dev->persist->pdev->dev,
+						 MLX4_MAILBOX_SIZE,
+						 MLX4_MAILBOX_SIZE, 0);
+		if (!priv->cmd.pool)
+			goto err;
+
+		flags |= MLX4_CMD_CLEANUP_POOL;
+	}
+
+	return 0;
+
+err:
+	mlx4_cmd_cleanup(dev, flags);
+	return -ENOMEM;
+}
+
+void mlx4_report_internal_err_comm_event(struct mlx4_dev *dev)
+{
+	struct mlx4_priv *priv = mlx4_priv(dev);
+	int slave;
+	u32 slave_read;
+
+	/* If the comm channel has not yet been initialized,
+	 * skip reporting the internal error event to all
+	 * the communication channels.
+	 */
+	if (!priv->mfunc.comm)
+		return;
+
+	/* Report an internal error event to all
+	 * communication channels.
+	 */
+	for (slave = 0; slave < dev->num_slaves; slave++) {
+		slave_read = swab32(readl(&priv->mfunc.comm[slave].slave_read));
+		slave_read |= (u32)COMM_CHAN_EVENT_INTERNAL_ERR;
+		__raw_writel((__force u32)cpu_to_be32(slave_read),
+			     &priv->mfunc.comm[slave].slave_read);
+		/* Make sure that our comm channel write doesn't
+		 * get mixed in with writes from another CPU.
+		 */
+		mmiowb();
+	}
+}
+
+void mlx4_multi_func_cleanup(struct mlx4_dev *dev)
+{
+	struct mlx4_priv *priv = mlx4_priv(dev);
+	int i, port;
+
+	if (mlx4_is_master(dev)) {
+		flush_workqueue(priv->mfunc.master.comm_wq);
+		destroy_workqueue(priv->mfunc.master.comm_wq);
+		for (i = 0; i < dev->num_slaves; i++) {
+			for (port = 1; port <= MLX4_MAX_PORTS; port++)
+				kfree(priv->mfunc.master.slave_state[i].vlan_filter[port]);
+		}
+		kfree(priv->mfunc.master.slave_state);
+		kfree(priv->mfunc.master.vf_admin);
+		kfree(priv->mfunc.master.vf_oper);
+		dev->num_slaves = 0;
+	}
+
+	iounmap(priv->mfunc.comm);
+	priv->mfunc.comm = NULL;
+}
+
+void mlx4_cmd_cleanup(struct mlx4_dev *dev, int cleanup_mask)
+{
+	struct mlx4_priv *priv = mlx4_priv(dev);
+
+	if (priv->cmd.pool && (cleanup_mask & MLX4_CMD_CLEANUP_POOL)) {
+		dma_pool_destroy(priv->cmd.pool);
+		priv->cmd.pool = NULL;
+	}
+
+	if (!mlx4_is_slave(dev) && priv->cmd.hcr &&
+	    (cleanup_mask & MLX4_CMD_CLEANUP_HCR)) {
+		iounmap(priv->cmd.hcr);
+		priv->cmd.hcr = NULL;
+	}
+	if (mlx4_is_mfunc(dev) && priv->mfunc.vhcr &&
+	    (cleanup_mask & MLX4_CMD_CLEANUP_VHCR)) {
+		dma_free_coherent(&dev->persist->pdev->dev, PAGE_SIZE,
+				  priv->mfunc.vhcr, priv->mfunc.vhcr_dma);
+		priv->mfunc.vhcr = NULL;
+	}
+	if (priv->cmd.initialized && (cleanup_mask & MLX4_CMD_CLEANUP_STRUCT))
+		priv->cmd.initialized = 0;
+}
+
+/*
+ * Switch to using events to issue FW commands (can only be called
+ * after event queue for command events has been initialized).
+ */
+int mlx4_cmd_use_events(struct mlx4_dev *dev)
+{
+	struct mlx4_priv *priv = mlx4_priv(dev);
+	int i;
+	int err = 0;
+
+	priv->cmd.context = kmalloc_array(priv->cmd.max_cmds,
+					  sizeof(struct mlx4_cmd_context),
+					  GFP_KERNEL);
+	if (!priv->cmd.context)
+		return -ENOMEM;
+
+	if (mlx4_is_mfunc(dev))
+		mutex_lock(&priv->cmd.slave_cmd_mutex);
+	down_write(&priv->cmd.switch_sem);
+	for (i = 0; i < priv->cmd.max_cmds; ++i) {
+		priv->cmd.context[i].token = i;
+		priv->cmd.context[i].next  = i + 1;
+		/* To support fatal error flow, initialize all
+		 * cmd contexts to allow simulating completions
+		 * with complete() at any time.
+		 */
+		init_completion(&priv->cmd.context[i].done);
+	}
+
+	priv->cmd.context[priv->cmd.max_cmds - 1].next = -1;
+	priv->cmd.free_head = 0;
+
+	sema_init(&priv->cmd.event_sem, priv->cmd.max_cmds);
+
+	for (priv->cmd.token_mask = 1;
+	     priv->cmd.token_mask < priv->cmd.max_cmds;
+	     priv->cmd.token_mask <<= 1)
+		; /* nothing */
+	--priv->cmd.token_mask;
+
+	down(&priv->cmd.poll_sem);
+	priv->cmd.use_events = 1;
+	up_write(&priv->cmd.switch_sem);
+	if (mlx4_is_mfunc(dev))
+		mutex_unlock(&priv->cmd.slave_cmd_mutex);
+
+	return err;
+}
+
+/*
+ * Switch back to polling (used when shutting down the device)
+ */
+void mlx4_cmd_use_polling(struct mlx4_dev *dev)
+{
+	struct mlx4_priv *priv = mlx4_priv(dev);
+	int i;
+
+	if (mlx4_is_mfunc(dev))
+		mutex_lock(&priv->cmd.slave_cmd_mutex);
+	down_write(&priv->cmd.switch_sem);
+	priv->cmd.use_events = 0;
+
+	for (i = 0; i < priv->cmd.max_cmds; ++i)
+		down(&priv->cmd.event_sem);
+
+	kfree(priv->cmd.context);
+	priv->cmd.context = NULL;
+
+	up(&priv->cmd.poll_sem);
+	up_write(&priv->cmd.switch_sem);
+	if (mlx4_is_mfunc(dev))
+		mutex_unlock(&priv->cmd.slave_cmd_mutex);
+}
+
+struct mlx4_cmd_mailbox *mlx4_alloc_cmd_mailbox(struct mlx4_dev *dev)
+{
+	struct mlx4_cmd_mailbox *mailbox;
+
+	mailbox = kmalloc(sizeof(*mailbox), GFP_KERNEL);
+	if (!mailbox)
+		return ERR_PTR(-ENOMEM);
+
+	mailbox->buf = dma_pool_zalloc(mlx4_priv(dev)->cmd.pool, GFP_KERNEL,
+				       &mailbox->dma);
+	if (!mailbox->buf) {
+		kfree(mailbox);
+		return ERR_PTR(-ENOMEM);
+	}
+
+	return mailbox;
+}
+EXPORT_SYMBOL_GPL(mlx4_alloc_cmd_mailbox);
+
+void mlx4_free_cmd_mailbox(struct mlx4_dev *dev,
+			   struct mlx4_cmd_mailbox *mailbox)
+{
+	if (!mailbox)
+		return;
+
+	dma_pool_free(mlx4_priv(dev)->cmd.pool, mailbox->buf, mailbox->dma);
+	kfree(mailbox);
+}
+EXPORT_SYMBOL_GPL(mlx4_free_cmd_mailbox);
+
+u32 mlx4_comm_get_version(void)
+{
+	 return ((u32) CMD_CHAN_IF_REV << 8) | (u32) CMD_CHAN_VER;
+}
+
+static int mlx4_get_slave_indx(struct mlx4_dev *dev, int vf)
+{
+	if ((vf < 0) || (vf >= dev->persist->num_vfs)) {
+		mlx4_err(dev, "Bad vf number:%d (number of activated vf: %d)\n",
+			 vf, dev->persist->num_vfs);
+		return -EINVAL;
+	}
+
+	return vf+1;
+}
+
+int mlx4_get_vf_indx(struct mlx4_dev *dev, int slave)
+{
+	if (slave < 1 || slave > dev->persist->num_vfs) {
+		mlx4_err(dev,
+			 "Bad slave number:%d (number of activated slaves: %lu)\n",
+			 slave, dev->num_slaves);
+		return -EINVAL;
+	}
+	return slave - 1;
+}
+
+void mlx4_cmd_wake_completions(struct mlx4_dev *dev)
+{
+	struct mlx4_priv *priv = mlx4_priv(dev);
+	struct mlx4_cmd_context *context;
+	int i;
+
+	spin_lock(&priv->cmd.context_lock);
+	if (priv->cmd.context) {
+		for (i = 0; i < priv->cmd.max_cmds; ++i) {
+			context = &priv->cmd.context[i];
+			context->fw_status = CMD_STAT_INTERNAL_ERR;
+			context->result    =
+				mlx4_status_to_errno(CMD_STAT_INTERNAL_ERR);
+			complete(&context->done);
+		}
+	}
+	spin_unlock(&priv->cmd.context_lock);
+}
+
+struct mlx4_active_ports mlx4_get_active_ports(struct mlx4_dev *dev, int slave)
+{
+	struct mlx4_active_ports actv_ports;
+	int vf;
+
+	bitmap_zero(actv_ports.ports, MLX4_MAX_PORTS);
+
+	if (slave == 0) {
+		bitmap_fill(actv_ports.ports, dev->caps.num_ports);
+		return actv_ports;
+	}
+
+	vf = mlx4_get_vf_indx(dev, slave);
+	if (vf < 0)
+		return actv_ports;
+
+	bitmap_set(actv_ports.ports, dev->dev_vfs[vf].min_port - 1,
+		   min((int)dev->dev_vfs[mlx4_get_vf_indx(dev, slave)].n_ports,
+		   dev->caps.num_ports));
+
+	return actv_ports;
+}
+EXPORT_SYMBOL_GPL(mlx4_get_active_ports);
+
+int mlx4_slave_convert_port(struct mlx4_dev *dev, int slave, int port)
+{
+	unsigned n;
+	struct mlx4_active_ports actv_ports = mlx4_get_active_ports(dev, slave);
+	unsigned m = bitmap_weight(actv_ports.ports, dev->caps.num_ports);
+
+	if (port <= 0 || port > m)
+		return -EINVAL;
+
+	n = find_first_bit(actv_ports.ports, dev->caps.num_ports);
+	if (port <= n)
+		port = n + 1;
+
+	return port;
+}
+EXPORT_SYMBOL_GPL(mlx4_slave_convert_port);
+
+int mlx4_phys_to_slave_port(struct mlx4_dev *dev, int slave, int port)
+{
+	struct mlx4_active_ports actv_ports = mlx4_get_active_ports(dev, slave);
+	if (test_bit(port - 1, actv_ports.ports))
+		return port -
+			find_first_bit(actv_ports.ports, dev->caps.num_ports);
+
+	return -1;
+}
+EXPORT_SYMBOL_GPL(mlx4_phys_to_slave_port);
+
+struct mlx4_slaves_pport mlx4_phys_to_slaves_pport(struct mlx4_dev *dev,
+						   int port)
+{
+	unsigned i;
+	struct mlx4_slaves_pport slaves_pport;
+
+	bitmap_zero(slaves_pport.slaves, MLX4_MFUNC_MAX);
+
+	if (port <= 0 || port > dev->caps.num_ports)
+		return slaves_pport;
+
+	for (i = 0; i < dev->persist->num_vfs + 1; i++) {
+		struct mlx4_active_ports actv_ports =
+			mlx4_get_active_ports(dev, i);
+		if (test_bit(port - 1, actv_ports.ports))
+			set_bit(i, slaves_pport.slaves);
+	}
+
+	return slaves_pport;
+}
+EXPORT_SYMBOL_GPL(mlx4_phys_to_slaves_pport);
+
+struct mlx4_slaves_pport mlx4_phys_to_slaves_pport_actv(
+		struct mlx4_dev *dev,
+		const struct mlx4_active_ports *crit_ports)
+{
+	unsigned i;
+	struct mlx4_slaves_pport slaves_pport;
+
+	bitmap_zero(slaves_pport.slaves, MLX4_MFUNC_MAX);
+
+	for (i = 0; i < dev->persist->num_vfs + 1; i++) {
+		struct mlx4_active_ports actv_ports =
+			mlx4_get_active_ports(dev, i);
+		if (bitmap_equal(crit_ports->ports, actv_ports.ports,
+				 dev->caps.num_ports))
+			set_bit(i, slaves_pport.slaves);
+	}
+
+	return slaves_pport;
+}
+EXPORT_SYMBOL_GPL(mlx4_phys_to_slaves_pport_actv);
+
+static int mlx4_slaves_closest_port(struct mlx4_dev *dev, int slave, int port)
+{
+	struct mlx4_active_ports actv_ports = mlx4_get_active_ports(dev, slave);
+	int min_port = find_first_bit(actv_ports.ports, dev->caps.num_ports)
+			+ 1;
+	int max_port = min_port +
+		bitmap_weight(actv_ports.ports, dev->caps.num_ports);
+
+	if (port < min_port)
+		port = min_port;
+	else if (port >= max_port)
+		port = max_port - 1;
+
+	return port;
+}
+
+static int mlx4_set_vport_qos(struct mlx4_priv *priv, int slave, int port,
+			      int max_tx_rate)
+{
+	int i;
+	int err;
+	struct mlx4_qos_manager *port_qos;
+	struct mlx4_dev *dev = &priv->dev;
+	struct mlx4_vport_qos_param vpp_qos[MLX4_NUM_UP];
+
+	port_qos = &priv->mfunc.master.qos_ctl[port];
+	memset(vpp_qos, 0, sizeof(struct mlx4_vport_qos_param) * MLX4_NUM_UP);
+
+	if (slave > port_qos->num_of_qos_vfs) {
+		mlx4_info(dev, "No available VPP resources for this VF\n");
+		return -EINVAL;
+	}
+
+	/* Query for default QoS values from Vport 0 is needed */
+	err = mlx4_SET_VPORT_QOS_get(dev, port, 0, vpp_qos);
+	if (err) {
+		mlx4_info(dev, "Failed to query Vport 0 QoS values\n");
+		return err;
+	}
+
+	for (i = 0; i < MLX4_NUM_UP; i++) {
+		if (test_bit(i, port_qos->priority_bm) && max_tx_rate) {
+			vpp_qos[i].max_avg_bw = max_tx_rate;
+			vpp_qos[i].enable = 1;
+		} else {
+			/* if user supplied tx_rate == 0, meaning no rate limit
+			 * configuration is required. so we are leaving the
+			 * value of max_avg_bw as queried from Vport 0.
+			 */
+			vpp_qos[i].enable = 0;
+		}
+	}
+
+	err = mlx4_SET_VPORT_QOS_set(dev, port, slave, vpp_qos);
+	if (err) {
+		mlx4_info(dev, "Failed to set Vport %d QoS values\n", slave);
+		return err;
+	}
+
+	return 0;
+}
+
+static bool mlx4_is_vf_vst_and_prio_qos(struct mlx4_dev *dev, int port,
+					struct mlx4_vport_state *vf_admin)
+{
+	struct mlx4_qos_manager *info;
+	struct mlx4_priv *priv = mlx4_priv(dev);
+
+	if (!mlx4_is_master(dev) ||
+	    !(dev->caps.flags2 & MLX4_DEV_CAP_FLAG2_QOS_VPP))
+		return false;
+
+	info = &priv->mfunc.master.qos_ctl[port];
+
+	if (vf_admin->default_vlan != MLX4_VGT &&
+	    test_bit(vf_admin->default_qos, info->priority_bm))
+		return true;
+
+	return false;
+}
+
+static bool mlx4_valid_vf_state_change(struct mlx4_dev *dev, int port,
+				       struct mlx4_vport_state *vf_admin,
+				       int vlan, int qos)
+{
+	struct mlx4_vport_state dummy_admin = {0};
+
+	if (!mlx4_is_vf_vst_and_prio_qos(dev, port, vf_admin) ||
+	    !vf_admin->tx_rate)
+		return true;
+
+	dummy_admin.default_qos = qos;
+	dummy_admin.default_vlan = vlan;
+
+	/* VF wants to move to other VST state which is valid with current
+	 * rate limit. Either differnt default vlan in VST or other
+	 * supported QoS priority. Otherwise we don't allow this change when
+	 * the TX rate is still configured.
+	 */
+	if (mlx4_is_vf_vst_and_prio_qos(dev, port, &dummy_admin))
+		return true;
+
+	mlx4_info(dev, "Cannot change VF state to %s while rate is set\n",
+		  (vlan == MLX4_VGT) ? "VGT" : "VST");
+
+	if (vlan != MLX4_VGT)
+		mlx4_info(dev, "VST priority %d not supported for QoS\n", qos);
+
+	mlx4_info(dev, "Please set rate to 0 prior to this VF state change\n");
+
+	return false;
+}
+
+int mlx4_set_vf_mac(struct mlx4_dev *dev, int port, int vf, u8 *mac)
+{
+	struct mlx4_priv *priv = mlx4_priv(dev);
+	struct mlx4_vport_state *s_info;
+	int slave;
+
+	if (!mlx4_is_master(dev))
+		return -EPROTONOSUPPORT;
+
+	if (is_multicast_ether_addr(mac))
+		return -EINVAL;
+
+	slave = mlx4_get_slave_indx(dev, vf);
+	if (slave < 0)
+		return -EINVAL;
+
+	port = mlx4_slaves_closest_port(dev, slave, port);
+	s_info = &priv->mfunc.master.vf_admin[slave].vport[port];
+
+	if (s_info->spoofchk && is_zero_ether_addr(mac)) {
+		mlx4_info(dev, "MAC invalidation is not allowed when spoofchk is on\n");
+		return -EPERM;
+	}
+
+	s_info->mac = mlx4_mac_to_u64(mac);
+	mlx4_info(dev, "default mac on vf %d port %d to %llX will take effect only after vf restart\n",
+		  vf, port, s_info->mac);
+	return 0;
+}
+EXPORT_SYMBOL_GPL(mlx4_set_vf_mac);
+
+
+int mlx4_set_vf_vlan(struct mlx4_dev *dev, int port, int vf, u16 vlan, u8 qos,
+		     __be16 proto)
+{
+	struct mlx4_priv *priv = mlx4_priv(dev);
+	struct mlx4_vport_state *vf_admin;
+	struct mlx4_slave_state *slave_state;
+	struct mlx4_vport_oper_state *vf_oper;
+	int slave;
+
+	if ((!mlx4_is_master(dev)) ||
+	    !(dev->caps.flags2 & MLX4_DEV_CAP_FLAG2_VLAN_CONTROL))
+		return -EPROTONOSUPPORT;
+
+	if ((vlan > 4095) || (qos > 7))
+		return -EINVAL;
+
+	if (proto == htons(ETH_P_8021AD) &&
+	    !(dev->caps.flags2 & MLX4_DEV_CAP_FLAG2_SVLAN_BY_QP))
+		return -EPROTONOSUPPORT;
+
+	if (proto != htons(ETH_P_8021Q) &&
+	    proto != htons(ETH_P_8021AD))
+		return -EINVAL;
+
+	if ((proto == htons(ETH_P_8021AD)) &&
+	    ((vlan == 0) || (vlan == MLX4_VGT)))
+		return -EINVAL;
+
+	slave = mlx4_get_slave_indx(dev, vf);
+	if (slave < 0)
+		return -EINVAL;
+
+	slave_state = &priv->mfunc.master.slave_state[slave];
+	if ((proto == htons(ETH_P_8021AD)) && (slave_state->active) &&
+	    (!slave_state->vst_qinq_supported)) {
+		mlx4_err(dev, "vf %d does not support VST QinQ mode\n", vf);
+		return -EPROTONOSUPPORT;
+	}
+	port = mlx4_slaves_closest_port(dev, slave, port);
+	vf_admin = &priv->mfunc.master.vf_admin[slave].vport[port];
+	vf_oper = &priv->mfunc.master.vf_oper[slave].vport[port];
+
+	if (!mlx4_valid_vf_state_change(dev, port, vf_admin, vlan, qos))
+		return -EPERM;
+
+	if ((0 == vlan) && (0 == qos))
+		vf_admin->default_vlan = MLX4_VGT;
+	else
+		vf_admin->default_vlan = vlan;
+	vf_admin->default_qos = qos;
+	vf_admin->vlan_proto = proto;
+
+	/* If rate was configured prior to VST, we saved the configured rate
+	 * in vf_admin->rate and now, if priority supported we enforce the QoS
+	 */
+	if (mlx4_is_vf_vst_and_prio_qos(dev, port, vf_admin) &&
+	    vf_admin->tx_rate)
+		vf_admin->qos_vport = slave;
+
+	/* Try to activate new vf state without restart,
+	 * this option is not supported while moving to VST QinQ mode.
+	 */
+	if ((proto == htons(ETH_P_8021AD) &&
+	     vf_oper->state.vlan_proto != proto) ||
+	    mlx4_master_immediate_activate_vlan_qos(priv, slave, port))
+		mlx4_info(dev,
+			  "updating vf %d port %d config will take effect on next VF restart\n",
+			  vf, port);
+	return 0;
+}
+EXPORT_SYMBOL_GPL(mlx4_set_vf_vlan);
+
+int mlx4_set_vf_rate(struct mlx4_dev *dev, int port, int vf, int min_tx_rate,
+		     int max_tx_rate)
+{
+	int err;
+	int slave;
+	struct mlx4_vport_state *vf_admin;
+	struct mlx4_priv *priv = mlx4_priv(dev);
+
+	if (!mlx4_is_master(dev) ||
+	    !(dev->caps.flags2 & MLX4_DEV_CAP_FLAG2_QOS_VPP))
+		return -EPROTONOSUPPORT;
+
+	if (min_tx_rate) {
+		mlx4_info(dev, "Minimum BW share not supported\n");
+		return -EPROTONOSUPPORT;
+	}
+
+	slave = mlx4_get_slave_indx(dev, vf);
+	if (slave < 0)
+		return -EINVAL;
+
+	port = mlx4_slaves_closest_port(dev, slave, port);
+	vf_admin = &priv->mfunc.master.vf_admin[slave].vport[port];
+
+	err = mlx4_set_vport_qos(priv, slave, port, max_tx_rate);
+	if (err) {
+		mlx4_info(dev, "vf %d failed to set rate %d\n", vf,
+			  max_tx_rate);
+		return err;
+	}
+
+	vf_admin->tx_rate = max_tx_rate;
+	/* if VF is not in supported mode (VST with supported prio),
+	 * we do not change vport configuration for its QPs, but save
+	 * the rate, so it will be enforced when it moves to supported
+	 * mode next time.
+	 */
+	if (!mlx4_is_vf_vst_and_prio_qos(dev, port, vf_admin)) {
+		mlx4_info(dev,
+			  "rate set for VF %d when not in valid state\n", vf);
+
+		if (vf_admin->default_vlan != MLX4_VGT)
+			mlx4_info(dev, "VST priority not supported by QoS\n");
+		else
+			mlx4_info(dev, "VF in VGT mode (needed VST)\n");
+
+		mlx4_info(dev,
+			  "rate %d take affect when VF moves to valid state\n",
+			  max_tx_rate);
+		return 0;
+	}
+
+	/* If user sets rate 0 assigning default vport for its QPs */
+	vf_admin->qos_vport = max_tx_rate ? slave : MLX4_VPP_DEFAULT_VPORT;
+
+	if (priv->mfunc.master.slave_state[slave].active &&
+	    dev->caps.flags2 & MLX4_DEV_CAP_FLAG2_UPDATE_QP)
+		mlx4_master_immediate_activate_vlan_qos(priv, slave, port);
+
+	return 0;
+}
+EXPORT_SYMBOL_GPL(mlx4_set_vf_rate);
+
+ /* mlx4_get_slave_default_vlan -
+ * return true if VST ( default vlan)
+ * if VST, will return vlan & qos (if not NULL)
+ */
+bool mlx4_get_slave_default_vlan(struct mlx4_dev *dev, int port, int slave,
+				 u16 *vlan, u8 *qos)
+{
+	struct mlx4_vport_oper_state *vp_oper;
+	struct mlx4_priv *priv;
+
+	priv = mlx4_priv(dev);
+	port = mlx4_slaves_closest_port(dev, slave, port);
+	vp_oper = &priv->mfunc.master.vf_oper[slave].vport[port];
+
+	if (MLX4_VGT != vp_oper->state.default_vlan) {
+		if (vlan)
+			*vlan = vp_oper->state.default_vlan;
+		if (qos)
+			*qos = vp_oper->state.default_qos;
+		return true;
+	}
+	return false;
+}
+EXPORT_SYMBOL_GPL(mlx4_get_slave_default_vlan);
+
+int mlx4_set_vf_spoofchk(struct mlx4_dev *dev, int port, int vf, bool setting)
+{
+	struct mlx4_priv *priv = mlx4_priv(dev);
+	struct mlx4_vport_state *s_info;
+	int slave;
+	u8 mac[ETH_ALEN];
+
+	if ((!mlx4_is_master(dev)) ||
+	    !(dev->caps.flags2 & MLX4_DEV_CAP_FLAG2_FSM))
+		return -EPROTONOSUPPORT;
+
+	slave = mlx4_get_slave_indx(dev, vf);
+	if (slave < 0)
+		return -EINVAL;
+
+	port = mlx4_slaves_closest_port(dev, slave, port);
+	s_info = &priv->mfunc.master.vf_admin[slave].vport[port];
+
+	mlx4_u64_to_mac(mac, s_info->mac);
+	if (setting && !is_valid_ether_addr(mac)) {
+		mlx4_info(dev, "Illegal MAC with spoofchk\n");
+		return -EPERM;
+	}
+
+	s_info->spoofchk = setting;
+
+	return 0;
+}
+EXPORT_SYMBOL_GPL(mlx4_set_vf_spoofchk);
+
+int mlx4_get_vf_config(struct mlx4_dev *dev, int port, int vf, struct ifla_vf_info *ivf)
+{
+	struct mlx4_priv *priv = mlx4_priv(dev);
+	struct mlx4_vport_state *s_info;
+	int slave;
+
+	if (!mlx4_is_master(dev))
+		return -EPROTONOSUPPORT;
+
+	slave = mlx4_get_slave_indx(dev, vf);
+	if (slave < 0)
+		return -EINVAL;
+
+	s_info = &priv->mfunc.master.vf_admin[slave].vport[port];
+	ivf->vf = vf;
+
+	/* need to convert it to a func */
+	ivf->mac[0] = ((s_info->mac >> (5*8)) & 0xff);
+	ivf->mac[1] = ((s_info->mac >> (4*8)) & 0xff);
+	ivf->mac[2] = ((s_info->mac >> (3*8)) & 0xff);
+	ivf->mac[3] = ((s_info->mac >> (2*8)) & 0xff);
+	ivf->mac[4] = ((s_info->mac >> (1*8)) & 0xff);
+	ivf->mac[5] = ((s_info->mac)  & 0xff);
+
+	ivf->vlan		= s_info->default_vlan;
+	ivf->qos		= s_info->default_qos;
+	ivf->vlan_proto		= s_info->vlan_proto;
+
+	if (mlx4_is_vf_vst_and_prio_qos(dev, port, s_info))
+		ivf->max_tx_rate = s_info->tx_rate;
+	else
+		ivf->max_tx_rate = 0;
+
+	ivf->min_tx_rate	= 0;
+	ivf->spoofchk		= s_info->spoofchk;
+	ivf->linkstate		= s_info->link_state;
+
+	return 0;
+}
+EXPORT_SYMBOL_GPL(mlx4_get_vf_config);
+
+int mlx4_set_vf_link_state(struct mlx4_dev *dev, int port, int vf, int link_state)
+{
+	struct mlx4_priv *priv = mlx4_priv(dev);
+	struct mlx4_vport_state *s_info;
+	int slave;
+	u8 link_stat_event;
+
+	slave = mlx4_get_slave_indx(dev, vf);
+	if (slave < 0)
+		return -EINVAL;
+
+	port = mlx4_slaves_closest_port(dev, slave, port);
+	switch (link_state) {
+	case IFLA_VF_LINK_STATE_AUTO:
+		/* get current link state */
+		if (!priv->sense.do_sense_port[port])
+			link_stat_event = MLX4_PORT_CHANGE_SUBTYPE_ACTIVE;
+		else
+			link_stat_event = MLX4_PORT_CHANGE_SUBTYPE_DOWN;
+	    break;
+
+	case IFLA_VF_LINK_STATE_ENABLE:
+		link_stat_event = MLX4_PORT_CHANGE_SUBTYPE_ACTIVE;
+	    break;
+
+	case IFLA_VF_LINK_STATE_DISABLE:
+		link_stat_event = MLX4_PORT_CHANGE_SUBTYPE_DOWN;
+	    break;
+
+	default:
+		mlx4_warn(dev, "unknown value for link_state %02x on slave %d port %d\n",
+			  link_state, slave, port);
+		return -EINVAL;
+	};
+	s_info = &priv->mfunc.master.vf_admin[slave].vport[port];
+	s_info->link_state = link_state;
+
+	/* send event */
+	mlx4_gen_port_state_change_eqe(dev, slave, port, link_stat_event);
+
+	if (mlx4_master_immediate_activate_vlan_qos(priv, slave, port))
+		mlx4_dbg(dev,
+			 "updating vf %d port %d no link state HW enforcement\n",
+			 vf, port);
+	return 0;
+}
+EXPORT_SYMBOL_GPL(mlx4_set_vf_link_state);
+
+int mlx4_get_counter_stats(struct mlx4_dev *dev, int counter_index,
+			   struct mlx4_counter *counter_stats, int reset)
+{
+	struct mlx4_cmd_mailbox *mailbox = NULL;
+	struct mlx4_counter *tmp_counter;
+	int err;
+	u32 if_stat_in_mod;
+
+	if (!counter_stats)
+		return -EINVAL;
+
+	if (counter_index == MLX4_SINK_COUNTER_INDEX(dev))
+		return 0;
+
+	mailbox = mlx4_alloc_cmd_mailbox(dev);
+	if (IS_ERR(mailbox))
+		return PTR_ERR(mailbox);
+
+	memset(mailbox->buf, 0, sizeof(struct mlx4_counter));
+	if_stat_in_mod = counter_index;
+	if (reset)
+		if_stat_in_mod |= MLX4_QUERY_IF_STAT_RESET;
+	err = mlx4_cmd_box(dev, 0, mailbox->dma,
+			   if_stat_in_mod, 0,
+			   MLX4_CMD_QUERY_IF_STAT,
+			   MLX4_CMD_TIME_CLASS_C,
+			   MLX4_CMD_NATIVE);
+	if (err) {
+		mlx4_dbg(dev, "%s: failed to read statistics for counter index %d\n",
+			 __func__, counter_index);
+		goto if_stat_out;
+	}
+	tmp_counter = (struct mlx4_counter *)mailbox->buf;
+	counter_stats->counter_mode = tmp_counter->counter_mode;
+	if (counter_stats->counter_mode == 0) {
+		counter_stats->rx_frames =
+			cpu_to_be64(be64_to_cpu(counter_stats->rx_frames) +
+				    be64_to_cpu(tmp_counter->rx_frames));
+		counter_stats->tx_frames =
+			cpu_to_be64(be64_to_cpu(counter_stats->tx_frames) +
+				    be64_to_cpu(tmp_counter->tx_frames));
+		counter_stats->rx_bytes =
+			cpu_to_be64(be64_to_cpu(counter_stats->rx_bytes) +
+				    be64_to_cpu(tmp_counter->rx_bytes));
+		counter_stats->tx_bytes =
+			cpu_to_be64(be64_to_cpu(counter_stats->tx_bytes) +
+				    be64_to_cpu(tmp_counter->tx_bytes));
+	}
+
+if_stat_out:
+	mlx4_free_cmd_mailbox(dev, mailbox);
+
+	return err;
+}
+EXPORT_SYMBOL_GPL(mlx4_get_counter_stats);
+
+int mlx4_get_vf_stats(struct mlx4_dev *dev, int port, int vf_idx,
+		      struct ifla_vf_stats *vf_stats)
+{
+	struct mlx4_counter tmp_vf_stats;
+	int slave;
+	int err = 0;
+
+	if (!vf_stats)
+		return -EINVAL;
+
+	if (!mlx4_is_master(dev))
+		return -EPROTONOSUPPORT;
+
+	slave = mlx4_get_slave_indx(dev, vf_idx);
+	if (slave < 0)
+		return -EINVAL;
+
+	port = mlx4_slaves_closest_port(dev, slave, port);
+	err = mlx4_calc_vf_counters(dev, slave, port, &tmp_vf_stats);
+	if (!err && tmp_vf_stats.counter_mode == 0) {
+		vf_stats->rx_packets = be64_to_cpu(tmp_vf_stats.rx_frames);
+		vf_stats->tx_packets = be64_to_cpu(tmp_vf_stats.tx_frames);
+		vf_stats->rx_bytes = be64_to_cpu(tmp_vf_stats.rx_bytes);
+		vf_stats->tx_bytes = be64_to_cpu(tmp_vf_stats.tx_bytes);
+	}
+
+	return err;
+}
+EXPORT_SYMBOL_GPL(mlx4_get_vf_stats);
+
+int mlx4_vf_smi_enabled(struct mlx4_dev *dev, int slave, int port)
+{
+	struct mlx4_priv *priv = mlx4_priv(dev);
+
+	if (slave < 1 || slave >= dev->num_slaves ||
+	    port < 1 || port > MLX4_MAX_PORTS)
+		return 0;
+
+	return priv->mfunc.master.vf_oper[slave].smi_enabled[port] ==
+		MLX4_VF_SMI_ENABLED;
+}
+EXPORT_SYMBOL_GPL(mlx4_vf_smi_enabled);
+
+int mlx4_vf_get_enable_smi_admin(struct mlx4_dev *dev, int slave, int port)
+{
+	struct mlx4_priv *priv = mlx4_priv(dev);
+
+	if (slave == mlx4_master_func_num(dev))
+		return 1;
+
+	if (slave < 1 || slave >= dev->num_slaves ||
+	    port < 1 || port > MLX4_MAX_PORTS)
+		return 0;
+
+	return priv->mfunc.master.vf_admin[slave].enable_smi[port] ==
+		MLX4_VF_SMI_ENABLED;
+}
+EXPORT_SYMBOL_GPL(mlx4_vf_get_enable_smi_admin);
+
+int mlx4_vf_set_enable_smi_admin(struct mlx4_dev *dev, int slave, int port,
+				 int enabled)
+{
+	struct mlx4_priv *priv = mlx4_priv(dev);
+	struct mlx4_active_ports actv_ports = mlx4_get_active_ports(
+			&priv->dev, slave);
+	int min_port = find_first_bit(actv_ports.ports,
+				      priv->dev.caps.num_ports) + 1;
+	int max_port = min_port - 1 +
+		bitmap_weight(actv_ports.ports, priv->dev.caps.num_ports);
+
+	if (slave == mlx4_master_func_num(dev))
+		return 0;
+
+	if (slave < 1 || slave >= dev->num_slaves ||
+	    port < 1 || port > MLX4_MAX_PORTS ||
+	    enabled < 0 || enabled > 1)
+		return -EINVAL;
+
+	if (min_port == max_port && dev->caps.num_ports > 1) {
+		mlx4_info(dev, "SMI access disallowed for single ported VFs\n");
+		return -EPROTONOSUPPORT;
+	}
+
+	priv->mfunc.master.vf_admin[slave].enable_smi[port] = enabled;
+	return 0;
+}
+EXPORT_SYMBOL_GPL(mlx4_vf_set_enable_smi_admin);
diff --git a/drivers/net/ethernet/mellanox/mlx4/cq.c b/drivers/net/ethernet/mellanox/mlx4/cq.c
new file mode 100644
index 000000000..d8e9a3231
--- /dev/null
+++ b/drivers/net/ethernet/mellanox/mlx4/cq.c
@@ -0,0 +1,421 @@
+/*
+ * Copyright (c) 2004, 2005 Topspin Communications.  All rights reserved.
+ * Copyright (c) 2005 Sun Microsystems, Inc. All rights reserved.
+ * Copyright (c) 2005, 2006, 2007 Cisco Systems, Inc. All rights reserved.
+ * Copyright (c) 2005, 2006, 2007, 2008 Mellanox Technologies. All rights reserved.
+ * Copyright (c) 2004 Voltaire, Inc. All rights reserved.
+ *
+ * This software is available to you under a choice of one of two
+ * licenses.  You may choose to be licensed under the terms of the GNU
+ * General Public License (GPL) Version 2, available from the file
+ * COPYING in the main directory of this source tree, or the
+ * OpenIB.org BSD license below:
+ *
+ *     Redistribution and use in source and binary forms, with or
+ *     without modification, are permitted provided that the following
+ *     conditions are met:
+ *
+ *      - Redistributions of source code must retain the above
+ *        copyright notice, this list of conditions and the following
+ *        disclaimer.
+ *
+ *      - Redistributions in binary form must reproduce the above
+ *        copyright notice, this list of conditions and the following
+ *        disclaimer in the documentation and/or other materials
+ *        provided with the distribution.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
+ * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
+ * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS
+ * BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN
+ * ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
+ * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+ * SOFTWARE.
+ */
+
+#include <linux/hardirq.h>
+#include <linux/export.h>
+
+#include <linux/mlx4/cmd.h>
+#include <linux/mlx4/cq.h>
+
+#include "mlx4.h"
+#include "icm.h"
+
+#define MLX4_CQ_STATUS_OK		( 0 << 28)
+#define MLX4_CQ_STATUS_OVERFLOW		( 9 << 28)
+#define MLX4_CQ_STATUS_WRITE_FAIL	(10 << 28)
+#define MLX4_CQ_FLAG_CC			( 1 << 18)
+#define MLX4_CQ_FLAG_OI			( 1 << 17)
+#define MLX4_CQ_STATE_ARMED		( 9 <<  8)
+#define MLX4_CQ_STATE_ARMED_SOL		( 6 <<  8)
+#define MLX4_EQ_STATE_FIRED		(10 <<  8)
+
+#define TASKLET_MAX_TIME 2
+#define TASKLET_MAX_TIME_JIFFIES msecs_to_jiffies(TASKLET_MAX_TIME)
+
+void mlx4_cq_tasklet_cb(unsigned long data)
+{
+	unsigned long flags;
+	unsigned long end = jiffies + TASKLET_MAX_TIME_JIFFIES;
+	struct mlx4_eq_tasklet *ctx = (struct mlx4_eq_tasklet *)data;
+	struct mlx4_cq *mcq, *temp;
+
+	spin_lock_irqsave(&ctx->lock, flags);
+	list_splice_tail_init(&ctx->list, &ctx->process_list);
+	spin_unlock_irqrestore(&ctx->lock, flags);
+
+	list_for_each_entry_safe(mcq, temp, &ctx->process_list, tasklet_ctx.list) {
+		list_del_init(&mcq->tasklet_ctx.list);
+		mcq->tasklet_ctx.comp(mcq);
+		if (refcount_dec_and_test(&mcq->refcount))
+			complete(&mcq->free);
+		if (time_after(jiffies, end))
+			break;
+	}
+
+	if (!list_empty(&ctx->process_list))
+		tasklet_schedule(&ctx->task);
+}
+
+static void mlx4_add_cq_to_tasklet(struct mlx4_cq *cq)
+{
+	struct mlx4_eq_tasklet *tasklet_ctx = cq->tasklet_ctx.priv;
+	unsigned long flags;
+	bool kick;
+
+	spin_lock_irqsave(&tasklet_ctx->lock, flags);
+	/* When migrating CQs between EQs will be implemented, please note
+	 * that you need to sync this point. It is possible that
+	 * while migrating a CQ, completions on the old EQs could
+	 * still arrive.
+	 */
+	if (list_empty_careful(&cq->tasklet_ctx.list)) {
+		refcount_inc(&cq->refcount);
+		kick = list_empty(&tasklet_ctx->list);
+		list_add_tail(&cq->tasklet_ctx.list, &tasklet_ctx->list);
+		if (kick)
+			tasklet_schedule(&tasklet_ctx->task);
+	}
+	spin_unlock_irqrestore(&tasklet_ctx->lock, flags);
+}
+
+void mlx4_cq_completion(struct mlx4_dev *dev, u32 cqn)
+{
+	struct mlx4_cq *cq;
+
+	rcu_read_lock();
+	cq = radix_tree_lookup(&mlx4_priv(dev)->cq_table.tree,
+			       cqn & (dev->caps.num_cqs - 1));
+	rcu_read_unlock();
+
+	if (!cq) {
+		mlx4_dbg(dev, "Completion event for bogus CQ %08x\n", cqn);
+		return;
+	}
+
+	/* Acessing the CQ outside of rcu_read_lock is safe, because
+	 * the CQ is freed only after interrupt handling is completed.
+	 */
+	++cq->arm_sn;
+
+	cq->comp(cq);
+}
+
+void mlx4_cq_event(struct mlx4_dev *dev, u32 cqn, int event_type)
+{
+	struct mlx4_cq_table *cq_table = &mlx4_priv(dev)->cq_table;
+	struct mlx4_cq *cq;
+
+	rcu_read_lock();
+	cq = radix_tree_lookup(&cq_table->tree, cqn & (dev->caps.num_cqs - 1));
+	rcu_read_unlock();
+
+	if (!cq) {
+		mlx4_dbg(dev, "Async event for bogus CQ %08x\n", cqn);
+		return;
+	}
+
+	/* Acessing the CQ outside of rcu_read_lock is safe, because
+	 * the CQ is freed only after interrupt handling is completed.
+	 */
+	cq->event(cq, event_type);
+}
+
+static int mlx4_SW2HW_CQ(struct mlx4_dev *dev, struct mlx4_cmd_mailbox *mailbox,
+			 int cq_num)
+{
+	return mlx4_cmd(dev, mailbox->dma, cq_num, 0,
+			MLX4_CMD_SW2HW_CQ, MLX4_CMD_TIME_CLASS_A,
+			MLX4_CMD_WRAPPED);
+}
+
+static int mlx4_MODIFY_CQ(struct mlx4_dev *dev, struct mlx4_cmd_mailbox *mailbox,
+			 int cq_num, u32 opmod)
+{
+	return mlx4_cmd(dev, mailbox->dma, cq_num, opmod, MLX4_CMD_MODIFY_CQ,
+			MLX4_CMD_TIME_CLASS_A, MLX4_CMD_WRAPPED);
+}
+
+static int mlx4_HW2SW_CQ(struct mlx4_dev *dev, struct mlx4_cmd_mailbox *mailbox,
+			 int cq_num)
+{
+	return mlx4_cmd_box(dev, 0, mailbox ? mailbox->dma : 0,
+			    cq_num, mailbox ? 0 : 1, MLX4_CMD_HW2SW_CQ,
+			    MLX4_CMD_TIME_CLASS_A, MLX4_CMD_WRAPPED);
+}
+
+int mlx4_cq_modify(struct mlx4_dev *dev, struct mlx4_cq *cq,
+		   u16 count, u16 period)
+{
+	struct mlx4_cmd_mailbox *mailbox;
+	struct mlx4_cq_context *cq_context;
+	int err;
+
+	mailbox = mlx4_alloc_cmd_mailbox(dev);
+	if (IS_ERR(mailbox))
+		return PTR_ERR(mailbox);
+
+	cq_context = mailbox->buf;
+	cq_context->cq_max_count = cpu_to_be16(count);
+	cq_context->cq_period    = cpu_to_be16(period);
+
+	err = mlx4_MODIFY_CQ(dev, mailbox, cq->cqn, 1);
+
+	mlx4_free_cmd_mailbox(dev, mailbox);
+	return err;
+}
+EXPORT_SYMBOL_GPL(mlx4_cq_modify);
+
+int mlx4_cq_resize(struct mlx4_dev *dev, struct mlx4_cq *cq,
+		   int entries, struct mlx4_mtt *mtt)
+{
+	struct mlx4_cmd_mailbox *mailbox;
+	struct mlx4_cq_context *cq_context;
+	u64 mtt_addr;
+	int err;
+
+	mailbox = mlx4_alloc_cmd_mailbox(dev);
+	if (IS_ERR(mailbox))
+		return PTR_ERR(mailbox);
+
+	cq_context = mailbox->buf;
+	cq_context->logsize_usrpage = cpu_to_be32(ilog2(entries) << 24);
+	cq_context->log_page_size   = mtt->page_shift - 12;
+	mtt_addr = mlx4_mtt_addr(dev, mtt);
+	cq_context->mtt_base_addr_h = mtt_addr >> 32;
+	cq_context->mtt_base_addr_l = cpu_to_be32(mtt_addr & 0xffffffff);
+
+	err = mlx4_MODIFY_CQ(dev, mailbox, cq->cqn, 0);
+
+	mlx4_free_cmd_mailbox(dev, mailbox);
+	return err;
+}
+EXPORT_SYMBOL_GPL(mlx4_cq_resize);
+
+int __mlx4_cq_alloc_icm(struct mlx4_dev *dev, int *cqn)
+{
+	struct mlx4_priv *priv = mlx4_priv(dev);
+	struct mlx4_cq_table *cq_table = &priv->cq_table;
+	int err;
+
+	*cqn = mlx4_bitmap_alloc(&cq_table->bitmap);
+	if (*cqn == -1)
+		return -ENOMEM;
+
+	err = mlx4_table_get(dev, &cq_table->table, *cqn);
+	if (err)
+		goto err_out;
+
+	err = mlx4_table_get(dev, &cq_table->cmpt_table, *cqn);
+	if (err)
+		goto err_put;
+	return 0;
+
+err_put:
+	mlx4_table_put(dev, &cq_table->table, *cqn);
+
+err_out:
+	mlx4_bitmap_free(&cq_table->bitmap, *cqn, MLX4_NO_RR);
+	return err;
+}
+
+static int mlx4_cq_alloc_icm(struct mlx4_dev *dev, int *cqn, u8 usage)
+{
+	u32 in_modifier = RES_CQ | (((u32)usage & 3) << 30);
+	u64 out_param;
+	int err;
+
+	if (mlx4_is_mfunc(dev)) {
+		err = mlx4_cmd_imm(dev, 0, &out_param, in_modifier,
+				   RES_OP_RESERVE_AND_MAP, MLX4_CMD_ALLOC_RES,
+				   MLX4_CMD_TIME_CLASS_A, MLX4_CMD_WRAPPED);
+		if (err)
+			return err;
+		else {
+			*cqn = get_param_l(&out_param);
+			return 0;
+		}
+	}
+	return __mlx4_cq_alloc_icm(dev, cqn);
+}
+
+void __mlx4_cq_free_icm(struct mlx4_dev *dev, int cqn)
+{
+	struct mlx4_priv *priv = mlx4_priv(dev);
+	struct mlx4_cq_table *cq_table = &priv->cq_table;
+
+	mlx4_table_put(dev, &cq_table->cmpt_table, cqn);
+	mlx4_table_put(dev, &cq_table->table, cqn);
+	mlx4_bitmap_free(&cq_table->bitmap, cqn, MLX4_NO_RR);
+}
+
+static void mlx4_cq_free_icm(struct mlx4_dev *dev, int cqn)
+{
+	u64 in_param = 0;
+	int err;
+
+	if (mlx4_is_mfunc(dev)) {
+		set_param_l(&in_param, cqn);
+		err = mlx4_cmd(dev, in_param, RES_CQ, RES_OP_RESERVE_AND_MAP,
+			       MLX4_CMD_FREE_RES,
+			       MLX4_CMD_TIME_CLASS_A, MLX4_CMD_WRAPPED);
+		if (err)
+			mlx4_warn(dev, "Failed freeing cq:%d\n", cqn);
+	} else
+		__mlx4_cq_free_icm(dev, cqn);
+}
+
+int mlx4_cq_alloc(struct mlx4_dev *dev, int nent,
+		  struct mlx4_mtt *mtt, struct mlx4_uar *uar, u64 db_rec,
+		  struct mlx4_cq *cq, unsigned vector, int collapsed,
+		  int timestamp_en)
+{
+	struct mlx4_priv *priv = mlx4_priv(dev);
+	struct mlx4_cq_table *cq_table = &priv->cq_table;
+	struct mlx4_cmd_mailbox *mailbox;
+	struct mlx4_cq_context *cq_context;
+	u64 mtt_addr;
+	int err;
+
+	if (vector >= dev->caps.num_comp_vectors)
+		return -EINVAL;
+
+	cq->vector = vector;
+
+	err = mlx4_cq_alloc_icm(dev, &cq->cqn, cq->usage);
+	if (err)
+		return err;
+
+	spin_lock(&cq_table->lock);
+	err = radix_tree_insert(&cq_table->tree, cq->cqn, cq);
+	spin_unlock(&cq_table->lock);
+	if (err)
+		goto err_icm;
+
+	mailbox = mlx4_alloc_cmd_mailbox(dev);
+	if (IS_ERR(mailbox)) {
+		err = PTR_ERR(mailbox);
+		goto err_radix;
+	}
+
+	cq_context = mailbox->buf;
+	cq_context->flags	    = cpu_to_be32(!!collapsed << 18);
+	if (timestamp_en)
+		cq_context->flags  |= cpu_to_be32(1 << 19);
+
+	cq_context->logsize_usrpage =
+		cpu_to_be32((ilog2(nent) << 24) |
+			    mlx4_to_hw_uar_index(dev, uar->index));
+	cq_context->comp_eqn	    = priv->eq_table.eq[MLX4_CQ_TO_EQ_VECTOR(vector)].eqn;
+	cq_context->log_page_size   = mtt->page_shift - MLX4_ICM_PAGE_SHIFT;
+
+	mtt_addr = mlx4_mtt_addr(dev, mtt);
+	cq_context->mtt_base_addr_h = mtt_addr >> 32;
+	cq_context->mtt_base_addr_l = cpu_to_be32(mtt_addr & 0xffffffff);
+	cq_context->db_rec_addr     = cpu_to_be64(db_rec);
+
+	err = mlx4_SW2HW_CQ(dev, mailbox, cq->cqn);
+	mlx4_free_cmd_mailbox(dev, mailbox);
+	if (err)
+		goto err_radix;
+
+	cq->cons_index = 0;
+	cq->arm_sn     = 1;
+	cq->uar        = uar;
+	refcount_set(&cq->refcount, 1);
+	init_completion(&cq->free);
+	cq->comp = mlx4_add_cq_to_tasklet;
+	cq->tasklet_ctx.priv =
+		&priv->eq_table.eq[MLX4_CQ_TO_EQ_VECTOR(vector)].tasklet_ctx;
+	INIT_LIST_HEAD(&cq->tasklet_ctx.list);
+
+
+	cq->irq = priv->eq_table.eq[MLX4_CQ_TO_EQ_VECTOR(vector)].irq;
+	return 0;
+
+err_radix:
+	spin_lock(&cq_table->lock);
+	radix_tree_delete(&cq_table->tree, cq->cqn);
+	spin_unlock(&cq_table->lock);
+
+err_icm:
+	mlx4_cq_free_icm(dev, cq->cqn);
+
+	return err;
+}
+EXPORT_SYMBOL_GPL(mlx4_cq_alloc);
+
+void mlx4_cq_free(struct mlx4_dev *dev, struct mlx4_cq *cq)
+{
+	struct mlx4_priv *priv = mlx4_priv(dev);
+	struct mlx4_cq_table *cq_table = &priv->cq_table;
+	int err;
+
+	err = mlx4_HW2SW_CQ(dev, NULL, cq->cqn);
+	if (err)
+		mlx4_warn(dev, "HW2SW_CQ failed (%d) for CQN %06x\n", err, cq->cqn);
+
+	spin_lock(&cq_table->lock);
+	radix_tree_delete(&cq_table->tree, cq->cqn);
+	spin_unlock(&cq_table->lock);
+
+	synchronize_irq(priv->eq_table.eq[MLX4_CQ_TO_EQ_VECTOR(cq->vector)].irq);
+	if (priv->eq_table.eq[MLX4_CQ_TO_EQ_VECTOR(cq->vector)].irq !=
+	    priv->eq_table.eq[MLX4_EQ_ASYNC].irq)
+		synchronize_irq(priv->eq_table.eq[MLX4_EQ_ASYNC].irq);
+
+	if (refcount_dec_and_test(&cq->refcount))
+		complete(&cq->free);
+	wait_for_completion(&cq->free);
+
+	mlx4_cq_free_icm(dev, cq->cqn);
+}
+EXPORT_SYMBOL_GPL(mlx4_cq_free);
+
+int mlx4_init_cq_table(struct mlx4_dev *dev)
+{
+	struct mlx4_cq_table *cq_table = &mlx4_priv(dev)->cq_table;
+	int err;
+
+	spin_lock_init(&cq_table->lock);
+	INIT_RADIX_TREE(&cq_table->tree, GFP_ATOMIC);
+	if (mlx4_is_slave(dev))
+		return 0;
+
+	err = mlx4_bitmap_init(&cq_table->bitmap, dev->caps.num_cqs,
+			       dev->caps.num_cqs - 1, dev->caps.reserved_cqs, 0);
+	if (err)
+		return err;
+
+	return 0;
+}
+
+void mlx4_cleanup_cq_table(struct mlx4_dev *dev)
+{
+	if (mlx4_is_slave(dev))
+		return;
+	/* Nothing to do to clean up radix_tree */
+	mlx4_bitmap_cleanup(&mlx4_priv(dev)->cq_table.bitmap);
+}
diff --git a/drivers/net/ethernet/mellanox/mlx4/crdump.c b/drivers/net/ethernet/mellanox/mlx4/crdump.c
new file mode 100644
index 000000000..88316c743
--- /dev/null
+++ b/drivers/net/ethernet/mellanox/mlx4/crdump.c
@@ -0,0 +1,239 @@
+/*
+ * Copyright (c) 2018, Mellanox Technologies. All rights reserved.
+ *
+ * This software is available to you under a choice of one of two
+ * licenses.  You may choose to be licensed under the terms of the GNU
+ * General Public License (GPL) Version 2, available from the file
+ * COPYING in the main directory of this source tree, or the
+ * OpenIB.org BSD license below:
+ *
+ *     Redistribution and use in source and binary forms, with or
+ *     without modification, are permitted provided that the following
+ *     conditions are met:
+ *
+ *      - Redistributions of source code must retain the above
+ *        copyright notice, this list of conditions and the following
+ *        disclaimer.
+ *
+ *      - Redistributions in binary form must reproduce the above
+ *        copyright notice, this list of conditions and the following
+ *        disclaimer in the documentation and/or other materials
+ *        provided with the distribution.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
+ * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
+ * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS
+ * BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN
+ * ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
+ * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+ * SOFTWARE.
+ */
+
+#include "mlx4.h"
+
+#define BAD_ACCESS			0xBADACCE5
+#define HEALTH_BUFFER_SIZE		0x40
+#define CR_ENABLE_BIT			swab32(BIT(6))
+#define CR_ENABLE_BIT_OFFSET		0xF3F04
+#define MAX_NUM_OF_DUMPS_TO_STORE	(8)
+
+static const char *region_cr_space_str = "cr-space";
+static const char *region_fw_health_str = "fw-health";
+
+/* Set to true in case cr enable bit was set to true before crdump */
+static bool crdump_enbale_bit_set;
+
+static void crdump_enable_crspace_access(struct mlx4_dev *dev,
+					 u8 __iomem *cr_space)
+{
+	/* Get current enable bit value */
+	crdump_enbale_bit_set =
+		readl(cr_space + CR_ENABLE_BIT_OFFSET) & CR_ENABLE_BIT;
+
+	/* Enable FW CR filter (set bit6 to 0) */
+	if (crdump_enbale_bit_set)
+		writel(readl(cr_space + CR_ENABLE_BIT_OFFSET) & ~CR_ENABLE_BIT,
+		       cr_space + CR_ENABLE_BIT_OFFSET);
+
+	/* Enable block volatile crspace accesses */
+	writel(swab32(1), cr_space + dev->caps.health_buffer_addrs +
+	       HEALTH_BUFFER_SIZE);
+}
+
+static void crdump_disable_crspace_access(struct mlx4_dev *dev,
+					  u8 __iomem *cr_space)
+{
+	/* Disable block volatile crspace accesses */
+	writel(0, cr_space + dev->caps.health_buffer_addrs +
+	       HEALTH_BUFFER_SIZE);
+
+	/* Restore FW CR filter value (set bit6 to original value) */
+	if (crdump_enbale_bit_set)
+		writel(readl(cr_space + CR_ENABLE_BIT_OFFSET) | CR_ENABLE_BIT,
+		       cr_space + CR_ENABLE_BIT_OFFSET);
+}
+
+static void mlx4_crdump_collect_crspace(struct mlx4_dev *dev,
+					u8 __iomem *cr_space,
+					u32 id)
+{
+	struct mlx4_fw_crdump *crdump = &dev->persist->crdump;
+	struct pci_dev *pdev = dev->persist->pdev;
+	unsigned long cr_res_size;
+	u8 *crspace_data;
+	int offset;
+	int err;
+
+	if (!crdump->region_crspace) {
+		mlx4_err(dev, "crdump: cr-space region is NULL\n");
+		return;
+	}
+
+	/* Try to collect CR space */
+	cr_res_size = pci_resource_len(pdev, 0);
+	crspace_data = kvmalloc(cr_res_size, GFP_KERNEL);
+	if (crspace_data) {
+		for (offset = 0; offset < cr_res_size; offset += 4)
+			*(u32 *)(crspace_data + offset) =
+					readl(cr_space + offset);
+
+		err = devlink_region_snapshot_create(crdump->region_crspace,
+						     cr_res_size, crspace_data,
+						     id, &kvfree);
+		if (err) {
+			kvfree(crspace_data);
+			mlx4_warn(dev, "crdump: devlink create %s snapshot id %d err %d\n",
+				  region_cr_space_str, id, err);
+		} else {
+			mlx4_info(dev, "crdump: added snapshot %d to devlink region %s\n",
+				  id, region_cr_space_str);
+		}
+	} else {
+		mlx4_err(dev, "crdump: Failed to allocate crspace buffer\n");
+	}
+}
+
+static void mlx4_crdump_collect_fw_health(struct mlx4_dev *dev,
+					  u8 __iomem *cr_space,
+					  u32 id)
+{
+	struct mlx4_fw_crdump *crdump = &dev->persist->crdump;
+	u8 *health_data;
+	int offset;
+	int err;
+
+	if (!crdump->region_fw_health) {
+		mlx4_err(dev, "crdump: fw-health region is NULL\n");
+		return;
+	}
+
+	/* Try to collect health buffer */
+	health_data = kvmalloc(HEALTH_BUFFER_SIZE, GFP_KERNEL);
+	if (health_data) {
+		u8 __iomem *health_buf_start =
+				cr_space + dev->caps.health_buffer_addrs;
+
+		for (offset = 0; offset < HEALTH_BUFFER_SIZE; offset += 4)
+			*(u32 *)(health_data + offset) =
+					readl(health_buf_start + offset);
+
+		err = devlink_region_snapshot_create(crdump->region_fw_health,
+						     HEALTH_BUFFER_SIZE,
+						     health_data,
+						     id, &kvfree);
+		if (err) {
+			kvfree(health_data);
+			mlx4_warn(dev, "crdump: devlink create %s snapshot id %d err %d\n",
+				  region_fw_health_str, id, err);
+		} else {
+			mlx4_info(dev, "crdump: added snapshot %d to devlink region %s\n",
+				  id, region_fw_health_str);
+		}
+	} else {
+		mlx4_err(dev, "crdump: Failed to allocate health buffer\n");
+	}
+}
+
+int mlx4_crdump_collect(struct mlx4_dev *dev)
+{
+	struct devlink *devlink = priv_to_devlink(mlx4_priv(dev));
+	struct mlx4_fw_crdump *crdump = &dev->persist->crdump;
+	struct pci_dev *pdev = dev->persist->pdev;
+	unsigned long cr_res_size;
+	u8 __iomem *cr_space;
+	u32 id;
+
+	if (!dev->caps.health_buffer_addrs) {
+		mlx4_info(dev, "crdump: FW doesn't support health buffer access, skipping\n");
+		return 0;
+	}
+
+	if (!crdump->snapshot_enable) {
+		mlx4_info(dev, "crdump: devlink snapshot disabled, skipping\n");
+		return 0;
+	}
+
+	cr_res_size = pci_resource_len(pdev, 0);
+
+	cr_space = ioremap(pci_resource_start(pdev, 0), cr_res_size);
+	if (!cr_space) {
+		mlx4_err(dev, "crdump: Failed to map pci cr region\n");
+		return -ENODEV;
+	}
+
+	crdump_enable_crspace_access(dev, cr_space);
+
+	/* Get the available snapshot ID for the dumps */
+	id = devlink_region_shapshot_id_get(devlink);
+
+	/* Try to capture dumps */
+	mlx4_crdump_collect_crspace(dev, cr_space, id);
+	mlx4_crdump_collect_fw_health(dev, cr_space, id);
+
+	crdump_disable_crspace_access(dev, cr_space);
+
+	iounmap(cr_space);
+	return 0;
+}
+
+int mlx4_crdump_init(struct mlx4_dev *dev)
+{
+	struct devlink *devlink = priv_to_devlink(mlx4_priv(dev));
+	struct mlx4_fw_crdump *crdump = &dev->persist->crdump;
+	struct pci_dev *pdev = dev->persist->pdev;
+
+	crdump->snapshot_enable = false;
+
+	/* Create cr-space region */
+	crdump->region_crspace =
+		devlink_region_create(devlink,
+				      region_cr_space_str,
+				      MAX_NUM_OF_DUMPS_TO_STORE,
+				      pci_resource_len(pdev, 0));
+	if (IS_ERR(crdump->region_crspace))
+		mlx4_warn(dev, "crdump: create devlink region %s err %ld\n",
+			  region_cr_space_str,
+			  PTR_ERR(crdump->region_crspace));
+
+	/* Create fw-health region */
+	crdump->region_fw_health =
+		devlink_region_create(devlink,
+				      region_fw_health_str,
+				      MAX_NUM_OF_DUMPS_TO_STORE,
+				      HEALTH_BUFFER_SIZE);
+	if (IS_ERR(crdump->region_fw_health))
+		mlx4_warn(dev, "crdump: create devlink region %s err %ld\n",
+			  region_fw_health_str,
+			  PTR_ERR(crdump->region_fw_health));
+
+	return 0;
+}
+
+void mlx4_crdump_end(struct mlx4_dev *dev)
+{
+	struct mlx4_fw_crdump *crdump = &dev->persist->crdump;
+
+	devlink_region_destroy(crdump->region_fw_health);
+	devlink_region_destroy(crdump->region_crspace);
+}
diff --git a/drivers/net/ethernet/mellanox/mlx4/en_clock.c b/drivers/net/ethernet/mellanox/mlx4/en_clock.c
new file mode 100644
index 000000000..024788549
--- /dev/null
+++ b/drivers/net/ethernet/mellanox/mlx4/en_clock.c
@@ -0,0 +1,303 @@
+/*
+ * Copyright (c) 2012 Mellanox Technologies. All rights reserved.
+ *
+ * This software is available to you under a choice of one of two
+ * licenses.  You may choose to be licensed under the terms of the GNU
+ * General Public License (GPL) Version 2, available from the file
+ * COPYING in the main directory of this source tree, or the
+ * OpenIB.org BSD license below:
+ *
+ *     Redistribution and use in source and binary forms, with or
+ *     without modification, are permitted provided that the following
+ *     conditions are met:
+ *
+ *      - Redistributions of source code must retain the above
+ *        copyright notice, this list of conditions and the following
+ *        disclaimer.
+ *
+ *      - Redistributions in binary form must reproduce the above
+ *        copyright notice, this list of conditions and the following
+ *        disclaimer in the documentation and/or other materials
+ *        provided with the distribution.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
+ * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
+ * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS
+ * BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN
+ * ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
+ * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+ * SOFTWARE.
+ *
+ */
+
+#include <linux/mlx4/device.h>
+#include <linux/clocksource.h>
+
+#include "mlx4_en.h"
+
+/* mlx4_en_read_clock - read raw cycle counter (to be used by time counter)
+ */
+static u64 mlx4_en_read_clock(const struct cyclecounter *tc)
+{
+	struct mlx4_en_dev *mdev =
+		container_of(tc, struct mlx4_en_dev, cycles);
+	struct mlx4_dev *dev = mdev->dev;
+
+	return mlx4_read_clock(dev) & tc->mask;
+}
+
+u64 mlx4_en_get_cqe_ts(struct mlx4_cqe *cqe)
+{
+	u64 hi, lo;
+	struct mlx4_ts_cqe *ts_cqe = (struct mlx4_ts_cqe *)cqe;
+
+	lo = (u64)be16_to_cpu(ts_cqe->timestamp_lo);
+	hi = ((u64)be32_to_cpu(ts_cqe->timestamp_hi) + !lo) << 16;
+
+	return hi | lo;
+}
+
+void mlx4_en_fill_hwtstamps(struct mlx4_en_dev *mdev,
+			    struct skb_shared_hwtstamps *hwts,
+			    u64 timestamp)
+{
+	unsigned int seq;
+	u64 nsec;
+
+	do {
+		seq = read_seqbegin(&mdev->clock_lock);
+		nsec = timecounter_cyc2time(&mdev->clock, timestamp);
+	} while (read_seqretry(&mdev->clock_lock, seq));
+
+	memset(hwts, 0, sizeof(struct skb_shared_hwtstamps));
+	hwts->hwtstamp = ns_to_ktime(nsec);
+}
+
+/**
+ * mlx4_en_remove_timestamp - disable PTP device
+ * @mdev: board private structure
+ *
+ * Stop the PTP support.
+ **/
+void mlx4_en_remove_timestamp(struct mlx4_en_dev *mdev)
+{
+	if (mdev->ptp_clock) {
+		ptp_clock_unregister(mdev->ptp_clock);
+		mdev->ptp_clock = NULL;
+		mlx4_info(mdev, "removed PHC\n");
+	}
+}
+
+#define MLX4_EN_WRAP_AROUND_SEC	10UL
+/* By scheduling the overflow check every 5 seconds, we have a reasonably
+ * good chance we wont miss a wrap around.
+ * TOTO: Use a timer instead of a work queue to increase the guarantee.
+ */
+#define MLX4_EN_OVERFLOW_PERIOD (MLX4_EN_WRAP_AROUND_SEC * HZ / 2)
+
+void mlx4_en_ptp_overflow_check(struct mlx4_en_dev *mdev)
+{
+	bool timeout = time_is_before_jiffies(mdev->last_overflow_check +
+					      MLX4_EN_OVERFLOW_PERIOD);
+	unsigned long flags;
+
+	if (timeout) {
+		write_seqlock_irqsave(&mdev->clock_lock, flags);
+		timecounter_read(&mdev->clock);
+		write_sequnlock_irqrestore(&mdev->clock_lock, flags);
+		mdev->last_overflow_check = jiffies;
+	}
+}
+
+/**
+ * mlx4_en_phc_adjfreq - adjust the frequency of the hardware clock
+ * @ptp: ptp clock structure
+ * @delta: Desired frequency change in parts per billion
+ *
+ * Adjust the frequency of the PHC cycle counter by the indicated delta from
+ * the base frequency.
+ **/
+static int mlx4_en_phc_adjfreq(struct ptp_clock_info *ptp, s32 delta)
+{
+	u64 adj;
+	u32 diff, mult;
+	int neg_adj = 0;
+	unsigned long flags;
+	struct mlx4_en_dev *mdev = container_of(ptp, struct mlx4_en_dev,
+						ptp_clock_info);
+
+	if (delta < 0) {
+		neg_adj = 1;
+		delta = -delta;
+	}
+	mult = mdev->nominal_c_mult;
+	adj = mult;
+	adj *= delta;
+	diff = div_u64(adj, 1000000000ULL);
+
+	write_seqlock_irqsave(&mdev->clock_lock, flags);
+	timecounter_read(&mdev->clock);
+	mdev->cycles.mult = neg_adj ? mult - diff : mult + diff;
+	write_sequnlock_irqrestore(&mdev->clock_lock, flags);
+
+	return 0;
+}
+
+/**
+ * mlx4_en_phc_adjtime - Shift the time of the hardware clock
+ * @ptp: ptp clock structure
+ * @delta: Desired change in nanoseconds
+ *
+ * Adjust the timer by resetting the timecounter structure.
+ **/
+static int mlx4_en_phc_adjtime(struct ptp_clock_info *ptp, s64 delta)
+{
+	struct mlx4_en_dev *mdev = container_of(ptp, struct mlx4_en_dev,
+						ptp_clock_info);
+	unsigned long flags;
+
+	write_seqlock_irqsave(&mdev->clock_lock, flags);
+	timecounter_adjtime(&mdev->clock, delta);
+	write_sequnlock_irqrestore(&mdev->clock_lock, flags);
+
+	return 0;
+}
+
+/**
+ * mlx4_en_phc_gettime - Reads the current time from the hardware clock
+ * @ptp: ptp clock structure
+ * @ts: timespec structure to hold the current time value
+ *
+ * Read the timecounter and return the correct value in ns after converting
+ * it into a struct timespec.
+ **/
+static int mlx4_en_phc_gettime(struct ptp_clock_info *ptp,
+			       struct timespec64 *ts)
+{
+	struct mlx4_en_dev *mdev = container_of(ptp, struct mlx4_en_dev,
+						ptp_clock_info);
+	unsigned long flags;
+	u64 ns;
+
+	write_seqlock_irqsave(&mdev->clock_lock, flags);
+	ns = timecounter_read(&mdev->clock);
+	write_sequnlock_irqrestore(&mdev->clock_lock, flags);
+
+	*ts = ns_to_timespec64(ns);
+
+	return 0;
+}
+
+/**
+ * mlx4_en_phc_settime - Set the current time on the hardware clock
+ * @ptp: ptp clock structure
+ * @ts: timespec containing the new time for the cycle counter
+ *
+ * Reset the timecounter to use a new base value instead of the kernel
+ * wall timer value.
+ **/
+static int mlx4_en_phc_settime(struct ptp_clock_info *ptp,
+			       const struct timespec64 *ts)
+{
+	struct mlx4_en_dev *mdev = container_of(ptp, struct mlx4_en_dev,
+						ptp_clock_info);
+	u64 ns = timespec64_to_ns(ts);
+	unsigned long flags;
+
+	/* reset the timecounter */
+	write_seqlock_irqsave(&mdev->clock_lock, flags);
+	timecounter_init(&mdev->clock, &mdev->cycles, ns);
+	write_sequnlock_irqrestore(&mdev->clock_lock, flags);
+
+	return 0;
+}
+
+/**
+ * mlx4_en_phc_enable - enable or disable an ancillary feature
+ * @ptp: ptp clock structure
+ * @request: Desired resource to enable or disable
+ * @on: Caller passes one to enable or zero to disable
+ *
+ * Enable (or disable) ancillary features of the PHC subsystem.
+ * Currently, no ancillary features are supported.
+ **/
+static int mlx4_en_phc_enable(struct ptp_clock_info __always_unused *ptp,
+			      struct ptp_clock_request __always_unused *request,
+			      int __always_unused on)
+{
+	return -EOPNOTSUPP;
+}
+
+static const struct ptp_clock_info mlx4_en_ptp_clock_info = {
+	.owner		= THIS_MODULE,
+	.max_adj	= 100000000,
+	.n_alarm	= 0,
+	.n_ext_ts	= 0,
+	.n_per_out	= 0,
+	.n_pins		= 0,
+	.pps		= 0,
+	.adjfreq	= mlx4_en_phc_adjfreq,
+	.adjtime	= mlx4_en_phc_adjtime,
+	.gettime64	= mlx4_en_phc_gettime,
+	.settime64	= mlx4_en_phc_settime,
+	.enable		= mlx4_en_phc_enable,
+};
+
+
+/* This function calculates the max shift that enables the user range
+ * of MLX4_EN_WRAP_AROUND_SEC values in the cycles register.
+ */
+static u32 freq_to_shift(u16 freq)
+{
+	u32 freq_khz = freq * 1000;
+	u64 max_val_cycles = freq_khz * 1000 * MLX4_EN_WRAP_AROUND_SEC;
+	u64 max_val_cycles_rounded = 1ULL << fls64(max_val_cycles - 1);
+	/* calculate max possible multiplier in order to fit in 64bit */
+	u64 max_mul = div64_u64(ULLONG_MAX, max_val_cycles_rounded);
+
+	/* This comes from the reverse of clocksource_khz2mult */
+	return ilog2(div_u64(max_mul * freq_khz, 1000000));
+}
+
+void mlx4_en_init_timestamp(struct mlx4_en_dev *mdev)
+{
+	struct mlx4_dev *dev = mdev->dev;
+	unsigned long flags;
+
+	/* mlx4_en_init_timestamp is called for each netdev.
+	 * mdev->ptp_clock is common for all ports, skip initialization if
+	 * was done for other port.
+	 */
+	if (mdev->ptp_clock)
+		return;
+
+	seqlock_init(&mdev->clock_lock);
+
+	memset(&mdev->cycles, 0, sizeof(mdev->cycles));
+	mdev->cycles.read = mlx4_en_read_clock;
+	mdev->cycles.mask = CLOCKSOURCE_MASK(48);
+	mdev->cycles.shift = freq_to_shift(dev->caps.hca_core_clock);
+	mdev->cycles.mult =
+		clocksource_khz2mult(1000 * dev->caps.hca_core_clock, mdev->cycles.shift);
+	mdev->nominal_c_mult = mdev->cycles.mult;
+
+	write_seqlock_irqsave(&mdev->clock_lock, flags);
+	timecounter_init(&mdev->clock, &mdev->cycles,
+			 ktime_to_ns(ktime_get_real()));
+	write_sequnlock_irqrestore(&mdev->clock_lock, flags);
+
+	/* Configure the PHC */
+	mdev->ptp_clock_info = mlx4_en_ptp_clock_info;
+	snprintf(mdev->ptp_clock_info.name, 16, "mlx4 ptp");
+
+	mdev->ptp_clock = ptp_clock_register(&mdev->ptp_clock_info,
+					     &mdev->pdev->dev);
+	if (IS_ERR(mdev->ptp_clock)) {
+		mdev->ptp_clock = NULL;
+		mlx4_err(mdev, "ptp_clock_register failed\n");
+	} else if (mdev->ptp_clock) {
+		mlx4_info(mdev, "registered PHC clock\n");
+	}
+
+}
diff --git a/drivers/net/ethernet/mellanox/mlx4/en_cq.c b/drivers/net/ethernet/mellanox/mlx4/en_cq.c
new file mode 100644
index 000000000..1e487acb4
--- /dev/null
+++ b/drivers/net/ethernet/mellanox/mlx4/en_cq.c
@@ -0,0 +1,218 @@
+/*
+ * Copyright (c) 2007 Mellanox Technologies. All rights reserved.
+ *
+ * This software is available to you under a choice of one of two
+ * licenses.  You may choose to be licensed under the terms of the GNU
+ * General Public License (GPL) Version 2, available from the file
+ * COPYING in the main directory of this source tree, or the
+ * OpenIB.org BSD license below:
+ *
+ *     Redistribution and use in source and binary forms, with or
+ *     without modification, are permitted provided that the following
+ *     conditions are met:
+ *
+ *      - Redistributions of source code must retain the above
+ *        copyright notice, this list of conditions and the following
+ *        disclaimer.
+ *
+ *      - Redistributions in binary form must reproduce the above
+ *        copyright notice, this list of conditions and the following
+ *        disclaimer in the documentation and/or other materials
+ *        provided with the distribution.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
+ * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
+ * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS
+ * BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN
+ * ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
+ * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+ * SOFTWARE.
+ *
+ */
+
+#include <linux/mlx4/cq.h>
+#include <linux/mlx4/qp.h>
+#include <linux/mlx4/cmd.h>
+
+#include "mlx4_en.h"
+
+static void mlx4_en_cq_event(struct mlx4_cq *cq, enum mlx4_event event)
+{
+	return;
+}
+
+
+int mlx4_en_create_cq(struct mlx4_en_priv *priv,
+		      struct mlx4_en_cq **pcq,
+		      int entries, int ring, enum cq_type mode,
+		      int node)
+{
+	struct mlx4_en_dev *mdev = priv->mdev;
+	struct mlx4_en_cq *cq;
+	int err;
+
+	cq = kzalloc_node(sizeof(*cq), GFP_KERNEL, node);
+	if (!cq) {
+		cq = kzalloc(sizeof(*cq), GFP_KERNEL);
+		if (!cq) {
+			en_err(priv, "Failed to allocate CQ structure\n");
+			return -ENOMEM;
+		}
+	}
+
+	cq->size = entries;
+	cq->buf_size = cq->size * mdev->dev->caps.cqe_size;
+
+	cq->ring = ring;
+	cq->type = mode;
+	cq->vector = mdev->dev->caps.num_comp_vectors;
+
+	/* Allocate HW buffers on provided NUMA node.
+	 * dev->numa_node is used in mtt range allocation flow.
+	 */
+	set_dev_node(&mdev->dev->persist->pdev->dev, node);
+	err = mlx4_alloc_hwq_res(mdev->dev, &cq->wqres,
+				cq->buf_size);
+	set_dev_node(&mdev->dev->persist->pdev->dev, mdev->dev->numa_node);
+	if (err)
+		goto err_cq;
+
+	cq->buf = (struct mlx4_cqe *)cq->wqres.buf.direct.buf;
+	*pcq = cq;
+
+	return 0;
+
+err_cq:
+	kfree(cq);
+	*pcq = NULL;
+	return err;
+}
+
+int mlx4_en_activate_cq(struct mlx4_en_priv *priv, struct mlx4_en_cq *cq,
+			int cq_idx)
+{
+	struct mlx4_en_dev *mdev = priv->mdev;
+	int err = 0;
+	int timestamp_en = 0;
+	bool assigned_eq = false;
+
+	cq->dev = mdev->pndev[priv->port];
+	cq->mcq.set_ci_db  = cq->wqres.db.db;
+	cq->mcq.arm_db     = cq->wqres.db.db + 1;
+	*cq->mcq.set_ci_db = 0;
+	*cq->mcq.arm_db    = 0;
+	memset(cq->buf, 0, cq->buf_size);
+
+	if (cq->type == RX) {
+		if (!mlx4_is_eq_vector_valid(mdev->dev, priv->port,
+					     cq->vector)) {
+			cq->vector = cpumask_first(priv->rx_ring[cq->ring]->affinity_mask);
+
+			err = mlx4_assign_eq(mdev->dev, priv->port,
+					     &cq->vector);
+			if (err) {
+				mlx4_err(mdev, "Failed assigning an EQ to CQ vector %d\n",
+					 cq->vector);
+				goto free_eq;
+			}
+
+			assigned_eq = true;
+		}
+
+		cq->irq_desc =
+			irq_to_desc(mlx4_eq_get_irq(mdev->dev,
+						    cq->vector));
+	} else {
+		/* For TX we use the same irq per
+		ring we assigned for the RX    */
+		struct mlx4_en_cq *rx_cq;
+
+		cq_idx = cq_idx % priv->rx_ring_num;
+		rx_cq = priv->rx_cq[cq_idx];
+		cq->vector = rx_cq->vector;
+	}
+
+	if (cq->type == RX)
+		cq->size = priv->rx_ring[cq->ring]->actual_size;
+
+	if ((cq->type != RX && priv->hwtstamp_config.tx_type) ||
+	    (cq->type == RX && priv->hwtstamp_config.rx_filter))
+		timestamp_en = 1;
+
+	cq->mcq.usage = MLX4_RES_USAGE_DRIVER;
+	err = mlx4_cq_alloc(mdev->dev, cq->size, &cq->wqres.mtt,
+			    &mdev->priv_uar, cq->wqres.db.dma, &cq->mcq,
+			    cq->vector, 0, timestamp_en);
+	if (err)
+		goto free_eq;
+
+	cq->mcq.event = mlx4_en_cq_event;
+
+	switch (cq->type) {
+	case TX:
+		cq->mcq.comp = mlx4_en_tx_irq;
+		netif_tx_napi_add(cq->dev, &cq->napi, mlx4_en_poll_tx_cq,
+				  NAPI_POLL_WEIGHT);
+		napi_enable(&cq->napi);
+		break;
+	case RX:
+		cq->mcq.comp = mlx4_en_rx_irq;
+		netif_napi_add(cq->dev, &cq->napi, mlx4_en_poll_rx_cq, 64);
+		napi_enable(&cq->napi);
+		break;
+	case TX_XDP:
+		/* nothing regarding napi, it's shared with rx ring */
+		cq->xdp_busy = false;
+		break;
+	}
+
+	return 0;
+
+free_eq:
+	if (assigned_eq)
+		mlx4_release_eq(mdev->dev, cq->vector);
+	cq->vector = mdev->dev->caps.num_comp_vectors;
+	return err;
+}
+
+void mlx4_en_destroy_cq(struct mlx4_en_priv *priv, struct mlx4_en_cq **pcq)
+{
+	struct mlx4_en_dev *mdev = priv->mdev;
+	struct mlx4_en_cq *cq = *pcq;
+
+	mlx4_free_hwq_res(mdev->dev, &cq->wqres, cq->buf_size);
+	if (mlx4_is_eq_vector_valid(mdev->dev, priv->port, cq->vector) &&
+	    cq->type == RX)
+		mlx4_release_eq(priv->mdev->dev, cq->vector);
+	cq->vector = 0;
+	cq->buf_size = 0;
+	cq->buf = NULL;
+	kfree(cq);
+	*pcq = NULL;
+}
+
+void mlx4_en_deactivate_cq(struct mlx4_en_priv *priv, struct mlx4_en_cq *cq)
+{
+	if (cq->type != TX_XDP) {
+		napi_disable(&cq->napi);
+		netif_napi_del(&cq->napi);
+	}
+
+	mlx4_cq_free(priv->mdev->dev, &cq->mcq);
+}
+
+/* Set rx cq moderation parameters */
+int mlx4_en_set_cq_moder(struct mlx4_en_priv *priv, struct mlx4_en_cq *cq)
+{
+	return mlx4_cq_modify(priv->mdev->dev, &cq->mcq,
+			      cq->moder_cnt, cq->moder_time);
+}
+
+void mlx4_en_arm_cq(struct mlx4_en_priv *priv, struct mlx4_en_cq *cq)
+{
+	mlx4_cq_arm(&cq->mcq, MLX4_CQ_DB_REQ_NOT, priv->mdev->uar_map,
+		    &priv->mdev->uar_lock);
+}
+
+
diff --git a/drivers/net/ethernet/mellanox/mlx4/en_dcb_nl.c b/drivers/net/ethernet/mellanox/mlx4/en_dcb_nl.c
new file mode 100644
index 000000000..752a72499
--- /dev/null
+++ b/drivers/net/ethernet/mellanox/mlx4/en_dcb_nl.c
@@ -0,0 +1,753 @@
+/*
+ * Copyright (c) 2011 Mellanox Technologies. All rights reserved.
+ *
+ * This software is available to you under a choice of one of two
+ * licenses.  You may choose to be licensed under the terms of the GNU
+ * General Public License (GPL) Version 2, available from the file
+ * COPYING in the main directory of this source tree, or the
+ * OpenIB.org BSD license below:
+ *
+ *     Redistribution and use in source and binary forms, with or
+ *     without modification, are permitted provided that the following
+ *     conditions are met:
+ *
+ *      - Redistributions of source code must retain the above
+ *        copyright notice, this list of conditions and the following
+ *        disclaimer.
+ *
+ *      - Redistributions in binary form must reproduce the above
+ *        copyright notice, this list of conditions and the following
+ *        disclaimer in the documentation and/or other materials
+ *        provided with the distribution.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
+ * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
+ * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS
+ * BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN
+ * ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
+ * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+ * SOFTWARE.
+ *
+ */
+
+#include <linux/dcbnl.h>
+#include <linux/math64.h>
+
+#include "mlx4_en.h"
+#include "fw_qos.h"
+
+enum {
+	MLX4_CEE_STATE_DOWN   = 0,
+	MLX4_CEE_STATE_UP     = 1,
+};
+
+/* Definitions for QCN
+ */
+
+struct mlx4_congestion_control_mb_prio_802_1_qau_params {
+	__be32 modify_enable_high;
+	__be32 modify_enable_low;
+	__be32 reserved1;
+	__be32 extended_enable;
+	__be32 rppp_max_rps;
+	__be32 rpg_time_reset;
+	__be32 rpg_byte_reset;
+	__be32 rpg_threshold;
+	__be32 rpg_max_rate;
+	__be32 rpg_ai_rate;
+	__be32 rpg_hai_rate;
+	__be32 rpg_gd;
+	__be32 rpg_min_dec_fac;
+	__be32 rpg_min_rate;
+	__be32 max_time_rise;
+	__be32 max_byte_rise;
+	__be32 max_qdelta;
+	__be32 min_qoffset;
+	__be32 gd_coefficient;
+	__be32 reserved2[5];
+	__be32 cp_sample_base;
+	__be32 reserved3[39];
+};
+
+struct mlx4_congestion_control_mb_prio_802_1_qau_statistics {
+	__be64 rppp_rp_centiseconds;
+	__be32 reserved1;
+	__be32 ignored_cnm;
+	__be32 rppp_created_rps;
+	__be32 estimated_total_rate;
+	__be32 max_active_rate_limiter_index;
+	__be32 dropped_cnms_busy_fw;
+	__be32 reserved2;
+	__be32 cnms_handled_successfully;
+	__be32 min_total_limiters_rate;
+	__be32 max_total_limiters_rate;
+	__be32 reserved3[4];
+};
+
+static u8 mlx4_en_dcbnl_getcap(struct net_device *dev, int capid, u8 *cap)
+{
+	struct mlx4_en_priv *priv = netdev_priv(dev);
+
+	switch (capid) {
+	case DCB_CAP_ATTR_PFC:
+		*cap = true;
+		break;
+	case DCB_CAP_ATTR_DCBX:
+		*cap = priv->dcbx_cap;
+		break;
+	case DCB_CAP_ATTR_PFC_TCS:
+		*cap = 1 <<  mlx4_max_tc(priv->mdev->dev);
+		break;
+	default:
+		*cap = false;
+		break;
+	}
+
+	return 0;
+}
+
+static u8 mlx4_en_dcbnl_getpfcstate(struct net_device *netdev)
+{
+	struct mlx4_en_priv *priv = netdev_priv(netdev);
+
+	return priv->cee_config.pfc_state;
+}
+
+static void mlx4_en_dcbnl_setpfcstate(struct net_device *netdev, u8 state)
+{
+	struct mlx4_en_priv *priv = netdev_priv(netdev);
+
+	priv->cee_config.pfc_state = state;
+}
+
+static void mlx4_en_dcbnl_get_pfc_cfg(struct net_device *netdev, int priority,
+				      u8 *setting)
+{
+	struct mlx4_en_priv *priv = netdev_priv(netdev);
+
+	*setting = priv->cee_config.dcb_pfc[priority];
+}
+
+static void mlx4_en_dcbnl_set_pfc_cfg(struct net_device *netdev, int priority,
+				      u8 setting)
+{
+	struct mlx4_en_priv *priv = netdev_priv(netdev);
+
+	priv->cee_config.dcb_pfc[priority] = setting;
+	priv->cee_config.pfc_state = true;
+}
+
+static int mlx4_en_dcbnl_getnumtcs(struct net_device *netdev, int tcid, u8 *num)
+{
+	struct mlx4_en_priv *priv = netdev_priv(netdev);
+
+	if (!(priv->flags & MLX4_EN_FLAG_DCB_ENABLED))
+		return -EINVAL;
+
+	if (tcid == DCB_NUMTCS_ATTR_PFC)
+		*num = mlx4_max_tc(priv->mdev->dev);
+	else
+		*num = 0;
+
+	return 0;
+}
+
+static u8 mlx4_en_dcbnl_set_all(struct net_device *netdev)
+{
+	struct mlx4_en_priv *priv = netdev_priv(netdev);
+	struct mlx4_en_port_profile *prof = priv->prof;
+	struct mlx4_en_dev *mdev = priv->mdev;
+	u8 tx_pause, tx_ppp, rx_pause, rx_ppp;
+
+	if (!(priv->dcbx_cap & DCB_CAP_DCBX_VER_CEE))
+		return 1;
+
+	if (priv->cee_config.pfc_state) {
+		int tc;
+		rx_ppp = prof->rx_ppp;
+		tx_ppp = prof->tx_ppp;
+
+		for (tc = 0; tc < CEE_DCBX_MAX_PRIO; tc++) {
+			u8 tc_mask = 1 << tc;
+
+			switch (priv->cee_config.dcb_pfc[tc]) {
+			case pfc_disabled:
+				tx_ppp &= ~tc_mask;
+				rx_ppp &= ~tc_mask;
+				break;
+			case pfc_enabled_full:
+				tx_ppp |= tc_mask;
+				rx_ppp |= tc_mask;
+				break;
+			case pfc_enabled_tx:
+				tx_ppp |= tc_mask;
+				rx_ppp &= ~tc_mask;
+				break;
+			case pfc_enabled_rx:
+				tx_ppp &= ~tc_mask;
+				rx_ppp |= tc_mask;
+				break;
+			default:
+				break;
+			}
+		}
+		rx_pause = !!(rx_ppp || tx_ppp) ? 0 : prof->rx_pause;
+		tx_pause = !!(rx_ppp || tx_ppp) ? 0 : prof->tx_pause;
+	} else {
+		rx_ppp = 0;
+		tx_ppp = 0;
+		rx_pause = prof->rx_pause;
+		tx_pause = prof->tx_pause;
+	}
+
+	if (mlx4_SET_PORT_general(mdev->dev, priv->port,
+				  priv->rx_skb_size + ETH_FCS_LEN,
+				  tx_pause, tx_ppp, rx_pause, rx_ppp)) {
+		en_err(priv, "Failed setting pause params\n");
+		return 1;
+	}
+
+	prof->tx_ppp = tx_ppp;
+	prof->rx_ppp = rx_ppp;
+	prof->tx_pause = tx_pause;
+	prof->rx_pause = rx_pause;
+
+	return 0;
+}
+
+static u8 mlx4_en_dcbnl_get_state(struct net_device *dev)
+{
+	struct mlx4_en_priv *priv = netdev_priv(dev);
+
+	if (priv->flags & MLX4_EN_FLAG_DCB_ENABLED)
+		return MLX4_CEE_STATE_UP;
+
+	return MLX4_CEE_STATE_DOWN;
+}
+
+static u8 mlx4_en_dcbnl_set_state(struct net_device *dev, u8 state)
+{
+	struct mlx4_en_priv *priv = netdev_priv(dev);
+	int num_tcs = 0;
+
+	if (!(priv->dcbx_cap & DCB_CAP_DCBX_VER_CEE))
+		return 1;
+
+	if (!!(state) == !!(priv->flags & MLX4_EN_FLAG_DCB_ENABLED))
+		return 0;
+
+	if (state) {
+		priv->flags |= MLX4_EN_FLAG_DCB_ENABLED;
+		num_tcs = IEEE_8021QAZ_MAX_TCS;
+	} else {
+		priv->flags &= ~MLX4_EN_FLAG_DCB_ENABLED;
+	}
+
+	if (mlx4_en_alloc_tx_queue_per_tc(dev, num_tcs))
+		return 1;
+
+	return 0;
+}
+
+/* On success returns a non-zero 802.1p user priority bitmap
+ * otherwise returns 0 as the invalid user priority bitmap to
+ * indicate an error.
+ */
+static int mlx4_en_dcbnl_getapp(struct net_device *netdev, u8 idtype, u16 id)
+{
+	struct mlx4_en_priv *priv = netdev_priv(netdev);
+	struct dcb_app app = {
+				.selector = idtype,
+				.protocol = id,
+			     };
+	if (!(priv->dcbx_cap & DCB_CAP_DCBX_VER_CEE))
+		return 0;
+
+	return dcb_getapp(netdev, &app);
+}
+
+static int mlx4_en_dcbnl_setapp(struct net_device *netdev, u8 idtype,
+				u16 id, u8 up)
+{
+	struct mlx4_en_priv *priv = netdev_priv(netdev);
+	struct dcb_app app;
+
+	if (!(priv->dcbx_cap & DCB_CAP_DCBX_VER_CEE))
+		return -EINVAL;
+
+	memset(&app, 0, sizeof(struct dcb_app));
+	app.selector = idtype;
+	app.protocol = id;
+	app.priority = up;
+
+	return dcb_setapp(netdev, &app);
+}
+
+static int mlx4_en_dcbnl_ieee_getets(struct net_device *dev,
+				   struct ieee_ets *ets)
+{
+	struct mlx4_en_priv *priv = netdev_priv(dev);
+	struct ieee_ets *my_ets = &priv->ets;
+
+	if (!my_ets)
+		return -EINVAL;
+
+	ets->ets_cap = IEEE_8021QAZ_MAX_TCS;
+	ets->cbs = my_ets->cbs;
+	memcpy(ets->tc_tx_bw, my_ets->tc_tx_bw, sizeof(ets->tc_tx_bw));
+	memcpy(ets->tc_tsa, my_ets->tc_tsa, sizeof(ets->tc_tsa));
+	memcpy(ets->prio_tc, my_ets->prio_tc, sizeof(ets->prio_tc));
+
+	return 0;
+}
+
+static int mlx4_en_ets_validate(struct mlx4_en_priv *priv, struct ieee_ets *ets)
+{
+	int i;
+	int total_ets_bw = 0;
+	int has_ets_tc = 0;
+
+	for (i = 0; i < IEEE_8021QAZ_MAX_TCS; i++) {
+		if (ets->prio_tc[i] >= MLX4_EN_NUM_UP_HIGH) {
+			en_err(priv, "Bad priority in UP <=> TC mapping. TC: %d, UP: %d\n",
+					i, ets->prio_tc[i]);
+			return -EINVAL;
+		}
+
+		switch (ets->tc_tsa[i]) {
+		case IEEE_8021QAZ_TSA_VENDOR:
+		case IEEE_8021QAZ_TSA_STRICT:
+			break;
+		case IEEE_8021QAZ_TSA_ETS:
+			has_ets_tc = 1;
+			total_ets_bw += ets->tc_tx_bw[i];
+			break;
+		default:
+			en_err(priv, "TC[%d]: Not supported TSA: %d\n",
+					i, ets->tc_tsa[i]);
+			return -EOPNOTSUPP;
+		}
+	}
+
+	if (has_ets_tc && total_ets_bw != MLX4_EN_BW_MAX) {
+		en_err(priv, "Bad ETS BW sum: %d. Should be exactly 100%%\n",
+				total_ets_bw);
+		return -EINVAL;
+	}
+
+	return 0;
+}
+
+static int mlx4_en_config_port_scheduler(struct mlx4_en_priv *priv,
+		struct ieee_ets *ets, u16 *ratelimit)
+{
+	struct mlx4_en_dev *mdev = priv->mdev;
+	int num_strict = 0;
+	int i;
+	__u8 tc_tx_bw[IEEE_8021QAZ_MAX_TCS] = { 0 };
+	__u8 pg[IEEE_8021QAZ_MAX_TCS] = { 0 };
+
+	ets = ets ?: &priv->ets;
+	ratelimit = ratelimit ?: priv->maxrate;
+
+	/* higher TC means higher priority => lower pg */
+	for (i = IEEE_8021QAZ_MAX_TCS - 1; i >= 0; i--) {
+		switch (ets->tc_tsa[i]) {
+		case IEEE_8021QAZ_TSA_VENDOR:
+			pg[i] = MLX4_EN_TC_VENDOR;
+			tc_tx_bw[i] = MLX4_EN_BW_MAX;
+			break;
+		case IEEE_8021QAZ_TSA_STRICT:
+			pg[i] = num_strict++;
+			tc_tx_bw[i] = MLX4_EN_BW_MAX;
+			break;
+		case IEEE_8021QAZ_TSA_ETS:
+			pg[i] = MLX4_EN_TC_ETS;
+			tc_tx_bw[i] = ets->tc_tx_bw[i] ?: MLX4_EN_BW_MIN;
+			break;
+		}
+	}
+
+	return mlx4_SET_PORT_SCHEDULER(mdev->dev, priv->port, tc_tx_bw, pg,
+			ratelimit);
+}
+
+static int
+mlx4_en_dcbnl_ieee_setets(struct net_device *dev, struct ieee_ets *ets)
+{
+	struct mlx4_en_priv *priv = netdev_priv(dev);
+	struct mlx4_en_dev *mdev = priv->mdev;
+	int err;
+
+	err = mlx4_en_ets_validate(priv, ets);
+	if (err)
+		return err;
+
+	err = mlx4_SET_PORT_PRIO2TC(mdev->dev, priv->port, ets->prio_tc);
+	if (err)
+		return err;
+
+	err = mlx4_en_config_port_scheduler(priv, ets, NULL);
+	if (err)
+		return err;
+
+	memcpy(&priv->ets, ets, sizeof(priv->ets));
+
+	return 0;
+}
+
+static int mlx4_en_dcbnl_ieee_getpfc(struct net_device *dev,
+		struct ieee_pfc *pfc)
+{
+	struct mlx4_en_priv *priv = netdev_priv(dev);
+
+	pfc->pfc_cap = IEEE_8021QAZ_MAX_TCS;
+	pfc->pfc_en = priv->prof->tx_ppp;
+
+	return 0;
+}
+
+static int mlx4_en_dcbnl_ieee_setpfc(struct net_device *dev,
+		struct ieee_pfc *pfc)
+{
+	struct mlx4_en_priv *priv = netdev_priv(dev);
+	struct mlx4_en_port_profile *prof = priv->prof;
+	struct mlx4_en_dev *mdev = priv->mdev;
+	u32 tx_pause, tx_ppp, rx_pause, rx_ppp;
+	int err;
+
+	en_dbg(DRV, priv, "cap: 0x%x en: 0x%x mbc: 0x%x delay: %d\n",
+			pfc->pfc_cap,
+			pfc->pfc_en,
+			pfc->mbc,
+			pfc->delay);
+
+	rx_pause = prof->rx_pause && !pfc->pfc_en;
+	tx_pause = prof->tx_pause && !pfc->pfc_en;
+	rx_ppp = pfc->pfc_en;
+	tx_ppp = pfc->pfc_en;
+
+	err = mlx4_SET_PORT_general(mdev->dev, priv->port,
+				    priv->rx_skb_size + ETH_FCS_LEN,
+				    tx_pause, tx_ppp, rx_pause, rx_ppp);
+	if (err) {
+		en_err(priv, "Failed setting pause params\n");
+		return err;
+	}
+
+	mlx4_en_update_pfc_stats_bitmap(mdev->dev, &priv->stats_bitmap,
+					rx_ppp, rx_pause, tx_ppp, tx_pause);
+
+	prof->tx_ppp = tx_ppp;
+	prof->rx_ppp = rx_ppp;
+	prof->rx_pause = rx_pause;
+	prof->tx_pause = tx_pause;
+
+	return err;
+}
+
+static u8 mlx4_en_dcbnl_getdcbx(struct net_device *dev)
+{
+	struct mlx4_en_priv *priv = netdev_priv(dev);
+
+	return priv->dcbx_cap;
+}
+
+static u8 mlx4_en_dcbnl_setdcbx(struct net_device *dev, u8 mode)
+{
+	struct mlx4_en_priv *priv = netdev_priv(dev);
+	struct ieee_ets ets = {0};
+	struct ieee_pfc pfc = {0};
+
+	if (mode == priv->dcbx_cap)
+		return 0;
+
+	if ((mode & DCB_CAP_DCBX_LLD_MANAGED) ||
+	    ((mode & DCB_CAP_DCBX_VER_IEEE) &&
+	     (mode & DCB_CAP_DCBX_VER_CEE)) ||
+	    !(mode & DCB_CAP_DCBX_HOST))
+		goto err;
+
+	priv->dcbx_cap = mode;
+
+	ets.ets_cap = IEEE_8021QAZ_MAX_TCS;
+	pfc.pfc_cap = IEEE_8021QAZ_MAX_TCS;
+
+	if (mode & DCB_CAP_DCBX_VER_IEEE) {
+		if (mlx4_en_dcbnl_ieee_setets(dev, &ets))
+			goto err;
+		if (mlx4_en_dcbnl_ieee_setpfc(dev, &pfc))
+			goto err;
+	} else if (mode & DCB_CAP_DCBX_VER_CEE) {
+		if (mlx4_en_dcbnl_set_all(dev))
+			goto err;
+	} else {
+		if (mlx4_en_dcbnl_ieee_setets(dev, &ets))
+			goto err;
+		if (mlx4_en_dcbnl_ieee_setpfc(dev, &pfc))
+			goto err;
+		if (mlx4_en_alloc_tx_queue_per_tc(dev, 0))
+			goto err;
+	}
+
+	return 0;
+err:
+	return 1;
+}
+
+#define MLX4_RATELIMIT_UNITS_IN_KB 100000 /* rate-limit HW unit in Kbps */
+static int mlx4_en_dcbnl_ieee_getmaxrate(struct net_device *dev,
+				   struct ieee_maxrate *maxrate)
+{
+	struct mlx4_en_priv *priv = netdev_priv(dev);
+	int i;
+
+	for (i = 0; i < IEEE_8021QAZ_MAX_TCS; i++)
+		maxrate->tc_maxrate[i] =
+			priv->maxrate[i] * MLX4_RATELIMIT_UNITS_IN_KB;
+
+	return 0;
+}
+
+static int mlx4_en_dcbnl_ieee_setmaxrate(struct net_device *dev,
+		struct ieee_maxrate *maxrate)
+{
+	struct mlx4_en_priv *priv = netdev_priv(dev);
+	u16 tmp[IEEE_8021QAZ_MAX_TCS];
+	int i, err;
+
+	for (i = 0; i < IEEE_8021QAZ_MAX_TCS; i++) {
+		/* Convert from Kbps into HW units, rounding result up.
+		 * Setting to 0, means unlimited BW.
+		 */
+		tmp[i] = div_u64(maxrate->tc_maxrate[i] +
+				 MLX4_RATELIMIT_UNITS_IN_KB - 1,
+				 MLX4_RATELIMIT_UNITS_IN_KB);
+	}
+
+	err = mlx4_en_config_port_scheduler(priv, NULL, tmp);
+	if (err)
+		return err;
+
+	memcpy(priv->maxrate, tmp, sizeof(priv->maxrate));
+
+	return 0;
+}
+
+#define RPG_ENABLE_BIT	31
+#define CN_TAG_BIT	30
+
+static int mlx4_en_dcbnl_ieee_getqcn(struct net_device *dev,
+				     struct ieee_qcn *qcn)
+{
+	struct mlx4_en_priv *priv = netdev_priv(dev);
+	struct mlx4_congestion_control_mb_prio_802_1_qau_params *hw_qcn;
+	struct mlx4_cmd_mailbox *mailbox_out = NULL;
+	u64 mailbox_in_dma = 0;
+	u32 inmod = 0;
+	int i, err;
+
+	if (!(priv->mdev->dev->caps.flags2 & MLX4_DEV_CAP_FLAG2_QCN))
+		return -EOPNOTSUPP;
+
+	mailbox_out = mlx4_alloc_cmd_mailbox(priv->mdev->dev);
+	if (IS_ERR(mailbox_out))
+		return -ENOMEM;
+	hw_qcn =
+	(struct mlx4_congestion_control_mb_prio_802_1_qau_params *)
+	mailbox_out->buf;
+
+	for (i = 0; i < IEEE_8021QAZ_MAX_TCS; i++) {
+		inmod = priv->port | ((1 << i) << 8) |
+			 (MLX4_CTRL_ALGO_802_1_QAU_REACTION_POINT << 16);
+		err = mlx4_cmd_box(priv->mdev->dev, mailbox_in_dma,
+				   mailbox_out->dma,
+				   inmod, MLX4_CONGESTION_CONTROL_GET_PARAMS,
+				   MLX4_CMD_CONGESTION_CTRL_OPCODE,
+				   MLX4_CMD_TIME_CLASS_C,
+				   MLX4_CMD_NATIVE);
+		if (err) {
+			mlx4_free_cmd_mailbox(priv->mdev->dev, mailbox_out);
+			return err;
+		}
+
+		qcn->rpg_enable[i] =
+			be32_to_cpu(hw_qcn->extended_enable) >> RPG_ENABLE_BIT;
+		qcn->rppp_max_rps[i] =
+			be32_to_cpu(hw_qcn->rppp_max_rps);
+		qcn->rpg_time_reset[i] =
+			be32_to_cpu(hw_qcn->rpg_time_reset);
+		qcn->rpg_byte_reset[i] =
+			be32_to_cpu(hw_qcn->rpg_byte_reset);
+		qcn->rpg_threshold[i] =
+			be32_to_cpu(hw_qcn->rpg_threshold);
+		qcn->rpg_max_rate[i] =
+			be32_to_cpu(hw_qcn->rpg_max_rate);
+		qcn->rpg_ai_rate[i] =
+			be32_to_cpu(hw_qcn->rpg_ai_rate);
+		qcn->rpg_hai_rate[i] =
+			be32_to_cpu(hw_qcn->rpg_hai_rate);
+		qcn->rpg_gd[i] =
+			be32_to_cpu(hw_qcn->rpg_gd);
+		qcn->rpg_min_dec_fac[i] =
+			be32_to_cpu(hw_qcn->rpg_min_dec_fac);
+		qcn->rpg_min_rate[i] =
+			be32_to_cpu(hw_qcn->rpg_min_rate);
+		qcn->cndd_state_machine[i] =
+			priv->cndd_state[i];
+	}
+	mlx4_free_cmd_mailbox(priv->mdev->dev, mailbox_out);
+	return 0;
+}
+
+static int mlx4_en_dcbnl_ieee_setqcn(struct net_device *dev,
+				     struct ieee_qcn *qcn)
+{
+	struct mlx4_en_priv *priv = netdev_priv(dev);
+	struct mlx4_congestion_control_mb_prio_802_1_qau_params *hw_qcn;
+	struct mlx4_cmd_mailbox *mailbox_in = NULL;
+	u64 mailbox_in_dma = 0;
+	u32 inmod = 0;
+	int i, err;
+#define MODIFY_ENABLE_HIGH_MASK 0xc0000000
+#define MODIFY_ENABLE_LOW_MASK 0xffc00000
+
+	if (!(priv->mdev->dev->caps.flags2 & MLX4_DEV_CAP_FLAG2_QCN))
+		return -EOPNOTSUPP;
+
+	mailbox_in = mlx4_alloc_cmd_mailbox(priv->mdev->dev);
+	if (IS_ERR(mailbox_in))
+		return -ENOMEM;
+
+	mailbox_in_dma = mailbox_in->dma;
+	hw_qcn =
+	(struct mlx4_congestion_control_mb_prio_802_1_qau_params *)mailbox_in->buf;
+	for (i = 0; i < IEEE_8021QAZ_MAX_TCS; i++) {
+		inmod = priv->port | ((1 << i) << 8) |
+			 (MLX4_CTRL_ALGO_802_1_QAU_REACTION_POINT << 16);
+
+		/* Before updating QCN parameter,
+		 * need to set it's modify enable bit to 1
+		 */
+
+		hw_qcn->modify_enable_high = cpu_to_be32(
+						MODIFY_ENABLE_HIGH_MASK);
+		hw_qcn->modify_enable_low = cpu_to_be32(MODIFY_ENABLE_LOW_MASK);
+
+		hw_qcn->extended_enable = cpu_to_be32(qcn->rpg_enable[i] << RPG_ENABLE_BIT);
+		hw_qcn->rppp_max_rps = cpu_to_be32(qcn->rppp_max_rps[i]);
+		hw_qcn->rpg_time_reset = cpu_to_be32(qcn->rpg_time_reset[i]);
+		hw_qcn->rpg_byte_reset = cpu_to_be32(qcn->rpg_byte_reset[i]);
+		hw_qcn->rpg_threshold = cpu_to_be32(qcn->rpg_threshold[i]);
+		hw_qcn->rpg_max_rate = cpu_to_be32(qcn->rpg_max_rate[i]);
+		hw_qcn->rpg_ai_rate = cpu_to_be32(qcn->rpg_ai_rate[i]);
+		hw_qcn->rpg_hai_rate = cpu_to_be32(qcn->rpg_hai_rate[i]);
+		hw_qcn->rpg_gd = cpu_to_be32(qcn->rpg_gd[i]);
+		hw_qcn->rpg_min_dec_fac = cpu_to_be32(qcn->rpg_min_dec_fac[i]);
+		hw_qcn->rpg_min_rate = cpu_to_be32(qcn->rpg_min_rate[i]);
+		priv->cndd_state[i] = qcn->cndd_state_machine[i];
+		if (qcn->cndd_state_machine[i] == DCB_CNDD_INTERIOR_READY)
+			hw_qcn->extended_enable |= cpu_to_be32(1 << CN_TAG_BIT);
+
+		err = mlx4_cmd(priv->mdev->dev, mailbox_in_dma, inmod,
+			       MLX4_CONGESTION_CONTROL_SET_PARAMS,
+			       MLX4_CMD_CONGESTION_CTRL_OPCODE,
+			       MLX4_CMD_TIME_CLASS_C,
+			       MLX4_CMD_NATIVE);
+		if (err) {
+			mlx4_free_cmd_mailbox(priv->mdev->dev, mailbox_in);
+			return err;
+		}
+	}
+	mlx4_free_cmd_mailbox(priv->mdev->dev, mailbox_in);
+	return 0;
+}
+
+static int mlx4_en_dcbnl_ieee_getqcnstats(struct net_device *dev,
+					  struct ieee_qcn_stats *qcn_stats)
+{
+	struct mlx4_en_priv *priv = netdev_priv(dev);
+	struct mlx4_congestion_control_mb_prio_802_1_qau_statistics *hw_qcn_stats;
+	struct mlx4_cmd_mailbox *mailbox_out = NULL;
+	u64 mailbox_in_dma = 0;
+	u32 inmod = 0;
+	int i, err;
+
+	if (!(priv->mdev->dev->caps.flags2 & MLX4_DEV_CAP_FLAG2_QCN))
+		return -EOPNOTSUPP;
+
+	mailbox_out = mlx4_alloc_cmd_mailbox(priv->mdev->dev);
+	if (IS_ERR(mailbox_out))
+		return -ENOMEM;
+
+	hw_qcn_stats =
+	(struct mlx4_congestion_control_mb_prio_802_1_qau_statistics *)
+	mailbox_out->buf;
+
+	for (i = 0; i < IEEE_8021QAZ_MAX_TCS; i++) {
+		inmod = priv->port | ((1 << i) << 8) |
+			 (MLX4_CTRL_ALGO_802_1_QAU_REACTION_POINT << 16);
+		err = mlx4_cmd_box(priv->mdev->dev, mailbox_in_dma,
+				   mailbox_out->dma, inmod,
+				   MLX4_CONGESTION_CONTROL_GET_STATISTICS,
+				   MLX4_CMD_CONGESTION_CTRL_OPCODE,
+				   MLX4_CMD_TIME_CLASS_C,
+				   MLX4_CMD_NATIVE);
+		if (err) {
+			mlx4_free_cmd_mailbox(priv->mdev->dev, mailbox_out);
+			return err;
+		}
+		qcn_stats->rppp_rp_centiseconds[i] =
+			be64_to_cpu(hw_qcn_stats->rppp_rp_centiseconds);
+		qcn_stats->rppp_created_rps[i] =
+			be32_to_cpu(hw_qcn_stats->rppp_created_rps);
+	}
+	mlx4_free_cmd_mailbox(priv->mdev->dev, mailbox_out);
+	return 0;
+}
+
+const struct dcbnl_rtnl_ops mlx4_en_dcbnl_ops = {
+	.ieee_getets		= mlx4_en_dcbnl_ieee_getets,
+	.ieee_setets		= mlx4_en_dcbnl_ieee_setets,
+	.ieee_getmaxrate	= mlx4_en_dcbnl_ieee_getmaxrate,
+	.ieee_setmaxrate	= mlx4_en_dcbnl_ieee_setmaxrate,
+	.ieee_getqcn		= mlx4_en_dcbnl_ieee_getqcn,
+	.ieee_setqcn		= mlx4_en_dcbnl_ieee_setqcn,
+	.ieee_getqcnstats	= mlx4_en_dcbnl_ieee_getqcnstats,
+	.ieee_getpfc		= mlx4_en_dcbnl_ieee_getpfc,
+	.ieee_setpfc		= mlx4_en_dcbnl_ieee_setpfc,
+
+	.getstate	= mlx4_en_dcbnl_get_state,
+	.setstate	= mlx4_en_dcbnl_set_state,
+	.getpfccfg	= mlx4_en_dcbnl_get_pfc_cfg,
+	.setpfccfg	= mlx4_en_dcbnl_set_pfc_cfg,
+	.setall		= mlx4_en_dcbnl_set_all,
+	.getcap		= mlx4_en_dcbnl_getcap,
+	.getnumtcs	= mlx4_en_dcbnl_getnumtcs,
+	.getpfcstate	= mlx4_en_dcbnl_getpfcstate,
+	.setpfcstate	= mlx4_en_dcbnl_setpfcstate,
+	.getapp		= mlx4_en_dcbnl_getapp,
+	.setapp		= mlx4_en_dcbnl_setapp,
+
+	.getdcbx	= mlx4_en_dcbnl_getdcbx,
+	.setdcbx	= mlx4_en_dcbnl_setdcbx,
+};
+
+const struct dcbnl_rtnl_ops mlx4_en_dcbnl_pfc_ops = {
+	.ieee_getpfc	= mlx4_en_dcbnl_ieee_getpfc,
+	.ieee_setpfc	= mlx4_en_dcbnl_ieee_setpfc,
+
+	.setstate	= mlx4_en_dcbnl_set_state,
+	.getpfccfg	= mlx4_en_dcbnl_get_pfc_cfg,
+	.setpfccfg	= mlx4_en_dcbnl_set_pfc_cfg,
+	.setall		= mlx4_en_dcbnl_set_all,
+	.getnumtcs	= mlx4_en_dcbnl_getnumtcs,
+	.getpfcstate	= mlx4_en_dcbnl_getpfcstate,
+	.setpfcstate	= mlx4_en_dcbnl_setpfcstate,
+	.getapp		= mlx4_en_dcbnl_getapp,
+	.setapp		= mlx4_en_dcbnl_setapp,
+
+	.getdcbx	= mlx4_en_dcbnl_getdcbx,
+	.setdcbx	= mlx4_en_dcbnl_setdcbx,
+};
diff --git a/drivers/net/ethernet/mellanox/mlx4/en_ethtool.c b/drivers/net/ethernet/mellanox/mlx4/en_ethtool.c
new file mode 100644
index 000000000..1d33fae52
--- /dev/null
+++ b/drivers/net/ethernet/mellanox/mlx4/en_ethtool.c
@@ -0,0 +1,2161 @@
+/*
+ * Copyright (c) 2007 Mellanox Technologies. All rights reserved.
+ *
+ * This software is available to you under a choice of one of two
+ * licenses.  You may choose to be licensed under the terms of the GNU
+ * General Public License (GPL) Version 2, available from the file
+ * COPYING in the main directory of this source tree, or the
+ * OpenIB.org BSD license below:
+ *
+ *     Redistribution and use in source and binary forms, with or
+ *     without modification, are permitted provided that the following
+ *     conditions are met:
+ *
+ *      - Redistributions of source code must retain the above
+ *        copyright notice, this list of conditions and the following
+ *        disclaimer.
+ *
+ *      - Redistributions in binary form must reproduce the above
+ *        copyright notice, this list of conditions and the following
+ *        disclaimer in the documentation and/or other materials
+ *        provided with the distribution.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
+ * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
+ * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS
+ * BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN
+ * ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
+ * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+ * SOFTWARE.
+ *
+ */
+
+#include <linux/kernel.h>
+#include <linux/ethtool.h>
+#include <linux/netdevice.h>
+#include <linux/mlx4/driver.h>
+#include <linux/mlx4/device.h>
+#include <linux/in.h>
+#include <net/ip.h>
+#include <linux/bitmap.h>
+
+#include "mlx4_en.h"
+#include "en_port.h"
+
+#define EN_ETHTOOL_QP_ATTACH (1ull << 63)
+#define EN_ETHTOOL_SHORT_MASK cpu_to_be16(0xffff)
+#define EN_ETHTOOL_WORD_MASK  cpu_to_be32(0xffffffff)
+
+int mlx4_en_moderation_update(struct mlx4_en_priv *priv)
+{
+	int i, t;
+	int err = 0;
+
+	for (t = 0 ; t < MLX4_EN_NUM_TX_TYPES; t++) {
+		for (i = 0; i < priv->tx_ring_num[t]; i++) {
+			priv->tx_cq[t][i]->moder_cnt = priv->tx_frames;
+			priv->tx_cq[t][i]->moder_time = priv->tx_usecs;
+			if (priv->port_up) {
+				err = mlx4_en_set_cq_moder(priv,
+							   priv->tx_cq[t][i]);
+				if (err)
+					return err;
+			}
+		}
+	}
+
+	if (priv->adaptive_rx_coal)
+		return 0;
+
+	for (i = 0; i < priv->rx_ring_num; i++) {
+		priv->rx_cq[i]->moder_cnt = priv->rx_frames;
+		priv->rx_cq[i]->moder_time = priv->rx_usecs;
+		priv->last_moder_time[i] = MLX4_EN_AUTO_CONF;
+		if (priv->port_up) {
+			err = mlx4_en_set_cq_moder(priv, priv->rx_cq[i]);
+			if (err)
+				return err;
+		}
+	}
+
+	return err;
+}
+
+static void
+mlx4_en_get_drvinfo(struct net_device *dev, struct ethtool_drvinfo *drvinfo)
+{
+	struct mlx4_en_priv *priv = netdev_priv(dev);
+	struct mlx4_en_dev *mdev = priv->mdev;
+
+	strlcpy(drvinfo->driver, DRV_NAME, sizeof(drvinfo->driver));
+	strlcpy(drvinfo->version, DRV_VERSION,
+		sizeof(drvinfo->version));
+	snprintf(drvinfo->fw_version, sizeof(drvinfo->fw_version),
+		"%d.%d.%d",
+		(u16) (mdev->dev->caps.fw_ver >> 32),
+		(u16) ((mdev->dev->caps.fw_ver >> 16) & 0xffff),
+		(u16) (mdev->dev->caps.fw_ver & 0xffff));
+	strlcpy(drvinfo->bus_info, pci_name(mdev->dev->persist->pdev),
+		sizeof(drvinfo->bus_info));
+}
+
+static const char mlx4_en_priv_flags[][ETH_GSTRING_LEN] = {
+	"blueflame",
+	"phv-bit"
+};
+
+static const char main_strings[][ETH_GSTRING_LEN] = {
+	/* main statistics */
+	"rx_packets", "tx_packets", "rx_bytes", "tx_bytes", "rx_errors",
+	"tx_errors", "rx_dropped", "tx_dropped", "multicast", "collisions",
+	"rx_length_errors", "rx_over_errors", "rx_crc_errors",
+	"rx_frame_errors", "rx_fifo_errors", "rx_missed_errors",
+	"tx_aborted_errors", "tx_carrier_errors", "tx_fifo_errors",
+	"tx_heartbeat_errors", "tx_window_errors",
+
+	/* port statistics */
+	"tso_packets",
+	"xmit_more",
+	"queue_stopped", "wake_queue", "tx_timeout", "rx_alloc_pages",
+	"rx_csum_good", "rx_csum_none", "rx_csum_complete", "tx_chksum_offload",
+
+	/* pf statistics */
+	"pf_rx_packets",
+	"pf_rx_bytes",
+	"pf_tx_packets",
+	"pf_tx_bytes",
+
+	/* priority flow control statistics rx */
+	"rx_pause_prio_0", "rx_pause_duration_prio_0",
+	"rx_pause_transition_prio_0",
+	"rx_pause_prio_1", "rx_pause_duration_prio_1",
+	"rx_pause_transition_prio_1",
+	"rx_pause_prio_2", "rx_pause_duration_prio_2",
+	"rx_pause_transition_prio_2",
+	"rx_pause_prio_3", "rx_pause_duration_prio_3",
+	"rx_pause_transition_prio_3",
+	"rx_pause_prio_4", "rx_pause_duration_prio_4",
+	"rx_pause_transition_prio_4",
+	"rx_pause_prio_5", "rx_pause_duration_prio_5",
+	"rx_pause_transition_prio_5",
+	"rx_pause_prio_6", "rx_pause_duration_prio_6",
+	"rx_pause_transition_prio_6",
+	"rx_pause_prio_7", "rx_pause_duration_prio_7",
+	"rx_pause_transition_prio_7",
+
+	/* flow control statistics rx */
+	"rx_pause", "rx_pause_duration", "rx_pause_transition",
+
+	/* priority flow control statistics tx */
+	"tx_pause_prio_0", "tx_pause_duration_prio_0",
+	"tx_pause_transition_prio_0",
+	"tx_pause_prio_1", "tx_pause_duration_prio_1",
+	"tx_pause_transition_prio_1",
+	"tx_pause_prio_2", "tx_pause_duration_prio_2",
+	"tx_pause_transition_prio_2",
+	"tx_pause_prio_3", "tx_pause_duration_prio_3",
+	"tx_pause_transition_prio_3",
+	"tx_pause_prio_4", "tx_pause_duration_prio_4",
+	"tx_pause_transition_prio_4",
+	"tx_pause_prio_5", "tx_pause_duration_prio_5",
+	"tx_pause_transition_prio_5",
+	"tx_pause_prio_6", "tx_pause_duration_prio_6",
+	"tx_pause_transition_prio_6",
+	"tx_pause_prio_7", "tx_pause_duration_prio_7",
+	"tx_pause_transition_prio_7",
+
+	/* flow control statistics tx */
+	"tx_pause", "tx_pause_duration", "tx_pause_transition",
+
+	/* packet statistics */
+	"rx_multicast_packets",
+	"rx_broadcast_packets",
+	"rx_jabbers",
+	"rx_in_range_length_error",
+	"rx_out_range_length_error",
+	"tx_multicast_packets",
+	"tx_broadcast_packets",
+	"rx_prio_0_packets", "rx_prio_0_bytes",
+	"rx_prio_1_packets", "rx_prio_1_bytes",
+	"rx_prio_2_packets", "rx_prio_2_bytes",
+	"rx_prio_3_packets", "rx_prio_3_bytes",
+	"rx_prio_4_packets", "rx_prio_4_bytes",
+	"rx_prio_5_packets", "rx_prio_5_bytes",
+	"rx_prio_6_packets", "rx_prio_6_bytes",
+	"rx_prio_7_packets", "rx_prio_7_bytes",
+	"rx_novlan_packets", "rx_novlan_bytes",
+	"tx_prio_0_packets", "tx_prio_0_bytes",
+	"tx_prio_1_packets", "tx_prio_1_bytes",
+	"tx_prio_2_packets", "tx_prio_2_bytes",
+	"tx_prio_3_packets", "tx_prio_3_bytes",
+	"tx_prio_4_packets", "tx_prio_4_bytes",
+	"tx_prio_5_packets", "tx_prio_5_bytes",
+	"tx_prio_6_packets", "tx_prio_6_bytes",
+	"tx_prio_7_packets", "tx_prio_7_bytes",
+	"tx_novlan_packets", "tx_novlan_bytes",
+
+	/* xdp statistics */
+	"rx_xdp_drop",
+	"rx_xdp_tx",
+	"rx_xdp_tx_full",
+
+	/* phy statistics */
+	"rx_packets_phy", "rx_bytes_phy",
+	"tx_packets_phy", "tx_bytes_phy",
+};
+
+static const char mlx4_en_test_names[][ETH_GSTRING_LEN]= {
+	"Interrupt Test",
+	"Link Test",
+	"Speed Test",
+	"Register Test",
+	"Loopback Test",
+};
+
+static u32 mlx4_en_get_msglevel(struct net_device *dev)
+{
+	return ((struct mlx4_en_priv *) netdev_priv(dev))->msg_enable;
+}
+
+static void mlx4_en_set_msglevel(struct net_device *dev, u32 val)
+{
+	((struct mlx4_en_priv *) netdev_priv(dev))->msg_enable = val;
+}
+
+static void mlx4_en_get_wol(struct net_device *netdev,
+			    struct ethtool_wolinfo *wol)
+{
+	struct mlx4_en_priv *priv = netdev_priv(netdev);
+	struct mlx4_caps *caps = &priv->mdev->dev->caps;
+	int err = 0;
+	u64 config = 0;
+	u64 mask;
+
+	if ((priv->port < 1) || (priv->port > 2)) {
+		en_err(priv, "Failed to get WoL information\n");
+		return;
+	}
+
+	mask = (priv->port == 1) ? MLX4_DEV_CAP_FLAG_WOL_PORT1 :
+		MLX4_DEV_CAP_FLAG_WOL_PORT2;
+
+	if (!(caps->flags & mask)) {
+		wol->supported = 0;
+		wol->wolopts = 0;
+		return;
+	}
+
+	if (caps->wol_port[priv->port])
+		wol->supported = WAKE_MAGIC;
+	else
+		wol->supported = 0;
+
+	err = mlx4_wol_read(priv->mdev->dev, &config, priv->port);
+	if (err) {
+		en_err(priv, "Failed to get WoL information\n");
+		return;
+	}
+
+	if ((config & MLX4_EN_WOL_ENABLED) && (config & MLX4_EN_WOL_MAGIC))
+		wol->wolopts = WAKE_MAGIC;
+	else
+		wol->wolopts = 0;
+}
+
+static int mlx4_en_set_wol(struct net_device *netdev,
+			    struct ethtool_wolinfo *wol)
+{
+	struct mlx4_en_priv *priv = netdev_priv(netdev);
+	u64 config = 0;
+	int err = 0;
+	u64 mask;
+
+	if ((priv->port < 1) || (priv->port > 2))
+		return -EOPNOTSUPP;
+
+	mask = (priv->port == 1) ? MLX4_DEV_CAP_FLAG_WOL_PORT1 :
+		MLX4_DEV_CAP_FLAG_WOL_PORT2;
+
+	if (!(priv->mdev->dev->caps.flags & mask))
+		return -EOPNOTSUPP;
+
+	if (wol->supported & ~WAKE_MAGIC)
+		return -EINVAL;
+
+	err = mlx4_wol_read(priv->mdev->dev, &config, priv->port);
+	if (err) {
+		en_err(priv, "Failed to get WoL info, unable to modify\n");
+		return err;
+	}
+
+	if (wol->wolopts & WAKE_MAGIC) {
+		config |= MLX4_EN_WOL_DO_MODIFY | MLX4_EN_WOL_ENABLED |
+				MLX4_EN_WOL_MAGIC;
+	} else {
+		config &= ~(MLX4_EN_WOL_ENABLED | MLX4_EN_WOL_MAGIC);
+		config |= MLX4_EN_WOL_DO_MODIFY;
+	}
+
+	err = mlx4_wol_write(priv->mdev->dev, config, priv->port);
+	if (err)
+		en_err(priv, "Failed to set WoL information\n");
+
+	return err;
+}
+
+struct bitmap_iterator {
+	unsigned long *stats_bitmap;
+	unsigned int count;
+	unsigned int iterator;
+	bool advance_array; /* if set, force no increments */
+};
+
+static inline void bitmap_iterator_init(struct bitmap_iterator *h,
+					unsigned long *stats_bitmap,
+					int count)
+{
+	h->iterator = 0;
+	h->advance_array = !bitmap_empty(stats_bitmap, count);
+	h->count = h->advance_array ? bitmap_weight(stats_bitmap, count)
+		: count;
+	h->stats_bitmap = stats_bitmap;
+}
+
+static inline int bitmap_iterator_test(struct bitmap_iterator *h)
+{
+	return !h->advance_array ? 1 : test_bit(h->iterator, h->stats_bitmap);
+}
+
+static inline int bitmap_iterator_inc(struct bitmap_iterator *h)
+{
+	return h->iterator++;
+}
+
+static inline unsigned int
+bitmap_iterator_count(struct bitmap_iterator *h)
+{
+	return h->count;
+}
+
+static int mlx4_en_get_sset_count(struct net_device *dev, int sset)
+{
+	struct mlx4_en_priv *priv = netdev_priv(dev);
+	struct bitmap_iterator it;
+
+	bitmap_iterator_init(&it, priv->stats_bitmap.bitmap, NUM_ALL_STATS);
+
+	switch (sset) {
+	case ETH_SS_STATS:
+		return bitmap_iterator_count(&it) +
+			(priv->tx_ring_num[TX] * 2) +
+			(priv->rx_ring_num * (3 + NUM_XDP_STATS));
+	case ETH_SS_TEST:
+		return MLX4_EN_NUM_SELF_TEST - !(priv->mdev->dev->caps.flags
+					& MLX4_DEV_CAP_FLAG_UC_LOOPBACK) * 2;
+	case ETH_SS_PRIV_FLAGS:
+		return ARRAY_SIZE(mlx4_en_priv_flags);
+	default:
+		return -EOPNOTSUPP;
+	}
+}
+
+static void mlx4_en_get_ethtool_stats(struct net_device *dev,
+		struct ethtool_stats *stats, uint64_t *data)
+{
+	struct mlx4_en_priv *priv = netdev_priv(dev);
+	int index = 0;
+	int i;
+	struct bitmap_iterator it;
+
+	bitmap_iterator_init(&it, priv->stats_bitmap.bitmap, NUM_ALL_STATS);
+
+	spin_lock_bh(&priv->stats_lock);
+
+	mlx4_en_fold_software_stats(dev);
+
+	for (i = 0; i < NUM_MAIN_STATS; i++, bitmap_iterator_inc(&it))
+		if (bitmap_iterator_test(&it))
+			data[index++] = ((unsigned long *)&dev->stats)[i];
+
+	for (i = 0; i < NUM_PORT_STATS; i++, bitmap_iterator_inc(&it))
+		if (bitmap_iterator_test(&it))
+			data[index++] = ((unsigned long *)&priv->port_stats)[i];
+
+	for (i = 0; i < NUM_PF_STATS; i++, bitmap_iterator_inc(&it))
+		if (bitmap_iterator_test(&it))
+			data[index++] =
+				((unsigned long *)&priv->pf_stats)[i];
+
+	for (i = 0; i < NUM_FLOW_PRIORITY_STATS_RX;
+	     i++, bitmap_iterator_inc(&it))
+		if (bitmap_iterator_test(&it))
+			data[index++] =
+				((u64 *)&priv->rx_priority_flowstats)[i];
+
+	for (i = 0; i < NUM_FLOW_STATS_RX; i++, bitmap_iterator_inc(&it))
+		if (bitmap_iterator_test(&it))
+			data[index++] = ((u64 *)&priv->rx_flowstats)[i];
+
+	for (i = 0; i < NUM_FLOW_PRIORITY_STATS_TX;
+	     i++, bitmap_iterator_inc(&it))
+		if (bitmap_iterator_test(&it))
+			data[index++] =
+				((u64 *)&priv->tx_priority_flowstats)[i];
+
+	for (i = 0; i < NUM_FLOW_STATS_TX; i++, bitmap_iterator_inc(&it))
+		if (bitmap_iterator_test(&it))
+			data[index++] = ((u64 *)&priv->tx_flowstats)[i];
+
+	for (i = 0; i < NUM_PKT_STATS; i++, bitmap_iterator_inc(&it))
+		if (bitmap_iterator_test(&it))
+			data[index++] = ((unsigned long *)&priv->pkstats)[i];
+
+	for (i = 0; i < NUM_XDP_STATS; i++, bitmap_iterator_inc(&it))
+		if (bitmap_iterator_test(&it))
+			data[index++] = ((unsigned long *)&priv->xdp_stats)[i];
+
+	for (i = 0; i < NUM_PHY_STATS; i++, bitmap_iterator_inc(&it))
+		if (bitmap_iterator_test(&it))
+			data[index++] = ((unsigned long *)&priv->phy_stats)[i];
+
+	for (i = 0; i < priv->tx_ring_num[TX]; i++) {
+		data[index++] = priv->tx_ring[TX][i]->packets;
+		data[index++] = priv->tx_ring[TX][i]->bytes;
+	}
+	for (i = 0; i < priv->rx_ring_num; i++) {
+		data[index++] = priv->rx_ring[i]->packets;
+		data[index++] = priv->rx_ring[i]->bytes;
+		data[index++] = priv->rx_ring[i]->dropped;
+		data[index++] = priv->rx_ring[i]->xdp_drop;
+		data[index++] = priv->rx_ring[i]->xdp_tx;
+		data[index++] = priv->rx_ring[i]->xdp_tx_full;
+	}
+	spin_unlock_bh(&priv->stats_lock);
+
+}
+
+static void mlx4_en_self_test(struct net_device *dev,
+			      struct ethtool_test *etest, u64 *buf)
+{
+	mlx4_en_ex_selftest(dev, &etest->flags, buf);
+}
+
+static void mlx4_en_get_strings(struct net_device *dev,
+				uint32_t stringset, uint8_t *data)
+{
+	struct mlx4_en_priv *priv = netdev_priv(dev);
+	int index = 0;
+	int i, strings = 0;
+	struct bitmap_iterator it;
+
+	bitmap_iterator_init(&it, priv->stats_bitmap.bitmap, NUM_ALL_STATS);
+
+	switch (stringset) {
+	case ETH_SS_TEST:
+		for (i = 0; i < MLX4_EN_NUM_SELF_TEST - 2; i++)
+			strcpy(data + i * ETH_GSTRING_LEN, mlx4_en_test_names[i]);
+		if (priv->mdev->dev->caps.flags & MLX4_DEV_CAP_FLAG_UC_LOOPBACK)
+			for (; i < MLX4_EN_NUM_SELF_TEST; i++)
+				strcpy(data + i * ETH_GSTRING_LEN, mlx4_en_test_names[i]);
+		break;
+
+	case ETH_SS_STATS:
+		/* Add main counters */
+		for (i = 0; i < NUM_MAIN_STATS; i++, strings++,
+		     bitmap_iterator_inc(&it))
+			if (bitmap_iterator_test(&it))
+				strcpy(data + (index++) * ETH_GSTRING_LEN,
+				       main_strings[strings]);
+
+		for (i = 0; i < NUM_PORT_STATS; i++, strings++,
+		     bitmap_iterator_inc(&it))
+			if (bitmap_iterator_test(&it))
+				strcpy(data + (index++) * ETH_GSTRING_LEN,
+				       main_strings[strings]);
+
+		for (i = 0; i < NUM_PF_STATS; i++, strings++,
+		     bitmap_iterator_inc(&it))
+			if (bitmap_iterator_test(&it))
+				strcpy(data + (index++) * ETH_GSTRING_LEN,
+				       main_strings[strings]);
+
+		for (i = 0; i < NUM_FLOW_STATS; i++, strings++,
+		     bitmap_iterator_inc(&it))
+			if (bitmap_iterator_test(&it))
+				strcpy(data + (index++) * ETH_GSTRING_LEN,
+				       main_strings[strings]);
+
+		for (i = 0; i < NUM_PKT_STATS; i++, strings++,
+		     bitmap_iterator_inc(&it))
+			if (bitmap_iterator_test(&it))
+				strcpy(data + (index++) * ETH_GSTRING_LEN,
+				       main_strings[strings]);
+
+		for (i = 0; i < NUM_XDP_STATS; i++, strings++,
+		     bitmap_iterator_inc(&it))
+			if (bitmap_iterator_test(&it))
+				strcpy(data + (index++) * ETH_GSTRING_LEN,
+				       main_strings[strings]);
+
+		for (i = 0; i < NUM_PHY_STATS; i++, strings++,
+		     bitmap_iterator_inc(&it))
+			if (bitmap_iterator_test(&it))
+				strcpy(data + (index++) * ETH_GSTRING_LEN,
+				       main_strings[strings]);
+
+		for (i = 0; i < priv->tx_ring_num[TX]; i++) {
+			sprintf(data + (index++) * ETH_GSTRING_LEN,
+				"tx%d_packets", i);
+			sprintf(data + (index++) * ETH_GSTRING_LEN,
+				"tx%d_bytes", i);
+		}
+		for (i = 0; i < priv->rx_ring_num; i++) {
+			sprintf(data + (index++) * ETH_GSTRING_LEN,
+				"rx%d_packets", i);
+			sprintf(data + (index++) * ETH_GSTRING_LEN,
+				"rx%d_bytes", i);
+			sprintf(data + (index++) * ETH_GSTRING_LEN,
+				"rx%d_dropped", i);
+			sprintf(data + (index++) * ETH_GSTRING_LEN,
+				"rx%d_xdp_drop", i);
+			sprintf(data + (index++) * ETH_GSTRING_LEN,
+				"rx%d_xdp_tx", i);
+			sprintf(data + (index++) * ETH_GSTRING_LEN,
+				"rx%d_xdp_tx_full", i);
+		}
+		break;
+	case ETH_SS_PRIV_FLAGS:
+		for (i = 0; i < ARRAY_SIZE(mlx4_en_priv_flags); i++)
+			strcpy(data + i * ETH_GSTRING_LEN,
+			       mlx4_en_priv_flags[i]);
+		break;
+
+	}
+}
+
+static u32 mlx4_en_autoneg_get(struct net_device *dev)
+{
+	struct mlx4_en_priv *priv = netdev_priv(dev);
+	struct mlx4_en_dev *mdev = priv->mdev;
+	u32 autoneg = AUTONEG_DISABLE;
+
+	if ((mdev->dev->caps.flags2 & MLX4_DEV_CAP_FLAG2_ETH_BACKPL_AN_REP) &&
+	    (priv->port_state.flags & MLX4_EN_PORT_ANE))
+		autoneg = AUTONEG_ENABLE;
+
+	return autoneg;
+}
+
+static void ptys2ethtool_update_supported_port(unsigned long *mask,
+					       struct mlx4_ptys_reg *ptys_reg)
+{
+	u32 eth_proto = be32_to_cpu(ptys_reg->eth_proto_cap);
+
+	if (eth_proto & (MLX4_PROT_MASK(MLX4_10GBASE_T)
+			 | MLX4_PROT_MASK(MLX4_1000BASE_T)
+			 | MLX4_PROT_MASK(MLX4_100BASE_TX))) {
+		__set_bit(ETHTOOL_LINK_MODE_TP_BIT, mask);
+	} else if (eth_proto & (MLX4_PROT_MASK(MLX4_10GBASE_CR)
+			 | MLX4_PROT_MASK(MLX4_10GBASE_SR)
+			 | MLX4_PROT_MASK(MLX4_56GBASE_SR4)
+			 | MLX4_PROT_MASK(MLX4_40GBASE_CR4)
+			 | MLX4_PROT_MASK(MLX4_40GBASE_SR4)
+			 | MLX4_PROT_MASK(MLX4_1000BASE_CX_SGMII))) {
+		__set_bit(ETHTOOL_LINK_MODE_FIBRE_BIT, mask);
+	} else if (eth_proto & (MLX4_PROT_MASK(MLX4_56GBASE_KR4)
+			 | MLX4_PROT_MASK(MLX4_40GBASE_KR4)
+			 | MLX4_PROT_MASK(MLX4_20GBASE_KR2)
+			 | MLX4_PROT_MASK(MLX4_10GBASE_KR)
+			 | MLX4_PROT_MASK(MLX4_10GBASE_KX4)
+			 | MLX4_PROT_MASK(MLX4_1000BASE_KX))) {
+		__set_bit(ETHTOOL_LINK_MODE_Backplane_BIT, mask);
+	}
+}
+
+static u32 ptys_get_active_port(struct mlx4_ptys_reg *ptys_reg)
+{
+	u32 eth_proto = be32_to_cpu(ptys_reg->eth_proto_oper);
+
+	if (!eth_proto) /* link down */
+		eth_proto = be32_to_cpu(ptys_reg->eth_proto_cap);
+
+	if (eth_proto & (MLX4_PROT_MASK(MLX4_10GBASE_T)
+			 | MLX4_PROT_MASK(MLX4_1000BASE_T)
+			 | MLX4_PROT_MASK(MLX4_100BASE_TX))) {
+			return PORT_TP;
+	}
+
+	if (eth_proto & (MLX4_PROT_MASK(MLX4_10GBASE_SR)
+			 | MLX4_PROT_MASK(MLX4_56GBASE_SR4)
+			 | MLX4_PROT_MASK(MLX4_40GBASE_SR4)
+			 | MLX4_PROT_MASK(MLX4_1000BASE_CX_SGMII))) {
+			return PORT_FIBRE;
+	}
+
+	if (eth_proto & (MLX4_PROT_MASK(MLX4_10GBASE_CR)
+			 | MLX4_PROT_MASK(MLX4_56GBASE_CR4)
+			 | MLX4_PROT_MASK(MLX4_40GBASE_CR4))) {
+			return PORT_DA;
+	}
+
+	if (eth_proto & (MLX4_PROT_MASK(MLX4_56GBASE_KR4)
+			 | MLX4_PROT_MASK(MLX4_40GBASE_KR4)
+			 | MLX4_PROT_MASK(MLX4_20GBASE_KR2)
+			 | MLX4_PROT_MASK(MLX4_10GBASE_KR)
+			 | MLX4_PROT_MASK(MLX4_10GBASE_KX4)
+			 | MLX4_PROT_MASK(MLX4_1000BASE_KX))) {
+			return PORT_NONE;
+	}
+	return PORT_OTHER;
+}
+
+#define MLX4_LINK_MODES_SZ \
+	(FIELD_SIZEOF(struct mlx4_ptys_reg, eth_proto_cap) * 8)
+
+enum ethtool_report {
+	SUPPORTED = 0,
+	ADVERTISED = 1,
+};
+
+struct ptys2ethtool_config {
+	__ETHTOOL_DECLARE_LINK_MODE_MASK(supported);
+	__ETHTOOL_DECLARE_LINK_MODE_MASK(advertised);
+	u32 speed;
+};
+
+static unsigned long *ptys2ethtool_link_mode(struct ptys2ethtool_config *cfg,
+					     enum ethtool_report report)
+{
+	switch (report) {
+	case SUPPORTED:
+		return cfg->supported;
+	case ADVERTISED:
+		return cfg->advertised;
+	}
+	return NULL;
+}
+
+#define MLX4_BUILD_PTYS2ETHTOOL_CONFIG(reg_, speed_, ...)		\
+	({								\
+		struct ptys2ethtool_config *cfg;			\
+		const unsigned int modes[] = { __VA_ARGS__ };		\
+		unsigned int i;						\
+		cfg = &ptys2ethtool_map[reg_];				\
+		cfg->speed = speed_;					\
+		bitmap_zero(cfg->supported,				\
+			    __ETHTOOL_LINK_MODE_MASK_NBITS);		\
+		bitmap_zero(cfg->advertised,				\
+			    __ETHTOOL_LINK_MODE_MASK_NBITS);		\
+		for (i = 0 ; i < ARRAY_SIZE(modes) ; ++i) {		\
+			__set_bit(modes[i], cfg->supported);		\
+			__set_bit(modes[i], cfg->advertised);		\
+		}							\
+	})
+
+/* Translates mlx4 link mode to equivalent ethtool Link modes/speed */
+static struct ptys2ethtool_config ptys2ethtool_map[MLX4_LINK_MODES_SZ];
+
+void __init mlx4_en_init_ptys2ethtool_map(void)
+{
+	MLX4_BUILD_PTYS2ETHTOOL_CONFIG(MLX4_100BASE_TX, SPEED_100,
+				       ETHTOOL_LINK_MODE_100baseT_Full_BIT);
+	MLX4_BUILD_PTYS2ETHTOOL_CONFIG(MLX4_1000BASE_T, SPEED_1000,
+				       ETHTOOL_LINK_MODE_1000baseT_Full_BIT);
+	MLX4_BUILD_PTYS2ETHTOOL_CONFIG(MLX4_1000BASE_CX_SGMII, SPEED_1000,
+				       ETHTOOL_LINK_MODE_1000baseX_Full_BIT);
+	MLX4_BUILD_PTYS2ETHTOOL_CONFIG(MLX4_1000BASE_KX, SPEED_1000,
+				       ETHTOOL_LINK_MODE_1000baseKX_Full_BIT);
+	MLX4_BUILD_PTYS2ETHTOOL_CONFIG(MLX4_10GBASE_T, SPEED_10000,
+				       ETHTOOL_LINK_MODE_10000baseT_Full_BIT);
+	MLX4_BUILD_PTYS2ETHTOOL_CONFIG(MLX4_10GBASE_CX4, SPEED_10000,
+				       ETHTOOL_LINK_MODE_10000baseKX4_Full_BIT);
+	MLX4_BUILD_PTYS2ETHTOOL_CONFIG(MLX4_10GBASE_KX4, SPEED_10000,
+				       ETHTOOL_LINK_MODE_10000baseKX4_Full_BIT);
+	MLX4_BUILD_PTYS2ETHTOOL_CONFIG(MLX4_10GBASE_KR, SPEED_10000,
+				       ETHTOOL_LINK_MODE_10000baseKR_Full_BIT);
+	MLX4_BUILD_PTYS2ETHTOOL_CONFIG(MLX4_10GBASE_CR, SPEED_10000,
+				       ETHTOOL_LINK_MODE_10000baseCR_Full_BIT);
+	MLX4_BUILD_PTYS2ETHTOOL_CONFIG(MLX4_10GBASE_SR, SPEED_10000,
+				       ETHTOOL_LINK_MODE_10000baseSR_Full_BIT);
+	MLX4_BUILD_PTYS2ETHTOOL_CONFIG(MLX4_20GBASE_KR2, SPEED_20000,
+				       ETHTOOL_LINK_MODE_20000baseMLD2_Full_BIT,
+				       ETHTOOL_LINK_MODE_20000baseKR2_Full_BIT);
+	MLX4_BUILD_PTYS2ETHTOOL_CONFIG(MLX4_40GBASE_CR4, SPEED_40000,
+				       ETHTOOL_LINK_MODE_40000baseCR4_Full_BIT);
+	MLX4_BUILD_PTYS2ETHTOOL_CONFIG(MLX4_40GBASE_KR4, SPEED_40000,
+				       ETHTOOL_LINK_MODE_40000baseKR4_Full_BIT);
+	MLX4_BUILD_PTYS2ETHTOOL_CONFIG(MLX4_40GBASE_SR4, SPEED_40000,
+				       ETHTOOL_LINK_MODE_40000baseSR4_Full_BIT);
+	MLX4_BUILD_PTYS2ETHTOOL_CONFIG(MLX4_56GBASE_KR4, SPEED_56000,
+				       ETHTOOL_LINK_MODE_56000baseKR4_Full_BIT);
+	MLX4_BUILD_PTYS2ETHTOOL_CONFIG(MLX4_56GBASE_CR4, SPEED_56000,
+				       ETHTOOL_LINK_MODE_56000baseCR4_Full_BIT);
+	MLX4_BUILD_PTYS2ETHTOOL_CONFIG(MLX4_56GBASE_SR4, SPEED_56000,
+				       ETHTOOL_LINK_MODE_56000baseSR4_Full_BIT);
+};
+
+static void ptys2ethtool_update_link_modes(unsigned long *link_modes,
+					   u32 eth_proto,
+					   enum ethtool_report report)
+{
+	int i;
+	for (i = 0; i < MLX4_LINK_MODES_SZ; i++) {
+		if (eth_proto & MLX4_PROT_MASK(i))
+			bitmap_or(link_modes, link_modes,
+				  ptys2ethtool_link_mode(&ptys2ethtool_map[i],
+							 report),
+				  __ETHTOOL_LINK_MODE_MASK_NBITS);
+	}
+}
+
+static u32 ethtool2ptys_link_modes(const unsigned long *link_modes,
+				   enum ethtool_report report)
+{
+	int i;
+	u32 ptys_modes = 0;
+
+	for (i = 0; i < MLX4_LINK_MODES_SZ; i++) {
+		if (bitmap_intersects(
+			    ptys2ethtool_link_mode(&ptys2ethtool_map[i],
+						   report),
+			    link_modes,
+			    __ETHTOOL_LINK_MODE_MASK_NBITS))
+			ptys_modes |= 1 << i;
+	}
+	return ptys_modes;
+}
+
+/* Convert actual speed (SPEED_XXX) to ptys link modes */
+static u32 speed2ptys_link_modes(u32 speed)
+{
+	int i;
+	u32 ptys_modes = 0;
+
+	for (i = 0; i < MLX4_LINK_MODES_SZ; i++) {
+		if (ptys2ethtool_map[i].speed == speed)
+			ptys_modes |= 1 << i;
+	}
+	return ptys_modes;
+}
+
+static int
+ethtool_get_ptys_link_ksettings(struct net_device *dev,
+				struct ethtool_link_ksettings *link_ksettings)
+{
+	struct mlx4_en_priv *priv = netdev_priv(dev);
+	struct mlx4_ptys_reg ptys_reg;
+	u32 eth_proto;
+	int ret;
+
+	memset(&ptys_reg, 0, sizeof(ptys_reg));
+	ptys_reg.local_port = priv->port;
+	ptys_reg.proto_mask = MLX4_PTYS_EN;
+	ret = mlx4_ACCESS_PTYS_REG(priv->mdev->dev,
+				   MLX4_ACCESS_REG_QUERY, &ptys_reg);
+	if (ret) {
+		en_warn(priv, "Failed to run mlx4_ACCESS_PTYS_REG status(%x)",
+			ret);
+		return ret;
+	}
+	en_dbg(DRV, priv, "ptys_reg.proto_mask       %x\n",
+	       ptys_reg.proto_mask);
+	en_dbg(DRV, priv, "ptys_reg.eth_proto_cap    %x\n",
+	       be32_to_cpu(ptys_reg.eth_proto_cap));
+	en_dbg(DRV, priv, "ptys_reg.eth_proto_admin  %x\n",
+	       be32_to_cpu(ptys_reg.eth_proto_admin));
+	en_dbg(DRV, priv, "ptys_reg.eth_proto_oper   %x\n",
+	       be32_to_cpu(ptys_reg.eth_proto_oper));
+	en_dbg(DRV, priv, "ptys_reg.eth_proto_lp_adv %x\n",
+	       be32_to_cpu(ptys_reg.eth_proto_lp_adv));
+
+	/* reset supported/advertising masks */
+	ethtool_link_ksettings_zero_link_mode(link_ksettings, supported);
+	ethtool_link_ksettings_zero_link_mode(link_ksettings, advertising);
+
+	ptys2ethtool_update_supported_port(link_ksettings->link_modes.supported,
+					   &ptys_reg);
+
+	eth_proto = be32_to_cpu(ptys_reg.eth_proto_cap);
+	ptys2ethtool_update_link_modes(link_ksettings->link_modes.supported,
+				       eth_proto, SUPPORTED);
+
+	eth_proto = be32_to_cpu(ptys_reg.eth_proto_admin);
+	ptys2ethtool_update_link_modes(link_ksettings->link_modes.advertising,
+				       eth_proto, ADVERTISED);
+
+	ethtool_link_ksettings_add_link_mode(link_ksettings, supported,
+					     Pause);
+	ethtool_link_ksettings_add_link_mode(link_ksettings, supported,
+					     Asym_Pause);
+
+	if (priv->prof->tx_pause)
+		ethtool_link_ksettings_add_link_mode(link_ksettings,
+						     advertising, Pause);
+	if (priv->prof->tx_pause ^ priv->prof->rx_pause)
+		ethtool_link_ksettings_add_link_mode(link_ksettings,
+						     advertising, Asym_Pause);
+
+	link_ksettings->base.port = ptys_get_active_port(&ptys_reg);
+
+	if (mlx4_en_autoneg_get(dev)) {
+		ethtool_link_ksettings_add_link_mode(link_ksettings,
+						     supported, Autoneg);
+		ethtool_link_ksettings_add_link_mode(link_ksettings,
+						     advertising, Autoneg);
+	}
+
+	link_ksettings->base.autoneg
+		= (priv->port_state.flags & MLX4_EN_PORT_ANC) ?
+		AUTONEG_ENABLE : AUTONEG_DISABLE;
+
+	eth_proto = be32_to_cpu(ptys_reg.eth_proto_lp_adv);
+
+	ethtool_link_ksettings_zero_link_mode(link_ksettings, lp_advertising);
+	ptys2ethtool_update_link_modes(
+		link_ksettings->link_modes.lp_advertising,
+		eth_proto, ADVERTISED);
+	if (priv->port_state.flags & MLX4_EN_PORT_ANC)
+		ethtool_link_ksettings_add_link_mode(link_ksettings,
+						     lp_advertising, Autoneg);
+
+	link_ksettings->base.phy_address = 0;
+	link_ksettings->base.mdio_support = 0;
+	link_ksettings->base.eth_tp_mdix = ETH_TP_MDI_INVALID;
+	link_ksettings->base.eth_tp_mdix_ctrl = ETH_TP_MDI_AUTO;
+
+	return ret;
+}
+
+static void
+ethtool_get_default_link_ksettings(
+	struct net_device *dev, struct ethtool_link_ksettings *link_ksettings)
+{
+	struct mlx4_en_priv *priv = netdev_priv(dev);
+	int trans_type;
+
+	link_ksettings->base.autoneg = AUTONEG_DISABLE;
+
+	ethtool_link_ksettings_zero_link_mode(link_ksettings, supported);
+	ethtool_link_ksettings_add_link_mode(link_ksettings, supported,
+					     10000baseT_Full);
+
+	ethtool_link_ksettings_zero_link_mode(link_ksettings, advertising);
+	ethtool_link_ksettings_add_link_mode(link_ksettings, advertising,
+					     10000baseT_Full);
+
+	trans_type = priv->port_state.transceiver;
+	if (trans_type > 0 && trans_type <= 0xC) {
+		link_ksettings->base.port = PORT_FIBRE;
+		ethtool_link_ksettings_add_link_mode(link_ksettings,
+						     supported, FIBRE);
+		ethtool_link_ksettings_add_link_mode(link_ksettings,
+						     advertising, FIBRE);
+	} else if (trans_type == 0x80 || trans_type == 0) {
+		link_ksettings->base.port = PORT_TP;
+		ethtool_link_ksettings_add_link_mode(link_ksettings,
+						     supported, TP);
+		ethtool_link_ksettings_add_link_mode(link_ksettings,
+						     advertising, TP);
+	} else  {
+		link_ksettings->base.port = -1;
+	}
+}
+
+static int
+mlx4_en_get_link_ksettings(struct net_device *dev,
+			   struct ethtool_link_ksettings *link_ksettings)
+{
+	struct mlx4_en_priv *priv = netdev_priv(dev);
+	int ret = -EINVAL;
+
+	if (mlx4_en_QUERY_PORT(priv->mdev, priv->port))
+		return -ENOMEM;
+
+	en_dbg(DRV, priv, "query port state.flags ANC(%x) ANE(%x)\n",
+	       priv->port_state.flags & MLX4_EN_PORT_ANC,
+	       priv->port_state.flags & MLX4_EN_PORT_ANE);
+
+	if (priv->mdev->dev->caps.flags2 & MLX4_DEV_CAP_FLAG2_ETH_PROT_CTRL)
+		ret = ethtool_get_ptys_link_ksettings(dev, link_ksettings);
+	if (ret) /* ETH PROT CRTL is not supported or PTYS CMD failed */
+		ethtool_get_default_link_ksettings(dev, link_ksettings);
+
+	if (netif_carrier_ok(dev)) {
+		link_ksettings->base.speed = priv->port_state.link_speed;
+		link_ksettings->base.duplex = DUPLEX_FULL;
+	} else {
+		link_ksettings->base.speed = SPEED_UNKNOWN;
+		link_ksettings->base.duplex = DUPLEX_UNKNOWN;
+	}
+	return 0;
+}
+
+/* Calculate PTYS admin according ethtool speed (SPEED_XXX) */
+static __be32 speed_set_ptys_admin(struct mlx4_en_priv *priv, u32 speed,
+				   __be32 proto_cap)
+{
+	__be32 proto_admin = 0;
+
+	if (!speed) { /* Speed = 0 ==> Reset Link modes */
+		proto_admin = proto_cap;
+		en_info(priv, "Speed was set to 0, Reset advertised Link Modes to default (%x)\n",
+			be32_to_cpu(proto_cap));
+	} else {
+		u32 ptys_link_modes = speed2ptys_link_modes(speed);
+
+		proto_admin = cpu_to_be32(ptys_link_modes) & proto_cap;
+		en_info(priv, "Setting Speed to %d\n", speed);
+	}
+	return proto_admin;
+}
+
+static int
+mlx4_en_set_link_ksettings(struct net_device *dev,
+			   const struct ethtool_link_ksettings *link_ksettings)
+{
+	struct mlx4_en_priv *priv = netdev_priv(dev);
+	struct mlx4_ptys_reg ptys_reg;
+	__be32 proto_admin;
+	u8 cur_autoneg;
+	int ret;
+
+	u32 ptys_adv = ethtool2ptys_link_modes(
+		link_ksettings->link_modes.advertising, ADVERTISED);
+	const int speed = link_ksettings->base.speed;
+
+	en_dbg(DRV, priv,
+	       "Set Speed=%d adv={%*pbl} autoneg=%d duplex=%d\n",
+	       speed, __ETHTOOL_LINK_MODE_MASK_NBITS,
+	       link_ksettings->link_modes.advertising,
+	       link_ksettings->base.autoneg,
+	       link_ksettings->base.duplex);
+
+	if (!(priv->mdev->dev->caps.flags2 &
+	      MLX4_DEV_CAP_FLAG2_ETH_PROT_CTRL) ||
+	    (link_ksettings->base.duplex == DUPLEX_HALF))
+		return -EINVAL;
+
+	memset(&ptys_reg, 0, sizeof(ptys_reg));
+	ptys_reg.local_port = priv->port;
+	ptys_reg.proto_mask = MLX4_PTYS_EN;
+	ret = mlx4_ACCESS_PTYS_REG(priv->mdev->dev,
+				   MLX4_ACCESS_REG_QUERY, &ptys_reg);
+	if (ret) {
+		en_warn(priv, "Failed to QUERY mlx4_ACCESS_PTYS_REG status(%x)\n",
+			ret);
+		return 0;
+	}
+
+	cur_autoneg = ptys_reg.flags & MLX4_PTYS_AN_DISABLE_ADMIN ?
+				AUTONEG_DISABLE : AUTONEG_ENABLE;
+
+	if (link_ksettings->base.autoneg == AUTONEG_DISABLE) {
+		proto_admin = speed_set_ptys_admin(priv, speed,
+						   ptys_reg.eth_proto_cap);
+		if ((be32_to_cpu(proto_admin) &
+		     (MLX4_PROT_MASK(MLX4_1000BASE_CX_SGMII) |
+		      MLX4_PROT_MASK(MLX4_1000BASE_KX))) &&
+		    (ptys_reg.flags & MLX4_PTYS_AN_DISABLE_CAP))
+			ptys_reg.flags |= MLX4_PTYS_AN_DISABLE_ADMIN;
+	} else {
+		proto_admin = cpu_to_be32(ptys_adv);
+		ptys_reg.flags &= ~MLX4_PTYS_AN_DISABLE_ADMIN;
+	}
+
+	proto_admin &= ptys_reg.eth_proto_cap;
+	if (!proto_admin) {
+		en_warn(priv, "Not supported link mode(s) requested, check supported link modes.\n");
+		return -EINVAL; /* nothing to change due to bad input */
+	}
+
+	if ((proto_admin == ptys_reg.eth_proto_admin) &&
+	    ((ptys_reg.flags & MLX4_PTYS_AN_DISABLE_CAP) &&
+	     (link_ksettings->base.autoneg == cur_autoneg)))
+		return 0; /* Nothing to change */
+
+	en_dbg(DRV, priv, "mlx4_ACCESS_PTYS_REG SET: ptys_reg.eth_proto_admin = 0x%x\n",
+	       be32_to_cpu(proto_admin));
+
+	ptys_reg.eth_proto_admin = proto_admin;
+	ret = mlx4_ACCESS_PTYS_REG(priv->mdev->dev, MLX4_ACCESS_REG_WRITE,
+				   &ptys_reg);
+	if (ret) {
+		en_warn(priv, "Failed to write mlx4_ACCESS_PTYS_REG eth_proto_admin(0x%x) status(0x%x)",
+			be32_to_cpu(ptys_reg.eth_proto_admin), ret);
+		return ret;
+	}
+
+	mutex_lock(&priv->mdev->state_lock);
+	if (priv->port_up) {
+		en_warn(priv, "Port link mode changed, restarting port...\n");
+		mlx4_en_stop_port(dev, 1);
+		if (mlx4_en_start_port(dev))
+			en_err(priv, "Failed restarting port %d\n", priv->port);
+	}
+	mutex_unlock(&priv->mdev->state_lock);
+	return 0;
+}
+
+static int mlx4_en_get_coalesce(struct net_device *dev,
+			      struct ethtool_coalesce *coal)
+{
+	struct mlx4_en_priv *priv = netdev_priv(dev);
+
+	coal->tx_coalesce_usecs = priv->tx_usecs;
+	coal->tx_max_coalesced_frames = priv->tx_frames;
+	coal->tx_max_coalesced_frames_irq = priv->tx_work_limit;
+
+	coal->rx_coalesce_usecs = priv->rx_usecs;
+	coal->rx_max_coalesced_frames = priv->rx_frames;
+
+	coal->pkt_rate_low = priv->pkt_rate_low;
+	coal->rx_coalesce_usecs_low = priv->rx_usecs_low;
+	coal->pkt_rate_high = priv->pkt_rate_high;
+	coal->rx_coalesce_usecs_high = priv->rx_usecs_high;
+	coal->rate_sample_interval = priv->sample_interval;
+	coal->use_adaptive_rx_coalesce = priv->adaptive_rx_coal;
+
+	return 0;
+}
+
+static int mlx4_en_set_coalesce(struct net_device *dev,
+			      struct ethtool_coalesce *coal)
+{
+	struct mlx4_en_priv *priv = netdev_priv(dev);
+
+	if (!coal->tx_max_coalesced_frames_irq)
+		return -EINVAL;
+
+	if (coal->tx_coalesce_usecs > MLX4_EN_MAX_COAL_TIME ||
+	    coal->rx_coalesce_usecs > MLX4_EN_MAX_COAL_TIME ||
+	    coal->rx_coalesce_usecs_low > MLX4_EN_MAX_COAL_TIME ||
+	    coal->rx_coalesce_usecs_high > MLX4_EN_MAX_COAL_TIME) {
+		netdev_info(dev, "%s: maximum coalesce time supported is %d usecs\n",
+			    __func__, MLX4_EN_MAX_COAL_TIME);
+		return -ERANGE;
+	}
+
+	if (coal->tx_max_coalesced_frames > MLX4_EN_MAX_COAL_PKTS ||
+	    coal->rx_max_coalesced_frames > MLX4_EN_MAX_COAL_PKTS) {
+		netdev_info(dev, "%s: maximum coalesced frames supported is %d\n",
+			    __func__, MLX4_EN_MAX_COAL_PKTS);
+		return -ERANGE;
+	}
+
+	priv->rx_frames = (coal->rx_max_coalesced_frames ==
+			   MLX4_EN_AUTO_CONF) ?
+				MLX4_EN_RX_COAL_TARGET :
+				coal->rx_max_coalesced_frames;
+	priv->rx_usecs = (coal->rx_coalesce_usecs ==
+			  MLX4_EN_AUTO_CONF) ?
+				MLX4_EN_RX_COAL_TIME :
+				coal->rx_coalesce_usecs;
+
+	/* Setting TX coalescing parameters */
+	if (coal->tx_coalesce_usecs != priv->tx_usecs ||
+	    coal->tx_max_coalesced_frames != priv->tx_frames) {
+		priv->tx_usecs = coal->tx_coalesce_usecs;
+		priv->tx_frames = coal->tx_max_coalesced_frames;
+	}
+
+	/* Set adaptive coalescing params */
+	priv->pkt_rate_low = coal->pkt_rate_low;
+	priv->rx_usecs_low = coal->rx_coalesce_usecs_low;
+	priv->pkt_rate_high = coal->pkt_rate_high;
+	priv->rx_usecs_high = coal->rx_coalesce_usecs_high;
+	priv->sample_interval = coal->rate_sample_interval;
+	priv->adaptive_rx_coal = coal->use_adaptive_rx_coalesce;
+	priv->tx_work_limit = coal->tx_max_coalesced_frames_irq;
+
+	return mlx4_en_moderation_update(priv);
+}
+
+static int mlx4_en_set_pauseparam(struct net_device *dev,
+				struct ethtool_pauseparam *pause)
+{
+	struct mlx4_en_priv *priv = netdev_priv(dev);
+	struct mlx4_en_dev *mdev = priv->mdev;
+	u8 tx_pause, tx_ppp, rx_pause, rx_ppp;
+	int err;
+
+	if (pause->autoneg)
+		return -EINVAL;
+
+	tx_pause = !!(pause->tx_pause);
+	rx_pause = !!(pause->rx_pause);
+	rx_ppp = (tx_pause || rx_pause) ? 0 : priv->prof->rx_ppp;
+	tx_ppp = (tx_pause || rx_pause) ? 0 : priv->prof->tx_ppp;
+
+	err = mlx4_SET_PORT_general(mdev->dev, priv->port,
+				    priv->rx_skb_size + ETH_FCS_LEN,
+				    tx_pause, tx_ppp, rx_pause, rx_ppp);
+	if (err) {
+		en_err(priv, "Failed setting pause params, err = %d\n", err);
+		return err;
+	}
+
+	mlx4_en_update_pfc_stats_bitmap(mdev->dev, &priv->stats_bitmap,
+					rx_ppp, rx_pause, tx_ppp, tx_pause);
+
+	priv->prof->tx_pause = tx_pause;
+	priv->prof->rx_pause = rx_pause;
+	priv->prof->tx_ppp = tx_ppp;
+	priv->prof->rx_ppp = rx_ppp;
+
+	return err;
+}
+
+static void mlx4_en_get_pauseparam(struct net_device *dev,
+				 struct ethtool_pauseparam *pause)
+{
+	struct mlx4_en_priv *priv = netdev_priv(dev);
+
+	pause->tx_pause = priv->prof->tx_pause;
+	pause->rx_pause = priv->prof->rx_pause;
+}
+
+static int mlx4_en_set_ringparam(struct net_device *dev,
+				 struct ethtool_ringparam *param)
+{
+	struct mlx4_en_priv *priv = netdev_priv(dev);
+	struct mlx4_en_dev *mdev = priv->mdev;
+	struct mlx4_en_port_profile new_prof;
+	struct mlx4_en_priv *tmp;
+	u32 rx_size, tx_size;
+	int port_up = 0;
+	int err = 0;
+
+	if (param->rx_jumbo_pending || param->rx_mini_pending)
+		return -EINVAL;
+
+	if (param->rx_pending < MLX4_EN_MIN_RX_SIZE) {
+		en_warn(priv, "%s: rx_pending (%d) < min (%d)\n",
+			__func__, param->rx_pending,
+			MLX4_EN_MIN_RX_SIZE);
+		return -EINVAL;
+	}
+	if (param->tx_pending < MLX4_EN_MIN_TX_SIZE) {
+		en_warn(priv, "%s: tx_pending (%d) < min (%lu)\n",
+			__func__, param->tx_pending,
+			MLX4_EN_MIN_TX_SIZE);
+		return -EINVAL;
+	}
+
+	rx_size = roundup_pow_of_two(param->rx_pending);
+	tx_size = roundup_pow_of_two(param->tx_pending);
+
+	if (rx_size == (priv->port_up ? priv->rx_ring[0]->actual_size :
+					priv->rx_ring[0]->size) &&
+	    tx_size == priv->tx_ring[TX][0]->size)
+		return 0;
+
+	tmp = kzalloc(sizeof(*tmp), GFP_KERNEL);
+	if (!tmp)
+		return -ENOMEM;
+
+	mutex_lock(&mdev->state_lock);
+	memcpy(&new_prof, priv->prof, sizeof(struct mlx4_en_port_profile));
+	new_prof.tx_ring_size = tx_size;
+	new_prof.rx_ring_size = rx_size;
+	err = mlx4_en_try_alloc_resources(priv, tmp, &new_prof, true);
+	if (err)
+		goto out;
+
+	if (priv->port_up) {
+		port_up = 1;
+		mlx4_en_stop_port(dev, 1);
+	}
+
+	mlx4_en_safe_replace_resources(priv, tmp);
+
+	if (port_up) {
+		err = mlx4_en_start_port(dev);
+		if (err)
+			en_err(priv, "Failed starting port\n");
+	}
+
+	err = mlx4_en_moderation_update(priv);
+out:
+	kfree(tmp);
+	mutex_unlock(&mdev->state_lock);
+	return err;
+}
+
+static void mlx4_en_get_ringparam(struct net_device *dev,
+				  struct ethtool_ringparam *param)
+{
+	struct mlx4_en_priv *priv = netdev_priv(dev);
+
+	memset(param, 0, sizeof(*param));
+	param->rx_max_pending = MLX4_EN_MAX_RX_SIZE;
+	param->tx_max_pending = MLX4_EN_MAX_TX_SIZE;
+	param->rx_pending = priv->port_up ?
+		priv->rx_ring[0]->actual_size : priv->rx_ring[0]->size;
+	param->tx_pending = priv->tx_ring[TX][0]->size;
+}
+
+static u32 mlx4_en_get_rxfh_indir_size(struct net_device *dev)
+{
+	struct mlx4_en_priv *priv = netdev_priv(dev);
+
+	return rounddown_pow_of_two(priv->rx_ring_num);
+}
+
+static u32 mlx4_en_get_rxfh_key_size(struct net_device *netdev)
+{
+	return MLX4_EN_RSS_KEY_SIZE;
+}
+
+static int mlx4_en_check_rxfh_func(struct net_device *dev, u8 hfunc)
+{
+	struct mlx4_en_priv *priv = netdev_priv(dev);
+
+	/* check if requested function is supported by the device */
+	if (hfunc == ETH_RSS_HASH_TOP) {
+		if (!(priv->mdev->dev->caps.flags2 & MLX4_DEV_CAP_FLAG2_RSS_TOP))
+			return -EINVAL;
+		if (!(dev->features & NETIF_F_RXHASH))
+			en_warn(priv, "Toeplitz hash function should be used in conjunction with RX hashing for optimal performance\n");
+		return 0;
+	} else if (hfunc == ETH_RSS_HASH_XOR) {
+		if (!(priv->mdev->dev->caps.flags2 & MLX4_DEV_CAP_FLAG2_RSS_XOR))
+			return -EINVAL;
+		if (dev->features & NETIF_F_RXHASH)
+			en_warn(priv, "Enabling both XOR Hash function and RX Hashing can limit RPS functionality\n");
+		return 0;
+	}
+
+	return -EINVAL;
+}
+
+static int mlx4_en_get_rxfh(struct net_device *dev, u32 *ring_index, u8 *key,
+			    u8 *hfunc)
+{
+	struct mlx4_en_priv *priv = netdev_priv(dev);
+	u32 n = mlx4_en_get_rxfh_indir_size(dev);
+	u32 i, rss_rings;
+	int err = 0;
+
+	rss_rings = priv->prof->rss_rings ?: n;
+	rss_rings = rounddown_pow_of_two(rss_rings);
+
+	for (i = 0; i < n; i++) {
+		if (!ring_index)
+			break;
+		ring_index[i] = i % rss_rings;
+	}
+	if (key)
+		memcpy(key, priv->rss_key, MLX4_EN_RSS_KEY_SIZE);
+	if (hfunc)
+		*hfunc = priv->rss_hash_fn;
+	return err;
+}
+
+static int mlx4_en_set_rxfh(struct net_device *dev, const u32 *ring_index,
+			    const u8 *key, const u8 hfunc)
+{
+	struct mlx4_en_priv *priv = netdev_priv(dev);
+	u32 n = mlx4_en_get_rxfh_indir_size(dev);
+	struct mlx4_en_dev *mdev = priv->mdev;
+	int port_up = 0;
+	int err = 0;
+	int i;
+	int rss_rings = 0;
+
+	/* Calculate RSS table size and make sure flows are spread evenly
+	 * between rings
+	 */
+	for (i = 0; i < n; i++) {
+		if (!ring_index)
+			break;
+		if (i > 0 && !ring_index[i] && !rss_rings)
+			rss_rings = i;
+
+		if (ring_index[i] != (i % (rss_rings ?: n)))
+			return -EINVAL;
+	}
+
+	if (!rss_rings)
+		rss_rings = n;
+
+	/* RSS table size must be an order of 2 */
+	if (!is_power_of_2(rss_rings))
+		return -EINVAL;
+
+	if (hfunc != ETH_RSS_HASH_NO_CHANGE) {
+		err = mlx4_en_check_rxfh_func(dev, hfunc);
+		if (err)
+			return err;
+	}
+
+	mutex_lock(&mdev->state_lock);
+	if (priv->port_up) {
+		port_up = 1;
+		mlx4_en_stop_port(dev, 1);
+	}
+
+	if (ring_index)
+		priv->prof->rss_rings = rss_rings;
+	if (key)
+		memcpy(priv->rss_key, key, MLX4_EN_RSS_KEY_SIZE);
+	if (hfunc !=  ETH_RSS_HASH_NO_CHANGE)
+		priv->rss_hash_fn = hfunc;
+
+	if (port_up) {
+		err = mlx4_en_start_port(dev);
+		if (err)
+			en_err(priv, "Failed starting port\n");
+	}
+
+	mutex_unlock(&mdev->state_lock);
+	return err;
+}
+
+#define all_zeros_or_all_ones(field)		\
+	((field) == 0 || (field) == (__force typeof(field))-1)
+
+static int mlx4_en_validate_flow(struct net_device *dev,
+				 struct ethtool_rxnfc *cmd)
+{
+	struct ethtool_usrip4_spec *l3_mask;
+	struct ethtool_tcpip4_spec *l4_mask;
+	struct ethhdr *eth_mask;
+
+	if (cmd->fs.location >= MAX_NUM_OF_FS_RULES)
+		return -EINVAL;
+
+	if (cmd->fs.flow_type & FLOW_MAC_EXT) {
+		/* dest mac mask must be ff:ff:ff:ff:ff:ff */
+		if (!is_broadcast_ether_addr(cmd->fs.m_ext.h_dest))
+			return -EINVAL;
+	}
+
+	switch (cmd->fs.flow_type & ~(FLOW_EXT | FLOW_MAC_EXT)) {
+	case TCP_V4_FLOW:
+	case UDP_V4_FLOW:
+		if (cmd->fs.m_u.tcp_ip4_spec.tos)
+			return -EINVAL;
+		l4_mask = &cmd->fs.m_u.tcp_ip4_spec;
+		/* don't allow mask which isn't all 0 or 1 */
+		if (!all_zeros_or_all_ones(l4_mask->ip4src) ||
+		    !all_zeros_or_all_ones(l4_mask->ip4dst) ||
+		    !all_zeros_or_all_ones(l4_mask->psrc) ||
+		    !all_zeros_or_all_ones(l4_mask->pdst))
+			return -EINVAL;
+		break;
+	case IP_USER_FLOW:
+		l3_mask = &cmd->fs.m_u.usr_ip4_spec;
+		if (l3_mask->l4_4_bytes || l3_mask->tos || l3_mask->proto ||
+		    cmd->fs.h_u.usr_ip4_spec.ip_ver != ETH_RX_NFC_IP4 ||
+		    (!l3_mask->ip4src && !l3_mask->ip4dst) ||
+		    !all_zeros_or_all_ones(l3_mask->ip4src) ||
+		    !all_zeros_or_all_ones(l3_mask->ip4dst))
+			return -EINVAL;
+		break;
+	case ETHER_FLOW:
+		eth_mask = &cmd->fs.m_u.ether_spec;
+		/* source mac mask must not be set */
+		if (!is_zero_ether_addr(eth_mask->h_source))
+			return -EINVAL;
+
+		/* dest mac mask must be ff:ff:ff:ff:ff:ff */
+		if (!is_broadcast_ether_addr(eth_mask->h_dest))
+			return -EINVAL;
+
+		if (!all_zeros_or_all_ones(eth_mask->h_proto))
+			return -EINVAL;
+		break;
+	default:
+		return -EINVAL;
+	}
+
+	if ((cmd->fs.flow_type & FLOW_EXT)) {
+		if (cmd->fs.m_ext.vlan_etype ||
+		    !((cmd->fs.m_ext.vlan_tci & cpu_to_be16(VLAN_VID_MASK)) ==
+		      0 ||
+		      (cmd->fs.m_ext.vlan_tci & cpu_to_be16(VLAN_VID_MASK)) ==
+		      cpu_to_be16(VLAN_VID_MASK)))
+			return -EINVAL;
+
+		if (cmd->fs.m_ext.vlan_tci) {
+			if (be16_to_cpu(cmd->fs.h_ext.vlan_tci) >= VLAN_N_VID)
+				return -EINVAL;
+
+		}
+	}
+
+	return 0;
+}
+
+static int mlx4_en_ethtool_add_mac_rule(struct ethtool_rxnfc *cmd,
+					struct list_head *rule_list_h,
+					struct mlx4_spec_list *spec_l2,
+					unsigned char *mac)
+{
+	int err = 0;
+	__be64 mac_msk = cpu_to_be64(MLX4_MAC_MASK << 16);
+
+	spec_l2->id = MLX4_NET_TRANS_RULE_ID_ETH;
+	memcpy(spec_l2->eth.dst_mac_msk, &mac_msk, ETH_ALEN);
+	memcpy(spec_l2->eth.dst_mac, mac, ETH_ALEN);
+
+	if ((cmd->fs.flow_type & FLOW_EXT) &&
+	    (cmd->fs.m_ext.vlan_tci & cpu_to_be16(VLAN_VID_MASK))) {
+		spec_l2->eth.vlan_id = cmd->fs.h_ext.vlan_tci;
+		spec_l2->eth.vlan_id_msk = cpu_to_be16(VLAN_VID_MASK);
+	}
+
+	list_add_tail(&spec_l2->list, rule_list_h);
+
+	return err;
+}
+
+static int mlx4_en_ethtool_add_mac_rule_by_ipv4(struct mlx4_en_priv *priv,
+						struct ethtool_rxnfc *cmd,
+						struct list_head *rule_list_h,
+						struct mlx4_spec_list *spec_l2,
+						__be32 ipv4_dst)
+{
+#ifdef CONFIG_INET
+	unsigned char mac[ETH_ALEN];
+
+	if (!ipv4_is_multicast(ipv4_dst)) {
+		if (cmd->fs.flow_type & FLOW_MAC_EXT)
+			memcpy(&mac, cmd->fs.h_ext.h_dest, ETH_ALEN);
+		else
+			memcpy(&mac, priv->dev->dev_addr, ETH_ALEN);
+	} else {
+		ip_eth_mc_map(ipv4_dst, mac);
+	}
+
+	return mlx4_en_ethtool_add_mac_rule(cmd, rule_list_h, spec_l2, &mac[0]);
+#else
+	return -EINVAL;
+#endif
+}
+
+static int add_ip_rule(struct mlx4_en_priv *priv,
+		       struct ethtool_rxnfc *cmd,
+		       struct list_head *list_h)
+{
+	int err;
+	struct mlx4_spec_list *spec_l2 = NULL;
+	struct mlx4_spec_list *spec_l3 = NULL;
+	struct ethtool_usrip4_spec *l3_mask = &cmd->fs.m_u.usr_ip4_spec;
+
+	spec_l3 = kzalloc(sizeof(*spec_l3), GFP_KERNEL);
+	spec_l2 = kzalloc(sizeof(*spec_l2), GFP_KERNEL);
+	if (!spec_l2 || !spec_l3) {
+		err = -ENOMEM;
+		goto free_spec;
+	}
+
+	err = mlx4_en_ethtool_add_mac_rule_by_ipv4(priv, cmd, list_h, spec_l2,
+						   cmd->fs.h_u.
+						   usr_ip4_spec.ip4dst);
+	if (err)
+		goto free_spec;
+	spec_l3->id = MLX4_NET_TRANS_RULE_ID_IPV4;
+	spec_l3->ipv4.src_ip = cmd->fs.h_u.usr_ip4_spec.ip4src;
+	if (l3_mask->ip4src)
+		spec_l3->ipv4.src_ip_msk = EN_ETHTOOL_WORD_MASK;
+	spec_l3->ipv4.dst_ip = cmd->fs.h_u.usr_ip4_spec.ip4dst;
+	if (l3_mask->ip4dst)
+		spec_l3->ipv4.dst_ip_msk = EN_ETHTOOL_WORD_MASK;
+	list_add_tail(&spec_l3->list, list_h);
+
+	return 0;
+
+free_spec:
+	kfree(spec_l2);
+	kfree(spec_l3);
+	return err;
+}
+
+static int add_tcp_udp_rule(struct mlx4_en_priv *priv,
+			     struct ethtool_rxnfc *cmd,
+			     struct list_head *list_h, int proto)
+{
+	int err;
+	struct mlx4_spec_list *spec_l2 = NULL;
+	struct mlx4_spec_list *spec_l3 = NULL;
+	struct mlx4_spec_list *spec_l4 = NULL;
+	struct ethtool_tcpip4_spec *l4_mask = &cmd->fs.m_u.tcp_ip4_spec;
+
+	spec_l2 = kzalloc(sizeof(*spec_l2), GFP_KERNEL);
+	spec_l3 = kzalloc(sizeof(*spec_l3), GFP_KERNEL);
+	spec_l4 = kzalloc(sizeof(*spec_l4), GFP_KERNEL);
+	if (!spec_l2 || !spec_l3 || !spec_l4) {
+		err = -ENOMEM;
+		goto free_spec;
+	}
+
+	spec_l3->id = MLX4_NET_TRANS_RULE_ID_IPV4;
+
+	if (proto == TCP_V4_FLOW) {
+		err = mlx4_en_ethtool_add_mac_rule_by_ipv4(priv, cmd, list_h,
+							   spec_l2,
+							   cmd->fs.h_u.
+							   tcp_ip4_spec.ip4dst);
+		if (err)
+			goto free_spec;
+		spec_l4->id = MLX4_NET_TRANS_RULE_ID_TCP;
+		spec_l3->ipv4.src_ip = cmd->fs.h_u.tcp_ip4_spec.ip4src;
+		spec_l3->ipv4.dst_ip = cmd->fs.h_u.tcp_ip4_spec.ip4dst;
+		spec_l4->tcp_udp.src_port = cmd->fs.h_u.tcp_ip4_spec.psrc;
+		spec_l4->tcp_udp.dst_port = cmd->fs.h_u.tcp_ip4_spec.pdst;
+	} else {
+		err = mlx4_en_ethtool_add_mac_rule_by_ipv4(priv, cmd, list_h,
+							   spec_l2,
+							   cmd->fs.h_u.
+							   udp_ip4_spec.ip4dst);
+		if (err)
+			goto free_spec;
+		spec_l4->id = MLX4_NET_TRANS_RULE_ID_UDP;
+		spec_l3->ipv4.src_ip = cmd->fs.h_u.udp_ip4_spec.ip4src;
+		spec_l3->ipv4.dst_ip = cmd->fs.h_u.udp_ip4_spec.ip4dst;
+		spec_l4->tcp_udp.src_port = cmd->fs.h_u.udp_ip4_spec.psrc;
+		spec_l4->tcp_udp.dst_port = cmd->fs.h_u.udp_ip4_spec.pdst;
+	}
+
+	if (l4_mask->ip4src)
+		spec_l3->ipv4.src_ip_msk = EN_ETHTOOL_WORD_MASK;
+	if (l4_mask->ip4dst)
+		spec_l3->ipv4.dst_ip_msk = EN_ETHTOOL_WORD_MASK;
+
+	if (l4_mask->psrc)
+		spec_l4->tcp_udp.src_port_msk = EN_ETHTOOL_SHORT_MASK;
+	if (l4_mask->pdst)
+		spec_l4->tcp_udp.dst_port_msk = EN_ETHTOOL_SHORT_MASK;
+
+	list_add_tail(&spec_l3->list, list_h);
+	list_add_tail(&spec_l4->list, list_h);
+
+	return 0;
+
+free_spec:
+	kfree(spec_l2);
+	kfree(spec_l3);
+	kfree(spec_l4);
+	return err;
+}
+
+static int mlx4_en_ethtool_to_net_trans_rule(struct net_device *dev,
+					     struct ethtool_rxnfc *cmd,
+					     struct list_head *rule_list_h)
+{
+	int err;
+	struct ethhdr *eth_spec;
+	struct mlx4_spec_list *spec_l2;
+	struct mlx4_en_priv *priv = netdev_priv(dev);
+
+	err = mlx4_en_validate_flow(dev, cmd);
+	if (err)
+		return err;
+
+	switch (cmd->fs.flow_type & ~(FLOW_EXT | FLOW_MAC_EXT)) {
+	case ETHER_FLOW:
+		spec_l2 = kzalloc(sizeof(*spec_l2), GFP_KERNEL);
+		if (!spec_l2)
+			return -ENOMEM;
+
+		eth_spec = &cmd->fs.h_u.ether_spec;
+		mlx4_en_ethtool_add_mac_rule(cmd, rule_list_h, spec_l2,
+					     &eth_spec->h_dest[0]);
+		spec_l2->eth.ether_type = eth_spec->h_proto;
+		if (eth_spec->h_proto)
+			spec_l2->eth.ether_type_enable = 1;
+		break;
+	case IP_USER_FLOW:
+		err = add_ip_rule(priv, cmd, rule_list_h);
+		break;
+	case TCP_V4_FLOW:
+		err = add_tcp_udp_rule(priv, cmd, rule_list_h, TCP_V4_FLOW);
+		break;
+	case UDP_V4_FLOW:
+		err = add_tcp_udp_rule(priv, cmd, rule_list_h, UDP_V4_FLOW);
+		break;
+	}
+
+	return err;
+}
+
+static int mlx4_en_flow_replace(struct net_device *dev,
+				struct ethtool_rxnfc *cmd)
+{
+	int err;
+	struct mlx4_en_priv *priv = netdev_priv(dev);
+	struct ethtool_flow_id *loc_rule;
+	struct mlx4_spec_list *spec, *tmp_spec;
+	u32 qpn;
+	u64 reg_id;
+
+	struct mlx4_net_trans_rule rule = {
+		.queue_mode = MLX4_NET_TRANS_Q_FIFO,
+		.exclusive = 0,
+		.allow_loopback = 1,
+		.promisc_mode = MLX4_FS_REGULAR,
+	};
+
+	rule.port = priv->port;
+	rule.priority = MLX4_DOMAIN_ETHTOOL | cmd->fs.location;
+	INIT_LIST_HEAD(&rule.list);
+
+	/* Allow direct QP attaches if the EN_ETHTOOL_QP_ATTACH flag is set */
+	if (cmd->fs.ring_cookie == RX_CLS_FLOW_DISC)
+		qpn = priv->drop_qp.qpn;
+	else if (cmd->fs.ring_cookie & EN_ETHTOOL_QP_ATTACH) {
+		qpn = cmd->fs.ring_cookie & (EN_ETHTOOL_QP_ATTACH - 1);
+	} else {
+		if (cmd->fs.ring_cookie >= priv->rx_ring_num) {
+			en_warn(priv, "rxnfc: RX ring (%llu) doesn't exist\n",
+				cmd->fs.ring_cookie);
+			return -EINVAL;
+		}
+		qpn = priv->rss_map.qps[cmd->fs.ring_cookie].qpn;
+		if (!qpn) {
+			en_warn(priv, "rxnfc: RX ring (%llu) is inactive\n",
+				cmd->fs.ring_cookie);
+			return -EINVAL;
+		}
+	}
+	rule.qpn = qpn;
+	err = mlx4_en_ethtool_to_net_trans_rule(dev, cmd, &rule.list);
+	if (err)
+		goto out_free_list;
+
+	loc_rule = &priv->ethtool_rules[cmd->fs.location];
+	if (loc_rule->id) {
+		err = mlx4_flow_detach(priv->mdev->dev, loc_rule->id);
+		if (err) {
+			en_err(priv, "Fail to detach network rule at location %d. registration id = %llx\n",
+			       cmd->fs.location, loc_rule->id);
+			goto out_free_list;
+		}
+		loc_rule->id = 0;
+		memset(&loc_rule->flow_spec, 0,
+		       sizeof(struct ethtool_rx_flow_spec));
+		list_del(&loc_rule->list);
+	}
+	err = mlx4_flow_attach(priv->mdev->dev, &rule, &reg_id);
+	if (err) {
+		en_err(priv, "Fail to attach network rule at location %d\n",
+		       cmd->fs.location);
+		goto out_free_list;
+	}
+	loc_rule->id = reg_id;
+	memcpy(&loc_rule->flow_spec, &cmd->fs,
+	       sizeof(struct ethtool_rx_flow_spec));
+	list_add_tail(&loc_rule->list, &priv->ethtool_list);
+
+out_free_list:
+	list_for_each_entry_safe(spec, tmp_spec, &rule.list, list) {
+		list_del(&spec->list);
+		kfree(spec);
+	}
+	return err;
+}
+
+static int mlx4_en_flow_detach(struct net_device *dev,
+			       struct ethtool_rxnfc *cmd)
+{
+	int err = 0;
+	struct ethtool_flow_id *rule;
+	struct mlx4_en_priv *priv = netdev_priv(dev);
+
+	if (cmd->fs.location >= MAX_NUM_OF_FS_RULES)
+		return -EINVAL;
+
+	rule = &priv->ethtool_rules[cmd->fs.location];
+	if (!rule->id) {
+		err =  -ENOENT;
+		goto out;
+	}
+
+	err = mlx4_flow_detach(priv->mdev->dev, rule->id);
+	if (err) {
+		en_err(priv, "Fail to detach network rule at location %d. registration id = 0x%llx\n",
+		       cmd->fs.location, rule->id);
+		goto out;
+	}
+	rule->id = 0;
+	memset(&rule->flow_spec, 0, sizeof(struct ethtool_rx_flow_spec));
+	list_del(&rule->list);
+out:
+	return err;
+
+}
+
+static int mlx4_en_get_flow(struct net_device *dev, struct ethtool_rxnfc *cmd,
+			    int loc)
+{
+	int err = 0;
+	struct ethtool_flow_id *rule;
+	struct mlx4_en_priv *priv = netdev_priv(dev);
+
+	if (loc < 0 || loc >= MAX_NUM_OF_FS_RULES)
+		return -EINVAL;
+
+	rule = &priv->ethtool_rules[loc];
+	if (rule->id)
+		memcpy(&cmd->fs, &rule->flow_spec,
+		       sizeof(struct ethtool_rx_flow_spec));
+	else
+		err = -ENOENT;
+
+	return err;
+}
+
+static int mlx4_en_get_num_flows(struct mlx4_en_priv *priv)
+{
+
+	int i, res = 0;
+	for (i = 0; i < MAX_NUM_OF_FS_RULES; i++) {
+		if (priv->ethtool_rules[i].id)
+			res++;
+	}
+	return res;
+
+}
+
+static int mlx4_en_get_rxnfc(struct net_device *dev, struct ethtool_rxnfc *cmd,
+			     u32 *rule_locs)
+{
+	struct mlx4_en_priv *priv = netdev_priv(dev);
+	struct mlx4_en_dev *mdev = priv->mdev;
+	int err = 0;
+	int i = 0, priority = 0;
+
+	if ((cmd->cmd == ETHTOOL_GRXCLSRLCNT ||
+	     cmd->cmd == ETHTOOL_GRXCLSRULE ||
+	     cmd->cmd == ETHTOOL_GRXCLSRLALL) &&
+	    (mdev->dev->caps.steering_mode !=
+	     MLX4_STEERING_MODE_DEVICE_MANAGED || !priv->port_up))
+		return -EINVAL;
+
+	switch (cmd->cmd) {
+	case ETHTOOL_GRXRINGS:
+		cmd->data = priv->rx_ring_num;
+		break;
+	case ETHTOOL_GRXCLSRLCNT:
+		cmd->rule_cnt = mlx4_en_get_num_flows(priv);
+		break;
+	case ETHTOOL_GRXCLSRULE:
+		err = mlx4_en_get_flow(dev, cmd, cmd->fs.location);
+		break;
+	case ETHTOOL_GRXCLSRLALL:
+		cmd->data = MAX_NUM_OF_FS_RULES;
+		while ((!err || err == -ENOENT) && priority < cmd->rule_cnt) {
+			err = mlx4_en_get_flow(dev, cmd, i);
+			if (!err)
+				rule_locs[priority++] = i;
+			i++;
+		}
+		err = 0;
+		break;
+	default:
+		err = -EOPNOTSUPP;
+		break;
+	}
+
+	return err;
+}
+
+static int mlx4_en_set_rxnfc(struct net_device *dev, struct ethtool_rxnfc *cmd)
+{
+	int err = 0;
+	struct mlx4_en_priv *priv = netdev_priv(dev);
+	struct mlx4_en_dev *mdev = priv->mdev;
+
+	if (mdev->dev->caps.steering_mode !=
+	    MLX4_STEERING_MODE_DEVICE_MANAGED || !priv->port_up)
+		return -EINVAL;
+
+	switch (cmd->cmd) {
+	case ETHTOOL_SRXCLSRLINS:
+		err = mlx4_en_flow_replace(dev, cmd);
+		break;
+	case ETHTOOL_SRXCLSRLDEL:
+		err = mlx4_en_flow_detach(dev, cmd);
+		break;
+	default:
+		en_warn(priv, "Unsupported ethtool command. (%d)\n", cmd->cmd);
+		return -EINVAL;
+	}
+
+	return err;
+}
+
+static int mlx4_en_get_max_num_rx_rings(struct net_device *dev)
+{
+	return min_t(int, num_online_cpus(), MAX_RX_RINGS);
+}
+
+static void mlx4_en_get_channels(struct net_device *dev,
+				 struct ethtool_channels *channel)
+{
+	struct mlx4_en_priv *priv = netdev_priv(dev);
+
+	channel->max_rx = mlx4_en_get_max_num_rx_rings(dev);
+	channel->max_tx = priv->mdev->profile.max_num_tx_rings_p_up;
+
+	channel->rx_count = priv->rx_ring_num;
+	channel->tx_count = priv->tx_ring_num[TX] /
+			    priv->prof->num_up;
+}
+
+static int mlx4_en_set_channels(struct net_device *dev,
+				struct ethtool_channels *channel)
+{
+	struct mlx4_en_priv *priv = netdev_priv(dev);
+	struct mlx4_en_dev *mdev = priv->mdev;
+	struct mlx4_en_port_profile new_prof;
+	struct mlx4_en_priv *tmp;
+	int total_tx_count;
+	int port_up = 0;
+	int xdp_count;
+	int err = 0;
+	u8 up;
+
+	if (!channel->tx_count || !channel->rx_count)
+		return -EINVAL;
+
+	tmp = kzalloc(sizeof(*tmp), GFP_KERNEL);
+	if (!tmp)
+		return -ENOMEM;
+
+	mutex_lock(&mdev->state_lock);
+	xdp_count = priv->tx_ring_num[TX_XDP] ? channel->rx_count : 0;
+	total_tx_count = channel->tx_count * priv->prof->num_up + xdp_count;
+	if (total_tx_count > MAX_TX_RINGS) {
+		err = -EINVAL;
+		en_err(priv,
+		       "Total number of TX and XDP rings (%d) exceeds the maximum supported (%d)\n",
+		       total_tx_count, MAX_TX_RINGS);
+		goto out;
+	}
+
+	memcpy(&new_prof, priv->prof, sizeof(struct mlx4_en_port_profile));
+	new_prof.num_tx_rings_p_up = channel->tx_count;
+	new_prof.tx_ring_num[TX] = channel->tx_count * priv->prof->num_up;
+	new_prof.tx_ring_num[TX_XDP] = xdp_count;
+	new_prof.rx_ring_num = channel->rx_count;
+
+	err = mlx4_en_try_alloc_resources(priv, tmp, &new_prof, true);
+	if (err)
+		goto out;
+
+	if (priv->port_up) {
+		port_up = 1;
+		mlx4_en_stop_port(dev, 1);
+	}
+
+	mlx4_en_safe_replace_resources(priv, tmp);
+
+	netif_set_real_num_rx_queues(dev, priv->rx_ring_num);
+
+	up = (priv->prof->num_up == MLX4_EN_NUM_UP_LOW) ?
+				    0 : priv->prof->num_up;
+	mlx4_en_setup_tc(dev, up);
+
+	en_warn(priv, "Using %d TX rings\n", priv->tx_ring_num[TX]);
+	en_warn(priv, "Using %d RX rings\n", priv->rx_ring_num);
+
+	if (port_up) {
+		err = mlx4_en_start_port(dev);
+		if (err)
+			en_err(priv, "Failed starting port\n");
+	}
+
+	err = mlx4_en_moderation_update(priv);
+out:
+	mutex_unlock(&mdev->state_lock);
+	kfree(tmp);
+	return err;
+}
+
+static int mlx4_en_get_ts_info(struct net_device *dev,
+			       struct ethtool_ts_info *info)
+{
+	struct mlx4_en_priv *priv = netdev_priv(dev);
+	struct mlx4_en_dev *mdev = priv->mdev;
+	int ret;
+
+	ret = ethtool_op_get_ts_info(dev, info);
+	if (ret)
+		return ret;
+
+	if (mdev->dev->caps.flags2 & MLX4_DEV_CAP_FLAG2_TS) {
+		info->so_timestamping |=
+			SOF_TIMESTAMPING_TX_HARDWARE |
+			SOF_TIMESTAMPING_RX_HARDWARE |
+			SOF_TIMESTAMPING_RAW_HARDWARE;
+
+		info->tx_types =
+			(1 << HWTSTAMP_TX_OFF) |
+			(1 << HWTSTAMP_TX_ON);
+
+		info->rx_filters =
+			(1 << HWTSTAMP_FILTER_NONE) |
+			(1 << HWTSTAMP_FILTER_ALL);
+
+		if (mdev->ptp_clock)
+			info->phc_index = ptp_clock_index(mdev->ptp_clock);
+	}
+
+	return ret;
+}
+
+static int mlx4_en_set_priv_flags(struct net_device *dev, u32 flags)
+{
+	struct mlx4_en_priv *priv = netdev_priv(dev);
+	struct mlx4_en_dev *mdev = priv->mdev;
+	bool bf_enabled_new = !!(flags & MLX4_EN_PRIV_FLAGS_BLUEFLAME);
+	bool bf_enabled_old = !!(priv->pflags & MLX4_EN_PRIV_FLAGS_BLUEFLAME);
+	bool phv_enabled_new = !!(flags & MLX4_EN_PRIV_FLAGS_PHV);
+	bool phv_enabled_old = !!(priv->pflags & MLX4_EN_PRIV_FLAGS_PHV);
+	int i;
+	int ret = 0;
+
+	if (bf_enabled_new != bf_enabled_old) {
+		int t;
+
+		if (bf_enabled_new) {
+			bool bf_supported = true;
+
+			for (t = 0; t < MLX4_EN_NUM_TX_TYPES; t++)
+				for (i = 0; i < priv->tx_ring_num[t]; i++)
+					bf_supported &=
+						priv->tx_ring[t][i]->bf_alloced;
+
+			if (!bf_supported) {
+				en_err(priv, "BlueFlame is not supported\n");
+				return -EINVAL;
+			}
+
+			priv->pflags |= MLX4_EN_PRIV_FLAGS_BLUEFLAME;
+		} else {
+			priv->pflags &= ~MLX4_EN_PRIV_FLAGS_BLUEFLAME;
+		}
+
+		for (t = 0; t < MLX4_EN_NUM_TX_TYPES; t++)
+			for (i = 0; i < priv->tx_ring_num[t]; i++)
+				priv->tx_ring[t][i]->bf_enabled =
+					bf_enabled_new;
+
+		en_info(priv, "BlueFlame %s\n",
+			bf_enabled_new ?  "Enabled" : "Disabled");
+	}
+
+	if (phv_enabled_new != phv_enabled_old) {
+		ret = set_phv_bit(mdev->dev, priv->port, (int)phv_enabled_new);
+		if (ret)
+			return ret;
+		else if (phv_enabled_new)
+			priv->pflags |= MLX4_EN_PRIV_FLAGS_PHV;
+		else
+			priv->pflags &= ~MLX4_EN_PRIV_FLAGS_PHV;
+		en_info(priv, "PHV bit %s\n",
+			phv_enabled_new ?  "Enabled" : "Disabled");
+	}
+	return 0;
+}
+
+static u32 mlx4_en_get_priv_flags(struct net_device *dev)
+{
+	struct mlx4_en_priv *priv = netdev_priv(dev);
+
+	return priv->pflags;
+}
+
+static int mlx4_en_get_tunable(struct net_device *dev,
+			       const struct ethtool_tunable *tuna,
+			       void *data)
+{
+	const struct mlx4_en_priv *priv = netdev_priv(dev);
+	int ret = 0;
+
+	switch (tuna->id) {
+	case ETHTOOL_TX_COPYBREAK:
+		*(u32 *)data = priv->prof->inline_thold;
+		break;
+	default:
+		ret = -EINVAL;
+		break;
+	}
+
+	return ret;
+}
+
+static int mlx4_en_set_tunable(struct net_device *dev,
+			       const struct ethtool_tunable *tuna,
+			       const void *data)
+{
+	struct mlx4_en_priv *priv = netdev_priv(dev);
+	int val, ret = 0;
+
+	switch (tuna->id) {
+	case ETHTOOL_TX_COPYBREAK:
+		val = *(u32 *)data;
+		if (val < MIN_PKT_LEN || val > MAX_INLINE)
+			ret = -EINVAL;
+		else
+			priv->prof->inline_thold = val;
+		break;
+	default:
+		ret = -EINVAL;
+		break;
+	}
+
+	return ret;
+}
+
+static int mlx4_en_get_module_info(struct net_device *dev,
+				   struct ethtool_modinfo *modinfo)
+{
+	struct mlx4_en_priv *priv = netdev_priv(dev);
+	struct mlx4_en_dev *mdev = priv->mdev;
+	int ret;
+	u8 data[4];
+
+	/* Read first 2 bytes to get Module & REV ID */
+	ret = mlx4_get_module_info(mdev->dev, priv->port,
+				   0/*offset*/, 2/*size*/, data);
+	if (ret < 2)
+		return -EIO;
+
+	switch (data[0] /* identifier */) {
+	case MLX4_MODULE_ID_QSFP:
+		modinfo->type = ETH_MODULE_SFF_8436;
+		modinfo->eeprom_len = ETH_MODULE_SFF_8436_LEN;
+		break;
+	case MLX4_MODULE_ID_QSFP_PLUS:
+		if (data[1] >= 0x3) { /* revision id */
+			modinfo->type = ETH_MODULE_SFF_8636;
+			modinfo->eeprom_len = ETH_MODULE_SFF_8636_LEN;
+		} else {
+			modinfo->type = ETH_MODULE_SFF_8436;
+			modinfo->eeprom_len = ETH_MODULE_SFF_8436_LEN;
+		}
+		break;
+	case MLX4_MODULE_ID_QSFP28:
+		modinfo->type = ETH_MODULE_SFF_8636;
+		modinfo->eeprom_len = ETH_MODULE_SFF_8636_LEN;
+		break;
+	case MLX4_MODULE_ID_SFP:
+		modinfo->type = ETH_MODULE_SFF_8472;
+		modinfo->eeprom_len = ETH_MODULE_SFF_8472_LEN;
+		break;
+	default:
+		return -EINVAL;
+	}
+
+	return 0;
+}
+
+static int mlx4_en_get_module_eeprom(struct net_device *dev,
+				     struct ethtool_eeprom *ee,
+				     u8 *data)
+{
+	struct mlx4_en_priv *priv = netdev_priv(dev);
+	struct mlx4_en_dev *mdev = priv->mdev;
+	int offset = ee->offset;
+	int i = 0, ret;
+
+	if (ee->len == 0)
+		return -EINVAL;
+
+	memset(data, 0, ee->len);
+
+	while (i < ee->len) {
+		en_dbg(DRV, priv,
+		       "mlx4_get_module_info i(%d) offset(%d) len(%d)\n",
+		       i, offset, ee->len - i);
+
+		ret = mlx4_get_module_info(mdev->dev, priv->port,
+					   offset, ee->len - i, data + i);
+
+		if (!ret) /* Done reading */
+			return 0;
+
+		if (ret < 0) {
+			en_err(priv,
+			       "mlx4_get_module_info i(%d) offset(%d) bytes_to_read(%d) - FAILED (0x%x)\n",
+			       i, offset, ee->len - i, ret);
+			return ret;
+		}
+
+		i += ret;
+		offset += ret;
+	}
+	return 0;
+}
+
+static int mlx4_en_set_phys_id(struct net_device *dev,
+			       enum ethtool_phys_id_state state)
+{
+	int err;
+	u16 beacon_duration;
+	struct mlx4_en_priv *priv = netdev_priv(dev);
+	struct mlx4_en_dev *mdev = priv->mdev;
+
+	if (!(mdev->dev->caps.flags2 & MLX4_DEV_CAP_FLAG2_PORT_BEACON))
+		return -EOPNOTSUPP;
+
+	switch (state) {
+	case ETHTOOL_ID_ACTIVE:
+		beacon_duration = PORT_BEACON_MAX_LIMIT;
+		break;
+	case ETHTOOL_ID_INACTIVE:
+		beacon_duration = 0;
+		break;
+	default:
+		return -EOPNOTSUPP;
+	}
+
+	err = mlx4_SET_PORT_BEACON(mdev->dev, priv->port, beacon_duration);
+	return err;
+}
+
+const struct ethtool_ops mlx4_en_ethtool_ops = {
+	.get_drvinfo = mlx4_en_get_drvinfo,
+	.get_link_ksettings = mlx4_en_get_link_ksettings,
+	.set_link_ksettings = mlx4_en_set_link_ksettings,
+	.get_link = ethtool_op_get_link,
+	.get_strings = mlx4_en_get_strings,
+	.get_sset_count = mlx4_en_get_sset_count,
+	.get_ethtool_stats = mlx4_en_get_ethtool_stats,
+	.self_test = mlx4_en_self_test,
+	.set_phys_id = mlx4_en_set_phys_id,
+	.get_wol = mlx4_en_get_wol,
+	.set_wol = mlx4_en_set_wol,
+	.get_msglevel = mlx4_en_get_msglevel,
+	.set_msglevel = mlx4_en_set_msglevel,
+	.get_coalesce = mlx4_en_get_coalesce,
+	.set_coalesce = mlx4_en_set_coalesce,
+	.get_pauseparam = mlx4_en_get_pauseparam,
+	.set_pauseparam = mlx4_en_set_pauseparam,
+	.get_ringparam = mlx4_en_get_ringparam,
+	.set_ringparam = mlx4_en_set_ringparam,
+	.get_rxnfc = mlx4_en_get_rxnfc,
+	.set_rxnfc = mlx4_en_set_rxnfc,
+	.get_rxfh_indir_size = mlx4_en_get_rxfh_indir_size,
+	.get_rxfh_key_size = mlx4_en_get_rxfh_key_size,
+	.get_rxfh = mlx4_en_get_rxfh,
+	.set_rxfh = mlx4_en_set_rxfh,
+	.get_channels = mlx4_en_get_channels,
+	.set_channels = mlx4_en_set_channels,
+	.get_ts_info = mlx4_en_get_ts_info,
+	.set_priv_flags = mlx4_en_set_priv_flags,
+	.get_priv_flags = mlx4_en_get_priv_flags,
+	.get_tunable		= mlx4_en_get_tunable,
+	.set_tunable		= mlx4_en_set_tunable,
+	.get_module_info = mlx4_en_get_module_info,
+	.get_module_eeprom = mlx4_en_get_module_eeprom
+};
+
+
+
+
+
diff --git a/drivers/net/ethernet/mellanox/mlx4/en_main.c b/drivers/net/ethernet/mellanox/mlx4/en_main.c
new file mode 100644
index 000000000..d25e16d2c
--- /dev/null
+++ b/drivers/net/ethernet/mellanox/mlx4/en_main.c
@@ -0,0 +1,394 @@
+/*
+ * Copyright (c) 2007 Mellanox Technologies. All rights reserved.
+ *
+ * This software is available to you under a choice of one of two
+ * licenses.  You may choose to be licensed under the terms of the GNU
+ * General Public License (GPL) Version 2, available from the file
+ * COPYING in the main directory of this source tree, or the
+ * OpenIB.org BSD license below:
+ *
+ *     Redistribution and use in source and binary forms, with or
+ *     without modification, are permitted provided that the following
+ *     conditions are met:
+ *
+ *      - Redistributions of source code must retain the above
+ *        copyright notice, this list of conditions and the following
+ *        disclaimer.
+ *
+ *      - Redistributions in binary form must reproduce the above
+ *        copyright notice, this list of conditions and the following
+ *        disclaimer in the documentation and/or other materials
+ *        provided with the distribution.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
+ * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
+ * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS
+ * BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN
+ * ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
+ * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+ * SOFTWARE.
+ *
+ */
+
+#include <linux/cpumask.h>
+#include <linux/module.h>
+#include <linux/delay.h>
+#include <linux/netdevice.h>
+#include <linux/slab.h>
+
+#include <linux/mlx4/driver.h>
+#include <linux/mlx4/device.h>
+#include <linux/mlx4/cmd.h>
+
+#include "mlx4_en.h"
+
+MODULE_AUTHOR("Liran Liss, Yevgeny Petrilin");
+MODULE_DESCRIPTION("Mellanox ConnectX HCA Ethernet driver");
+MODULE_LICENSE("Dual BSD/GPL");
+MODULE_VERSION(DRV_VERSION);
+
+static const char mlx4_en_version[] =
+	DRV_NAME ": Mellanox ConnectX HCA Ethernet driver v"
+	DRV_VERSION "\n";
+
+#define MLX4_EN_PARM_INT(X, def_val, desc) \
+	static unsigned int X = def_val;\
+	module_param(X , uint, 0444); \
+	MODULE_PARM_DESC(X, desc);
+
+
+/*
+ * Device scope module parameters
+ */
+
+/* Enable RSS UDP traffic */
+MLX4_EN_PARM_INT(udp_rss, 1,
+		 "Enable RSS for incoming UDP traffic or disabled (0)");
+
+/* Priority pausing */
+MLX4_EN_PARM_INT(pfctx, 0, "Priority based Flow Control policy on TX[7:0]."
+			   " Per priority bit mask");
+MLX4_EN_PARM_INT(pfcrx, 0, "Priority based Flow Control policy on RX[7:0]."
+			   " Per priority bit mask");
+
+MLX4_EN_PARM_INT(inline_thold, MAX_INLINE,
+		 "Threshold for using inline data (range: 17-104, default: 104)");
+
+#define MAX_PFC_TX     0xff
+#define MAX_PFC_RX     0xff
+
+void en_print(const char *level, const struct mlx4_en_priv *priv,
+	      const char *format, ...)
+{
+	va_list args;
+	struct va_format vaf;
+
+	va_start(args, format);
+
+	vaf.fmt = format;
+	vaf.va = &args;
+	if (priv->registered)
+		printk("%s%s: %s: %pV",
+		       level, DRV_NAME, priv->dev->name, &vaf);
+	else
+		printk("%s%s: %s: Port %d: %pV",
+		       level, DRV_NAME, dev_name(&priv->mdev->pdev->dev),
+		       priv->port, &vaf);
+	va_end(args);
+}
+
+void mlx4_en_update_loopback_state(struct net_device *dev,
+				   netdev_features_t features)
+{
+	struct mlx4_en_priv *priv = netdev_priv(dev);
+
+	if (features & NETIF_F_LOOPBACK)
+		priv->ctrl_flags |= cpu_to_be32(MLX4_WQE_CTRL_FORCE_LOOPBACK);
+	else
+		priv->ctrl_flags &= cpu_to_be32(~MLX4_WQE_CTRL_FORCE_LOOPBACK);
+
+	priv->flags &= ~(MLX4_EN_FLAG_RX_FILTER_NEEDED|
+			MLX4_EN_FLAG_ENABLE_HW_LOOPBACK);
+
+	/* Drop the packet if SRIOV is not enabled
+	 * and not performing the selftest or flb disabled
+	 */
+	if (mlx4_is_mfunc(priv->mdev->dev) &&
+	    !(features & NETIF_F_LOOPBACK) && !priv->validate_loopback)
+		priv->flags |= MLX4_EN_FLAG_RX_FILTER_NEEDED;
+
+	/* Set dmac in Tx WQE if we are in SRIOV mode or if loopback selftest
+	 * is requested
+	 */
+	if (mlx4_is_mfunc(priv->mdev->dev) || priv->validate_loopback)
+		priv->flags |= MLX4_EN_FLAG_ENABLE_HW_LOOPBACK;
+
+	mutex_lock(&priv->mdev->state_lock);
+	if ((priv->mdev->dev->caps.flags2 &
+	     MLX4_DEV_CAP_FLAG2_UPDATE_QP_SRC_CHECK_LB) &&
+	    priv->rss_map.indir_qp && priv->rss_map.indir_qp->qpn) {
+		int i;
+		int err = 0;
+		int loopback = !!(features & NETIF_F_LOOPBACK);
+
+		for (i = 0; i < priv->rx_ring_num; i++) {
+			int ret;
+
+			ret = mlx4_en_change_mcast_lb(priv,
+						      &priv->rss_map.qps[i],
+						      loopback);
+			if (!err)
+				err = ret;
+		}
+		if (err)
+			mlx4_warn(priv->mdev, "failed to change mcast loopback\n");
+	}
+	mutex_unlock(&priv->mdev->state_lock);
+}
+
+static void mlx4_en_get_profile(struct mlx4_en_dev *mdev)
+{
+	struct mlx4_en_profile *params = &mdev->profile;
+	int i;
+
+	params->udp_rss = udp_rss;
+	params->max_num_tx_rings_p_up = mlx4_low_memory_profile() ?
+		MLX4_EN_MIN_TX_RING_P_UP :
+		min_t(int, num_online_cpus(), MLX4_EN_MAX_TX_RING_P_UP);
+
+	if (params->udp_rss && !(mdev->dev->caps.flags
+					& MLX4_DEV_CAP_FLAG_UDP_RSS)) {
+		mlx4_warn(mdev, "UDP RSS is not supported on this device\n");
+		params->udp_rss = 0;
+	}
+	for (i = 1; i <= MLX4_MAX_PORTS; i++) {
+		params->prof[i].rx_pause = !(pfcrx || pfctx);
+		params->prof[i].rx_ppp = pfcrx;
+		params->prof[i].tx_pause = !(pfcrx || pfctx);
+		params->prof[i].tx_ppp = pfctx;
+		params->prof[i].tx_ring_size = MLX4_EN_DEF_TX_RING_SIZE;
+		params->prof[i].rx_ring_size = MLX4_EN_DEF_RX_RING_SIZE;
+		params->prof[i].num_up = MLX4_EN_NUM_UP_LOW;
+		params->prof[i].num_tx_rings_p_up = params->max_num_tx_rings_p_up;
+		params->prof[i].tx_ring_num[TX] = params->max_num_tx_rings_p_up *
+			params->prof[i].num_up;
+		params->prof[i].rss_rings = 0;
+		params->prof[i].inline_thold = inline_thold;
+	}
+}
+
+static void *mlx4_en_get_netdev(struct mlx4_dev *dev, void *ctx, u8 port)
+{
+	struct mlx4_en_dev *endev = ctx;
+
+	return endev->pndev[port];
+}
+
+static void mlx4_en_event(struct mlx4_dev *dev, void *endev_ptr,
+			  enum mlx4_dev_event event, unsigned long port)
+{
+	struct mlx4_en_dev *mdev = (struct mlx4_en_dev *) endev_ptr;
+	struct mlx4_en_priv *priv;
+
+	switch (event) {
+	case MLX4_DEV_EVENT_PORT_UP:
+	case MLX4_DEV_EVENT_PORT_DOWN:
+		if (!mdev->pndev[port])
+			return;
+		priv = netdev_priv(mdev->pndev[port]);
+		/* To prevent races, we poll the link state in a separate
+		  task rather than changing it here */
+		priv->link_state = event;
+		queue_work(mdev->workqueue, &priv->linkstate_task);
+		break;
+
+	case MLX4_DEV_EVENT_CATASTROPHIC_ERROR:
+		mlx4_err(mdev, "Internal error detected, restarting device\n");
+		break;
+
+	case MLX4_DEV_EVENT_SLAVE_INIT:
+	case MLX4_DEV_EVENT_SLAVE_SHUTDOWN:
+		break;
+	default:
+		if (port < 1 || port > dev->caps.num_ports ||
+		    !mdev->pndev[port])
+			return;
+		mlx4_warn(mdev, "Unhandled event %d for port %d\n", event,
+			  (int) port);
+	}
+}
+
+static void mlx4_en_remove(struct mlx4_dev *dev, void *endev_ptr)
+{
+	struct mlx4_en_dev *mdev = endev_ptr;
+	int i;
+
+	mutex_lock(&mdev->state_lock);
+	mdev->device_up = false;
+	mutex_unlock(&mdev->state_lock);
+
+	mlx4_foreach_port(i, dev, MLX4_PORT_TYPE_ETH)
+		if (mdev->pndev[i])
+			mlx4_en_destroy_netdev(mdev->pndev[i]);
+
+	flush_workqueue(mdev->workqueue);
+	destroy_workqueue(mdev->workqueue);
+	(void) mlx4_mr_free(dev, &mdev->mr);
+	iounmap(mdev->uar_map);
+	mlx4_uar_free(dev, &mdev->priv_uar);
+	mlx4_pd_free(dev, mdev->priv_pdn);
+	if (mdev->nb.notifier_call)
+		unregister_netdevice_notifier(&mdev->nb);
+	kfree(mdev);
+}
+
+static void mlx4_en_activate(struct mlx4_dev *dev, void *ctx)
+{
+	int i;
+	struct mlx4_en_dev *mdev = ctx;
+
+	/* Create a netdev for each port */
+	mlx4_foreach_port(i, dev, MLX4_PORT_TYPE_ETH) {
+		mlx4_info(mdev, "Activating port:%d\n", i);
+		if (mlx4_en_init_netdev(mdev, i, &mdev->profile.prof[i]))
+			mdev->pndev[i] = NULL;
+	}
+
+	/* register notifier */
+	mdev->nb.notifier_call = mlx4_en_netdev_event;
+	if (register_netdevice_notifier(&mdev->nb)) {
+		mdev->nb.notifier_call = NULL;
+		mlx4_err(mdev, "Failed to create notifier\n");
+	}
+}
+
+static void *mlx4_en_add(struct mlx4_dev *dev)
+{
+	struct mlx4_en_dev *mdev;
+	int i;
+
+	printk_once(KERN_INFO "%s", mlx4_en_version);
+
+	mdev = kzalloc(sizeof(*mdev), GFP_KERNEL);
+	if (!mdev)
+		goto err_free_res;
+
+	if (mlx4_pd_alloc(dev, &mdev->priv_pdn))
+		goto err_free_dev;
+
+	if (mlx4_uar_alloc(dev, &mdev->priv_uar))
+		goto err_pd;
+
+	mdev->uar_map = ioremap((phys_addr_t) mdev->priv_uar.pfn << PAGE_SHIFT,
+				PAGE_SIZE);
+	if (!mdev->uar_map)
+		goto err_uar;
+	spin_lock_init(&mdev->uar_lock);
+
+	mdev->dev = dev;
+	mdev->dma_device = &dev->persist->pdev->dev;
+	mdev->pdev = dev->persist->pdev;
+	mdev->device_up = false;
+
+	mdev->LSO_support = !!(dev->caps.flags & (1 << 15));
+	if (!mdev->LSO_support)
+		mlx4_warn(mdev, "LSO not supported, please upgrade to later FW version to enable LSO\n");
+
+	if (mlx4_mr_alloc(mdev->dev, mdev->priv_pdn, 0, ~0ull,
+			 MLX4_PERM_LOCAL_WRITE |  MLX4_PERM_LOCAL_READ,
+			 0, 0, &mdev->mr)) {
+		mlx4_err(mdev, "Failed allocating memory region\n");
+		goto err_map;
+	}
+	if (mlx4_mr_enable(mdev->dev, &mdev->mr)) {
+		mlx4_err(mdev, "Failed enabling memory region\n");
+		goto err_mr;
+	}
+
+	/* Build device profile according to supplied module parameters */
+	mlx4_en_get_profile(mdev);
+
+	/* Configure which ports to start according to module parameters */
+	mdev->port_cnt = 0;
+	mlx4_foreach_port(i, dev, MLX4_PORT_TYPE_ETH)
+		mdev->port_cnt++;
+
+	/* Set default number of RX rings*/
+	mlx4_en_set_num_rx_rings(mdev);
+
+	/* Create our own workqueue for reset/multicast tasks
+	 * Note: we cannot use the shared workqueue because of deadlocks caused
+	 *       by the rtnl lock */
+	mdev->workqueue = create_singlethread_workqueue("mlx4_en");
+	if (!mdev->workqueue)
+		goto err_mr;
+
+	/* At this stage all non-port specific tasks are complete:
+	 * mark the card state as up */
+	mutex_init(&mdev->state_lock);
+	mdev->device_up = true;
+
+	return mdev;
+
+err_mr:
+	(void) mlx4_mr_free(dev, &mdev->mr);
+err_map:
+	if (mdev->uar_map)
+		iounmap(mdev->uar_map);
+err_uar:
+	mlx4_uar_free(dev, &mdev->priv_uar);
+err_pd:
+	mlx4_pd_free(dev, mdev->priv_pdn);
+err_free_dev:
+	kfree(mdev);
+err_free_res:
+	return NULL;
+}
+
+static struct mlx4_interface mlx4_en_interface = {
+	.add		= mlx4_en_add,
+	.remove		= mlx4_en_remove,
+	.event		= mlx4_en_event,
+	.get_dev	= mlx4_en_get_netdev,
+	.protocol	= MLX4_PROT_ETH,
+	.activate	= mlx4_en_activate,
+};
+
+static void mlx4_en_verify_params(void)
+{
+	if (pfctx > MAX_PFC_TX) {
+		pr_warn("mlx4_en: WARNING: illegal module parameter pfctx 0x%x - should be in range 0-0x%x, will be changed to default (0)\n",
+			pfctx, MAX_PFC_TX);
+		pfctx = 0;
+	}
+
+	if (pfcrx > MAX_PFC_RX) {
+		pr_warn("mlx4_en: WARNING: illegal module parameter pfcrx 0x%x - should be in range 0-0x%x, will be changed to default (0)\n",
+			pfcrx, MAX_PFC_RX);
+		pfcrx = 0;
+	}
+
+	if (inline_thold < MIN_PKT_LEN || inline_thold > MAX_INLINE) {
+		pr_warn("mlx4_en: WARNING: illegal module parameter inline_thold %d - should be in range %d-%d, will be changed to default (%d)\n",
+			inline_thold, MIN_PKT_LEN, MAX_INLINE, MAX_INLINE);
+		inline_thold = MAX_INLINE;
+	}
+}
+
+static int __init mlx4_en_init(void)
+{
+	mlx4_en_verify_params();
+	mlx4_en_init_ptys2ethtool_map();
+
+	return mlx4_register_interface(&mlx4_en_interface);
+}
+
+static void __exit mlx4_en_cleanup(void)
+{
+	mlx4_unregister_interface(&mlx4_en_interface);
+}
+
+module_init(mlx4_en_init);
+module_exit(mlx4_en_cleanup);
+
diff --git a/drivers/net/ethernet/mellanox/mlx4/en_netdev.c b/drivers/net/ethernet/mellanox/mlx4/en_netdev.c
new file mode 100644
index 000000000..14d46afd2
--- /dev/null
+++ b/drivers/net/ethernet/mellanox/mlx4/en_netdev.c
@@ -0,0 +1,3686 @@
+/*
+ * Copyright (c) 2007 Mellanox Technologies. All rights reserved.
+ *
+ * This software is available to you under a choice of one of two
+ * licenses.  You may choose to be licensed under the terms of the GNU
+ * General Public License (GPL) Version 2, available from the file
+ * COPYING in the main directory of this source tree, or the
+ * OpenIB.org BSD license below:
+ *
+ *     Redistribution and use in source and binary forms, with or
+ *     without modification, are permitted provided that the following
+ *     conditions are met:
+ *
+ *      - Redistributions of source code must retain the above
+ *        copyright notice, this list of conditions and the following
+ *        disclaimer.
+ *
+ *      - Redistributions in binary form must reproduce the above
+ *        copyright notice, this list of conditions and the following
+ *        disclaimer in the documentation and/or other materials
+ *        provided with the distribution.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
+ * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
+ * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS
+ * BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN
+ * ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
+ * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+ * SOFTWARE.
+ *
+ */
+
+#include <linux/bpf.h>
+#include <linux/etherdevice.h>
+#include <linux/tcp.h>
+#include <linux/if_vlan.h>
+#include <linux/delay.h>
+#include <linux/slab.h>
+#include <linux/hash.h>
+#include <net/ip.h>
+#include <net/busy_poll.h>
+#include <net/vxlan.h>
+#include <net/devlink.h>
+
+#include <linux/mlx4/driver.h>
+#include <linux/mlx4/device.h>
+#include <linux/mlx4/cmd.h>
+#include <linux/mlx4/cq.h>
+
+#include "mlx4_en.h"
+#include "en_port.h"
+
+#define MLX4_EN_MAX_XDP_MTU ((int)(PAGE_SIZE - ETH_HLEN - (2 * VLAN_HLEN) - \
+				   XDP_PACKET_HEADROOM))
+
+int mlx4_en_setup_tc(struct net_device *dev, u8 up)
+{
+	struct mlx4_en_priv *priv = netdev_priv(dev);
+	int i;
+	unsigned int offset = 0;
+
+	if (up && up != MLX4_EN_NUM_UP_HIGH)
+		return -EINVAL;
+
+	netdev_set_num_tc(dev, up);
+	netif_set_real_num_tx_queues(dev, priv->tx_ring_num[TX]);
+	/* Partition Tx queues evenly amongst UP's */
+	for (i = 0; i < up; i++) {
+		netdev_set_tc_queue(dev, i, priv->num_tx_rings_p_up, offset);
+		offset += priv->num_tx_rings_p_up;
+	}
+
+#ifdef CONFIG_MLX4_EN_DCB
+	if (!mlx4_is_slave(priv->mdev->dev)) {
+		if (up) {
+			if (priv->dcbx_cap)
+				priv->flags |= MLX4_EN_FLAG_DCB_ENABLED;
+		} else {
+			priv->flags &= ~MLX4_EN_FLAG_DCB_ENABLED;
+			priv->cee_config.pfc_state = false;
+		}
+	}
+#endif /* CONFIG_MLX4_EN_DCB */
+
+	return 0;
+}
+
+int mlx4_en_alloc_tx_queue_per_tc(struct net_device *dev, u8 tc)
+{
+	struct mlx4_en_priv *priv = netdev_priv(dev);
+	struct mlx4_en_dev *mdev = priv->mdev;
+	struct mlx4_en_port_profile new_prof;
+	struct mlx4_en_priv *tmp;
+	int total_count;
+	int port_up = 0;
+	int err = 0;
+
+	tmp = kzalloc(sizeof(*tmp), GFP_KERNEL);
+	if (!tmp)
+		return -ENOMEM;
+
+	mutex_lock(&mdev->state_lock);
+	memcpy(&new_prof, priv->prof, sizeof(struct mlx4_en_port_profile));
+	new_prof.num_up = (tc == 0) ? MLX4_EN_NUM_UP_LOW :
+				      MLX4_EN_NUM_UP_HIGH;
+	new_prof.tx_ring_num[TX] = new_prof.num_tx_rings_p_up *
+				   new_prof.num_up;
+	total_count = new_prof.tx_ring_num[TX] + new_prof.tx_ring_num[TX_XDP];
+	if (total_count > MAX_TX_RINGS) {
+		err = -EINVAL;
+		en_err(priv,
+		       "Total number of TX and XDP rings (%d) exceeds the maximum supported (%d)\n",
+		       total_count, MAX_TX_RINGS);
+		goto out;
+	}
+	err = mlx4_en_try_alloc_resources(priv, tmp, &new_prof, true);
+	if (err)
+		goto out;
+
+	if (priv->port_up) {
+		port_up = 1;
+		mlx4_en_stop_port(dev, 1);
+	}
+
+	mlx4_en_safe_replace_resources(priv, tmp);
+	if (port_up) {
+		err = mlx4_en_start_port(dev);
+		if (err) {
+			en_err(priv, "Failed starting port for setup TC\n");
+			goto out;
+		}
+	}
+
+	err = mlx4_en_setup_tc(dev, tc);
+out:
+	mutex_unlock(&mdev->state_lock);
+	kfree(tmp);
+	return err;
+}
+
+static int __mlx4_en_setup_tc(struct net_device *dev, enum tc_setup_type type,
+			      void *type_data)
+{
+	struct tc_mqprio_qopt *mqprio = type_data;
+
+	if (type != TC_SETUP_QDISC_MQPRIO)
+		return -EOPNOTSUPP;
+
+	if (mqprio->num_tc && mqprio->num_tc != MLX4_EN_NUM_UP_HIGH)
+		return -EINVAL;
+
+	mqprio->hw = TC_MQPRIO_HW_OFFLOAD_TCS;
+
+	return mlx4_en_alloc_tx_queue_per_tc(dev, mqprio->num_tc);
+}
+
+#ifdef CONFIG_RFS_ACCEL
+
+struct mlx4_en_filter {
+	struct list_head next;
+	struct work_struct work;
+
+	u8     ip_proto;
+	__be32 src_ip;
+	__be32 dst_ip;
+	__be16 src_port;
+	__be16 dst_port;
+
+	int rxq_index;
+	struct mlx4_en_priv *priv;
+	u32 flow_id;			/* RFS infrastructure id */
+	int id;				/* mlx4_en driver id */
+	u64 reg_id;			/* Flow steering API id */
+	u8 activated;			/* Used to prevent expiry before filter
+					 * is attached
+					 */
+	struct hlist_node filter_chain;
+};
+
+static void mlx4_en_filter_rfs_expire(struct mlx4_en_priv *priv);
+
+static enum mlx4_net_trans_rule_id mlx4_ip_proto_to_trans_rule_id(u8 ip_proto)
+{
+	switch (ip_proto) {
+	case IPPROTO_UDP:
+		return MLX4_NET_TRANS_RULE_ID_UDP;
+	case IPPROTO_TCP:
+		return MLX4_NET_TRANS_RULE_ID_TCP;
+	default:
+		return MLX4_NET_TRANS_RULE_NUM;
+	}
+};
+
+/* Must not acquire state_lock, as its corresponding work_sync
+ * is done under it.
+ */
+static void mlx4_en_filter_work(struct work_struct *work)
+{
+	struct mlx4_en_filter *filter = container_of(work,
+						     struct mlx4_en_filter,
+						     work);
+	struct mlx4_en_priv *priv = filter->priv;
+	struct mlx4_spec_list spec_tcp_udp = {
+		.id = mlx4_ip_proto_to_trans_rule_id(filter->ip_proto),
+		{
+			.tcp_udp = {
+				.dst_port = filter->dst_port,
+				.dst_port_msk = (__force __be16)-1,
+				.src_port = filter->src_port,
+				.src_port_msk = (__force __be16)-1,
+			},
+		},
+	};
+	struct mlx4_spec_list spec_ip = {
+		.id = MLX4_NET_TRANS_RULE_ID_IPV4,
+		{
+			.ipv4 = {
+				.dst_ip = filter->dst_ip,
+				.dst_ip_msk = (__force __be32)-1,
+				.src_ip = filter->src_ip,
+				.src_ip_msk = (__force __be32)-1,
+			},
+		},
+	};
+	struct mlx4_spec_list spec_eth = {
+		.id = MLX4_NET_TRANS_RULE_ID_ETH,
+	};
+	struct mlx4_net_trans_rule rule = {
+		.list = LIST_HEAD_INIT(rule.list),
+		.queue_mode = MLX4_NET_TRANS_Q_LIFO,
+		.exclusive = 1,
+		.allow_loopback = 1,
+		.promisc_mode = MLX4_FS_REGULAR,
+		.port = priv->port,
+		.priority = MLX4_DOMAIN_RFS,
+	};
+	int rc;
+	__be64 mac_mask = cpu_to_be64(MLX4_MAC_MASK << 16);
+
+	if (spec_tcp_udp.id >= MLX4_NET_TRANS_RULE_NUM) {
+		en_warn(priv, "RFS: ignoring unsupported ip protocol (%d)\n",
+			filter->ip_proto);
+		goto ignore;
+	}
+	list_add_tail(&spec_eth.list, &rule.list);
+	list_add_tail(&spec_ip.list, &rule.list);
+	list_add_tail(&spec_tcp_udp.list, &rule.list);
+
+	rule.qpn = priv->rss_map.qps[filter->rxq_index].qpn;
+	memcpy(spec_eth.eth.dst_mac, priv->dev->dev_addr, ETH_ALEN);
+	memcpy(spec_eth.eth.dst_mac_msk, &mac_mask, ETH_ALEN);
+
+	filter->activated = 0;
+
+	if (filter->reg_id) {
+		rc = mlx4_flow_detach(priv->mdev->dev, filter->reg_id);
+		if (rc && rc != -ENOENT)
+			en_err(priv, "Error detaching flow. rc = %d\n", rc);
+	}
+
+	rc = mlx4_flow_attach(priv->mdev->dev, &rule, &filter->reg_id);
+	if (rc)
+		en_err(priv, "Error attaching flow. err = %d\n", rc);
+
+ignore:
+	mlx4_en_filter_rfs_expire(priv);
+
+	filter->activated = 1;
+}
+
+static inline struct hlist_head *
+filter_hash_bucket(struct mlx4_en_priv *priv, __be32 src_ip, __be32 dst_ip,
+		   __be16 src_port, __be16 dst_port)
+{
+	unsigned long l;
+	int bucket_idx;
+
+	l = (__force unsigned long)src_port |
+	    ((__force unsigned long)dst_port << 2);
+	l ^= (__force unsigned long)(src_ip ^ dst_ip);
+
+	bucket_idx = hash_long(l, MLX4_EN_FILTER_HASH_SHIFT);
+
+	return &priv->filter_hash[bucket_idx];
+}
+
+static struct mlx4_en_filter *
+mlx4_en_filter_alloc(struct mlx4_en_priv *priv, int rxq_index, __be32 src_ip,
+		     __be32 dst_ip, u8 ip_proto, __be16 src_port,
+		     __be16 dst_port, u32 flow_id)
+{
+	struct mlx4_en_filter *filter = NULL;
+
+	filter = kzalloc(sizeof(struct mlx4_en_filter), GFP_ATOMIC);
+	if (!filter)
+		return NULL;
+
+	filter->priv = priv;
+	filter->rxq_index = rxq_index;
+	INIT_WORK(&filter->work, mlx4_en_filter_work);
+
+	filter->src_ip = src_ip;
+	filter->dst_ip = dst_ip;
+	filter->ip_proto = ip_proto;
+	filter->src_port = src_port;
+	filter->dst_port = dst_port;
+
+	filter->flow_id = flow_id;
+
+	filter->id = priv->last_filter_id++ % RPS_NO_FILTER;
+
+	list_add_tail(&filter->next, &priv->filters);
+	hlist_add_head(&filter->filter_chain,
+		       filter_hash_bucket(priv, src_ip, dst_ip, src_port,
+					  dst_port));
+
+	return filter;
+}
+
+static void mlx4_en_filter_free(struct mlx4_en_filter *filter)
+{
+	struct mlx4_en_priv *priv = filter->priv;
+	int rc;
+
+	list_del(&filter->next);
+
+	rc = mlx4_flow_detach(priv->mdev->dev, filter->reg_id);
+	if (rc && rc != -ENOENT)
+		en_err(priv, "Error detaching flow. rc = %d\n", rc);
+
+	kfree(filter);
+}
+
+static inline struct mlx4_en_filter *
+mlx4_en_filter_find(struct mlx4_en_priv *priv, __be32 src_ip, __be32 dst_ip,
+		    u8 ip_proto, __be16 src_port, __be16 dst_port)
+{
+	struct mlx4_en_filter *filter;
+	struct mlx4_en_filter *ret = NULL;
+
+	hlist_for_each_entry(filter,
+			     filter_hash_bucket(priv, src_ip, dst_ip,
+						src_port, dst_port),
+			     filter_chain) {
+		if (filter->src_ip == src_ip &&
+		    filter->dst_ip == dst_ip &&
+		    filter->ip_proto == ip_proto &&
+		    filter->src_port == src_port &&
+		    filter->dst_port == dst_port) {
+			ret = filter;
+			break;
+		}
+	}
+
+	return ret;
+}
+
+static int
+mlx4_en_filter_rfs(struct net_device *net_dev, const struct sk_buff *skb,
+		   u16 rxq_index, u32 flow_id)
+{
+	struct mlx4_en_priv *priv = netdev_priv(net_dev);
+	struct mlx4_en_filter *filter;
+	const struct iphdr *ip;
+	const __be16 *ports;
+	u8 ip_proto;
+	__be32 src_ip;
+	__be32 dst_ip;
+	__be16 src_port;
+	__be16 dst_port;
+	int nhoff = skb_network_offset(skb);
+	int ret = 0;
+
+	if (skb->encapsulation)
+		return -EPROTONOSUPPORT;
+
+	if (skb->protocol != htons(ETH_P_IP))
+		return -EPROTONOSUPPORT;
+
+	ip = (const struct iphdr *)(skb->data + nhoff);
+	if (ip_is_fragment(ip))
+		return -EPROTONOSUPPORT;
+
+	if ((ip->protocol != IPPROTO_TCP) && (ip->protocol != IPPROTO_UDP))
+		return -EPROTONOSUPPORT;
+	ports = (const __be16 *)(skb->data + nhoff + 4 * ip->ihl);
+
+	ip_proto = ip->protocol;
+	src_ip = ip->saddr;
+	dst_ip = ip->daddr;
+	src_port = ports[0];
+	dst_port = ports[1];
+
+	spin_lock_bh(&priv->filters_lock);
+	filter = mlx4_en_filter_find(priv, src_ip, dst_ip, ip_proto,
+				     src_port, dst_port);
+	if (filter) {
+		if (filter->rxq_index == rxq_index)
+			goto out;
+
+		filter->rxq_index = rxq_index;
+	} else {
+		filter = mlx4_en_filter_alloc(priv, rxq_index,
+					      src_ip, dst_ip, ip_proto,
+					      src_port, dst_port, flow_id);
+		if (!filter) {
+			ret = -ENOMEM;
+			goto err;
+		}
+	}
+
+	queue_work(priv->mdev->workqueue, &filter->work);
+
+out:
+	ret = filter->id;
+err:
+	spin_unlock_bh(&priv->filters_lock);
+
+	return ret;
+}
+
+void mlx4_en_cleanup_filters(struct mlx4_en_priv *priv)
+{
+	struct mlx4_en_filter *filter, *tmp;
+	LIST_HEAD(del_list);
+
+	spin_lock_bh(&priv->filters_lock);
+	list_for_each_entry_safe(filter, tmp, &priv->filters, next) {
+		list_move(&filter->next, &del_list);
+		hlist_del(&filter->filter_chain);
+	}
+	spin_unlock_bh(&priv->filters_lock);
+
+	list_for_each_entry_safe(filter, tmp, &del_list, next) {
+		cancel_work_sync(&filter->work);
+		mlx4_en_filter_free(filter);
+	}
+}
+
+static void mlx4_en_filter_rfs_expire(struct mlx4_en_priv *priv)
+{
+	struct mlx4_en_filter *filter = NULL, *tmp, *last_filter = NULL;
+	LIST_HEAD(del_list);
+	int i = 0;
+
+	spin_lock_bh(&priv->filters_lock);
+	list_for_each_entry_safe(filter, tmp, &priv->filters, next) {
+		if (i > MLX4_EN_FILTER_EXPIRY_QUOTA)
+			break;
+
+		if (filter->activated &&
+		    !work_pending(&filter->work) &&
+		    rps_may_expire_flow(priv->dev,
+					filter->rxq_index, filter->flow_id,
+					filter->id)) {
+			list_move(&filter->next, &del_list);
+			hlist_del(&filter->filter_chain);
+		} else
+			last_filter = filter;
+
+		i++;
+	}
+
+	if (last_filter && (&last_filter->next != priv->filters.next))
+		list_move(&priv->filters, &last_filter->next);
+
+	spin_unlock_bh(&priv->filters_lock);
+
+	list_for_each_entry_safe(filter, tmp, &del_list, next)
+		mlx4_en_filter_free(filter);
+}
+#endif
+
+static int mlx4_en_vlan_rx_add_vid(struct net_device *dev,
+				   __be16 proto, u16 vid)
+{
+	struct mlx4_en_priv *priv = netdev_priv(dev);
+	struct mlx4_en_dev *mdev = priv->mdev;
+	int err;
+	int idx;
+
+	en_dbg(HW, priv, "adding VLAN:%d\n", vid);
+
+	set_bit(vid, priv->active_vlans);
+
+	/* Add VID to port VLAN filter */
+	mutex_lock(&mdev->state_lock);
+	if (mdev->device_up && priv->port_up) {
+		err = mlx4_SET_VLAN_FLTR(mdev->dev, priv);
+		if (err) {
+			en_err(priv, "Failed configuring VLAN filter\n");
+			goto out;
+		}
+	}
+	err = mlx4_register_vlan(mdev->dev, priv->port, vid, &idx);
+	if (err)
+		en_dbg(HW, priv, "Failed adding vlan %d\n", vid);
+
+out:
+	mutex_unlock(&mdev->state_lock);
+	return err;
+}
+
+static int mlx4_en_vlan_rx_kill_vid(struct net_device *dev,
+				    __be16 proto, u16 vid)
+{
+	struct mlx4_en_priv *priv = netdev_priv(dev);
+	struct mlx4_en_dev *mdev = priv->mdev;
+	int err = 0;
+
+	en_dbg(HW, priv, "Killing VID:%d\n", vid);
+
+	clear_bit(vid, priv->active_vlans);
+
+	/* Remove VID from port VLAN filter */
+	mutex_lock(&mdev->state_lock);
+	mlx4_unregister_vlan(mdev->dev, priv->port, vid);
+
+	if (mdev->device_up && priv->port_up) {
+		err = mlx4_SET_VLAN_FLTR(mdev->dev, priv);
+		if (err)
+			en_err(priv, "Failed configuring VLAN filter\n");
+	}
+	mutex_unlock(&mdev->state_lock);
+
+	return err;
+}
+
+static void mlx4_en_u64_to_mac(unsigned char dst_mac[ETH_ALEN + 2], u64 src_mac)
+{
+	int i;
+	for (i = ETH_ALEN - 1; i >= 0; --i) {
+		dst_mac[i] = src_mac & 0xff;
+		src_mac >>= 8;
+	}
+	memset(&dst_mac[ETH_ALEN], 0, 2);
+}
+
+
+static int mlx4_en_tunnel_steer_add(struct mlx4_en_priv *priv, unsigned char *addr,
+				    int qpn, u64 *reg_id)
+{
+	int err;
+
+	if (priv->mdev->dev->caps.tunnel_offload_mode != MLX4_TUNNEL_OFFLOAD_MODE_VXLAN ||
+	    priv->mdev->dev->caps.dmfs_high_steer_mode == MLX4_STEERING_DMFS_A0_STATIC)
+		return 0; /* do nothing */
+
+	err = mlx4_tunnel_steer_add(priv->mdev->dev, addr, priv->port, qpn,
+				    MLX4_DOMAIN_NIC, reg_id);
+	if (err) {
+		en_err(priv, "failed to add vxlan steering rule, err %d\n", err);
+		return err;
+	}
+	en_dbg(DRV, priv, "added vxlan steering rule, mac %pM reg_id %llx\n", addr, *reg_id);
+	return 0;
+}
+
+
+static int mlx4_en_uc_steer_add(struct mlx4_en_priv *priv,
+				unsigned char *mac, int *qpn, u64 *reg_id)
+{
+	struct mlx4_en_dev *mdev = priv->mdev;
+	struct mlx4_dev *dev = mdev->dev;
+	int err;
+
+	switch (dev->caps.steering_mode) {
+	case MLX4_STEERING_MODE_B0: {
+		struct mlx4_qp qp;
+		u8 gid[16] = {0};
+
+		qp.qpn = *qpn;
+		memcpy(&gid[10], mac, ETH_ALEN);
+		gid[5] = priv->port;
+
+		err = mlx4_unicast_attach(dev, &qp, gid, 0, MLX4_PROT_ETH);
+		break;
+	}
+	case MLX4_STEERING_MODE_DEVICE_MANAGED: {
+		struct mlx4_spec_list spec_eth = { {NULL} };
+		__be64 mac_mask = cpu_to_be64(MLX4_MAC_MASK << 16);
+
+		struct mlx4_net_trans_rule rule = {
+			.queue_mode = MLX4_NET_TRANS_Q_FIFO,
+			.exclusive = 0,
+			.allow_loopback = 1,
+			.promisc_mode = MLX4_FS_REGULAR,
+			.priority = MLX4_DOMAIN_NIC,
+		};
+
+		rule.port = priv->port;
+		rule.qpn = *qpn;
+		INIT_LIST_HEAD(&rule.list);
+
+		spec_eth.id = MLX4_NET_TRANS_RULE_ID_ETH;
+		memcpy(spec_eth.eth.dst_mac, mac, ETH_ALEN);
+		memcpy(spec_eth.eth.dst_mac_msk, &mac_mask, ETH_ALEN);
+		list_add_tail(&spec_eth.list, &rule.list);
+
+		err = mlx4_flow_attach(dev, &rule, reg_id);
+		break;
+	}
+	default:
+		return -EINVAL;
+	}
+	if (err)
+		en_warn(priv, "Failed Attaching Unicast\n");
+
+	return err;
+}
+
+static void mlx4_en_uc_steer_release(struct mlx4_en_priv *priv,
+				     unsigned char *mac, int qpn, u64 reg_id)
+{
+	struct mlx4_en_dev *mdev = priv->mdev;
+	struct mlx4_dev *dev = mdev->dev;
+
+	switch (dev->caps.steering_mode) {
+	case MLX4_STEERING_MODE_B0: {
+		struct mlx4_qp qp;
+		u8 gid[16] = {0};
+
+		qp.qpn = qpn;
+		memcpy(&gid[10], mac, ETH_ALEN);
+		gid[5] = priv->port;
+
+		mlx4_unicast_detach(dev, &qp, gid, MLX4_PROT_ETH);
+		break;
+	}
+	case MLX4_STEERING_MODE_DEVICE_MANAGED: {
+		mlx4_flow_detach(dev, reg_id);
+		break;
+	}
+	default:
+		en_err(priv, "Invalid steering mode.\n");
+	}
+}
+
+static int mlx4_en_get_qp(struct mlx4_en_priv *priv)
+{
+	struct mlx4_en_dev *mdev = priv->mdev;
+	struct mlx4_dev *dev = mdev->dev;
+	int index = 0;
+	int err = 0;
+	int *qpn = &priv->base_qpn;
+	u64 mac = mlx4_mac_to_u64(priv->dev->dev_addr);
+
+	en_dbg(DRV, priv, "Registering MAC: %pM for adding\n",
+	       priv->dev->dev_addr);
+	index = mlx4_register_mac(dev, priv->port, mac);
+	if (index < 0) {
+		err = index;
+		en_err(priv, "Failed adding MAC: %pM\n",
+		       priv->dev->dev_addr);
+		return err;
+	}
+
+	en_info(priv, "Steering Mode %d\n", dev->caps.steering_mode);
+
+	if (dev->caps.steering_mode == MLX4_STEERING_MODE_A0) {
+		int base_qpn = mlx4_get_base_qpn(dev, priv->port);
+		*qpn = base_qpn + index;
+		return 0;
+	}
+
+	err = mlx4_qp_reserve_range(dev, 1, 1, qpn, MLX4_RESERVE_A0_QP,
+				    MLX4_RES_USAGE_DRIVER);
+	en_dbg(DRV, priv, "Reserved qp %d\n", *qpn);
+	if (err) {
+		en_err(priv, "Failed to reserve qp for mac registration\n");
+		mlx4_unregister_mac(dev, priv->port, mac);
+		return err;
+	}
+
+	return 0;
+}
+
+static void mlx4_en_put_qp(struct mlx4_en_priv *priv)
+{
+	struct mlx4_en_dev *mdev = priv->mdev;
+	struct mlx4_dev *dev = mdev->dev;
+	int qpn = priv->base_qpn;
+
+	if (dev->caps.steering_mode == MLX4_STEERING_MODE_A0) {
+		u64 mac = mlx4_mac_to_u64(priv->dev->dev_addr);
+		en_dbg(DRV, priv, "Registering MAC: %pM for deleting\n",
+		       priv->dev->dev_addr);
+		mlx4_unregister_mac(dev, priv->port, mac);
+	} else {
+		en_dbg(DRV, priv, "Releasing qp: port %d, qpn %d\n",
+		       priv->port, qpn);
+		mlx4_qp_release_range(dev, qpn, 1);
+		priv->flags &= ~MLX4_EN_FLAG_FORCE_PROMISC;
+	}
+}
+
+static int mlx4_en_replace_mac(struct mlx4_en_priv *priv, int qpn,
+			       unsigned char *new_mac, unsigned char *prev_mac)
+{
+	struct mlx4_en_dev *mdev = priv->mdev;
+	struct mlx4_dev *dev = mdev->dev;
+	int err = 0;
+	u64 new_mac_u64 = mlx4_mac_to_u64(new_mac);
+
+	if (dev->caps.steering_mode != MLX4_STEERING_MODE_A0) {
+		struct hlist_head *bucket;
+		unsigned int mac_hash;
+		struct mlx4_mac_entry *entry;
+		struct hlist_node *tmp;
+		u64 prev_mac_u64 = mlx4_mac_to_u64(prev_mac);
+
+		bucket = &priv->mac_hash[prev_mac[MLX4_EN_MAC_HASH_IDX]];
+		hlist_for_each_entry_safe(entry, tmp, bucket, hlist) {
+			if (ether_addr_equal_64bits(entry->mac, prev_mac)) {
+				mlx4_en_uc_steer_release(priv, entry->mac,
+							 qpn, entry->reg_id);
+				mlx4_unregister_mac(dev, priv->port,
+						    prev_mac_u64);
+				hlist_del_rcu(&entry->hlist);
+				synchronize_rcu();
+				memcpy(entry->mac, new_mac, ETH_ALEN);
+				entry->reg_id = 0;
+				mac_hash = new_mac[MLX4_EN_MAC_HASH_IDX];
+				hlist_add_head_rcu(&entry->hlist,
+						   &priv->mac_hash[mac_hash]);
+				mlx4_register_mac(dev, priv->port, new_mac_u64);
+				err = mlx4_en_uc_steer_add(priv, new_mac,
+							   &qpn,
+							   &entry->reg_id);
+				if (err)
+					return err;
+				if (priv->tunnel_reg_id) {
+					mlx4_flow_detach(priv->mdev->dev, priv->tunnel_reg_id);
+					priv->tunnel_reg_id = 0;
+				}
+				err = mlx4_en_tunnel_steer_add(priv, new_mac, qpn,
+							       &priv->tunnel_reg_id);
+				return err;
+			}
+		}
+		return -EINVAL;
+	}
+
+	return __mlx4_replace_mac(dev, priv->port, qpn, new_mac_u64);
+}
+
+static void mlx4_en_update_user_mac(struct mlx4_en_priv *priv,
+				    unsigned char new_mac[ETH_ALEN + 2])
+{
+	struct mlx4_en_dev *mdev = priv->mdev;
+	int err;
+
+	if (!(mdev->dev->caps.flags2 & MLX4_DEV_CAP_FLAG2_USER_MAC_EN))
+		return;
+
+	err = mlx4_SET_PORT_user_mac(mdev->dev, priv->port, new_mac);
+	if (err)
+		en_err(priv, "Failed to pass user MAC(%pM) to Firmware for port %d, with error %d\n",
+		       new_mac, priv->port, err);
+}
+
+static int mlx4_en_do_set_mac(struct mlx4_en_priv *priv,
+			      unsigned char new_mac[ETH_ALEN + 2])
+{
+	int err = 0;
+
+	if (priv->port_up) {
+		/* Remove old MAC and insert the new one */
+		err = mlx4_en_replace_mac(priv, priv->base_qpn,
+					  new_mac, priv->current_mac);
+		if (err)
+			en_err(priv, "Failed changing HW MAC address\n");
+	} else
+		en_dbg(HW, priv, "Port is down while registering mac, exiting...\n");
+
+	if (!err)
+		memcpy(priv->current_mac, new_mac, sizeof(priv->current_mac));
+
+	return err;
+}
+
+static int mlx4_en_set_mac(struct net_device *dev, void *addr)
+{
+	struct mlx4_en_priv *priv = netdev_priv(dev);
+	struct mlx4_en_dev *mdev = priv->mdev;
+	struct sockaddr *saddr = addr;
+	unsigned char new_mac[ETH_ALEN + 2];
+	int err;
+
+	if (!is_valid_ether_addr(saddr->sa_data))
+		return -EADDRNOTAVAIL;
+
+	mutex_lock(&mdev->state_lock);
+	memcpy(new_mac, saddr->sa_data, ETH_ALEN);
+	err = mlx4_en_do_set_mac(priv, new_mac);
+	if (err)
+		goto out;
+
+	memcpy(dev->dev_addr, saddr->sa_data, ETH_ALEN);
+	mlx4_en_update_user_mac(priv, new_mac);
+out:
+	mutex_unlock(&mdev->state_lock);
+
+	return err;
+}
+
+static void mlx4_en_clear_list(struct net_device *dev)
+{
+	struct mlx4_en_priv *priv = netdev_priv(dev);
+	struct mlx4_en_mc_list *tmp, *mc_to_del;
+
+	list_for_each_entry_safe(mc_to_del, tmp, &priv->mc_list, list) {
+		list_del(&mc_to_del->list);
+		kfree(mc_to_del);
+	}
+}
+
+static void mlx4_en_cache_mclist(struct net_device *dev)
+{
+	struct mlx4_en_priv *priv = netdev_priv(dev);
+	struct netdev_hw_addr *ha;
+	struct mlx4_en_mc_list *tmp;
+
+	mlx4_en_clear_list(dev);
+	netdev_for_each_mc_addr(ha, dev) {
+		tmp = kzalloc(sizeof(struct mlx4_en_mc_list), GFP_ATOMIC);
+		if (!tmp) {
+			mlx4_en_clear_list(dev);
+			return;
+		}
+		memcpy(tmp->addr, ha->addr, ETH_ALEN);
+		list_add_tail(&tmp->list, &priv->mc_list);
+	}
+}
+
+static void update_mclist_flags(struct mlx4_en_priv *priv,
+				struct list_head *dst,
+				struct list_head *src)
+{
+	struct mlx4_en_mc_list *dst_tmp, *src_tmp, *new_mc;
+	bool found;
+
+	/* Find all the entries that should be removed from dst,
+	 * These are the entries that are not found in src
+	 */
+	list_for_each_entry(dst_tmp, dst, list) {
+		found = false;
+		list_for_each_entry(src_tmp, src, list) {
+			if (ether_addr_equal(dst_tmp->addr, src_tmp->addr)) {
+				found = true;
+				break;
+			}
+		}
+		if (!found)
+			dst_tmp->action = MCLIST_REM;
+	}
+
+	/* Add entries that exist in src but not in dst
+	 * mark them as need to add
+	 */
+	list_for_each_entry(src_tmp, src, list) {
+		found = false;
+		list_for_each_entry(dst_tmp, dst, list) {
+			if (ether_addr_equal(dst_tmp->addr, src_tmp->addr)) {
+				dst_tmp->action = MCLIST_NONE;
+				found = true;
+				break;
+			}
+		}
+		if (!found) {
+			new_mc = kmemdup(src_tmp,
+					 sizeof(struct mlx4_en_mc_list),
+					 GFP_KERNEL);
+			if (!new_mc)
+				return;
+
+			new_mc->action = MCLIST_ADD;
+			list_add_tail(&new_mc->list, dst);
+		}
+	}
+}
+
+static void mlx4_en_set_rx_mode(struct net_device *dev)
+{
+	struct mlx4_en_priv *priv = netdev_priv(dev);
+
+	if (!priv->port_up)
+		return;
+
+	queue_work(priv->mdev->workqueue, &priv->rx_mode_task);
+}
+
+static void mlx4_en_set_promisc_mode(struct mlx4_en_priv *priv,
+				     struct mlx4_en_dev *mdev)
+{
+	int err = 0;
+
+	if (!(priv->flags & MLX4_EN_FLAG_PROMISC)) {
+		if (netif_msg_rx_status(priv))
+			en_warn(priv, "Entering promiscuous mode\n");
+		priv->flags |= MLX4_EN_FLAG_PROMISC;
+
+		/* Enable promiscouos mode */
+		switch (mdev->dev->caps.steering_mode) {
+		case MLX4_STEERING_MODE_DEVICE_MANAGED:
+			err = mlx4_flow_steer_promisc_add(mdev->dev,
+							  priv->port,
+							  priv->base_qpn,
+							  MLX4_FS_ALL_DEFAULT);
+			if (err)
+				en_err(priv, "Failed enabling promiscuous mode\n");
+			priv->flags |= MLX4_EN_FLAG_MC_PROMISC;
+			break;
+
+		case MLX4_STEERING_MODE_B0:
+			err = mlx4_unicast_promisc_add(mdev->dev,
+						       priv->base_qpn,
+						       priv->port);
+			if (err)
+				en_err(priv, "Failed enabling unicast promiscuous mode\n");
+
+			/* Add the default qp number as multicast
+			 * promisc
+			 */
+			if (!(priv->flags & MLX4_EN_FLAG_MC_PROMISC)) {
+				err = mlx4_multicast_promisc_add(mdev->dev,
+								 priv->base_qpn,
+								 priv->port);
+				if (err)
+					en_err(priv, "Failed enabling multicast promiscuous mode\n");
+				priv->flags |= MLX4_EN_FLAG_MC_PROMISC;
+			}
+			break;
+
+		case MLX4_STEERING_MODE_A0:
+			err = mlx4_SET_PORT_qpn_calc(mdev->dev,
+						     priv->port,
+						     priv->base_qpn,
+						     1);
+			if (err)
+				en_err(priv, "Failed enabling promiscuous mode\n");
+			break;
+		}
+
+		/* Disable port multicast filter (unconditionally) */
+		err = mlx4_SET_MCAST_FLTR(mdev->dev, priv->port, 0,
+					  0, MLX4_MCAST_DISABLE);
+		if (err)
+			en_err(priv, "Failed disabling multicast filter\n");
+	}
+}
+
+static void mlx4_en_clear_promisc_mode(struct mlx4_en_priv *priv,
+				       struct mlx4_en_dev *mdev)
+{
+	int err = 0;
+
+	if (netif_msg_rx_status(priv))
+		en_warn(priv, "Leaving promiscuous mode\n");
+	priv->flags &= ~MLX4_EN_FLAG_PROMISC;
+
+	/* Disable promiscouos mode */
+	switch (mdev->dev->caps.steering_mode) {
+	case MLX4_STEERING_MODE_DEVICE_MANAGED:
+		err = mlx4_flow_steer_promisc_remove(mdev->dev,
+						     priv->port,
+						     MLX4_FS_ALL_DEFAULT);
+		if (err)
+			en_err(priv, "Failed disabling promiscuous mode\n");
+		priv->flags &= ~MLX4_EN_FLAG_MC_PROMISC;
+		break;
+
+	case MLX4_STEERING_MODE_B0:
+		err = mlx4_unicast_promisc_remove(mdev->dev,
+						  priv->base_qpn,
+						  priv->port);
+		if (err)
+			en_err(priv, "Failed disabling unicast promiscuous mode\n");
+		/* Disable Multicast promisc */
+		if (priv->flags & MLX4_EN_FLAG_MC_PROMISC) {
+			err = mlx4_multicast_promisc_remove(mdev->dev,
+							    priv->base_qpn,
+							    priv->port);
+			if (err)
+				en_err(priv, "Failed disabling multicast promiscuous mode\n");
+			priv->flags &= ~MLX4_EN_FLAG_MC_PROMISC;
+		}
+		break;
+
+	case MLX4_STEERING_MODE_A0:
+		err = mlx4_SET_PORT_qpn_calc(mdev->dev,
+					     priv->port,
+					     priv->base_qpn, 0);
+		if (err)
+			en_err(priv, "Failed disabling promiscuous mode\n");
+		break;
+	}
+}
+
+static void mlx4_en_do_multicast(struct mlx4_en_priv *priv,
+				 struct net_device *dev,
+				 struct mlx4_en_dev *mdev)
+{
+	struct mlx4_en_mc_list *mclist, *tmp;
+	u64 mcast_addr = 0;
+	u8 mc_list[16] = {0};
+	int err = 0;
+
+	/* Enable/disable the multicast filter according to IFF_ALLMULTI */
+	if (dev->flags & IFF_ALLMULTI) {
+		err = mlx4_SET_MCAST_FLTR(mdev->dev, priv->port, 0,
+					  0, MLX4_MCAST_DISABLE);
+		if (err)
+			en_err(priv, "Failed disabling multicast filter\n");
+
+		/* Add the default qp number as multicast promisc */
+		if (!(priv->flags & MLX4_EN_FLAG_MC_PROMISC)) {
+			switch (mdev->dev->caps.steering_mode) {
+			case MLX4_STEERING_MODE_DEVICE_MANAGED:
+				err = mlx4_flow_steer_promisc_add(mdev->dev,
+								  priv->port,
+								  priv->base_qpn,
+								  MLX4_FS_MC_DEFAULT);
+				break;
+
+			case MLX4_STEERING_MODE_B0:
+				err = mlx4_multicast_promisc_add(mdev->dev,
+								 priv->base_qpn,
+								 priv->port);
+				break;
+
+			case MLX4_STEERING_MODE_A0:
+				break;
+			}
+			if (err)
+				en_err(priv, "Failed entering multicast promisc mode\n");
+			priv->flags |= MLX4_EN_FLAG_MC_PROMISC;
+		}
+	} else {
+		/* Disable Multicast promisc */
+		if (priv->flags & MLX4_EN_FLAG_MC_PROMISC) {
+			switch (mdev->dev->caps.steering_mode) {
+			case MLX4_STEERING_MODE_DEVICE_MANAGED:
+				err = mlx4_flow_steer_promisc_remove(mdev->dev,
+								     priv->port,
+								     MLX4_FS_MC_DEFAULT);
+				break;
+
+			case MLX4_STEERING_MODE_B0:
+				err = mlx4_multicast_promisc_remove(mdev->dev,
+								    priv->base_qpn,
+								    priv->port);
+				break;
+
+			case MLX4_STEERING_MODE_A0:
+				break;
+			}
+			if (err)
+				en_err(priv, "Failed disabling multicast promiscuous mode\n");
+			priv->flags &= ~MLX4_EN_FLAG_MC_PROMISC;
+		}
+
+		err = mlx4_SET_MCAST_FLTR(mdev->dev, priv->port, 0,
+					  0, MLX4_MCAST_DISABLE);
+		if (err)
+			en_err(priv, "Failed disabling multicast filter\n");
+
+		/* Flush mcast filter and init it with broadcast address */
+		mlx4_SET_MCAST_FLTR(mdev->dev, priv->port, ETH_BCAST,
+				    1, MLX4_MCAST_CONFIG);
+
+		/* Update multicast list - we cache all addresses so they won't
+		 * change while HW is updated holding the command semaphor */
+		netif_addr_lock_bh(dev);
+		mlx4_en_cache_mclist(dev);
+		netif_addr_unlock_bh(dev);
+		list_for_each_entry(mclist, &priv->mc_list, list) {
+			mcast_addr = mlx4_mac_to_u64(mclist->addr);
+			mlx4_SET_MCAST_FLTR(mdev->dev, priv->port,
+					    mcast_addr, 0, MLX4_MCAST_CONFIG);
+		}
+		err = mlx4_SET_MCAST_FLTR(mdev->dev, priv->port, 0,
+					  0, MLX4_MCAST_ENABLE);
+		if (err)
+			en_err(priv, "Failed enabling multicast filter\n");
+
+		update_mclist_flags(priv, &priv->curr_list, &priv->mc_list);
+		list_for_each_entry_safe(mclist, tmp, &priv->curr_list, list) {
+			if (mclist->action == MCLIST_REM) {
+				/* detach this address and delete from list */
+				memcpy(&mc_list[10], mclist->addr, ETH_ALEN);
+				mc_list[5] = priv->port;
+				err = mlx4_multicast_detach(mdev->dev,
+							    priv->rss_map.indir_qp,
+							    mc_list,
+							    MLX4_PROT_ETH,
+							    mclist->reg_id);
+				if (err)
+					en_err(priv, "Fail to detach multicast address\n");
+
+				if (mclist->tunnel_reg_id) {
+					err = mlx4_flow_detach(priv->mdev->dev, mclist->tunnel_reg_id);
+					if (err)
+						en_err(priv, "Failed to detach multicast address\n");
+				}
+
+				/* remove from list */
+				list_del(&mclist->list);
+				kfree(mclist);
+			} else if (mclist->action == MCLIST_ADD) {
+				/* attach the address */
+				memcpy(&mc_list[10], mclist->addr, ETH_ALEN);
+				/* needed for B0 steering support */
+				mc_list[5] = priv->port;
+				err = mlx4_multicast_attach(mdev->dev,
+							    priv->rss_map.indir_qp,
+							    mc_list,
+							    priv->port, 0,
+							    MLX4_PROT_ETH,
+							    &mclist->reg_id);
+				if (err)
+					en_err(priv, "Fail to attach multicast address\n");
+
+				err = mlx4_en_tunnel_steer_add(priv, &mc_list[10], priv->base_qpn,
+							       &mclist->tunnel_reg_id);
+				if (err)
+					en_err(priv, "Failed to attach multicast address\n");
+			}
+		}
+	}
+}
+
+static void mlx4_en_do_uc_filter(struct mlx4_en_priv *priv,
+				 struct net_device *dev,
+				 struct mlx4_en_dev *mdev)
+{
+	struct netdev_hw_addr *ha;
+	struct mlx4_mac_entry *entry;
+	struct hlist_node *tmp;
+	bool found;
+	u64 mac;
+	int err = 0;
+	struct hlist_head *bucket;
+	unsigned int i;
+	int removed = 0;
+	u32 prev_flags;
+
+	/* Note that we do not need to protect our mac_hash traversal with rcu,
+	 * since all modification code is protected by mdev->state_lock
+	 */
+
+	/* find what to remove */
+	for (i = 0; i < MLX4_EN_MAC_HASH_SIZE; ++i) {
+		bucket = &priv->mac_hash[i];
+		hlist_for_each_entry_safe(entry, tmp, bucket, hlist) {
+			found = false;
+			netdev_for_each_uc_addr(ha, dev) {
+				if (ether_addr_equal_64bits(entry->mac,
+							    ha->addr)) {
+					found = true;
+					break;
+				}
+			}
+
+			/* MAC address of the port is not in uc list */
+			if (ether_addr_equal_64bits(entry->mac,
+						    priv->current_mac))
+				found = true;
+
+			if (!found) {
+				mac = mlx4_mac_to_u64(entry->mac);
+				mlx4_en_uc_steer_release(priv, entry->mac,
+							 priv->base_qpn,
+							 entry->reg_id);
+				mlx4_unregister_mac(mdev->dev, priv->port, mac);
+
+				hlist_del_rcu(&entry->hlist);
+				kfree_rcu(entry, rcu);
+				en_dbg(DRV, priv, "Removed MAC %pM on port:%d\n",
+				       entry->mac, priv->port);
+				++removed;
+			}
+		}
+	}
+
+	/* if we didn't remove anything, there is no use in trying to add
+	 * again once we are in a forced promisc mode state
+	 */
+	if ((priv->flags & MLX4_EN_FLAG_FORCE_PROMISC) && 0 == removed)
+		return;
+
+	prev_flags = priv->flags;
+	priv->flags &= ~MLX4_EN_FLAG_FORCE_PROMISC;
+
+	/* find what to add */
+	netdev_for_each_uc_addr(ha, dev) {
+		found = false;
+		bucket = &priv->mac_hash[ha->addr[MLX4_EN_MAC_HASH_IDX]];
+		hlist_for_each_entry(entry, bucket, hlist) {
+			if (ether_addr_equal_64bits(entry->mac, ha->addr)) {
+				found = true;
+				break;
+			}
+		}
+
+		if (!found) {
+			entry = kmalloc(sizeof(*entry), GFP_KERNEL);
+			if (!entry) {
+				en_err(priv, "Failed adding MAC %pM on port:%d (out of memory)\n",
+				       ha->addr, priv->port);
+				priv->flags |= MLX4_EN_FLAG_FORCE_PROMISC;
+				break;
+			}
+			mac = mlx4_mac_to_u64(ha->addr);
+			memcpy(entry->mac, ha->addr, ETH_ALEN);
+			err = mlx4_register_mac(mdev->dev, priv->port, mac);
+			if (err < 0) {
+				en_err(priv, "Failed registering MAC %pM on port %d: %d\n",
+				       ha->addr, priv->port, err);
+				kfree(entry);
+				priv->flags |= MLX4_EN_FLAG_FORCE_PROMISC;
+				break;
+			}
+			err = mlx4_en_uc_steer_add(priv, ha->addr,
+						   &priv->base_qpn,
+						   &entry->reg_id);
+			if (err) {
+				en_err(priv, "Failed adding MAC %pM on port %d: %d\n",
+				       ha->addr, priv->port, err);
+				mlx4_unregister_mac(mdev->dev, priv->port, mac);
+				kfree(entry);
+				priv->flags |= MLX4_EN_FLAG_FORCE_PROMISC;
+				break;
+			} else {
+				unsigned int mac_hash;
+				en_dbg(DRV, priv, "Added MAC %pM on port:%d\n",
+				       ha->addr, priv->port);
+				mac_hash = ha->addr[MLX4_EN_MAC_HASH_IDX];
+				bucket = &priv->mac_hash[mac_hash];
+				hlist_add_head_rcu(&entry->hlist, bucket);
+			}
+		}
+	}
+
+	if (priv->flags & MLX4_EN_FLAG_FORCE_PROMISC) {
+		en_warn(priv, "Forcing promiscuous mode on port:%d\n",
+			priv->port);
+	} else if (prev_flags & MLX4_EN_FLAG_FORCE_PROMISC) {
+		en_warn(priv, "Stop forcing promiscuous mode on port:%d\n",
+			priv->port);
+	}
+}
+
+static void mlx4_en_do_set_rx_mode(struct work_struct *work)
+{
+	struct mlx4_en_priv *priv = container_of(work, struct mlx4_en_priv,
+						 rx_mode_task);
+	struct mlx4_en_dev *mdev = priv->mdev;
+	struct net_device *dev = priv->dev;
+
+	mutex_lock(&mdev->state_lock);
+	if (!mdev->device_up) {
+		en_dbg(HW, priv, "Card is not up, ignoring rx mode change.\n");
+		goto out;
+	}
+	if (!priv->port_up) {
+		en_dbg(HW, priv, "Port is down, ignoring rx mode change.\n");
+		goto out;
+	}
+
+	if (!netif_carrier_ok(dev)) {
+		if (!mlx4_en_QUERY_PORT(mdev, priv->port)) {
+			if (priv->port_state.link_state) {
+				priv->last_link_state = MLX4_DEV_EVENT_PORT_UP;
+				netif_carrier_on(dev);
+				en_dbg(LINK, priv, "Link Up\n");
+			}
+		}
+	}
+
+	if (dev->priv_flags & IFF_UNICAST_FLT)
+		mlx4_en_do_uc_filter(priv, dev, mdev);
+
+	/* Promsicuous mode: disable all filters */
+	if ((dev->flags & IFF_PROMISC) ||
+	    (priv->flags & MLX4_EN_FLAG_FORCE_PROMISC)) {
+		mlx4_en_set_promisc_mode(priv, mdev);
+		goto out;
+	}
+
+	/* Not in promiscuous mode */
+	if (priv->flags & MLX4_EN_FLAG_PROMISC)
+		mlx4_en_clear_promisc_mode(priv, mdev);
+
+	mlx4_en_do_multicast(priv, dev, mdev);
+out:
+	mutex_unlock(&mdev->state_lock);
+}
+
+static int mlx4_en_set_rss_steer_rules(struct mlx4_en_priv *priv)
+{
+	u64 reg_id;
+	int err = 0;
+	int *qpn = &priv->base_qpn;
+	struct mlx4_mac_entry *entry;
+
+	err = mlx4_en_uc_steer_add(priv, priv->dev->dev_addr, qpn, &reg_id);
+	if (err)
+		return err;
+
+	err = mlx4_en_tunnel_steer_add(priv, priv->dev->dev_addr, *qpn,
+				       &priv->tunnel_reg_id);
+	if (err)
+		goto tunnel_err;
+
+	entry = kmalloc(sizeof(*entry), GFP_KERNEL);
+	if (!entry) {
+		err = -ENOMEM;
+		goto alloc_err;
+	}
+
+	memcpy(entry->mac, priv->dev->dev_addr, sizeof(entry->mac));
+	memcpy(priv->current_mac, entry->mac, sizeof(priv->current_mac));
+	entry->reg_id = reg_id;
+	hlist_add_head_rcu(&entry->hlist,
+			   &priv->mac_hash[entry->mac[MLX4_EN_MAC_HASH_IDX]]);
+
+	return 0;
+
+alloc_err:
+	if (priv->tunnel_reg_id)
+		mlx4_flow_detach(priv->mdev->dev, priv->tunnel_reg_id);
+
+tunnel_err:
+	mlx4_en_uc_steer_release(priv, priv->dev->dev_addr, *qpn, reg_id);
+	return err;
+}
+
+static void mlx4_en_delete_rss_steer_rules(struct mlx4_en_priv *priv)
+{
+	u64 mac;
+	unsigned int i;
+	int qpn = priv->base_qpn;
+	struct hlist_head *bucket;
+	struct hlist_node *tmp;
+	struct mlx4_mac_entry *entry;
+
+	for (i = 0; i < MLX4_EN_MAC_HASH_SIZE; ++i) {
+		bucket = &priv->mac_hash[i];
+		hlist_for_each_entry_safe(entry, tmp, bucket, hlist) {
+			mac = mlx4_mac_to_u64(entry->mac);
+			en_dbg(DRV, priv, "Registering MAC:%pM for deleting\n",
+			       entry->mac);
+			mlx4_en_uc_steer_release(priv, entry->mac,
+						 qpn, entry->reg_id);
+
+			mlx4_unregister_mac(priv->mdev->dev, priv->port, mac);
+			hlist_del_rcu(&entry->hlist);
+			kfree_rcu(entry, rcu);
+		}
+	}
+
+	if (priv->tunnel_reg_id) {
+		mlx4_flow_detach(priv->mdev->dev, priv->tunnel_reg_id);
+		priv->tunnel_reg_id = 0;
+	}
+}
+
+static void mlx4_en_tx_timeout(struct net_device *dev)
+{
+	struct mlx4_en_priv *priv = netdev_priv(dev);
+	struct mlx4_en_dev *mdev = priv->mdev;
+	int i;
+
+	if (netif_msg_timer(priv))
+		en_warn(priv, "Tx timeout called on port:%d\n", priv->port);
+
+	for (i = 0; i < priv->tx_ring_num[TX]; i++) {
+		struct mlx4_en_tx_ring *tx_ring = priv->tx_ring[TX][i];
+
+		if (!netif_tx_queue_stopped(netdev_get_tx_queue(dev, i)))
+			continue;
+		en_warn(priv, "TX timeout on queue: %d, QP: 0x%x, CQ: 0x%x, Cons: 0x%x, Prod: 0x%x\n",
+			i, tx_ring->qpn, tx_ring->sp_cqn,
+			tx_ring->cons, tx_ring->prod);
+	}
+
+	priv->port_stats.tx_timeout++;
+	if (!test_and_set_bit(MLX4_EN_STATE_FLAG_RESTARTING, &priv->state)) {
+		en_dbg(DRV, priv, "Scheduling port restart\n");
+		queue_work(mdev->workqueue, &priv->restart_task);
+	}
+}
+
+
+static void
+mlx4_en_get_stats64(struct net_device *dev, struct rtnl_link_stats64 *stats)
+{
+	struct mlx4_en_priv *priv = netdev_priv(dev);
+
+	spin_lock_bh(&priv->stats_lock);
+	mlx4_en_fold_software_stats(dev);
+	netdev_stats_to_stats64(stats, &dev->stats);
+	spin_unlock_bh(&priv->stats_lock);
+}
+
+static void mlx4_en_set_default_moderation(struct mlx4_en_priv *priv)
+{
+	struct mlx4_en_cq *cq;
+	int i, t;
+
+	/* If we haven't received a specific coalescing setting
+	 * (module param), we set the moderation parameters as follows:
+	 * - moder_cnt is set to the number of mtu sized packets to
+	 *   satisfy our coalescing target.
+	 * - moder_time is set to a fixed value.
+	 */
+	priv->rx_frames = MLX4_EN_RX_COAL_TARGET;
+	priv->rx_usecs = MLX4_EN_RX_COAL_TIME;
+	priv->tx_frames = MLX4_EN_TX_COAL_PKTS;
+	priv->tx_usecs = MLX4_EN_TX_COAL_TIME;
+	en_dbg(INTR, priv, "Default coalescing params for mtu:%d - rx_frames:%d rx_usecs:%d\n",
+	       priv->dev->mtu, priv->rx_frames, priv->rx_usecs);
+
+	/* Setup cq moderation params */
+	for (i = 0; i < priv->rx_ring_num; i++) {
+		cq = priv->rx_cq[i];
+		cq->moder_cnt = priv->rx_frames;
+		cq->moder_time = priv->rx_usecs;
+		priv->last_moder_time[i] = MLX4_EN_AUTO_CONF;
+		priv->last_moder_packets[i] = 0;
+		priv->last_moder_bytes[i] = 0;
+	}
+
+	for (t = 0 ; t < MLX4_EN_NUM_TX_TYPES; t++) {
+		for (i = 0; i < priv->tx_ring_num[t]; i++) {
+			cq = priv->tx_cq[t][i];
+			cq->moder_cnt = priv->tx_frames;
+			cq->moder_time = priv->tx_usecs;
+		}
+	}
+
+	/* Reset auto-moderation params */
+	priv->pkt_rate_low = MLX4_EN_RX_RATE_LOW;
+	priv->rx_usecs_low = MLX4_EN_RX_COAL_TIME_LOW;
+	priv->pkt_rate_high = MLX4_EN_RX_RATE_HIGH;
+	priv->rx_usecs_high = MLX4_EN_RX_COAL_TIME_HIGH;
+	priv->sample_interval = MLX4_EN_SAMPLE_INTERVAL;
+	priv->adaptive_rx_coal = 1;
+	priv->last_moder_jiffies = 0;
+	priv->last_moder_tx_packets = 0;
+}
+
+static void mlx4_en_auto_moderation(struct mlx4_en_priv *priv)
+{
+	unsigned long period = (unsigned long) (jiffies - priv->last_moder_jiffies);
+	u32 pkt_rate_high, pkt_rate_low;
+	struct mlx4_en_cq *cq;
+	unsigned long packets;
+	unsigned long rate;
+	unsigned long avg_pkt_size;
+	unsigned long rx_packets;
+	unsigned long rx_bytes;
+	unsigned long rx_pkt_diff;
+	int moder_time;
+	int ring, err;
+
+	if (!priv->adaptive_rx_coal || period < priv->sample_interval * HZ)
+		return;
+
+	pkt_rate_low = READ_ONCE(priv->pkt_rate_low);
+	pkt_rate_high = READ_ONCE(priv->pkt_rate_high);
+
+	for (ring = 0; ring < priv->rx_ring_num; ring++) {
+		rx_packets = READ_ONCE(priv->rx_ring[ring]->packets);
+		rx_bytes = READ_ONCE(priv->rx_ring[ring]->bytes);
+
+		rx_pkt_diff = rx_packets - priv->last_moder_packets[ring];
+		packets = rx_pkt_diff;
+		rate = packets * HZ / period;
+		avg_pkt_size = packets ? (rx_bytes -
+				priv->last_moder_bytes[ring]) / packets : 0;
+
+		/* Apply auto-moderation only when packet rate
+		 * exceeds a rate that it matters */
+		if (rate > (MLX4_EN_RX_RATE_THRESH / priv->rx_ring_num) &&
+		    avg_pkt_size > MLX4_EN_AVG_PKT_SMALL) {
+			if (rate <= pkt_rate_low)
+				moder_time = priv->rx_usecs_low;
+			else if (rate >= pkt_rate_high)
+				moder_time = priv->rx_usecs_high;
+			else
+				moder_time = (rate - pkt_rate_low) *
+					(priv->rx_usecs_high - priv->rx_usecs_low) /
+					(pkt_rate_high - pkt_rate_low) +
+					priv->rx_usecs_low;
+		} else {
+			moder_time = priv->rx_usecs_low;
+		}
+
+		cq = priv->rx_cq[ring];
+		if (moder_time != priv->last_moder_time[ring] ||
+		    cq->moder_cnt != priv->rx_frames) {
+			priv->last_moder_time[ring] = moder_time;
+			cq->moder_time = moder_time;
+			cq->moder_cnt = priv->rx_frames;
+			err = mlx4_en_set_cq_moder(priv, cq);
+			if (err)
+				en_err(priv, "Failed modifying moderation for cq:%d\n",
+				       ring);
+		}
+		priv->last_moder_packets[ring] = rx_packets;
+		priv->last_moder_bytes[ring] = rx_bytes;
+	}
+
+	priv->last_moder_jiffies = jiffies;
+}
+
+static void mlx4_en_do_get_stats(struct work_struct *work)
+{
+	struct delayed_work *delay = to_delayed_work(work);
+	struct mlx4_en_priv *priv = container_of(delay, struct mlx4_en_priv,
+						 stats_task);
+	struct mlx4_en_dev *mdev = priv->mdev;
+	int err;
+
+	mutex_lock(&mdev->state_lock);
+	if (mdev->device_up) {
+		if (priv->port_up) {
+			err = mlx4_en_DUMP_ETH_STATS(mdev, priv->port, 0);
+			if (err)
+				en_dbg(HW, priv, "Could not update stats\n");
+
+			mlx4_en_auto_moderation(priv);
+		}
+
+		queue_delayed_work(mdev->workqueue, &priv->stats_task, STATS_DELAY);
+	}
+	if (mdev->mac_removed[MLX4_MAX_PORTS + 1 - priv->port]) {
+		mlx4_en_do_set_mac(priv, priv->current_mac);
+		mdev->mac_removed[MLX4_MAX_PORTS + 1 - priv->port] = 0;
+	}
+	mutex_unlock(&mdev->state_lock);
+}
+
+/* mlx4_en_service_task - Run service task for tasks that needed to be done
+ * periodically
+ */
+static void mlx4_en_service_task(struct work_struct *work)
+{
+	struct delayed_work *delay = to_delayed_work(work);
+	struct mlx4_en_priv *priv = container_of(delay, struct mlx4_en_priv,
+						 service_task);
+	struct mlx4_en_dev *mdev = priv->mdev;
+
+	mutex_lock(&mdev->state_lock);
+	if (mdev->device_up) {
+		if (mdev->dev->caps.flags2 & MLX4_DEV_CAP_FLAG2_TS)
+			mlx4_en_ptp_overflow_check(mdev);
+
+		mlx4_en_recover_from_oom(priv);
+		queue_delayed_work(mdev->workqueue, &priv->service_task,
+				   SERVICE_TASK_DELAY);
+	}
+	mutex_unlock(&mdev->state_lock);
+}
+
+static void mlx4_en_linkstate(struct work_struct *work)
+{
+	struct mlx4_en_priv *priv = container_of(work, struct mlx4_en_priv,
+						 linkstate_task);
+	struct mlx4_en_dev *mdev = priv->mdev;
+	int linkstate = priv->link_state;
+
+	mutex_lock(&mdev->state_lock);
+	/* If observable port state changed set carrier state and
+	 * report to system log */
+	if (priv->last_link_state != linkstate) {
+		if (linkstate == MLX4_DEV_EVENT_PORT_DOWN) {
+			en_info(priv, "Link Down\n");
+			netif_carrier_off(priv->dev);
+		} else {
+			en_info(priv, "Link Up\n");
+			netif_carrier_on(priv->dev);
+		}
+	}
+	priv->last_link_state = linkstate;
+	mutex_unlock(&mdev->state_lock);
+}
+
+static int mlx4_en_init_affinity_hint(struct mlx4_en_priv *priv, int ring_idx)
+{
+	struct mlx4_en_rx_ring *ring = priv->rx_ring[ring_idx];
+	int numa_node = priv->mdev->dev->numa_node;
+
+	if (!zalloc_cpumask_var(&ring->affinity_mask, GFP_KERNEL))
+		return -ENOMEM;
+
+	cpumask_set_cpu(cpumask_local_spread(ring_idx, numa_node),
+			ring->affinity_mask);
+	return 0;
+}
+
+static void mlx4_en_free_affinity_hint(struct mlx4_en_priv *priv, int ring_idx)
+{
+	free_cpumask_var(priv->rx_ring[ring_idx]->affinity_mask);
+}
+
+static void mlx4_en_init_recycle_ring(struct mlx4_en_priv *priv,
+				      int tx_ring_idx)
+{
+	struct mlx4_en_tx_ring *tx_ring = priv->tx_ring[TX_XDP][tx_ring_idx];
+	int rr_index = tx_ring_idx;
+
+	tx_ring->free_tx_desc = mlx4_en_recycle_tx_desc;
+	tx_ring->recycle_ring = priv->rx_ring[rr_index];
+	en_dbg(DRV, priv, "Set tx_ring[%d][%d]->recycle_ring = rx_ring[%d]\n",
+	       TX_XDP, tx_ring_idx, rr_index);
+}
+
+int mlx4_en_start_port(struct net_device *dev)
+{
+	struct mlx4_en_priv *priv = netdev_priv(dev);
+	struct mlx4_en_dev *mdev = priv->mdev;
+	struct mlx4_en_cq *cq;
+	struct mlx4_en_tx_ring *tx_ring;
+	int rx_index = 0;
+	int err = 0;
+	int i, t;
+	int j;
+	u8 mc_list[16] = {0};
+
+	if (priv->port_up) {
+		en_dbg(DRV, priv, "start port called while port already up\n");
+		return 0;
+	}
+
+	INIT_LIST_HEAD(&priv->mc_list);
+	INIT_LIST_HEAD(&priv->curr_list);
+	INIT_LIST_HEAD(&priv->ethtool_list);
+	memset(&priv->ethtool_rules[0], 0,
+	       sizeof(struct ethtool_flow_id) * MAX_NUM_OF_FS_RULES);
+
+	/* Calculate Rx buf size */
+	dev->mtu = min(dev->mtu, priv->max_mtu);
+	mlx4_en_calc_rx_buf(dev);
+	en_dbg(DRV, priv, "Rx buf size:%d\n", priv->rx_skb_size);
+
+	/* Configure rx cq's and rings */
+	err = mlx4_en_activate_rx_rings(priv);
+	if (err) {
+		en_err(priv, "Failed to activate RX rings\n");
+		return err;
+	}
+	for (i = 0; i < priv->rx_ring_num; i++) {
+		cq = priv->rx_cq[i];
+
+		err = mlx4_en_init_affinity_hint(priv, i);
+		if (err) {
+			en_err(priv, "Failed preparing IRQ affinity hint\n");
+			goto cq_err;
+		}
+
+		err = mlx4_en_activate_cq(priv, cq, i);
+		if (err) {
+			en_err(priv, "Failed activating Rx CQ\n");
+			mlx4_en_free_affinity_hint(priv, i);
+			goto cq_err;
+		}
+
+		for (j = 0; j < cq->size; j++) {
+			struct mlx4_cqe *cqe = NULL;
+
+			cqe = mlx4_en_get_cqe(cq->buf, j, priv->cqe_size) +
+			      priv->cqe_factor;
+			cqe->owner_sr_opcode = MLX4_CQE_OWNER_MASK;
+		}
+
+		err = mlx4_en_set_cq_moder(priv, cq);
+		if (err) {
+			en_err(priv, "Failed setting cq moderation parameters\n");
+			mlx4_en_deactivate_cq(priv, cq);
+			mlx4_en_free_affinity_hint(priv, i);
+			goto cq_err;
+		}
+		mlx4_en_arm_cq(priv, cq);
+		priv->rx_ring[i]->cqn = cq->mcq.cqn;
+		++rx_index;
+	}
+
+	/* Set qp number */
+	en_dbg(DRV, priv, "Getting qp number for port %d\n", priv->port);
+	err = mlx4_en_get_qp(priv);
+	if (err) {
+		en_err(priv, "Failed getting eth qp\n");
+		goto cq_err;
+	}
+	mdev->mac_removed[priv->port] = 0;
+
+	priv->counter_index =
+			mlx4_get_default_counter_index(mdev->dev, priv->port);
+
+	err = mlx4_en_config_rss_steer(priv);
+	if (err) {
+		en_err(priv, "Failed configuring rss steering\n");
+		goto mac_err;
+	}
+
+	err = mlx4_en_create_drop_qp(priv);
+	if (err)
+		goto rss_err;
+
+	/* Configure tx cq's and rings */
+	for (t = 0 ; t < MLX4_EN_NUM_TX_TYPES; t++) {
+		u8 num_tx_rings_p_up = t == TX ?
+			priv->num_tx_rings_p_up : priv->tx_ring_num[t];
+
+		for (i = 0; i < priv->tx_ring_num[t]; i++) {
+			/* Configure cq */
+			cq = priv->tx_cq[t][i];
+			err = mlx4_en_activate_cq(priv, cq, i);
+			if (err) {
+				en_err(priv, "Failed allocating Tx CQ\n");
+				goto tx_err;
+			}
+			err = mlx4_en_set_cq_moder(priv, cq);
+			if (err) {
+				en_err(priv, "Failed setting cq moderation parameters\n");
+				mlx4_en_deactivate_cq(priv, cq);
+				goto tx_err;
+			}
+			en_dbg(DRV, priv,
+			       "Resetting index of collapsed CQ:%d to -1\n", i);
+			cq->buf->wqe_index = cpu_to_be16(0xffff);
+
+			/* Configure ring */
+			tx_ring = priv->tx_ring[t][i];
+			err = mlx4_en_activate_tx_ring(priv, tx_ring,
+						       cq->mcq.cqn,
+						       i / num_tx_rings_p_up);
+			if (err) {
+				en_err(priv, "Failed allocating Tx ring\n");
+				mlx4_en_deactivate_cq(priv, cq);
+				goto tx_err;
+			}
+			clear_bit(MLX4_EN_TX_RING_STATE_RECOVERING, &tx_ring->state);
+			if (t != TX_XDP) {
+				tx_ring->tx_queue = netdev_get_tx_queue(dev, i);
+				tx_ring->recycle_ring = NULL;
+
+				/* Arm CQ for TX completions */
+				mlx4_en_arm_cq(priv, cq);
+
+			} else {
+				mlx4_en_init_tx_xdp_ring_descs(priv, tx_ring);
+				mlx4_en_init_recycle_ring(priv, i);
+				/* XDP TX CQ should never be armed */
+			}
+
+			/* Set initial ownership of all Tx TXBBs to SW (1) */
+			for (j = 0; j < tx_ring->buf_size; j += STAMP_STRIDE)
+				*((u32 *)(tx_ring->buf + j)) = 0xffffffff;
+		}
+	}
+
+	/* Configure port */
+	err = mlx4_SET_PORT_general(mdev->dev, priv->port,
+				    priv->rx_skb_size + ETH_FCS_LEN,
+				    priv->prof->tx_pause,
+				    priv->prof->tx_ppp,
+				    priv->prof->rx_pause,
+				    priv->prof->rx_ppp);
+	if (err) {
+		en_err(priv, "Failed setting port general configurations for port %d, with error %d\n",
+		       priv->port, err);
+		goto tx_err;
+	}
+
+	err = mlx4_SET_PORT_user_mtu(mdev->dev, priv->port, dev->mtu);
+	if (err) {
+		en_err(priv, "Failed to pass user MTU(%d) to Firmware for port %d, with error %d\n",
+		       dev->mtu, priv->port, err);
+		goto tx_err;
+	}
+
+	/* Set default qp number */
+	err = mlx4_SET_PORT_qpn_calc(mdev->dev, priv->port, priv->base_qpn, 0);
+	if (err) {
+		en_err(priv, "Failed setting default qp numbers\n");
+		goto tx_err;
+	}
+
+	if (mdev->dev->caps.tunnel_offload_mode == MLX4_TUNNEL_OFFLOAD_MODE_VXLAN) {
+		err = mlx4_SET_PORT_VXLAN(mdev->dev, priv->port, VXLAN_STEER_BY_OUTER_MAC, 1);
+		if (err) {
+			en_err(priv, "Failed setting port L2 tunnel configuration, err %d\n",
+			       err);
+			goto tx_err;
+		}
+	}
+
+	/* Init port */
+	en_dbg(HW, priv, "Initializing port\n");
+	err = mlx4_INIT_PORT(mdev->dev, priv->port);
+	if (err) {
+		en_err(priv, "Failed Initializing port\n");
+		goto tx_err;
+	}
+
+	/* Set Unicast and VXLAN steering rules */
+	if (mdev->dev->caps.steering_mode != MLX4_STEERING_MODE_A0 &&
+	    mlx4_en_set_rss_steer_rules(priv))
+		mlx4_warn(mdev, "Failed setting steering rules\n");
+
+	/* Attach rx QP to bradcast address */
+	eth_broadcast_addr(&mc_list[10]);
+	mc_list[5] = priv->port; /* needed for B0 steering support */
+	if (mlx4_multicast_attach(mdev->dev, priv->rss_map.indir_qp, mc_list,
+				  priv->port, 0, MLX4_PROT_ETH,
+				  &priv->broadcast_id))
+		mlx4_warn(mdev, "Failed Attaching Broadcast\n");
+
+	/* Must redo promiscuous mode setup. */
+	priv->flags &= ~(MLX4_EN_FLAG_PROMISC | MLX4_EN_FLAG_MC_PROMISC);
+
+	/* Schedule multicast task to populate multicast list */
+	queue_work(mdev->workqueue, &priv->rx_mode_task);
+
+	if (priv->mdev->dev->caps.tunnel_offload_mode == MLX4_TUNNEL_OFFLOAD_MODE_VXLAN)
+		udp_tunnel_get_rx_info(dev);
+
+	priv->port_up = true;
+
+	/* Process all completions if exist to prevent
+	 * the queues freezing if they are full
+	 */
+	for (i = 0; i < priv->rx_ring_num; i++) {
+		local_bh_disable();
+		napi_schedule(&priv->rx_cq[i]->napi);
+		local_bh_enable();
+	}
+
+	clear_bit(MLX4_EN_STATE_FLAG_RESTARTING, &priv->state);
+	netif_tx_start_all_queues(dev);
+	netif_device_attach(dev);
+
+	return 0;
+
+tx_err:
+	if (t == MLX4_EN_NUM_TX_TYPES) {
+		t--;
+		i = priv->tx_ring_num[t];
+	}
+	while (t >= 0) {
+		while (i--) {
+			mlx4_en_deactivate_tx_ring(priv, priv->tx_ring[t][i]);
+			mlx4_en_deactivate_cq(priv, priv->tx_cq[t][i]);
+		}
+		if (!t--)
+			break;
+		i = priv->tx_ring_num[t];
+	}
+	mlx4_en_destroy_drop_qp(priv);
+rss_err:
+	mlx4_en_release_rss_steer(priv);
+mac_err:
+	mlx4_en_put_qp(priv);
+cq_err:
+	while (rx_index--) {
+		mlx4_en_deactivate_cq(priv, priv->rx_cq[rx_index]);
+		mlx4_en_free_affinity_hint(priv, rx_index);
+	}
+	for (i = 0; i < priv->rx_ring_num; i++)
+		mlx4_en_deactivate_rx_ring(priv, priv->rx_ring[i]);
+
+	return err; /* need to close devices */
+}
+
+
+void mlx4_en_stop_port(struct net_device *dev, int detach)
+{
+	struct mlx4_en_priv *priv = netdev_priv(dev);
+	struct mlx4_en_dev *mdev = priv->mdev;
+	struct mlx4_en_mc_list *mclist, *tmp;
+	struct ethtool_flow_id *flow, *tmp_flow;
+	int i, t;
+	u8 mc_list[16] = {0};
+
+	if (!priv->port_up) {
+		en_dbg(DRV, priv, "stop port called while port already down\n");
+		return;
+	}
+
+	/* close port*/
+	mlx4_CLOSE_PORT(mdev->dev, priv->port);
+
+	/* Synchronize with tx routine */
+	netif_tx_lock_bh(dev);
+	if (detach)
+		netif_device_detach(dev);
+	netif_tx_stop_all_queues(dev);
+	netif_tx_unlock_bh(dev);
+
+	netif_tx_disable(dev);
+
+	spin_lock_bh(&priv->stats_lock);
+	mlx4_en_fold_software_stats(dev);
+	/* Set port as not active */
+	priv->port_up = false;
+	spin_unlock_bh(&priv->stats_lock);
+
+	priv->counter_index = MLX4_SINK_COUNTER_INDEX(mdev->dev);
+
+	/* Promsicuous mode */
+	if (mdev->dev->caps.steering_mode ==
+	    MLX4_STEERING_MODE_DEVICE_MANAGED) {
+		priv->flags &= ~(MLX4_EN_FLAG_PROMISC |
+				 MLX4_EN_FLAG_MC_PROMISC);
+		mlx4_flow_steer_promisc_remove(mdev->dev,
+					       priv->port,
+					       MLX4_FS_ALL_DEFAULT);
+		mlx4_flow_steer_promisc_remove(mdev->dev,
+					       priv->port,
+					       MLX4_FS_MC_DEFAULT);
+	} else if (priv->flags & MLX4_EN_FLAG_PROMISC) {
+		priv->flags &= ~MLX4_EN_FLAG_PROMISC;
+
+		/* Disable promiscouos mode */
+		mlx4_unicast_promisc_remove(mdev->dev, priv->base_qpn,
+					    priv->port);
+
+		/* Disable Multicast promisc */
+		if (priv->flags & MLX4_EN_FLAG_MC_PROMISC) {
+			mlx4_multicast_promisc_remove(mdev->dev, priv->base_qpn,
+						      priv->port);
+			priv->flags &= ~MLX4_EN_FLAG_MC_PROMISC;
+		}
+	}
+
+	/* Detach All multicasts */
+	eth_broadcast_addr(&mc_list[10]);
+	mc_list[5] = priv->port; /* needed for B0 steering support */
+	mlx4_multicast_detach(mdev->dev, priv->rss_map.indir_qp, mc_list,
+			      MLX4_PROT_ETH, priv->broadcast_id);
+	list_for_each_entry(mclist, &priv->curr_list, list) {
+		memcpy(&mc_list[10], mclist->addr, ETH_ALEN);
+		mc_list[5] = priv->port;
+		mlx4_multicast_detach(mdev->dev, priv->rss_map.indir_qp,
+				      mc_list, MLX4_PROT_ETH, mclist->reg_id);
+		if (mclist->tunnel_reg_id)
+			mlx4_flow_detach(mdev->dev, mclist->tunnel_reg_id);
+	}
+	mlx4_en_clear_list(dev);
+	list_for_each_entry_safe(mclist, tmp, &priv->curr_list, list) {
+		list_del(&mclist->list);
+		kfree(mclist);
+	}
+
+	/* Flush multicast filter */
+	mlx4_SET_MCAST_FLTR(mdev->dev, priv->port, 0, 1, MLX4_MCAST_CONFIG);
+
+	/* Remove flow steering rules for the port*/
+	if (mdev->dev->caps.steering_mode ==
+	    MLX4_STEERING_MODE_DEVICE_MANAGED) {
+		ASSERT_RTNL();
+		list_for_each_entry_safe(flow, tmp_flow,
+					 &priv->ethtool_list, list) {
+			mlx4_flow_detach(mdev->dev, flow->id);
+			list_del(&flow->list);
+		}
+	}
+
+	mlx4_en_destroy_drop_qp(priv);
+
+	/* Free TX Rings */
+	for (t = 0; t < MLX4_EN_NUM_TX_TYPES; t++) {
+		for (i = 0; i < priv->tx_ring_num[t]; i++) {
+			mlx4_en_deactivate_tx_ring(priv, priv->tx_ring[t][i]);
+			mlx4_en_deactivate_cq(priv, priv->tx_cq[t][i]);
+		}
+	}
+	msleep(10);
+
+	for (t = 0; t < MLX4_EN_NUM_TX_TYPES; t++)
+		for (i = 0; i < priv->tx_ring_num[t]; i++)
+			mlx4_en_free_tx_buf(dev, priv->tx_ring[t][i]);
+
+	if (mdev->dev->caps.steering_mode != MLX4_STEERING_MODE_A0)
+		mlx4_en_delete_rss_steer_rules(priv);
+
+	/* Free RSS qps */
+	mlx4_en_release_rss_steer(priv);
+
+	/* Unregister Mac address for the port */
+	mlx4_en_put_qp(priv);
+	if (!(mdev->dev->caps.flags2 & MLX4_DEV_CAP_FLAG2_REASSIGN_MAC_EN))
+		mdev->mac_removed[priv->port] = 1;
+
+	/* Free RX Rings */
+	for (i = 0; i < priv->rx_ring_num; i++) {
+		struct mlx4_en_cq *cq = priv->rx_cq[i];
+
+		napi_synchronize(&cq->napi);
+		mlx4_en_deactivate_rx_ring(priv, priv->rx_ring[i]);
+		mlx4_en_deactivate_cq(priv, cq);
+
+		mlx4_en_free_affinity_hint(priv, i);
+	}
+}
+
+static void mlx4_en_restart(struct work_struct *work)
+{
+	struct mlx4_en_priv *priv = container_of(work, struct mlx4_en_priv,
+						 restart_task);
+	struct mlx4_en_dev *mdev = priv->mdev;
+	struct net_device *dev = priv->dev;
+
+	en_dbg(DRV, priv, "Watchdog task called for port %d\n", priv->port);
+
+	rtnl_lock();
+	mutex_lock(&mdev->state_lock);
+	if (priv->port_up) {
+		mlx4_en_stop_port(dev, 1);
+		if (mlx4_en_start_port(dev))
+			en_err(priv, "Failed restarting port %d\n", priv->port);
+	}
+	mutex_unlock(&mdev->state_lock);
+	rtnl_unlock();
+}
+
+static void mlx4_en_clear_stats(struct net_device *dev)
+{
+	struct mlx4_en_priv *priv = netdev_priv(dev);
+	struct mlx4_en_dev *mdev = priv->mdev;
+	struct mlx4_en_tx_ring **tx_ring;
+	int i;
+
+	if (!mlx4_is_slave(mdev->dev))
+		if (mlx4_en_DUMP_ETH_STATS(mdev, priv->port, 1))
+			en_dbg(HW, priv, "Failed dumping statistics\n");
+
+	memset(&priv->pstats, 0, sizeof(priv->pstats));
+	memset(&priv->pkstats, 0, sizeof(priv->pkstats));
+	memset(&priv->port_stats, 0, sizeof(priv->port_stats));
+	memset(&priv->rx_flowstats, 0, sizeof(priv->rx_flowstats));
+	memset(&priv->tx_flowstats, 0, sizeof(priv->tx_flowstats));
+	memset(&priv->rx_priority_flowstats, 0,
+	       sizeof(priv->rx_priority_flowstats));
+	memset(&priv->tx_priority_flowstats, 0,
+	       sizeof(priv->tx_priority_flowstats));
+	memset(&priv->pf_stats, 0, sizeof(priv->pf_stats));
+
+	tx_ring = priv->tx_ring[TX];
+	for (i = 0; i < priv->tx_ring_num[TX]; i++) {
+		tx_ring[i]->bytes = 0;
+		tx_ring[i]->packets = 0;
+		tx_ring[i]->tx_csum = 0;
+		tx_ring[i]->tx_dropped = 0;
+		tx_ring[i]->queue_stopped = 0;
+		tx_ring[i]->wake_queue = 0;
+		tx_ring[i]->tso_packets = 0;
+		tx_ring[i]->xmit_more = 0;
+	}
+	for (i = 0; i < priv->rx_ring_num; i++) {
+		priv->rx_ring[i]->bytes = 0;
+		priv->rx_ring[i]->packets = 0;
+		priv->rx_ring[i]->csum_ok = 0;
+		priv->rx_ring[i]->csum_none = 0;
+		priv->rx_ring[i]->csum_complete = 0;
+	}
+}
+
+static int mlx4_en_open(struct net_device *dev)
+{
+	struct mlx4_en_priv *priv = netdev_priv(dev);
+	struct mlx4_en_dev *mdev = priv->mdev;
+	int err = 0;
+
+	mutex_lock(&mdev->state_lock);
+
+	if (!mdev->device_up) {
+		en_err(priv, "Cannot open - device down/disabled\n");
+		err = -EBUSY;
+		goto out;
+	}
+
+	/* Reset HW statistics and SW counters */
+	mlx4_en_clear_stats(dev);
+
+	err = mlx4_en_start_port(dev);
+	if (err)
+		en_err(priv, "Failed starting port:%d\n", priv->port);
+
+out:
+	mutex_unlock(&mdev->state_lock);
+	return err;
+}
+
+
+static int mlx4_en_close(struct net_device *dev)
+{
+	struct mlx4_en_priv *priv = netdev_priv(dev);
+	struct mlx4_en_dev *mdev = priv->mdev;
+
+	en_dbg(IFDOWN, priv, "Close port called\n");
+
+	mutex_lock(&mdev->state_lock);
+
+	mlx4_en_stop_port(dev, 0);
+	netif_carrier_off(dev);
+
+	mutex_unlock(&mdev->state_lock);
+	return 0;
+}
+
+static void mlx4_en_free_resources(struct mlx4_en_priv *priv)
+{
+	int i, t;
+
+#ifdef CONFIG_RFS_ACCEL
+	priv->dev->rx_cpu_rmap = NULL;
+#endif
+
+	for (t = 0; t < MLX4_EN_NUM_TX_TYPES; t++) {
+		for (i = 0; i < priv->tx_ring_num[t]; i++) {
+			if (priv->tx_ring[t] && priv->tx_ring[t][i])
+				mlx4_en_destroy_tx_ring(priv,
+							&priv->tx_ring[t][i]);
+			if (priv->tx_cq[t] && priv->tx_cq[t][i])
+				mlx4_en_destroy_cq(priv, &priv->tx_cq[t][i]);
+		}
+		kfree(priv->tx_ring[t]);
+		kfree(priv->tx_cq[t]);
+	}
+
+	for (i = 0; i < priv->rx_ring_num; i++) {
+		if (priv->rx_ring[i])
+			mlx4_en_destroy_rx_ring(priv, &priv->rx_ring[i],
+				priv->prof->rx_ring_size, priv->stride);
+		if (priv->rx_cq[i])
+			mlx4_en_destroy_cq(priv, &priv->rx_cq[i]);
+	}
+
+}
+
+static int mlx4_en_alloc_resources(struct mlx4_en_priv *priv)
+{
+	struct mlx4_en_port_profile *prof = priv->prof;
+	int i, t;
+	int node;
+
+	/* Create tx Rings */
+	for (t = 0; t < MLX4_EN_NUM_TX_TYPES; t++) {
+		for (i = 0; i < priv->tx_ring_num[t]; i++) {
+			node = cpu_to_node(i % num_online_cpus());
+			if (mlx4_en_create_cq(priv, &priv->tx_cq[t][i],
+					      prof->tx_ring_size, i, t, node))
+				goto err;
+
+			if (mlx4_en_create_tx_ring(priv, &priv->tx_ring[t][i],
+						   prof->tx_ring_size,
+						   TXBB_SIZE, node, i))
+				goto err;
+		}
+	}
+
+	/* Create rx Rings */
+	for (i = 0; i < priv->rx_ring_num; i++) {
+		node = cpu_to_node(i % num_online_cpus());
+		if (mlx4_en_create_cq(priv, &priv->rx_cq[i],
+				      prof->rx_ring_size, i, RX, node))
+			goto err;
+
+		if (mlx4_en_create_rx_ring(priv, &priv->rx_ring[i],
+					   prof->rx_ring_size, priv->stride,
+					   node, i))
+			goto err;
+
+	}
+
+#ifdef CONFIG_RFS_ACCEL
+	priv->dev->rx_cpu_rmap = mlx4_get_cpu_rmap(priv->mdev->dev, priv->port);
+#endif
+
+	return 0;
+
+err:
+	en_err(priv, "Failed to allocate NIC resources\n");
+	for (i = 0; i < priv->rx_ring_num; i++) {
+		if (priv->rx_ring[i])
+			mlx4_en_destroy_rx_ring(priv, &priv->rx_ring[i],
+						prof->rx_ring_size,
+						priv->stride);
+		if (priv->rx_cq[i])
+			mlx4_en_destroy_cq(priv, &priv->rx_cq[i]);
+	}
+	for (t = 0; t < MLX4_EN_NUM_TX_TYPES; t++) {
+		for (i = 0; i < priv->tx_ring_num[t]; i++) {
+			if (priv->tx_ring[t][i])
+				mlx4_en_destroy_tx_ring(priv,
+							&priv->tx_ring[t][i]);
+			if (priv->tx_cq[t][i])
+				mlx4_en_destroy_cq(priv, &priv->tx_cq[t][i]);
+		}
+	}
+	return -ENOMEM;
+}
+
+
+static int mlx4_en_copy_priv(struct mlx4_en_priv *dst,
+			     struct mlx4_en_priv *src,
+			     struct mlx4_en_port_profile *prof)
+{
+	int t;
+
+	memcpy(&dst->hwtstamp_config, &prof->hwtstamp_config,
+	       sizeof(dst->hwtstamp_config));
+	dst->num_tx_rings_p_up = prof->num_tx_rings_p_up;
+	dst->rx_ring_num = prof->rx_ring_num;
+	dst->flags = prof->flags;
+	dst->mdev = src->mdev;
+	dst->port = src->port;
+	dst->dev = src->dev;
+	dst->prof = prof;
+	dst->stride = roundup_pow_of_two(sizeof(struct mlx4_en_rx_desc) +
+					 DS_SIZE * MLX4_EN_MAX_RX_FRAGS);
+
+	for (t = 0; t < MLX4_EN_NUM_TX_TYPES; t++) {
+		dst->tx_ring_num[t] = prof->tx_ring_num[t];
+		if (!dst->tx_ring_num[t])
+			continue;
+
+		dst->tx_ring[t] = kcalloc(MAX_TX_RINGS,
+					  sizeof(struct mlx4_en_tx_ring *),
+					  GFP_KERNEL);
+		if (!dst->tx_ring[t])
+			goto err_free_tx;
+
+		dst->tx_cq[t] = kcalloc(MAX_TX_RINGS,
+					sizeof(struct mlx4_en_cq *),
+					GFP_KERNEL);
+		if (!dst->tx_cq[t]) {
+			kfree(dst->tx_ring[t]);
+			goto err_free_tx;
+		}
+	}
+
+	return 0;
+
+err_free_tx:
+	while (t--) {
+		kfree(dst->tx_ring[t]);
+		kfree(dst->tx_cq[t]);
+	}
+	return -ENOMEM;
+}
+
+static void mlx4_en_update_priv(struct mlx4_en_priv *dst,
+				struct mlx4_en_priv *src)
+{
+	int t;
+	memcpy(dst->rx_ring, src->rx_ring,
+	       sizeof(struct mlx4_en_rx_ring *) * src->rx_ring_num);
+	memcpy(dst->rx_cq, src->rx_cq,
+	       sizeof(struct mlx4_en_cq *) * src->rx_ring_num);
+	memcpy(&dst->hwtstamp_config, &src->hwtstamp_config,
+	       sizeof(dst->hwtstamp_config));
+	for (t = 0; t < MLX4_EN_NUM_TX_TYPES; t++) {
+		dst->tx_ring_num[t] = src->tx_ring_num[t];
+		dst->tx_ring[t] = src->tx_ring[t];
+		dst->tx_cq[t] = src->tx_cq[t];
+	}
+	dst->num_tx_rings_p_up = src->num_tx_rings_p_up;
+	dst->rx_ring_num = src->rx_ring_num;
+	memcpy(dst->prof, src->prof, sizeof(struct mlx4_en_port_profile));
+}
+
+int mlx4_en_try_alloc_resources(struct mlx4_en_priv *priv,
+				struct mlx4_en_priv *tmp,
+				struct mlx4_en_port_profile *prof,
+				bool carry_xdp_prog)
+{
+	struct bpf_prog *xdp_prog;
+	int i, t, ret;
+
+	ret = mlx4_en_copy_priv(tmp, priv, prof);
+	if (ret) {
+		en_warn(priv, "%s: mlx4_en_copy_priv() failed, return\n",
+			__func__);
+		return ret;
+	}
+
+	if (mlx4_en_alloc_resources(tmp)) {
+		en_warn(priv,
+			"%s: Resource allocation failed, using previous configuration\n",
+			__func__);
+		for (t = 0; t < MLX4_EN_NUM_TX_TYPES; t++) {
+			kfree(tmp->tx_ring[t]);
+			kfree(tmp->tx_cq[t]);
+		}
+		return -ENOMEM;
+	}
+
+	/* All rx_rings has the same xdp_prog.  Pick the first one. */
+	xdp_prog = rcu_dereference_protected(
+		priv->rx_ring[0]->xdp_prog,
+		lockdep_is_held(&priv->mdev->state_lock));
+
+	if (xdp_prog && carry_xdp_prog) {
+		xdp_prog = bpf_prog_add(xdp_prog, tmp->rx_ring_num);
+		if (IS_ERR(xdp_prog)) {
+			mlx4_en_free_resources(tmp);
+			return PTR_ERR(xdp_prog);
+		}
+		for (i = 0; i < tmp->rx_ring_num; i++)
+			rcu_assign_pointer(tmp->rx_ring[i]->xdp_prog,
+					   xdp_prog);
+	}
+
+	return 0;
+}
+
+void mlx4_en_safe_replace_resources(struct mlx4_en_priv *priv,
+				    struct mlx4_en_priv *tmp)
+{
+	mlx4_en_free_resources(priv);
+	mlx4_en_update_priv(priv, tmp);
+}
+
+void mlx4_en_destroy_netdev(struct net_device *dev)
+{
+	struct mlx4_en_priv *priv = netdev_priv(dev);
+	struct mlx4_en_dev *mdev = priv->mdev;
+
+	en_dbg(DRV, priv, "Destroying netdev on port:%d\n", priv->port);
+
+	/* Unregister device - this will close the port if it was up */
+	if (priv->registered) {
+		devlink_port_type_clear(mlx4_get_devlink_port(mdev->dev,
+							      priv->port));
+		unregister_netdev(dev);
+	}
+
+	if (priv->allocated)
+		mlx4_free_hwq_res(mdev->dev, &priv->res, MLX4_EN_PAGE_SIZE);
+
+	cancel_delayed_work(&priv->stats_task);
+	cancel_delayed_work(&priv->service_task);
+	/* flush any pending task for this netdev */
+	flush_workqueue(mdev->workqueue);
+
+	if (mdev->dev->caps.flags2 & MLX4_DEV_CAP_FLAG2_TS)
+		mlx4_en_remove_timestamp(mdev);
+
+	/* Detach the netdev so tasks would not attempt to access it */
+	mutex_lock(&mdev->state_lock);
+	mdev->pndev[priv->port] = NULL;
+	mdev->upper[priv->port] = NULL;
+
+#ifdef CONFIG_RFS_ACCEL
+	mlx4_en_cleanup_filters(priv);
+#endif
+
+	mlx4_en_free_resources(priv);
+	mutex_unlock(&mdev->state_lock);
+
+	free_netdev(dev);
+}
+
+static bool mlx4_en_check_xdp_mtu(struct net_device *dev, int mtu)
+{
+	struct mlx4_en_priv *priv = netdev_priv(dev);
+
+	if (mtu > MLX4_EN_MAX_XDP_MTU) {
+		en_err(priv, "mtu:%d > max:%d when XDP prog is attached\n",
+		       mtu, MLX4_EN_MAX_XDP_MTU);
+		return false;
+	}
+
+	return true;
+}
+
+static int mlx4_en_change_mtu(struct net_device *dev, int new_mtu)
+{
+	struct mlx4_en_priv *priv = netdev_priv(dev);
+	struct mlx4_en_dev *mdev = priv->mdev;
+	int err = 0;
+
+	en_dbg(DRV, priv, "Change MTU called - current:%d new:%d\n",
+		 dev->mtu, new_mtu);
+
+	if (priv->tx_ring_num[TX_XDP] &&
+	    !mlx4_en_check_xdp_mtu(dev, new_mtu))
+		return -EOPNOTSUPP;
+
+	dev->mtu = new_mtu;
+
+	if (netif_running(dev)) {
+		mutex_lock(&mdev->state_lock);
+		if (!mdev->device_up) {
+			/* NIC is probably restarting - let restart task reset
+			 * the port */
+			en_dbg(DRV, priv, "Change MTU called with card down!?\n");
+		} else {
+			mlx4_en_stop_port(dev, 1);
+			err = mlx4_en_start_port(dev);
+			if (err) {
+				en_err(priv, "Failed restarting port:%d\n",
+					 priv->port);
+				if (!test_and_set_bit(MLX4_EN_STATE_FLAG_RESTARTING,
+						      &priv->state))
+					queue_work(mdev->workqueue, &priv->restart_task);
+			}
+		}
+		mutex_unlock(&mdev->state_lock);
+	}
+	return 0;
+}
+
+static int mlx4_en_hwtstamp_set(struct net_device *dev, struct ifreq *ifr)
+{
+	struct mlx4_en_priv *priv = netdev_priv(dev);
+	struct mlx4_en_dev *mdev = priv->mdev;
+	struct hwtstamp_config config;
+
+	if (copy_from_user(&config, ifr->ifr_data, sizeof(config)))
+		return -EFAULT;
+
+	/* reserved for future extensions */
+	if (config.flags)
+		return -EINVAL;
+
+	/* device doesn't support time stamping */
+	if (!(mdev->dev->caps.flags2 & MLX4_DEV_CAP_FLAG2_TS))
+		return -EINVAL;
+
+	/* TX HW timestamp */
+	switch (config.tx_type) {
+	case HWTSTAMP_TX_OFF:
+	case HWTSTAMP_TX_ON:
+		break;
+	default:
+		return -ERANGE;
+	}
+
+	/* RX HW timestamp */
+	switch (config.rx_filter) {
+	case HWTSTAMP_FILTER_NONE:
+		break;
+	case HWTSTAMP_FILTER_ALL:
+	case HWTSTAMP_FILTER_SOME:
+	case HWTSTAMP_FILTER_PTP_V1_L4_EVENT:
+	case HWTSTAMP_FILTER_PTP_V1_L4_SYNC:
+	case HWTSTAMP_FILTER_PTP_V1_L4_DELAY_REQ:
+	case HWTSTAMP_FILTER_PTP_V2_L4_EVENT:
+	case HWTSTAMP_FILTER_PTP_V2_L4_SYNC:
+	case HWTSTAMP_FILTER_PTP_V2_L4_DELAY_REQ:
+	case HWTSTAMP_FILTER_PTP_V2_L2_EVENT:
+	case HWTSTAMP_FILTER_PTP_V2_L2_SYNC:
+	case HWTSTAMP_FILTER_PTP_V2_L2_DELAY_REQ:
+	case HWTSTAMP_FILTER_PTP_V2_EVENT:
+	case HWTSTAMP_FILTER_PTP_V2_SYNC:
+	case HWTSTAMP_FILTER_PTP_V2_DELAY_REQ:
+	case HWTSTAMP_FILTER_NTP_ALL:
+		config.rx_filter = HWTSTAMP_FILTER_ALL;
+		break;
+	default:
+		return -ERANGE;
+	}
+
+	if (mlx4_en_reset_config(dev, config, dev->features)) {
+		config.tx_type = HWTSTAMP_TX_OFF;
+		config.rx_filter = HWTSTAMP_FILTER_NONE;
+	}
+
+	return copy_to_user(ifr->ifr_data, &config,
+			    sizeof(config)) ? -EFAULT : 0;
+}
+
+static int mlx4_en_hwtstamp_get(struct net_device *dev, struct ifreq *ifr)
+{
+	struct mlx4_en_priv *priv = netdev_priv(dev);
+
+	return copy_to_user(ifr->ifr_data, &priv->hwtstamp_config,
+			    sizeof(priv->hwtstamp_config)) ? -EFAULT : 0;
+}
+
+static int mlx4_en_ioctl(struct net_device *dev, struct ifreq *ifr, int cmd)
+{
+	switch (cmd) {
+	case SIOCSHWTSTAMP:
+		return mlx4_en_hwtstamp_set(dev, ifr);
+	case SIOCGHWTSTAMP:
+		return mlx4_en_hwtstamp_get(dev, ifr);
+	default:
+		return -EOPNOTSUPP;
+	}
+}
+
+static netdev_features_t mlx4_en_fix_features(struct net_device *netdev,
+					      netdev_features_t features)
+{
+	struct mlx4_en_priv *en_priv = netdev_priv(netdev);
+	struct mlx4_en_dev *mdev = en_priv->mdev;
+
+	/* Since there is no support for separate RX C-TAG/S-TAG vlan accel
+	 * enable/disable make sure S-TAG flag is always in same state as
+	 * C-TAG.
+	 */
+	if (features & NETIF_F_HW_VLAN_CTAG_RX &&
+	    !(mdev->dev->caps.flags2 & MLX4_DEV_CAP_FLAG2_SKIP_OUTER_VLAN))
+		features |= NETIF_F_HW_VLAN_STAG_RX;
+	else
+		features &= ~NETIF_F_HW_VLAN_STAG_RX;
+
+	return features;
+}
+
+static int mlx4_en_set_features(struct net_device *netdev,
+		netdev_features_t features)
+{
+	struct mlx4_en_priv *priv = netdev_priv(netdev);
+	bool reset = false;
+	int ret = 0;
+
+	if (DEV_FEATURE_CHANGED(netdev, features, NETIF_F_RXFCS)) {
+		en_info(priv, "Turn %s RX-FCS\n",
+			(features & NETIF_F_RXFCS) ? "ON" : "OFF");
+		reset = true;
+	}
+
+	if (DEV_FEATURE_CHANGED(netdev, features, NETIF_F_RXALL)) {
+		u8 ignore_fcs_value = (features & NETIF_F_RXALL) ? 1 : 0;
+
+		en_info(priv, "Turn %s RX-ALL\n",
+			ignore_fcs_value ? "ON" : "OFF");
+		ret = mlx4_SET_PORT_fcs_check(priv->mdev->dev,
+					      priv->port, ignore_fcs_value);
+		if (ret)
+			return ret;
+	}
+
+	if (DEV_FEATURE_CHANGED(netdev, features, NETIF_F_HW_VLAN_CTAG_RX)) {
+		en_info(priv, "Turn %s RX vlan strip offload\n",
+			(features & NETIF_F_HW_VLAN_CTAG_RX) ? "ON" : "OFF");
+		reset = true;
+	}
+
+	if (DEV_FEATURE_CHANGED(netdev, features, NETIF_F_HW_VLAN_CTAG_TX))
+		en_info(priv, "Turn %s TX vlan strip offload\n",
+			(features & NETIF_F_HW_VLAN_CTAG_TX) ? "ON" : "OFF");
+
+	if (DEV_FEATURE_CHANGED(netdev, features, NETIF_F_HW_VLAN_STAG_TX))
+		en_info(priv, "Turn %s TX S-VLAN strip offload\n",
+			(features & NETIF_F_HW_VLAN_STAG_TX) ? "ON" : "OFF");
+
+	if (DEV_FEATURE_CHANGED(netdev, features, NETIF_F_LOOPBACK)) {
+		en_info(priv, "Turn %s loopback\n",
+			(features & NETIF_F_LOOPBACK) ? "ON" : "OFF");
+		mlx4_en_update_loopback_state(netdev, features);
+	}
+
+	if (reset) {
+		ret = mlx4_en_reset_config(netdev, priv->hwtstamp_config,
+					   features);
+		if (ret)
+			return ret;
+	}
+
+	return 0;
+}
+
+static int mlx4_en_set_vf_mac(struct net_device *dev, int queue, u8 *mac)
+{
+	struct mlx4_en_priv *en_priv = netdev_priv(dev);
+	struct mlx4_en_dev *mdev = en_priv->mdev;
+
+	return mlx4_set_vf_mac(mdev->dev, en_priv->port, queue, mac);
+}
+
+static int mlx4_en_set_vf_vlan(struct net_device *dev, int vf, u16 vlan, u8 qos,
+			       __be16 vlan_proto)
+{
+	struct mlx4_en_priv *en_priv = netdev_priv(dev);
+	struct mlx4_en_dev *mdev = en_priv->mdev;
+
+	return mlx4_set_vf_vlan(mdev->dev, en_priv->port, vf, vlan, qos,
+				vlan_proto);
+}
+
+static int mlx4_en_set_vf_rate(struct net_device *dev, int vf, int min_tx_rate,
+			       int max_tx_rate)
+{
+	struct mlx4_en_priv *en_priv = netdev_priv(dev);
+	struct mlx4_en_dev *mdev = en_priv->mdev;
+
+	return mlx4_set_vf_rate(mdev->dev, en_priv->port, vf, min_tx_rate,
+				max_tx_rate);
+}
+
+static int mlx4_en_set_vf_spoofchk(struct net_device *dev, int vf, bool setting)
+{
+	struct mlx4_en_priv *en_priv = netdev_priv(dev);
+	struct mlx4_en_dev *mdev = en_priv->mdev;
+
+	return mlx4_set_vf_spoofchk(mdev->dev, en_priv->port, vf, setting);
+}
+
+static int mlx4_en_get_vf_config(struct net_device *dev, int vf, struct ifla_vf_info *ivf)
+{
+	struct mlx4_en_priv *en_priv = netdev_priv(dev);
+	struct mlx4_en_dev *mdev = en_priv->mdev;
+
+	return mlx4_get_vf_config(mdev->dev, en_priv->port, vf, ivf);
+}
+
+static int mlx4_en_set_vf_link_state(struct net_device *dev, int vf, int link_state)
+{
+	struct mlx4_en_priv *en_priv = netdev_priv(dev);
+	struct mlx4_en_dev *mdev = en_priv->mdev;
+
+	return mlx4_set_vf_link_state(mdev->dev, en_priv->port, vf, link_state);
+}
+
+static int mlx4_en_get_vf_stats(struct net_device *dev, int vf,
+				struct ifla_vf_stats *vf_stats)
+{
+	struct mlx4_en_priv *en_priv = netdev_priv(dev);
+	struct mlx4_en_dev *mdev = en_priv->mdev;
+
+	return mlx4_get_vf_stats(mdev->dev, en_priv->port, vf, vf_stats);
+}
+
+#define PORT_ID_BYTE_LEN 8
+static int mlx4_en_get_phys_port_id(struct net_device *dev,
+				    struct netdev_phys_item_id *ppid)
+{
+	struct mlx4_en_priv *priv = netdev_priv(dev);
+	struct mlx4_dev *mdev = priv->mdev->dev;
+	int i;
+	u64 phys_port_id = mdev->caps.phys_port_id[priv->port];
+
+	if (!phys_port_id)
+		return -EOPNOTSUPP;
+
+	ppid->id_len = sizeof(phys_port_id);
+	for (i = PORT_ID_BYTE_LEN - 1; i >= 0; --i) {
+		ppid->id[i] =  phys_port_id & 0xff;
+		phys_port_id >>= 8;
+	}
+	return 0;
+}
+
+static void mlx4_en_add_vxlan_offloads(struct work_struct *work)
+{
+	int ret;
+	struct mlx4_en_priv *priv = container_of(work, struct mlx4_en_priv,
+						 vxlan_add_task);
+
+	ret = mlx4_config_vxlan_port(priv->mdev->dev, priv->vxlan_port);
+	if (ret)
+		goto out;
+
+	ret = mlx4_SET_PORT_VXLAN(priv->mdev->dev, priv->port,
+				  VXLAN_STEER_BY_OUTER_MAC, 1);
+out:
+	if (ret) {
+		en_err(priv, "failed setting L2 tunnel configuration ret %d\n", ret);
+		return;
+	}
+
+	/* set offloads */
+	priv->dev->hw_enc_features |= NETIF_F_IP_CSUM | NETIF_F_IPV6_CSUM |
+				      NETIF_F_RXCSUM |
+				      NETIF_F_TSO | NETIF_F_TSO6 |
+				      NETIF_F_GSO_UDP_TUNNEL |
+				      NETIF_F_GSO_UDP_TUNNEL_CSUM |
+				      NETIF_F_GSO_PARTIAL;
+}
+
+static void mlx4_en_del_vxlan_offloads(struct work_struct *work)
+{
+	int ret;
+	struct mlx4_en_priv *priv = container_of(work, struct mlx4_en_priv,
+						 vxlan_del_task);
+	/* unset offloads */
+	priv->dev->hw_enc_features &= ~(NETIF_F_IP_CSUM | NETIF_F_IPV6_CSUM |
+					NETIF_F_RXCSUM |
+					NETIF_F_TSO | NETIF_F_TSO6 |
+					NETIF_F_GSO_UDP_TUNNEL |
+					NETIF_F_GSO_UDP_TUNNEL_CSUM |
+					NETIF_F_GSO_PARTIAL);
+
+	ret = mlx4_SET_PORT_VXLAN(priv->mdev->dev, priv->port,
+				  VXLAN_STEER_BY_OUTER_MAC, 0);
+	if (ret)
+		en_err(priv, "failed setting L2 tunnel configuration ret %d\n", ret);
+
+	priv->vxlan_port = 0;
+}
+
+static void mlx4_en_add_vxlan_port(struct  net_device *dev,
+				   struct udp_tunnel_info *ti)
+{
+	struct mlx4_en_priv *priv = netdev_priv(dev);
+	__be16 port = ti->port;
+	__be16 current_port;
+
+	if (ti->type != UDP_TUNNEL_TYPE_VXLAN)
+		return;
+
+	if (ti->sa_family != AF_INET)
+		return;
+
+	if (priv->mdev->dev->caps.tunnel_offload_mode != MLX4_TUNNEL_OFFLOAD_MODE_VXLAN)
+		return;
+
+	current_port = priv->vxlan_port;
+	if (current_port && current_port != port) {
+		en_warn(priv, "vxlan port %d configured, can't add port %d\n",
+			ntohs(current_port), ntohs(port));
+		return;
+	}
+
+	priv->vxlan_port = port;
+	queue_work(priv->mdev->workqueue, &priv->vxlan_add_task);
+}
+
+static void mlx4_en_del_vxlan_port(struct  net_device *dev,
+				   struct udp_tunnel_info *ti)
+{
+	struct mlx4_en_priv *priv = netdev_priv(dev);
+	__be16 port = ti->port;
+	__be16 current_port;
+
+	if (ti->type != UDP_TUNNEL_TYPE_VXLAN)
+		return;
+
+	if (ti->sa_family != AF_INET)
+		return;
+
+	if (priv->mdev->dev->caps.tunnel_offload_mode != MLX4_TUNNEL_OFFLOAD_MODE_VXLAN)
+		return;
+
+	current_port = priv->vxlan_port;
+	if (current_port != port) {
+		en_dbg(DRV, priv, "vxlan port %d isn't configured, ignoring\n", ntohs(port));
+		return;
+	}
+
+	queue_work(priv->mdev->workqueue, &priv->vxlan_del_task);
+}
+
+static netdev_features_t mlx4_en_features_check(struct sk_buff *skb,
+						struct net_device *dev,
+						netdev_features_t features)
+{
+	features = vlan_features_check(skb, features);
+	features = vxlan_features_check(skb, features);
+
+	/* The ConnectX-3 doesn't support outer IPv6 checksums but it does
+	 * support inner IPv6 checksums and segmentation so  we need to
+	 * strip that feature if this is an IPv6 encapsulated frame.
+	 */
+	if (skb->encapsulation &&
+	    (skb->ip_summed == CHECKSUM_PARTIAL)) {
+		struct mlx4_en_priv *priv = netdev_priv(dev);
+
+		if (!priv->vxlan_port ||
+		    (ip_hdr(skb)->version != 4) ||
+		    (udp_hdr(skb)->dest != priv->vxlan_port))
+			features &= ~(NETIF_F_CSUM_MASK | NETIF_F_GSO_MASK);
+	}
+
+	return features;
+}
+
+static int mlx4_en_set_tx_maxrate(struct net_device *dev, int queue_index, u32 maxrate)
+{
+	struct mlx4_en_priv *priv = netdev_priv(dev);
+	struct mlx4_en_tx_ring *tx_ring = priv->tx_ring[TX][queue_index];
+	struct mlx4_update_qp_params params;
+	int err;
+
+	if (!(priv->mdev->dev->caps.flags2 & MLX4_DEV_CAP_FLAG2_QP_RATE_LIMIT))
+		return -EOPNOTSUPP;
+
+	/* rate provided to us in Mbs, check if it fits into 12 bits, if not use Gbs */
+	if (maxrate >> 12) {
+		params.rate_unit = MLX4_QP_RATE_LIMIT_GBS;
+		params.rate_val  = maxrate / 1000;
+	} else if (maxrate) {
+		params.rate_unit = MLX4_QP_RATE_LIMIT_MBS;
+		params.rate_val  = maxrate;
+	} else { /* zero serves to revoke the QP rate-limitation */
+		params.rate_unit = 0;
+		params.rate_val  = 0;
+	}
+
+	err = mlx4_update_qp(priv->mdev->dev, tx_ring->qpn, MLX4_UPDATE_QP_RATE_LIMIT,
+			     &params);
+	return err;
+}
+
+static int mlx4_xdp_set(struct net_device *dev, struct bpf_prog *prog)
+{
+	struct mlx4_en_priv *priv = netdev_priv(dev);
+	struct mlx4_en_dev *mdev = priv->mdev;
+	struct mlx4_en_port_profile new_prof;
+	struct bpf_prog *old_prog;
+	struct mlx4_en_priv *tmp;
+	int tx_changed = 0;
+	int xdp_ring_num;
+	int port_up = 0;
+	int err;
+	int i;
+
+	xdp_ring_num = prog ? priv->rx_ring_num : 0;
+
+	/* No need to reconfigure buffers when simply swapping the
+	 * program for a new one.
+	 */
+	if (priv->tx_ring_num[TX_XDP] == xdp_ring_num) {
+		if (prog) {
+			prog = bpf_prog_add(prog, priv->rx_ring_num - 1);
+			if (IS_ERR(prog))
+				return PTR_ERR(prog);
+		}
+		mutex_lock(&mdev->state_lock);
+		for (i = 0; i < priv->rx_ring_num; i++) {
+			old_prog = rcu_dereference_protected(
+					priv->rx_ring[i]->xdp_prog,
+					lockdep_is_held(&mdev->state_lock));
+			rcu_assign_pointer(priv->rx_ring[i]->xdp_prog, prog);
+			if (old_prog)
+				bpf_prog_put(old_prog);
+		}
+		mutex_unlock(&mdev->state_lock);
+		return 0;
+	}
+
+	if (!mlx4_en_check_xdp_mtu(dev, dev->mtu))
+		return -EOPNOTSUPP;
+
+	tmp = kzalloc(sizeof(*tmp), GFP_KERNEL);
+	if (!tmp)
+		return -ENOMEM;
+
+	if (prog) {
+		prog = bpf_prog_add(prog, priv->rx_ring_num - 1);
+		if (IS_ERR(prog)) {
+			err = PTR_ERR(prog);
+			goto out;
+		}
+	}
+
+	mutex_lock(&mdev->state_lock);
+	memcpy(&new_prof, priv->prof, sizeof(struct mlx4_en_port_profile));
+	new_prof.tx_ring_num[TX_XDP] = xdp_ring_num;
+
+	if (priv->tx_ring_num[TX] + xdp_ring_num > MAX_TX_RINGS) {
+		tx_changed = 1;
+		new_prof.tx_ring_num[TX] =
+			MAX_TX_RINGS - ALIGN(xdp_ring_num, priv->prof->num_up);
+		en_warn(priv, "Reducing the number of TX rings, to not exceed the max total rings number.\n");
+	}
+
+	err = mlx4_en_try_alloc_resources(priv, tmp, &new_prof, false);
+	if (err) {
+		if (prog)
+			bpf_prog_sub(prog, priv->rx_ring_num - 1);
+		goto unlock_out;
+	}
+
+	if (priv->port_up) {
+		port_up = 1;
+		mlx4_en_stop_port(dev, 1);
+	}
+
+	mlx4_en_safe_replace_resources(priv, tmp);
+	if (tx_changed)
+		netif_set_real_num_tx_queues(dev, priv->tx_ring_num[TX]);
+
+	for (i = 0; i < priv->rx_ring_num; i++) {
+		old_prog = rcu_dereference_protected(
+					priv->rx_ring[i]->xdp_prog,
+					lockdep_is_held(&mdev->state_lock));
+		rcu_assign_pointer(priv->rx_ring[i]->xdp_prog, prog);
+		if (old_prog)
+			bpf_prog_put(old_prog);
+	}
+
+	if (port_up) {
+		err = mlx4_en_start_port(dev);
+		if (err) {
+			en_err(priv, "Failed starting port %d for XDP change\n",
+			       priv->port);
+			if (!test_and_set_bit(MLX4_EN_STATE_FLAG_RESTARTING, &priv->state))
+				queue_work(mdev->workqueue, &priv->restart_task);
+		}
+	}
+
+unlock_out:
+	mutex_unlock(&mdev->state_lock);
+out:
+	kfree(tmp);
+	return err;
+}
+
+static u32 mlx4_xdp_query(struct net_device *dev)
+{
+	struct mlx4_en_priv *priv = netdev_priv(dev);
+	struct mlx4_en_dev *mdev = priv->mdev;
+	const struct bpf_prog *xdp_prog;
+	u32 prog_id = 0;
+
+	if (!priv->tx_ring_num[TX_XDP])
+		return prog_id;
+
+	mutex_lock(&mdev->state_lock);
+	xdp_prog = rcu_dereference_protected(
+		priv->rx_ring[0]->xdp_prog,
+		lockdep_is_held(&mdev->state_lock));
+	if (xdp_prog)
+		prog_id = xdp_prog->aux->id;
+	mutex_unlock(&mdev->state_lock);
+
+	return prog_id;
+}
+
+static int mlx4_xdp(struct net_device *dev, struct netdev_bpf *xdp)
+{
+	switch (xdp->command) {
+	case XDP_SETUP_PROG:
+		return mlx4_xdp_set(dev, xdp->prog);
+	case XDP_QUERY_PROG:
+		xdp->prog_id = mlx4_xdp_query(dev);
+		return 0;
+	default:
+		return -EINVAL;
+	}
+}
+
+static const struct net_device_ops mlx4_netdev_ops = {
+	.ndo_open		= mlx4_en_open,
+	.ndo_stop		= mlx4_en_close,
+	.ndo_start_xmit		= mlx4_en_xmit,
+	.ndo_select_queue	= mlx4_en_select_queue,
+	.ndo_get_stats64	= mlx4_en_get_stats64,
+	.ndo_set_rx_mode	= mlx4_en_set_rx_mode,
+	.ndo_set_mac_address	= mlx4_en_set_mac,
+	.ndo_validate_addr	= eth_validate_addr,
+	.ndo_change_mtu		= mlx4_en_change_mtu,
+	.ndo_do_ioctl		= mlx4_en_ioctl,
+	.ndo_tx_timeout		= mlx4_en_tx_timeout,
+	.ndo_vlan_rx_add_vid	= mlx4_en_vlan_rx_add_vid,
+	.ndo_vlan_rx_kill_vid	= mlx4_en_vlan_rx_kill_vid,
+	.ndo_set_features	= mlx4_en_set_features,
+	.ndo_fix_features	= mlx4_en_fix_features,
+	.ndo_setup_tc		= __mlx4_en_setup_tc,
+#ifdef CONFIG_RFS_ACCEL
+	.ndo_rx_flow_steer	= mlx4_en_filter_rfs,
+#endif
+	.ndo_get_phys_port_id	= mlx4_en_get_phys_port_id,
+	.ndo_udp_tunnel_add	= mlx4_en_add_vxlan_port,
+	.ndo_udp_tunnel_del	= mlx4_en_del_vxlan_port,
+	.ndo_features_check	= mlx4_en_features_check,
+	.ndo_set_tx_maxrate	= mlx4_en_set_tx_maxrate,
+	.ndo_bpf		= mlx4_xdp,
+};
+
+static const struct net_device_ops mlx4_netdev_ops_master = {
+	.ndo_open		= mlx4_en_open,
+	.ndo_stop		= mlx4_en_close,
+	.ndo_start_xmit		= mlx4_en_xmit,
+	.ndo_select_queue	= mlx4_en_select_queue,
+	.ndo_get_stats64	= mlx4_en_get_stats64,
+	.ndo_set_rx_mode	= mlx4_en_set_rx_mode,
+	.ndo_set_mac_address	= mlx4_en_set_mac,
+	.ndo_validate_addr	= eth_validate_addr,
+	.ndo_change_mtu		= mlx4_en_change_mtu,
+	.ndo_tx_timeout		= mlx4_en_tx_timeout,
+	.ndo_vlan_rx_add_vid	= mlx4_en_vlan_rx_add_vid,
+	.ndo_vlan_rx_kill_vid	= mlx4_en_vlan_rx_kill_vid,
+	.ndo_set_vf_mac		= mlx4_en_set_vf_mac,
+	.ndo_set_vf_vlan	= mlx4_en_set_vf_vlan,
+	.ndo_set_vf_rate	= mlx4_en_set_vf_rate,
+	.ndo_set_vf_spoofchk	= mlx4_en_set_vf_spoofchk,
+	.ndo_set_vf_link_state	= mlx4_en_set_vf_link_state,
+	.ndo_get_vf_stats       = mlx4_en_get_vf_stats,
+	.ndo_get_vf_config	= mlx4_en_get_vf_config,
+	.ndo_set_features	= mlx4_en_set_features,
+	.ndo_fix_features	= mlx4_en_fix_features,
+	.ndo_setup_tc		= __mlx4_en_setup_tc,
+#ifdef CONFIG_RFS_ACCEL
+	.ndo_rx_flow_steer	= mlx4_en_filter_rfs,
+#endif
+	.ndo_get_phys_port_id	= mlx4_en_get_phys_port_id,
+	.ndo_udp_tunnel_add	= mlx4_en_add_vxlan_port,
+	.ndo_udp_tunnel_del	= mlx4_en_del_vxlan_port,
+	.ndo_features_check	= mlx4_en_features_check,
+	.ndo_set_tx_maxrate	= mlx4_en_set_tx_maxrate,
+	.ndo_bpf		= mlx4_xdp,
+};
+
+struct mlx4_en_bond {
+	struct work_struct work;
+	struct mlx4_en_priv *priv;
+	int is_bonded;
+	struct mlx4_port_map port_map;
+};
+
+static void mlx4_en_bond_work(struct work_struct *work)
+{
+	struct mlx4_en_bond *bond = container_of(work,
+						     struct mlx4_en_bond,
+						     work);
+	int err = 0;
+	struct mlx4_dev *dev = bond->priv->mdev->dev;
+
+	if (bond->is_bonded) {
+		if (!mlx4_is_bonded(dev)) {
+			err = mlx4_bond(dev);
+			if (err)
+				en_err(bond->priv, "Fail to bond device\n");
+		}
+		if (!err) {
+			err = mlx4_port_map_set(dev, &bond->port_map);
+			if (err)
+				en_err(bond->priv, "Fail to set port map [%d][%d]: %d\n",
+				       bond->port_map.port1,
+				       bond->port_map.port2,
+				       err);
+		}
+	} else if (mlx4_is_bonded(dev)) {
+		err = mlx4_unbond(dev);
+		if (err)
+			en_err(bond->priv, "Fail to unbond device\n");
+	}
+	dev_put(bond->priv->dev);
+	kfree(bond);
+}
+
+static int mlx4_en_queue_bond_work(struct mlx4_en_priv *priv, int is_bonded,
+				   u8 v2p_p1, u8 v2p_p2)
+{
+	struct mlx4_en_bond *bond = NULL;
+
+	bond = kzalloc(sizeof(*bond), GFP_ATOMIC);
+	if (!bond)
+		return -ENOMEM;
+
+	INIT_WORK(&bond->work, mlx4_en_bond_work);
+	bond->priv = priv;
+	bond->is_bonded = is_bonded;
+	bond->port_map.port1 = v2p_p1;
+	bond->port_map.port2 = v2p_p2;
+	dev_hold(priv->dev);
+	queue_work(priv->mdev->workqueue, &bond->work);
+	return 0;
+}
+
+int mlx4_en_netdev_event(struct notifier_block *this,
+			 unsigned long event, void *ptr)
+{
+	struct net_device *ndev = netdev_notifier_info_to_dev(ptr);
+	u8 port = 0;
+	struct mlx4_en_dev *mdev;
+	struct mlx4_dev *dev;
+	int i, num_eth_ports = 0;
+	bool do_bond = true;
+	struct mlx4_en_priv *priv;
+	u8 v2p_port1 = 0;
+	u8 v2p_port2 = 0;
+
+	if (!net_eq(dev_net(ndev), &init_net))
+		return NOTIFY_DONE;
+
+	mdev = container_of(this, struct mlx4_en_dev, nb);
+	dev = mdev->dev;
+
+	/* Go into this mode only when two network devices set on two ports
+	 * of the same mlx4 device are slaves of the same bonding master
+	 */
+	mlx4_foreach_port(i, dev, MLX4_PORT_TYPE_ETH) {
+		++num_eth_ports;
+		if (!port && (mdev->pndev[i] == ndev))
+			port = i;
+		mdev->upper[i] = mdev->pndev[i] ?
+			netdev_master_upper_dev_get(mdev->pndev[i]) : NULL;
+		/* condition not met: network device is a slave */
+		if (!mdev->upper[i])
+			do_bond = false;
+		if (num_eth_ports < 2)
+			continue;
+		/* condition not met: same master */
+		if (mdev->upper[i] != mdev->upper[i-1])
+			do_bond = false;
+	}
+	/* condition not met: 2 salves */
+	do_bond = (num_eth_ports ==  2) ? do_bond : false;
+
+	/* handle only events that come with enough info */
+	if ((do_bond && (event != NETDEV_BONDING_INFO)) || !port)
+		return NOTIFY_DONE;
+
+	priv = netdev_priv(ndev);
+	if (do_bond) {
+		struct netdev_notifier_bonding_info *notifier_info = ptr;
+		struct netdev_bonding_info *bonding_info =
+			&notifier_info->bonding_info;
+
+		/* required mode 1, 2 or 4 */
+		if ((bonding_info->master.bond_mode != BOND_MODE_ACTIVEBACKUP) &&
+		    (bonding_info->master.bond_mode != BOND_MODE_XOR) &&
+		    (bonding_info->master.bond_mode != BOND_MODE_8023AD))
+			do_bond = false;
+
+		/* require exactly 2 slaves */
+		if (bonding_info->master.num_slaves != 2)
+			do_bond = false;
+
+		/* calc v2p */
+		if (do_bond) {
+			if (bonding_info->master.bond_mode ==
+			    BOND_MODE_ACTIVEBACKUP) {
+				/* in active-backup mode virtual ports are
+				 * mapped to the physical port of the active
+				 * slave */
+				if (bonding_info->slave.state ==
+				    BOND_STATE_BACKUP) {
+					if (port == 1) {
+						v2p_port1 = 2;
+						v2p_port2 = 2;
+					} else {
+						v2p_port1 = 1;
+						v2p_port2 = 1;
+					}
+				} else { /* BOND_STATE_ACTIVE */
+					if (port == 1) {
+						v2p_port1 = 1;
+						v2p_port2 = 1;
+					} else {
+						v2p_port1 = 2;
+						v2p_port2 = 2;
+					}
+				}
+			} else { /* Active-Active */
+				/* in active-active mode a virtual port is
+				 * mapped to the native physical port if and only
+				 * if the physical port is up */
+				__s8 link = bonding_info->slave.link;
+
+				if (port == 1)
+					v2p_port2 = 2;
+				else
+					v2p_port1 = 1;
+				if ((link == BOND_LINK_UP) ||
+				    (link == BOND_LINK_FAIL)) {
+					if (port == 1)
+						v2p_port1 = 1;
+					else
+						v2p_port2 = 2;
+				} else { /* BOND_LINK_DOWN || BOND_LINK_BACK */
+					if (port == 1)
+						v2p_port1 = 2;
+					else
+						v2p_port2 = 1;
+				}
+			}
+		}
+	}
+
+	mlx4_en_queue_bond_work(priv, do_bond,
+				v2p_port1, v2p_port2);
+
+	return NOTIFY_DONE;
+}
+
+void mlx4_en_update_pfc_stats_bitmap(struct mlx4_dev *dev,
+				     struct mlx4_en_stats_bitmap *stats_bitmap,
+				     u8 rx_ppp, u8 rx_pause,
+				     u8 tx_ppp, u8 tx_pause)
+{
+	int last_i = NUM_MAIN_STATS + NUM_PORT_STATS + NUM_PF_STATS;
+
+	if (!mlx4_is_slave(dev) &&
+	    (dev->caps.flags2 & MLX4_DEV_CAP_FLAG2_FLOWSTATS_EN)) {
+		mutex_lock(&stats_bitmap->mutex);
+		bitmap_clear(stats_bitmap->bitmap, last_i, NUM_FLOW_STATS);
+
+		if (rx_ppp)
+			bitmap_set(stats_bitmap->bitmap, last_i,
+				   NUM_FLOW_PRIORITY_STATS_RX);
+		last_i += NUM_FLOW_PRIORITY_STATS_RX;
+
+		if (rx_pause && !(rx_ppp))
+			bitmap_set(stats_bitmap->bitmap, last_i,
+				   NUM_FLOW_STATS_RX);
+		last_i += NUM_FLOW_STATS_RX;
+
+		if (tx_ppp)
+			bitmap_set(stats_bitmap->bitmap, last_i,
+				   NUM_FLOW_PRIORITY_STATS_TX);
+		last_i += NUM_FLOW_PRIORITY_STATS_TX;
+
+		if (tx_pause && !(tx_ppp))
+			bitmap_set(stats_bitmap->bitmap, last_i,
+				   NUM_FLOW_STATS_TX);
+		last_i += NUM_FLOW_STATS_TX;
+
+		mutex_unlock(&stats_bitmap->mutex);
+	}
+}
+
+void mlx4_en_set_stats_bitmap(struct mlx4_dev *dev,
+			      struct mlx4_en_stats_bitmap *stats_bitmap,
+			      u8 rx_ppp, u8 rx_pause,
+			      u8 tx_ppp, u8 tx_pause)
+{
+	int last_i = 0;
+
+	mutex_init(&stats_bitmap->mutex);
+	bitmap_zero(stats_bitmap->bitmap, NUM_ALL_STATS);
+
+	if (mlx4_is_slave(dev)) {
+		bitmap_set(stats_bitmap->bitmap, last_i +
+					 MLX4_FIND_NETDEV_STAT(rx_packets), 1);
+		bitmap_set(stats_bitmap->bitmap, last_i +
+					 MLX4_FIND_NETDEV_STAT(tx_packets), 1);
+		bitmap_set(stats_bitmap->bitmap, last_i +
+					 MLX4_FIND_NETDEV_STAT(rx_bytes), 1);
+		bitmap_set(stats_bitmap->bitmap, last_i +
+					 MLX4_FIND_NETDEV_STAT(tx_bytes), 1);
+		bitmap_set(stats_bitmap->bitmap, last_i +
+					 MLX4_FIND_NETDEV_STAT(rx_dropped), 1);
+		bitmap_set(stats_bitmap->bitmap, last_i +
+					 MLX4_FIND_NETDEV_STAT(tx_dropped), 1);
+	} else {
+		bitmap_set(stats_bitmap->bitmap, last_i, NUM_MAIN_STATS);
+	}
+	last_i += NUM_MAIN_STATS;
+
+	bitmap_set(stats_bitmap->bitmap, last_i, NUM_PORT_STATS);
+	last_i += NUM_PORT_STATS;
+
+	if (mlx4_is_master(dev))
+		bitmap_set(stats_bitmap->bitmap, last_i,
+			   NUM_PF_STATS);
+	last_i += NUM_PF_STATS;
+
+	mlx4_en_update_pfc_stats_bitmap(dev, stats_bitmap,
+					rx_ppp, rx_pause,
+					tx_ppp, tx_pause);
+	last_i += NUM_FLOW_STATS;
+
+	if (!mlx4_is_slave(dev))
+		bitmap_set(stats_bitmap->bitmap, last_i, NUM_PKT_STATS);
+	last_i += NUM_PKT_STATS;
+
+	bitmap_set(stats_bitmap->bitmap, last_i, NUM_XDP_STATS);
+	last_i += NUM_XDP_STATS;
+
+	if (!mlx4_is_slave(dev))
+		bitmap_set(stats_bitmap->bitmap, last_i, NUM_PHY_STATS);
+	last_i += NUM_PHY_STATS;
+}
+
+int mlx4_en_init_netdev(struct mlx4_en_dev *mdev, int port,
+			struct mlx4_en_port_profile *prof)
+{
+	struct net_device *dev;
+	struct mlx4_en_priv *priv;
+	int i, t;
+	int err;
+
+	dev = alloc_etherdev_mqs(sizeof(struct mlx4_en_priv),
+				 MAX_TX_RINGS, MAX_RX_RINGS);
+	if (dev == NULL)
+		return -ENOMEM;
+
+	netif_set_real_num_tx_queues(dev, prof->tx_ring_num[TX]);
+	netif_set_real_num_rx_queues(dev, prof->rx_ring_num);
+
+	SET_NETDEV_DEV(dev, &mdev->dev->persist->pdev->dev);
+	dev->dev_port = port - 1;
+
+	/*
+	 * Initialize driver private data
+	 */
+
+	priv = netdev_priv(dev);
+	memset(priv, 0, sizeof(struct mlx4_en_priv));
+	priv->counter_index = MLX4_SINK_COUNTER_INDEX(mdev->dev);
+	spin_lock_init(&priv->stats_lock);
+	INIT_WORK(&priv->rx_mode_task, mlx4_en_do_set_rx_mode);
+	INIT_WORK(&priv->restart_task, mlx4_en_restart);
+	INIT_WORK(&priv->linkstate_task, mlx4_en_linkstate);
+	INIT_DELAYED_WORK(&priv->stats_task, mlx4_en_do_get_stats);
+	INIT_DELAYED_WORK(&priv->service_task, mlx4_en_service_task);
+	INIT_WORK(&priv->vxlan_add_task, mlx4_en_add_vxlan_offloads);
+	INIT_WORK(&priv->vxlan_del_task, mlx4_en_del_vxlan_offloads);
+#ifdef CONFIG_RFS_ACCEL
+	INIT_LIST_HEAD(&priv->filters);
+	spin_lock_init(&priv->filters_lock);
+#endif
+
+	priv->dev = dev;
+	priv->mdev = mdev;
+	priv->ddev = &mdev->pdev->dev;
+	priv->prof = prof;
+	priv->port = port;
+	priv->port_up = false;
+	priv->flags = prof->flags;
+	priv->pflags = MLX4_EN_PRIV_FLAGS_BLUEFLAME;
+	priv->ctrl_flags = cpu_to_be32(MLX4_WQE_CTRL_CQ_UPDATE |
+			MLX4_WQE_CTRL_SOLICITED);
+	priv->num_tx_rings_p_up = mdev->profile.max_num_tx_rings_p_up;
+	priv->tx_work_limit = MLX4_EN_DEFAULT_TX_WORK;
+	netdev_rss_key_fill(priv->rss_key, sizeof(priv->rss_key));
+
+	for (t = 0; t < MLX4_EN_NUM_TX_TYPES; t++) {
+		priv->tx_ring_num[t] = prof->tx_ring_num[t];
+		if (!priv->tx_ring_num[t])
+			continue;
+
+		priv->tx_ring[t] = kcalloc(MAX_TX_RINGS,
+					   sizeof(struct mlx4_en_tx_ring *),
+					   GFP_KERNEL);
+		if (!priv->tx_ring[t]) {
+			err = -ENOMEM;
+			goto out;
+		}
+		priv->tx_cq[t] = kcalloc(MAX_TX_RINGS,
+					 sizeof(struct mlx4_en_cq *),
+					 GFP_KERNEL);
+		if (!priv->tx_cq[t]) {
+			err = -ENOMEM;
+			goto out;
+		}
+	}
+	priv->rx_ring_num = prof->rx_ring_num;
+	priv->cqe_factor = (mdev->dev->caps.cqe_size == 64) ? 1 : 0;
+	priv->cqe_size = mdev->dev->caps.cqe_size;
+	priv->mac_index = -1;
+	priv->msg_enable = MLX4_EN_MSG_LEVEL;
+#ifdef CONFIG_MLX4_EN_DCB
+	if (!mlx4_is_slave(priv->mdev->dev)) {
+		u8 prio;
+
+		for (prio = 0; prio < IEEE_8021QAZ_MAX_TCS; ++prio) {
+			priv->ets.prio_tc[prio] = prio;
+			priv->ets.tc_tsa[prio]  = IEEE_8021QAZ_TSA_VENDOR;
+		}
+
+		priv->dcbx_cap = DCB_CAP_DCBX_VER_CEE | DCB_CAP_DCBX_HOST |
+			DCB_CAP_DCBX_VER_IEEE;
+		priv->flags |= MLX4_EN_DCB_ENABLED;
+		priv->cee_config.pfc_state = false;
+
+		for (i = 0; i < MLX4_EN_NUM_UP_HIGH; i++)
+			priv->cee_config.dcb_pfc[i] = pfc_disabled;
+
+		if (mdev->dev->caps.flags2 & MLX4_DEV_CAP_FLAG2_ETS_CFG) {
+			dev->dcbnl_ops = &mlx4_en_dcbnl_ops;
+		} else {
+			en_info(priv, "enabling only PFC DCB ops\n");
+			dev->dcbnl_ops = &mlx4_en_dcbnl_pfc_ops;
+		}
+	}
+#endif
+
+	for (i = 0; i < MLX4_EN_MAC_HASH_SIZE; ++i)
+		INIT_HLIST_HEAD(&priv->mac_hash[i]);
+
+	/* Query for default mac and max mtu */
+	priv->max_mtu = mdev->dev->caps.eth_mtu_cap[priv->port];
+
+	if (mdev->dev->caps.rx_checksum_flags_port[priv->port] &
+	    MLX4_RX_CSUM_MODE_VAL_NON_TCP_UDP)
+		priv->flags |= MLX4_EN_FLAG_RX_CSUM_NON_TCP_UDP;
+
+	/* Set default MAC */
+	dev->addr_len = ETH_ALEN;
+	mlx4_en_u64_to_mac(dev->dev_addr, mdev->dev->caps.def_mac[priv->port]);
+	if (!is_valid_ether_addr(dev->dev_addr)) {
+		en_err(priv, "Port: %d, invalid mac burned: %pM, quiting\n",
+		       priv->port, dev->dev_addr);
+		err = -EINVAL;
+		goto out;
+	} else if (mlx4_is_slave(priv->mdev->dev) &&
+		   (priv->mdev->dev->port_random_macs & 1 << priv->port)) {
+		/* Random MAC was assigned in mlx4_slave_cap
+		 * in mlx4_core module
+		 */
+		dev->addr_assign_type |= NET_ADDR_RANDOM;
+		en_warn(priv, "Assigned random MAC address %pM\n", dev->dev_addr);
+	}
+
+	memcpy(priv->current_mac, dev->dev_addr, sizeof(priv->current_mac));
+
+	priv->stride = roundup_pow_of_two(sizeof(struct mlx4_en_rx_desc) +
+					  DS_SIZE * MLX4_EN_MAX_RX_FRAGS);
+	err = mlx4_en_alloc_resources(priv);
+	if (err)
+		goto out;
+
+	/* Initialize time stamping config */
+	priv->hwtstamp_config.flags = 0;
+	priv->hwtstamp_config.tx_type = HWTSTAMP_TX_OFF;
+	priv->hwtstamp_config.rx_filter = HWTSTAMP_FILTER_NONE;
+
+	/* Allocate page for receive rings */
+	err = mlx4_alloc_hwq_res(mdev->dev, &priv->res,
+				MLX4_EN_PAGE_SIZE);
+	if (err) {
+		en_err(priv, "Failed to allocate page for rx qps\n");
+		goto out;
+	}
+	priv->allocated = 1;
+
+	/*
+	 * Initialize netdev entry points
+	 */
+	if (mlx4_is_master(priv->mdev->dev))
+		dev->netdev_ops = &mlx4_netdev_ops_master;
+	else
+		dev->netdev_ops = &mlx4_netdev_ops;
+	dev->watchdog_timeo = MLX4_EN_WATCHDOG_TIMEOUT;
+	netif_set_real_num_tx_queues(dev, priv->tx_ring_num[TX]);
+	netif_set_real_num_rx_queues(dev, priv->rx_ring_num);
+
+	dev->ethtool_ops = &mlx4_en_ethtool_ops;
+
+	/*
+	 * Set driver features
+	 */
+	dev->hw_features = NETIF_F_SG | NETIF_F_IP_CSUM | NETIF_F_IPV6_CSUM;
+	if (mdev->LSO_support)
+		dev->hw_features |= NETIF_F_TSO | NETIF_F_TSO6;
+
+	dev->vlan_features = dev->hw_features;
+
+	dev->hw_features |= NETIF_F_RXCSUM | NETIF_F_RXHASH;
+	dev->features = dev->hw_features | NETIF_F_HIGHDMA |
+			NETIF_F_HW_VLAN_CTAG_TX | NETIF_F_HW_VLAN_CTAG_RX |
+			NETIF_F_HW_VLAN_CTAG_FILTER;
+	dev->hw_features |= NETIF_F_LOOPBACK |
+			NETIF_F_HW_VLAN_CTAG_TX | NETIF_F_HW_VLAN_CTAG_RX;
+
+	if (!(mdev->dev->caps.flags2 & MLX4_DEV_CAP_FLAG2_SKIP_OUTER_VLAN)) {
+		dev->features |= NETIF_F_HW_VLAN_STAG_RX |
+			NETIF_F_HW_VLAN_STAG_FILTER;
+		dev->hw_features |= NETIF_F_HW_VLAN_STAG_RX;
+	}
+
+	if (mlx4_is_slave(mdev->dev)) {
+		bool vlan_offload_disabled;
+		int phv;
+
+		err = get_phv_bit(mdev->dev, port, &phv);
+		if (!err && phv) {
+			dev->hw_features |= NETIF_F_HW_VLAN_STAG_TX;
+			priv->pflags |= MLX4_EN_PRIV_FLAGS_PHV;
+		}
+		err = mlx4_get_is_vlan_offload_disabled(mdev->dev, port,
+							&vlan_offload_disabled);
+		if (!err && vlan_offload_disabled) {
+			dev->hw_features &= ~(NETIF_F_HW_VLAN_CTAG_TX |
+					      NETIF_F_HW_VLAN_CTAG_RX |
+					      NETIF_F_HW_VLAN_STAG_TX |
+					      NETIF_F_HW_VLAN_STAG_RX);
+			dev->features &= ~(NETIF_F_HW_VLAN_CTAG_TX |
+					   NETIF_F_HW_VLAN_CTAG_RX |
+					   NETIF_F_HW_VLAN_STAG_TX |
+					   NETIF_F_HW_VLAN_STAG_RX);
+		}
+	} else {
+		if (mdev->dev->caps.flags2 & MLX4_DEV_CAP_FLAG2_PHV_EN &&
+		    !(mdev->dev->caps.flags2 &
+		      MLX4_DEV_CAP_FLAG2_SKIP_OUTER_VLAN))
+			dev->hw_features |= NETIF_F_HW_VLAN_STAG_TX;
+	}
+
+	if (mdev->dev->caps.flags & MLX4_DEV_CAP_FLAG_FCS_KEEP)
+		dev->hw_features |= NETIF_F_RXFCS;
+
+	if (mdev->dev->caps.flags2 & MLX4_DEV_CAP_FLAG2_IGNORE_FCS)
+		dev->hw_features |= NETIF_F_RXALL;
+
+	if (mdev->dev->caps.steering_mode ==
+	    MLX4_STEERING_MODE_DEVICE_MANAGED &&
+	    mdev->dev->caps.dmfs_high_steer_mode != MLX4_STEERING_DMFS_A0_STATIC)
+		dev->hw_features |= NETIF_F_NTUPLE;
+
+	if (mdev->dev->caps.steering_mode != MLX4_STEERING_MODE_A0)
+		dev->priv_flags |= IFF_UNICAST_FLT;
+
+	/* Setting a default hash function value */
+	if (mdev->dev->caps.flags2 & MLX4_DEV_CAP_FLAG2_RSS_TOP) {
+		priv->rss_hash_fn = ETH_RSS_HASH_TOP;
+	} else if (mdev->dev->caps.flags2 & MLX4_DEV_CAP_FLAG2_RSS_XOR) {
+		priv->rss_hash_fn = ETH_RSS_HASH_XOR;
+	} else {
+		en_warn(priv,
+			"No RSS hash capabilities exposed, using Toeplitz\n");
+		priv->rss_hash_fn = ETH_RSS_HASH_TOP;
+	}
+
+	if (mdev->dev->caps.tunnel_offload_mode == MLX4_TUNNEL_OFFLOAD_MODE_VXLAN) {
+		dev->hw_features |= NETIF_F_GSO_UDP_TUNNEL |
+				    NETIF_F_GSO_UDP_TUNNEL_CSUM |
+				    NETIF_F_GSO_PARTIAL;
+		dev->features    |= NETIF_F_GSO_UDP_TUNNEL |
+				    NETIF_F_GSO_UDP_TUNNEL_CSUM |
+				    NETIF_F_GSO_PARTIAL;
+		dev->gso_partial_features = NETIF_F_GSO_UDP_TUNNEL_CSUM;
+	}
+
+	/* MTU range: 68 - hw-specific max */
+	dev->min_mtu = ETH_MIN_MTU;
+	dev->max_mtu = priv->max_mtu;
+
+	mdev->pndev[port] = dev;
+	mdev->upper[port] = NULL;
+
+	netif_carrier_off(dev);
+	mlx4_en_set_default_moderation(priv);
+
+	en_warn(priv, "Using %d TX rings\n", prof->tx_ring_num[TX]);
+	en_warn(priv, "Using %d RX rings\n", prof->rx_ring_num);
+
+	mlx4_en_update_loopback_state(priv->dev, priv->dev->features);
+
+	/* Configure port */
+	mlx4_en_calc_rx_buf(dev);
+	err = mlx4_SET_PORT_general(mdev->dev, priv->port,
+				    priv->rx_skb_size + ETH_FCS_LEN,
+				    prof->tx_pause, prof->tx_ppp,
+				    prof->rx_pause, prof->rx_ppp);
+	if (err) {
+		en_err(priv, "Failed setting port general configurations for port %d, with error %d\n",
+		       priv->port, err);
+		goto out;
+	}
+
+	if (mdev->dev->caps.tunnel_offload_mode == MLX4_TUNNEL_OFFLOAD_MODE_VXLAN) {
+		err = mlx4_SET_PORT_VXLAN(mdev->dev, priv->port, VXLAN_STEER_BY_OUTER_MAC, 1);
+		if (err) {
+			en_err(priv, "Failed setting port L2 tunnel configuration, err %d\n",
+			       err);
+			goto out;
+		}
+	}
+
+	/* Init port */
+	en_warn(priv, "Initializing port\n");
+	err = mlx4_INIT_PORT(mdev->dev, priv->port);
+	if (err) {
+		en_err(priv, "Failed Initializing port\n");
+		goto out;
+	}
+	queue_delayed_work(mdev->workqueue, &priv->stats_task, STATS_DELAY);
+
+	/* Initialize time stamp mechanism */
+	if (mdev->dev->caps.flags2 & MLX4_DEV_CAP_FLAG2_TS)
+		mlx4_en_init_timestamp(mdev);
+
+	queue_delayed_work(mdev->workqueue, &priv->service_task,
+			   SERVICE_TASK_DELAY);
+
+	mlx4_en_set_stats_bitmap(mdev->dev, &priv->stats_bitmap,
+				 mdev->profile.prof[priv->port].rx_ppp,
+				 mdev->profile.prof[priv->port].rx_pause,
+				 mdev->profile.prof[priv->port].tx_ppp,
+				 mdev->profile.prof[priv->port].tx_pause);
+
+	err = register_netdev(dev);
+	if (err) {
+		en_err(priv, "Netdev registration failed for port %d\n", port);
+		goto out;
+	}
+
+	priv->registered = 1;
+	devlink_port_type_eth_set(mlx4_get_devlink_port(mdev->dev, priv->port),
+				  dev);
+
+	return 0;
+
+out:
+	mlx4_en_destroy_netdev(dev);
+	return err;
+}
+
+int mlx4_en_reset_config(struct net_device *dev,
+			 struct hwtstamp_config ts_config,
+			 netdev_features_t features)
+{
+	struct mlx4_en_priv *priv = netdev_priv(dev);
+	struct mlx4_en_dev *mdev = priv->mdev;
+	struct mlx4_en_port_profile new_prof;
+	struct mlx4_en_priv *tmp;
+	int port_up = 0;
+	int err = 0;
+
+	if (priv->hwtstamp_config.tx_type == ts_config.tx_type &&
+	    priv->hwtstamp_config.rx_filter == ts_config.rx_filter &&
+	    !DEV_FEATURE_CHANGED(dev, features, NETIF_F_HW_VLAN_CTAG_RX) &&
+	    !DEV_FEATURE_CHANGED(dev, features, NETIF_F_RXFCS))
+		return 0; /* Nothing to change */
+
+	if (DEV_FEATURE_CHANGED(dev, features, NETIF_F_HW_VLAN_CTAG_RX) &&
+	    (features & NETIF_F_HW_VLAN_CTAG_RX) &&
+	    (priv->hwtstamp_config.rx_filter != HWTSTAMP_FILTER_NONE)) {
+		en_warn(priv, "Can't turn ON rx vlan offload while time-stamping rx filter is ON\n");
+		return -EINVAL;
+	}
+
+	tmp = kzalloc(sizeof(*tmp), GFP_KERNEL);
+	if (!tmp)
+		return -ENOMEM;
+
+	mutex_lock(&mdev->state_lock);
+
+	memcpy(&new_prof, priv->prof, sizeof(struct mlx4_en_port_profile));
+	memcpy(&new_prof.hwtstamp_config, &ts_config, sizeof(ts_config));
+
+	err = mlx4_en_try_alloc_resources(priv, tmp, &new_prof, true);
+	if (err)
+		goto out;
+
+	if (priv->port_up) {
+		port_up = 1;
+		mlx4_en_stop_port(dev, 1);
+	}
+
+	mlx4_en_safe_replace_resources(priv, tmp);
+
+	if (DEV_FEATURE_CHANGED(dev, features, NETIF_F_HW_VLAN_CTAG_RX)) {
+		if (features & NETIF_F_HW_VLAN_CTAG_RX)
+			dev->features |= NETIF_F_HW_VLAN_CTAG_RX;
+		else
+			dev->features &= ~NETIF_F_HW_VLAN_CTAG_RX;
+	} else if (ts_config.rx_filter == HWTSTAMP_FILTER_NONE) {
+		/* RX time-stamping is OFF, update the RX vlan offload
+		 * to the latest wanted state
+		 */
+		if (dev->wanted_features & NETIF_F_HW_VLAN_CTAG_RX)
+			dev->features |= NETIF_F_HW_VLAN_CTAG_RX;
+		else
+			dev->features &= ~NETIF_F_HW_VLAN_CTAG_RX;
+	}
+
+	if (DEV_FEATURE_CHANGED(dev, features, NETIF_F_RXFCS)) {
+		if (features & NETIF_F_RXFCS)
+			dev->features |= NETIF_F_RXFCS;
+		else
+			dev->features &= ~NETIF_F_RXFCS;
+	}
+
+	/* RX vlan offload and RX time-stamping can't co-exist !
+	 * Regardless of the caller's choice,
+	 * Turn Off RX vlan offload in case of time-stamping is ON
+	 */
+	if (ts_config.rx_filter != HWTSTAMP_FILTER_NONE) {
+		if (dev->features & NETIF_F_HW_VLAN_CTAG_RX)
+			en_warn(priv, "Turning off RX vlan offload since RX time-stamping is ON\n");
+		dev->features &= ~NETIF_F_HW_VLAN_CTAG_RX;
+	}
+
+	if (port_up) {
+		err = mlx4_en_start_port(dev);
+		if (err)
+			en_err(priv, "Failed starting port\n");
+	}
+
+	if (!err)
+		err = mlx4_en_moderation_update(priv);
+out:
+	mutex_unlock(&mdev->state_lock);
+	kfree(tmp);
+	if (!err)
+		netdev_features_change(dev);
+	return err;
+}
diff --git a/drivers/net/ethernet/mellanox/mlx4/en_port.c b/drivers/net/ethernet/mellanox/mlx4/en_port.c
new file mode 100644
index 000000000..0158b88be
--- /dev/null
+++ b/drivers/net/ethernet/mellanox/mlx4/en_port.c
@@ -0,0 +1,432 @@
+/*
+ * Copyright (c) 2007 Mellanox Technologies. All rights reserved.
+ *
+ * This software is available to you under a choice of one of two
+ * licenses.  You may choose to be licensed under the terms of the GNU
+ * General Public License (GPL) Version 2, available from the file
+ * COPYING in the main directory of this source tree, or the
+ * OpenIB.org BSD license below:
+ *
+ *     Redistribution and use in source and binary forms, with or
+ *     without modification, are permitted provided that the following
+ *     conditions are met:
+ *
+ *      - Redistributions of source code must retain the above
+ *        copyright notice, this list of conditions and the following
+ *        disclaimer.
+ *
+ *      - Redistributions in binary form must reproduce the above
+ *        copyright notice, this list of conditions and the following
+ *        disclaimer in the documentation and/or other materials
+ *        provided with the distribution.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
+ * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
+ * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS
+ * BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN
+ * ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
+ * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+ * SOFTWARE.
+ *
+ */
+
+
+#include <linux/if_vlan.h>
+
+#include <linux/mlx4/device.h>
+#include <linux/mlx4/cmd.h>
+
+#include "en_port.h"
+#include "mlx4_en.h"
+
+
+int mlx4_SET_VLAN_FLTR(struct mlx4_dev *dev, struct mlx4_en_priv *priv)
+{
+	struct mlx4_cmd_mailbox *mailbox;
+	struct mlx4_set_vlan_fltr_mbox *filter;
+	int i;
+	int j;
+	int index = 0;
+	u32 entry;
+	int err = 0;
+
+	mailbox = mlx4_alloc_cmd_mailbox(dev);
+	if (IS_ERR(mailbox))
+		return PTR_ERR(mailbox);
+
+	filter = mailbox->buf;
+	for (i = VLAN_FLTR_SIZE - 1; i >= 0; i--) {
+		entry = 0;
+		for (j = 0; j < 32; j++)
+			if (test_bit(index++, priv->active_vlans))
+				entry |= 1 << j;
+		filter->entry[i] = cpu_to_be32(entry);
+	}
+	err = mlx4_cmd(dev, mailbox->dma, priv->port, 0, MLX4_CMD_SET_VLAN_FLTR,
+		       MLX4_CMD_TIME_CLASS_B, MLX4_CMD_WRAPPED);
+	mlx4_free_cmd_mailbox(dev, mailbox);
+	return err;
+}
+
+int mlx4_en_QUERY_PORT(struct mlx4_en_dev *mdev, u8 port)
+{
+	struct mlx4_en_query_port_context *qport_context;
+	struct mlx4_en_priv *priv = netdev_priv(mdev->pndev[port]);
+	struct mlx4_en_port_state *state = &priv->port_state;
+	struct mlx4_cmd_mailbox *mailbox;
+	int err;
+
+	mailbox = mlx4_alloc_cmd_mailbox(mdev->dev);
+	if (IS_ERR(mailbox))
+		return PTR_ERR(mailbox);
+	err = mlx4_cmd_box(mdev->dev, 0, mailbox->dma, port, 0,
+			   MLX4_CMD_QUERY_PORT, MLX4_CMD_TIME_CLASS_B,
+			   MLX4_CMD_WRAPPED);
+	if (err)
+		goto out;
+	qport_context = mailbox->buf;
+
+	/* This command is always accessed from Ethtool context
+	 * already synchronized, no need in locking */
+	state->link_state = !!(qport_context->link_up & MLX4_EN_LINK_UP_MASK);
+	switch (qport_context->link_speed & MLX4_EN_SPEED_MASK) {
+	case MLX4_EN_100M_SPEED:
+		state->link_speed = SPEED_100;
+		break;
+	case MLX4_EN_1G_SPEED:
+		state->link_speed = SPEED_1000;
+		break;
+	case MLX4_EN_10G_SPEED_XAUI:
+	case MLX4_EN_10G_SPEED_XFI:
+		state->link_speed = SPEED_10000;
+		break;
+	case MLX4_EN_20G_SPEED:
+		state->link_speed = SPEED_20000;
+		break;
+	case MLX4_EN_40G_SPEED:
+		state->link_speed = SPEED_40000;
+		break;
+	case MLX4_EN_56G_SPEED:
+		state->link_speed = SPEED_56000;
+		break;
+	default:
+		state->link_speed = -1;
+		break;
+	}
+
+	state->transceiver = qport_context->transceiver;
+
+	state->flags = 0; /* Reset and recalculate the port flags */
+	state->flags |= (qport_context->link_up & MLX4_EN_ANC_MASK) ?
+		MLX4_EN_PORT_ANC : 0;
+	state->flags |= (qport_context->autoneg & MLX4_EN_AUTONEG_MASK) ?
+		MLX4_EN_PORT_ANE : 0;
+
+out:
+	mlx4_free_cmd_mailbox(mdev->dev, mailbox);
+	return err;
+}
+
+/* Each counter set is located in struct mlx4_en_stat_out_mbox
+ * with a const offset between its prio components.
+ * This function runs over a counter set and sum all of it's prio components.
+ */
+static unsigned long en_stats_adder(__be64 *start, __be64 *next, int num)
+{
+	__be64 *curr = start;
+	unsigned long ret = 0;
+	int i;
+	int offset = next - start;
+
+	for (i = 0; i < num; i++) {
+		ret += be64_to_cpu(*curr);
+		curr += offset;
+	}
+
+	return ret;
+}
+
+void mlx4_en_fold_software_stats(struct net_device *dev)
+{
+	struct mlx4_en_priv *priv = netdev_priv(dev);
+	struct mlx4_en_dev *mdev = priv->mdev;
+	unsigned long packets, bytes;
+	int i;
+
+	if (!priv->port_up || mlx4_is_master(mdev->dev))
+		return;
+
+	packets = 0;
+	bytes = 0;
+	for (i = 0; i < priv->rx_ring_num; i++) {
+		const struct mlx4_en_rx_ring *ring = priv->rx_ring[i];
+
+		packets += READ_ONCE(ring->packets);
+		bytes   += READ_ONCE(ring->bytes);
+	}
+	dev->stats.rx_packets = packets;
+	dev->stats.rx_bytes = bytes;
+
+	packets = 0;
+	bytes = 0;
+	for (i = 0; i < priv->tx_ring_num[TX]; i++) {
+		const struct mlx4_en_tx_ring *ring = priv->tx_ring[TX][i];
+
+		packets += READ_ONCE(ring->packets);
+		bytes   += READ_ONCE(ring->bytes);
+	}
+	dev->stats.tx_packets = packets;
+	dev->stats.tx_bytes = bytes;
+}
+
+int mlx4_en_DUMP_ETH_STATS(struct mlx4_en_dev *mdev, u8 port, u8 reset)
+{
+	struct mlx4_counter tmp_counter_stats;
+	struct mlx4_en_stat_out_mbox *mlx4_en_stats;
+	struct mlx4_en_stat_out_flow_control_mbox *flowstats;
+	struct net_device *dev = mdev->pndev[port];
+	struct mlx4_en_priv *priv = netdev_priv(dev);
+	struct net_device_stats *stats = &dev->stats;
+	struct mlx4_cmd_mailbox *mailbox, *mailbox_priority;
+	u64 in_mod = reset << 8 | port;
+	int err;
+	int i, counter_index;
+	unsigned long sw_tx_dropped = 0;
+	unsigned long sw_rx_dropped = 0;
+
+	mailbox = mlx4_alloc_cmd_mailbox(mdev->dev);
+	if (IS_ERR(mailbox))
+		return PTR_ERR(mailbox);
+
+	mailbox_priority = mlx4_alloc_cmd_mailbox(mdev->dev);
+	if (IS_ERR(mailbox_priority)) {
+		mlx4_free_cmd_mailbox(mdev->dev, mailbox);
+		return PTR_ERR(mailbox_priority);
+	}
+
+	err = mlx4_cmd_box(mdev->dev, 0, mailbox->dma, in_mod, 0,
+			   MLX4_CMD_DUMP_ETH_STATS, MLX4_CMD_TIME_CLASS_B,
+			   MLX4_CMD_NATIVE);
+	if (err)
+		goto out;
+
+	mlx4_en_stats = mailbox->buf;
+
+	memset(&tmp_counter_stats, 0, sizeof(tmp_counter_stats));
+	counter_index = mlx4_get_default_counter_index(mdev->dev, port);
+	err = mlx4_get_counter_stats(mdev->dev, counter_index,
+				     &tmp_counter_stats, reset);
+
+	/* 0xffs indicates invalid value */
+	memset(mailbox_priority->buf, 0xff,
+	       sizeof(*flowstats) * MLX4_NUM_PRIORITIES);
+
+	if (mdev->dev->caps.flags2 & MLX4_DEV_CAP_FLAG2_FLOWSTATS_EN) {
+		memset(mailbox_priority->buf, 0,
+		       sizeof(*flowstats) * MLX4_NUM_PRIORITIES);
+		err = mlx4_cmd_box(mdev->dev, 0, mailbox_priority->dma,
+				   in_mod | MLX4_DUMP_ETH_STATS_FLOW_CONTROL,
+				   0, MLX4_CMD_DUMP_ETH_STATS,
+				   MLX4_CMD_TIME_CLASS_B, MLX4_CMD_NATIVE);
+		if (err)
+			goto out;
+	}
+
+	flowstats = mailbox_priority->buf;
+
+	spin_lock_bh(&priv->stats_lock);
+
+	mlx4_en_fold_software_stats(dev);
+
+	priv->port_stats.rx_chksum_good = 0;
+	priv->port_stats.rx_chksum_none = 0;
+	priv->port_stats.rx_chksum_complete = 0;
+	priv->port_stats.rx_alloc_pages = 0;
+	priv->xdp_stats.rx_xdp_drop    = 0;
+	priv->xdp_stats.rx_xdp_tx      = 0;
+	priv->xdp_stats.rx_xdp_tx_full = 0;
+	for (i = 0; i < priv->rx_ring_num; i++) {
+		const struct mlx4_en_rx_ring *ring = priv->rx_ring[i];
+
+		sw_rx_dropped			+= READ_ONCE(ring->dropped);
+		priv->port_stats.rx_chksum_good += READ_ONCE(ring->csum_ok);
+		priv->port_stats.rx_chksum_none += READ_ONCE(ring->csum_none);
+		priv->port_stats.rx_chksum_complete += READ_ONCE(ring->csum_complete);
+		priv->port_stats.rx_alloc_pages += READ_ONCE(ring->rx_alloc_pages);
+		priv->xdp_stats.rx_xdp_drop	+= READ_ONCE(ring->xdp_drop);
+		priv->xdp_stats.rx_xdp_tx	+= READ_ONCE(ring->xdp_tx);
+		priv->xdp_stats.rx_xdp_tx_full	+= READ_ONCE(ring->xdp_tx_full);
+	}
+	priv->port_stats.tx_chksum_offload = 0;
+	priv->port_stats.queue_stopped = 0;
+	priv->port_stats.wake_queue = 0;
+	priv->port_stats.tso_packets = 0;
+	priv->port_stats.xmit_more = 0;
+
+	for (i = 0; i < priv->tx_ring_num[TX]; i++) {
+		const struct mlx4_en_tx_ring *ring = priv->tx_ring[TX][i];
+
+		sw_tx_dropped			   += READ_ONCE(ring->tx_dropped);
+		priv->port_stats.tx_chksum_offload += READ_ONCE(ring->tx_csum);
+		priv->port_stats.queue_stopped     += READ_ONCE(ring->queue_stopped);
+		priv->port_stats.wake_queue        += READ_ONCE(ring->wake_queue);
+		priv->port_stats.tso_packets       += READ_ONCE(ring->tso_packets);
+		priv->port_stats.xmit_more         += READ_ONCE(ring->xmit_more);
+	}
+
+	if (!mlx4_is_slave(mdev->dev)) {
+		struct mlx4_en_phy_stats *p_stats = &priv->phy_stats;
+
+		p_stats->rx_packets_phy =
+			en_stats_adder(&mlx4_en_stats->RTOT_prio_0,
+				       &mlx4_en_stats->RTOT_prio_1,
+				       NUM_PRIORITIES);
+		p_stats->tx_packets_phy =
+			en_stats_adder(&mlx4_en_stats->TTOT_prio_0,
+				       &mlx4_en_stats->TTOT_prio_1,
+				       NUM_PRIORITIES);
+		p_stats->rx_bytes_phy =
+			en_stats_adder(&mlx4_en_stats->ROCT_prio_0,
+				       &mlx4_en_stats->ROCT_prio_1,
+				       NUM_PRIORITIES);
+		p_stats->tx_bytes_phy =
+			en_stats_adder(&mlx4_en_stats->TOCT_prio_0,
+				       &mlx4_en_stats->TOCT_prio_1,
+				       NUM_PRIORITIES);
+		if (mlx4_is_master(mdev->dev)) {
+			stats->rx_packets = p_stats->rx_packets_phy;
+			stats->tx_packets = p_stats->tx_packets_phy;
+			stats->rx_bytes = p_stats->rx_bytes_phy;
+			stats->tx_bytes = p_stats->tx_bytes_phy;
+		}
+	}
+
+	/* net device stats */
+	stats->rx_errors = be64_to_cpu(mlx4_en_stats->PCS) +
+			   be32_to_cpu(mlx4_en_stats->RJBBR) +
+			   be32_to_cpu(mlx4_en_stats->RCRC) +
+			   be32_to_cpu(mlx4_en_stats->RRUNT) +
+			   be64_to_cpu(mlx4_en_stats->RInRangeLengthErr) +
+			   be64_to_cpu(mlx4_en_stats->ROutRangeLengthErr) +
+			   be32_to_cpu(mlx4_en_stats->RSHORT) +
+			   en_stats_adder(&mlx4_en_stats->RGIANT_prio_0,
+					  &mlx4_en_stats->RGIANT_prio_1,
+					  NUM_PRIORITIES);
+	stats->tx_errors = en_stats_adder(&mlx4_en_stats->TGIANT_prio_0,
+					  &mlx4_en_stats->TGIANT_prio_1,
+					  NUM_PRIORITIES);
+	stats->multicast = en_stats_adder(&mlx4_en_stats->MCAST_prio_0,
+					  &mlx4_en_stats->MCAST_prio_1,
+					  NUM_PRIORITIES);
+	stats->rx_dropped = be32_to_cpu(mlx4_en_stats->RDROP) +
+			    sw_rx_dropped;
+	stats->rx_length_errors = be32_to_cpu(mlx4_en_stats->RdropLength);
+	stats->rx_crc_errors = be32_to_cpu(mlx4_en_stats->RCRC);
+	stats->rx_fifo_errors = be32_to_cpu(mlx4_en_stats->RdropOvflw);
+	stats->tx_dropped = be32_to_cpu(mlx4_en_stats->TDROP) +
+			    sw_tx_dropped;
+
+	/* RX stats */
+	priv->pkstats.rx_multicast_packets = stats->multicast;
+	priv->pkstats.rx_broadcast_packets =
+			en_stats_adder(&mlx4_en_stats->RBCAST_prio_0,
+				       &mlx4_en_stats->RBCAST_prio_1,
+				       NUM_PRIORITIES);
+	priv->pkstats.rx_jabbers = be32_to_cpu(mlx4_en_stats->RJBBR);
+	priv->pkstats.rx_in_range_length_error =
+		be64_to_cpu(mlx4_en_stats->RInRangeLengthErr);
+	priv->pkstats.rx_out_range_length_error =
+		be64_to_cpu(mlx4_en_stats->ROutRangeLengthErr);
+
+	/* Tx stats */
+	priv->pkstats.tx_multicast_packets =
+		en_stats_adder(&mlx4_en_stats->TMCAST_prio_0,
+			       &mlx4_en_stats->TMCAST_prio_1,
+			       NUM_PRIORITIES);
+	priv->pkstats.tx_broadcast_packets =
+		en_stats_adder(&mlx4_en_stats->TBCAST_prio_0,
+			       &mlx4_en_stats->TBCAST_prio_1,
+			       NUM_PRIORITIES);
+
+	priv->pkstats.rx_prio[0][0] = be64_to_cpu(mlx4_en_stats->RTOT_prio_0);
+	priv->pkstats.rx_prio[0][1] = be64_to_cpu(mlx4_en_stats->ROCT_prio_0);
+	priv->pkstats.rx_prio[1][0] = be64_to_cpu(mlx4_en_stats->RTOT_prio_1);
+	priv->pkstats.rx_prio[1][1] = be64_to_cpu(mlx4_en_stats->ROCT_prio_1);
+	priv->pkstats.rx_prio[2][0] = be64_to_cpu(mlx4_en_stats->RTOT_prio_2);
+	priv->pkstats.rx_prio[2][1] = be64_to_cpu(mlx4_en_stats->ROCT_prio_2);
+	priv->pkstats.rx_prio[3][0] = be64_to_cpu(mlx4_en_stats->RTOT_prio_3);
+	priv->pkstats.rx_prio[3][1] = be64_to_cpu(mlx4_en_stats->ROCT_prio_3);
+	priv->pkstats.rx_prio[4][0] = be64_to_cpu(mlx4_en_stats->RTOT_prio_4);
+	priv->pkstats.rx_prio[4][1] = be64_to_cpu(mlx4_en_stats->ROCT_prio_4);
+	priv->pkstats.rx_prio[5][0] = be64_to_cpu(mlx4_en_stats->RTOT_prio_5);
+	priv->pkstats.rx_prio[5][1] = be64_to_cpu(mlx4_en_stats->ROCT_prio_5);
+	priv->pkstats.rx_prio[6][0] = be64_to_cpu(mlx4_en_stats->RTOT_prio_6);
+	priv->pkstats.rx_prio[6][1] = be64_to_cpu(mlx4_en_stats->ROCT_prio_6);
+	priv->pkstats.rx_prio[7][0] = be64_to_cpu(mlx4_en_stats->RTOT_prio_7);
+	priv->pkstats.rx_prio[7][1] = be64_to_cpu(mlx4_en_stats->ROCT_prio_7);
+	priv->pkstats.rx_prio[8][0] = be64_to_cpu(mlx4_en_stats->RTOT_novlan);
+	priv->pkstats.rx_prio[8][1] = be64_to_cpu(mlx4_en_stats->ROCT_novlan);
+	priv->pkstats.tx_prio[0][0] = be64_to_cpu(mlx4_en_stats->TTOT_prio_0);
+	priv->pkstats.tx_prio[0][1] = be64_to_cpu(mlx4_en_stats->TOCT_prio_0);
+	priv->pkstats.tx_prio[1][0] = be64_to_cpu(mlx4_en_stats->TTOT_prio_1);
+	priv->pkstats.tx_prio[1][1] = be64_to_cpu(mlx4_en_stats->TOCT_prio_1);
+	priv->pkstats.tx_prio[2][0] = be64_to_cpu(mlx4_en_stats->TTOT_prio_2);
+	priv->pkstats.tx_prio[2][1] = be64_to_cpu(mlx4_en_stats->TOCT_prio_2);
+	priv->pkstats.tx_prio[3][0] = be64_to_cpu(mlx4_en_stats->TTOT_prio_3);
+	priv->pkstats.tx_prio[3][1] = be64_to_cpu(mlx4_en_stats->TOCT_prio_3);
+	priv->pkstats.tx_prio[4][0] = be64_to_cpu(mlx4_en_stats->TTOT_prio_4);
+	priv->pkstats.tx_prio[4][1] = be64_to_cpu(mlx4_en_stats->TOCT_prio_4);
+	priv->pkstats.tx_prio[5][0] = be64_to_cpu(mlx4_en_stats->TTOT_prio_5);
+	priv->pkstats.tx_prio[5][1] = be64_to_cpu(mlx4_en_stats->TOCT_prio_5);
+	priv->pkstats.tx_prio[6][0] = be64_to_cpu(mlx4_en_stats->TTOT_prio_6);
+	priv->pkstats.tx_prio[6][1] = be64_to_cpu(mlx4_en_stats->TOCT_prio_6);
+	priv->pkstats.tx_prio[7][0] = be64_to_cpu(mlx4_en_stats->TTOT_prio_7);
+	priv->pkstats.tx_prio[7][1] = be64_to_cpu(mlx4_en_stats->TOCT_prio_7);
+	priv->pkstats.tx_prio[8][0] = be64_to_cpu(mlx4_en_stats->TTOT_novlan);
+	priv->pkstats.tx_prio[8][1] = be64_to_cpu(mlx4_en_stats->TOCT_novlan);
+
+	if (tmp_counter_stats.counter_mode == 0) {
+		priv->pf_stats.rx_bytes   = be64_to_cpu(tmp_counter_stats.rx_bytes);
+		priv->pf_stats.tx_bytes   = be64_to_cpu(tmp_counter_stats.tx_bytes);
+		priv->pf_stats.rx_packets = be64_to_cpu(tmp_counter_stats.rx_frames);
+		priv->pf_stats.tx_packets = be64_to_cpu(tmp_counter_stats.tx_frames);
+	}
+
+	for (i = 0; i < MLX4_NUM_PRIORITIES; i++)	{
+		priv->rx_priority_flowstats[i].rx_pause =
+			be64_to_cpu(flowstats[i].rx_pause);
+		priv->rx_priority_flowstats[i].rx_pause_duration =
+			be64_to_cpu(flowstats[i].rx_pause_duration);
+		priv->rx_priority_flowstats[i].rx_pause_transition =
+			be64_to_cpu(flowstats[i].rx_pause_transition);
+		priv->tx_priority_flowstats[i].tx_pause =
+			be64_to_cpu(flowstats[i].tx_pause);
+		priv->tx_priority_flowstats[i].tx_pause_duration =
+			be64_to_cpu(flowstats[i].tx_pause_duration);
+		priv->tx_priority_flowstats[i].tx_pause_transition =
+			be64_to_cpu(flowstats[i].tx_pause_transition);
+	}
+
+	/* if pfc is not in use, all priorities counters have the same value */
+	priv->rx_flowstats.rx_pause =
+		be64_to_cpu(flowstats[0].rx_pause);
+	priv->rx_flowstats.rx_pause_duration =
+		be64_to_cpu(flowstats[0].rx_pause_duration);
+	priv->rx_flowstats.rx_pause_transition =
+		be64_to_cpu(flowstats[0].rx_pause_transition);
+	priv->tx_flowstats.tx_pause =
+		be64_to_cpu(flowstats[0].tx_pause);
+	priv->tx_flowstats.tx_pause_duration =
+		be64_to_cpu(flowstats[0].tx_pause_duration);
+	priv->tx_flowstats.tx_pause_transition =
+		be64_to_cpu(flowstats[0].tx_pause_transition);
+
+	spin_unlock_bh(&priv->stats_lock);
+
+out:
+	mlx4_free_cmd_mailbox(mdev->dev, mailbox);
+	mlx4_free_cmd_mailbox(mdev->dev, mailbox_priority);
+	return err;
+}
+
diff --git a/drivers/net/ethernet/mellanox/mlx4/en_port.h b/drivers/net/ethernet/mellanox/mlx4/en_port.h
new file mode 100644
index 000000000..930f961fe
--- /dev/null
+++ b/drivers/net/ethernet/mellanox/mlx4/en_port.h
@@ -0,0 +1,586 @@
+/*
+ * Copyright (c) 2007 Mellanox Technologies. All rights reserved.
+ *
+ * This software is available to you under a choice of one of two
+ * licenses.  You may choose to be licensed under the terms of the GNU
+ * General Public License (GPL) Version 2, available from the file
+ * COPYING in the main directory of this source tree, or the
+ * OpenIB.org BSD license below:
+ *
+ *     Redistribution and use in source and binary forms, with or
+ *     without modification, are permitted provided that the following
+ *     conditions are met:
+ *
+ *      - Redistributions of source code must retain the above
+ *        copyright notice, this list of conditions and the following
+ *        disclaimer.
+ *
+ *      - Redistributions in binary form must reproduce the above
+ *        copyright notice, this list of conditions and the following
+ *        disclaimer in the documentation and/or other materials
+ *        provided with the distribution.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
+ * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
+ * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS
+ * BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN
+ * ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
+ * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+ * SOFTWARE.
+ *
+ */
+
+#ifndef _MLX4_EN_PORT_H_
+#define _MLX4_EN_PORT_H_
+
+
+#define SET_PORT_PROMISC_SHIFT	31
+#define SET_PORT_MC_PROMISC_SHIFT	30
+
+#define MLX4_EN_NUM_TC		8
+
+#define VLAN_FLTR_SIZE	128
+struct mlx4_set_vlan_fltr_mbox {
+	__be32 entry[VLAN_FLTR_SIZE];
+};
+
+
+enum {
+	MLX4_MCAST_CONFIG       = 0,
+	MLX4_MCAST_DISABLE      = 1,
+	MLX4_MCAST_ENABLE       = 2,
+};
+
+enum mlx4_link_mode {
+	MLX4_1000BASE_CX_SGMII	 = 0,
+	MLX4_1000BASE_KX	 = 1,
+	MLX4_10GBASE_CX4	 = 2,
+	MLX4_10GBASE_KX4	 = 3,
+	MLX4_10GBASE_KR		 = 4,
+	MLX4_20GBASE_KR2	 = 5,
+	MLX4_40GBASE_CR4	 = 6,
+	MLX4_40GBASE_KR4	 = 7,
+	MLX4_56GBASE_KR4	 = 8,
+	MLX4_10GBASE_CR		 = 12,
+	MLX4_10GBASE_SR		 = 13,
+	MLX4_40GBASE_SR4	 = 15,
+	MLX4_56GBASE_CR4	 = 17,
+	MLX4_56GBASE_SR4	 = 18,
+	MLX4_100BASE_TX		 = 24,
+	MLX4_1000BASE_T		 = 25,
+	MLX4_10GBASE_T		 = 26,
+};
+
+#define MLX4_PROT_MASK(link_mode) (1<<link_mode)
+
+enum {
+	MLX4_EN_100M_SPEED	= 0x04,
+	MLX4_EN_10G_SPEED_XAUI	= 0x00,
+	MLX4_EN_10G_SPEED_XFI	= 0x01,
+	MLX4_EN_1G_SPEED	= 0x02,
+	MLX4_EN_20G_SPEED	= 0x08,
+	MLX4_EN_40G_SPEED	= 0x40,
+	MLX4_EN_56G_SPEED	= 0x20,
+	MLX4_EN_OTHER_SPEED	= 0x0f,
+};
+
+struct mlx4_en_query_port_context {
+	u8 link_up;
+#define MLX4_EN_LINK_UP_MASK	0x80
+#define MLX4_EN_ANC_MASK	0x40
+	u8 autoneg;
+#define MLX4_EN_AUTONEG_MASK	0x80
+	__be16 mtu;
+	u8 reserved2;
+	u8 link_speed;
+#define MLX4_EN_SPEED_MASK	0x6f
+	u16 reserved3[5];
+	__be64 mac;
+	u8 transceiver;
+};
+
+
+struct mlx4_en_stat_out_mbox {
+	/* Received frames with a length of 64 octets */
+	__be64 R64_prio_0;
+	__be64 R64_prio_1;
+	__be64 R64_prio_2;
+	__be64 R64_prio_3;
+	__be64 R64_prio_4;
+	__be64 R64_prio_5;
+	__be64 R64_prio_6;
+	__be64 R64_prio_7;
+	__be64 R64_novlan;
+	/* Received frames with a length of 127 octets */
+	__be64 R127_prio_0;
+	__be64 R127_prio_1;
+	__be64 R127_prio_2;
+	__be64 R127_prio_3;
+	__be64 R127_prio_4;
+	__be64 R127_prio_5;
+	__be64 R127_prio_6;
+	__be64 R127_prio_7;
+	__be64 R127_novlan;
+	/* Received frames with a length of 255 octets */
+	__be64 R255_prio_0;
+	__be64 R255_prio_1;
+	__be64 R255_prio_2;
+	__be64 R255_prio_3;
+	__be64 R255_prio_4;
+	__be64 R255_prio_5;
+	__be64 R255_prio_6;
+	__be64 R255_prio_7;
+	__be64 R255_novlan;
+	/* Received frames with a length of 511 octets */
+	__be64 R511_prio_0;
+	__be64 R511_prio_1;
+	__be64 R511_prio_2;
+	__be64 R511_prio_3;
+	__be64 R511_prio_4;
+	__be64 R511_prio_5;
+	__be64 R511_prio_6;
+	__be64 R511_prio_7;
+	__be64 R511_novlan;
+	/* Received frames with a length of 1023 octets */
+	__be64 R1023_prio_0;
+	__be64 R1023_prio_1;
+	__be64 R1023_prio_2;
+	__be64 R1023_prio_3;
+	__be64 R1023_prio_4;
+	__be64 R1023_prio_5;
+	__be64 R1023_prio_6;
+	__be64 R1023_prio_7;
+	__be64 R1023_novlan;
+	/* Received frames with a length of 1518 octets */
+	__be64 R1518_prio_0;
+	__be64 R1518_prio_1;
+	__be64 R1518_prio_2;
+	__be64 R1518_prio_3;
+	__be64 R1518_prio_4;
+	__be64 R1518_prio_5;
+	__be64 R1518_prio_6;
+	__be64 R1518_prio_7;
+	__be64 R1518_novlan;
+	/* Received frames with a length of 1522 octets */
+	__be64 R1522_prio_0;
+	__be64 R1522_prio_1;
+	__be64 R1522_prio_2;
+	__be64 R1522_prio_3;
+	__be64 R1522_prio_4;
+	__be64 R1522_prio_5;
+	__be64 R1522_prio_6;
+	__be64 R1522_prio_7;
+	__be64 R1522_novlan;
+	/* Received frames with a length of 1548 octets */
+	__be64 R1548_prio_0;
+	__be64 R1548_prio_1;
+	__be64 R1548_prio_2;
+	__be64 R1548_prio_3;
+	__be64 R1548_prio_4;
+	__be64 R1548_prio_5;
+	__be64 R1548_prio_6;
+	__be64 R1548_prio_7;
+	__be64 R1548_novlan;
+	/* Received frames with a length of 1548 < octets < MTU */
+	__be64 R2MTU_prio_0;
+	__be64 R2MTU_prio_1;
+	__be64 R2MTU_prio_2;
+	__be64 R2MTU_prio_3;
+	__be64 R2MTU_prio_4;
+	__be64 R2MTU_prio_5;
+	__be64 R2MTU_prio_6;
+	__be64 R2MTU_prio_7;
+	__be64 R2MTU_novlan;
+	/* Received frames with a length of MTU< octets and good CRC */
+	__be64 RGIANT_prio_0;
+	__be64 RGIANT_prio_1;
+	__be64 RGIANT_prio_2;
+	__be64 RGIANT_prio_3;
+	__be64 RGIANT_prio_4;
+	__be64 RGIANT_prio_5;
+	__be64 RGIANT_prio_6;
+	__be64 RGIANT_prio_7;
+	__be64 RGIANT_novlan;
+	/* Received broadcast frames with good CRC */
+	__be64 RBCAST_prio_0;
+	__be64 RBCAST_prio_1;
+	__be64 RBCAST_prio_2;
+	__be64 RBCAST_prio_3;
+	__be64 RBCAST_prio_4;
+	__be64 RBCAST_prio_5;
+	__be64 RBCAST_prio_6;
+	__be64 RBCAST_prio_7;
+	__be64 RBCAST_novlan;
+	/* Received multicast frames with good CRC */
+	__be64 MCAST_prio_0;
+	__be64 MCAST_prio_1;
+	__be64 MCAST_prio_2;
+	__be64 MCAST_prio_3;
+	__be64 MCAST_prio_4;
+	__be64 MCAST_prio_5;
+	__be64 MCAST_prio_6;
+	__be64 MCAST_prio_7;
+	__be64 MCAST_novlan;
+	/* Received unicast not short or GIANT frames with good CRC */
+	__be64 RTOTG_prio_0;
+	__be64 RTOTG_prio_1;
+	__be64 RTOTG_prio_2;
+	__be64 RTOTG_prio_3;
+	__be64 RTOTG_prio_4;
+	__be64 RTOTG_prio_5;
+	__be64 RTOTG_prio_6;
+	__be64 RTOTG_prio_7;
+	__be64 RTOTG_novlan;
+
+	/* Count of total octets of received frames, includes framing characters */
+	__be64 RTTLOCT_prio_0;
+	/* Count of total octets of received frames, not including framing
+	   characters */
+	__be64 RTTLOCT_NOFRM_prio_0;
+	/* Count of Total number of octets received
+	   (only for frames without errors) */
+	__be64 ROCT_prio_0;
+
+	__be64 RTTLOCT_prio_1;
+	__be64 RTTLOCT_NOFRM_prio_1;
+	__be64 ROCT_prio_1;
+
+	__be64 RTTLOCT_prio_2;
+	__be64 RTTLOCT_NOFRM_prio_2;
+	__be64 ROCT_prio_2;
+
+	__be64 RTTLOCT_prio_3;
+	__be64 RTTLOCT_NOFRM_prio_3;
+	__be64 ROCT_prio_3;
+
+	__be64 RTTLOCT_prio_4;
+	__be64 RTTLOCT_NOFRM_prio_4;
+	__be64 ROCT_prio_4;
+
+	__be64 RTTLOCT_prio_5;
+	__be64 RTTLOCT_NOFRM_prio_5;
+	__be64 ROCT_prio_5;
+
+	__be64 RTTLOCT_prio_6;
+	__be64 RTTLOCT_NOFRM_prio_6;
+	__be64 ROCT_prio_6;
+
+	__be64 RTTLOCT_prio_7;
+	__be64 RTTLOCT_NOFRM_prio_7;
+	__be64 ROCT_prio_7;
+
+	__be64 RTTLOCT_novlan;
+	__be64 RTTLOCT_NOFRM_novlan;
+	__be64 ROCT_novlan;
+
+	/* Count of Total received frames including bad frames */
+	__be64 RTOT_prio_0;
+	/* Count of  Total number of received frames with 802.1Q encapsulation */
+	__be64 R1Q_prio_0;
+	__be64 reserved1;
+
+	__be64 RTOT_prio_1;
+	__be64 R1Q_prio_1;
+	__be64 reserved2;
+
+	__be64 RTOT_prio_2;
+	__be64 R1Q_prio_2;
+	__be64 reserved3;
+
+	__be64 RTOT_prio_3;
+	__be64 R1Q_prio_3;
+	__be64 reserved4;
+
+	__be64 RTOT_prio_4;
+	__be64 R1Q_prio_4;
+	__be64 reserved5;
+
+	__be64 RTOT_prio_5;
+	__be64 R1Q_prio_5;
+	__be64 reserved6;
+
+	__be64 RTOT_prio_6;
+	__be64 R1Q_prio_6;
+	__be64 reserved7;
+
+	__be64 RTOT_prio_7;
+	__be64 R1Q_prio_7;
+	__be64 reserved8;
+
+	__be64 RTOT_novlan;
+	__be64 R1Q_novlan;
+	__be64 reserved9;
+
+	/* Total number of Successfully Received Control Frames */
+	__be64 RCNTL;
+	__be64 reserved10;
+	__be64 reserved11;
+	__be64 reserved12;
+	/* Count of received frames with a length/type field  value between 46
+	   (42 for VLANtagged frames) and 1500 (also 1500 for VLAN-tagged frames),
+	   inclusive */
+	__be64 RInRangeLengthErr;
+	/* Count of received frames with length/type field between 1501 and 1535
+	   decimal, inclusive */
+	__be64 ROutRangeLengthErr;
+	/* Count of received frames that are longer than max allowed size for
+	   802.3 frames (1518/1522) */
+	__be64 RFrmTooLong;
+	/* Count frames received with PCS error */
+	__be64 PCS;
+
+	/* Transmit frames with a length of 64 octets */
+	__be64 T64_prio_0;
+	__be64 T64_prio_1;
+	__be64 T64_prio_2;
+	__be64 T64_prio_3;
+	__be64 T64_prio_4;
+	__be64 T64_prio_5;
+	__be64 T64_prio_6;
+	__be64 T64_prio_7;
+	__be64 T64_novlan;
+	__be64 T64_loopbk;
+	/* Transmit frames with a length of 65 to 127 octets. */
+	__be64 T127_prio_0;
+	__be64 T127_prio_1;
+	__be64 T127_prio_2;
+	__be64 T127_prio_3;
+	__be64 T127_prio_4;
+	__be64 T127_prio_5;
+	__be64 T127_prio_6;
+	__be64 T127_prio_7;
+	__be64 T127_novlan;
+	__be64 T127_loopbk;
+	/* Transmit frames with a length of 128 to 255 octets */
+	__be64 T255_prio_0;
+	__be64 T255_prio_1;
+	__be64 T255_prio_2;
+	__be64 T255_prio_3;
+	__be64 T255_prio_4;
+	__be64 T255_prio_5;
+	__be64 T255_prio_6;
+	__be64 T255_prio_7;
+	__be64 T255_novlan;
+	__be64 T255_loopbk;
+	/* Transmit frames with a length of 256 to 511 octets */
+	__be64 T511_prio_0;
+	__be64 T511_prio_1;
+	__be64 T511_prio_2;
+	__be64 T511_prio_3;
+	__be64 T511_prio_4;
+	__be64 T511_prio_5;
+	__be64 T511_prio_6;
+	__be64 T511_prio_7;
+	__be64 T511_novlan;
+	__be64 T511_loopbk;
+	/* Transmit frames with a length of 512 to 1023 octets */
+	__be64 T1023_prio_0;
+	__be64 T1023_prio_1;
+	__be64 T1023_prio_2;
+	__be64 T1023_prio_3;
+	__be64 T1023_prio_4;
+	__be64 T1023_prio_5;
+	__be64 T1023_prio_6;
+	__be64 T1023_prio_7;
+	__be64 T1023_novlan;
+	__be64 T1023_loopbk;
+	/* Transmit frames with a length of 1024 to 1518 octets */
+	__be64 T1518_prio_0;
+	__be64 T1518_prio_1;
+	__be64 T1518_prio_2;
+	__be64 T1518_prio_3;
+	__be64 T1518_prio_4;
+	__be64 T1518_prio_5;
+	__be64 T1518_prio_6;
+	__be64 T1518_prio_7;
+	__be64 T1518_novlan;
+	__be64 T1518_loopbk;
+	/* Counts transmit frames with a length of 1519 to 1522 bytes */
+	__be64 T1522_prio_0;
+	__be64 T1522_prio_1;
+	__be64 T1522_prio_2;
+	__be64 T1522_prio_3;
+	__be64 T1522_prio_4;
+	__be64 T1522_prio_5;
+	__be64 T1522_prio_6;
+	__be64 T1522_prio_7;
+	__be64 T1522_novlan;
+	__be64 T1522_loopbk;
+	/* Transmit frames with a length of 1523 to 1548 octets */
+	__be64 T1548_prio_0;
+	__be64 T1548_prio_1;
+	__be64 T1548_prio_2;
+	__be64 T1548_prio_3;
+	__be64 T1548_prio_4;
+	__be64 T1548_prio_5;
+	__be64 T1548_prio_6;
+	__be64 T1548_prio_7;
+	__be64 T1548_novlan;
+	__be64 T1548_loopbk;
+	/* Counts transmit frames with a length of 1549 to MTU bytes */
+	__be64 T2MTU_prio_0;
+	__be64 T2MTU_prio_1;
+	__be64 T2MTU_prio_2;
+	__be64 T2MTU_prio_3;
+	__be64 T2MTU_prio_4;
+	__be64 T2MTU_prio_5;
+	__be64 T2MTU_prio_6;
+	__be64 T2MTU_prio_7;
+	__be64 T2MTU_novlan;
+	__be64 T2MTU_loopbk;
+	/* Transmit frames with a length greater than MTU octets and a good CRC. */
+	__be64 TGIANT_prio_0;
+	__be64 TGIANT_prio_1;
+	__be64 TGIANT_prio_2;
+	__be64 TGIANT_prio_3;
+	__be64 TGIANT_prio_4;
+	__be64 TGIANT_prio_5;
+	__be64 TGIANT_prio_6;
+	__be64 TGIANT_prio_7;
+	__be64 TGIANT_novlan;
+	__be64 TGIANT_loopbk;
+	/* Transmit broadcast frames with a good CRC */
+	__be64 TBCAST_prio_0;
+	__be64 TBCAST_prio_1;
+	__be64 TBCAST_prio_2;
+	__be64 TBCAST_prio_3;
+	__be64 TBCAST_prio_4;
+	__be64 TBCAST_prio_5;
+	__be64 TBCAST_prio_6;
+	__be64 TBCAST_prio_7;
+	__be64 TBCAST_novlan;
+	__be64 TBCAST_loopbk;
+	/* Transmit multicast frames with a good CRC */
+	__be64 TMCAST_prio_0;
+	__be64 TMCAST_prio_1;
+	__be64 TMCAST_prio_2;
+	__be64 TMCAST_prio_3;
+	__be64 TMCAST_prio_4;
+	__be64 TMCAST_prio_5;
+	__be64 TMCAST_prio_6;
+	__be64 TMCAST_prio_7;
+	__be64 TMCAST_novlan;
+	__be64 TMCAST_loopbk;
+	/* Transmit good frames that are neither broadcast nor multicast */
+	__be64 TTOTG_prio_0;
+	__be64 TTOTG_prio_1;
+	__be64 TTOTG_prio_2;
+	__be64 TTOTG_prio_3;
+	__be64 TTOTG_prio_4;
+	__be64 TTOTG_prio_5;
+	__be64 TTOTG_prio_6;
+	__be64 TTOTG_prio_7;
+	__be64 TTOTG_novlan;
+	__be64 TTOTG_loopbk;
+
+	/* total octets of transmitted frames, including framing characters */
+	__be64 TTTLOCT_prio_0;
+	/* total octets of transmitted frames, not including framing characters */
+	__be64 TTTLOCT_NOFRM_prio_0;
+	/* ifOutOctets */
+	__be64 TOCT_prio_0;
+
+	__be64 TTTLOCT_prio_1;
+	__be64 TTTLOCT_NOFRM_prio_1;
+	__be64 TOCT_prio_1;
+
+	__be64 TTTLOCT_prio_2;
+	__be64 TTTLOCT_NOFRM_prio_2;
+	__be64 TOCT_prio_2;
+
+	__be64 TTTLOCT_prio_3;
+	__be64 TTTLOCT_NOFRM_prio_3;
+	__be64 TOCT_prio_3;
+
+	__be64 TTTLOCT_prio_4;
+	__be64 TTTLOCT_NOFRM_prio_4;
+	__be64 TOCT_prio_4;
+
+	__be64 TTTLOCT_prio_5;
+	__be64 TTTLOCT_NOFRM_prio_5;
+	__be64 TOCT_prio_5;
+
+	__be64 TTTLOCT_prio_6;
+	__be64 TTTLOCT_NOFRM_prio_6;
+	__be64 TOCT_prio_6;
+
+	__be64 TTTLOCT_prio_7;
+	__be64 TTTLOCT_NOFRM_prio_7;
+	__be64 TOCT_prio_7;
+
+	__be64 TTTLOCT_novlan;
+	__be64 TTTLOCT_NOFRM_novlan;
+	__be64 TOCT_novlan;
+
+	__be64 TTTLOCT_loopbk;
+	__be64 TTTLOCT_NOFRM_loopbk;
+	__be64 TOCT_loopbk;
+
+	/* Total frames transmitted with a good CRC that are not aborted  */
+	__be64 TTOT_prio_0;
+	/* Total number of frames transmitted with 802.1Q encapsulation */
+	__be64 T1Q_prio_0;
+	__be64 reserved13;
+
+	__be64 TTOT_prio_1;
+	__be64 T1Q_prio_1;
+	__be64 reserved14;
+
+	__be64 TTOT_prio_2;
+	__be64 T1Q_prio_2;
+	__be64 reserved15;
+
+	__be64 TTOT_prio_3;
+	__be64 T1Q_prio_3;
+	__be64 reserved16;
+
+	__be64 TTOT_prio_4;
+	__be64 T1Q_prio_4;
+	__be64 reserved17;
+
+	__be64 TTOT_prio_5;
+	__be64 T1Q_prio_5;
+	__be64 reserved18;
+
+	__be64 TTOT_prio_6;
+	__be64 T1Q_prio_6;
+	__be64 reserved19;
+
+	__be64 TTOT_prio_7;
+	__be64 T1Q_prio_7;
+	__be64 reserved20;
+
+	__be64 TTOT_novlan;
+	__be64 T1Q_novlan;
+	__be64 reserved21;
+
+	__be64 TTOT_loopbk;
+	__be64 T1Q_loopbk;
+	__be64 reserved22;
+
+	/* Received frames with a length greater than MTU octets and a bad CRC */
+	__be32 RJBBR;
+	/* Received frames with a bad CRC that are not runts, jabbers,
+	   or alignment errors */
+	__be32 RCRC;
+	/* Received frames with SFD with a length of less than 64 octets and a
+	   bad CRC */
+	__be32 RRUNT;
+	/* Received frames with a length less than 64 octets and a good CRC */
+	__be32 RSHORT;
+	/* Total Number of Received Packets Dropped */
+	__be32 RDROP;
+	/* Drop due to overflow  */
+	__be32 RdropOvflw;
+	/* Drop due to overflow */
+	__be32 RdropLength;
+	/* Total of good frames. Does not include frames received with
+	   frame-too-long, FCS, or length errors */
+	__be32 RTOTFRMS;
+	/* Total dropped Xmited packets */
+	__be32 TDROP;
+};
+
+
+#endif
diff --git a/drivers/net/ethernet/mellanox/mlx4/en_resources.c b/drivers/net/ethernet/mellanox/mlx4/en_resources.c
new file mode 100644
index 000000000..6883ac75d
--- /dev/null
+++ b/drivers/net/ethernet/mellanox/mlx4/en_resources.c
@@ -0,0 +1,115 @@
+/*
+ * Copyright (c) 2007 Mellanox Technologies. All rights reserved.
+ *
+ * This software is available to you under a choice of one of two
+ * licenses.  You may choose to be licensed under the terms of the GNU
+ * General Public License (GPL) Version 2, available from the file
+ * COPYING in the main directory of this source tree, or the
+ * OpenIB.org BSD license below:
+ *
+ *     Redistribution and use in source and binary forms, with or
+ *     without modification, are permitted provided that the following
+ *     conditions are met:
+ *
+ *      - Redistributions of source code must retain the above
+ *        copyright notice, this list of conditions and the following
+ *        disclaimer.
+ *
+ *      - Redistributions in binary form must reproduce the above
+ *        copyright notice, this list of conditions and the following
+ *        disclaimer in the documentation and/or other materials
+ *        provided with the distribution.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
+ * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
+ * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS
+ * BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN
+ * ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
+ * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+ * SOFTWARE.
+ *
+ */
+
+#include <linux/slab.h>
+#include <linux/vmalloc.h>
+#include <linux/mlx4/qp.h>
+
+#include "mlx4_en.h"
+
+void mlx4_en_fill_qp_context(struct mlx4_en_priv *priv, int size, int stride,
+			     int is_tx, int rss, int qpn, int cqn,
+			     int user_prio, struct mlx4_qp_context *context)
+{
+	struct mlx4_en_dev *mdev = priv->mdev;
+	struct net_device *dev = priv->dev;
+
+	memset(context, 0, sizeof(*context));
+	context->flags = cpu_to_be32(7 << 16 | rss << MLX4_RSS_QPC_FLAG_OFFSET);
+	context->pd = cpu_to_be32(mdev->priv_pdn);
+	context->mtu_msgmax = 0xff;
+	if (!is_tx && !rss)
+		context->rq_size_stride = ilog2(size) << 3 | (ilog2(stride) - 4);
+	if (is_tx) {
+		context->sq_size_stride = ilog2(size) << 3 | (ilog2(stride) - 4);
+		if (mdev->dev->caps.flags2 & MLX4_DEV_CAP_FLAG2_PORT_REMAP)
+			context->params2 |= cpu_to_be32(MLX4_QP_BIT_FPP);
+
+	} else {
+		context->sq_size_stride = ilog2(TXBB_SIZE) - 4;
+	}
+	context->usr_page = cpu_to_be32(mlx4_to_hw_uar_index(mdev->dev,
+					mdev->priv_uar.index));
+	context->local_qpn = cpu_to_be32(qpn);
+	context->pri_path.ackto = 1 & 0x07;
+	context->pri_path.sched_queue = 0x83 | (priv->port - 1) << 6;
+	/* force user priority per tx ring */
+	if (user_prio >= 0 && priv->prof->num_up == MLX4_EN_NUM_UP_HIGH) {
+		context->pri_path.sched_queue |= user_prio << 3;
+		context->pri_path.feup = MLX4_FEUP_FORCE_ETH_UP;
+	}
+	context->pri_path.counter_index = priv->counter_index;
+	context->cqn_send = cpu_to_be32(cqn);
+	context->cqn_recv = cpu_to_be32(cqn);
+	if (!rss &&
+	    (mdev->dev->caps.flags2 & MLX4_DEV_CAP_FLAG2_LB_SRC_CHK) &&
+	    context->pri_path.counter_index !=
+			    MLX4_SINK_COUNTER_INDEX(mdev->dev)) {
+		/* disable multicast loopback to qp with same counter */
+		if (!(dev->features & NETIF_F_LOOPBACK))
+			context->pri_path.fl |= MLX4_FL_ETH_SRC_CHECK_MC_LB;
+		context->pri_path.control |= MLX4_CTRL_ETH_SRC_CHECK_IF_COUNTER;
+	}
+	context->db_rec_addr = cpu_to_be64(priv->res.db.dma << 2);
+	if (!(dev->features & NETIF_F_HW_VLAN_CTAG_RX))
+		context->param3 |= cpu_to_be32(1 << 30);
+
+	if (!is_tx && !rss &&
+	    (mdev->dev->caps.tunnel_offload_mode ==  MLX4_TUNNEL_OFFLOAD_MODE_VXLAN)) {
+		en_dbg(HW, priv, "Setting RX qp %x tunnel mode to RX tunneled & non-tunneled\n", qpn);
+		context->srqn = cpu_to_be32(7 << 28); /* this fills bits 30:28 */
+	}
+}
+
+int mlx4_en_change_mcast_lb(struct mlx4_en_priv *priv, struct mlx4_qp *qp,
+			    int loopback)
+{
+	int ret;
+	struct mlx4_update_qp_params qp_params;
+
+	memset(&qp_params, 0, sizeof(qp_params));
+	if (!loopback)
+		qp_params.flags = MLX4_UPDATE_QP_PARAMS_FLAGS_ETH_CHECK_MC_LB;
+
+	ret = mlx4_update_qp(priv->mdev->dev, qp->qpn,
+			     MLX4_UPDATE_QP_ETH_SRC_CHECK_MC_LB,
+			     &qp_params);
+
+	return ret;
+}
+
+void mlx4_en_sqp_event(struct mlx4_qp *qp, enum mlx4_event event)
+{
+    return;
+}
+
diff --git a/drivers/net/ethernet/mellanox/mlx4/en_rx.c b/drivers/net/ethernet/mellanox/mlx4/en_rx.c
new file mode 100644
index 000000000..f509a6ce3
--- /dev/null
+++ b/drivers/net/ethernet/mellanox/mlx4/en_rx.c
@@ -0,0 +1,1287 @@
+/*
+ * Copyright (c) 2007 Mellanox Technologies. All rights reserved.
+ *
+ * This software is available to you under a choice of one of two
+ * licenses.  You may choose to be licensed under the terms of the GNU
+ * General Public License (GPL) Version 2, available from the file
+ * COPYING in the main directory of this source tree, or the
+ * OpenIB.org BSD license below:
+ *
+ *     Redistribution and use in source and binary forms, with or
+ *     without modification, are permitted provided that the following
+ *     conditions are met:
+ *
+ *      - Redistributions of source code must retain the above
+ *        copyright notice, this list of conditions and the following
+ *        disclaimer.
+ *
+ *      - Redistributions in binary form must reproduce the above
+ *        copyright notice, this list of conditions and the following
+ *        disclaimer in the documentation and/or other materials
+ *        provided with the distribution.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
+ * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
+ * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS
+ * BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN
+ * ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
+ * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+ * SOFTWARE.
+ *
+ */
+
+#include <net/busy_poll.h>
+#include <linux/bpf.h>
+#include <linux/bpf_trace.h>
+#include <linux/mlx4/cq.h>
+#include <linux/slab.h>
+#include <linux/mlx4/qp.h>
+#include <linux/skbuff.h>
+#include <linux/rculist.h>
+#include <linux/if_ether.h>
+#include <linux/if_vlan.h>
+#include <linux/vmalloc.h>
+#include <linux/irq.h>
+
+#if IS_ENABLED(CONFIG_IPV6)
+#include <net/ip6_checksum.h>
+#endif
+
+#include "mlx4_en.h"
+
+static int mlx4_alloc_page(struct mlx4_en_priv *priv,
+			   struct mlx4_en_rx_alloc *frag,
+			   gfp_t gfp)
+{
+	struct page *page;
+	dma_addr_t dma;
+
+	page = alloc_page(gfp);
+	if (unlikely(!page))
+		return -ENOMEM;
+	dma = dma_map_page(priv->ddev, page, 0, PAGE_SIZE, priv->dma_dir);
+	if (unlikely(dma_mapping_error(priv->ddev, dma))) {
+		__free_page(page);
+		return -ENOMEM;
+	}
+	frag->page = page;
+	frag->dma = dma;
+	frag->page_offset = priv->rx_headroom;
+	return 0;
+}
+
+static int mlx4_en_alloc_frags(struct mlx4_en_priv *priv,
+			       struct mlx4_en_rx_ring *ring,
+			       struct mlx4_en_rx_desc *rx_desc,
+			       struct mlx4_en_rx_alloc *frags,
+			       gfp_t gfp)
+{
+	int i;
+
+	for (i = 0; i < priv->num_frags; i++, frags++) {
+		if (!frags->page) {
+			if (mlx4_alloc_page(priv, frags, gfp))
+				return -ENOMEM;
+			ring->rx_alloc_pages++;
+		}
+		rx_desc->data[i].addr = cpu_to_be64(frags->dma +
+						    frags->page_offset);
+	}
+	return 0;
+}
+
+static void mlx4_en_free_frag(const struct mlx4_en_priv *priv,
+			      struct mlx4_en_rx_alloc *frag)
+{
+	if (frag->page) {
+		dma_unmap_page(priv->ddev, frag->dma,
+			       PAGE_SIZE, priv->dma_dir);
+		__free_page(frag->page);
+	}
+	/* We need to clear all fields, otherwise a change of priv->log_rx_info
+	 * could lead to see garbage later in frag->page.
+	 */
+	memset(frag, 0, sizeof(*frag));
+}
+
+static void mlx4_en_init_rx_desc(const struct mlx4_en_priv *priv,
+				 struct mlx4_en_rx_ring *ring, int index)
+{
+	struct mlx4_en_rx_desc *rx_desc = ring->buf + ring->stride * index;
+	int possible_frags;
+	int i;
+
+	/* Set size and memtype fields */
+	for (i = 0; i < priv->num_frags; i++) {
+		rx_desc->data[i].byte_count =
+			cpu_to_be32(priv->frag_info[i].frag_size);
+		rx_desc->data[i].lkey = cpu_to_be32(priv->mdev->mr.key);
+	}
+
+	/* If the number of used fragments does not fill up the ring stride,
+	 * remaining (unused) fragments must be padded with null address/size
+	 * and a special memory key */
+	possible_frags = (ring->stride - sizeof(struct mlx4_en_rx_desc)) / DS_SIZE;
+	for (i = priv->num_frags; i < possible_frags; i++) {
+		rx_desc->data[i].byte_count = 0;
+		rx_desc->data[i].lkey = cpu_to_be32(MLX4_EN_MEMTYPE_PAD);
+		rx_desc->data[i].addr = 0;
+	}
+}
+
+static int mlx4_en_prepare_rx_desc(struct mlx4_en_priv *priv,
+				   struct mlx4_en_rx_ring *ring, int index,
+				   gfp_t gfp)
+{
+	struct mlx4_en_rx_desc *rx_desc = ring->buf +
+		(index << ring->log_stride);
+	struct mlx4_en_rx_alloc *frags = ring->rx_info +
+					(index << priv->log_rx_info);
+	if (likely(ring->page_cache.index > 0)) {
+		/* XDP uses a single page per frame */
+		if (!frags->page) {
+			ring->page_cache.index--;
+			frags->page = ring->page_cache.buf[ring->page_cache.index].page;
+			frags->dma  = ring->page_cache.buf[ring->page_cache.index].dma;
+		}
+		frags->page_offset = XDP_PACKET_HEADROOM;
+		rx_desc->data[0].addr = cpu_to_be64(frags->dma +
+						    XDP_PACKET_HEADROOM);
+		return 0;
+	}
+
+	return mlx4_en_alloc_frags(priv, ring, rx_desc, frags, gfp);
+}
+
+static bool mlx4_en_is_ring_empty(const struct mlx4_en_rx_ring *ring)
+{
+	return ring->prod == ring->cons;
+}
+
+static inline void mlx4_en_update_rx_prod_db(struct mlx4_en_rx_ring *ring)
+{
+	*ring->wqres.db.db = cpu_to_be32(ring->prod & 0xffff);
+}
+
+/* slow path */
+static void mlx4_en_free_rx_desc(const struct mlx4_en_priv *priv,
+				 struct mlx4_en_rx_ring *ring,
+				 int index)
+{
+	struct mlx4_en_rx_alloc *frags;
+	int nr;
+
+	frags = ring->rx_info + (index << priv->log_rx_info);
+	for (nr = 0; nr < priv->num_frags; nr++) {
+		en_dbg(DRV, priv, "Freeing fragment:%d\n", nr);
+		mlx4_en_free_frag(priv, frags + nr);
+	}
+}
+
+/* Function not in fast-path */
+static int mlx4_en_fill_rx_buffers(struct mlx4_en_priv *priv)
+{
+	struct mlx4_en_rx_ring *ring;
+	int ring_ind;
+	int buf_ind;
+	int new_size;
+
+	for (buf_ind = 0; buf_ind < priv->prof->rx_ring_size; buf_ind++) {
+		for (ring_ind = 0; ring_ind < priv->rx_ring_num; ring_ind++) {
+			ring = priv->rx_ring[ring_ind];
+
+			if (mlx4_en_prepare_rx_desc(priv, ring,
+						    ring->actual_size,
+						    GFP_KERNEL)) {
+				if (ring->actual_size < MLX4_EN_MIN_RX_SIZE) {
+					en_err(priv, "Failed to allocate enough rx buffers\n");
+					return -ENOMEM;
+				} else {
+					new_size = rounddown_pow_of_two(ring->actual_size);
+					en_warn(priv, "Only %d buffers allocated reducing ring size to %d\n",
+						ring->actual_size, new_size);
+					goto reduce_rings;
+				}
+			}
+			ring->actual_size++;
+			ring->prod++;
+		}
+	}
+	return 0;
+
+reduce_rings:
+	for (ring_ind = 0; ring_ind < priv->rx_ring_num; ring_ind++) {
+		ring = priv->rx_ring[ring_ind];
+		while (ring->actual_size > new_size) {
+			ring->actual_size--;
+			ring->prod--;
+			mlx4_en_free_rx_desc(priv, ring, ring->actual_size);
+		}
+	}
+
+	return 0;
+}
+
+static void mlx4_en_free_rx_buf(struct mlx4_en_priv *priv,
+				struct mlx4_en_rx_ring *ring)
+{
+	int index;
+
+	en_dbg(DRV, priv, "Freeing Rx buf - cons:%d prod:%d\n",
+	       ring->cons, ring->prod);
+
+	/* Unmap and free Rx buffers */
+	for (index = 0; index < ring->size; index++) {
+		en_dbg(DRV, priv, "Processing descriptor:%d\n", index);
+		mlx4_en_free_rx_desc(priv, ring, index);
+	}
+	ring->cons = 0;
+	ring->prod = 0;
+}
+
+void mlx4_en_set_num_rx_rings(struct mlx4_en_dev *mdev)
+{
+	int i;
+	int num_of_eqs;
+	int num_rx_rings;
+	struct mlx4_dev *dev = mdev->dev;
+
+	mlx4_foreach_port(i, dev, MLX4_PORT_TYPE_ETH) {
+		num_of_eqs = max_t(int, MIN_RX_RINGS,
+				   min_t(int,
+					 mlx4_get_eqs_per_port(mdev->dev, i),
+					 DEF_RX_RINGS));
+
+		num_rx_rings = mlx4_low_memory_profile() ? MIN_RX_RINGS :
+			min_t(int, num_of_eqs, num_online_cpus());
+		mdev->profile.prof[i].rx_ring_num =
+			rounddown_pow_of_two(num_rx_rings);
+	}
+}
+
+int mlx4_en_create_rx_ring(struct mlx4_en_priv *priv,
+			   struct mlx4_en_rx_ring **pring,
+			   u32 size, u16 stride, int node, int queue_index)
+{
+	struct mlx4_en_dev *mdev = priv->mdev;
+	struct mlx4_en_rx_ring *ring;
+	int err = -ENOMEM;
+	int tmp;
+
+	ring = kzalloc_node(sizeof(*ring), GFP_KERNEL, node);
+	if (!ring) {
+		ring = kzalloc(sizeof(*ring), GFP_KERNEL);
+		if (!ring) {
+			en_err(priv, "Failed to allocate RX ring structure\n");
+			return -ENOMEM;
+		}
+	}
+
+	ring->prod = 0;
+	ring->cons = 0;
+	ring->size = size;
+	ring->size_mask = size - 1;
+	ring->stride = stride;
+	ring->log_stride = ffs(ring->stride) - 1;
+	ring->buf_size = ring->size * ring->stride + TXBB_SIZE;
+
+	if (xdp_rxq_info_reg(&ring->xdp_rxq, priv->dev, queue_index) < 0)
+		goto err_ring;
+
+	tmp = size * roundup_pow_of_two(MLX4_EN_MAX_RX_FRAGS *
+					sizeof(struct mlx4_en_rx_alloc));
+	ring->rx_info = kvzalloc_node(tmp, GFP_KERNEL, node);
+	if (!ring->rx_info) {
+		err = -ENOMEM;
+		goto err_xdp_info;
+	}
+
+	en_dbg(DRV, priv, "Allocated rx_info ring at addr:%p size:%d\n",
+		 ring->rx_info, tmp);
+
+	/* Allocate HW buffers on provided NUMA node */
+	set_dev_node(&mdev->dev->persist->pdev->dev, node);
+	err = mlx4_alloc_hwq_res(mdev->dev, &ring->wqres, ring->buf_size);
+	set_dev_node(&mdev->dev->persist->pdev->dev, mdev->dev->numa_node);
+	if (err)
+		goto err_info;
+
+	ring->buf = ring->wqres.buf.direct.buf;
+
+	ring->hwtstamp_rx_filter = priv->hwtstamp_config.rx_filter;
+
+	*pring = ring;
+	return 0;
+
+err_info:
+	kvfree(ring->rx_info);
+	ring->rx_info = NULL;
+err_xdp_info:
+	xdp_rxq_info_unreg(&ring->xdp_rxq);
+err_ring:
+	kfree(ring);
+	*pring = NULL;
+
+	return err;
+}
+
+int mlx4_en_activate_rx_rings(struct mlx4_en_priv *priv)
+{
+	struct mlx4_en_rx_ring *ring;
+	int i;
+	int ring_ind;
+	int err;
+	int stride = roundup_pow_of_two(sizeof(struct mlx4_en_rx_desc) +
+					DS_SIZE * priv->num_frags);
+
+	for (ring_ind = 0; ring_ind < priv->rx_ring_num; ring_ind++) {
+		ring = priv->rx_ring[ring_ind];
+
+		ring->prod = 0;
+		ring->cons = 0;
+		ring->actual_size = 0;
+		ring->cqn = priv->rx_cq[ring_ind]->mcq.cqn;
+
+		ring->stride = stride;
+		if (ring->stride <= TXBB_SIZE) {
+			/* Stamp first unused send wqe */
+			__be32 *ptr = (__be32 *)ring->buf;
+			__be32 stamp = cpu_to_be32(1 << STAMP_SHIFT);
+			*ptr = stamp;
+			/* Move pointer to start of rx section */
+			ring->buf += TXBB_SIZE;
+		}
+
+		ring->log_stride = ffs(ring->stride) - 1;
+		ring->buf_size = ring->size * ring->stride;
+
+		memset(ring->buf, 0, ring->buf_size);
+		mlx4_en_update_rx_prod_db(ring);
+
+		/* Initialize all descriptors */
+		for (i = 0; i < ring->size; i++)
+			mlx4_en_init_rx_desc(priv, ring, i);
+	}
+	err = mlx4_en_fill_rx_buffers(priv);
+	if (err)
+		goto err_buffers;
+
+	for (ring_ind = 0; ring_ind < priv->rx_ring_num; ring_ind++) {
+		ring = priv->rx_ring[ring_ind];
+
+		ring->size_mask = ring->actual_size - 1;
+		mlx4_en_update_rx_prod_db(ring);
+	}
+
+	return 0;
+
+err_buffers:
+	for (ring_ind = 0; ring_ind < priv->rx_ring_num; ring_ind++)
+		mlx4_en_free_rx_buf(priv, priv->rx_ring[ring_ind]);
+
+	ring_ind = priv->rx_ring_num - 1;
+	while (ring_ind >= 0) {
+		if (priv->rx_ring[ring_ind]->stride <= TXBB_SIZE)
+			priv->rx_ring[ring_ind]->buf -= TXBB_SIZE;
+		ring_ind--;
+	}
+	return err;
+}
+
+/* We recover from out of memory by scheduling our napi poll
+ * function (mlx4_en_process_cq), which tries to allocate
+ * all missing RX buffers (call to mlx4_en_refill_rx_buffers).
+ */
+void mlx4_en_recover_from_oom(struct mlx4_en_priv *priv)
+{
+	int ring;
+
+	if (!priv->port_up)
+		return;
+
+	for (ring = 0; ring < priv->rx_ring_num; ring++) {
+		if (mlx4_en_is_ring_empty(priv->rx_ring[ring])) {
+			local_bh_disable();
+			napi_reschedule(&priv->rx_cq[ring]->napi);
+			local_bh_enable();
+		}
+	}
+}
+
+/* When the rx ring is running in page-per-packet mode, a released frame can go
+ * directly into a small cache, to avoid unmapping or touching the page
+ * allocator. In bpf prog performance scenarios, buffers are either forwarded
+ * or dropped, never converted to skbs, so every page can come directly from
+ * this cache when it is sized to be a multiple of the napi budget.
+ */
+bool mlx4_en_rx_recycle(struct mlx4_en_rx_ring *ring,
+			struct mlx4_en_rx_alloc *frame)
+{
+	struct mlx4_en_page_cache *cache = &ring->page_cache;
+
+	if (cache->index >= MLX4_EN_CACHE_SIZE)
+		return false;
+
+	cache->buf[cache->index].page = frame->page;
+	cache->buf[cache->index].dma = frame->dma;
+	cache->index++;
+	return true;
+}
+
+void mlx4_en_destroy_rx_ring(struct mlx4_en_priv *priv,
+			     struct mlx4_en_rx_ring **pring,
+			     u32 size, u16 stride)
+{
+	struct mlx4_en_dev *mdev = priv->mdev;
+	struct mlx4_en_rx_ring *ring = *pring;
+	struct bpf_prog *old_prog;
+
+	old_prog = rcu_dereference_protected(
+					ring->xdp_prog,
+					lockdep_is_held(&mdev->state_lock));
+	if (old_prog)
+		bpf_prog_put(old_prog);
+	xdp_rxq_info_unreg(&ring->xdp_rxq);
+	mlx4_free_hwq_res(mdev->dev, &ring->wqres, size * stride + TXBB_SIZE);
+	kvfree(ring->rx_info);
+	ring->rx_info = NULL;
+	kfree(ring);
+	*pring = NULL;
+}
+
+void mlx4_en_deactivate_rx_ring(struct mlx4_en_priv *priv,
+				struct mlx4_en_rx_ring *ring)
+{
+	int i;
+
+	for (i = 0; i < ring->page_cache.index; i++) {
+		dma_unmap_page(priv->ddev, ring->page_cache.buf[i].dma,
+			       PAGE_SIZE, priv->dma_dir);
+		put_page(ring->page_cache.buf[i].page);
+	}
+	ring->page_cache.index = 0;
+	mlx4_en_free_rx_buf(priv, ring);
+	if (ring->stride <= TXBB_SIZE)
+		ring->buf -= TXBB_SIZE;
+}
+
+
+static int mlx4_en_complete_rx_desc(struct mlx4_en_priv *priv,
+				    struct mlx4_en_rx_alloc *frags,
+				    struct sk_buff *skb,
+				    int length)
+{
+	const struct mlx4_en_frag_info *frag_info = priv->frag_info;
+	unsigned int truesize = 0;
+	bool release = true;
+	int nr, frag_size;
+	struct page *page;
+	dma_addr_t dma;
+
+	/* Collect used fragments while replacing them in the HW descriptors */
+	for (nr = 0;; frags++) {
+		frag_size = min_t(int, length, frag_info->frag_size);
+
+		page = frags->page;
+		if (unlikely(!page))
+			goto fail;
+
+		dma = frags->dma;
+		dma_sync_single_range_for_cpu(priv->ddev, dma, frags->page_offset,
+					      frag_size, priv->dma_dir);
+
+		__skb_fill_page_desc(skb, nr, page, frags->page_offset,
+				     frag_size);
+
+		truesize += frag_info->frag_stride;
+		if (frag_info->frag_stride == PAGE_SIZE / 2) {
+			frags->page_offset ^= PAGE_SIZE / 2;
+			release = page_count(page) != 1 ||
+				  page_is_pfmemalloc(page) ||
+				  page_to_nid(page) != numa_mem_id();
+		} else if (!priv->rx_headroom) {
+			/* rx_headroom for non XDP setup is always 0.
+			 * When XDP is set, the above condition will
+			 * guarantee page is always released.
+			 */
+			u32 sz_align = ALIGN(frag_size, SMP_CACHE_BYTES);
+
+			frags->page_offset += sz_align;
+			release = frags->page_offset + frag_info->frag_size > PAGE_SIZE;
+		}
+		if (release) {
+			dma_unmap_page(priv->ddev, dma, PAGE_SIZE, priv->dma_dir);
+			frags->page = NULL;
+		} else {
+			page_ref_inc(page);
+		}
+
+		nr++;
+		length -= frag_size;
+		if (!length)
+			break;
+		frag_info++;
+	}
+	skb->truesize += truesize;
+	return nr;
+
+fail:
+	while (nr > 0) {
+		nr--;
+		__skb_frag_unref(skb_shinfo(skb)->frags + nr);
+	}
+	return 0;
+}
+
+static void validate_loopback(struct mlx4_en_priv *priv, void *va)
+{
+	const unsigned char *data = va + ETH_HLEN;
+	int i;
+
+	for (i = 0; i < MLX4_LOOPBACK_TEST_PAYLOAD; i++) {
+		if (data[i] != (unsigned char)i)
+			return;
+	}
+	/* Loopback found */
+	priv->loopback_ok = 1;
+}
+
+static void mlx4_en_refill_rx_buffers(struct mlx4_en_priv *priv,
+				      struct mlx4_en_rx_ring *ring)
+{
+	u32 missing = ring->actual_size - (ring->prod - ring->cons);
+
+	/* Try to batch allocations, but not too much. */
+	if (missing < 8)
+		return;
+	do {
+		if (mlx4_en_prepare_rx_desc(priv, ring,
+					    ring->prod & ring->size_mask,
+					    GFP_ATOMIC | __GFP_MEMALLOC))
+			break;
+		ring->prod++;
+	} while (likely(--missing));
+
+	mlx4_en_update_rx_prod_db(ring);
+}
+
+/* When hardware doesn't strip the vlan, we need to calculate the checksum
+ * over it and add it to the hardware's checksum calculation
+ */
+static inline __wsum get_fixed_vlan_csum(__wsum hw_checksum,
+					 struct vlan_hdr *vlanh)
+{
+	return csum_add(hw_checksum, *(__wsum *)vlanh);
+}
+
+/* Although the stack expects checksum which doesn't include the pseudo
+ * header, the HW adds it. To address that, we are subtracting the pseudo
+ * header checksum from the checksum value provided by the HW.
+ */
+static int get_fixed_ipv4_csum(__wsum hw_checksum, struct sk_buff *skb,
+			       struct iphdr *iph)
+{
+	__u16 length_for_csum = 0;
+	__wsum csum_pseudo_header = 0;
+	__u8 ipproto = iph->protocol;
+
+	if (unlikely(ipproto == IPPROTO_SCTP))
+		return -1;
+
+	length_for_csum = (be16_to_cpu(iph->tot_len) - (iph->ihl << 2));
+	csum_pseudo_header = csum_tcpudp_nofold(iph->saddr, iph->daddr,
+						length_for_csum, ipproto, 0);
+	skb->csum = csum_sub(hw_checksum, csum_pseudo_header);
+	return 0;
+}
+
+#if IS_ENABLED(CONFIG_IPV6)
+/* In IPv6 packets, hw_checksum lacks 6 bytes from IPv6 header:
+ * 4 first bytes : priority, version, flow_lbl
+ * and 2 additional bytes : nexthdr, hop_limit.
+ */
+static int get_fixed_ipv6_csum(__wsum hw_checksum, struct sk_buff *skb,
+			       struct ipv6hdr *ipv6h)
+{
+	__u8 nexthdr = ipv6h->nexthdr;
+	__wsum temp;
+
+	if (unlikely(nexthdr == IPPROTO_FRAGMENT ||
+		     nexthdr == IPPROTO_HOPOPTS ||
+		     nexthdr == IPPROTO_SCTP))
+		return -1;
+
+	/* priority, version, flow_lbl */
+	temp = csum_add(hw_checksum, *(__wsum *)ipv6h);
+	/* nexthdr and hop_limit */
+	skb->csum = csum_add(temp, (__force __wsum)*(__be16 *)&ipv6h->nexthdr);
+	return 0;
+}
+#endif
+
+#define short_frame(size) ((size) <= ETH_ZLEN + ETH_FCS_LEN)
+
+/* We reach this function only after checking that any of
+ * the (IPv4 | IPv6) bits are set in cqe->status.
+ */
+static int check_csum(struct mlx4_cqe *cqe, struct sk_buff *skb, void *va,
+		      netdev_features_t dev_features)
+{
+	__wsum hw_checksum = 0;
+	void *hdr;
+
+	/* CQE csum doesn't cover padding octets in short ethernet
+	 * frames. And the pad field is appended prior to calculating
+	 * and appending the FCS field.
+	 *
+	 * Detecting these padded frames requires to verify and parse
+	 * IP headers, so we simply force all those small frames to skip
+	 * checksum complete.
+	 */
+	if (short_frame(skb->len))
+		return -EINVAL;
+
+	hdr = (u8 *)va + sizeof(struct ethhdr);
+	hw_checksum = csum_unfold((__force __sum16)cqe->checksum);
+
+	if (cqe->vlan_my_qpn & cpu_to_be32(MLX4_CQE_CVLAN_PRESENT_MASK) &&
+	    !(dev_features & NETIF_F_HW_VLAN_CTAG_RX)) {
+		hw_checksum = get_fixed_vlan_csum(hw_checksum, hdr);
+		hdr += sizeof(struct vlan_hdr);
+	}
+
+#if IS_ENABLED(CONFIG_IPV6)
+	if (cqe->status & cpu_to_be16(MLX4_CQE_STATUS_IPV6))
+		return get_fixed_ipv6_csum(hw_checksum, skb, hdr);
+#endif
+	return get_fixed_ipv4_csum(hw_checksum, skb, hdr);
+}
+
+#if IS_ENABLED(CONFIG_IPV6)
+#define MLX4_CQE_STATUS_IP_ANY (MLX4_CQE_STATUS_IPV4 | MLX4_CQE_STATUS_IPV6)
+#else
+#define MLX4_CQE_STATUS_IP_ANY (MLX4_CQE_STATUS_IPV4)
+#endif
+
+int mlx4_en_process_rx_cq(struct net_device *dev, struct mlx4_en_cq *cq, int budget)
+{
+	struct mlx4_en_priv *priv = netdev_priv(dev);
+	int factor = priv->cqe_factor;
+	struct mlx4_en_rx_ring *ring;
+	struct bpf_prog *xdp_prog;
+	int cq_ring = cq->ring;
+	bool doorbell_pending;
+	struct mlx4_cqe *cqe;
+	struct xdp_buff xdp;
+	int polled = 0;
+	int index;
+
+	if (unlikely(!priv->port_up || budget <= 0))
+		return 0;
+
+	ring = priv->rx_ring[cq_ring];
+
+	/* Protect accesses to: ring->xdp_prog, priv->mac_hash list */
+	rcu_read_lock();
+	xdp_prog = rcu_dereference(ring->xdp_prog);
+	xdp.rxq = &ring->xdp_rxq;
+	doorbell_pending = 0;
+
+	/* We assume a 1:1 mapping between CQEs and Rx descriptors, so Rx
+	 * descriptor offset can be deduced from the CQE index instead of
+	 * reading 'cqe->index' */
+	index = cq->mcq.cons_index & ring->size_mask;
+	cqe = mlx4_en_get_cqe(cq->buf, index, priv->cqe_size) + factor;
+
+	/* Process all completed CQEs */
+	while (XNOR(cqe->owner_sr_opcode & MLX4_CQE_OWNER_MASK,
+		    cq->mcq.cons_index & cq->size)) {
+		struct mlx4_en_rx_alloc *frags;
+		enum pkt_hash_types hash_type;
+		struct sk_buff *skb;
+		unsigned int length;
+		int ip_summed;
+		void *va;
+		int nr;
+
+		frags = ring->rx_info + (index << priv->log_rx_info);
+		va = page_address(frags[0].page) + frags[0].page_offset;
+		prefetchw(va);
+		/*
+		 * make sure we read the CQE after we read the ownership bit
+		 */
+		dma_rmb();
+
+		/* Drop packet on bad receive or bad checksum */
+		if (unlikely((cqe->owner_sr_opcode & MLX4_CQE_OPCODE_MASK) ==
+						MLX4_CQE_OPCODE_ERROR)) {
+			en_err(priv, "CQE completed in error - vendor syndrom:%d syndrom:%d\n",
+			       ((struct mlx4_err_cqe *)cqe)->vendor_err_syndrome,
+			       ((struct mlx4_err_cqe *)cqe)->syndrome);
+			goto next;
+		}
+		if (unlikely(cqe->badfcs_enc & MLX4_CQE_BAD_FCS)) {
+			en_dbg(RX_ERR, priv, "Accepted frame with bad FCS\n");
+			goto next;
+		}
+
+		/* Check if we need to drop the packet if SRIOV is not enabled
+		 * and not performing the selftest or flb disabled
+		 */
+		if (priv->flags & MLX4_EN_FLAG_RX_FILTER_NEEDED) {
+			const struct ethhdr *ethh = va;
+			dma_addr_t dma;
+			/* Get pointer to first fragment since we haven't
+			 * skb yet and cast it to ethhdr struct
+			 */
+			dma = frags[0].dma + frags[0].page_offset;
+			dma_sync_single_for_cpu(priv->ddev, dma, sizeof(*ethh),
+						DMA_FROM_DEVICE);
+
+			if (is_multicast_ether_addr(ethh->h_dest)) {
+				struct mlx4_mac_entry *entry;
+				struct hlist_head *bucket;
+				unsigned int mac_hash;
+
+				/* Drop the packet, since HW loopback-ed it */
+				mac_hash = ethh->h_source[MLX4_EN_MAC_HASH_IDX];
+				bucket = &priv->mac_hash[mac_hash];
+				hlist_for_each_entry_rcu(entry, bucket, hlist) {
+					if (ether_addr_equal_64bits(entry->mac,
+								    ethh->h_source))
+						goto next;
+				}
+			}
+		}
+
+		if (unlikely(priv->validate_loopback)) {
+			validate_loopback(priv, va);
+			goto next;
+		}
+
+		/*
+		 * Packet is OK - process it.
+		 */
+		length = be32_to_cpu(cqe->byte_cnt);
+		length -= ring->fcs_del;
+
+		/* A bpf program gets first chance to drop the packet. It may
+		 * read bytes but not past the end of the frag.
+		 */
+		if (xdp_prog) {
+			dma_addr_t dma;
+			void *orig_data;
+			u32 act;
+
+			dma = frags[0].dma + frags[0].page_offset;
+			dma_sync_single_for_cpu(priv->ddev, dma,
+						priv->frag_info[0].frag_size,
+						DMA_FROM_DEVICE);
+
+			xdp.data_hard_start = va - frags[0].page_offset;
+			xdp.data = va;
+			xdp_set_data_meta_invalid(&xdp);
+			xdp.data_end = xdp.data + length;
+			orig_data = xdp.data;
+
+			act = bpf_prog_run_xdp(xdp_prog, &xdp);
+
+			length = xdp.data_end - xdp.data;
+			if (xdp.data != orig_data) {
+				frags[0].page_offset = xdp.data -
+					xdp.data_hard_start;
+				va = xdp.data;
+			}
+
+			switch (act) {
+			case XDP_PASS:
+				break;
+			case XDP_TX:
+				if (likely(!mlx4_en_xmit_frame(ring, frags, priv,
+							length, cq_ring,
+							&doorbell_pending))) {
+					frags[0].page = NULL;
+					goto next;
+				}
+				trace_xdp_exception(dev, xdp_prog, act);
+				goto xdp_drop_no_cnt; /* Drop on xmit failure */
+			default:
+				bpf_warn_invalid_xdp_action(act);
+				/* fall through */
+			case XDP_ABORTED:
+				trace_xdp_exception(dev, xdp_prog, act);
+				/* fall through */
+			case XDP_DROP:
+				ring->xdp_drop++;
+xdp_drop_no_cnt:
+				goto next;
+			}
+		}
+
+		ring->bytes += length;
+		ring->packets++;
+
+		skb = napi_get_frags(&cq->napi);
+		if (unlikely(!skb))
+			goto next;
+
+		if (unlikely(ring->hwtstamp_rx_filter == HWTSTAMP_FILTER_ALL)) {
+			u64 timestamp = mlx4_en_get_cqe_ts(cqe);
+
+			mlx4_en_fill_hwtstamps(priv->mdev, skb_hwtstamps(skb),
+					       timestamp);
+		}
+		skb_record_rx_queue(skb, cq_ring);
+
+		if (likely(dev->features & NETIF_F_RXCSUM)) {
+			/* TODO: For IP non TCP/UDP packets when csum complete is
+			 * not an option (not supported or any other reason) we can
+			 * actually check cqe IPOK status bit and report
+			 * CHECKSUM_UNNECESSARY rather than CHECKSUM_NONE
+			 */
+			if ((cqe->status & cpu_to_be16(MLX4_CQE_STATUS_TCP |
+						       MLX4_CQE_STATUS_UDP)) &&
+			    (cqe->status & cpu_to_be16(MLX4_CQE_STATUS_IPOK)) &&
+			    cqe->checksum == cpu_to_be16(0xffff)) {
+				bool l2_tunnel;
+
+				l2_tunnel = (dev->hw_enc_features & NETIF_F_RXCSUM) &&
+					(cqe->vlan_my_qpn & cpu_to_be32(MLX4_CQE_L2_TUNNEL));
+				ip_summed = CHECKSUM_UNNECESSARY;
+				hash_type = PKT_HASH_TYPE_L4;
+				if (l2_tunnel)
+					skb->csum_level = 1;
+				ring->csum_ok++;
+			} else {
+				if (!(priv->flags & MLX4_EN_FLAG_RX_CSUM_NON_TCP_UDP &&
+				      (cqe->status & cpu_to_be16(MLX4_CQE_STATUS_IP_ANY))))
+					goto csum_none;
+				if (check_csum(cqe, skb, va, dev->features))
+					goto csum_none;
+				ip_summed = CHECKSUM_COMPLETE;
+				hash_type = PKT_HASH_TYPE_L3;
+				ring->csum_complete++;
+			}
+		} else {
+csum_none:
+			ip_summed = CHECKSUM_NONE;
+			hash_type = PKT_HASH_TYPE_L3;
+			ring->csum_none++;
+		}
+		skb->ip_summed = ip_summed;
+		if (dev->features & NETIF_F_RXHASH)
+			skb_set_hash(skb,
+				     be32_to_cpu(cqe->immed_rss_invalid),
+				     hash_type);
+
+		if ((cqe->vlan_my_qpn &
+		     cpu_to_be32(MLX4_CQE_CVLAN_PRESENT_MASK)) &&
+		    (dev->features & NETIF_F_HW_VLAN_CTAG_RX))
+			__vlan_hwaccel_put_tag(skb, htons(ETH_P_8021Q),
+					       be16_to_cpu(cqe->sl_vid));
+		else if ((cqe->vlan_my_qpn &
+			  cpu_to_be32(MLX4_CQE_SVLAN_PRESENT_MASK)) &&
+			 (dev->features & NETIF_F_HW_VLAN_STAG_RX))
+			__vlan_hwaccel_put_tag(skb, htons(ETH_P_8021AD),
+					       be16_to_cpu(cqe->sl_vid));
+
+		nr = mlx4_en_complete_rx_desc(priv, frags, skb, length);
+		if (likely(nr)) {
+			skb_shinfo(skb)->nr_frags = nr;
+			skb->len = length;
+			skb->data_len = length;
+			napi_gro_frags(&cq->napi);
+		} else {
+			skb->vlan_tci = 0;
+			skb_clear_hash(skb);
+		}
+next:
+		++cq->mcq.cons_index;
+		index = (cq->mcq.cons_index) & ring->size_mask;
+		cqe = mlx4_en_get_cqe(cq->buf, index, priv->cqe_size) + factor;
+		if (unlikely(++polled == budget))
+			break;
+	}
+
+	rcu_read_unlock();
+
+	if (likely(polled)) {
+		if (doorbell_pending) {
+			priv->tx_cq[TX_XDP][cq_ring]->xdp_busy = true;
+			mlx4_en_xmit_doorbell(priv->tx_ring[TX_XDP][cq_ring]);
+		}
+
+		mlx4_cq_set_ci(&cq->mcq);
+		wmb(); /* ensure HW sees CQ consumer before we post new buffers */
+		ring->cons = cq->mcq.cons_index;
+	}
+	AVG_PERF_COUNTER(priv->pstats.rx_coal_avg, polled);
+
+	mlx4_en_refill_rx_buffers(priv, ring);
+
+	return polled;
+}
+
+
+void mlx4_en_rx_irq(struct mlx4_cq *mcq)
+{
+	struct mlx4_en_cq *cq = container_of(mcq, struct mlx4_en_cq, mcq);
+	struct mlx4_en_priv *priv = netdev_priv(cq->dev);
+
+	if (likely(priv->port_up))
+		napi_schedule_irqoff(&cq->napi);
+	else
+		mlx4_en_arm_cq(priv, cq);
+}
+
+/* Rx CQ polling - called by NAPI */
+int mlx4_en_poll_rx_cq(struct napi_struct *napi, int budget)
+{
+	struct mlx4_en_cq *cq = container_of(napi, struct mlx4_en_cq, napi);
+	struct net_device *dev = cq->dev;
+	struct mlx4_en_priv *priv = netdev_priv(dev);
+	struct mlx4_en_cq *xdp_tx_cq = NULL;
+	bool clean_complete = true;
+	int done;
+
+	if (!budget)
+		return 0;
+
+	if (priv->tx_ring_num[TX_XDP]) {
+		xdp_tx_cq = priv->tx_cq[TX_XDP][cq->ring];
+		if (xdp_tx_cq->xdp_busy) {
+			clean_complete = mlx4_en_process_tx_cq(dev, xdp_tx_cq,
+							       budget);
+			xdp_tx_cq->xdp_busy = !clean_complete;
+		}
+	}
+
+	done = mlx4_en_process_rx_cq(dev, cq, budget);
+
+	/* If we used up all the quota - we're probably not done yet... */
+	if (done == budget || !clean_complete) {
+		const struct cpumask *aff;
+		struct irq_data *idata;
+		int cpu_curr;
+
+		/* in case we got here because of !clean_complete */
+		done = budget;
+
+		INC_PERF_COUNTER(priv->pstats.napi_quota);
+
+		cpu_curr = smp_processor_id();
+		idata = irq_desc_get_irq_data(cq->irq_desc);
+		aff = irq_data_get_affinity_mask(idata);
+
+		if (likely(cpumask_test_cpu(cpu_curr, aff)))
+			return budget;
+
+		/* Current cpu is not according to smp_irq_affinity -
+		 * probably affinity changed. Need to stop this NAPI
+		 * poll, and restart it on the right CPU.
+		 * Try to avoid returning a too small value (like 0),
+		 * to not fool net_rx_action() and its netdev_budget
+		 */
+		if (done)
+			done--;
+	}
+	/* Done for now */
+	if (likely(napi_complete_done(napi, done)))
+		mlx4_en_arm_cq(priv, cq);
+	return done;
+}
+
+void mlx4_en_calc_rx_buf(struct net_device *dev)
+{
+	struct mlx4_en_priv *priv = netdev_priv(dev);
+	int eff_mtu = MLX4_EN_EFF_MTU(dev->mtu);
+	int i = 0;
+
+	/* bpf requires buffers to be set up as 1 packet per page.
+	 * This only works when num_frags == 1.
+	 */
+	if (priv->tx_ring_num[TX_XDP]) {
+		priv->frag_info[0].frag_size = eff_mtu;
+		/* This will gain efficient xdp frame recycling at the
+		 * expense of more costly truesize accounting
+		 */
+		priv->frag_info[0].frag_stride = PAGE_SIZE;
+		priv->dma_dir = PCI_DMA_BIDIRECTIONAL;
+		priv->rx_headroom = XDP_PACKET_HEADROOM;
+		i = 1;
+	} else {
+		int frag_size_max = 2048, buf_size = 0;
+
+		/* should not happen, right ? */
+		if (eff_mtu > PAGE_SIZE + (MLX4_EN_MAX_RX_FRAGS - 1) * 2048)
+			frag_size_max = PAGE_SIZE;
+
+		while (buf_size < eff_mtu) {
+			int frag_stride, frag_size = eff_mtu - buf_size;
+			int pad, nb;
+
+			if (i < MLX4_EN_MAX_RX_FRAGS - 1)
+				frag_size = min(frag_size, frag_size_max);
+
+			priv->frag_info[i].frag_size = frag_size;
+			frag_stride = ALIGN(frag_size, SMP_CACHE_BYTES);
+			/* We can only pack 2 1536-bytes frames in on 4K page
+			 * Therefore, each frame would consume more bytes (truesize)
+			 */
+			nb = PAGE_SIZE / frag_stride;
+			pad = (PAGE_SIZE - nb * frag_stride) / nb;
+			pad &= ~(SMP_CACHE_BYTES - 1);
+			priv->frag_info[i].frag_stride = frag_stride + pad;
+
+			buf_size += frag_size;
+			i++;
+		}
+		priv->dma_dir = PCI_DMA_FROMDEVICE;
+		priv->rx_headroom = 0;
+	}
+
+	priv->num_frags = i;
+	priv->rx_skb_size = eff_mtu;
+	priv->log_rx_info = ROUNDUP_LOG2(i * sizeof(struct mlx4_en_rx_alloc));
+
+	en_dbg(DRV, priv, "Rx buffer scatter-list (effective-mtu:%d num_frags:%d):\n",
+	       eff_mtu, priv->num_frags);
+	for (i = 0; i < priv->num_frags; i++) {
+		en_dbg(DRV,
+		       priv,
+		       "  frag:%d - size:%d stride:%d\n",
+		       i,
+		       priv->frag_info[i].frag_size,
+		       priv->frag_info[i].frag_stride);
+	}
+}
+
+/* RSS related functions */
+
+static int mlx4_en_config_rss_qp(struct mlx4_en_priv *priv, int qpn,
+				 struct mlx4_en_rx_ring *ring,
+				 enum mlx4_qp_state *state,
+				 struct mlx4_qp *qp)
+{
+	struct mlx4_en_dev *mdev = priv->mdev;
+	struct mlx4_qp_context *context;
+	int err = 0;
+
+	context = kmalloc(sizeof(*context), GFP_KERNEL);
+	if (!context)
+		return -ENOMEM;
+
+	err = mlx4_qp_alloc(mdev->dev, qpn, qp);
+	if (err) {
+		en_err(priv, "Failed to allocate qp #%x\n", qpn);
+		goto out;
+	}
+	qp->event = mlx4_en_sqp_event;
+
+	memset(context, 0, sizeof(*context));
+	mlx4_en_fill_qp_context(priv, ring->actual_size, ring->stride, 0, 0,
+				qpn, ring->cqn, -1, context);
+	context->db_rec_addr = cpu_to_be64(ring->wqres.db.dma);
+
+	/* Cancel FCS removal if FW allows */
+	if (mdev->dev->caps.flags & MLX4_DEV_CAP_FLAG_FCS_KEEP) {
+		context->param3 |= cpu_to_be32(1 << 29);
+		if (priv->dev->features & NETIF_F_RXFCS)
+			ring->fcs_del = 0;
+		else
+			ring->fcs_del = ETH_FCS_LEN;
+	} else
+		ring->fcs_del = 0;
+
+	err = mlx4_qp_to_ready(mdev->dev, &ring->wqres.mtt, context, qp, state);
+	if (err) {
+		mlx4_qp_remove(mdev->dev, qp);
+		mlx4_qp_free(mdev->dev, qp);
+	}
+	mlx4_en_update_rx_prod_db(ring);
+out:
+	kfree(context);
+	return err;
+}
+
+int mlx4_en_create_drop_qp(struct mlx4_en_priv *priv)
+{
+	int err;
+	u32 qpn;
+
+	err = mlx4_qp_reserve_range(priv->mdev->dev, 1, 1, &qpn,
+				    MLX4_RESERVE_A0_QP,
+				    MLX4_RES_USAGE_DRIVER);
+	if (err) {
+		en_err(priv, "Failed reserving drop qpn\n");
+		return err;
+	}
+	err = mlx4_qp_alloc(priv->mdev->dev, qpn, &priv->drop_qp);
+	if (err) {
+		en_err(priv, "Failed allocating drop qp\n");
+		mlx4_qp_release_range(priv->mdev->dev, qpn, 1);
+		return err;
+	}
+
+	return 0;
+}
+
+void mlx4_en_destroy_drop_qp(struct mlx4_en_priv *priv)
+{
+	u32 qpn;
+
+	qpn = priv->drop_qp.qpn;
+	mlx4_qp_remove(priv->mdev->dev, &priv->drop_qp);
+	mlx4_qp_free(priv->mdev->dev, &priv->drop_qp);
+	mlx4_qp_release_range(priv->mdev->dev, qpn, 1);
+}
+
+/* Allocate rx qp's and configure them according to rss map */
+int mlx4_en_config_rss_steer(struct mlx4_en_priv *priv)
+{
+	struct mlx4_en_dev *mdev = priv->mdev;
+	struct mlx4_en_rss_map *rss_map = &priv->rss_map;
+	struct mlx4_qp_context context;
+	struct mlx4_rss_context *rss_context;
+	int rss_rings;
+	void *ptr;
+	u8 rss_mask = (MLX4_RSS_IPV4 | MLX4_RSS_TCP_IPV4 | MLX4_RSS_IPV6 |
+			MLX4_RSS_TCP_IPV6);
+	int i, qpn;
+	int err = 0;
+	int good_qps = 0;
+	u8 flags;
+
+	en_dbg(DRV, priv, "Configuring rss steering\n");
+
+	flags = priv->rx_ring_num == 1 ? MLX4_RESERVE_A0_QP : 0;
+	err = mlx4_qp_reserve_range(mdev->dev, priv->rx_ring_num,
+				    priv->rx_ring_num,
+				    &rss_map->base_qpn, flags,
+				    MLX4_RES_USAGE_DRIVER);
+	if (err) {
+		en_err(priv, "Failed reserving %d qps\n", priv->rx_ring_num);
+		return err;
+	}
+
+	for (i = 0; i < priv->rx_ring_num; i++) {
+		qpn = rss_map->base_qpn + i;
+		err = mlx4_en_config_rss_qp(priv, qpn, priv->rx_ring[i],
+					    &rss_map->state[i],
+					    &rss_map->qps[i]);
+		if (err)
+			goto rss_err;
+
+		++good_qps;
+	}
+
+	if (priv->rx_ring_num == 1) {
+		rss_map->indir_qp = &rss_map->qps[0];
+		priv->base_qpn = rss_map->indir_qp->qpn;
+		en_info(priv, "Optimized Non-RSS steering\n");
+		return 0;
+	}
+
+	rss_map->indir_qp = kzalloc(sizeof(*rss_map->indir_qp), GFP_KERNEL);
+	if (!rss_map->indir_qp) {
+		err = -ENOMEM;
+		goto rss_err;
+	}
+
+	/* Configure RSS indirection qp */
+	err = mlx4_qp_alloc(mdev->dev, priv->base_qpn, rss_map->indir_qp);
+	if (err) {
+		en_err(priv, "Failed to allocate RSS indirection QP\n");
+		goto qp_alloc_err;
+	}
+
+	rss_map->indir_qp->event = mlx4_en_sqp_event;
+	mlx4_en_fill_qp_context(priv, 0, 0, 0, 1, priv->base_qpn,
+				priv->rx_ring[0]->cqn, -1, &context);
+
+	if (!priv->prof->rss_rings || priv->prof->rss_rings > priv->rx_ring_num)
+		rss_rings = priv->rx_ring_num;
+	else
+		rss_rings = priv->prof->rss_rings;
+
+	ptr = ((void *) &context) + offsetof(struct mlx4_qp_context, pri_path)
+					+ MLX4_RSS_OFFSET_IN_QPC_PRI_PATH;
+	rss_context = ptr;
+	rss_context->base_qpn = cpu_to_be32(ilog2(rss_rings) << 24 |
+					    (rss_map->base_qpn));
+	rss_context->default_qpn = cpu_to_be32(rss_map->base_qpn);
+	if (priv->mdev->profile.udp_rss) {
+		rss_mask |=  MLX4_RSS_UDP_IPV4 | MLX4_RSS_UDP_IPV6;
+		rss_context->base_qpn_udp = rss_context->default_qpn;
+	}
+
+	if (mdev->dev->caps.tunnel_offload_mode == MLX4_TUNNEL_OFFLOAD_MODE_VXLAN) {
+		en_info(priv, "Setting RSS context tunnel type to RSS on inner headers\n");
+		rss_mask |= MLX4_RSS_BY_INNER_HEADERS;
+	}
+
+	rss_context->flags = rss_mask;
+	rss_context->hash_fn = MLX4_RSS_HASH_TOP;
+	if (priv->rss_hash_fn == ETH_RSS_HASH_XOR) {
+		rss_context->hash_fn = MLX4_RSS_HASH_XOR;
+	} else if (priv->rss_hash_fn == ETH_RSS_HASH_TOP) {
+		rss_context->hash_fn = MLX4_RSS_HASH_TOP;
+		memcpy(rss_context->rss_key, priv->rss_key,
+		       MLX4_EN_RSS_KEY_SIZE);
+	} else {
+		en_err(priv, "Unknown RSS hash function requested\n");
+		err = -EINVAL;
+		goto indir_err;
+	}
+
+	err = mlx4_qp_to_ready(mdev->dev, &priv->res.mtt, &context,
+			       rss_map->indir_qp, &rss_map->indir_state);
+	if (err)
+		goto indir_err;
+
+	return 0;
+
+indir_err:
+	mlx4_qp_modify(mdev->dev, NULL, rss_map->indir_state,
+		       MLX4_QP_STATE_RST, NULL, 0, 0, rss_map->indir_qp);
+	mlx4_qp_remove(mdev->dev, rss_map->indir_qp);
+	mlx4_qp_free(mdev->dev, rss_map->indir_qp);
+qp_alloc_err:
+	kfree(rss_map->indir_qp);
+	rss_map->indir_qp = NULL;
+rss_err:
+	for (i = 0; i < good_qps; i++) {
+		mlx4_qp_modify(mdev->dev, NULL, rss_map->state[i],
+			       MLX4_QP_STATE_RST, NULL, 0, 0, &rss_map->qps[i]);
+		mlx4_qp_remove(mdev->dev, &rss_map->qps[i]);
+		mlx4_qp_free(mdev->dev, &rss_map->qps[i]);
+	}
+	mlx4_qp_release_range(mdev->dev, rss_map->base_qpn, priv->rx_ring_num);
+	return err;
+}
+
+void mlx4_en_release_rss_steer(struct mlx4_en_priv *priv)
+{
+	struct mlx4_en_dev *mdev = priv->mdev;
+	struct mlx4_en_rss_map *rss_map = &priv->rss_map;
+	int i;
+
+	if (priv->rx_ring_num > 1) {
+		mlx4_qp_modify(mdev->dev, NULL, rss_map->indir_state,
+			       MLX4_QP_STATE_RST, NULL, 0, 0,
+			       rss_map->indir_qp);
+		mlx4_qp_remove(mdev->dev, rss_map->indir_qp);
+		mlx4_qp_free(mdev->dev, rss_map->indir_qp);
+		kfree(rss_map->indir_qp);
+		rss_map->indir_qp = NULL;
+	}
+
+	for (i = 0; i < priv->rx_ring_num; i++) {
+		mlx4_qp_modify(mdev->dev, NULL, rss_map->state[i],
+			       MLX4_QP_STATE_RST, NULL, 0, 0, &rss_map->qps[i]);
+		mlx4_qp_remove(mdev->dev, &rss_map->qps[i]);
+		mlx4_qp_free(mdev->dev, &rss_map->qps[i]);
+	}
+	mlx4_qp_release_range(mdev->dev, rss_map->base_qpn, priv->rx_ring_num);
+}
diff --git a/drivers/net/ethernet/mellanox/mlx4/en_selftest.c b/drivers/net/ethernet/mellanox/mlx4/en_selftest.c
new file mode 100644
index 000000000..946d9db7c
--- /dev/null
+++ b/drivers/net/ethernet/mellanox/mlx4/en_selftest.c
@@ -0,0 +1,204 @@
+/*
+ * Copyright (c) 2007 Mellanox Technologies. All rights reserved.
+ *
+ * This software is available to you under a choice of one of two
+ * licenses.  You may choose to be licensed under the terms of the GNU
+ * General Public License (GPL) Version 2, available from the file
+ * COPYING in the main directory of this source tree, or the
+ * OpenIB.org BSD license below:
+ *
+ *     Redistribution and use in source and binary forms, with or
+ *     without modification, are permitted provided that the following
+ *     conditions are met:
+ *
+ *      - Redistributions of source code must retain the above
+ *        copyright notice, this list of conditions and the following
+ *        disclaimer.
+ *
+ *      - Redistributions in binary form must reproduce the above
+ *        copyright notice, this list of conditions and the following
+ *        disclaimer in the documentation and/or other materials
+ *        provided with the distribution.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
+ * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
+ * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS
+ * BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN
+ * ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
+ * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+ * SOFTWARE.
+ *
+ */
+
+#include <linux/kernel.h>
+#include <linux/ethtool.h>
+#include <linux/netdevice.h>
+#include <linux/delay.h>
+#include <linux/mlx4/driver.h>
+
+#include "mlx4_en.h"
+
+
+static int mlx4_en_test_registers(struct mlx4_en_priv *priv)
+{
+	return mlx4_cmd(priv->mdev->dev, 0, 0, 0, MLX4_CMD_HW_HEALTH_CHECK,
+			MLX4_CMD_TIME_CLASS_A, MLX4_CMD_WRAPPED);
+}
+
+static int mlx4_en_test_loopback_xmit(struct mlx4_en_priv *priv)
+{
+	struct sk_buff *skb;
+	struct ethhdr *ethh;
+	unsigned char *packet;
+	unsigned int packet_size = MLX4_LOOPBACK_TEST_PAYLOAD;
+	unsigned int i;
+	int err;
+
+
+	/* build the pkt before xmit */
+	skb = netdev_alloc_skb(priv->dev, MLX4_LOOPBACK_TEST_PAYLOAD + ETH_HLEN + NET_IP_ALIGN);
+	if (!skb)
+		return -ENOMEM;
+
+	skb_reserve(skb, NET_IP_ALIGN);
+
+	ethh = skb_put(skb, sizeof(struct ethhdr));
+	packet = skb_put(skb, packet_size);
+	memcpy(ethh->h_dest, priv->dev->dev_addr, ETH_ALEN);
+	eth_zero_addr(ethh->h_source);
+	ethh->h_proto = htons(ETH_P_ARP);
+	skb_reset_mac_header(skb);
+	for (i = 0; i < packet_size; ++i)	/* fill our packet */
+		packet[i] = (unsigned char)(i & 0xff);
+
+	/* xmit the pkt */
+	err = mlx4_en_xmit(skb, priv->dev);
+	return err;
+}
+
+static int mlx4_en_test_loopback(struct mlx4_en_priv *priv)
+{
+	u32 loopback_ok = 0;
+	int i;
+
+        priv->loopback_ok = 0;
+	priv->validate_loopback = 1;
+
+	mlx4_en_update_loopback_state(priv->dev, priv->dev->features);
+
+	/* xmit */
+	if (mlx4_en_test_loopback_xmit(priv)) {
+		en_err(priv, "Transmitting loopback packet failed\n");
+		goto mlx4_en_test_loopback_exit;
+	}
+
+	/* polling for result */
+	for (i = 0; i < MLX4_EN_LOOPBACK_RETRIES; ++i) {
+		msleep(MLX4_EN_LOOPBACK_TIMEOUT);
+		if (priv->loopback_ok) {
+			loopback_ok = 1;
+			break;
+		}
+	}
+	if (!loopback_ok)
+		en_err(priv, "Loopback packet didn't arrive\n");
+
+mlx4_en_test_loopback_exit:
+
+	priv->validate_loopback = 0;
+
+	mlx4_en_update_loopback_state(priv->dev, priv->dev->features);
+	return !loopback_ok;
+}
+
+static int mlx4_en_test_interrupts(struct mlx4_en_priv *priv)
+{
+	struct mlx4_en_dev *mdev = priv->mdev;
+	int err = 0;
+	int i = 0;
+
+	err = mlx4_test_async(mdev->dev);
+	/* When not in MSI_X or slave, test only async */
+	if (!(mdev->dev->flags & MLX4_FLAG_MSI_X) || mlx4_is_slave(mdev->dev))
+		return err;
+
+	/* A loop over all completion vectors of current port,
+	 * for each vector check whether it works by mapping command
+	 * completions to that vector and performing a NOP command
+	 */
+	for (i = 0; i < priv->rx_ring_num; i++) {
+		err = mlx4_test_interrupt(mdev->dev, priv->rx_cq[i]->vector);
+		if (err)
+			break;
+	}
+
+	return err;
+}
+
+static int mlx4_en_test_link(struct mlx4_en_priv *priv)
+{
+	if (mlx4_en_QUERY_PORT(priv->mdev, priv->port))
+		return -ENOMEM;
+	if (priv->port_state.link_state == 1)
+		return 0;
+	else
+		return 1;
+}
+
+static int mlx4_en_test_speed(struct mlx4_en_priv *priv)
+{
+
+	if (mlx4_en_QUERY_PORT(priv->mdev, priv->port))
+		return -ENOMEM;
+
+	/* The device supports 100M, 1G, 10G, 20G, 40G and 56G speed */
+	if (priv->port_state.link_speed != SPEED_100 &&
+	    priv->port_state.link_speed != SPEED_1000 &&
+	    priv->port_state.link_speed != SPEED_10000 &&
+	    priv->port_state.link_speed != SPEED_20000 &&
+	    priv->port_state.link_speed != SPEED_40000 &&
+	    priv->port_state.link_speed != SPEED_56000)
+		return priv->port_state.link_speed;
+
+	return 0;
+}
+
+
+void mlx4_en_ex_selftest(struct net_device *dev, u32 *flags, u64 *buf)
+{
+	struct mlx4_en_priv *priv = netdev_priv(dev);
+	int i, carrier_ok;
+
+	memset(buf, 0, sizeof(u64) * MLX4_EN_NUM_SELF_TEST);
+
+	if (*flags & ETH_TEST_FL_OFFLINE) {
+		/* disable the interface */
+		carrier_ok = netif_carrier_ok(dev);
+
+		netif_carrier_off(dev);
+		/* Wait until all tx queues are empty.
+		 * there should not be any additional incoming traffic
+		 * since we turned the carrier off */
+		msleep(200);
+
+		if (priv->mdev->dev->caps.flags &
+					MLX4_DEV_CAP_FLAG_UC_LOOPBACK) {
+			buf[3] = mlx4_en_test_registers(priv);
+			if (priv->port_up && dev->mtu >= MLX4_SELFTEST_LB_MIN_MTU)
+				buf[4] = mlx4_en_test_loopback(priv);
+		}
+
+		if (carrier_ok)
+			netif_carrier_on(dev);
+
+	}
+	buf[0] = mlx4_en_test_interrupts(priv);
+	buf[1] = mlx4_en_test_link(priv);
+	buf[2] = mlx4_en_test_speed(priv);
+
+	for (i = 0; i < MLX4_EN_NUM_SELF_TEST; i++) {
+		if (buf[i])
+			*flags |= ETH_TEST_FL_FAILED;
+	}
+}
diff --git a/drivers/net/ethernet/mellanox/mlx4/en_tx.c b/drivers/net/ethernet/mellanox/mlx4/en_tx.c
new file mode 100644
index 000000000..6517e53da
--- /dev/null
+++ b/drivers/net/ethernet/mellanox/mlx4/en_tx.c
@@ -0,0 +1,1211 @@
+/*
+ * Copyright (c) 2007 Mellanox Technologies. All rights reserved.
+ *
+ * This software is available to you under a choice of one of two
+ * licenses.  You may choose to be licensed under the terms of the GNU
+ * General Public License (GPL) Version 2, available from the file
+ * COPYING in the main directory of this source tree, or the
+ * OpenIB.org BSD license below:
+ *
+ *     Redistribution and use in source and binary forms, with or
+ *     without modification, are permitted provided that the following
+ *     conditions are met:
+ *
+ *      - Redistributions of source code must retain the above
+ *        copyright notice, this list of conditions and the following
+ *        disclaimer.
+ *
+ *      - Redistributions in binary form must reproduce the above
+ *        copyright notice, this list of conditions and the following
+ *        disclaimer in the documentation and/or other materials
+ *        provided with the distribution.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
+ * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
+ * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS
+ * BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN
+ * ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
+ * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+ * SOFTWARE.
+ *
+ */
+
+#include <asm/page.h>
+#include <linux/mlx4/cq.h>
+#include <linux/slab.h>
+#include <linux/mlx4/qp.h>
+#include <linux/skbuff.h>
+#include <linux/if_vlan.h>
+#include <linux/prefetch.h>
+#include <linux/vmalloc.h>
+#include <linux/tcp.h>
+#include <linux/ip.h>
+#include <linux/ipv6.h>
+#include <linux/moduleparam.h>
+
+#include "mlx4_en.h"
+
+int mlx4_en_create_tx_ring(struct mlx4_en_priv *priv,
+			   struct mlx4_en_tx_ring **pring, u32 size,
+			   u16 stride, int node, int queue_index)
+{
+	struct mlx4_en_dev *mdev = priv->mdev;
+	struct mlx4_en_tx_ring *ring;
+	int tmp;
+	int err;
+
+	ring = kzalloc_node(sizeof(*ring), GFP_KERNEL, node);
+	if (!ring) {
+		ring = kzalloc(sizeof(*ring), GFP_KERNEL);
+		if (!ring) {
+			en_err(priv, "Failed allocating TX ring\n");
+			return -ENOMEM;
+		}
+	}
+
+	ring->size = size;
+	ring->size_mask = size - 1;
+	ring->sp_stride = stride;
+	ring->full_size = ring->size - HEADROOM - MAX_DESC_TXBBS;
+
+	tmp = size * sizeof(struct mlx4_en_tx_info);
+	ring->tx_info = kvmalloc_node(tmp, GFP_KERNEL, node);
+	if (!ring->tx_info) {
+		err = -ENOMEM;
+		goto err_ring;
+	}
+
+	en_dbg(DRV, priv, "Allocated tx_info ring at addr:%p size:%d\n",
+		 ring->tx_info, tmp);
+
+	ring->bounce_buf = kmalloc_node(MAX_DESC_SIZE, GFP_KERNEL, node);
+	if (!ring->bounce_buf) {
+		ring->bounce_buf = kmalloc(MAX_DESC_SIZE, GFP_KERNEL);
+		if (!ring->bounce_buf) {
+			err = -ENOMEM;
+			goto err_info;
+		}
+	}
+	ring->buf_size = ALIGN(size * ring->sp_stride, MLX4_EN_PAGE_SIZE);
+
+	/* Allocate HW buffers on provided NUMA node */
+	set_dev_node(&mdev->dev->persist->pdev->dev, node);
+	err = mlx4_alloc_hwq_res(mdev->dev, &ring->sp_wqres, ring->buf_size);
+	set_dev_node(&mdev->dev->persist->pdev->dev, mdev->dev->numa_node);
+	if (err) {
+		en_err(priv, "Failed allocating hwq resources\n");
+		goto err_bounce;
+	}
+
+	ring->buf = ring->sp_wqres.buf.direct.buf;
+
+	en_dbg(DRV, priv, "Allocated TX ring (addr:%p) - buf:%p size:%d buf_size:%d dma:%llx\n",
+	       ring, ring->buf, ring->size, ring->buf_size,
+	       (unsigned long long) ring->sp_wqres.buf.direct.map);
+
+	err = mlx4_qp_reserve_range(mdev->dev, 1, 1, &ring->qpn,
+				    MLX4_RESERVE_ETH_BF_QP,
+				    MLX4_RES_USAGE_DRIVER);
+	if (err) {
+		en_err(priv, "failed reserving qp for TX ring\n");
+		goto err_hwq_res;
+	}
+
+	err = mlx4_qp_alloc(mdev->dev, ring->qpn, &ring->sp_qp);
+	if (err) {
+		en_err(priv, "Failed allocating qp %d\n", ring->qpn);
+		goto err_reserve;
+	}
+	ring->sp_qp.event = mlx4_en_sqp_event;
+
+	err = mlx4_bf_alloc(mdev->dev, &ring->bf, node);
+	if (err) {
+		en_dbg(DRV, priv, "working without blueflame (%d)\n", err);
+		ring->bf.uar = &mdev->priv_uar;
+		ring->bf.uar->map = mdev->uar_map;
+		ring->bf_enabled = false;
+		ring->bf_alloced = false;
+		priv->pflags &= ~MLX4_EN_PRIV_FLAGS_BLUEFLAME;
+	} else {
+		ring->bf_alloced = true;
+		ring->bf_enabled = !!(priv->pflags &
+				      MLX4_EN_PRIV_FLAGS_BLUEFLAME);
+	}
+
+	ring->hwtstamp_tx_type = priv->hwtstamp_config.tx_type;
+	ring->queue_index = queue_index;
+
+	if (queue_index < priv->num_tx_rings_p_up)
+		cpumask_set_cpu(cpumask_local_spread(queue_index,
+						     priv->mdev->dev->numa_node),
+				&ring->sp_affinity_mask);
+
+	*pring = ring;
+	return 0;
+
+err_reserve:
+	mlx4_qp_release_range(mdev->dev, ring->qpn, 1);
+err_hwq_res:
+	mlx4_free_hwq_res(mdev->dev, &ring->sp_wqres, ring->buf_size);
+err_bounce:
+	kfree(ring->bounce_buf);
+	ring->bounce_buf = NULL;
+err_info:
+	kvfree(ring->tx_info);
+	ring->tx_info = NULL;
+err_ring:
+	kfree(ring);
+	*pring = NULL;
+	return err;
+}
+
+void mlx4_en_destroy_tx_ring(struct mlx4_en_priv *priv,
+			     struct mlx4_en_tx_ring **pring)
+{
+	struct mlx4_en_dev *mdev = priv->mdev;
+	struct mlx4_en_tx_ring *ring = *pring;
+	en_dbg(DRV, priv, "Destroying tx ring, qpn: %d\n", ring->qpn);
+
+	if (ring->bf_alloced)
+		mlx4_bf_free(mdev->dev, &ring->bf);
+	mlx4_qp_remove(mdev->dev, &ring->sp_qp);
+	mlx4_qp_free(mdev->dev, &ring->sp_qp);
+	mlx4_qp_release_range(priv->mdev->dev, ring->qpn, 1);
+	mlx4_free_hwq_res(mdev->dev, &ring->sp_wqres, ring->buf_size);
+	kfree(ring->bounce_buf);
+	ring->bounce_buf = NULL;
+	kvfree(ring->tx_info);
+	ring->tx_info = NULL;
+	kfree(ring);
+	*pring = NULL;
+}
+
+int mlx4_en_activate_tx_ring(struct mlx4_en_priv *priv,
+			     struct mlx4_en_tx_ring *ring,
+			     int cq, int user_prio)
+{
+	struct mlx4_en_dev *mdev = priv->mdev;
+	int err;
+
+	ring->sp_cqn = cq;
+	ring->prod = 0;
+	ring->cons = 0xffffffff;
+	ring->last_nr_txbb = 1;
+	memset(ring->tx_info, 0, ring->size * sizeof(struct mlx4_en_tx_info));
+	memset(ring->buf, 0, ring->buf_size);
+	ring->free_tx_desc = mlx4_en_free_tx_desc;
+
+	ring->sp_qp_state = MLX4_QP_STATE_RST;
+	ring->doorbell_qpn = cpu_to_be32(ring->sp_qp.qpn << 8);
+	ring->mr_key = cpu_to_be32(mdev->mr.key);
+
+	mlx4_en_fill_qp_context(priv, ring->size, ring->sp_stride, 1, 0, ring->qpn,
+				ring->sp_cqn, user_prio, &ring->sp_context);
+	if (ring->bf_alloced)
+		ring->sp_context.usr_page =
+			cpu_to_be32(mlx4_to_hw_uar_index(mdev->dev,
+							 ring->bf.uar->index));
+
+	err = mlx4_qp_to_ready(mdev->dev, &ring->sp_wqres.mtt, &ring->sp_context,
+			       &ring->sp_qp, &ring->sp_qp_state);
+	if (!cpumask_empty(&ring->sp_affinity_mask))
+		netif_set_xps_queue(priv->dev, &ring->sp_affinity_mask,
+				    ring->queue_index);
+
+	return err;
+}
+
+void mlx4_en_deactivate_tx_ring(struct mlx4_en_priv *priv,
+				struct mlx4_en_tx_ring *ring)
+{
+	struct mlx4_en_dev *mdev = priv->mdev;
+
+	mlx4_qp_modify(mdev->dev, NULL, ring->sp_qp_state,
+		       MLX4_QP_STATE_RST, NULL, 0, 0, &ring->sp_qp);
+}
+
+static inline bool mlx4_en_is_tx_ring_full(struct mlx4_en_tx_ring *ring)
+{
+	return ring->prod - ring->cons > ring->full_size;
+}
+
+static void mlx4_en_stamp_wqe(struct mlx4_en_priv *priv,
+			      struct mlx4_en_tx_ring *ring, int index,
+			      u8 owner)
+{
+	__be32 stamp = cpu_to_be32(STAMP_VAL | (!!owner << STAMP_SHIFT));
+	struct mlx4_en_tx_desc *tx_desc = ring->buf + (index << LOG_TXBB_SIZE);
+	struct mlx4_en_tx_info *tx_info = &ring->tx_info[index];
+	void *end = ring->buf + ring->buf_size;
+	__be32 *ptr = (__be32 *)tx_desc;
+	int i;
+
+	/* Optimize the common case when there are no wraparounds */
+	if (likely((void *)tx_desc +
+		   (tx_info->nr_txbb << LOG_TXBB_SIZE) <= end)) {
+		/* Stamp the freed descriptor */
+		for (i = 0; i < tx_info->nr_txbb << LOG_TXBB_SIZE;
+		     i += STAMP_STRIDE) {
+			*ptr = stamp;
+			ptr += STAMP_DWORDS;
+		}
+	} else {
+		/* Stamp the freed descriptor */
+		for (i = 0; i < tx_info->nr_txbb << LOG_TXBB_SIZE;
+		     i += STAMP_STRIDE) {
+			*ptr = stamp;
+			ptr += STAMP_DWORDS;
+			if ((void *)ptr >= end) {
+				ptr = ring->buf;
+				stamp ^= cpu_to_be32(0x80000000);
+			}
+		}
+	}
+}
+
+
+u32 mlx4_en_free_tx_desc(struct mlx4_en_priv *priv,
+			 struct mlx4_en_tx_ring *ring,
+			 int index, u64 timestamp,
+			 int napi_mode)
+{
+	struct mlx4_en_tx_info *tx_info = &ring->tx_info[index];
+	struct mlx4_en_tx_desc *tx_desc = ring->buf + (index << LOG_TXBB_SIZE);
+	struct mlx4_wqe_data_seg *data = (void *) tx_desc + tx_info->data_offset;
+	void *end = ring->buf + ring->buf_size;
+	struct sk_buff *skb = tx_info->skb;
+	int nr_maps = tx_info->nr_maps;
+	int i;
+
+	/* We do not touch skb here, so prefetch skb->users location
+	 * to speedup consume_skb()
+	 */
+	prefetchw(&skb->users);
+
+	if (unlikely(timestamp)) {
+		struct skb_shared_hwtstamps hwts;
+
+		mlx4_en_fill_hwtstamps(priv->mdev, &hwts, timestamp);
+		skb_tstamp_tx(skb, &hwts);
+	}
+
+	if (!tx_info->inl) {
+		if (tx_info->linear)
+			dma_unmap_single(priv->ddev,
+					 tx_info->map0_dma,
+					 tx_info->map0_byte_count,
+					 PCI_DMA_TODEVICE);
+		else
+			dma_unmap_page(priv->ddev,
+				       tx_info->map0_dma,
+				       tx_info->map0_byte_count,
+				       PCI_DMA_TODEVICE);
+		/* Optimize the common case when there are no wraparounds */
+		if (likely((void *)tx_desc +
+			   (tx_info->nr_txbb << LOG_TXBB_SIZE) <= end)) {
+			for (i = 1; i < nr_maps; i++) {
+				data++;
+				dma_unmap_page(priv->ddev,
+					(dma_addr_t)be64_to_cpu(data->addr),
+					be32_to_cpu(data->byte_count),
+					PCI_DMA_TODEVICE);
+			}
+		} else {
+			if ((void *)data >= end)
+				data = ring->buf + ((void *)data - end);
+
+			for (i = 1; i < nr_maps; i++) {
+				data++;
+				/* Check for wraparound before unmapping */
+				if ((void *) data >= end)
+					data = ring->buf;
+				dma_unmap_page(priv->ddev,
+					(dma_addr_t)be64_to_cpu(data->addr),
+					be32_to_cpu(data->byte_count),
+					PCI_DMA_TODEVICE);
+			}
+		}
+	}
+	napi_consume_skb(skb, napi_mode);
+
+	return tx_info->nr_txbb;
+}
+
+u32 mlx4_en_recycle_tx_desc(struct mlx4_en_priv *priv,
+			    struct mlx4_en_tx_ring *ring,
+			    int index, u64 timestamp,
+			    int napi_mode)
+{
+	struct mlx4_en_tx_info *tx_info = &ring->tx_info[index];
+	struct mlx4_en_rx_alloc frame = {
+		.page = tx_info->page,
+		.dma = tx_info->map0_dma,
+	};
+
+	if (!napi_mode || !mlx4_en_rx_recycle(ring->recycle_ring, &frame)) {
+		dma_unmap_page(priv->ddev, tx_info->map0_dma,
+			       PAGE_SIZE, priv->dma_dir);
+		put_page(tx_info->page);
+	}
+
+	return tx_info->nr_txbb;
+}
+
+int mlx4_en_free_tx_buf(struct net_device *dev, struct mlx4_en_tx_ring *ring)
+{
+	struct mlx4_en_priv *priv = netdev_priv(dev);
+	int cnt = 0;
+
+	/* Skip last polled descriptor */
+	ring->cons += ring->last_nr_txbb;
+	en_dbg(DRV, priv, "Freeing Tx buf - cons:0x%x prod:0x%x\n",
+		 ring->cons, ring->prod);
+
+	if ((u32) (ring->prod - ring->cons) > ring->size) {
+		if (netif_msg_tx_err(priv))
+			en_warn(priv, "Tx consumer passed producer!\n");
+		return 0;
+	}
+
+	while (ring->cons != ring->prod) {
+		ring->last_nr_txbb = ring->free_tx_desc(priv, ring,
+						ring->cons & ring->size_mask,
+						0, 0 /* Non-NAPI caller */);
+		ring->cons += ring->last_nr_txbb;
+		cnt++;
+	}
+
+	if (ring->tx_queue)
+		netdev_tx_reset_queue(ring->tx_queue);
+
+	if (cnt)
+		en_dbg(DRV, priv, "Freed %d uncompleted tx descriptors\n", cnt);
+
+	return cnt;
+}
+
+static void mlx4_en_handle_err_cqe(struct mlx4_en_priv *priv, struct mlx4_err_cqe *err_cqe,
+				   u16 cqe_index, struct mlx4_en_tx_ring *ring)
+{
+	struct mlx4_en_dev *mdev = priv->mdev;
+	struct mlx4_en_tx_info *tx_info;
+	struct mlx4_en_tx_desc *tx_desc;
+	u16 wqe_index;
+	int desc_size;
+
+	en_err(priv, "CQE error - cqn 0x%x, ci 0x%x, vendor syndrome: 0x%x syndrome: 0x%x\n",
+	       ring->sp_cqn, cqe_index, err_cqe->vendor_err_syndrome, err_cqe->syndrome);
+	print_hex_dump(KERN_WARNING, "", DUMP_PREFIX_OFFSET, 16, 1, err_cqe, sizeof(*err_cqe),
+		       false);
+
+	wqe_index = be16_to_cpu(err_cqe->wqe_index) & ring->size_mask;
+	tx_info = &ring->tx_info[wqe_index];
+	desc_size = tx_info->nr_txbb << LOG_TXBB_SIZE;
+	en_err(priv, "Related WQE - qpn 0x%x, wqe index 0x%x, wqe size 0x%x\n", ring->qpn,
+	       wqe_index, desc_size);
+	tx_desc = ring->buf + (wqe_index << LOG_TXBB_SIZE);
+	print_hex_dump(KERN_WARNING, "", DUMP_PREFIX_OFFSET, 16, 1, tx_desc, desc_size, false);
+
+	if (test_and_set_bit(MLX4_EN_STATE_FLAG_RESTARTING, &priv->state))
+		return;
+
+	en_err(priv, "Scheduling port restart\n");
+	queue_work(mdev->workqueue, &priv->restart_task);
+}
+
+bool mlx4_en_process_tx_cq(struct net_device *dev,
+			   struct mlx4_en_cq *cq, int napi_budget)
+{
+	struct mlx4_en_priv *priv = netdev_priv(dev);
+	struct mlx4_cq *mcq = &cq->mcq;
+	struct mlx4_en_tx_ring *ring = priv->tx_ring[cq->type][cq->ring];
+	struct mlx4_cqe *cqe;
+	u16 index, ring_index, stamp_index;
+	u32 txbbs_skipped = 0;
+	u32 txbbs_stamp = 0;
+	u32 cons_index = mcq->cons_index;
+	int size = cq->size;
+	u32 size_mask = ring->size_mask;
+	struct mlx4_cqe *buf = cq->buf;
+	u32 packets = 0;
+	u32 bytes = 0;
+	int factor = priv->cqe_factor;
+	int done = 0;
+	int budget = priv->tx_work_limit;
+	u32 last_nr_txbb;
+	u32 ring_cons;
+
+	if (unlikely(!priv->port_up))
+		return true;
+
+	netdev_txq_bql_complete_prefetchw(ring->tx_queue);
+
+	index = cons_index & size_mask;
+	cqe = mlx4_en_get_cqe(buf, index, priv->cqe_size) + factor;
+	last_nr_txbb = READ_ONCE(ring->last_nr_txbb);
+	ring_cons = READ_ONCE(ring->cons);
+	ring_index = ring_cons & size_mask;
+	stamp_index = ring_index;
+
+	/* Process all completed CQEs */
+	while (XNOR(cqe->owner_sr_opcode & MLX4_CQE_OWNER_MASK,
+			cons_index & size) && (done < budget)) {
+		u16 new_index;
+
+		/*
+		 * make sure we read the CQE after we read the
+		 * ownership bit
+		 */
+		dma_rmb();
+
+		if (unlikely((cqe->owner_sr_opcode & MLX4_CQE_OPCODE_MASK) ==
+			     MLX4_CQE_OPCODE_ERROR))
+			if (!test_and_set_bit(MLX4_EN_TX_RING_STATE_RECOVERING, &ring->state))
+				mlx4_en_handle_err_cqe(priv, (struct mlx4_err_cqe *)cqe, index,
+						       ring);
+
+		/* Skip over last polled CQE */
+		new_index = be16_to_cpu(cqe->wqe_index) & size_mask;
+
+		do {
+			u64 timestamp = 0;
+
+			txbbs_skipped += last_nr_txbb;
+			ring_index = (ring_index + last_nr_txbb) & size_mask;
+
+			if (unlikely(ring->tx_info[ring_index].ts_requested))
+				timestamp = mlx4_en_get_cqe_ts(cqe);
+
+			/* free next descriptor */
+			last_nr_txbb = ring->free_tx_desc(
+					priv, ring, ring_index,
+					timestamp, napi_budget);
+
+			mlx4_en_stamp_wqe(priv, ring, stamp_index,
+					  !!((ring_cons + txbbs_stamp) &
+						ring->size));
+			stamp_index = ring_index;
+			txbbs_stamp = txbbs_skipped;
+			packets++;
+			bytes += ring->tx_info[ring_index].nr_bytes;
+		} while ((++done < budget) && (ring_index != new_index));
+
+		++cons_index;
+		index = cons_index & size_mask;
+		cqe = mlx4_en_get_cqe(buf, index, priv->cqe_size) + factor;
+	}
+
+	/*
+	 * To prevent CQ overflow we first update CQ consumer and only then
+	 * the ring consumer.
+	 */
+	mcq->cons_index = cons_index;
+	mlx4_cq_set_ci(mcq);
+	wmb();
+
+	/* we want to dirty this cache line once */
+	WRITE_ONCE(ring->last_nr_txbb, last_nr_txbb);
+	WRITE_ONCE(ring->cons, ring_cons + txbbs_skipped);
+
+	if (cq->type == TX_XDP)
+		return done < budget;
+
+	netdev_tx_completed_queue(ring->tx_queue, packets, bytes);
+
+	/* Wakeup Tx queue if this stopped, and ring is not full.
+	 */
+	if (netif_tx_queue_stopped(ring->tx_queue) &&
+	    !mlx4_en_is_tx_ring_full(ring)) {
+		netif_tx_wake_queue(ring->tx_queue);
+		ring->wake_queue++;
+	}
+
+	return done < budget;
+}
+
+void mlx4_en_tx_irq(struct mlx4_cq *mcq)
+{
+	struct mlx4_en_cq *cq = container_of(mcq, struct mlx4_en_cq, mcq);
+	struct mlx4_en_priv *priv = netdev_priv(cq->dev);
+
+	if (likely(priv->port_up))
+		napi_schedule_irqoff(&cq->napi);
+	else
+		mlx4_en_arm_cq(priv, cq);
+}
+
+/* TX CQ polling - called by NAPI */
+int mlx4_en_poll_tx_cq(struct napi_struct *napi, int budget)
+{
+	struct mlx4_en_cq *cq = container_of(napi, struct mlx4_en_cq, napi);
+	struct net_device *dev = cq->dev;
+	struct mlx4_en_priv *priv = netdev_priv(dev);
+	bool clean_complete;
+
+	clean_complete = mlx4_en_process_tx_cq(dev, cq, budget);
+	if (!clean_complete)
+		return budget;
+
+	napi_complete(napi);
+	mlx4_en_arm_cq(priv, cq);
+
+	return 0;
+}
+
+static struct mlx4_en_tx_desc *mlx4_en_bounce_to_desc(struct mlx4_en_priv *priv,
+						      struct mlx4_en_tx_ring *ring,
+						      u32 index,
+						      unsigned int desc_size)
+{
+	u32 copy = (ring->size - index) << LOG_TXBB_SIZE;
+	int i;
+
+	for (i = desc_size - copy - 4; i >= 0; i -= 4) {
+		if ((i & (TXBB_SIZE - 1)) == 0)
+			wmb();
+
+		*((u32 *) (ring->buf + i)) =
+			*((u32 *) (ring->bounce_buf + copy + i));
+	}
+
+	for (i = copy - 4; i >= 4 ; i -= 4) {
+		if ((i & (TXBB_SIZE - 1)) == 0)
+			wmb();
+
+		*((u32 *)(ring->buf + (index << LOG_TXBB_SIZE) + i)) =
+			*((u32 *) (ring->bounce_buf + i));
+	}
+
+	/* Return real descriptor location */
+	return ring->buf + (index << LOG_TXBB_SIZE);
+}
+
+/* Decide if skb can be inlined in tx descriptor to avoid dma mapping
+ *
+ * It seems strange we do not simply use skb_copy_bits().
+ * This would allow to inline all skbs iff skb->len <= inline_thold
+ *
+ * Note that caller already checked skb was not a gso packet
+ */
+static bool is_inline(int inline_thold, const struct sk_buff *skb,
+		      const struct skb_shared_info *shinfo,
+		      void **pfrag)
+{
+	void *ptr;
+
+	if (skb->len > inline_thold || !inline_thold)
+		return false;
+
+	if (shinfo->nr_frags == 1) {
+		ptr = skb_frag_address_safe(&shinfo->frags[0]);
+		if (unlikely(!ptr))
+			return false;
+		*pfrag = ptr;
+		return true;
+	}
+	if (shinfo->nr_frags)
+		return false;
+	return true;
+}
+
+static int inline_size(const struct sk_buff *skb)
+{
+	if (skb->len + CTRL_SIZE + sizeof(struct mlx4_wqe_inline_seg)
+	    <= MLX4_INLINE_ALIGN)
+		return ALIGN(skb->len + CTRL_SIZE +
+			     sizeof(struct mlx4_wqe_inline_seg), 16);
+	else
+		return ALIGN(skb->len + CTRL_SIZE + 2 *
+			     sizeof(struct mlx4_wqe_inline_seg), 16);
+}
+
+static int get_real_size(const struct sk_buff *skb,
+			 const struct skb_shared_info *shinfo,
+			 struct net_device *dev,
+			 int *lso_header_size,
+			 bool *inline_ok,
+			 void **pfrag)
+{
+	struct mlx4_en_priv *priv = netdev_priv(dev);
+	int real_size;
+
+	if (shinfo->gso_size) {
+		*inline_ok = false;
+		if (skb->encapsulation)
+			*lso_header_size = (skb_inner_transport_header(skb) - skb->data) + inner_tcp_hdrlen(skb);
+		else
+			*lso_header_size = skb_transport_offset(skb) + tcp_hdrlen(skb);
+		real_size = CTRL_SIZE + shinfo->nr_frags * DS_SIZE +
+			ALIGN(*lso_header_size + 4, DS_SIZE);
+		if (unlikely(*lso_header_size != skb_headlen(skb))) {
+			/* We add a segment for the skb linear buffer only if
+			 * it contains data */
+			if (*lso_header_size < skb_headlen(skb))
+				real_size += DS_SIZE;
+			else {
+				if (netif_msg_tx_err(priv))
+					en_warn(priv, "Non-linear headers\n");
+				return 0;
+			}
+		}
+	} else {
+		*lso_header_size = 0;
+		*inline_ok = is_inline(priv->prof->inline_thold, skb,
+				       shinfo, pfrag);
+
+		if (*inline_ok)
+			real_size = inline_size(skb);
+		else
+			real_size = CTRL_SIZE +
+				    (shinfo->nr_frags + 1) * DS_SIZE;
+	}
+
+	return real_size;
+}
+
+static void build_inline_wqe(struct mlx4_en_tx_desc *tx_desc,
+			     const struct sk_buff *skb,
+			     const struct skb_shared_info *shinfo,
+			     void *fragptr)
+{
+	struct mlx4_wqe_inline_seg *inl = &tx_desc->inl;
+	int spc = MLX4_INLINE_ALIGN - CTRL_SIZE - sizeof(*inl);
+	unsigned int hlen = skb_headlen(skb);
+
+	if (skb->len <= spc) {
+		if (likely(skb->len >= MIN_PKT_LEN)) {
+			inl->byte_count = cpu_to_be32(1 << 31 | skb->len);
+		} else {
+			inl->byte_count = cpu_to_be32(1 << 31 | MIN_PKT_LEN);
+			memset(((void *)(inl + 1)) + skb->len, 0,
+			       MIN_PKT_LEN - skb->len);
+		}
+		skb_copy_from_linear_data(skb, inl + 1, hlen);
+		if (shinfo->nr_frags)
+			memcpy(((void *)(inl + 1)) + hlen, fragptr,
+			       skb_frag_size(&shinfo->frags[0]));
+
+	} else {
+		inl->byte_count = cpu_to_be32(1 << 31 | spc);
+		if (hlen <= spc) {
+			skb_copy_from_linear_data(skb, inl + 1, hlen);
+			if (hlen < spc) {
+				memcpy(((void *)(inl + 1)) + hlen,
+				       fragptr, spc - hlen);
+				fragptr +=  spc - hlen;
+			}
+			inl = (void *) (inl + 1) + spc;
+			memcpy(((void *)(inl + 1)), fragptr, skb->len - spc);
+		} else {
+			skb_copy_from_linear_data(skb, inl + 1, spc);
+			inl = (void *) (inl + 1) + spc;
+			skb_copy_from_linear_data_offset(skb, spc, inl + 1,
+							 hlen - spc);
+			if (shinfo->nr_frags)
+				memcpy(((void *)(inl + 1)) + hlen - spc,
+				       fragptr,
+				       skb_frag_size(&shinfo->frags[0]));
+		}
+
+		dma_wmb();
+		inl->byte_count = cpu_to_be32(1 << 31 | (skb->len - spc));
+	}
+}
+
+u16 mlx4_en_select_queue(struct net_device *dev, struct sk_buff *skb,
+			 struct net_device *sb_dev,
+			 select_queue_fallback_t fallback)
+{
+	struct mlx4_en_priv *priv = netdev_priv(dev);
+	u16 rings_p_up = priv->num_tx_rings_p_up;
+
+	if (netdev_get_num_tc(dev))
+		return fallback(dev, skb, NULL);
+
+	return fallback(dev, skb, NULL) % rings_p_up;
+}
+
+static void mlx4_bf_copy(void __iomem *dst, const void *src,
+			 unsigned int bytecnt)
+{
+	__iowrite64_copy(dst, src, bytecnt / 8);
+}
+
+void mlx4_en_xmit_doorbell(struct mlx4_en_tx_ring *ring)
+{
+	wmb();
+	/* Since there is no iowrite*_native() that writes the
+	 * value as is, without byteswapping - using the one
+	 * the doesn't do byteswapping in the relevant arch
+	 * endianness.
+	 */
+#if defined(__LITTLE_ENDIAN)
+	iowrite32(
+#else
+	iowrite32be(
+#endif
+		  (__force u32)ring->doorbell_qpn,
+		  ring->bf.uar->map + MLX4_SEND_DOORBELL);
+}
+
+static void mlx4_en_tx_write_desc(struct mlx4_en_tx_ring *ring,
+				  struct mlx4_en_tx_desc *tx_desc,
+				  union mlx4_wqe_qpn_vlan qpn_vlan,
+				  int desc_size, int bf_index,
+				  __be32 op_own, bool bf_ok,
+				  bool send_doorbell)
+{
+	tx_desc->ctrl.qpn_vlan = qpn_vlan;
+
+	if (bf_ok) {
+		op_own |= htonl((bf_index & 0xffff) << 8);
+		/* Ensure new descriptor hits memory
+		 * before setting ownership of this descriptor to HW
+		 */
+		dma_wmb();
+		tx_desc->ctrl.owner_opcode = op_own;
+
+		wmb();
+
+		mlx4_bf_copy(ring->bf.reg + ring->bf.offset, &tx_desc->ctrl,
+			     desc_size);
+
+		wmb();
+
+		ring->bf.offset ^= ring->bf.buf_size;
+	} else {
+		/* Ensure new descriptor hits memory
+		 * before setting ownership of this descriptor to HW
+		 */
+		dma_wmb();
+		tx_desc->ctrl.owner_opcode = op_own;
+		if (send_doorbell)
+			mlx4_en_xmit_doorbell(ring);
+		else
+			ring->xmit_more++;
+	}
+}
+
+static bool mlx4_en_build_dma_wqe(struct mlx4_en_priv *priv,
+				  struct skb_shared_info *shinfo,
+				  struct mlx4_wqe_data_seg *data,
+				  struct sk_buff *skb,
+				  int lso_header_size,
+				  __be32 mr_key,
+				  struct mlx4_en_tx_info *tx_info)
+{
+	struct device *ddev = priv->ddev;
+	dma_addr_t dma = 0;
+	u32 byte_count = 0;
+	int i_frag;
+
+	/* Map fragments if any */
+	for (i_frag = shinfo->nr_frags - 1; i_frag >= 0; i_frag--) {
+		const struct skb_frag_struct *frag;
+
+		frag = &shinfo->frags[i_frag];
+		byte_count = skb_frag_size(frag);
+		dma = skb_frag_dma_map(ddev, frag,
+				       0, byte_count,
+				       DMA_TO_DEVICE);
+		if (dma_mapping_error(ddev, dma))
+			goto tx_drop_unmap;
+
+		data->addr = cpu_to_be64(dma);
+		data->lkey = mr_key;
+		dma_wmb();
+		data->byte_count = cpu_to_be32(byte_count);
+		--data;
+	}
+
+	/* Map linear part if needed */
+	if (tx_info->linear) {
+		byte_count = skb_headlen(skb) - lso_header_size;
+
+		dma = dma_map_single(ddev, skb->data +
+				     lso_header_size, byte_count,
+				     PCI_DMA_TODEVICE);
+		if (dma_mapping_error(ddev, dma))
+			goto tx_drop_unmap;
+
+		data->addr = cpu_to_be64(dma);
+		data->lkey = mr_key;
+		dma_wmb();
+		data->byte_count = cpu_to_be32(byte_count);
+	}
+	/* tx completion can avoid cache line miss for common cases */
+	tx_info->map0_dma = dma;
+	tx_info->map0_byte_count = byte_count;
+
+	return true;
+
+tx_drop_unmap:
+	en_err(priv, "DMA mapping error\n");
+
+	while (++i_frag < shinfo->nr_frags) {
+		++data;
+		dma_unmap_page(ddev, (dma_addr_t)be64_to_cpu(data->addr),
+			       be32_to_cpu(data->byte_count),
+			       PCI_DMA_TODEVICE);
+	}
+
+	return false;
+}
+
+netdev_tx_t mlx4_en_xmit(struct sk_buff *skb, struct net_device *dev)
+{
+	struct skb_shared_info *shinfo = skb_shinfo(skb);
+	struct mlx4_en_priv *priv = netdev_priv(dev);
+	union mlx4_wqe_qpn_vlan	qpn_vlan = {};
+	struct mlx4_en_tx_ring *ring;
+	struct mlx4_en_tx_desc *tx_desc;
+	struct mlx4_wqe_data_seg *data;
+	struct mlx4_en_tx_info *tx_info;
+	u32 __maybe_unused ring_cons;
+	int tx_ind;
+	int nr_txbb;
+	int desc_size;
+	int real_size;
+	u32 index, bf_index;
+	__be32 op_own;
+	int lso_header_size;
+	void *fragptr = NULL;
+	bool bounce = false;
+	bool send_doorbell;
+	bool stop_queue;
+	bool inline_ok;
+	u8 data_offset;
+	bool bf_ok;
+
+	tx_ind = skb_get_queue_mapping(skb);
+	ring = priv->tx_ring[TX][tx_ind];
+
+	if (unlikely(!priv->port_up))
+		goto tx_drop;
+
+	/* fetch ring->cons far ahead before needing it to avoid stall */
+	ring_cons = READ_ONCE(ring->cons);
+
+	real_size = get_real_size(skb, shinfo, dev, &lso_header_size,
+				  &inline_ok, &fragptr);
+	if (unlikely(!real_size))
+		goto tx_drop_count;
+
+	/* Align descriptor to TXBB size */
+	desc_size = ALIGN(real_size, TXBB_SIZE);
+	nr_txbb = desc_size >> LOG_TXBB_SIZE;
+	if (unlikely(nr_txbb > MAX_DESC_TXBBS)) {
+		if (netif_msg_tx_err(priv))
+			en_warn(priv, "Oversized header or SG list\n");
+		goto tx_drop_count;
+	}
+
+	bf_ok = ring->bf_enabled;
+	if (skb_vlan_tag_present(skb)) {
+		u16 vlan_proto;
+
+		qpn_vlan.vlan_tag = cpu_to_be16(skb_vlan_tag_get(skb));
+		vlan_proto = be16_to_cpu(skb->vlan_proto);
+		if (vlan_proto == ETH_P_8021AD)
+			qpn_vlan.ins_vlan = MLX4_WQE_CTRL_INS_SVLAN;
+		else if (vlan_proto == ETH_P_8021Q)
+			qpn_vlan.ins_vlan = MLX4_WQE_CTRL_INS_CVLAN;
+		else
+			qpn_vlan.ins_vlan = 0;
+		bf_ok = false;
+	}
+
+	netdev_txq_bql_enqueue_prefetchw(ring->tx_queue);
+
+	/* Track current inflight packets for performance analysis */
+	AVG_PERF_COUNTER(priv->pstats.inflight_avg,
+			 (u32)(ring->prod - ring_cons - 1));
+
+	/* Packet is good - grab an index and transmit it */
+	index = ring->prod & ring->size_mask;
+	bf_index = ring->prod;
+
+	/* See if we have enough space for whole descriptor TXBB for setting
+	 * SW ownership on next descriptor; if not, use a bounce buffer. */
+	if (likely(index + nr_txbb <= ring->size))
+		tx_desc = ring->buf + (index << LOG_TXBB_SIZE);
+	else {
+		tx_desc = (struct mlx4_en_tx_desc *) ring->bounce_buf;
+		bounce = true;
+		bf_ok = false;
+	}
+
+	/* Save skb in tx_info ring */
+	tx_info = &ring->tx_info[index];
+	tx_info->skb = skb;
+	tx_info->nr_txbb = nr_txbb;
+
+	if (!lso_header_size) {
+		data = &tx_desc->data;
+		data_offset = offsetof(struct mlx4_en_tx_desc, data);
+	} else {
+		int lso_align = ALIGN(lso_header_size + 4, DS_SIZE);
+
+		data = (void *)&tx_desc->lso + lso_align;
+		data_offset = offsetof(struct mlx4_en_tx_desc, lso) + lso_align;
+	}
+
+	/* valid only for none inline segments */
+	tx_info->data_offset = data_offset;
+
+	tx_info->inl = inline_ok;
+
+	tx_info->linear = lso_header_size < skb_headlen(skb) && !inline_ok;
+
+	tx_info->nr_maps = shinfo->nr_frags + tx_info->linear;
+	data += tx_info->nr_maps - 1;
+
+	if (!tx_info->inl)
+		if (!mlx4_en_build_dma_wqe(priv, shinfo, data, skb,
+					   lso_header_size, ring->mr_key,
+					   tx_info))
+			goto tx_drop_count;
+
+	/*
+	 * For timestamping add flag to skb_shinfo and
+	 * set flag for further reference
+	 */
+	tx_info->ts_requested = 0;
+	if (unlikely(ring->hwtstamp_tx_type == HWTSTAMP_TX_ON &&
+		     shinfo->tx_flags & SKBTX_HW_TSTAMP)) {
+		shinfo->tx_flags |= SKBTX_IN_PROGRESS;
+		tx_info->ts_requested = 1;
+	}
+
+	/* Prepare ctrl segement apart opcode+ownership, which depends on
+	 * whether LSO is used */
+	tx_desc->ctrl.srcrb_flags = priv->ctrl_flags;
+	if (likely(skb->ip_summed == CHECKSUM_PARTIAL)) {
+		if (!skb->encapsulation)
+			tx_desc->ctrl.srcrb_flags |= cpu_to_be32(MLX4_WQE_CTRL_IP_CSUM |
+								 MLX4_WQE_CTRL_TCP_UDP_CSUM);
+		else
+			tx_desc->ctrl.srcrb_flags |= cpu_to_be32(MLX4_WQE_CTRL_IP_CSUM);
+		ring->tx_csum++;
+	}
+
+	if (priv->flags & MLX4_EN_FLAG_ENABLE_HW_LOOPBACK) {
+		struct ethhdr *ethh;
+
+		/* Copy dst mac address to wqe. This allows loopback in eSwitch,
+		 * so that VFs and PF can communicate with each other
+		 */
+		ethh = (struct ethhdr *)skb->data;
+		tx_desc->ctrl.srcrb_flags16[0] = get_unaligned((__be16 *)ethh->h_dest);
+		tx_desc->ctrl.imm = get_unaligned((__be32 *)(ethh->h_dest + 2));
+	}
+
+	/* Handle LSO (TSO) packets */
+	if (lso_header_size) {
+		int i;
+
+		/* Mark opcode as LSO */
+		op_own = cpu_to_be32(MLX4_OPCODE_LSO | (1 << 6)) |
+			((ring->prod & ring->size) ?
+				cpu_to_be32(MLX4_EN_BIT_DESC_OWN) : 0);
+
+		/* Fill in the LSO prefix */
+		tx_desc->lso.mss_hdr_size = cpu_to_be32(
+			shinfo->gso_size << 16 | lso_header_size);
+
+		/* Copy headers;
+		 * note that we already verified that it is linear */
+		memcpy(tx_desc->lso.header, skb->data, lso_header_size);
+
+		ring->tso_packets++;
+
+		i = shinfo->gso_segs;
+		tx_info->nr_bytes = skb->len + (i - 1) * lso_header_size;
+		ring->packets += i;
+	} else {
+		/* Normal (Non LSO) packet */
+		op_own = cpu_to_be32(MLX4_OPCODE_SEND) |
+			((ring->prod & ring->size) ?
+			 cpu_to_be32(MLX4_EN_BIT_DESC_OWN) : 0);
+		tx_info->nr_bytes = max_t(unsigned int, skb->len, ETH_ZLEN);
+		ring->packets++;
+	}
+	ring->bytes += tx_info->nr_bytes;
+	netdev_tx_sent_queue(ring->tx_queue, tx_info->nr_bytes);
+	AVG_PERF_COUNTER(priv->pstats.tx_pktsz_avg, skb->len);
+
+	if (tx_info->inl)
+		build_inline_wqe(tx_desc, skb, shinfo, fragptr);
+
+	if (skb->encapsulation) {
+		union {
+			struct iphdr *v4;
+			struct ipv6hdr *v6;
+			unsigned char *hdr;
+		} ip;
+		u8 proto;
+
+		ip.hdr = skb_inner_network_header(skb);
+		proto = (ip.v4->version == 4) ? ip.v4->protocol :
+						ip.v6->nexthdr;
+
+		if (proto == IPPROTO_TCP || proto == IPPROTO_UDP)
+			op_own |= cpu_to_be32(MLX4_WQE_CTRL_IIP | MLX4_WQE_CTRL_ILP);
+		else
+			op_own |= cpu_to_be32(MLX4_WQE_CTRL_IIP);
+	}
+
+	ring->prod += nr_txbb;
+
+	/* If we used a bounce buffer then copy descriptor back into place */
+	if (unlikely(bounce))
+		tx_desc = mlx4_en_bounce_to_desc(priv, ring, index, desc_size);
+
+	skb_tx_timestamp(skb);
+
+	/* Check available TXBBs And 2K spare for prefetch */
+	stop_queue = mlx4_en_is_tx_ring_full(ring);
+	if (unlikely(stop_queue)) {
+		netif_tx_stop_queue(ring->tx_queue);
+		ring->queue_stopped++;
+	}
+	send_doorbell = !skb->xmit_more || netif_xmit_stopped(ring->tx_queue);
+
+	real_size = (real_size / 16) & 0x3f;
+
+	bf_ok &= desc_size <= MAX_BF && send_doorbell;
+
+	if (bf_ok)
+		qpn_vlan.bf_qpn = ring->doorbell_qpn | cpu_to_be32(real_size);
+	else
+		qpn_vlan.fence_size = real_size;
+
+	mlx4_en_tx_write_desc(ring, tx_desc, qpn_vlan, desc_size, bf_index,
+			      op_own, bf_ok, send_doorbell);
+
+	if (unlikely(stop_queue)) {
+		/* If queue was emptied after the if (stop_queue) , and before
+		 * the netif_tx_stop_queue() - need to wake the queue,
+		 * or else it will remain stopped forever.
+		 * Need a memory barrier to make sure ring->cons was not
+		 * updated before queue was stopped.
+		 */
+		smp_rmb();
+
+		ring_cons = READ_ONCE(ring->cons);
+		if (unlikely(!mlx4_en_is_tx_ring_full(ring))) {
+			netif_tx_wake_queue(ring->tx_queue);
+			ring->wake_queue++;
+		}
+	}
+	return NETDEV_TX_OK;
+
+tx_drop_count:
+	ring->tx_dropped++;
+tx_drop:
+	dev_kfree_skb_any(skb);
+	return NETDEV_TX_OK;
+}
+
+#define MLX4_EN_XDP_TX_NRTXBB  1
+#define MLX4_EN_XDP_TX_REAL_SZ (((CTRL_SIZE + MLX4_EN_XDP_TX_NRTXBB * DS_SIZE) \
+				 / 16) & 0x3f)
+
+void mlx4_en_init_tx_xdp_ring_descs(struct mlx4_en_priv *priv,
+				    struct mlx4_en_tx_ring *ring)
+{
+	int i;
+
+	for (i = 0; i < ring->size; i++) {
+		struct mlx4_en_tx_info *tx_info = &ring->tx_info[i];
+		struct mlx4_en_tx_desc *tx_desc = ring->buf +
+			(i << LOG_TXBB_SIZE);
+
+		tx_info->map0_byte_count = PAGE_SIZE;
+		tx_info->nr_txbb = MLX4_EN_XDP_TX_NRTXBB;
+		tx_info->data_offset = offsetof(struct mlx4_en_tx_desc, data);
+		tx_info->ts_requested = 0;
+		tx_info->nr_maps = 1;
+		tx_info->linear = 1;
+		tx_info->inl = 0;
+
+		tx_desc->data.lkey = ring->mr_key;
+		tx_desc->ctrl.qpn_vlan.fence_size = MLX4_EN_XDP_TX_REAL_SZ;
+		tx_desc->ctrl.srcrb_flags = priv->ctrl_flags;
+	}
+}
+
+netdev_tx_t mlx4_en_xmit_frame(struct mlx4_en_rx_ring *rx_ring,
+			       struct mlx4_en_rx_alloc *frame,
+			       struct mlx4_en_priv *priv, unsigned int length,
+			       int tx_ind, bool *doorbell_pending)
+{
+	struct mlx4_en_tx_desc *tx_desc;
+	struct mlx4_en_tx_info *tx_info;
+	struct mlx4_wqe_data_seg *data;
+	struct mlx4_en_tx_ring *ring;
+	dma_addr_t dma;
+	__be32 op_own;
+	int index;
+
+	if (unlikely(!priv->port_up))
+		goto tx_drop;
+
+	ring = priv->tx_ring[TX_XDP][tx_ind];
+
+	if (unlikely(mlx4_en_is_tx_ring_full(ring)))
+		goto tx_drop_count;
+
+	index = ring->prod & ring->size_mask;
+	tx_info = &ring->tx_info[index];
+
+	/* Track current inflight packets for performance analysis */
+	AVG_PERF_COUNTER(priv->pstats.inflight_avg,
+			 (u32)(ring->prod - READ_ONCE(ring->cons) - 1));
+
+	tx_desc = ring->buf + (index << LOG_TXBB_SIZE);
+	data = &tx_desc->data;
+
+	dma = frame->dma;
+
+	tx_info->page = frame->page;
+	frame->page = NULL;
+	tx_info->map0_dma = dma;
+	tx_info->nr_bytes = max_t(unsigned int, length, ETH_ZLEN);
+
+	dma_sync_single_range_for_device(priv->ddev, dma, frame->page_offset,
+					 length, PCI_DMA_TODEVICE);
+
+	data->addr = cpu_to_be64(dma + frame->page_offset);
+	dma_wmb();
+	data->byte_count = cpu_to_be32(length);
+
+	/* tx completion can avoid cache line miss for common cases */
+
+	op_own = cpu_to_be32(MLX4_OPCODE_SEND) |
+		((ring->prod & ring->size) ?
+		 cpu_to_be32(MLX4_EN_BIT_DESC_OWN) : 0);
+
+	rx_ring->xdp_tx++;
+	AVG_PERF_COUNTER(priv->pstats.tx_pktsz_avg, length);
+
+	ring->prod += MLX4_EN_XDP_TX_NRTXBB;
+
+	/* Ensure new descriptor hits memory
+	 * before setting ownership of this descriptor to HW
+	 */
+	dma_wmb();
+	tx_desc->ctrl.owner_opcode = op_own;
+	ring->xmit_more++;
+
+	*doorbell_pending = true;
+
+	return NETDEV_TX_OK;
+
+tx_drop_count:
+	rx_ring->xdp_tx_full++;
+	*doorbell_pending = true;
+tx_drop:
+	return NETDEV_TX_BUSY;
+}
diff --git a/drivers/net/ethernet/mellanox/mlx4/eq.c b/drivers/net/ethernet/mellanox/mlx4/eq.c
new file mode 100644
index 000000000..2df92dbd3
--- /dev/null
+++ b/drivers/net/ethernet/mellanox/mlx4/eq.c
@@ -0,0 +1,1567 @@
+/*
+ * Copyright (c) 2005, 2006, 2007, 2008 Mellanox Technologies. All rights reserved.
+ * Copyright (c) 2005, 2006, 2007 Cisco Systems, Inc. All rights reserved.
+ *
+ * This software is available to you under a choice of one of two
+ * licenses.  You may choose to be licensed under the terms of the GNU
+ * General Public License (GPL) Version 2, available from the file
+ * COPYING in the main directory of this source tree, or the
+ * OpenIB.org BSD license below:
+ *
+ *     Redistribution and use in source and binary forms, with or
+ *     without modification, are permitted provided that the following
+ *     conditions are met:
+ *
+ *	- Redistributions of source code must retain the above
+ *	  copyright notice, this list of conditions and the following
+ *	  disclaimer.
+ *
+ *	- Redistributions in binary form must reproduce the above
+ *	  copyright notice, this list of conditions and the following
+ *	  disclaimer in the documentation and/or other materials
+ *	  provided with the distribution.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
+ * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
+ * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS
+ * BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN
+ * ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
+ * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+ * SOFTWARE.
+ */
+
+#include <linux/interrupt.h>
+#include <linux/slab.h>
+#include <linux/export.h>
+#include <linux/mm.h>
+#include <linux/dma-mapping.h>
+
+#include <linux/mlx4/cmd.h>
+#include <linux/cpu_rmap.h>
+
+#include "mlx4.h"
+#include "fw.h"
+
+enum {
+	MLX4_IRQNAME_SIZE	= 32
+};
+
+enum {
+	MLX4_NUM_ASYNC_EQE	= 0x100,
+	MLX4_NUM_SPARE_EQE	= 0x80,
+	MLX4_EQ_ENTRY_SIZE	= 0x20
+};
+
+#define MLX4_EQ_STATUS_OK	   ( 0 << 28)
+#define MLX4_EQ_STATUS_WRITE_FAIL  (10 << 28)
+#define MLX4_EQ_OWNER_SW	   ( 0 << 24)
+#define MLX4_EQ_OWNER_HW	   ( 1 << 24)
+#define MLX4_EQ_FLAG_EC		   ( 1 << 18)
+#define MLX4_EQ_FLAG_OI		   ( 1 << 17)
+#define MLX4_EQ_STATE_ARMED	   ( 9 <<  8)
+#define MLX4_EQ_STATE_FIRED	   (10 <<  8)
+#define MLX4_EQ_STATE_ALWAYS_ARMED (11 <<  8)
+
+#define MLX4_ASYNC_EVENT_MASK ((1ull << MLX4_EVENT_TYPE_PATH_MIG)	    | \
+			       (1ull << MLX4_EVENT_TYPE_COMM_EST)	    | \
+			       (1ull << MLX4_EVENT_TYPE_SQ_DRAINED)	    | \
+			       (1ull << MLX4_EVENT_TYPE_CQ_ERROR)	    | \
+			       (1ull << MLX4_EVENT_TYPE_WQ_CATAS_ERROR)	    | \
+			       (1ull << MLX4_EVENT_TYPE_EEC_CATAS_ERROR)    | \
+			       (1ull << MLX4_EVENT_TYPE_PATH_MIG_FAILED)    | \
+			       (1ull << MLX4_EVENT_TYPE_WQ_INVAL_REQ_ERROR) | \
+			       (1ull << MLX4_EVENT_TYPE_WQ_ACCESS_ERROR)    | \
+			       (1ull << MLX4_EVENT_TYPE_PORT_CHANGE)	    | \
+			       (1ull << MLX4_EVENT_TYPE_ECC_DETECT)	    | \
+			       (1ull << MLX4_EVENT_TYPE_SRQ_CATAS_ERROR)    | \
+			       (1ull << MLX4_EVENT_TYPE_SRQ_QP_LAST_WQE)    | \
+			       (1ull << MLX4_EVENT_TYPE_SRQ_LIMIT)	    | \
+			       (1ull << MLX4_EVENT_TYPE_CMD)		    | \
+			       (1ull << MLX4_EVENT_TYPE_OP_REQUIRED)	    | \
+			       (1ull << MLX4_EVENT_TYPE_COMM_CHANNEL)       | \
+			       (1ull << MLX4_EVENT_TYPE_FLR_EVENT)	    | \
+			       (1ull << MLX4_EVENT_TYPE_FATAL_WARNING))
+
+static u64 get_async_ev_mask(struct mlx4_dev *dev)
+{
+	u64 async_ev_mask = MLX4_ASYNC_EVENT_MASK;
+	if (dev->caps.flags & MLX4_DEV_CAP_FLAG_PORT_MNG_CHG_EV)
+		async_ev_mask |= (1ull << MLX4_EVENT_TYPE_PORT_MNG_CHG_EVENT);
+	if (dev->caps.flags2 & MLX4_DEV_CAP_FLAG2_RECOVERABLE_ERROR_EVENT)
+		async_ev_mask |= (1ull << MLX4_EVENT_TYPE_RECOVERABLE_ERROR_EVENT);
+
+	return async_ev_mask;
+}
+
+static void eq_set_ci(struct mlx4_eq *eq, int req_not)
+{
+	__raw_writel((__force u32) cpu_to_be32((eq->cons_index & 0xffffff) |
+					       req_not << 31),
+		     eq->doorbell);
+	/* We still want ordering, just not swabbing, so add a barrier */
+	mb();
+}
+
+static struct mlx4_eqe *get_eqe(struct mlx4_eq *eq, u32 entry, u8 eqe_factor,
+				u8 eqe_size)
+{
+	/* (entry & (eq->nent - 1)) gives us a cyclic array */
+	unsigned long offset = (entry & (eq->nent - 1)) * eqe_size;
+	/* CX3 is capable of extending the EQE from 32 to 64 bytes with
+	 * strides of 64B,128B and 256B.
+	 * When 64B EQE is used, the first (in the lower addresses)
+	 * 32 bytes in the 64 byte EQE are reserved and the next 32 bytes
+	 * contain the legacy EQE information.
+	 * In all other cases, the first 32B contains the legacy EQE info.
+	 */
+	return eq->page_list[offset / PAGE_SIZE].buf + (offset + (eqe_factor ? MLX4_EQ_ENTRY_SIZE : 0)) % PAGE_SIZE;
+}
+
+static struct mlx4_eqe *next_eqe_sw(struct mlx4_eq *eq, u8 eqe_factor, u8 size)
+{
+	struct mlx4_eqe *eqe = get_eqe(eq, eq->cons_index, eqe_factor, size);
+	return !!(eqe->owner & 0x80) ^ !!(eq->cons_index & eq->nent) ? NULL : eqe;
+}
+
+static struct mlx4_eqe *next_slave_event_eqe(struct mlx4_slave_event_eq *slave_eq)
+{
+	struct mlx4_eqe *eqe =
+		&slave_eq->event_eqe[slave_eq->cons & (SLAVE_EVENT_EQ_SIZE - 1)];
+	return (!!(eqe->owner & 0x80) ^
+		!!(slave_eq->cons & SLAVE_EVENT_EQ_SIZE)) ?
+		eqe : NULL;
+}
+
+void mlx4_gen_slave_eqe(struct work_struct *work)
+{
+	struct mlx4_mfunc_master_ctx *master =
+		container_of(work, struct mlx4_mfunc_master_ctx,
+			     slave_event_work);
+	struct mlx4_mfunc *mfunc =
+		container_of(master, struct mlx4_mfunc, master);
+	struct mlx4_priv *priv = container_of(mfunc, struct mlx4_priv, mfunc);
+	struct mlx4_dev *dev = &priv->dev;
+	struct mlx4_slave_event_eq *slave_eq = &mfunc->master.slave_eq;
+	struct mlx4_eqe *eqe;
+	u8 slave;
+	int i, phys_port, slave_port;
+
+	for (eqe = next_slave_event_eqe(slave_eq); eqe;
+	      eqe = next_slave_event_eqe(slave_eq)) {
+		slave = eqe->slave_id;
+
+		if (eqe->type == MLX4_EVENT_TYPE_PORT_CHANGE &&
+		    eqe->subtype == MLX4_PORT_CHANGE_SUBTYPE_DOWN &&
+		    mlx4_is_bonded(dev)) {
+			struct mlx4_port_cap port_cap;
+
+			if (!mlx4_QUERY_PORT(dev, 1, &port_cap) && port_cap.link_state)
+				goto consume;
+
+			if (!mlx4_QUERY_PORT(dev, 2, &port_cap) && port_cap.link_state)
+				goto consume;
+		}
+		/* All active slaves need to receive the event */
+		if (slave == ALL_SLAVES) {
+			for (i = 0; i <= dev->persist->num_vfs; i++) {
+				phys_port = 0;
+				if (eqe->type == MLX4_EVENT_TYPE_PORT_MNG_CHG_EVENT &&
+				    eqe->subtype == MLX4_DEV_PMC_SUBTYPE_PORT_INFO) {
+					phys_port  = eqe->event.port_mgmt_change.port;
+					slave_port = mlx4_phys_to_slave_port(dev, i, phys_port);
+					if (slave_port < 0) /* VF doesn't have this port */
+						continue;
+					eqe->event.port_mgmt_change.port = slave_port;
+				}
+				if (mlx4_GEN_EQE(dev, i, eqe))
+					mlx4_warn(dev, "Failed to generate event for slave %d\n",
+						  i);
+				if (phys_port)
+					eqe->event.port_mgmt_change.port = phys_port;
+			}
+		} else {
+			if (mlx4_GEN_EQE(dev, slave, eqe))
+				mlx4_warn(dev, "Failed to generate event for slave %d\n",
+					  slave);
+		}
+consume:
+		++slave_eq->cons;
+	}
+}
+
+
+static void slave_event(struct mlx4_dev *dev, u8 slave, struct mlx4_eqe *eqe)
+{
+	struct mlx4_priv *priv = mlx4_priv(dev);
+	struct mlx4_slave_event_eq *slave_eq = &priv->mfunc.master.slave_eq;
+	struct mlx4_eqe *s_eqe;
+	unsigned long flags;
+
+	spin_lock_irqsave(&slave_eq->event_lock, flags);
+	s_eqe = &slave_eq->event_eqe[slave_eq->prod & (SLAVE_EVENT_EQ_SIZE - 1)];
+	if ((!!(s_eqe->owner & 0x80)) ^
+	    (!!(slave_eq->prod & SLAVE_EVENT_EQ_SIZE))) {
+		mlx4_warn(dev, "Master failed to generate an EQE for slave: %d. No free EQE on slave events queue\n",
+			  slave);
+		spin_unlock_irqrestore(&slave_eq->event_lock, flags);
+		return;
+	}
+
+	memcpy(s_eqe, eqe, sizeof(struct mlx4_eqe) - 1);
+	s_eqe->slave_id = slave;
+	/* ensure all information is written before setting the ownersip bit */
+	dma_wmb();
+	s_eqe->owner = !!(slave_eq->prod & SLAVE_EVENT_EQ_SIZE) ? 0x0 : 0x80;
+	++slave_eq->prod;
+
+	queue_work(priv->mfunc.master.comm_wq,
+		   &priv->mfunc.master.slave_event_work);
+	spin_unlock_irqrestore(&slave_eq->event_lock, flags);
+}
+
+static void mlx4_slave_event(struct mlx4_dev *dev, int slave,
+			     struct mlx4_eqe *eqe)
+{
+	struct mlx4_priv *priv = mlx4_priv(dev);
+
+	if (slave < 0 || slave > dev->persist->num_vfs ||
+	    slave == dev->caps.function ||
+	    !priv->mfunc.master.slave_state[slave].active)
+		return;
+
+	slave_event(dev, slave, eqe);
+}
+
+#if defined(CONFIG_SMP)
+static void mlx4_set_eq_affinity_hint(struct mlx4_priv *priv, int vec)
+{
+	int hint_err;
+	struct mlx4_dev *dev = &priv->dev;
+	struct mlx4_eq *eq = &priv->eq_table.eq[vec];
+
+	if (!cpumask_available(eq->affinity_mask) ||
+	    cpumask_empty(eq->affinity_mask))
+		return;
+
+	hint_err = irq_set_affinity_hint(eq->irq, eq->affinity_mask);
+	if (hint_err)
+		mlx4_warn(dev, "irq_set_affinity_hint failed, err %d\n", hint_err);
+}
+#endif
+
+int mlx4_gen_pkey_eqe(struct mlx4_dev *dev, int slave, u8 port)
+{
+	struct mlx4_eqe eqe;
+
+	struct mlx4_priv *priv = mlx4_priv(dev);
+	struct mlx4_slave_state *s_slave = &priv->mfunc.master.slave_state[slave];
+
+	if (!s_slave->active)
+		return 0;
+
+	memset(&eqe, 0, sizeof(eqe));
+
+	eqe.type = MLX4_EVENT_TYPE_PORT_MNG_CHG_EVENT;
+	eqe.subtype = MLX4_DEV_PMC_SUBTYPE_PKEY_TABLE;
+	eqe.event.port_mgmt_change.port = mlx4_phys_to_slave_port(dev, slave, port);
+
+	return mlx4_GEN_EQE(dev, slave, &eqe);
+}
+EXPORT_SYMBOL(mlx4_gen_pkey_eqe);
+
+int mlx4_gen_guid_change_eqe(struct mlx4_dev *dev, int slave, u8 port)
+{
+	struct mlx4_eqe eqe;
+
+	/*don't send if we don't have the that slave */
+	if (dev->persist->num_vfs < slave)
+		return 0;
+	memset(&eqe, 0, sizeof(eqe));
+
+	eqe.type = MLX4_EVENT_TYPE_PORT_MNG_CHG_EVENT;
+	eqe.subtype = MLX4_DEV_PMC_SUBTYPE_GUID_INFO;
+	eqe.event.port_mgmt_change.port = mlx4_phys_to_slave_port(dev, slave, port);
+
+	return mlx4_GEN_EQE(dev, slave, &eqe);
+}
+EXPORT_SYMBOL(mlx4_gen_guid_change_eqe);
+
+int mlx4_gen_port_state_change_eqe(struct mlx4_dev *dev, int slave, u8 port,
+				   u8 port_subtype_change)
+{
+	struct mlx4_eqe eqe;
+	u8 slave_port = mlx4_phys_to_slave_port(dev, slave, port);
+
+	/*don't send if we don't have the that slave */
+	if (dev->persist->num_vfs < slave)
+		return 0;
+	memset(&eqe, 0, sizeof(eqe));
+
+	eqe.type = MLX4_EVENT_TYPE_PORT_CHANGE;
+	eqe.subtype = port_subtype_change;
+	eqe.event.port_change.port = cpu_to_be32(slave_port << 28);
+
+	mlx4_dbg(dev, "%s: sending: %d to slave: %d on port: %d\n", __func__,
+		 port_subtype_change, slave, port);
+	return mlx4_GEN_EQE(dev, slave, &eqe);
+}
+EXPORT_SYMBOL(mlx4_gen_port_state_change_eqe);
+
+enum slave_port_state mlx4_get_slave_port_state(struct mlx4_dev *dev, int slave, u8 port)
+{
+	struct mlx4_priv *priv = mlx4_priv(dev);
+	struct mlx4_slave_state *s_state = priv->mfunc.master.slave_state;
+	struct mlx4_active_ports actv_ports = mlx4_get_active_ports(dev, slave);
+
+	if (slave >= dev->num_slaves || port > dev->caps.num_ports ||
+	    port <= 0 || !test_bit(port - 1, actv_ports.ports)) {
+		pr_err("%s: Error: asking for slave:%d, port:%d\n",
+		       __func__, slave, port);
+		return SLAVE_PORT_DOWN;
+	}
+	return s_state[slave].port_state[port];
+}
+EXPORT_SYMBOL(mlx4_get_slave_port_state);
+
+static int mlx4_set_slave_port_state(struct mlx4_dev *dev, int slave, u8 port,
+				     enum slave_port_state state)
+{
+	struct mlx4_priv *priv = mlx4_priv(dev);
+	struct mlx4_slave_state *s_state = priv->mfunc.master.slave_state;
+	struct mlx4_active_ports actv_ports = mlx4_get_active_ports(dev, slave);
+
+	if (slave >= dev->num_slaves || port > dev->caps.num_ports ||
+	    port <= 0 || !test_bit(port - 1, actv_ports.ports)) {
+		pr_err("%s: Error: asking for slave:%d, port:%d\n",
+		       __func__, slave, port);
+		return -1;
+	}
+	s_state[slave].port_state[port] = state;
+
+	return 0;
+}
+
+static void set_all_slave_state(struct mlx4_dev *dev, u8 port, int event)
+{
+	int i;
+	enum slave_port_gen_event gen_event;
+	struct mlx4_slaves_pport slaves_pport = mlx4_phys_to_slaves_pport(dev,
+									  port);
+
+	for (i = 0; i < dev->persist->num_vfs + 1; i++)
+		if (test_bit(i, slaves_pport.slaves))
+			set_and_calc_slave_port_state(dev, i, port,
+						      event, &gen_event);
+}
+/**************************************************************************
+	The function get as input the new event to that port,
+	and according to the prev state change the slave's port state.
+	The events are:
+		MLX4_PORT_STATE_DEV_EVENT_PORT_DOWN,
+		MLX4_PORT_STATE_DEV_EVENT_PORT_UP
+		MLX4_PORT_STATE_IB_EVENT_GID_VALID
+		MLX4_PORT_STATE_IB_EVENT_GID_INVALID
+***************************************************************************/
+int set_and_calc_slave_port_state(struct mlx4_dev *dev, int slave,
+				  u8 port, int event,
+				  enum slave_port_gen_event *gen_event)
+{
+	struct mlx4_priv *priv = mlx4_priv(dev);
+	struct mlx4_slave_state *ctx = NULL;
+	unsigned long flags;
+	int ret = -1;
+	struct mlx4_active_ports actv_ports = mlx4_get_active_ports(dev, slave);
+	enum slave_port_state cur_state =
+		mlx4_get_slave_port_state(dev, slave, port);
+
+	*gen_event = SLAVE_PORT_GEN_EVENT_NONE;
+
+	if (slave >= dev->num_slaves || port > dev->caps.num_ports ||
+	    port <= 0 || !test_bit(port - 1, actv_ports.ports)) {
+		pr_err("%s: Error: asking for slave:%d, port:%d\n",
+		       __func__, slave, port);
+		return ret;
+	}
+
+	ctx = &priv->mfunc.master.slave_state[slave];
+	spin_lock_irqsave(&ctx->lock, flags);
+
+	switch (cur_state) {
+	case SLAVE_PORT_DOWN:
+		if (MLX4_PORT_STATE_DEV_EVENT_PORT_UP == event)
+			mlx4_set_slave_port_state(dev, slave, port,
+						  SLAVE_PENDING_UP);
+		break;
+	case SLAVE_PENDING_UP:
+		if (MLX4_PORT_STATE_DEV_EVENT_PORT_DOWN == event)
+			mlx4_set_slave_port_state(dev, slave, port,
+						  SLAVE_PORT_DOWN);
+		else if (MLX4_PORT_STATE_IB_PORT_STATE_EVENT_GID_VALID == event) {
+			mlx4_set_slave_port_state(dev, slave, port,
+						  SLAVE_PORT_UP);
+			*gen_event = SLAVE_PORT_GEN_EVENT_UP;
+		}
+		break;
+	case SLAVE_PORT_UP:
+		if (MLX4_PORT_STATE_DEV_EVENT_PORT_DOWN == event) {
+			mlx4_set_slave_port_state(dev, slave, port,
+						  SLAVE_PORT_DOWN);
+			*gen_event = SLAVE_PORT_GEN_EVENT_DOWN;
+		} else if (MLX4_PORT_STATE_IB_EVENT_GID_INVALID ==
+				event) {
+			mlx4_set_slave_port_state(dev, slave, port,
+						  SLAVE_PENDING_UP);
+			*gen_event = SLAVE_PORT_GEN_EVENT_DOWN;
+		}
+		break;
+	default:
+		pr_err("%s: BUG!!! UNKNOWN state: slave:%d, port:%d\n",
+		       __func__, slave, port);
+		goto out;
+	}
+	ret = mlx4_get_slave_port_state(dev, slave, port);
+
+out:
+	spin_unlock_irqrestore(&ctx->lock, flags);
+	return ret;
+}
+
+EXPORT_SYMBOL(set_and_calc_slave_port_state);
+
+int mlx4_gen_slaves_port_mgt_ev(struct mlx4_dev *dev, u8 port, int attr)
+{
+	struct mlx4_eqe eqe;
+
+	memset(&eqe, 0, sizeof(eqe));
+
+	eqe.type = MLX4_EVENT_TYPE_PORT_MNG_CHG_EVENT;
+	eqe.subtype = MLX4_DEV_PMC_SUBTYPE_PORT_INFO;
+	eqe.event.port_mgmt_change.port = port;
+	eqe.event.port_mgmt_change.params.port_info.changed_attr =
+		cpu_to_be32((u32) attr);
+
+	slave_event(dev, ALL_SLAVES, &eqe);
+	return 0;
+}
+EXPORT_SYMBOL(mlx4_gen_slaves_port_mgt_ev);
+
+void mlx4_master_handle_slave_flr(struct work_struct *work)
+{
+	struct mlx4_mfunc_master_ctx *master =
+		container_of(work, struct mlx4_mfunc_master_ctx,
+			     slave_flr_event_work);
+	struct mlx4_mfunc *mfunc =
+		container_of(master, struct mlx4_mfunc, master);
+	struct mlx4_priv *priv =
+		container_of(mfunc, struct mlx4_priv, mfunc);
+	struct mlx4_dev *dev = &priv->dev;
+	struct mlx4_slave_state *slave_state = priv->mfunc.master.slave_state;
+	int i;
+	int err;
+	unsigned long flags;
+
+	mlx4_dbg(dev, "mlx4_handle_slave_flr\n");
+
+	for (i = 0 ; i < dev->num_slaves; i++) {
+
+		if (MLX4_COMM_CMD_FLR == slave_state[i].last_cmd) {
+			mlx4_dbg(dev, "mlx4_handle_slave_flr: clean slave: %d\n",
+				 i);
+			/* In case of 'Reset flow' FLR can be generated for
+			 * a slave before mlx4_load_one is done.
+			 * make sure interface is up before trying to delete
+			 * slave resources which weren't allocated yet.
+			 */
+			if (dev->persist->interface_state &
+			    MLX4_INTERFACE_STATE_UP)
+				mlx4_delete_all_resources_for_slave(dev, i);
+			/*return the slave to running mode*/
+			spin_lock_irqsave(&priv->mfunc.master.slave_state_lock, flags);
+			slave_state[i].last_cmd = MLX4_COMM_CMD_RESET;
+			slave_state[i].is_slave_going_down = 0;
+			spin_unlock_irqrestore(&priv->mfunc.master.slave_state_lock, flags);
+			/*notify the FW:*/
+			err = mlx4_cmd(dev, 0, i, 0, MLX4_CMD_INFORM_FLR_DONE,
+				       MLX4_CMD_TIME_CLASS_A, MLX4_CMD_WRAPPED);
+			if (err)
+				mlx4_warn(dev, "Failed to notify FW on FLR done (slave:%d)\n",
+					  i);
+		}
+	}
+}
+
+static int mlx4_eq_int(struct mlx4_dev *dev, struct mlx4_eq *eq)
+{
+	struct mlx4_priv *priv = mlx4_priv(dev);
+	struct mlx4_eqe *eqe;
+	int cqn;
+	int eqes_found = 0;
+	int set_ci = 0;
+	int port;
+	int slave = 0;
+	int ret;
+	u32 flr_slave;
+	u8 update_slave_state;
+	int i;
+	enum slave_port_gen_event gen_event;
+	unsigned long flags;
+	struct mlx4_vport_state *s_info;
+	int eqe_size = dev->caps.eqe_size;
+
+	while ((eqe = next_eqe_sw(eq, dev->caps.eqe_factor, eqe_size))) {
+		/*
+		 * Make sure we read EQ entry contents after we've
+		 * checked the ownership bit.
+		 */
+		dma_rmb();
+
+		switch (eqe->type) {
+		case MLX4_EVENT_TYPE_COMP:
+			cqn = be32_to_cpu(eqe->event.comp.cqn) & 0xffffff;
+			mlx4_cq_completion(dev, cqn);
+			break;
+
+		case MLX4_EVENT_TYPE_PATH_MIG:
+		case MLX4_EVENT_TYPE_COMM_EST:
+		case MLX4_EVENT_TYPE_SQ_DRAINED:
+		case MLX4_EVENT_TYPE_SRQ_QP_LAST_WQE:
+		case MLX4_EVENT_TYPE_WQ_CATAS_ERROR:
+		case MLX4_EVENT_TYPE_PATH_MIG_FAILED:
+		case MLX4_EVENT_TYPE_WQ_INVAL_REQ_ERROR:
+		case MLX4_EVENT_TYPE_WQ_ACCESS_ERROR:
+			mlx4_dbg(dev, "event %d arrived\n", eqe->type);
+			if (mlx4_is_master(dev)) {
+				/* forward only to slave owning the QP */
+				ret = mlx4_get_slave_from_resource_id(dev,
+						RES_QP,
+						be32_to_cpu(eqe->event.qp.qpn)
+						& 0xffffff, &slave);
+				if (ret && ret != -ENOENT) {
+					mlx4_dbg(dev, "QP event %02x(%02x) on EQ %d at index %u: could not get slave id (%d)\n",
+						 eqe->type, eqe->subtype,
+						 eq->eqn, eq->cons_index, ret);
+					break;
+				}
+
+				if (!ret && slave != dev->caps.function) {
+					mlx4_slave_event(dev, slave, eqe);
+					break;
+				}
+
+			}
+			mlx4_qp_event(dev, be32_to_cpu(eqe->event.qp.qpn) &
+				      0xffffff, eqe->type);
+			break;
+
+		case MLX4_EVENT_TYPE_SRQ_LIMIT:
+			mlx4_dbg(dev, "%s: MLX4_EVENT_TYPE_SRQ_LIMIT. srq_no=0x%x, eq 0x%x\n",
+				 __func__, be32_to_cpu(eqe->event.srq.srqn),
+				 eq->eqn);
+		case MLX4_EVENT_TYPE_SRQ_CATAS_ERROR:
+			if (mlx4_is_master(dev)) {
+				/* forward only to slave owning the SRQ */
+				ret = mlx4_get_slave_from_resource_id(dev,
+						RES_SRQ,
+						be32_to_cpu(eqe->event.srq.srqn)
+						& 0xffffff,
+						&slave);
+				if (ret && ret != -ENOENT) {
+					mlx4_warn(dev, "SRQ event %02x(%02x) on EQ %d at index %u: could not get slave id (%d)\n",
+						  eqe->type, eqe->subtype,
+						  eq->eqn, eq->cons_index, ret);
+					break;
+				}
+				if (eqe->type ==
+				    MLX4_EVENT_TYPE_SRQ_CATAS_ERROR)
+					mlx4_warn(dev, "%s: slave:%d, srq_no:0x%x, event: %02x(%02x)\n",
+						  __func__, slave,
+						  be32_to_cpu(eqe->event.srq.srqn),
+						  eqe->type, eqe->subtype);
+
+				if (!ret && slave != dev->caps.function) {
+					if (eqe->type ==
+					    MLX4_EVENT_TYPE_SRQ_CATAS_ERROR)
+						mlx4_warn(dev, "%s: sending event %02x(%02x) to slave:%d\n",
+							  __func__, eqe->type,
+							  eqe->subtype, slave);
+					mlx4_slave_event(dev, slave, eqe);
+					break;
+				}
+			}
+			mlx4_srq_event(dev, be32_to_cpu(eqe->event.srq.srqn) &
+				       0xffffff, eqe->type);
+			break;
+
+		case MLX4_EVENT_TYPE_CMD:
+			mlx4_cmd_event(dev,
+				       be16_to_cpu(eqe->event.cmd.token),
+				       eqe->event.cmd.status,
+				       be64_to_cpu(eqe->event.cmd.out_param));
+			break;
+
+		case MLX4_EVENT_TYPE_PORT_CHANGE: {
+			struct mlx4_slaves_pport slaves_port;
+			port = be32_to_cpu(eqe->event.port_change.port) >> 28;
+			slaves_port = mlx4_phys_to_slaves_pport(dev, port);
+			if (eqe->subtype == MLX4_PORT_CHANGE_SUBTYPE_DOWN) {
+				mlx4_dispatch_event(dev, MLX4_DEV_EVENT_PORT_DOWN,
+						    port);
+				mlx4_priv(dev)->sense.do_sense_port[port] = 1;
+				if (!mlx4_is_master(dev))
+					break;
+				for (i = 0; i < dev->persist->num_vfs + 1;
+				     i++) {
+					int reported_port = mlx4_is_bonded(dev) ? 1 : mlx4_phys_to_slave_port(dev, i, port);
+
+					if (!test_bit(i, slaves_port.slaves) && !mlx4_is_bonded(dev))
+						continue;
+					if (dev->caps.port_type[port] == MLX4_PORT_TYPE_ETH) {
+						if (i == mlx4_master_func_num(dev))
+							continue;
+						mlx4_dbg(dev, "%s: Sending MLX4_PORT_CHANGE_SUBTYPE_DOWN to slave: %d, port:%d\n",
+							 __func__, i, port);
+						s_info = &priv->mfunc.master.vf_oper[i].vport[port].state;
+						if (IFLA_VF_LINK_STATE_AUTO == s_info->link_state) {
+							eqe->event.port_change.port =
+								cpu_to_be32(
+								(be32_to_cpu(eqe->event.port_change.port) & 0xFFFFFFF)
+								| (reported_port << 28));
+							mlx4_slave_event(dev, i, eqe);
+						}
+					} else {  /* IB port */
+						set_and_calc_slave_port_state(dev, i, port,
+									      MLX4_PORT_STATE_DEV_EVENT_PORT_DOWN,
+									      &gen_event);
+						/*we can be in pending state, then do not send port_down event*/
+						if (SLAVE_PORT_GEN_EVENT_DOWN ==  gen_event) {
+							if (i == mlx4_master_func_num(dev))
+								continue;
+							eqe->event.port_change.port =
+								cpu_to_be32(
+								(be32_to_cpu(eqe->event.port_change.port) & 0xFFFFFFF)
+								| (mlx4_phys_to_slave_port(dev, i, port) << 28));
+							mlx4_slave_event(dev, i, eqe);
+						}
+					}
+				}
+			} else {
+				mlx4_dispatch_event(dev, MLX4_DEV_EVENT_PORT_UP, port);
+
+				mlx4_priv(dev)->sense.do_sense_port[port] = 0;
+
+				if (!mlx4_is_master(dev))
+					break;
+				if (dev->caps.port_type[port] == MLX4_PORT_TYPE_ETH)
+					for (i = 0;
+					     i < dev->persist->num_vfs + 1;
+					     i++) {
+						int reported_port = mlx4_is_bonded(dev) ? 1 : mlx4_phys_to_slave_port(dev, i, port);
+
+						if (!test_bit(i, slaves_port.slaves) && !mlx4_is_bonded(dev))
+							continue;
+						if (i == mlx4_master_func_num(dev))
+							continue;
+						s_info = &priv->mfunc.master.vf_oper[i].vport[port].state;
+						if (IFLA_VF_LINK_STATE_AUTO == s_info->link_state) {
+							eqe->event.port_change.port =
+								cpu_to_be32(
+								(be32_to_cpu(eqe->event.port_change.port) & 0xFFFFFFF)
+								| (reported_port << 28));
+							mlx4_slave_event(dev, i, eqe);
+						}
+					}
+				else /* IB port */
+					/* port-up event will be sent to a slave when the
+					 * slave's alias-guid is set. This is done in alias_GUID.c
+					 */
+					set_all_slave_state(dev, port, MLX4_DEV_EVENT_PORT_UP);
+			}
+			break;
+		}
+
+		case MLX4_EVENT_TYPE_CQ_ERROR:
+			mlx4_warn(dev, "CQ %s on CQN %06x\n",
+				  eqe->event.cq_err.syndrome == 1 ?
+				  "overrun" : "access violation",
+				  be32_to_cpu(eqe->event.cq_err.cqn) & 0xffffff);
+			if (mlx4_is_master(dev)) {
+				ret = mlx4_get_slave_from_resource_id(dev,
+					RES_CQ,
+					be32_to_cpu(eqe->event.cq_err.cqn)
+					& 0xffffff, &slave);
+				if (ret && ret != -ENOENT) {
+					mlx4_dbg(dev, "CQ event %02x(%02x) on EQ %d at index %u: could not get slave id (%d)\n",
+						 eqe->type, eqe->subtype,
+						 eq->eqn, eq->cons_index, ret);
+					break;
+				}
+
+				if (!ret && slave != dev->caps.function) {
+					mlx4_slave_event(dev, slave, eqe);
+					break;
+				}
+			}
+			mlx4_cq_event(dev,
+				      be32_to_cpu(eqe->event.cq_err.cqn)
+				      & 0xffffff,
+				      eqe->type);
+			break;
+
+		case MLX4_EVENT_TYPE_EQ_OVERFLOW:
+			mlx4_warn(dev, "EQ overrun on EQN %d\n", eq->eqn);
+			break;
+
+		case MLX4_EVENT_TYPE_OP_REQUIRED:
+			atomic_inc(&priv->opreq_count);
+			/* FW commands can't be executed from interrupt context
+			 * working in deferred task
+			 */
+			queue_work(mlx4_wq, &priv->opreq_task);
+			break;
+
+		case MLX4_EVENT_TYPE_COMM_CHANNEL:
+			if (!mlx4_is_master(dev)) {
+				mlx4_warn(dev, "Received comm channel event for non master device\n");
+				break;
+			}
+			memcpy(&priv->mfunc.master.comm_arm_bit_vector,
+			       eqe->event.comm_channel_arm.bit_vec,
+			       sizeof(eqe->event.comm_channel_arm.bit_vec));
+			queue_work(priv->mfunc.master.comm_wq,
+				   &priv->mfunc.master.comm_work);
+			break;
+
+		case MLX4_EVENT_TYPE_FLR_EVENT:
+			flr_slave = be32_to_cpu(eqe->event.flr_event.slave_id);
+			if (!mlx4_is_master(dev)) {
+				mlx4_warn(dev, "Non-master function received FLR event\n");
+				break;
+			}
+
+			mlx4_dbg(dev, "FLR event for slave: %d\n", flr_slave);
+
+			if (flr_slave >= dev->num_slaves) {
+				mlx4_warn(dev,
+					  "Got FLR for unknown function: %d\n",
+					  flr_slave);
+				update_slave_state = 0;
+			} else
+				update_slave_state = 1;
+
+			spin_lock_irqsave(&priv->mfunc.master.slave_state_lock, flags);
+			if (update_slave_state) {
+				priv->mfunc.master.slave_state[flr_slave].active = false;
+				priv->mfunc.master.slave_state[flr_slave].last_cmd = MLX4_COMM_CMD_FLR;
+				priv->mfunc.master.slave_state[flr_slave].is_slave_going_down = 1;
+			}
+			spin_unlock_irqrestore(&priv->mfunc.master.slave_state_lock, flags);
+			mlx4_dispatch_event(dev, MLX4_DEV_EVENT_SLAVE_SHUTDOWN,
+					    flr_slave);
+			queue_work(priv->mfunc.master.comm_wq,
+				   &priv->mfunc.master.slave_flr_event_work);
+			break;
+
+		case MLX4_EVENT_TYPE_FATAL_WARNING:
+			if (eqe->subtype == MLX4_FATAL_WARNING_SUBTYPE_WARMING) {
+				if (mlx4_is_master(dev))
+					for (i = 0; i < dev->num_slaves; i++) {
+						mlx4_dbg(dev, "%s: Sending MLX4_FATAL_WARNING_SUBTYPE_WARMING to slave: %d\n",
+							 __func__, i);
+						if (i == dev->caps.function)
+							continue;
+						mlx4_slave_event(dev, i, eqe);
+					}
+				mlx4_err(dev, "Temperature Threshold was reached! Threshold: %d celsius degrees; Current Temperature: %d\n",
+					 be16_to_cpu(eqe->event.warming.warning_threshold),
+					 be16_to_cpu(eqe->event.warming.current_temperature));
+			} else
+				mlx4_warn(dev, "Unhandled event FATAL WARNING (%02x), subtype %02x on EQ %d at index %u. owner=%x, nent=0x%x, slave=%x, ownership=%s\n",
+					  eqe->type, eqe->subtype, eq->eqn,
+					  eq->cons_index, eqe->owner, eq->nent,
+					  eqe->slave_id,
+					  !!(eqe->owner & 0x80) ^
+					  !!(eq->cons_index & eq->nent) ? "HW" : "SW");
+
+			break;
+
+		case MLX4_EVENT_TYPE_PORT_MNG_CHG_EVENT:
+			mlx4_dispatch_event(dev, MLX4_DEV_EVENT_PORT_MGMT_CHANGE,
+					    (unsigned long) eqe);
+			break;
+
+		case MLX4_EVENT_TYPE_RECOVERABLE_ERROR_EVENT:
+			switch (eqe->subtype) {
+			case MLX4_RECOVERABLE_ERROR_EVENT_SUBTYPE_BAD_CABLE:
+				mlx4_warn(dev, "Bad cable detected on port %u\n",
+					  eqe->event.bad_cable.port);
+				break;
+			case MLX4_RECOVERABLE_ERROR_EVENT_SUBTYPE_UNSUPPORTED_CABLE:
+				mlx4_warn(dev, "Unsupported cable detected\n");
+				break;
+			default:
+				mlx4_dbg(dev,
+					 "Unhandled recoverable error event detected: %02x(%02x) on EQ %d at index %u. owner=%x, nent=0x%x, ownership=%s\n",
+					 eqe->type, eqe->subtype, eq->eqn,
+					 eq->cons_index, eqe->owner, eq->nent,
+					 !!(eqe->owner & 0x80) ^
+					 !!(eq->cons_index & eq->nent) ? "HW" : "SW");
+				break;
+			}
+			break;
+
+		case MLX4_EVENT_TYPE_EEC_CATAS_ERROR:
+		case MLX4_EVENT_TYPE_ECC_DETECT:
+		default:
+			mlx4_warn(dev, "Unhandled event %02x(%02x) on EQ %d at index %u. owner=%x, nent=0x%x, slave=%x, ownership=%s\n",
+				  eqe->type, eqe->subtype, eq->eqn,
+				  eq->cons_index, eqe->owner, eq->nent,
+				  eqe->slave_id,
+				  !!(eqe->owner & 0x80) ^
+				  !!(eq->cons_index & eq->nent) ? "HW" : "SW");
+			break;
+		};
+
+		++eq->cons_index;
+		eqes_found = 1;
+		++set_ci;
+
+		/*
+		 * The HCA will think the queue has overflowed if we
+		 * don't tell it we've been processing events.  We
+		 * create our EQs with MLX4_NUM_SPARE_EQE extra
+		 * entries, so we must update our consumer index at
+		 * least that often.
+		 */
+		if (unlikely(set_ci >= MLX4_NUM_SPARE_EQE)) {
+			eq_set_ci(eq, 0);
+			set_ci = 0;
+		}
+	}
+
+	eq_set_ci(eq, 1);
+
+	return eqes_found;
+}
+
+static irqreturn_t mlx4_interrupt(int irq, void *dev_ptr)
+{
+	struct mlx4_dev *dev = dev_ptr;
+	struct mlx4_priv *priv = mlx4_priv(dev);
+	int work = 0;
+	int i;
+
+	writel(priv->eq_table.clr_mask, priv->eq_table.clr_int);
+
+	for (i = 0; i < dev->caps.num_comp_vectors + 1; ++i)
+		work |= mlx4_eq_int(dev, &priv->eq_table.eq[i]);
+
+	return IRQ_RETVAL(work);
+}
+
+static irqreturn_t mlx4_msi_x_interrupt(int irq, void *eq_ptr)
+{
+	struct mlx4_eq  *eq  = eq_ptr;
+	struct mlx4_dev *dev = eq->dev;
+
+	mlx4_eq_int(dev, eq);
+
+	/* MSI-X vectors always belong to us */
+	return IRQ_HANDLED;
+}
+
+int mlx4_MAP_EQ_wrapper(struct mlx4_dev *dev, int slave,
+			struct mlx4_vhcr *vhcr,
+			struct mlx4_cmd_mailbox *inbox,
+			struct mlx4_cmd_mailbox *outbox,
+			struct mlx4_cmd_info *cmd)
+{
+	struct mlx4_priv *priv = mlx4_priv(dev);
+	struct mlx4_slave_event_eq_info *event_eq =
+		priv->mfunc.master.slave_state[slave].event_eq;
+	u32 in_modifier = vhcr->in_modifier;
+	u32 eqn = in_modifier & 0x3FF;
+	u64 in_param =  vhcr->in_param;
+	int err = 0;
+	int i;
+
+	if (slave == dev->caps.function)
+		err = mlx4_cmd(dev, in_param, (in_modifier & 0x80000000) | eqn,
+			       0, MLX4_CMD_MAP_EQ, MLX4_CMD_TIME_CLASS_B,
+			       MLX4_CMD_NATIVE);
+	if (!err)
+		for (i = 0; i < MLX4_EVENT_TYPES_NUM; ++i)
+			if (in_param & (1LL << i))
+				event_eq[i].eqn = in_modifier >> 31 ? -1 : eqn;
+
+	return err;
+}
+
+static int mlx4_MAP_EQ(struct mlx4_dev *dev, u64 event_mask, int unmap,
+			int eq_num)
+{
+	return mlx4_cmd(dev, event_mask, (unmap << 31) | eq_num,
+			0, MLX4_CMD_MAP_EQ, MLX4_CMD_TIME_CLASS_B,
+			MLX4_CMD_WRAPPED);
+}
+
+static int mlx4_SW2HW_EQ(struct mlx4_dev *dev, struct mlx4_cmd_mailbox *mailbox,
+			 int eq_num)
+{
+	return mlx4_cmd(dev, mailbox->dma, eq_num, 0,
+			MLX4_CMD_SW2HW_EQ, MLX4_CMD_TIME_CLASS_A,
+			MLX4_CMD_WRAPPED);
+}
+
+static int mlx4_HW2SW_EQ(struct mlx4_dev *dev,  int eq_num)
+{
+	return mlx4_cmd(dev, 0, eq_num, 1, MLX4_CMD_HW2SW_EQ,
+			MLX4_CMD_TIME_CLASS_A, MLX4_CMD_WRAPPED);
+}
+
+static int mlx4_num_eq_uar(struct mlx4_dev *dev)
+{
+	/*
+	 * Each UAR holds 4 EQ doorbells.  To figure out how many UARs
+	 * we need to map, take the difference of highest index and
+	 * the lowest index we'll use and add 1.
+	 */
+	return (dev->caps.num_comp_vectors + 1 + dev->caps.reserved_eqs) / 4 -
+		dev->caps.reserved_eqs / 4 + 1;
+}
+
+static void __iomem *mlx4_get_eq_uar(struct mlx4_dev *dev, struct mlx4_eq *eq)
+{
+	struct mlx4_priv *priv = mlx4_priv(dev);
+	int index;
+
+	index = eq->eqn / 4 - dev->caps.reserved_eqs / 4;
+
+	if (!priv->eq_table.uar_map[index]) {
+		priv->eq_table.uar_map[index] =
+			ioremap(
+				pci_resource_start(dev->persist->pdev, 2) +
+				((eq->eqn / 4) << (dev->uar_page_shift)),
+				(1 << (dev->uar_page_shift)));
+		if (!priv->eq_table.uar_map[index]) {
+			mlx4_err(dev, "Couldn't map EQ doorbell for EQN 0x%06x\n",
+				 eq->eqn);
+			return NULL;
+		}
+	}
+
+	return priv->eq_table.uar_map[index] + 0x800 + 8 * (eq->eqn % 4);
+}
+
+static void mlx4_unmap_uar(struct mlx4_dev *dev)
+{
+	struct mlx4_priv *priv = mlx4_priv(dev);
+	int i;
+
+	for (i = 0; i < mlx4_num_eq_uar(dev); ++i)
+		if (priv->eq_table.uar_map[i]) {
+			iounmap(priv->eq_table.uar_map[i]);
+			priv->eq_table.uar_map[i] = NULL;
+		}
+}
+
+static int mlx4_create_eq(struct mlx4_dev *dev, int nent,
+			  u8 intr, struct mlx4_eq *eq)
+{
+	struct mlx4_priv *priv = mlx4_priv(dev);
+	struct mlx4_cmd_mailbox *mailbox;
+	struct mlx4_eq_context *eq_context;
+	int npages;
+	u64 *dma_list = NULL;
+	dma_addr_t t;
+	u64 mtt_addr;
+	int err = -ENOMEM;
+	int i;
+
+	eq->dev   = dev;
+	eq->nent  = roundup_pow_of_two(max(nent, 2));
+	/* CX3 is capable of extending the CQE/EQE from 32 to 64 bytes, with
+	 * strides of 64B,128B and 256B.
+	 */
+	npages = PAGE_ALIGN(eq->nent * dev->caps.eqe_size) / PAGE_SIZE;
+
+	eq->page_list = kmalloc_array(npages, sizeof(*eq->page_list),
+				      GFP_KERNEL);
+	if (!eq->page_list)
+		goto err_out;
+
+	for (i = 0; i < npages; ++i)
+		eq->page_list[i].buf = NULL;
+
+	dma_list = kmalloc_array(npages, sizeof(*dma_list), GFP_KERNEL);
+	if (!dma_list)
+		goto err_out_free;
+
+	mailbox = mlx4_alloc_cmd_mailbox(dev);
+	if (IS_ERR(mailbox))
+		goto err_out_free;
+	eq_context = mailbox->buf;
+
+	for (i = 0; i < npages; ++i) {
+		eq->page_list[i].buf = dma_alloc_coherent(&dev->persist->
+							  pdev->dev,
+							  PAGE_SIZE, &t,
+							  GFP_KERNEL);
+		if (!eq->page_list[i].buf)
+			goto err_out_free_pages;
+
+		dma_list[i] = t;
+		eq->page_list[i].map = t;
+
+		memset(eq->page_list[i].buf, 0, PAGE_SIZE);
+	}
+
+	eq->eqn = mlx4_bitmap_alloc(&priv->eq_table.bitmap);
+	if (eq->eqn == -1)
+		goto err_out_free_pages;
+
+	eq->doorbell = mlx4_get_eq_uar(dev, eq);
+	if (!eq->doorbell) {
+		err = -ENOMEM;
+		goto err_out_free_eq;
+	}
+
+	err = mlx4_mtt_init(dev, npages, PAGE_SHIFT, &eq->mtt);
+	if (err)
+		goto err_out_free_eq;
+
+	err = mlx4_write_mtt(dev, &eq->mtt, 0, npages, dma_list);
+	if (err)
+		goto err_out_free_mtt;
+
+	eq_context->flags	  = cpu_to_be32(MLX4_EQ_STATUS_OK   |
+						MLX4_EQ_STATE_ARMED);
+	eq_context->log_eq_size	  = ilog2(eq->nent);
+	eq_context->intr	  = intr;
+	eq_context->log_page_size = PAGE_SHIFT - MLX4_ICM_PAGE_SHIFT;
+
+	mtt_addr = mlx4_mtt_addr(dev, &eq->mtt);
+	eq_context->mtt_base_addr_h = mtt_addr >> 32;
+	eq_context->mtt_base_addr_l = cpu_to_be32(mtt_addr & 0xffffffff);
+
+	err = mlx4_SW2HW_EQ(dev, mailbox, eq->eqn);
+	if (err) {
+		mlx4_warn(dev, "SW2HW_EQ failed (%d)\n", err);
+		goto err_out_free_mtt;
+	}
+
+	kfree(dma_list);
+	mlx4_free_cmd_mailbox(dev, mailbox);
+
+	eq->cons_index = 0;
+
+	INIT_LIST_HEAD(&eq->tasklet_ctx.list);
+	INIT_LIST_HEAD(&eq->tasklet_ctx.process_list);
+	spin_lock_init(&eq->tasklet_ctx.lock);
+	tasklet_init(&eq->tasklet_ctx.task, mlx4_cq_tasklet_cb,
+		     (unsigned long)&eq->tasklet_ctx);
+
+	return err;
+
+err_out_free_mtt:
+	mlx4_mtt_cleanup(dev, &eq->mtt);
+
+err_out_free_eq:
+	mlx4_bitmap_free(&priv->eq_table.bitmap, eq->eqn, MLX4_USE_RR);
+
+err_out_free_pages:
+	for (i = 0; i < npages; ++i)
+		if (eq->page_list[i].buf)
+			dma_free_coherent(&dev->persist->pdev->dev, PAGE_SIZE,
+					  eq->page_list[i].buf,
+					  eq->page_list[i].map);
+
+	mlx4_free_cmd_mailbox(dev, mailbox);
+
+err_out_free:
+	kfree(eq->page_list);
+	kfree(dma_list);
+
+err_out:
+	return err;
+}
+
+static void mlx4_free_eq(struct mlx4_dev *dev,
+			 struct mlx4_eq *eq)
+{
+	struct mlx4_priv *priv = mlx4_priv(dev);
+	int err;
+	int i;
+	/* CX3 is capable of extending the CQE/EQE from 32 to 64 bytes, with
+	 * strides of 64B,128B and 256B
+	 */
+	int npages = PAGE_ALIGN(dev->caps.eqe_size  * eq->nent) / PAGE_SIZE;
+
+	err = mlx4_HW2SW_EQ(dev, eq->eqn);
+	if (err)
+		mlx4_warn(dev, "HW2SW_EQ failed (%d)\n", err);
+
+	synchronize_irq(eq->irq);
+	tasklet_disable(&eq->tasklet_ctx.task);
+
+	mlx4_mtt_cleanup(dev, &eq->mtt);
+	for (i = 0; i < npages; ++i)
+		dma_free_coherent(&dev->persist->pdev->dev, PAGE_SIZE,
+				  eq->page_list[i].buf,
+				  eq->page_list[i].map);
+
+	kfree(eq->page_list);
+	mlx4_bitmap_free(&priv->eq_table.bitmap, eq->eqn, MLX4_USE_RR);
+}
+
+static void mlx4_free_irqs(struct mlx4_dev *dev)
+{
+	struct mlx4_eq_table *eq_table = &mlx4_priv(dev)->eq_table;
+	int	i;
+
+	if (eq_table->have_irq)
+		free_irq(dev->persist->pdev->irq, dev);
+
+	for (i = 0; i < dev->caps.num_comp_vectors + 1; ++i)
+		if (eq_table->eq[i].have_irq) {
+			free_cpumask_var(eq_table->eq[i].affinity_mask);
+#if defined(CONFIG_SMP)
+			irq_set_affinity_hint(eq_table->eq[i].irq, NULL);
+#endif
+			free_irq(eq_table->eq[i].irq, eq_table->eq + i);
+			eq_table->eq[i].have_irq = 0;
+		}
+
+	kfree(eq_table->irq_names);
+}
+
+static int mlx4_map_clr_int(struct mlx4_dev *dev)
+{
+	struct mlx4_priv *priv = mlx4_priv(dev);
+
+	priv->clr_base = ioremap(pci_resource_start(dev->persist->pdev,
+				 priv->fw.clr_int_bar) +
+				 priv->fw.clr_int_base, MLX4_CLR_INT_SIZE);
+	if (!priv->clr_base) {
+		mlx4_err(dev, "Couldn't map interrupt clear register, aborting\n");
+		return -ENOMEM;
+	}
+
+	return 0;
+}
+
+static void mlx4_unmap_clr_int(struct mlx4_dev *dev)
+{
+	struct mlx4_priv *priv = mlx4_priv(dev);
+
+	iounmap(priv->clr_base);
+}
+
+int mlx4_alloc_eq_table(struct mlx4_dev *dev)
+{
+	struct mlx4_priv *priv = mlx4_priv(dev);
+
+	priv->eq_table.eq = kcalloc(dev->caps.num_eqs - dev->caps.reserved_eqs,
+				    sizeof(*priv->eq_table.eq), GFP_KERNEL);
+	if (!priv->eq_table.eq)
+		return -ENOMEM;
+
+	return 0;
+}
+
+void mlx4_free_eq_table(struct mlx4_dev *dev)
+{
+	kfree(mlx4_priv(dev)->eq_table.eq);
+}
+
+int mlx4_init_eq_table(struct mlx4_dev *dev)
+{
+	struct mlx4_priv *priv = mlx4_priv(dev);
+	int err;
+	int i;
+
+	priv->eq_table.uar_map = kcalloc(mlx4_num_eq_uar(dev),
+					 sizeof(*priv->eq_table.uar_map),
+					 GFP_KERNEL);
+	if (!priv->eq_table.uar_map) {
+		err = -ENOMEM;
+		goto err_out_free;
+	}
+
+	err = mlx4_bitmap_init(&priv->eq_table.bitmap,
+			       roundup_pow_of_two(dev->caps.num_eqs),
+			       dev->caps.num_eqs - 1,
+			       dev->caps.reserved_eqs,
+			       roundup_pow_of_two(dev->caps.num_eqs) -
+			       dev->caps.num_eqs);
+	if (err)
+		goto err_out_free;
+
+	for (i = 0; i < mlx4_num_eq_uar(dev); ++i)
+		priv->eq_table.uar_map[i] = NULL;
+
+	if (!mlx4_is_slave(dev)) {
+		err = mlx4_map_clr_int(dev);
+		if (err)
+			goto err_out_bitmap;
+
+		priv->eq_table.clr_mask =
+			swab32(1 << (priv->eq_table.inta_pin & 31));
+		priv->eq_table.clr_int  = priv->clr_base +
+			(priv->eq_table.inta_pin < 32 ? 4 : 0);
+	}
+
+	priv->eq_table.irq_names =
+		kmalloc_array(MLX4_IRQNAME_SIZE,
+			      (dev->caps.num_comp_vectors + 1),
+			      GFP_KERNEL);
+	if (!priv->eq_table.irq_names) {
+		err = -ENOMEM;
+		goto err_out_clr_int;
+	}
+
+	for (i = 0; i < dev->caps.num_comp_vectors + 1; ++i) {
+		if (i == MLX4_EQ_ASYNC) {
+			err = mlx4_create_eq(dev,
+					     MLX4_NUM_ASYNC_EQE + MLX4_NUM_SPARE_EQE,
+					     0, &priv->eq_table.eq[MLX4_EQ_ASYNC]);
+		} else {
+			struct mlx4_eq	*eq = &priv->eq_table.eq[i];
+#ifdef CONFIG_RFS_ACCEL
+			int port = find_first_bit(eq->actv_ports.ports,
+						  dev->caps.num_ports) + 1;
+
+			if (port <= dev->caps.num_ports) {
+				struct mlx4_port_info *info =
+					&mlx4_priv(dev)->port[port];
+
+				if (!info->rmap) {
+					info->rmap = alloc_irq_cpu_rmap(
+						mlx4_get_eqs_per_port(dev, port));
+					if (!info->rmap) {
+						mlx4_warn(dev, "Failed to allocate cpu rmap\n");
+						err = -ENOMEM;
+						goto err_out_unmap;
+					}
+				}
+
+				err = irq_cpu_rmap_add(
+					info->rmap, eq->irq);
+				if (err)
+					mlx4_warn(dev, "Failed adding irq rmap\n");
+			}
+#endif
+			err = mlx4_create_eq(dev, dev->quotas.cq +
+					     MLX4_NUM_SPARE_EQE,
+					     (dev->flags & MLX4_FLAG_MSI_X) ?
+					     i + 1 - !!(i > MLX4_EQ_ASYNC) : 0,
+					     eq);
+		}
+		if (err)
+			goto err_out_unmap;
+	}
+
+	if (dev->flags & MLX4_FLAG_MSI_X) {
+		const char *eq_name;
+
+		snprintf(priv->eq_table.irq_names +
+			 MLX4_EQ_ASYNC * MLX4_IRQNAME_SIZE,
+			 MLX4_IRQNAME_SIZE,
+			 "mlx4-async@pci:%s",
+			 pci_name(dev->persist->pdev));
+		eq_name = priv->eq_table.irq_names +
+			MLX4_EQ_ASYNC * MLX4_IRQNAME_SIZE;
+
+		err = request_irq(priv->eq_table.eq[MLX4_EQ_ASYNC].irq,
+				  mlx4_msi_x_interrupt, 0, eq_name,
+				  priv->eq_table.eq + MLX4_EQ_ASYNC);
+		if (err)
+			goto err_out_unmap;
+
+		priv->eq_table.eq[MLX4_EQ_ASYNC].have_irq = 1;
+	} else {
+		snprintf(priv->eq_table.irq_names,
+			 MLX4_IRQNAME_SIZE,
+			 DRV_NAME "@pci:%s",
+			 pci_name(dev->persist->pdev));
+		err = request_irq(dev->persist->pdev->irq, mlx4_interrupt,
+				  IRQF_SHARED, priv->eq_table.irq_names, dev);
+		if (err)
+			goto err_out_unmap;
+
+		priv->eq_table.have_irq = 1;
+	}
+
+	err = mlx4_MAP_EQ(dev, get_async_ev_mask(dev), 0,
+			  priv->eq_table.eq[MLX4_EQ_ASYNC].eqn);
+	if (err)
+		mlx4_warn(dev, "MAP_EQ for async EQ %d failed (%d)\n",
+			   priv->eq_table.eq[MLX4_EQ_ASYNC].eqn, err);
+
+	/* arm ASYNC eq */
+	eq_set_ci(&priv->eq_table.eq[MLX4_EQ_ASYNC], 1);
+
+	return 0;
+
+err_out_unmap:
+	while (i > 0)
+		mlx4_free_eq(dev, &priv->eq_table.eq[--i]);
+#ifdef CONFIG_RFS_ACCEL
+	for (i = 1; i <= dev->caps.num_ports; i++) {
+		if (mlx4_priv(dev)->port[i].rmap) {
+			free_irq_cpu_rmap(mlx4_priv(dev)->port[i].rmap);
+			mlx4_priv(dev)->port[i].rmap = NULL;
+		}
+	}
+#endif
+	mlx4_free_irqs(dev);
+
+err_out_clr_int:
+	if (!mlx4_is_slave(dev))
+		mlx4_unmap_clr_int(dev);
+
+err_out_bitmap:
+	mlx4_unmap_uar(dev);
+	mlx4_bitmap_cleanup(&priv->eq_table.bitmap);
+
+err_out_free:
+	kfree(priv->eq_table.uar_map);
+
+	return err;
+}
+
+void mlx4_cleanup_eq_table(struct mlx4_dev *dev)
+{
+	struct mlx4_priv *priv = mlx4_priv(dev);
+	int i;
+
+	mlx4_MAP_EQ(dev, get_async_ev_mask(dev), 1,
+		    priv->eq_table.eq[MLX4_EQ_ASYNC].eqn);
+
+#ifdef CONFIG_RFS_ACCEL
+	for (i = 1; i <= dev->caps.num_ports; i++) {
+		if (mlx4_priv(dev)->port[i].rmap) {
+			free_irq_cpu_rmap(mlx4_priv(dev)->port[i].rmap);
+			mlx4_priv(dev)->port[i].rmap = NULL;
+		}
+	}
+#endif
+	mlx4_free_irqs(dev);
+
+	for (i = 0; i < dev->caps.num_comp_vectors + 1; ++i)
+		mlx4_free_eq(dev, &priv->eq_table.eq[i]);
+
+	if (!mlx4_is_slave(dev))
+		mlx4_unmap_clr_int(dev);
+
+	mlx4_unmap_uar(dev);
+	mlx4_bitmap_cleanup(&priv->eq_table.bitmap);
+
+	kfree(priv->eq_table.uar_map);
+}
+
+/* A test that verifies that we can accept interrupts
+ * on the vector allocated for asynchronous events
+ */
+int mlx4_test_async(struct mlx4_dev *dev)
+{
+	return mlx4_NOP(dev);
+}
+EXPORT_SYMBOL(mlx4_test_async);
+
+/* A test that verifies that we can accept interrupts
+ * on the given irq vector of the tested port.
+ * Interrupts are checked using the NOP command.
+ */
+int mlx4_test_interrupt(struct mlx4_dev *dev, int vector)
+{
+	struct mlx4_priv *priv = mlx4_priv(dev);
+	int err;
+
+	/* Temporary use polling for command completions */
+	mlx4_cmd_use_polling(dev);
+
+	/* Map the new eq to handle all asynchronous events */
+	err = mlx4_MAP_EQ(dev, get_async_ev_mask(dev), 0,
+			  priv->eq_table.eq[MLX4_CQ_TO_EQ_VECTOR(vector)].eqn);
+	if (err) {
+		mlx4_warn(dev, "Failed mapping eq for interrupt test\n");
+		goto out;
+	}
+
+	/* Go back to using events */
+	mlx4_cmd_use_events(dev);
+	err = mlx4_NOP(dev);
+
+	/* Return to default */
+	mlx4_cmd_use_polling(dev);
+out:
+	mlx4_MAP_EQ(dev, get_async_ev_mask(dev), 0,
+		    priv->eq_table.eq[MLX4_EQ_ASYNC].eqn);
+	mlx4_cmd_use_events(dev);
+
+	return err;
+}
+EXPORT_SYMBOL(mlx4_test_interrupt);
+
+bool mlx4_is_eq_vector_valid(struct mlx4_dev *dev, u8 port, int vector)
+{
+	struct mlx4_priv *priv = mlx4_priv(dev);
+
+	vector = MLX4_CQ_TO_EQ_VECTOR(vector);
+	if (vector < 0 || (vector >= dev->caps.num_comp_vectors + 1) ||
+	    (vector == MLX4_EQ_ASYNC))
+		return false;
+
+	return test_bit(port - 1, priv->eq_table.eq[vector].actv_ports.ports);
+}
+EXPORT_SYMBOL(mlx4_is_eq_vector_valid);
+
+u32 mlx4_get_eqs_per_port(struct mlx4_dev *dev, u8 port)
+{
+	struct mlx4_priv *priv = mlx4_priv(dev);
+	unsigned int i;
+	unsigned int sum = 0;
+
+	for (i = 0; i < dev->caps.num_comp_vectors + 1; i++)
+		sum += !!test_bit(port - 1,
+				  priv->eq_table.eq[i].actv_ports.ports);
+
+	return sum;
+}
+EXPORT_SYMBOL(mlx4_get_eqs_per_port);
+
+int mlx4_is_eq_shared(struct mlx4_dev *dev, int vector)
+{
+	struct mlx4_priv *priv = mlx4_priv(dev);
+
+	vector = MLX4_CQ_TO_EQ_VECTOR(vector);
+	if (vector <= 0 || (vector >= dev->caps.num_comp_vectors + 1))
+		return -EINVAL;
+
+	return !!(bitmap_weight(priv->eq_table.eq[vector].actv_ports.ports,
+				dev->caps.num_ports) > 1);
+}
+EXPORT_SYMBOL(mlx4_is_eq_shared);
+
+struct cpu_rmap *mlx4_get_cpu_rmap(struct mlx4_dev *dev, int port)
+{
+	return mlx4_priv(dev)->port[port].rmap;
+}
+EXPORT_SYMBOL(mlx4_get_cpu_rmap);
+
+int mlx4_assign_eq(struct mlx4_dev *dev, u8 port, int *vector)
+{
+	struct mlx4_priv *priv = mlx4_priv(dev);
+	int err = 0, i = 0;
+	u32 min_ref_count_val = (u32)-1;
+	int requested_vector = MLX4_CQ_TO_EQ_VECTOR(*vector);
+	int *prequested_vector = NULL;
+
+
+	mutex_lock(&priv->msix_ctl.pool_lock);
+	if (requested_vector < (dev->caps.num_comp_vectors + 1) &&
+	    (requested_vector >= 0) &&
+	    (requested_vector != MLX4_EQ_ASYNC)) {
+		if (test_bit(port - 1,
+			     priv->eq_table.eq[requested_vector].actv_ports.ports)) {
+			prequested_vector = &requested_vector;
+		} else {
+			struct mlx4_eq *eq;
+
+			for (i = 1; i < port;
+			     requested_vector += mlx4_get_eqs_per_port(dev, i++))
+				;
+
+			eq = &priv->eq_table.eq[requested_vector];
+			if (requested_vector < dev->caps.num_comp_vectors + 1 &&
+			    test_bit(port - 1, eq->actv_ports.ports)) {
+				prequested_vector = &requested_vector;
+			}
+		}
+	}
+
+	if  (!prequested_vector) {
+		requested_vector = -1;
+		for (i = 0; min_ref_count_val && i < dev->caps.num_comp_vectors + 1;
+		     i++) {
+			struct mlx4_eq *eq = &priv->eq_table.eq[i];
+
+			if (min_ref_count_val > eq->ref_count &&
+			    test_bit(port - 1, eq->actv_ports.ports)) {
+				min_ref_count_val = eq->ref_count;
+				requested_vector = i;
+			}
+		}
+
+		if (requested_vector < 0) {
+			err = -ENOSPC;
+			goto err_unlock;
+		}
+
+		prequested_vector = &requested_vector;
+	}
+
+	if (!test_bit(*prequested_vector, priv->msix_ctl.pool_bm) &&
+	    dev->flags & MLX4_FLAG_MSI_X) {
+		set_bit(*prequested_vector, priv->msix_ctl.pool_bm);
+		snprintf(priv->eq_table.irq_names +
+			 *prequested_vector * MLX4_IRQNAME_SIZE,
+			 MLX4_IRQNAME_SIZE, "mlx4-%d@%s",
+			 *prequested_vector, dev_name(&dev->persist->pdev->dev));
+
+		err = request_irq(priv->eq_table.eq[*prequested_vector].irq,
+				  mlx4_msi_x_interrupt, 0,
+				  &priv->eq_table.irq_names[*prequested_vector << 5],
+				  priv->eq_table.eq + *prequested_vector);
+
+		if (err) {
+			clear_bit(*prequested_vector, priv->msix_ctl.pool_bm);
+			*prequested_vector = -1;
+		} else {
+#if defined(CONFIG_SMP)
+			mlx4_set_eq_affinity_hint(priv, *prequested_vector);
+#endif
+			eq_set_ci(&priv->eq_table.eq[*prequested_vector], 1);
+			priv->eq_table.eq[*prequested_vector].have_irq = 1;
+		}
+	}
+
+	if (!err && *prequested_vector >= 0)
+		priv->eq_table.eq[*prequested_vector].ref_count++;
+
+err_unlock:
+	mutex_unlock(&priv->msix_ctl.pool_lock);
+
+	if (!err && *prequested_vector >= 0)
+		*vector = MLX4_EQ_TO_CQ_VECTOR(*prequested_vector);
+	else
+		*vector = 0;
+
+	return err;
+}
+EXPORT_SYMBOL(mlx4_assign_eq);
+
+int mlx4_eq_get_irq(struct mlx4_dev *dev, int cq_vec)
+{
+	struct mlx4_priv *priv = mlx4_priv(dev);
+
+	return priv->eq_table.eq[MLX4_CQ_TO_EQ_VECTOR(cq_vec)].irq;
+}
+EXPORT_SYMBOL(mlx4_eq_get_irq);
+
+void mlx4_release_eq(struct mlx4_dev *dev, int vec)
+{
+	struct mlx4_priv *priv = mlx4_priv(dev);
+	int eq_vec = MLX4_CQ_TO_EQ_VECTOR(vec);
+
+	mutex_lock(&priv->msix_ctl.pool_lock);
+	priv->eq_table.eq[eq_vec].ref_count--;
+
+	/* once we allocated EQ, we don't release it because it might be binded
+	 * to cpu_rmap.
+	 */
+	mutex_unlock(&priv->msix_ctl.pool_lock);
+}
+EXPORT_SYMBOL(mlx4_release_eq);
+
diff --git a/drivers/net/ethernet/mellanox/mlx4/fw.c b/drivers/net/ethernet/mellanox/mlx4/fw.c
new file mode 100644
index 000000000..ce57df04a
--- /dev/null
+++ b/drivers/net/ethernet/mellanox/mlx4/fw.c
@@ -0,0 +1,3108 @@
+/*
+ * Copyright (c) 2004, 2005 Topspin Communications.  All rights reserved.
+ * Copyright (c) 2005, 2006, 2007, 2008 Mellanox Technologies. All rights reserved.
+ * Copyright (c) 2005, 2006, 2007 Cisco Systems, Inc.  All rights reserved.
+ *
+ * This software is available to you under a choice of one of two
+ * licenses.  You may choose to be licensed under the terms of the GNU
+ * General Public License (GPL) Version 2, available from the file
+ * COPYING in the main directory of this source tree, or the
+ * OpenIB.org BSD license below:
+ *
+ *     Redistribution and use in source and binary forms, with or
+ *     without modification, are permitted provided that the following
+ *     conditions are met:
+ *
+ *      - Redistributions of source code must retain the above
+ *        copyright notice, this list of conditions and the following
+ *        disclaimer.
+ *
+ *      - Redistributions in binary form must reproduce the above
+ *        copyright notice, this list of conditions and the following
+ *        disclaimer in the documentation and/or other materials
+ *        provided with the distribution.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
+ * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
+ * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS
+ * BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN
+ * ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
+ * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+ * SOFTWARE.
+ */
+
+#include <linux/etherdevice.h>
+#include <linux/mlx4/cmd.h>
+#include <linux/module.h>
+#include <linux/cache.h>
+#include <linux/kernel.h>
+#include <uapi/rdma/mlx4-abi.h>
+
+#include "fw.h"
+#include "icm.h"
+
+enum {
+	MLX4_COMMAND_INTERFACE_MIN_REV		= 2,
+	MLX4_COMMAND_INTERFACE_MAX_REV		= 3,
+	MLX4_COMMAND_INTERFACE_NEW_PORT_CMDS	= 3,
+};
+
+extern void __buggy_use_of_MLX4_GET(void);
+extern void __buggy_use_of_MLX4_PUT(void);
+
+static bool enable_qos;
+module_param(enable_qos, bool, 0444);
+MODULE_PARM_DESC(enable_qos, "Enable Enhanced QoS support (default: off)");
+
+#define MLX4_GET(dest, source, offset)				      \
+	do {							      \
+		void *__p = (char *) (source) + (offset);	      \
+		__be64 val;                                           \
+		switch (sizeof(dest)) {				      \
+		case 1: (dest) = *(u8 *) __p;	    break;	      \
+		case 2: (dest) = be16_to_cpup(__p); break;	      \
+		case 4: (dest) = be32_to_cpup(__p); break;	      \
+		case 8: val = get_unaligned((__be64 *)__p);           \
+			(dest) = be64_to_cpu(val);  break;            \
+		default: __buggy_use_of_MLX4_GET();		      \
+		}						      \
+	} while (0)
+
+#define MLX4_PUT(dest, source, offset)				      \
+	do {							      \
+		void *__d = ((char *) (dest) + (offset));	      \
+		switch (sizeof(source)) {			      \
+		case 1: *(u8 *) __d = (source);		       break; \
+		case 2:	*(__be16 *) __d = cpu_to_be16(source); break; \
+		case 4:	*(__be32 *) __d = cpu_to_be32(source); break; \
+		case 8:	*(__be64 *) __d = cpu_to_be64(source); break; \
+		default: __buggy_use_of_MLX4_PUT();		      \
+		}						      \
+	} while (0)
+
+static void dump_dev_cap_flags(struct mlx4_dev *dev, u64 flags)
+{
+	static const char *fname[] = {
+		[ 0] = "RC transport",
+		[ 1] = "UC transport",
+		[ 2] = "UD transport",
+		[ 3] = "XRC transport",
+		[ 6] = "SRQ support",
+		[ 7] = "IPoIB checksum offload",
+		[ 8] = "P_Key violation counter",
+		[ 9] = "Q_Key violation counter",
+		[12] = "Dual Port Different Protocol (DPDP) support",
+		[15] = "Big LSO headers",
+		[16] = "MW support",
+		[17] = "APM support",
+		[18] = "Atomic ops support",
+		[19] = "Raw multicast support",
+		[20] = "Address vector port checking support",
+		[21] = "UD multicast support",
+		[30] = "IBoE support",
+		[32] = "Unicast loopback support",
+		[34] = "FCS header control",
+		[37] = "Wake On LAN (port1) support",
+		[38] = "Wake On LAN (port2) support",
+		[40] = "UDP RSS support",
+		[41] = "Unicast VEP steering support",
+		[42] = "Multicast VEP steering support",
+		[48] = "Counters support",
+		[52] = "RSS IP fragments support",
+		[53] = "Port ETS Scheduler support",
+		[55] = "Port link type sensing support",
+		[59] = "Port management change event support",
+		[61] = "64 byte EQE support",
+		[62] = "64 byte CQE support",
+	};
+	int i;
+
+	mlx4_dbg(dev, "DEV_CAP flags:\n");
+	for (i = 0; i < ARRAY_SIZE(fname); ++i)
+		if (fname[i] && (flags & (1LL << i)))
+			mlx4_dbg(dev, "    %s\n", fname[i]);
+}
+
+static void dump_dev_cap_flags2(struct mlx4_dev *dev, u64 flags)
+{
+	static const char * const fname[] = {
+		[0] = "RSS support",
+		[1] = "RSS Toeplitz Hash Function support",
+		[2] = "RSS XOR Hash Function support",
+		[3] = "Device managed flow steering support",
+		[4] = "Automatic MAC reassignment support",
+		[5] = "Time stamping support",
+		[6] = "VST (control vlan insertion/stripping) support",
+		[7] = "FSM (MAC anti-spoofing) support",
+		[8] = "Dynamic QP updates support",
+		[9] = "Device managed flow steering IPoIB support",
+		[10] = "TCP/IP offloads/flow-steering for VXLAN support",
+		[11] = "MAD DEMUX (Secure-Host) support",
+		[12] = "Large cache line (>64B) CQE stride support",
+		[13] = "Large cache line (>64B) EQE stride support",
+		[14] = "Ethernet protocol control support",
+		[15] = "Ethernet Backplane autoneg support",
+		[16] = "CONFIG DEV support",
+		[17] = "Asymmetric EQs support",
+		[18] = "More than 80 VFs support",
+		[19] = "Performance optimized for limited rule configuration flow steering support",
+		[20] = "Recoverable error events support",
+		[21] = "Port Remap support",
+		[22] = "QCN support",
+		[23] = "QP rate limiting support",
+		[24] = "Ethernet Flow control statistics support",
+		[25] = "Granular QoS per VF support",
+		[26] = "Port ETS Scheduler support",
+		[27] = "Port beacon support",
+		[28] = "RX-ALL support",
+		[29] = "802.1ad offload support",
+		[31] = "Modifying loopback source checks using UPDATE_QP support",
+		[32] = "Loopback source checks support",
+		[33] = "RoCEv2 support",
+		[34] = "DMFS Sniffer support (UC & MC)",
+		[35] = "Diag counters per port",
+		[36] = "QinQ VST mode support",
+		[37] = "sl to vl mapping table change event support",
+		[38] = "user MAC support",
+		[39] = "Report driver version to FW support",
+	};
+	int i;
+
+	for (i = 0; i < ARRAY_SIZE(fname); ++i)
+		if (fname[i] && (flags & (1LL << i)))
+			mlx4_dbg(dev, "    %s\n", fname[i]);
+}
+
+int mlx4_MOD_STAT_CFG(struct mlx4_dev *dev, struct mlx4_mod_stat_cfg *cfg)
+{
+	struct mlx4_cmd_mailbox *mailbox;
+	u32 *inbox;
+	int err = 0;
+
+#define MOD_STAT_CFG_IN_SIZE		0x100
+
+#define MOD_STAT_CFG_PG_SZ_M_OFFSET	0x002
+#define MOD_STAT_CFG_PG_SZ_OFFSET	0x003
+
+	mailbox = mlx4_alloc_cmd_mailbox(dev);
+	if (IS_ERR(mailbox))
+		return PTR_ERR(mailbox);
+	inbox = mailbox->buf;
+
+	MLX4_PUT(inbox, cfg->log_pg_sz, MOD_STAT_CFG_PG_SZ_OFFSET);
+	MLX4_PUT(inbox, cfg->log_pg_sz_m, MOD_STAT_CFG_PG_SZ_M_OFFSET);
+
+	err = mlx4_cmd(dev, mailbox->dma, 0, 0, MLX4_CMD_MOD_STAT_CFG,
+			MLX4_CMD_TIME_CLASS_A, MLX4_CMD_NATIVE);
+
+	mlx4_free_cmd_mailbox(dev, mailbox);
+	return err;
+}
+
+int mlx4_QUERY_FUNC(struct mlx4_dev *dev, struct mlx4_func *func, int slave)
+{
+	struct mlx4_cmd_mailbox *mailbox;
+	u32 *outbox;
+	u8 in_modifier;
+	u8 field;
+	u16 field16;
+	int err;
+
+#define QUERY_FUNC_BUS_OFFSET			0x00
+#define QUERY_FUNC_DEVICE_OFFSET		0x01
+#define QUERY_FUNC_FUNCTION_OFFSET		0x01
+#define QUERY_FUNC_PHYSICAL_FUNCTION_OFFSET	0x03
+#define QUERY_FUNC_RSVD_EQS_OFFSET		0x04
+#define QUERY_FUNC_MAX_EQ_OFFSET		0x06
+#define QUERY_FUNC_RSVD_UARS_OFFSET		0x0b
+
+	mailbox = mlx4_alloc_cmd_mailbox(dev);
+	if (IS_ERR(mailbox))
+		return PTR_ERR(mailbox);
+	outbox = mailbox->buf;
+
+	in_modifier = slave;
+
+	err = mlx4_cmd_box(dev, 0, mailbox->dma, in_modifier, 0,
+			   MLX4_CMD_QUERY_FUNC,
+			   MLX4_CMD_TIME_CLASS_A,
+			   MLX4_CMD_NATIVE);
+	if (err)
+		goto out;
+
+	MLX4_GET(field, outbox, QUERY_FUNC_BUS_OFFSET);
+	func->bus = field & 0xf;
+	MLX4_GET(field, outbox, QUERY_FUNC_DEVICE_OFFSET);
+	func->device = field & 0xf1;
+	MLX4_GET(field, outbox, QUERY_FUNC_FUNCTION_OFFSET);
+	func->function = field & 0x7;
+	MLX4_GET(field, outbox, QUERY_FUNC_PHYSICAL_FUNCTION_OFFSET);
+	func->physical_function = field & 0xf;
+	MLX4_GET(field16, outbox, QUERY_FUNC_RSVD_EQS_OFFSET);
+	func->rsvd_eqs = field16 & 0xffff;
+	MLX4_GET(field16, outbox, QUERY_FUNC_MAX_EQ_OFFSET);
+	func->max_eq = field16 & 0xffff;
+	MLX4_GET(field, outbox, QUERY_FUNC_RSVD_UARS_OFFSET);
+	func->rsvd_uars = field & 0x0f;
+
+	mlx4_dbg(dev, "Bus: %d, Device: %d, Function: %d, Physical function: %d, Max EQs: %d, Reserved EQs: %d, Reserved UARs: %d\n",
+		 func->bus, func->device, func->function, func->physical_function,
+		 func->max_eq, func->rsvd_eqs, func->rsvd_uars);
+
+out:
+	mlx4_free_cmd_mailbox(dev, mailbox);
+	return err;
+}
+
+static int mlx4_activate_vst_qinq(struct mlx4_priv *priv, int slave, int port)
+{
+	struct mlx4_vport_oper_state *vp_oper;
+	struct mlx4_vport_state *vp_admin;
+	int err;
+
+	vp_oper = &priv->mfunc.master.vf_oper[slave].vport[port];
+	vp_admin = &priv->mfunc.master.vf_admin[slave].vport[port];
+
+	if (vp_admin->default_vlan != vp_oper->state.default_vlan) {
+		err = __mlx4_register_vlan(&priv->dev, port,
+					   vp_admin->default_vlan,
+					   &vp_oper->vlan_idx);
+		if (err) {
+			vp_oper->vlan_idx = NO_INDX;
+			mlx4_warn(&priv->dev,
+				  "No vlan resources slave %d, port %d\n",
+				  slave, port);
+			return err;
+		}
+		mlx4_dbg(&priv->dev, "alloc vlan %d idx  %d slave %d port %d\n",
+			 (int)(vp_oper->state.default_vlan),
+			 vp_oper->vlan_idx, slave, port);
+	}
+	vp_oper->state.vlan_proto   = vp_admin->vlan_proto;
+	vp_oper->state.default_vlan = vp_admin->default_vlan;
+	vp_oper->state.default_qos  = vp_admin->default_qos;
+
+	return 0;
+}
+
+static int mlx4_handle_vst_qinq(struct mlx4_priv *priv, int slave, int port)
+{
+	struct mlx4_vport_oper_state *vp_oper;
+	struct mlx4_slave_state *slave_state;
+	struct mlx4_vport_state *vp_admin;
+	int err;
+
+	vp_oper = &priv->mfunc.master.vf_oper[slave].vport[port];
+	vp_admin = &priv->mfunc.master.vf_admin[slave].vport[port];
+	slave_state = &priv->mfunc.master.slave_state[slave];
+
+	if ((vp_admin->vlan_proto != htons(ETH_P_8021AD)) ||
+	    (!slave_state->active))
+		return 0;
+
+	if (vp_oper->state.vlan_proto == vp_admin->vlan_proto &&
+	    vp_oper->state.default_vlan == vp_admin->default_vlan &&
+	    vp_oper->state.default_qos == vp_admin->default_qos)
+		return 0;
+
+	if (!slave_state->vst_qinq_supported) {
+		/* Warn and revert the request to set vst QinQ mode */
+		vp_admin->vlan_proto   = vp_oper->state.vlan_proto;
+		vp_admin->default_vlan = vp_oper->state.default_vlan;
+		vp_admin->default_qos  = vp_oper->state.default_qos;
+
+		mlx4_warn(&priv->dev,
+			  "Slave %d does not support VST QinQ mode\n", slave);
+		return 0;
+	}
+
+	err = mlx4_activate_vst_qinq(priv, slave, port);
+	return err;
+}
+
+int mlx4_QUERY_FUNC_CAP_wrapper(struct mlx4_dev *dev, int slave,
+				struct mlx4_vhcr *vhcr,
+				struct mlx4_cmd_mailbox *inbox,
+				struct mlx4_cmd_mailbox *outbox,
+				struct mlx4_cmd_info *cmd)
+{
+	struct mlx4_priv *priv = mlx4_priv(dev);
+	u8	field, port;
+	u32	size, proxy_qp, qkey;
+	int	err = 0;
+	struct mlx4_func func;
+
+#define QUERY_FUNC_CAP_FLAGS_OFFSET		0x0
+#define QUERY_FUNC_CAP_NUM_PORTS_OFFSET		0x1
+#define QUERY_FUNC_CAP_PF_BHVR_OFFSET		0x4
+#define QUERY_FUNC_CAP_FMR_OFFSET		0x8
+#define QUERY_FUNC_CAP_QP_QUOTA_OFFSET_DEP	0x10
+#define QUERY_FUNC_CAP_CQ_QUOTA_OFFSET_DEP	0x14
+#define QUERY_FUNC_CAP_SRQ_QUOTA_OFFSET_DEP	0x18
+#define QUERY_FUNC_CAP_MPT_QUOTA_OFFSET_DEP	0x20
+#define QUERY_FUNC_CAP_MTT_QUOTA_OFFSET_DEP	0x24
+#define QUERY_FUNC_CAP_MCG_QUOTA_OFFSET_DEP	0x28
+#define QUERY_FUNC_CAP_MAX_EQ_OFFSET		0x2c
+#define QUERY_FUNC_CAP_RESERVED_EQ_OFFSET	0x30
+#define QUERY_FUNC_CAP_QP_RESD_LKEY_OFFSET	0x48
+
+#define QUERY_FUNC_CAP_QP_QUOTA_OFFSET		0x50
+#define QUERY_FUNC_CAP_CQ_QUOTA_OFFSET		0x54
+#define QUERY_FUNC_CAP_SRQ_QUOTA_OFFSET		0x58
+#define QUERY_FUNC_CAP_MPT_QUOTA_OFFSET		0x60
+#define QUERY_FUNC_CAP_MTT_QUOTA_OFFSET		0x64
+#define QUERY_FUNC_CAP_MCG_QUOTA_OFFSET		0x68
+
+#define QUERY_FUNC_CAP_EXTRA_FLAGS_OFFSET	0x6c
+
+#define QUERY_FUNC_CAP_FMR_FLAG			0x80
+#define QUERY_FUNC_CAP_FLAG_RDMA		0x40
+#define QUERY_FUNC_CAP_FLAG_ETH			0x80
+#define QUERY_FUNC_CAP_FLAG_QUOTAS		0x10
+#define QUERY_FUNC_CAP_FLAG_RESD_LKEY		0x08
+#define QUERY_FUNC_CAP_FLAG_VALID_MAILBOX	0x04
+
+#define QUERY_FUNC_CAP_EXTRA_FLAGS_BF_QP_ALLOC_FLAG	(1UL << 31)
+#define QUERY_FUNC_CAP_EXTRA_FLAGS_A0_QP_ALLOC_FLAG	(1UL << 30)
+
+/* when opcode modifier = 1 */
+#define QUERY_FUNC_CAP_PHYS_PORT_OFFSET		0x3
+#define QUERY_FUNC_CAP_PRIV_VF_QKEY_OFFSET	0x4
+#define QUERY_FUNC_CAP_FLAGS0_OFFSET		0x8
+#define QUERY_FUNC_CAP_FLAGS1_OFFSET		0xc
+
+#define QUERY_FUNC_CAP_QP0_TUNNEL		0x10
+#define QUERY_FUNC_CAP_QP0_PROXY		0x14
+#define QUERY_FUNC_CAP_QP1_TUNNEL		0x18
+#define QUERY_FUNC_CAP_QP1_PROXY		0x1c
+#define QUERY_FUNC_CAP_PHYS_PORT_ID		0x28
+
+#define QUERY_FUNC_CAP_FLAGS1_FORCE_MAC		0x40
+#define QUERY_FUNC_CAP_FLAGS1_FORCE_VLAN	0x80
+#define QUERY_FUNC_CAP_FLAGS1_NIC_INFO			0x10
+#define QUERY_FUNC_CAP_VF_ENABLE_QP0		0x08
+
+#define QUERY_FUNC_CAP_FLAGS0_FORCE_PHY_WQE_GID 0x80
+#define QUERY_FUNC_CAP_PHV_BIT			0x40
+#define QUERY_FUNC_CAP_VLAN_OFFLOAD_DISABLE	0x20
+
+#define QUERY_FUNC_CAP_SUPPORTS_VST_QINQ	BIT(30)
+#define QUERY_FUNC_CAP_SUPPORTS_NON_POWER_OF_2_NUM_EQS BIT(31)
+
+	if (vhcr->op_modifier == 1) {
+		struct mlx4_active_ports actv_ports =
+			mlx4_get_active_ports(dev, slave);
+		int converted_port = mlx4_slave_convert_port(
+				dev, slave, vhcr->in_modifier);
+		struct mlx4_vport_oper_state *vp_oper;
+
+		if (converted_port < 0)
+			return -EINVAL;
+
+		vhcr->in_modifier = converted_port;
+		/* phys-port = logical-port */
+		field = vhcr->in_modifier -
+			find_first_bit(actv_ports.ports, dev->caps.num_ports);
+		MLX4_PUT(outbox->buf, field, QUERY_FUNC_CAP_PHYS_PORT_OFFSET);
+
+		port = vhcr->in_modifier;
+		proxy_qp = dev->phys_caps.base_proxy_sqpn + 8 * slave + port - 1;
+
+		/* Set nic_info bit to mark new fields support */
+		field  = QUERY_FUNC_CAP_FLAGS1_NIC_INFO;
+
+		if (mlx4_vf_smi_enabled(dev, slave, port) &&
+		    !mlx4_get_parav_qkey(dev, proxy_qp, &qkey)) {
+			field |= QUERY_FUNC_CAP_VF_ENABLE_QP0;
+			MLX4_PUT(outbox->buf, qkey,
+				 QUERY_FUNC_CAP_PRIV_VF_QKEY_OFFSET);
+		}
+		MLX4_PUT(outbox->buf, field, QUERY_FUNC_CAP_FLAGS1_OFFSET);
+
+		/* size is now the QP number */
+		size = dev->phys_caps.base_tunnel_sqpn + 8 * slave + port - 1;
+		MLX4_PUT(outbox->buf, size, QUERY_FUNC_CAP_QP0_TUNNEL);
+
+		size += 2;
+		MLX4_PUT(outbox->buf, size, QUERY_FUNC_CAP_QP1_TUNNEL);
+
+		MLX4_PUT(outbox->buf, proxy_qp, QUERY_FUNC_CAP_QP0_PROXY);
+		proxy_qp += 2;
+		MLX4_PUT(outbox->buf, proxy_qp, QUERY_FUNC_CAP_QP1_PROXY);
+
+		MLX4_PUT(outbox->buf, dev->caps.phys_port_id[vhcr->in_modifier],
+			 QUERY_FUNC_CAP_PHYS_PORT_ID);
+
+		vp_oper = &priv->mfunc.master.vf_oper[slave].vport[port];
+		err = mlx4_handle_vst_qinq(priv, slave, port);
+		if (err)
+			return err;
+
+		field = 0;
+		if (dev->caps.phv_bit[port])
+			field |= QUERY_FUNC_CAP_PHV_BIT;
+		if (vp_oper->state.vlan_proto == htons(ETH_P_8021AD))
+			field |= QUERY_FUNC_CAP_VLAN_OFFLOAD_DISABLE;
+		MLX4_PUT(outbox->buf, field, QUERY_FUNC_CAP_FLAGS0_OFFSET);
+
+	} else if (vhcr->op_modifier == 0) {
+		struct mlx4_active_ports actv_ports =
+			mlx4_get_active_ports(dev, slave);
+		struct mlx4_slave_state *slave_state =
+			&priv->mfunc.master.slave_state[slave];
+
+		/* enable rdma and ethernet interfaces, new quota locations,
+		 * and reserved lkey
+		 */
+		field = (QUERY_FUNC_CAP_FLAG_ETH | QUERY_FUNC_CAP_FLAG_RDMA |
+			 QUERY_FUNC_CAP_FLAG_QUOTAS | QUERY_FUNC_CAP_FLAG_VALID_MAILBOX |
+			 QUERY_FUNC_CAP_FLAG_RESD_LKEY);
+		MLX4_PUT(outbox->buf, field, QUERY_FUNC_CAP_FLAGS_OFFSET);
+
+		field = min(
+			bitmap_weight(actv_ports.ports, dev->caps.num_ports),
+			dev->caps.num_ports);
+		MLX4_PUT(outbox->buf, field, QUERY_FUNC_CAP_NUM_PORTS_OFFSET);
+
+		size = dev->caps.function_caps; /* set PF behaviours */
+		MLX4_PUT(outbox->buf, size, QUERY_FUNC_CAP_PF_BHVR_OFFSET);
+
+		field = 0; /* protected FMR support not available as yet */
+		MLX4_PUT(outbox->buf, field, QUERY_FUNC_CAP_FMR_OFFSET);
+
+		size = priv->mfunc.master.res_tracker.res_alloc[RES_QP].quota[slave];
+		MLX4_PUT(outbox->buf, size, QUERY_FUNC_CAP_QP_QUOTA_OFFSET);
+		size = dev->caps.num_qps;
+		MLX4_PUT(outbox->buf, size, QUERY_FUNC_CAP_QP_QUOTA_OFFSET_DEP);
+
+		size = priv->mfunc.master.res_tracker.res_alloc[RES_SRQ].quota[slave];
+		MLX4_PUT(outbox->buf, size, QUERY_FUNC_CAP_SRQ_QUOTA_OFFSET);
+		size = dev->caps.num_srqs;
+		MLX4_PUT(outbox->buf, size, QUERY_FUNC_CAP_SRQ_QUOTA_OFFSET_DEP);
+
+		size = priv->mfunc.master.res_tracker.res_alloc[RES_CQ].quota[slave];
+		MLX4_PUT(outbox->buf, size, QUERY_FUNC_CAP_CQ_QUOTA_OFFSET);
+		size = dev->caps.num_cqs;
+		MLX4_PUT(outbox->buf, size, QUERY_FUNC_CAP_CQ_QUOTA_OFFSET_DEP);
+
+		if (!(dev->caps.flags2 & MLX4_DEV_CAP_FLAG2_SYS_EQS) ||
+		    mlx4_QUERY_FUNC(dev, &func, slave)) {
+			size = vhcr->in_modifier &
+				QUERY_FUNC_CAP_SUPPORTS_NON_POWER_OF_2_NUM_EQS ?
+				dev->caps.num_eqs :
+				rounddown_pow_of_two(dev->caps.num_eqs);
+			MLX4_PUT(outbox->buf, size, QUERY_FUNC_CAP_MAX_EQ_OFFSET);
+			size = dev->caps.reserved_eqs;
+			MLX4_PUT(outbox->buf, size, QUERY_FUNC_CAP_RESERVED_EQ_OFFSET);
+		} else {
+			size = vhcr->in_modifier &
+				QUERY_FUNC_CAP_SUPPORTS_NON_POWER_OF_2_NUM_EQS ?
+				func.max_eq :
+				rounddown_pow_of_two(func.max_eq);
+			MLX4_PUT(outbox->buf, size, QUERY_FUNC_CAP_MAX_EQ_OFFSET);
+			size = func.rsvd_eqs;
+			MLX4_PUT(outbox->buf, size, QUERY_FUNC_CAP_RESERVED_EQ_OFFSET);
+		}
+
+		size = priv->mfunc.master.res_tracker.res_alloc[RES_MPT].quota[slave];
+		MLX4_PUT(outbox->buf, size, QUERY_FUNC_CAP_MPT_QUOTA_OFFSET);
+		size = dev->caps.num_mpts;
+		MLX4_PUT(outbox->buf, size, QUERY_FUNC_CAP_MPT_QUOTA_OFFSET_DEP);
+
+		size = priv->mfunc.master.res_tracker.res_alloc[RES_MTT].quota[slave];
+		MLX4_PUT(outbox->buf, size, QUERY_FUNC_CAP_MTT_QUOTA_OFFSET);
+		size = dev->caps.num_mtts;
+		MLX4_PUT(outbox->buf, size, QUERY_FUNC_CAP_MTT_QUOTA_OFFSET_DEP);
+
+		size = dev->caps.num_mgms + dev->caps.num_amgms;
+		MLX4_PUT(outbox->buf, size, QUERY_FUNC_CAP_MCG_QUOTA_OFFSET);
+		MLX4_PUT(outbox->buf, size, QUERY_FUNC_CAP_MCG_QUOTA_OFFSET_DEP);
+
+		size = QUERY_FUNC_CAP_EXTRA_FLAGS_BF_QP_ALLOC_FLAG |
+			QUERY_FUNC_CAP_EXTRA_FLAGS_A0_QP_ALLOC_FLAG;
+		MLX4_PUT(outbox->buf, size, QUERY_FUNC_CAP_EXTRA_FLAGS_OFFSET);
+
+		size = dev->caps.reserved_lkey + ((slave << 8) & 0xFF00);
+		MLX4_PUT(outbox->buf, size, QUERY_FUNC_CAP_QP_RESD_LKEY_OFFSET);
+
+		if (vhcr->in_modifier & QUERY_FUNC_CAP_SUPPORTS_VST_QINQ)
+			slave_state->vst_qinq_supported = true;
+
+	} else
+		err = -EINVAL;
+
+	return err;
+}
+
+int mlx4_QUERY_FUNC_CAP(struct mlx4_dev *dev, u8 gen_or_port,
+			struct mlx4_func_cap *func_cap)
+{
+	struct mlx4_cmd_mailbox *mailbox;
+	u32			*outbox;
+	u8			field, op_modifier;
+	u32			size, qkey;
+	int			err = 0, quotas = 0;
+	u32                     in_modifier;
+	u32			slave_caps;
+
+	op_modifier = !!gen_or_port; /* 0 = general, 1 = logical port */
+	slave_caps = QUERY_FUNC_CAP_SUPPORTS_VST_QINQ |
+		QUERY_FUNC_CAP_SUPPORTS_NON_POWER_OF_2_NUM_EQS;
+	in_modifier = op_modifier ? gen_or_port : slave_caps;
+
+	mailbox = mlx4_alloc_cmd_mailbox(dev);
+	if (IS_ERR(mailbox))
+		return PTR_ERR(mailbox);
+
+	err = mlx4_cmd_box(dev, 0, mailbox->dma, in_modifier, op_modifier,
+			   MLX4_CMD_QUERY_FUNC_CAP,
+			   MLX4_CMD_TIME_CLASS_A, MLX4_CMD_WRAPPED);
+	if (err)
+		goto out;
+
+	outbox = mailbox->buf;
+
+	if (!op_modifier) {
+		MLX4_GET(field, outbox, QUERY_FUNC_CAP_FLAGS_OFFSET);
+		if (!(field & (QUERY_FUNC_CAP_FLAG_ETH | QUERY_FUNC_CAP_FLAG_RDMA))) {
+			mlx4_err(dev, "The host supports neither eth nor rdma interfaces\n");
+			err = -EPROTONOSUPPORT;
+			goto out;
+		}
+		func_cap->flags = field;
+		quotas = !!(func_cap->flags & QUERY_FUNC_CAP_FLAG_QUOTAS);
+
+		MLX4_GET(field, outbox, QUERY_FUNC_CAP_NUM_PORTS_OFFSET);
+		func_cap->num_ports = field;
+
+		MLX4_GET(size, outbox, QUERY_FUNC_CAP_PF_BHVR_OFFSET);
+		func_cap->pf_context_behaviour = size;
+
+		if (quotas) {
+			MLX4_GET(size, outbox, QUERY_FUNC_CAP_QP_QUOTA_OFFSET);
+			func_cap->qp_quota = size & 0xFFFFFF;
+
+			MLX4_GET(size, outbox, QUERY_FUNC_CAP_SRQ_QUOTA_OFFSET);
+			func_cap->srq_quota = size & 0xFFFFFF;
+
+			MLX4_GET(size, outbox, QUERY_FUNC_CAP_CQ_QUOTA_OFFSET);
+			func_cap->cq_quota = size & 0xFFFFFF;
+
+			MLX4_GET(size, outbox, QUERY_FUNC_CAP_MPT_QUOTA_OFFSET);
+			func_cap->mpt_quota = size & 0xFFFFFF;
+
+			MLX4_GET(size, outbox, QUERY_FUNC_CAP_MTT_QUOTA_OFFSET);
+			func_cap->mtt_quota = size & 0xFFFFFF;
+
+			MLX4_GET(size, outbox, QUERY_FUNC_CAP_MCG_QUOTA_OFFSET);
+			func_cap->mcg_quota = size & 0xFFFFFF;
+
+		} else {
+			MLX4_GET(size, outbox, QUERY_FUNC_CAP_QP_QUOTA_OFFSET_DEP);
+			func_cap->qp_quota = size & 0xFFFFFF;
+
+			MLX4_GET(size, outbox, QUERY_FUNC_CAP_SRQ_QUOTA_OFFSET_DEP);
+			func_cap->srq_quota = size & 0xFFFFFF;
+
+			MLX4_GET(size, outbox, QUERY_FUNC_CAP_CQ_QUOTA_OFFSET_DEP);
+			func_cap->cq_quota = size & 0xFFFFFF;
+
+			MLX4_GET(size, outbox, QUERY_FUNC_CAP_MPT_QUOTA_OFFSET_DEP);
+			func_cap->mpt_quota = size & 0xFFFFFF;
+
+			MLX4_GET(size, outbox, QUERY_FUNC_CAP_MTT_QUOTA_OFFSET_DEP);
+			func_cap->mtt_quota = size & 0xFFFFFF;
+
+			MLX4_GET(size, outbox, QUERY_FUNC_CAP_MCG_QUOTA_OFFSET_DEP);
+			func_cap->mcg_quota = size & 0xFFFFFF;
+		}
+		MLX4_GET(size, outbox, QUERY_FUNC_CAP_MAX_EQ_OFFSET);
+		func_cap->max_eq = size & 0xFFFFFF;
+
+		MLX4_GET(size, outbox, QUERY_FUNC_CAP_RESERVED_EQ_OFFSET);
+		func_cap->reserved_eq = size & 0xFFFFFF;
+
+		if (func_cap->flags & QUERY_FUNC_CAP_FLAG_RESD_LKEY) {
+			MLX4_GET(size, outbox, QUERY_FUNC_CAP_QP_RESD_LKEY_OFFSET);
+			func_cap->reserved_lkey = size;
+		} else {
+			func_cap->reserved_lkey = 0;
+		}
+
+		func_cap->extra_flags = 0;
+
+		/* Mailbox data from 0x6c and onward should only be treated if
+		 * QUERY_FUNC_CAP_FLAG_VALID_MAILBOX is set in func_cap->flags
+		 */
+		if (func_cap->flags & QUERY_FUNC_CAP_FLAG_VALID_MAILBOX) {
+			MLX4_GET(size, outbox, QUERY_FUNC_CAP_EXTRA_FLAGS_OFFSET);
+			if (size & QUERY_FUNC_CAP_EXTRA_FLAGS_BF_QP_ALLOC_FLAG)
+				func_cap->extra_flags |= MLX4_QUERY_FUNC_FLAGS_BF_RES_QP;
+			if (size & QUERY_FUNC_CAP_EXTRA_FLAGS_A0_QP_ALLOC_FLAG)
+				func_cap->extra_flags |= MLX4_QUERY_FUNC_FLAGS_A0_RES_QP;
+		}
+
+		goto out;
+	}
+
+	/* logical port query */
+	if (gen_or_port > dev->caps.num_ports) {
+		err = -EINVAL;
+		goto out;
+	}
+
+	MLX4_GET(func_cap->flags1, outbox, QUERY_FUNC_CAP_FLAGS1_OFFSET);
+	if (dev->caps.port_type[gen_or_port] == MLX4_PORT_TYPE_ETH) {
+		if (func_cap->flags1 & QUERY_FUNC_CAP_FLAGS1_FORCE_VLAN) {
+			mlx4_err(dev, "VLAN is enforced on this port\n");
+			err = -EPROTONOSUPPORT;
+			goto out;
+		}
+
+		if (func_cap->flags1 & QUERY_FUNC_CAP_FLAGS1_FORCE_MAC) {
+			mlx4_err(dev, "Force mac is enabled on this port\n");
+			err = -EPROTONOSUPPORT;
+			goto out;
+		}
+	} else if (dev->caps.port_type[gen_or_port] == MLX4_PORT_TYPE_IB) {
+		MLX4_GET(field, outbox, QUERY_FUNC_CAP_FLAGS0_OFFSET);
+		if (field & QUERY_FUNC_CAP_FLAGS0_FORCE_PHY_WQE_GID) {
+			mlx4_err(dev, "phy_wqe_gid is enforced on this ib port\n");
+			err = -EPROTONOSUPPORT;
+			goto out;
+		}
+	}
+
+	MLX4_GET(field, outbox, QUERY_FUNC_CAP_PHYS_PORT_OFFSET);
+	func_cap->physical_port = field;
+	if (func_cap->physical_port != gen_or_port) {
+		err = -EINVAL;
+		goto out;
+	}
+
+	if (func_cap->flags1 & QUERY_FUNC_CAP_VF_ENABLE_QP0) {
+		MLX4_GET(qkey, outbox, QUERY_FUNC_CAP_PRIV_VF_QKEY_OFFSET);
+		func_cap->spec_qps.qp0_qkey = qkey;
+	} else {
+		func_cap->spec_qps.qp0_qkey = 0;
+	}
+
+	MLX4_GET(size, outbox, QUERY_FUNC_CAP_QP0_TUNNEL);
+	func_cap->spec_qps.qp0_tunnel = size & 0xFFFFFF;
+
+	MLX4_GET(size, outbox, QUERY_FUNC_CAP_QP0_PROXY);
+	func_cap->spec_qps.qp0_proxy = size & 0xFFFFFF;
+
+	MLX4_GET(size, outbox, QUERY_FUNC_CAP_QP1_TUNNEL);
+	func_cap->spec_qps.qp1_tunnel = size & 0xFFFFFF;
+
+	MLX4_GET(size, outbox, QUERY_FUNC_CAP_QP1_PROXY);
+	func_cap->spec_qps.qp1_proxy = size & 0xFFFFFF;
+
+	if (func_cap->flags1 & QUERY_FUNC_CAP_FLAGS1_NIC_INFO)
+		MLX4_GET(func_cap->phys_port_id, outbox,
+			 QUERY_FUNC_CAP_PHYS_PORT_ID);
+
+	MLX4_GET(func_cap->flags0, outbox, QUERY_FUNC_CAP_FLAGS0_OFFSET);
+
+	/* All other resources are allocated by the master, but we still report
+	 * 'num' and 'reserved' capabilities as follows:
+	 * - num remains the maximum resource index
+	 * - 'num - reserved' is the total available objects of a resource, but
+	 *   resource indices may be less than 'reserved'
+	 * TODO: set per-resource quotas */
+
+out:
+	mlx4_free_cmd_mailbox(dev, mailbox);
+
+	return err;
+}
+
+static void disable_unsupported_roce_caps(void *buf);
+
+int mlx4_QUERY_DEV_CAP(struct mlx4_dev *dev, struct mlx4_dev_cap *dev_cap)
+{
+	struct mlx4_cmd_mailbox *mailbox;
+	u32 *outbox;
+	u8 field;
+	u32 field32, flags, ext_flags;
+	u16 size;
+	u16 stat_rate;
+	int err;
+	int i;
+
+#define QUERY_DEV_CAP_OUT_SIZE		       0x100
+#define QUERY_DEV_CAP_MAX_SRQ_SZ_OFFSET		0x10
+#define QUERY_DEV_CAP_MAX_QP_SZ_OFFSET		0x11
+#define QUERY_DEV_CAP_RSVD_QP_OFFSET		0x12
+#define QUERY_DEV_CAP_MAX_QP_OFFSET		0x13
+#define QUERY_DEV_CAP_RSVD_SRQ_OFFSET		0x14
+#define QUERY_DEV_CAP_MAX_SRQ_OFFSET		0x15
+#define QUERY_DEV_CAP_RSVD_EEC_OFFSET		0x16
+#define QUERY_DEV_CAP_MAX_EEC_OFFSET		0x17
+#define QUERY_DEV_CAP_MAX_CQ_SZ_OFFSET		0x19
+#define QUERY_DEV_CAP_RSVD_CQ_OFFSET		0x1a
+#define QUERY_DEV_CAP_MAX_CQ_OFFSET		0x1b
+#define QUERY_DEV_CAP_MAX_MPT_OFFSET		0x1d
+#define QUERY_DEV_CAP_RSVD_EQ_OFFSET		0x1e
+#define QUERY_DEV_CAP_MAX_EQ_OFFSET		0x1f
+#define QUERY_DEV_CAP_RSVD_MTT_OFFSET		0x20
+#define QUERY_DEV_CAP_MAX_MRW_SZ_OFFSET		0x21
+#define QUERY_DEV_CAP_RSVD_MRW_OFFSET		0x22
+#define QUERY_DEV_CAP_MAX_MTT_SEG_OFFSET	0x23
+#define QUERY_DEV_CAP_NUM_SYS_EQ_OFFSET		0x26
+#define QUERY_DEV_CAP_MAX_AV_OFFSET		0x27
+#define QUERY_DEV_CAP_MAX_REQ_QP_OFFSET		0x29
+#define QUERY_DEV_CAP_MAX_RES_QP_OFFSET		0x2b
+#define QUERY_DEV_CAP_MAX_GSO_OFFSET		0x2d
+#define QUERY_DEV_CAP_RSS_OFFSET		0x2e
+#define QUERY_DEV_CAP_MAX_RDMA_OFFSET		0x2f
+#define QUERY_DEV_CAP_RSZ_SRQ_OFFSET		0x33
+#define QUERY_DEV_CAP_PORT_BEACON_OFFSET	0x34
+#define QUERY_DEV_CAP_ACK_DELAY_OFFSET		0x35
+#define QUERY_DEV_CAP_MTU_WIDTH_OFFSET		0x36
+#define QUERY_DEV_CAP_VL_PORT_OFFSET		0x37
+#define QUERY_DEV_CAP_MAX_MSG_SZ_OFFSET		0x38
+#define QUERY_DEV_CAP_MAX_GID_OFFSET		0x3b
+#define QUERY_DEV_CAP_RATE_SUPPORT_OFFSET	0x3c
+#define QUERY_DEV_CAP_CQ_TS_SUPPORT_OFFSET	0x3e
+#define QUERY_DEV_CAP_MAX_PKEY_OFFSET		0x3f
+#define QUERY_DEV_CAP_EXT_FLAGS_OFFSET		0x40
+#define QUERY_DEV_CAP_WOL_OFFSET		0x43
+#define QUERY_DEV_CAP_FLAGS_OFFSET		0x44
+#define QUERY_DEV_CAP_RSVD_UAR_OFFSET		0x48
+#define QUERY_DEV_CAP_UAR_SZ_OFFSET		0x49
+#define QUERY_DEV_CAP_PAGE_SZ_OFFSET		0x4b
+#define QUERY_DEV_CAP_BF_OFFSET			0x4c
+#define QUERY_DEV_CAP_LOG_BF_REG_SZ_OFFSET	0x4d
+#define QUERY_DEV_CAP_LOG_MAX_BF_REGS_PER_PAGE_OFFSET	0x4e
+#define QUERY_DEV_CAP_LOG_MAX_BF_PAGES_OFFSET	0x4f
+#define QUERY_DEV_CAP_MAX_SG_SQ_OFFSET		0x51
+#define QUERY_DEV_CAP_MAX_DESC_SZ_SQ_OFFSET	0x52
+#define QUERY_DEV_CAP_MAX_SG_RQ_OFFSET		0x55
+#define QUERY_DEV_CAP_MAX_DESC_SZ_RQ_OFFSET	0x56
+#define QUERY_DEV_CAP_USER_MAC_EN_OFFSET	0x5C
+#define QUERY_DEV_CAP_SVLAN_BY_QP_OFFSET	0x5D
+#define QUERY_DEV_CAP_MAX_QP_MCG_OFFSET		0x61
+#define QUERY_DEV_CAP_RSVD_MCG_OFFSET		0x62
+#define QUERY_DEV_CAP_MAX_MCG_OFFSET		0x63
+#define QUERY_DEV_CAP_RSVD_PD_OFFSET		0x64
+#define QUERY_DEV_CAP_MAX_PD_OFFSET		0x65
+#define QUERY_DEV_CAP_RSVD_XRC_OFFSET		0x66
+#define QUERY_DEV_CAP_MAX_XRC_OFFSET		0x67
+#define QUERY_DEV_CAP_MAX_COUNTERS_OFFSET	0x68
+#define QUERY_DEV_CAP_PORT_FLOWSTATS_COUNTERS_OFFSET	0x70
+#define QUERY_DEV_CAP_EXT_2_FLAGS_OFFSET	0x70
+#define QUERY_DEV_CAP_FLOW_STEERING_IPOIB_OFFSET	0x74
+#define QUERY_DEV_CAP_FLOW_STEERING_RANGE_EN_OFFSET	0x76
+#define QUERY_DEV_CAP_FLOW_STEERING_MAX_QP_OFFSET	0x77
+#define QUERY_DEV_CAP_SL2VL_EVENT_OFFSET	0x78
+#define QUERY_DEV_CAP_CQ_EQ_CACHE_LINE_STRIDE	0x7a
+#define QUERY_DEV_CAP_ECN_QCN_VER_OFFSET	0x7b
+#define QUERY_DEV_CAP_RDMARC_ENTRY_SZ_OFFSET	0x80
+#define QUERY_DEV_CAP_QPC_ENTRY_SZ_OFFSET	0x82
+#define QUERY_DEV_CAP_AUX_ENTRY_SZ_OFFSET	0x84
+#define QUERY_DEV_CAP_ALTC_ENTRY_SZ_OFFSET	0x86
+#define QUERY_DEV_CAP_EQC_ENTRY_SZ_OFFSET	0x88
+#define QUERY_DEV_CAP_CQC_ENTRY_SZ_OFFSET	0x8a
+#define QUERY_DEV_CAP_SRQ_ENTRY_SZ_OFFSET	0x8c
+#define QUERY_DEV_CAP_C_MPT_ENTRY_SZ_OFFSET	0x8e
+#define QUERY_DEV_CAP_MTT_ENTRY_SZ_OFFSET	0x90
+#define QUERY_DEV_CAP_D_MPT_ENTRY_SZ_OFFSET	0x92
+#define QUERY_DEV_CAP_BMME_FLAGS_OFFSET		0x94
+#define QUERY_DEV_CAP_CONFIG_DEV_OFFSET		0x94
+#define QUERY_DEV_CAP_PHV_EN_OFFSET		0x96
+#define QUERY_DEV_CAP_RSVD_LKEY_OFFSET		0x98
+#define QUERY_DEV_CAP_MAX_ICM_SZ_OFFSET		0xa0
+#define QUERY_DEV_CAP_ETH_BACKPL_OFFSET		0x9c
+#define QUERY_DEV_CAP_DIAG_RPRT_PER_PORT	0x9c
+#define QUERY_DEV_CAP_FW_REASSIGN_MAC		0x9d
+#define QUERY_DEV_CAP_VXLAN			0x9e
+#define QUERY_DEV_CAP_MAD_DEMUX_OFFSET		0xb0
+#define QUERY_DEV_CAP_DMFS_HIGH_RATE_QPN_BASE_OFFSET	0xa8
+#define QUERY_DEV_CAP_DMFS_HIGH_RATE_QPN_RANGE_OFFSET	0xac
+#define QUERY_DEV_CAP_MAP_CLOCK_TO_USER 0xc1
+#define QUERY_DEV_CAP_QP_RATE_LIMIT_NUM_OFFSET	0xcc
+#define QUERY_DEV_CAP_QP_RATE_LIMIT_MAX_OFFSET	0xd0
+#define QUERY_DEV_CAP_QP_RATE_LIMIT_MIN_OFFSET	0xd2
+#define QUERY_DEV_CAP_HEALTH_BUFFER_ADDRESS_OFFSET	0xe4
+
+	dev_cap->flags2 = 0;
+	mailbox = mlx4_alloc_cmd_mailbox(dev);
+	if (IS_ERR(mailbox))
+		return PTR_ERR(mailbox);
+	outbox = mailbox->buf;
+
+	err = mlx4_cmd_box(dev, 0, mailbox->dma, 0, 0, MLX4_CMD_QUERY_DEV_CAP,
+			   MLX4_CMD_TIME_CLASS_A, MLX4_CMD_NATIVE);
+	if (err)
+		goto out;
+
+	if (mlx4_is_mfunc(dev))
+		disable_unsupported_roce_caps(outbox);
+	MLX4_GET(field, outbox, QUERY_DEV_CAP_MAP_CLOCK_TO_USER);
+	dev_cap->map_clock_to_user = field & 0x80;
+	MLX4_GET(field, outbox, QUERY_DEV_CAP_RSVD_QP_OFFSET);
+	dev_cap->reserved_qps = 1 << (field & 0xf);
+	MLX4_GET(field, outbox, QUERY_DEV_CAP_MAX_QP_OFFSET);
+	dev_cap->max_qps = 1 << (field & 0x1f);
+	MLX4_GET(field, outbox, QUERY_DEV_CAP_RSVD_SRQ_OFFSET);
+	dev_cap->reserved_srqs = 1 << (field >> 4);
+	MLX4_GET(field, outbox, QUERY_DEV_CAP_MAX_SRQ_OFFSET);
+	dev_cap->max_srqs = 1 << (field & 0x1f);
+	MLX4_GET(field, outbox, QUERY_DEV_CAP_MAX_CQ_SZ_OFFSET);
+	dev_cap->max_cq_sz = 1 << field;
+	MLX4_GET(field, outbox, QUERY_DEV_CAP_RSVD_CQ_OFFSET);
+	dev_cap->reserved_cqs = 1 << (field & 0xf);
+	MLX4_GET(field, outbox, QUERY_DEV_CAP_MAX_CQ_OFFSET);
+	dev_cap->max_cqs = 1 << (field & 0x1f);
+	MLX4_GET(field, outbox, QUERY_DEV_CAP_MAX_MPT_OFFSET);
+	dev_cap->max_mpts = 1 << (field & 0x3f);
+	MLX4_GET(field, outbox, QUERY_DEV_CAP_RSVD_EQ_OFFSET);
+	dev_cap->reserved_eqs = 1 << (field & 0xf);
+	MLX4_GET(field, outbox, QUERY_DEV_CAP_MAX_EQ_OFFSET);
+	dev_cap->max_eqs = 1 << (field & 0xf);
+	MLX4_GET(field, outbox, QUERY_DEV_CAP_RSVD_MTT_OFFSET);
+	dev_cap->reserved_mtts = 1 << (field >> 4);
+	MLX4_GET(field, outbox, QUERY_DEV_CAP_RSVD_MRW_OFFSET);
+	dev_cap->reserved_mrws = 1 << (field & 0xf);
+	MLX4_GET(size, outbox, QUERY_DEV_CAP_NUM_SYS_EQ_OFFSET);
+	dev_cap->num_sys_eqs = size & 0xfff;
+	MLX4_GET(field, outbox, QUERY_DEV_CAP_MAX_REQ_QP_OFFSET);
+	dev_cap->max_requester_per_qp = 1 << (field & 0x3f);
+	MLX4_GET(field, outbox, QUERY_DEV_CAP_MAX_RES_QP_OFFSET);
+	dev_cap->max_responder_per_qp = 1 << (field & 0x3f);
+	MLX4_GET(field, outbox, QUERY_DEV_CAP_MAX_GSO_OFFSET);
+	field &= 0x1f;
+	if (!field)
+		dev_cap->max_gso_sz = 0;
+	else
+		dev_cap->max_gso_sz = 1 << field;
+
+	MLX4_GET(field, outbox, QUERY_DEV_CAP_RSS_OFFSET);
+	if (field & 0x20)
+		dev_cap->flags2 |= MLX4_DEV_CAP_FLAG2_RSS_XOR;
+	if (field & 0x10)
+		dev_cap->flags2 |= MLX4_DEV_CAP_FLAG2_RSS_TOP;
+	field &= 0xf;
+	if (field) {
+		dev_cap->flags2 |= MLX4_DEV_CAP_FLAG2_RSS;
+		dev_cap->max_rss_tbl_sz = 1 << field;
+	} else
+		dev_cap->max_rss_tbl_sz = 0;
+	MLX4_GET(field, outbox, QUERY_DEV_CAP_MAX_RDMA_OFFSET);
+	dev_cap->max_rdma_global = 1 << (field & 0x3f);
+	MLX4_GET(field, outbox, QUERY_DEV_CAP_ACK_DELAY_OFFSET);
+	dev_cap->local_ca_ack_delay = field & 0x1f;
+	MLX4_GET(field, outbox, QUERY_DEV_CAP_VL_PORT_OFFSET);
+	dev_cap->num_ports = field & 0xf;
+	MLX4_GET(field, outbox, QUERY_DEV_CAP_MAX_MSG_SZ_OFFSET);
+	dev_cap->max_msg_sz = 1 << (field & 0x1f);
+	MLX4_GET(field, outbox, QUERY_DEV_CAP_PORT_FLOWSTATS_COUNTERS_OFFSET);
+	if (field & 0x10)
+		dev_cap->flags2 |= MLX4_DEV_CAP_FLAG2_FLOWSTATS_EN;
+	MLX4_GET(field, outbox, QUERY_DEV_CAP_FLOW_STEERING_RANGE_EN_OFFSET);
+	if (field & 0x80)
+		dev_cap->flags2 |= MLX4_DEV_CAP_FLAG2_FS_EN;
+	dev_cap->fs_log_max_ucast_qp_range_size = field & 0x1f;
+	if (field & 0x20)
+		dev_cap->flags2 |= MLX4_DEV_CAP_FLAG2_DMFS_UC_MC_SNIFFER;
+	MLX4_GET(field, outbox, QUERY_DEV_CAP_PORT_BEACON_OFFSET);
+	if (field & 0x80)
+		dev_cap->flags2 |= MLX4_DEV_CAP_FLAG2_PORT_BEACON;
+	MLX4_GET(field, outbox, QUERY_DEV_CAP_FLOW_STEERING_IPOIB_OFFSET);
+	if (field & 0x80)
+		dev_cap->flags2 |= MLX4_DEV_CAP_FLAG2_DMFS_IPOIB;
+	MLX4_GET(field, outbox, QUERY_DEV_CAP_FLOW_STEERING_MAX_QP_OFFSET);
+	dev_cap->fs_max_num_qp_per_entry = field;
+	MLX4_GET(field, outbox, QUERY_DEV_CAP_SL2VL_EVENT_OFFSET);
+	if (field & (1 << 5))
+		dev_cap->flags2 |= MLX4_DEV_CAP_FLAG2_SL_TO_VL_CHANGE_EVENT;
+	MLX4_GET(field, outbox, QUERY_DEV_CAP_ECN_QCN_VER_OFFSET);
+	if (field & 0x1)
+		dev_cap->flags2 |= MLX4_DEV_CAP_FLAG2_QCN;
+	MLX4_GET(stat_rate, outbox, QUERY_DEV_CAP_RATE_SUPPORT_OFFSET);
+	dev_cap->stat_rate_support = stat_rate;
+	MLX4_GET(field, outbox, QUERY_DEV_CAP_CQ_TS_SUPPORT_OFFSET);
+	if (field & 0x80)
+		dev_cap->flags2 |= MLX4_DEV_CAP_FLAG2_TS;
+	MLX4_GET(ext_flags, outbox, QUERY_DEV_CAP_EXT_FLAGS_OFFSET);
+	MLX4_GET(flags, outbox, QUERY_DEV_CAP_FLAGS_OFFSET);
+	dev_cap->flags = flags | (u64)ext_flags << 32;
+	MLX4_GET(field, outbox, QUERY_DEV_CAP_WOL_OFFSET);
+	dev_cap->wol_port[1] = !!(field & 0x20);
+	dev_cap->wol_port[2] = !!(field & 0x40);
+	MLX4_GET(field, outbox, QUERY_DEV_CAP_RSVD_UAR_OFFSET);
+	dev_cap->reserved_uars = field >> 4;
+	MLX4_GET(field, outbox, QUERY_DEV_CAP_UAR_SZ_OFFSET);
+	dev_cap->uar_size = 1 << ((field & 0x3f) + 20);
+	MLX4_GET(field, outbox, QUERY_DEV_CAP_PAGE_SZ_OFFSET);
+	dev_cap->min_page_sz = 1 << field;
+
+	MLX4_GET(field, outbox, QUERY_DEV_CAP_BF_OFFSET);
+	if (field & 0x80) {
+		MLX4_GET(field, outbox, QUERY_DEV_CAP_LOG_BF_REG_SZ_OFFSET);
+		dev_cap->bf_reg_size = 1 << (field & 0x1f);
+		MLX4_GET(field, outbox, QUERY_DEV_CAP_LOG_MAX_BF_REGS_PER_PAGE_OFFSET);
+		if ((1 << (field & 0x3f)) > (PAGE_SIZE / dev_cap->bf_reg_size))
+			field = 3;
+		dev_cap->bf_regs_per_page = 1 << (field & 0x3f);
+	} else {
+		dev_cap->bf_reg_size = 0;
+	}
+
+	MLX4_GET(field, outbox, QUERY_DEV_CAP_MAX_SG_SQ_OFFSET);
+	dev_cap->max_sq_sg = field;
+	MLX4_GET(size, outbox, QUERY_DEV_CAP_MAX_DESC_SZ_SQ_OFFSET);
+	dev_cap->max_sq_desc_sz = size;
+
+	MLX4_GET(field, outbox, QUERY_DEV_CAP_USER_MAC_EN_OFFSET);
+	if (field & (1 << 2))
+		dev_cap->flags2 |= MLX4_DEV_CAP_FLAG2_USER_MAC_EN;
+	MLX4_GET(field, outbox, QUERY_DEV_CAP_SVLAN_BY_QP_OFFSET);
+	if (field & 0x1)
+		dev_cap->flags2 |= MLX4_DEV_CAP_FLAG2_SVLAN_BY_QP;
+	MLX4_GET(field, outbox, QUERY_DEV_CAP_MAX_QP_MCG_OFFSET);
+	dev_cap->max_qp_per_mcg = 1 << field;
+	MLX4_GET(field, outbox, QUERY_DEV_CAP_RSVD_MCG_OFFSET);
+	dev_cap->reserved_mgms = field & 0xf;
+	MLX4_GET(field, outbox, QUERY_DEV_CAP_MAX_MCG_OFFSET);
+	dev_cap->max_mcgs = 1 << field;
+	MLX4_GET(field, outbox, QUERY_DEV_CAP_RSVD_PD_OFFSET);
+	dev_cap->reserved_pds = field >> 4;
+	MLX4_GET(field, outbox, QUERY_DEV_CAP_MAX_PD_OFFSET);
+	dev_cap->max_pds = 1 << (field & 0x3f);
+	MLX4_GET(field, outbox, QUERY_DEV_CAP_RSVD_XRC_OFFSET);
+	dev_cap->reserved_xrcds = field >> 4;
+	MLX4_GET(field, outbox, QUERY_DEV_CAP_MAX_XRC_OFFSET);
+	dev_cap->max_xrcds = 1 << (field & 0x1f);
+
+	MLX4_GET(size, outbox, QUERY_DEV_CAP_RDMARC_ENTRY_SZ_OFFSET);
+	dev_cap->rdmarc_entry_sz = size;
+	MLX4_GET(size, outbox, QUERY_DEV_CAP_QPC_ENTRY_SZ_OFFSET);
+	dev_cap->qpc_entry_sz = size;
+	MLX4_GET(size, outbox, QUERY_DEV_CAP_AUX_ENTRY_SZ_OFFSET);
+	dev_cap->aux_entry_sz = size;
+	MLX4_GET(size, outbox, QUERY_DEV_CAP_ALTC_ENTRY_SZ_OFFSET);
+	dev_cap->altc_entry_sz = size;
+	MLX4_GET(size, outbox, QUERY_DEV_CAP_EQC_ENTRY_SZ_OFFSET);
+	dev_cap->eqc_entry_sz = size;
+	MLX4_GET(size, outbox, QUERY_DEV_CAP_CQC_ENTRY_SZ_OFFSET);
+	dev_cap->cqc_entry_sz = size;
+	MLX4_GET(size, outbox, QUERY_DEV_CAP_SRQ_ENTRY_SZ_OFFSET);
+	dev_cap->srq_entry_sz = size;
+	MLX4_GET(size, outbox, QUERY_DEV_CAP_C_MPT_ENTRY_SZ_OFFSET);
+	dev_cap->cmpt_entry_sz = size;
+	MLX4_GET(size, outbox, QUERY_DEV_CAP_MTT_ENTRY_SZ_OFFSET);
+	dev_cap->mtt_entry_sz = size;
+	MLX4_GET(size, outbox, QUERY_DEV_CAP_D_MPT_ENTRY_SZ_OFFSET);
+	dev_cap->dmpt_entry_sz = size;
+
+	MLX4_GET(field, outbox, QUERY_DEV_CAP_MAX_SRQ_SZ_OFFSET);
+	dev_cap->max_srq_sz = 1 << field;
+	MLX4_GET(field, outbox, QUERY_DEV_CAP_MAX_QP_SZ_OFFSET);
+	dev_cap->max_qp_sz = 1 << field;
+	MLX4_GET(field, outbox, QUERY_DEV_CAP_RSZ_SRQ_OFFSET);
+	dev_cap->resize_srq = field & 1;
+	MLX4_GET(field, outbox, QUERY_DEV_CAP_MAX_SG_RQ_OFFSET);
+	dev_cap->max_rq_sg = field;
+	MLX4_GET(size, outbox, QUERY_DEV_CAP_MAX_DESC_SZ_RQ_OFFSET);
+	dev_cap->max_rq_desc_sz = size;
+	MLX4_GET(field, outbox, QUERY_DEV_CAP_CQ_EQ_CACHE_LINE_STRIDE);
+	if (field & (1 << 4))
+		dev_cap->flags2 |= MLX4_DEV_CAP_FLAG2_QOS_VPP;
+	if (field & (1 << 5))
+		dev_cap->flags2 |= MLX4_DEV_CAP_FLAG2_ETH_PROT_CTRL;
+	if (field & (1 << 6))
+		dev_cap->flags2 |= MLX4_DEV_CAP_FLAG2_CQE_STRIDE;
+	if (field & (1 << 7))
+		dev_cap->flags2 |= MLX4_DEV_CAP_FLAG2_EQE_STRIDE;
+	MLX4_GET(dev_cap->bmme_flags, outbox,
+		 QUERY_DEV_CAP_BMME_FLAGS_OFFSET);
+	if (dev_cap->bmme_flags & MLX4_FLAG_ROCE_V1_V2)
+		dev_cap->flags2 |= MLX4_DEV_CAP_FLAG2_ROCE_V1_V2;
+	if (dev_cap->bmme_flags & MLX4_FLAG_PORT_REMAP)
+		dev_cap->flags2 |= MLX4_DEV_CAP_FLAG2_PORT_REMAP;
+	MLX4_GET(field, outbox, QUERY_DEV_CAP_CONFIG_DEV_OFFSET);
+	if (field & 0x20)
+		dev_cap->flags2 |= MLX4_DEV_CAP_FLAG2_CONFIG_DEV;
+	if (field & (1 << 2))
+		dev_cap->flags2 |= MLX4_DEV_CAP_FLAG2_IGNORE_FCS;
+	MLX4_GET(field, outbox, QUERY_DEV_CAP_PHV_EN_OFFSET);
+	if (field & 0x80)
+		dev_cap->flags2 |= MLX4_DEV_CAP_FLAG2_PHV_EN;
+	if (field & 0x40)
+		dev_cap->flags2 |= MLX4_DEV_CAP_FLAG2_SKIP_OUTER_VLAN;
+
+	MLX4_GET(dev_cap->reserved_lkey, outbox,
+		 QUERY_DEV_CAP_RSVD_LKEY_OFFSET);
+	MLX4_GET(field32, outbox, QUERY_DEV_CAP_ETH_BACKPL_OFFSET);
+	if (field32 & (1 << 0))
+		dev_cap->flags2 |= MLX4_DEV_CAP_FLAG2_ETH_BACKPL_AN_REP;
+	if (field32 & (1 << 7))
+		dev_cap->flags2 |= MLX4_DEV_CAP_FLAG2_RECOVERABLE_ERROR_EVENT;
+	if (field32 & (1 << 8))
+		dev_cap->flags2 |= MLX4_DEV_CAP_FLAG2_DRIVER_VERSION_TO_FW;
+	MLX4_GET(field32, outbox, QUERY_DEV_CAP_DIAG_RPRT_PER_PORT);
+	if (field32 & (1 << 17))
+		dev_cap->flags2 |= MLX4_DEV_CAP_FLAG2_DIAG_PER_PORT;
+	MLX4_GET(field, outbox, QUERY_DEV_CAP_FW_REASSIGN_MAC);
+	if (field & 1<<6)
+		dev_cap->flags2 |= MLX4_DEV_CAP_FLAG2_REASSIGN_MAC_EN;
+	MLX4_GET(field, outbox, QUERY_DEV_CAP_VXLAN);
+	if (field & 1<<3)
+		dev_cap->flags2 |= MLX4_DEV_CAP_FLAG2_VXLAN_OFFLOADS;
+	if (field & (1 << 5))
+		dev_cap->flags2 |= MLX4_DEV_CAP_FLAG2_ETS_CFG;
+	MLX4_GET(dev_cap->max_icm_sz, outbox,
+		 QUERY_DEV_CAP_MAX_ICM_SZ_OFFSET);
+	if (dev_cap->flags & MLX4_DEV_CAP_FLAG_COUNTERS)
+		MLX4_GET(dev_cap->max_counters, outbox,
+			 QUERY_DEV_CAP_MAX_COUNTERS_OFFSET);
+
+	MLX4_GET(field32, outbox,
+		 QUERY_DEV_CAP_MAD_DEMUX_OFFSET);
+	if (field32 & (1 << 0))
+		dev_cap->flags2 |= MLX4_DEV_CAP_FLAG2_MAD_DEMUX;
+
+	MLX4_GET(dev_cap->dmfs_high_rate_qpn_base, outbox,
+		 QUERY_DEV_CAP_DMFS_HIGH_RATE_QPN_BASE_OFFSET);
+	dev_cap->dmfs_high_rate_qpn_base &= MGM_QPN_MASK;
+	MLX4_GET(dev_cap->dmfs_high_rate_qpn_range, outbox,
+		 QUERY_DEV_CAP_DMFS_HIGH_RATE_QPN_RANGE_OFFSET);
+	dev_cap->dmfs_high_rate_qpn_range &= MGM_QPN_MASK;
+
+	MLX4_GET(size, outbox, QUERY_DEV_CAP_QP_RATE_LIMIT_NUM_OFFSET);
+	dev_cap->rl_caps.num_rates = size;
+	if (dev_cap->rl_caps.num_rates) {
+		dev_cap->flags2 |= MLX4_DEV_CAP_FLAG2_QP_RATE_LIMIT;
+		MLX4_GET(size, outbox, QUERY_DEV_CAP_QP_RATE_LIMIT_MAX_OFFSET);
+		dev_cap->rl_caps.max_val  = size & 0xfff;
+		dev_cap->rl_caps.max_unit = size >> 14;
+		MLX4_GET(size, outbox, QUERY_DEV_CAP_QP_RATE_LIMIT_MIN_OFFSET);
+		dev_cap->rl_caps.min_val  = size & 0xfff;
+		dev_cap->rl_caps.min_unit = size >> 14;
+	}
+
+	MLX4_GET(dev_cap->health_buffer_addrs, outbox,
+		 QUERY_DEV_CAP_HEALTH_BUFFER_ADDRESS_OFFSET);
+
+	MLX4_GET(field32, outbox, QUERY_DEV_CAP_EXT_2_FLAGS_OFFSET);
+	if (field32 & (1 << 16))
+		dev_cap->flags2 |= MLX4_DEV_CAP_FLAG2_UPDATE_QP;
+	if (field32 & (1 << 18))
+		dev_cap->flags2 |= MLX4_DEV_CAP_FLAG2_UPDATE_QP_SRC_CHECK_LB;
+	if (field32 & (1 << 19))
+		dev_cap->flags2 |= MLX4_DEV_CAP_FLAG2_LB_SRC_CHK;
+	if (field32 & (1 << 26))
+		dev_cap->flags2 |= MLX4_DEV_CAP_FLAG2_VLAN_CONTROL;
+	if (field32 & (1 << 20))
+		dev_cap->flags2 |= MLX4_DEV_CAP_FLAG2_FSM;
+	if (field32 & (1 << 21))
+		dev_cap->flags2 |= MLX4_DEV_CAP_FLAG2_80_VFS;
+
+	for (i = 1; i <= dev_cap->num_ports; i++) {
+		err = mlx4_QUERY_PORT(dev, i, dev_cap->port_cap + i);
+		if (err)
+			goto out;
+	}
+
+	/*
+	 * Each UAR has 4 EQ doorbells; so if a UAR is reserved, then
+	 * we can't use any EQs whose doorbell falls on that page,
+	 * even if the EQ itself isn't reserved.
+	 */
+	if (dev_cap->num_sys_eqs == 0)
+		dev_cap->reserved_eqs = max(dev_cap->reserved_uars * 4,
+					    dev_cap->reserved_eqs);
+	else
+		dev_cap->flags2 |= MLX4_DEV_CAP_FLAG2_SYS_EQS;
+
+out:
+	mlx4_free_cmd_mailbox(dev, mailbox);
+	return err;
+}
+
+void mlx4_dev_cap_dump(struct mlx4_dev *dev, struct mlx4_dev_cap *dev_cap)
+{
+	if (dev_cap->bf_reg_size > 0)
+		mlx4_dbg(dev, "BlueFlame available (reg size %d, regs/page %d)\n",
+			 dev_cap->bf_reg_size, dev_cap->bf_regs_per_page);
+	else
+		mlx4_dbg(dev, "BlueFlame not available\n");
+
+	mlx4_dbg(dev, "Base MM extensions: flags %08x, rsvd L_Key %08x\n",
+		 dev_cap->bmme_flags, dev_cap->reserved_lkey);
+	mlx4_dbg(dev, "Max ICM size %lld MB\n",
+		 (unsigned long long) dev_cap->max_icm_sz >> 20);
+	mlx4_dbg(dev, "Max QPs: %d, reserved QPs: %d, entry size: %d\n",
+		 dev_cap->max_qps, dev_cap->reserved_qps, dev_cap->qpc_entry_sz);
+	mlx4_dbg(dev, "Max SRQs: %d, reserved SRQs: %d, entry size: %d\n",
+		 dev_cap->max_srqs, dev_cap->reserved_srqs, dev_cap->srq_entry_sz);
+	mlx4_dbg(dev, "Max CQs: %d, reserved CQs: %d, entry size: %d\n",
+		 dev_cap->max_cqs, dev_cap->reserved_cqs, dev_cap->cqc_entry_sz);
+	mlx4_dbg(dev, "Num sys EQs: %d, max EQs: %d, reserved EQs: %d, entry size: %d\n",
+		 dev_cap->num_sys_eqs, dev_cap->max_eqs, dev_cap->reserved_eqs,
+		 dev_cap->eqc_entry_sz);
+	mlx4_dbg(dev, "reserved MPTs: %d, reserved MTTs: %d\n",
+		 dev_cap->reserved_mrws, dev_cap->reserved_mtts);
+	mlx4_dbg(dev, "Max PDs: %d, reserved PDs: %d, reserved UARs: %d\n",
+		 dev_cap->max_pds, dev_cap->reserved_pds, dev_cap->reserved_uars);
+	mlx4_dbg(dev, "Max QP/MCG: %d, reserved MGMs: %d\n",
+		 dev_cap->max_pds, dev_cap->reserved_mgms);
+	mlx4_dbg(dev, "Max CQEs: %d, max WQEs: %d, max SRQ WQEs: %d\n",
+		 dev_cap->max_cq_sz, dev_cap->max_qp_sz, dev_cap->max_srq_sz);
+	mlx4_dbg(dev, "Local CA ACK delay: %d, max MTU: %d, port width cap: %d\n",
+		 dev_cap->local_ca_ack_delay, 128 << dev_cap->port_cap[1].ib_mtu,
+		 dev_cap->port_cap[1].max_port_width);
+	mlx4_dbg(dev, "Max SQ desc size: %d, max SQ S/G: %d\n",
+		 dev_cap->max_sq_desc_sz, dev_cap->max_sq_sg);
+	mlx4_dbg(dev, "Max RQ desc size: %d, max RQ S/G: %d\n",
+		 dev_cap->max_rq_desc_sz, dev_cap->max_rq_sg);
+	mlx4_dbg(dev, "Max GSO size: %d\n", dev_cap->max_gso_sz);
+	mlx4_dbg(dev, "Max counters: %d\n", dev_cap->max_counters);
+	mlx4_dbg(dev, "Max RSS Table size: %d\n", dev_cap->max_rss_tbl_sz);
+	mlx4_dbg(dev, "DMFS high rate steer QPn base: %d\n",
+		 dev_cap->dmfs_high_rate_qpn_base);
+	mlx4_dbg(dev, "DMFS high rate steer QPn range: %d\n",
+		 dev_cap->dmfs_high_rate_qpn_range);
+
+	if (dev_cap->flags2 & MLX4_DEV_CAP_FLAG2_QP_RATE_LIMIT) {
+		struct mlx4_rate_limit_caps *rl_caps = &dev_cap->rl_caps;
+
+		mlx4_dbg(dev, "QP Rate-Limit: #rates %d, unit/val max %d/%d, min %d/%d\n",
+			 rl_caps->num_rates, rl_caps->max_unit, rl_caps->max_val,
+			 rl_caps->min_unit, rl_caps->min_val);
+	}
+
+	dump_dev_cap_flags(dev, dev_cap->flags);
+	dump_dev_cap_flags2(dev, dev_cap->flags2);
+}
+
+int mlx4_QUERY_PORT(struct mlx4_dev *dev, int port, struct mlx4_port_cap *port_cap)
+{
+	struct mlx4_cmd_mailbox *mailbox;
+	u32 *outbox;
+	u8 field;
+	u32 field32;
+	int err;
+
+	mailbox = mlx4_alloc_cmd_mailbox(dev);
+	if (IS_ERR(mailbox))
+		return PTR_ERR(mailbox);
+	outbox = mailbox->buf;
+
+	if (dev->flags & MLX4_FLAG_OLD_PORT_CMDS) {
+		err = mlx4_cmd_box(dev, 0, mailbox->dma, 0, 0, MLX4_CMD_QUERY_DEV_CAP,
+				   MLX4_CMD_TIME_CLASS_A,
+				   MLX4_CMD_NATIVE);
+
+		if (err)
+			goto out;
+
+		MLX4_GET(field, outbox, QUERY_DEV_CAP_VL_PORT_OFFSET);
+		port_cap->max_vl	   = field >> 4;
+		MLX4_GET(field, outbox, QUERY_DEV_CAP_MTU_WIDTH_OFFSET);
+		port_cap->ib_mtu	   = field >> 4;
+		port_cap->max_port_width = field & 0xf;
+		MLX4_GET(field, outbox, QUERY_DEV_CAP_MAX_GID_OFFSET);
+		port_cap->max_gids	   = 1 << (field & 0xf);
+		MLX4_GET(field, outbox, QUERY_DEV_CAP_MAX_PKEY_OFFSET);
+		port_cap->max_pkeys	   = 1 << (field & 0xf);
+	} else {
+#define QUERY_PORT_SUPPORTED_TYPE_OFFSET	0x00
+#define QUERY_PORT_MTU_OFFSET			0x01
+#define QUERY_PORT_ETH_MTU_OFFSET		0x02
+#define QUERY_PORT_WIDTH_OFFSET			0x06
+#define QUERY_PORT_MAX_GID_PKEY_OFFSET		0x07
+#define QUERY_PORT_MAX_MACVLAN_OFFSET		0x0a
+#define QUERY_PORT_MAX_VL_OFFSET		0x0b
+#define QUERY_PORT_MAC_OFFSET			0x10
+#define QUERY_PORT_TRANS_VENDOR_OFFSET		0x18
+#define QUERY_PORT_WAVELENGTH_OFFSET		0x1c
+#define QUERY_PORT_TRANS_CODE_OFFSET		0x20
+
+		err = mlx4_cmd_box(dev, 0, mailbox->dma, port, 0, MLX4_CMD_QUERY_PORT,
+				   MLX4_CMD_TIME_CLASS_B, MLX4_CMD_NATIVE);
+		if (err)
+			goto out;
+
+		MLX4_GET(field, outbox, QUERY_PORT_SUPPORTED_TYPE_OFFSET);
+		port_cap->link_state = (field & 0x80) >> 7;
+		port_cap->supported_port_types = field & 3;
+		port_cap->suggested_type = (field >> 3) & 1;
+		port_cap->default_sense = (field >> 4) & 1;
+		port_cap->dmfs_optimized_state = (field >> 5) & 1;
+		MLX4_GET(field, outbox, QUERY_PORT_MTU_OFFSET);
+		port_cap->ib_mtu	   = field & 0xf;
+		MLX4_GET(field, outbox, QUERY_PORT_WIDTH_OFFSET);
+		port_cap->max_port_width = field & 0xf;
+		MLX4_GET(field, outbox, QUERY_PORT_MAX_GID_PKEY_OFFSET);
+		port_cap->max_gids	   = 1 << (field >> 4);
+		port_cap->max_pkeys	   = 1 << (field & 0xf);
+		MLX4_GET(field, outbox, QUERY_PORT_MAX_VL_OFFSET);
+		port_cap->max_vl	   = field & 0xf;
+		port_cap->max_tc_eth	   = field >> 4;
+		MLX4_GET(field, outbox, QUERY_PORT_MAX_MACVLAN_OFFSET);
+		port_cap->log_max_macs  = field & 0xf;
+		port_cap->log_max_vlans = field >> 4;
+		MLX4_GET(port_cap->eth_mtu, outbox, QUERY_PORT_ETH_MTU_OFFSET);
+		MLX4_GET(port_cap->def_mac, outbox, QUERY_PORT_MAC_OFFSET);
+		MLX4_GET(field32, outbox, QUERY_PORT_TRANS_VENDOR_OFFSET);
+		port_cap->trans_type = field32 >> 24;
+		port_cap->vendor_oui = field32 & 0xffffff;
+		MLX4_GET(port_cap->wavelength, outbox, QUERY_PORT_WAVELENGTH_OFFSET);
+		MLX4_GET(port_cap->trans_code, outbox, QUERY_PORT_TRANS_CODE_OFFSET);
+	}
+
+out:
+	mlx4_free_cmd_mailbox(dev, mailbox);
+	return err;
+}
+
+#define DEV_CAP_EXT_2_FLAG_PFC_COUNTERS	(1 << 28)
+#define DEV_CAP_EXT_2_FLAG_VLAN_CONTROL (1 << 26)
+#define DEV_CAP_EXT_2_FLAG_80_VFS	(1 << 21)
+#define DEV_CAP_EXT_2_FLAG_FSM		(1 << 20)
+
+int mlx4_QUERY_DEV_CAP_wrapper(struct mlx4_dev *dev, int slave,
+			       struct mlx4_vhcr *vhcr,
+			       struct mlx4_cmd_mailbox *inbox,
+			       struct mlx4_cmd_mailbox *outbox,
+			       struct mlx4_cmd_info *cmd)
+{
+	u64	flags;
+	int	err = 0;
+	u8	field;
+	u16	field16;
+	u32	bmme_flags, field32;
+	int	real_port;
+	int	slave_port;
+	int	first_port;
+	struct mlx4_active_ports actv_ports;
+
+	err = mlx4_cmd_box(dev, 0, outbox->dma, 0, 0, MLX4_CMD_QUERY_DEV_CAP,
+			   MLX4_CMD_TIME_CLASS_A, MLX4_CMD_NATIVE);
+	if (err)
+		return err;
+
+	disable_unsupported_roce_caps(outbox->buf);
+	/* add port mng change event capability and disable mw type 1
+	 * unconditionally to slaves
+	 */
+	MLX4_GET(flags, outbox->buf, QUERY_DEV_CAP_EXT_FLAGS_OFFSET);
+	flags |= MLX4_DEV_CAP_FLAG_PORT_MNG_CHG_EV;
+	flags &= ~MLX4_DEV_CAP_FLAG_MEM_WINDOW;
+	actv_ports = mlx4_get_active_ports(dev, slave);
+	first_port = find_first_bit(actv_ports.ports, dev->caps.num_ports);
+	for (slave_port = 0, real_port = first_port;
+	     real_port < first_port +
+	     bitmap_weight(actv_ports.ports, dev->caps.num_ports);
+	     ++real_port, ++slave_port) {
+		if (flags & (MLX4_DEV_CAP_FLAG_WOL_PORT1 << real_port))
+			flags |= MLX4_DEV_CAP_FLAG_WOL_PORT1 << slave_port;
+		else
+			flags &= ~(MLX4_DEV_CAP_FLAG_WOL_PORT1 << slave_port);
+	}
+	for (; slave_port < dev->caps.num_ports; ++slave_port)
+		flags &= ~(MLX4_DEV_CAP_FLAG_WOL_PORT1 << slave_port);
+
+	/* Not exposing RSS IP fragments to guests */
+	flags &= ~MLX4_DEV_CAP_FLAG_RSS_IP_FRAG;
+	MLX4_PUT(outbox->buf, flags, QUERY_DEV_CAP_EXT_FLAGS_OFFSET);
+
+	MLX4_GET(field, outbox->buf, QUERY_DEV_CAP_VL_PORT_OFFSET);
+	field &= ~0x0F;
+	field |= bitmap_weight(actv_ports.ports, dev->caps.num_ports) & 0x0F;
+	MLX4_PUT(outbox->buf, field, QUERY_DEV_CAP_VL_PORT_OFFSET);
+
+	/* For guests, disable timestamp */
+	MLX4_GET(field, outbox->buf, QUERY_DEV_CAP_CQ_TS_SUPPORT_OFFSET);
+	field &= 0x7f;
+	MLX4_PUT(outbox->buf, field, QUERY_DEV_CAP_CQ_TS_SUPPORT_OFFSET);
+
+	/* For guests, disable vxlan tunneling and QoS support */
+	MLX4_GET(field, outbox->buf, QUERY_DEV_CAP_VXLAN);
+	field &= 0xd7;
+	MLX4_PUT(outbox->buf, field, QUERY_DEV_CAP_VXLAN);
+
+	/* For guests, disable port BEACON */
+	MLX4_GET(field, outbox->buf, QUERY_DEV_CAP_PORT_BEACON_OFFSET);
+	field &= 0x7f;
+	MLX4_PUT(outbox->buf, field, QUERY_DEV_CAP_PORT_BEACON_OFFSET);
+
+	/* For guests, report Blueflame disabled */
+	MLX4_GET(field, outbox->buf, QUERY_DEV_CAP_BF_OFFSET);
+	field &= 0x7f;
+	MLX4_PUT(outbox->buf, field, QUERY_DEV_CAP_BF_OFFSET);
+
+	/* For guests, disable mw type 2 and port remap*/
+	MLX4_GET(bmme_flags, outbox->buf, QUERY_DEV_CAP_BMME_FLAGS_OFFSET);
+	bmme_flags &= ~MLX4_BMME_FLAG_TYPE_2_WIN;
+	bmme_flags &= ~MLX4_FLAG_PORT_REMAP;
+	MLX4_PUT(outbox->buf, bmme_flags, QUERY_DEV_CAP_BMME_FLAGS_OFFSET);
+
+	/* turn off device-managed steering capability if not enabled */
+	if (dev->caps.steering_mode != MLX4_STEERING_MODE_DEVICE_MANAGED) {
+		MLX4_GET(field, outbox->buf,
+			 QUERY_DEV_CAP_FLOW_STEERING_RANGE_EN_OFFSET);
+		field &= 0x7f;
+		MLX4_PUT(outbox->buf, field,
+			 QUERY_DEV_CAP_FLOW_STEERING_RANGE_EN_OFFSET);
+	}
+
+	/* turn off ipoib managed steering for guests */
+	MLX4_GET(field, outbox->buf, QUERY_DEV_CAP_FLOW_STEERING_IPOIB_OFFSET);
+	field &= ~0x80;
+	MLX4_PUT(outbox->buf, field, QUERY_DEV_CAP_FLOW_STEERING_IPOIB_OFFSET);
+
+	/* turn off host side virt features (VST, FSM, etc) for guests */
+	MLX4_GET(field32, outbox->buf, QUERY_DEV_CAP_EXT_2_FLAGS_OFFSET);
+	field32 &= ~(DEV_CAP_EXT_2_FLAG_VLAN_CONTROL | DEV_CAP_EXT_2_FLAG_80_VFS |
+		     DEV_CAP_EXT_2_FLAG_FSM | DEV_CAP_EXT_2_FLAG_PFC_COUNTERS);
+	MLX4_PUT(outbox->buf, field32, QUERY_DEV_CAP_EXT_2_FLAGS_OFFSET);
+
+	/* turn off QCN for guests */
+	MLX4_GET(field, outbox->buf, QUERY_DEV_CAP_ECN_QCN_VER_OFFSET);
+	field &= 0xfe;
+	MLX4_PUT(outbox->buf, field, QUERY_DEV_CAP_ECN_QCN_VER_OFFSET);
+
+	/* turn off QP max-rate limiting for guests */
+	field16 = 0;
+	MLX4_PUT(outbox->buf, field16, QUERY_DEV_CAP_QP_RATE_LIMIT_NUM_OFFSET);
+
+	/* turn off QoS per VF support for guests */
+	MLX4_GET(field, outbox->buf, QUERY_DEV_CAP_CQ_EQ_CACHE_LINE_STRIDE);
+	field &= 0xef;
+	MLX4_PUT(outbox->buf, field, QUERY_DEV_CAP_CQ_EQ_CACHE_LINE_STRIDE);
+
+	/* turn off ignore FCS feature for guests */
+	MLX4_GET(field, outbox->buf, QUERY_DEV_CAP_CONFIG_DEV_OFFSET);
+	field &= 0xfb;
+	MLX4_PUT(outbox->buf, field, QUERY_DEV_CAP_CONFIG_DEV_OFFSET);
+
+	return 0;
+}
+
+static void disable_unsupported_roce_caps(void *buf)
+{
+	u32 flags;
+
+	MLX4_GET(flags, buf, QUERY_DEV_CAP_EXT_FLAGS_OFFSET);
+	flags &= ~(1UL << 31);
+	MLX4_PUT(buf, flags, QUERY_DEV_CAP_EXT_FLAGS_OFFSET);
+	MLX4_GET(flags, buf, QUERY_DEV_CAP_EXT_2_FLAGS_OFFSET);
+	flags &= ~(1UL << 24);
+	MLX4_PUT(buf, flags, QUERY_DEV_CAP_EXT_2_FLAGS_OFFSET);
+	MLX4_GET(flags, buf, QUERY_DEV_CAP_BMME_FLAGS_OFFSET);
+	flags &= ~(MLX4_FLAG_ROCE_V1_V2);
+	MLX4_PUT(buf, flags, QUERY_DEV_CAP_BMME_FLAGS_OFFSET);
+}
+
+int mlx4_QUERY_PORT_wrapper(struct mlx4_dev *dev, int slave,
+			    struct mlx4_vhcr *vhcr,
+			    struct mlx4_cmd_mailbox *inbox,
+			    struct mlx4_cmd_mailbox *outbox,
+			    struct mlx4_cmd_info *cmd)
+{
+	struct mlx4_priv *priv = mlx4_priv(dev);
+	u64 def_mac;
+	u8 port_type;
+	u16 short_field;
+	int err;
+	int admin_link_state;
+	int port = mlx4_slave_convert_port(dev, slave,
+					   vhcr->in_modifier & 0xFF);
+
+#define MLX4_VF_PORT_NO_LINK_SENSE_MASK	0xE0
+#define MLX4_PORT_LINK_UP_MASK		0x80
+#define QUERY_PORT_CUR_MAX_PKEY_OFFSET	0x0c
+#define QUERY_PORT_CUR_MAX_GID_OFFSET	0x0e
+
+	if (port < 0)
+		return -EINVAL;
+
+	/* Protect against untrusted guests: enforce that this is the
+	 * QUERY_PORT general query.
+	 */
+	if (vhcr->op_modifier || vhcr->in_modifier & ~0xFF)
+		return -EINVAL;
+
+	vhcr->in_modifier = port;
+
+	err = mlx4_cmd_box(dev, 0, outbox->dma, vhcr->in_modifier, 0,
+			   MLX4_CMD_QUERY_PORT, MLX4_CMD_TIME_CLASS_B,
+			   MLX4_CMD_NATIVE);
+
+	if (!err && dev->caps.function != slave) {
+		def_mac = priv->mfunc.master.vf_oper[slave].vport[vhcr->in_modifier].state.mac;
+		MLX4_PUT(outbox->buf, def_mac, QUERY_PORT_MAC_OFFSET);
+
+		/* get port type - currently only eth is enabled */
+		MLX4_GET(port_type, outbox->buf,
+			 QUERY_PORT_SUPPORTED_TYPE_OFFSET);
+
+		/* No link sensing allowed */
+		port_type &= MLX4_VF_PORT_NO_LINK_SENSE_MASK;
+		/* set port type to currently operating port type */
+		port_type |= (dev->caps.port_type[vhcr->in_modifier] & 0x3);
+
+		admin_link_state = priv->mfunc.master.vf_oper[slave].vport[vhcr->in_modifier].state.link_state;
+		if (IFLA_VF_LINK_STATE_ENABLE == admin_link_state)
+			port_type |= MLX4_PORT_LINK_UP_MASK;
+		else if (IFLA_VF_LINK_STATE_DISABLE == admin_link_state)
+			port_type &= ~MLX4_PORT_LINK_UP_MASK;
+		else if (IFLA_VF_LINK_STATE_AUTO == admin_link_state && mlx4_is_bonded(dev)) {
+			int other_port = (port == 1) ? 2 : 1;
+			struct mlx4_port_cap port_cap;
+
+			err = mlx4_QUERY_PORT(dev, other_port, &port_cap);
+			if (err)
+				goto out;
+			port_type |= (port_cap.link_state << 7);
+		}
+
+		MLX4_PUT(outbox->buf, port_type,
+			 QUERY_PORT_SUPPORTED_TYPE_OFFSET);
+
+		if (dev->caps.port_type[vhcr->in_modifier] == MLX4_PORT_TYPE_ETH)
+			short_field = mlx4_get_slave_num_gids(dev, slave, port);
+		else
+			short_field = 1; /* slave max gids */
+		MLX4_PUT(outbox->buf, short_field,
+			 QUERY_PORT_CUR_MAX_GID_OFFSET);
+
+		short_field = dev->caps.pkey_table_len[vhcr->in_modifier];
+		MLX4_PUT(outbox->buf, short_field,
+			 QUERY_PORT_CUR_MAX_PKEY_OFFSET);
+	}
+out:
+	return err;
+}
+
+int mlx4_get_slave_pkey_gid_tbl_len(struct mlx4_dev *dev, u8 port,
+				    int *gid_tbl_len, int *pkey_tbl_len)
+{
+	struct mlx4_cmd_mailbox *mailbox;
+	u32			*outbox;
+	u16			field;
+	int			err;
+
+	mailbox = mlx4_alloc_cmd_mailbox(dev);
+	if (IS_ERR(mailbox))
+		return PTR_ERR(mailbox);
+
+	err =  mlx4_cmd_box(dev, 0, mailbox->dma, port, 0,
+			    MLX4_CMD_QUERY_PORT, MLX4_CMD_TIME_CLASS_B,
+			    MLX4_CMD_WRAPPED);
+	if (err)
+		goto out;
+
+	outbox = mailbox->buf;
+
+	MLX4_GET(field, outbox, QUERY_PORT_CUR_MAX_GID_OFFSET);
+	*gid_tbl_len = field;
+
+	MLX4_GET(field, outbox, QUERY_PORT_CUR_MAX_PKEY_OFFSET);
+	*pkey_tbl_len = field;
+
+out:
+	mlx4_free_cmd_mailbox(dev, mailbox);
+	return err;
+}
+EXPORT_SYMBOL(mlx4_get_slave_pkey_gid_tbl_len);
+
+int mlx4_map_cmd(struct mlx4_dev *dev, u16 op, struct mlx4_icm *icm, u64 virt)
+{
+	struct mlx4_cmd_mailbox *mailbox;
+	struct mlx4_icm_iter iter;
+	__be64 *pages;
+	int lg;
+	int nent = 0;
+	int i;
+	int err = 0;
+	int ts = 0, tc = 0;
+
+	mailbox = mlx4_alloc_cmd_mailbox(dev);
+	if (IS_ERR(mailbox))
+		return PTR_ERR(mailbox);
+	pages = mailbox->buf;
+
+	for (mlx4_icm_first(icm, &iter);
+	     !mlx4_icm_last(&iter);
+	     mlx4_icm_next(&iter)) {
+		/*
+		 * We have to pass pages that are aligned to their
+		 * size, so find the least significant 1 in the
+		 * address or size and use that as our log2 size.
+		 */
+		lg = ffs(mlx4_icm_addr(&iter) | mlx4_icm_size(&iter)) - 1;
+		if (lg < MLX4_ICM_PAGE_SHIFT) {
+			mlx4_warn(dev, "Got FW area not aligned to %d (%llx/%lx)\n",
+				  MLX4_ICM_PAGE_SIZE,
+				  (unsigned long long) mlx4_icm_addr(&iter),
+				  mlx4_icm_size(&iter));
+			err = -EINVAL;
+			goto out;
+		}
+
+		for (i = 0; i < mlx4_icm_size(&iter) >> lg; ++i) {
+			if (virt != -1) {
+				pages[nent * 2] = cpu_to_be64(virt);
+				virt += 1ULL << lg;
+			}
+
+			pages[nent * 2 + 1] =
+				cpu_to_be64((mlx4_icm_addr(&iter) + (i << lg)) |
+					    (lg - MLX4_ICM_PAGE_SHIFT));
+			ts += 1 << (lg - 10);
+			++tc;
+
+			if (++nent == MLX4_MAILBOX_SIZE / 16) {
+				err = mlx4_cmd(dev, mailbox->dma, nent, 0, op,
+						MLX4_CMD_TIME_CLASS_B,
+						MLX4_CMD_NATIVE);
+				if (err)
+					goto out;
+				nent = 0;
+			}
+		}
+	}
+
+	if (nent)
+		err = mlx4_cmd(dev, mailbox->dma, nent, 0, op,
+			       MLX4_CMD_TIME_CLASS_B, MLX4_CMD_NATIVE);
+	if (err)
+		goto out;
+
+	switch (op) {
+	case MLX4_CMD_MAP_FA:
+		mlx4_dbg(dev, "Mapped %d chunks/%d KB for FW\n", tc, ts);
+		break;
+	case MLX4_CMD_MAP_ICM_AUX:
+		mlx4_dbg(dev, "Mapped %d chunks/%d KB for ICM aux\n", tc, ts);
+		break;
+	case MLX4_CMD_MAP_ICM:
+		mlx4_dbg(dev, "Mapped %d chunks/%d KB at %llx for ICM\n",
+			 tc, ts, (unsigned long long) virt - (ts << 10));
+		break;
+	}
+
+out:
+	mlx4_free_cmd_mailbox(dev, mailbox);
+	return err;
+}
+
+int mlx4_MAP_FA(struct mlx4_dev *dev, struct mlx4_icm *icm)
+{
+	return mlx4_map_cmd(dev, MLX4_CMD_MAP_FA, icm, -1);
+}
+
+int mlx4_UNMAP_FA(struct mlx4_dev *dev)
+{
+	return mlx4_cmd(dev, 0, 0, 0, MLX4_CMD_UNMAP_FA,
+			MLX4_CMD_TIME_CLASS_B, MLX4_CMD_NATIVE);
+}
+
+
+int mlx4_RUN_FW(struct mlx4_dev *dev)
+{
+	return mlx4_cmd(dev, 0, 0, 0, MLX4_CMD_RUN_FW,
+			MLX4_CMD_TIME_CLASS_A, MLX4_CMD_NATIVE);
+}
+
+int mlx4_QUERY_FW(struct mlx4_dev *dev)
+{
+	struct mlx4_fw  *fw  = &mlx4_priv(dev)->fw;
+	struct mlx4_cmd *cmd = &mlx4_priv(dev)->cmd;
+	struct mlx4_cmd_mailbox *mailbox;
+	u32 *outbox;
+	int err = 0;
+	u64 fw_ver;
+	u16 cmd_if_rev;
+	u8 lg;
+
+#define QUERY_FW_OUT_SIZE             0x100
+#define QUERY_FW_VER_OFFSET            0x00
+#define QUERY_FW_PPF_ID		       0x09
+#define QUERY_FW_CMD_IF_REV_OFFSET     0x0a
+#define QUERY_FW_MAX_CMD_OFFSET        0x0f
+#define QUERY_FW_ERR_START_OFFSET      0x30
+#define QUERY_FW_ERR_SIZE_OFFSET       0x38
+#define QUERY_FW_ERR_BAR_OFFSET        0x3c
+
+#define QUERY_FW_SIZE_OFFSET           0x00
+#define QUERY_FW_CLR_INT_BASE_OFFSET   0x20
+#define QUERY_FW_CLR_INT_BAR_OFFSET    0x28
+
+#define QUERY_FW_COMM_BASE_OFFSET      0x40
+#define QUERY_FW_COMM_BAR_OFFSET       0x48
+
+#define QUERY_FW_CLOCK_OFFSET	       0x50
+#define QUERY_FW_CLOCK_BAR	       0x58
+
+	mailbox = mlx4_alloc_cmd_mailbox(dev);
+	if (IS_ERR(mailbox))
+		return PTR_ERR(mailbox);
+	outbox = mailbox->buf;
+
+	err = mlx4_cmd_box(dev, 0, mailbox->dma, 0, 0, MLX4_CMD_QUERY_FW,
+			    MLX4_CMD_TIME_CLASS_A, MLX4_CMD_NATIVE);
+	if (err)
+		goto out;
+
+	MLX4_GET(fw_ver, outbox, QUERY_FW_VER_OFFSET);
+	/*
+	 * FW subminor version is at more significant bits than minor
+	 * version, so swap here.
+	 */
+	dev->caps.fw_ver = (fw_ver & 0xffff00000000ull) |
+		((fw_ver & 0xffff0000ull) >> 16) |
+		((fw_ver & 0x0000ffffull) << 16);
+
+	MLX4_GET(lg, outbox, QUERY_FW_PPF_ID);
+	dev->caps.function = lg;
+
+	if (mlx4_is_slave(dev))
+		goto out;
+
+
+	MLX4_GET(cmd_if_rev, outbox, QUERY_FW_CMD_IF_REV_OFFSET);
+	if (cmd_if_rev < MLX4_COMMAND_INTERFACE_MIN_REV ||
+	    cmd_if_rev > MLX4_COMMAND_INTERFACE_MAX_REV) {
+		mlx4_err(dev, "Installed FW has unsupported command interface revision %d\n",
+			 cmd_if_rev);
+		mlx4_err(dev, "(Installed FW version is %d.%d.%03d)\n",
+			 (int) (dev->caps.fw_ver >> 32),
+			 (int) (dev->caps.fw_ver >> 16) & 0xffff,
+			 (int) dev->caps.fw_ver & 0xffff);
+		mlx4_err(dev, "This driver version supports only revisions %d to %d\n",
+			 MLX4_COMMAND_INTERFACE_MIN_REV, MLX4_COMMAND_INTERFACE_MAX_REV);
+		err = -ENODEV;
+		goto out;
+	}
+
+	if (cmd_if_rev < MLX4_COMMAND_INTERFACE_NEW_PORT_CMDS)
+		dev->flags |= MLX4_FLAG_OLD_PORT_CMDS;
+
+	MLX4_GET(lg, outbox, QUERY_FW_MAX_CMD_OFFSET);
+	cmd->max_cmds = 1 << lg;
+
+	mlx4_dbg(dev, "FW version %d.%d.%03d (cmd intf rev %d), max commands %d\n",
+		 (int) (dev->caps.fw_ver >> 32),
+		 (int) (dev->caps.fw_ver >> 16) & 0xffff,
+		 (int) dev->caps.fw_ver & 0xffff,
+		 cmd_if_rev, cmd->max_cmds);
+
+	MLX4_GET(fw->catas_offset, outbox, QUERY_FW_ERR_START_OFFSET);
+	MLX4_GET(fw->catas_size,   outbox, QUERY_FW_ERR_SIZE_OFFSET);
+	MLX4_GET(fw->catas_bar,    outbox, QUERY_FW_ERR_BAR_OFFSET);
+	fw->catas_bar = (fw->catas_bar >> 6) * 2;
+
+	mlx4_dbg(dev, "Catastrophic error buffer at 0x%llx, size 0x%x, BAR %d\n",
+		 (unsigned long long) fw->catas_offset, fw->catas_size, fw->catas_bar);
+
+	MLX4_GET(fw->fw_pages,     outbox, QUERY_FW_SIZE_OFFSET);
+	MLX4_GET(fw->clr_int_base, outbox, QUERY_FW_CLR_INT_BASE_OFFSET);
+	MLX4_GET(fw->clr_int_bar,  outbox, QUERY_FW_CLR_INT_BAR_OFFSET);
+	fw->clr_int_bar = (fw->clr_int_bar >> 6) * 2;
+
+	MLX4_GET(fw->comm_base, outbox, QUERY_FW_COMM_BASE_OFFSET);
+	MLX4_GET(fw->comm_bar,  outbox, QUERY_FW_COMM_BAR_OFFSET);
+	fw->comm_bar = (fw->comm_bar >> 6) * 2;
+	mlx4_dbg(dev, "Communication vector bar:%d offset:0x%llx\n",
+		 fw->comm_bar, fw->comm_base);
+	mlx4_dbg(dev, "FW size %d KB\n", fw->fw_pages >> 2);
+
+	MLX4_GET(fw->clock_offset, outbox, QUERY_FW_CLOCK_OFFSET);
+	MLX4_GET(fw->clock_bar,    outbox, QUERY_FW_CLOCK_BAR);
+	fw->clock_bar = (fw->clock_bar >> 6) * 2;
+	mlx4_dbg(dev, "Internal clock bar:%d offset:0x%llx\n",
+		 fw->clock_bar, fw->clock_offset);
+
+	/*
+	 * Round up number of system pages needed in case
+	 * MLX4_ICM_PAGE_SIZE < PAGE_SIZE.
+	 */
+	fw->fw_pages =
+		ALIGN(fw->fw_pages, PAGE_SIZE / MLX4_ICM_PAGE_SIZE) >>
+		(PAGE_SHIFT - MLX4_ICM_PAGE_SHIFT);
+
+	mlx4_dbg(dev, "Clear int @ %llx, BAR %d\n",
+		 (unsigned long long) fw->clr_int_base, fw->clr_int_bar);
+
+out:
+	mlx4_free_cmd_mailbox(dev, mailbox);
+	return err;
+}
+
+int mlx4_QUERY_FW_wrapper(struct mlx4_dev *dev, int slave,
+			  struct mlx4_vhcr *vhcr,
+			  struct mlx4_cmd_mailbox *inbox,
+			  struct mlx4_cmd_mailbox *outbox,
+			  struct mlx4_cmd_info *cmd)
+{
+	u8 *outbuf;
+	int err;
+
+	outbuf = outbox->buf;
+	err = mlx4_cmd_box(dev, 0, outbox->dma, 0, 0, MLX4_CMD_QUERY_FW,
+			    MLX4_CMD_TIME_CLASS_A, MLX4_CMD_NATIVE);
+	if (err)
+		return err;
+
+	/* for slaves, set pci PPF ID to invalid and zero out everything
+	 * else except FW version */
+	outbuf[0] = outbuf[1] = 0;
+	memset(&outbuf[8], 0, QUERY_FW_OUT_SIZE - 8);
+	outbuf[QUERY_FW_PPF_ID] = MLX4_INVALID_SLAVE_ID;
+
+	return 0;
+}
+
+static void get_board_id(void *vsd, char *board_id)
+{
+	int i;
+
+#define VSD_OFFSET_SIG1		0x00
+#define VSD_OFFSET_SIG2		0xde
+#define VSD_OFFSET_MLX_BOARD_ID	0xd0
+#define VSD_OFFSET_TS_BOARD_ID	0x20
+
+#define VSD_SIGNATURE_TOPSPIN	0x5ad
+
+	memset(board_id, 0, MLX4_BOARD_ID_LEN);
+
+	if (be16_to_cpup(vsd + VSD_OFFSET_SIG1) == VSD_SIGNATURE_TOPSPIN &&
+	    be16_to_cpup(vsd + VSD_OFFSET_SIG2) == VSD_SIGNATURE_TOPSPIN) {
+		strlcpy(board_id, vsd + VSD_OFFSET_TS_BOARD_ID, MLX4_BOARD_ID_LEN);
+	} else {
+		/*
+		 * The board ID is a string but the firmware byte
+		 * swaps each 4-byte word before passing it back to
+		 * us.  Therefore we need to swab it before printing.
+		 */
+		u32 *bid_u32 = (u32 *)board_id;
+
+		for (i = 0; i < 4; ++i) {
+			u32 *addr;
+			u32 val;
+
+			addr = (u32 *) (vsd + VSD_OFFSET_MLX_BOARD_ID + i * 4);
+			val = get_unaligned(addr);
+			val = swab32(val);
+			put_unaligned(val, &bid_u32[i]);
+		}
+	}
+}
+
+int mlx4_QUERY_ADAPTER(struct mlx4_dev *dev, struct mlx4_adapter *adapter)
+{
+	struct mlx4_cmd_mailbox *mailbox;
+	u32 *outbox;
+	int err;
+
+#define QUERY_ADAPTER_OUT_SIZE             0x100
+#define QUERY_ADAPTER_INTA_PIN_OFFSET      0x10
+#define QUERY_ADAPTER_VSD_OFFSET           0x20
+
+	mailbox = mlx4_alloc_cmd_mailbox(dev);
+	if (IS_ERR(mailbox))
+		return PTR_ERR(mailbox);
+	outbox = mailbox->buf;
+
+	err = mlx4_cmd_box(dev, 0, mailbox->dma, 0, 0, MLX4_CMD_QUERY_ADAPTER,
+			   MLX4_CMD_TIME_CLASS_A, MLX4_CMD_NATIVE);
+	if (err)
+		goto out;
+
+	MLX4_GET(adapter->inta_pin, outbox,    QUERY_ADAPTER_INTA_PIN_OFFSET);
+
+	get_board_id(outbox + QUERY_ADAPTER_VSD_OFFSET / 4,
+		     adapter->board_id);
+
+out:
+	mlx4_free_cmd_mailbox(dev, mailbox);
+	return err;
+}
+
+int mlx4_INIT_HCA(struct mlx4_dev *dev, struct mlx4_init_hca_param *param)
+{
+	struct mlx4_cmd_mailbox *mailbox;
+	__be32 *inbox;
+	int err;
+	static const u8 a0_dmfs_hw_steering[] =  {
+		[MLX4_STEERING_DMFS_A0_DEFAULT]		= 0,
+		[MLX4_STEERING_DMFS_A0_DYNAMIC]		= 1,
+		[MLX4_STEERING_DMFS_A0_STATIC]		= 2,
+		[MLX4_STEERING_DMFS_A0_DISABLE]		= 3
+	};
+
+#define INIT_HCA_IN_SIZE		 0x200
+#define INIT_HCA_VERSION_OFFSET		 0x000
+#define	 INIT_HCA_VERSION		 2
+#define INIT_HCA_VXLAN_OFFSET		 0x0c
+#define INIT_HCA_CACHELINE_SZ_OFFSET	 0x0e
+#define INIT_HCA_FLAGS_OFFSET		 0x014
+#define INIT_HCA_RECOVERABLE_ERROR_EVENT_OFFSET 0x018
+#define INIT_HCA_QPC_OFFSET		 0x020
+#define	 INIT_HCA_QPC_BASE_OFFSET	 (INIT_HCA_QPC_OFFSET + 0x10)
+#define	 INIT_HCA_LOG_QP_OFFSET		 (INIT_HCA_QPC_OFFSET + 0x17)
+#define	 INIT_HCA_SRQC_BASE_OFFSET	 (INIT_HCA_QPC_OFFSET + 0x28)
+#define	 INIT_HCA_LOG_SRQ_OFFSET	 (INIT_HCA_QPC_OFFSET + 0x2f)
+#define	 INIT_HCA_CQC_BASE_OFFSET	 (INIT_HCA_QPC_OFFSET + 0x30)
+#define	 INIT_HCA_LOG_CQ_OFFSET		 (INIT_HCA_QPC_OFFSET + 0x37)
+#define	 INIT_HCA_EQE_CQE_OFFSETS	 (INIT_HCA_QPC_OFFSET + 0x38)
+#define	 INIT_HCA_EQE_CQE_STRIDE_OFFSET  (INIT_HCA_QPC_OFFSET + 0x3b)
+#define	 INIT_HCA_ALTC_BASE_OFFSET	 (INIT_HCA_QPC_OFFSET + 0x40)
+#define	 INIT_HCA_AUXC_BASE_OFFSET	 (INIT_HCA_QPC_OFFSET + 0x50)
+#define	 INIT_HCA_EQC_BASE_OFFSET	 (INIT_HCA_QPC_OFFSET + 0x60)
+#define	 INIT_HCA_LOG_EQ_OFFSET		 (INIT_HCA_QPC_OFFSET + 0x67)
+#define	INIT_HCA_NUM_SYS_EQS_OFFSET	(INIT_HCA_QPC_OFFSET + 0x6a)
+#define	 INIT_HCA_RDMARC_BASE_OFFSET	 (INIT_HCA_QPC_OFFSET + 0x70)
+#define	 INIT_HCA_LOG_RD_OFFSET		 (INIT_HCA_QPC_OFFSET + 0x77)
+#define INIT_HCA_MCAST_OFFSET		 0x0c0
+#define	 INIT_HCA_MC_BASE_OFFSET	 (INIT_HCA_MCAST_OFFSET + 0x00)
+#define	 INIT_HCA_LOG_MC_ENTRY_SZ_OFFSET (INIT_HCA_MCAST_OFFSET + 0x13)
+#define	 INIT_HCA_LOG_MC_HASH_SZ_OFFSET	 (INIT_HCA_MCAST_OFFSET + 0x17)
+#define  INIT_HCA_UC_STEERING_OFFSET	 (INIT_HCA_MCAST_OFFSET + 0x18)
+#define	 INIT_HCA_LOG_MC_TABLE_SZ_OFFSET (INIT_HCA_MCAST_OFFSET + 0x1b)
+#define  INIT_HCA_DEVICE_MANAGED_FLOW_STEERING_EN	0x6
+#define  INIT_HCA_DRIVER_VERSION_OFFSET   0x140
+#define  INIT_HCA_DRIVER_VERSION_SZ       0x40
+#define  INIT_HCA_FS_PARAM_OFFSET         0x1d0
+#define  INIT_HCA_FS_BASE_OFFSET          (INIT_HCA_FS_PARAM_OFFSET + 0x00)
+#define  INIT_HCA_FS_LOG_ENTRY_SZ_OFFSET  (INIT_HCA_FS_PARAM_OFFSET + 0x13)
+#define  INIT_HCA_FS_A0_OFFSET		  (INIT_HCA_FS_PARAM_OFFSET + 0x18)
+#define  INIT_HCA_FS_LOG_TABLE_SZ_OFFSET  (INIT_HCA_FS_PARAM_OFFSET + 0x1b)
+#define  INIT_HCA_FS_ETH_BITS_OFFSET      (INIT_HCA_FS_PARAM_OFFSET + 0x21)
+#define  INIT_HCA_FS_ETH_NUM_ADDRS_OFFSET (INIT_HCA_FS_PARAM_OFFSET + 0x22)
+#define  INIT_HCA_FS_IB_BITS_OFFSET       (INIT_HCA_FS_PARAM_OFFSET + 0x25)
+#define  INIT_HCA_FS_IB_NUM_ADDRS_OFFSET  (INIT_HCA_FS_PARAM_OFFSET + 0x26)
+#define INIT_HCA_TPT_OFFSET		 0x0f0
+#define	 INIT_HCA_DMPT_BASE_OFFSET	 (INIT_HCA_TPT_OFFSET + 0x00)
+#define  INIT_HCA_TPT_MW_OFFSET		 (INIT_HCA_TPT_OFFSET + 0x08)
+#define	 INIT_HCA_LOG_MPT_SZ_OFFSET	 (INIT_HCA_TPT_OFFSET + 0x0b)
+#define	 INIT_HCA_MTT_BASE_OFFSET	 (INIT_HCA_TPT_OFFSET + 0x10)
+#define	 INIT_HCA_CMPT_BASE_OFFSET	 (INIT_HCA_TPT_OFFSET + 0x18)
+#define INIT_HCA_UAR_OFFSET		 0x120
+#define	 INIT_HCA_LOG_UAR_SZ_OFFSET	 (INIT_HCA_UAR_OFFSET + 0x0a)
+#define  INIT_HCA_UAR_PAGE_SZ_OFFSET     (INIT_HCA_UAR_OFFSET + 0x0b)
+
+	mailbox = mlx4_alloc_cmd_mailbox(dev);
+	if (IS_ERR(mailbox))
+		return PTR_ERR(mailbox);
+	inbox = mailbox->buf;
+
+	*((u8 *) mailbox->buf + INIT_HCA_VERSION_OFFSET) = INIT_HCA_VERSION;
+
+	*((u8 *) mailbox->buf + INIT_HCA_CACHELINE_SZ_OFFSET) =
+		((ilog2(cache_line_size()) - 4) << 5) | (1 << 4);
+
+#if defined(__LITTLE_ENDIAN)
+	*(inbox + INIT_HCA_FLAGS_OFFSET / 4) &= ~cpu_to_be32(1 << 1);
+#elif defined(__BIG_ENDIAN)
+	*(inbox + INIT_HCA_FLAGS_OFFSET / 4) |= cpu_to_be32(1 << 1);
+#else
+#error Host endianness not defined
+#endif
+	/* Check port for UD address vector: */
+	*(inbox + INIT_HCA_FLAGS_OFFSET / 4) |= cpu_to_be32(1);
+
+	/* Enable IPoIB checksumming if we can: */
+	if (dev->caps.flags & MLX4_DEV_CAP_FLAG_IPOIB_CSUM)
+		*(inbox + INIT_HCA_FLAGS_OFFSET / 4) |= cpu_to_be32(1 << 3);
+
+	/* Enable QoS support if module parameter set */
+	if (dev->caps.flags2 & MLX4_DEV_CAP_FLAG2_ETS_CFG && enable_qos)
+		*(inbox + INIT_HCA_FLAGS_OFFSET / 4) |= cpu_to_be32(1 << 2);
+
+	/* enable counters */
+	if (dev->caps.flags & MLX4_DEV_CAP_FLAG_COUNTERS)
+		*(inbox + INIT_HCA_FLAGS_OFFSET / 4) |= cpu_to_be32(1 << 4);
+
+	/* Enable RSS spread to fragmented IP packets when supported */
+	if (dev->caps.flags & MLX4_DEV_CAP_FLAG_RSS_IP_FRAG)
+		*(inbox + INIT_HCA_FLAGS_OFFSET / 4) |= cpu_to_be32(1 << 13);
+
+	/* CX3 is capable of extending CQEs/EQEs from 32 to 64 bytes */
+	if (dev->caps.flags & MLX4_DEV_CAP_FLAG_64B_EQE) {
+		*(inbox + INIT_HCA_EQE_CQE_OFFSETS / 4) |= cpu_to_be32(1 << 29);
+		dev->caps.eqe_size   = 64;
+		dev->caps.eqe_factor = 1;
+	} else {
+		dev->caps.eqe_size   = 32;
+		dev->caps.eqe_factor = 0;
+	}
+
+	if (dev->caps.flags & MLX4_DEV_CAP_FLAG_64B_CQE) {
+		*(inbox + INIT_HCA_EQE_CQE_OFFSETS / 4) |= cpu_to_be32(1 << 30);
+		dev->caps.cqe_size   = 64;
+		dev->caps.userspace_caps |= MLX4_USER_DEV_CAP_LARGE_CQE;
+	} else {
+		dev->caps.cqe_size   = 32;
+	}
+
+	/* CX3 is capable of extending CQEs\EQEs to strides larger than 64B */
+	if ((dev->caps.flags2 & MLX4_DEV_CAP_FLAG2_EQE_STRIDE) &&
+	    (dev->caps.flags2 & MLX4_DEV_CAP_FLAG2_CQE_STRIDE)) {
+		dev->caps.eqe_size = cache_line_size();
+		dev->caps.cqe_size = cache_line_size();
+		dev->caps.eqe_factor = 0;
+		MLX4_PUT(inbox, (u8)((ilog2(dev->caps.eqe_size) - 5) << 4 |
+				      (ilog2(dev->caps.eqe_size) - 5)),
+			 INIT_HCA_EQE_CQE_STRIDE_OFFSET);
+
+		/* User still need to know to support CQE > 32B */
+		dev->caps.userspace_caps |= MLX4_USER_DEV_CAP_LARGE_CQE;
+	}
+
+	if (dev->caps.flags2 & MLX4_DEV_CAP_FLAG2_RECOVERABLE_ERROR_EVENT)
+		*(inbox + INIT_HCA_RECOVERABLE_ERROR_EVENT_OFFSET / 4) |= cpu_to_be32(1 << 31);
+
+	if (dev->caps.flags2 & MLX4_DEV_CAP_FLAG2_DRIVER_VERSION_TO_FW) {
+		u8 *dst = (u8 *)(inbox + INIT_HCA_DRIVER_VERSION_OFFSET / 4);
+
+		strncpy(dst, DRV_NAME_FOR_FW, INIT_HCA_DRIVER_VERSION_SZ - 1);
+		mlx4_dbg(dev, "Reporting Driver Version to FW: %s\n", dst);
+	}
+
+	/* QPC/EEC/CQC/EQC/RDMARC attributes */
+
+	MLX4_PUT(inbox, param->qpc_base,      INIT_HCA_QPC_BASE_OFFSET);
+	MLX4_PUT(inbox, param->log_num_qps,   INIT_HCA_LOG_QP_OFFSET);
+	MLX4_PUT(inbox, param->srqc_base,     INIT_HCA_SRQC_BASE_OFFSET);
+	MLX4_PUT(inbox, param->log_num_srqs,  INIT_HCA_LOG_SRQ_OFFSET);
+	MLX4_PUT(inbox, param->cqc_base,      INIT_HCA_CQC_BASE_OFFSET);
+	MLX4_PUT(inbox, param->log_num_cqs,   INIT_HCA_LOG_CQ_OFFSET);
+	MLX4_PUT(inbox, param->altc_base,     INIT_HCA_ALTC_BASE_OFFSET);
+	MLX4_PUT(inbox, param->auxc_base,     INIT_HCA_AUXC_BASE_OFFSET);
+	MLX4_PUT(inbox, param->eqc_base,      INIT_HCA_EQC_BASE_OFFSET);
+	MLX4_PUT(inbox, param->log_num_eqs,   INIT_HCA_LOG_EQ_OFFSET);
+	MLX4_PUT(inbox, param->num_sys_eqs,   INIT_HCA_NUM_SYS_EQS_OFFSET);
+	MLX4_PUT(inbox, param->rdmarc_base,   INIT_HCA_RDMARC_BASE_OFFSET);
+	MLX4_PUT(inbox, param->log_rd_per_qp, INIT_HCA_LOG_RD_OFFSET);
+
+	/* steering attributes */
+	if (dev->caps.steering_mode ==
+	    MLX4_STEERING_MODE_DEVICE_MANAGED) {
+		*(inbox + INIT_HCA_FLAGS_OFFSET / 4) |=
+			cpu_to_be32(1 <<
+				    INIT_HCA_DEVICE_MANAGED_FLOW_STEERING_EN);
+
+		MLX4_PUT(inbox, param->mc_base, INIT_HCA_FS_BASE_OFFSET);
+		MLX4_PUT(inbox, param->log_mc_entry_sz,
+			 INIT_HCA_FS_LOG_ENTRY_SZ_OFFSET);
+		MLX4_PUT(inbox, param->log_mc_table_sz,
+			 INIT_HCA_FS_LOG_TABLE_SZ_OFFSET);
+		/* Enable Ethernet flow steering
+		 * with udp unicast and tcp unicast
+		 */
+		if (dev->caps.dmfs_high_steer_mode !=
+		    MLX4_STEERING_DMFS_A0_STATIC)
+			MLX4_PUT(inbox,
+				 (u8)(MLX4_FS_UDP_UC_EN | MLX4_FS_TCP_UC_EN),
+				 INIT_HCA_FS_ETH_BITS_OFFSET);
+		MLX4_PUT(inbox, (u16) MLX4_FS_NUM_OF_L2_ADDR,
+			 INIT_HCA_FS_ETH_NUM_ADDRS_OFFSET);
+		/* Enable IPoIB flow steering
+		 * with udp unicast and tcp unicast
+		 */
+		MLX4_PUT(inbox, (u8) (MLX4_FS_UDP_UC_EN | MLX4_FS_TCP_UC_EN),
+			 INIT_HCA_FS_IB_BITS_OFFSET);
+		MLX4_PUT(inbox, (u16) MLX4_FS_NUM_OF_L2_ADDR,
+			 INIT_HCA_FS_IB_NUM_ADDRS_OFFSET);
+
+		if (dev->caps.dmfs_high_steer_mode !=
+		    MLX4_STEERING_DMFS_A0_NOT_SUPPORTED)
+			MLX4_PUT(inbox,
+				 ((u8)(a0_dmfs_hw_steering[dev->caps.dmfs_high_steer_mode]
+				       << 6)),
+				 INIT_HCA_FS_A0_OFFSET);
+	} else {
+		MLX4_PUT(inbox, param->mc_base,	INIT_HCA_MC_BASE_OFFSET);
+		MLX4_PUT(inbox, param->log_mc_entry_sz,
+			 INIT_HCA_LOG_MC_ENTRY_SZ_OFFSET);
+		MLX4_PUT(inbox, param->log_mc_hash_sz,
+			 INIT_HCA_LOG_MC_HASH_SZ_OFFSET);
+		MLX4_PUT(inbox, param->log_mc_table_sz,
+			 INIT_HCA_LOG_MC_TABLE_SZ_OFFSET);
+		if (dev->caps.steering_mode == MLX4_STEERING_MODE_B0)
+			MLX4_PUT(inbox, (u8) (1 << 3),
+				 INIT_HCA_UC_STEERING_OFFSET);
+	}
+
+	/* TPT attributes */
+
+	MLX4_PUT(inbox, param->dmpt_base,  INIT_HCA_DMPT_BASE_OFFSET);
+	MLX4_PUT(inbox, param->mw_enabled, INIT_HCA_TPT_MW_OFFSET);
+	MLX4_PUT(inbox, param->log_mpt_sz, INIT_HCA_LOG_MPT_SZ_OFFSET);
+	MLX4_PUT(inbox, param->mtt_base,   INIT_HCA_MTT_BASE_OFFSET);
+	MLX4_PUT(inbox, param->cmpt_base,  INIT_HCA_CMPT_BASE_OFFSET);
+
+	/* UAR attributes */
+
+	MLX4_PUT(inbox, param->uar_page_sz,	INIT_HCA_UAR_PAGE_SZ_OFFSET);
+	MLX4_PUT(inbox, param->log_uar_sz,      INIT_HCA_LOG_UAR_SZ_OFFSET);
+
+	/* set parser VXLAN attributes */
+	if (dev->caps.flags2 & MLX4_DEV_CAP_FLAG2_VXLAN_OFFLOADS) {
+		u8 parser_params = 0;
+		MLX4_PUT(inbox, parser_params,	INIT_HCA_VXLAN_OFFSET);
+	}
+
+	err = mlx4_cmd(dev, mailbox->dma, 0, 0, MLX4_CMD_INIT_HCA,
+		       MLX4_CMD_TIME_CLASS_C, MLX4_CMD_NATIVE);
+
+	if (err)
+		mlx4_err(dev, "INIT_HCA returns %d\n", err);
+
+	mlx4_free_cmd_mailbox(dev, mailbox);
+	return err;
+}
+
+int mlx4_QUERY_HCA(struct mlx4_dev *dev,
+		   struct mlx4_init_hca_param *param)
+{
+	struct mlx4_cmd_mailbox *mailbox;
+	__be32 *outbox;
+	u64 qword_field;
+	u32 dword_field;
+	u16 word_field;
+	u8 byte_field;
+	int err;
+	static const u8 a0_dmfs_query_hw_steering[] =  {
+		[0] = MLX4_STEERING_DMFS_A0_DEFAULT,
+		[1] = MLX4_STEERING_DMFS_A0_DYNAMIC,
+		[2] = MLX4_STEERING_DMFS_A0_STATIC,
+		[3] = MLX4_STEERING_DMFS_A0_DISABLE
+	};
+
+#define QUERY_HCA_GLOBAL_CAPS_OFFSET	0x04
+#define QUERY_HCA_CORE_CLOCK_OFFSET	0x0c
+
+	mailbox = mlx4_alloc_cmd_mailbox(dev);
+	if (IS_ERR(mailbox))
+		return PTR_ERR(mailbox);
+	outbox = mailbox->buf;
+
+	err = mlx4_cmd_box(dev, 0, mailbox->dma, 0, 0,
+			   MLX4_CMD_QUERY_HCA,
+			   MLX4_CMD_TIME_CLASS_B,
+			   !mlx4_is_slave(dev));
+	if (err)
+		goto out;
+
+	MLX4_GET(param->global_caps, outbox, QUERY_HCA_GLOBAL_CAPS_OFFSET);
+	MLX4_GET(param->hca_core_clock, outbox, QUERY_HCA_CORE_CLOCK_OFFSET);
+
+	/* QPC/EEC/CQC/EQC/RDMARC attributes */
+
+	MLX4_GET(qword_field, outbox, INIT_HCA_QPC_BASE_OFFSET);
+	param->qpc_base = qword_field & ~((u64)0x1f);
+	MLX4_GET(byte_field, outbox, INIT_HCA_LOG_QP_OFFSET);
+	param->log_num_qps = byte_field & 0x1f;
+	MLX4_GET(qword_field, outbox, INIT_HCA_SRQC_BASE_OFFSET);
+	param->srqc_base = qword_field & ~((u64)0x1f);
+	MLX4_GET(byte_field, outbox, INIT_HCA_LOG_SRQ_OFFSET);
+	param->log_num_srqs = byte_field & 0x1f;
+	MLX4_GET(qword_field, outbox, INIT_HCA_CQC_BASE_OFFSET);
+	param->cqc_base = qword_field & ~((u64)0x1f);
+	MLX4_GET(byte_field, outbox, INIT_HCA_LOG_CQ_OFFSET);
+	param->log_num_cqs = byte_field & 0x1f;
+	MLX4_GET(qword_field, outbox, INIT_HCA_ALTC_BASE_OFFSET);
+	param->altc_base = qword_field;
+	MLX4_GET(qword_field, outbox, INIT_HCA_AUXC_BASE_OFFSET);
+	param->auxc_base = qword_field;
+	MLX4_GET(qword_field, outbox, INIT_HCA_EQC_BASE_OFFSET);
+	param->eqc_base = qword_field & ~((u64)0x1f);
+	MLX4_GET(byte_field, outbox, INIT_HCA_LOG_EQ_OFFSET);
+	param->log_num_eqs = byte_field & 0x1f;
+	MLX4_GET(word_field, outbox, INIT_HCA_NUM_SYS_EQS_OFFSET);
+	param->num_sys_eqs = word_field & 0xfff;
+	MLX4_GET(qword_field, outbox, INIT_HCA_RDMARC_BASE_OFFSET);
+	param->rdmarc_base = qword_field & ~((u64)0x1f);
+	MLX4_GET(byte_field, outbox, INIT_HCA_LOG_RD_OFFSET);
+	param->log_rd_per_qp = byte_field & 0x7;
+
+	MLX4_GET(dword_field, outbox, INIT_HCA_FLAGS_OFFSET);
+	if (dword_field & (1 << INIT_HCA_DEVICE_MANAGED_FLOW_STEERING_EN)) {
+		param->steering_mode = MLX4_STEERING_MODE_DEVICE_MANAGED;
+	} else {
+		MLX4_GET(byte_field, outbox, INIT_HCA_UC_STEERING_OFFSET);
+		if (byte_field & 0x8)
+			param->steering_mode = MLX4_STEERING_MODE_B0;
+		else
+			param->steering_mode = MLX4_STEERING_MODE_A0;
+	}
+
+	if (dword_field & (1 << 13))
+		param->rss_ip_frags = 1;
+
+	/* steering attributes */
+	if (param->steering_mode == MLX4_STEERING_MODE_DEVICE_MANAGED) {
+		MLX4_GET(param->mc_base, outbox, INIT_HCA_FS_BASE_OFFSET);
+		MLX4_GET(byte_field, outbox, INIT_HCA_FS_LOG_ENTRY_SZ_OFFSET);
+		param->log_mc_entry_sz = byte_field & 0x1f;
+		MLX4_GET(byte_field, outbox, INIT_HCA_FS_LOG_TABLE_SZ_OFFSET);
+		param->log_mc_table_sz = byte_field & 0x1f;
+		MLX4_GET(byte_field, outbox, INIT_HCA_FS_A0_OFFSET);
+		param->dmfs_high_steer_mode =
+			a0_dmfs_query_hw_steering[(byte_field >> 6) & 3];
+	} else {
+		MLX4_GET(param->mc_base, outbox, INIT_HCA_MC_BASE_OFFSET);
+		MLX4_GET(byte_field, outbox, INIT_HCA_LOG_MC_ENTRY_SZ_OFFSET);
+		param->log_mc_entry_sz = byte_field & 0x1f;
+		MLX4_GET(byte_field,  outbox, INIT_HCA_LOG_MC_HASH_SZ_OFFSET);
+		param->log_mc_hash_sz = byte_field & 0x1f;
+		MLX4_GET(byte_field, outbox, INIT_HCA_LOG_MC_TABLE_SZ_OFFSET);
+		param->log_mc_table_sz = byte_field & 0x1f;
+	}
+
+	/* CX3 is capable of extending CQEs/EQEs from 32 to 64 bytes */
+	MLX4_GET(byte_field, outbox, INIT_HCA_EQE_CQE_OFFSETS);
+	if (byte_field & 0x20) /* 64-bytes eqe enabled */
+		param->dev_cap_enabled |= MLX4_DEV_CAP_64B_EQE_ENABLED;
+	if (byte_field & 0x40) /* 64-bytes cqe enabled */
+		param->dev_cap_enabled |= MLX4_DEV_CAP_64B_CQE_ENABLED;
+
+	/* CX3 is capable of extending CQEs\EQEs to strides larger than 64B */
+	MLX4_GET(byte_field, outbox, INIT_HCA_EQE_CQE_STRIDE_OFFSET);
+	if (byte_field) {
+		param->dev_cap_enabled |= MLX4_DEV_CAP_EQE_STRIDE_ENABLED;
+		param->dev_cap_enabled |= MLX4_DEV_CAP_CQE_STRIDE_ENABLED;
+		param->cqe_size = 1 << ((byte_field &
+					 MLX4_CQE_SIZE_MASK_STRIDE) + 5);
+		param->eqe_size = 1 << (((byte_field &
+					  MLX4_EQE_SIZE_MASK_STRIDE) >> 4) + 5);
+	}
+
+	/* TPT attributes */
+
+	MLX4_GET(param->dmpt_base,  outbox, INIT_HCA_DMPT_BASE_OFFSET);
+	MLX4_GET(byte_field, outbox, INIT_HCA_TPT_MW_OFFSET);
+	param->mw_enabled = byte_field >> 7;
+	MLX4_GET(byte_field, outbox, INIT_HCA_LOG_MPT_SZ_OFFSET);
+	param->log_mpt_sz = byte_field & 0x3f;
+	MLX4_GET(param->mtt_base,   outbox, INIT_HCA_MTT_BASE_OFFSET);
+	MLX4_GET(param->cmpt_base,  outbox, INIT_HCA_CMPT_BASE_OFFSET);
+
+	/* UAR attributes */
+
+	MLX4_GET(param->uar_page_sz, outbox, INIT_HCA_UAR_PAGE_SZ_OFFSET);
+	MLX4_GET(byte_field, outbox, INIT_HCA_LOG_UAR_SZ_OFFSET);
+	param->log_uar_sz = byte_field & 0xf;
+
+	/* phv_check enable */
+	MLX4_GET(byte_field, outbox, INIT_HCA_CACHELINE_SZ_OFFSET);
+	if (byte_field & 0x2)
+		param->phv_check_en = 1;
+out:
+	mlx4_free_cmd_mailbox(dev, mailbox);
+
+	return err;
+}
+
+static int mlx4_hca_core_clock_update(struct mlx4_dev *dev)
+{
+	struct mlx4_cmd_mailbox *mailbox;
+	__be32 *outbox;
+	int err;
+
+	mailbox = mlx4_alloc_cmd_mailbox(dev);
+	if (IS_ERR(mailbox)) {
+		mlx4_warn(dev, "hca_core_clock mailbox allocation failed\n");
+		return PTR_ERR(mailbox);
+	}
+	outbox = mailbox->buf;
+
+	err = mlx4_cmd_box(dev, 0, mailbox->dma, 0, 0,
+			   MLX4_CMD_QUERY_HCA,
+			   MLX4_CMD_TIME_CLASS_B,
+			   !mlx4_is_slave(dev));
+	if (err) {
+		mlx4_warn(dev, "hca_core_clock update failed\n");
+		goto out;
+	}
+
+	MLX4_GET(dev->caps.hca_core_clock, outbox, QUERY_HCA_CORE_CLOCK_OFFSET);
+
+out:
+	mlx4_free_cmd_mailbox(dev, mailbox);
+
+	return err;
+}
+
+/* for IB-type ports only in SRIOV mode. Checks that both proxy QP0
+ * and real QP0 are active, so that the paravirtualized QP0 is ready
+ * to operate */
+static int check_qp0_state(struct mlx4_dev *dev, int function, int port)
+{
+	struct mlx4_priv *priv = mlx4_priv(dev);
+	/* irrelevant if not infiniband */
+	if (priv->mfunc.master.qp0_state[port].proxy_qp0_active &&
+	    priv->mfunc.master.qp0_state[port].qp0_active)
+		return 1;
+	return 0;
+}
+
+int mlx4_INIT_PORT_wrapper(struct mlx4_dev *dev, int slave,
+			   struct mlx4_vhcr *vhcr,
+			   struct mlx4_cmd_mailbox *inbox,
+			   struct mlx4_cmd_mailbox *outbox,
+			   struct mlx4_cmd_info *cmd)
+{
+	struct mlx4_priv *priv = mlx4_priv(dev);
+	int port = mlx4_slave_convert_port(dev, slave, vhcr->in_modifier);
+	int err;
+
+	if (port < 0)
+		return -EINVAL;
+
+	if (priv->mfunc.master.slave_state[slave].init_port_mask & (1 << port))
+		return 0;
+
+	if (dev->caps.port_mask[port] != MLX4_PORT_TYPE_IB) {
+		/* Enable port only if it was previously disabled */
+		if (!priv->mfunc.master.init_port_ref[port]) {
+			err = mlx4_cmd(dev, 0, port, 0, MLX4_CMD_INIT_PORT,
+				       MLX4_CMD_TIME_CLASS_A, MLX4_CMD_NATIVE);
+			if (err)
+				return err;
+		}
+		priv->mfunc.master.slave_state[slave].init_port_mask |= (1 << port);
+	} else {
+		if (slave == mlx4_master_func_num(dev)) {
+			if (check_qp0_state(dev, slave, port) &&
+			    !priv->mfunc.master.qp0_state[port].port_active) {
+				err = mlx4_cmd(dev, 0, port, 0, MLX4_CMD_INIT_PORT,
+					       MLX4_CMD_TIME_CLASS_A, MLX4_CMD_NATIVE);
+				if (err)
+					return err;
+				priv->mfunc.master.qp0_state[port].port_active = 1;
+				priv->mfunc.master.slave_state[slave].init_port_mask |= (1 << port);
+			}
+		} else
+			priv->mfunc.master.slave_state[slave].init_port_mask |= (1 << port);
+	}
+	++priv->mfunc.master.init_port_ref[port];
+	return 0;
+}
+
+int mlx4_INIT_PORT(struct mlx4_dev *dev, int port)
+{
+	struct mlx4_cmd_mailbox *mailbox;
+	u32 *inbox;
+	int err;
+	u32 flags;
+	u16 field;
+
+	if (dev->flags & MLX4_FLAG_OLD_PORT_CMDS) {
+#define INIT_PORT_IN_SIZE          256
+#define INIT_PORT_FLAGS_OFFSET     0x00
+#define INIT_PORT_FLAG_SIG         (1 << 18)
+#define INIT_PORT_FLAG_NG          (1 << 17)
+#define INIT_PORT_FLAG_G0          (1 << 16)
+#define INIT_PORT_VL_SHIFT         4
+#define INIT_PORT_PORT_WIDTH_SHIFT 8
+#define INIT_PORT_MTU_OFFSET       0x04
+#define INIT_PORT_MAX_GID_OFFSET   0x06
+#define INIT_PORT_MAX_PKEY_OFFSET  0x0a
+#define INIT_PORT_GUID0_OFFSET     0x10
+#define INIT_PORT_NODE_GUID_OFFSET 0x18
+#define INIT_PORT_SI_GUID_OFFSET   0x20
+
+		mailbox = mlx4_alloc_cmd_mailbox(dev);
+		if (IS_ERR(mailbox))
+			return PTR_ERR(mailbox);
+		inbox = mailbox->buf;
+
+		flags = 0;
+		flags |= (dev->caps.vl_cap[port] & 0xf) << INIT_PORT_VL_SHIFT;
+		flags |= (dev->caps.port_width_cap[port] & 0xf) << INIT_PORT_PORT_WIDTH_SHIFT;
+		MLX4_PUT(inbox, flags,		  INIT_PORT_FLAGS_OFFSET);
+
+		field = 128 << dev->caps.ib_mtu_cap[port];
+		MLX4_PUT(inbox, field, INIT_PORT_MTU_OFFSET);
+		field = dev->caps.gid_table_len[port];
+		MLX4_PUT(inbox, field, INIT_PORT_MAX_GID_OFFSET);
+		field = dev->caps.pkey_table_len[port];
+		MLX4_PUT(inbox, field, INIT_PORT_MAX_PKEY_OFFSET);
+
+		err = mlx4_cmd(dev, mailbox->dma, port, 0, MLX4_CMD_INIT_PORT,
+			       MLX4_CMD_TIME_CLASS_A, MLX4_CMD_NATIVE);
+
+		mlx4_free_cmd_mailbox(dev, mailbox);
+	} else
+		err = mlx4_cmd(dev, 0, port, 0, MLX4_CMD_INIT_PORT,
+			       MLX4_CMD_TIME_CLASS_A, MLX4_CMD_WRAPPED);
+
+	if (!err)
+		mlx4_hca_core_clock_update(dev);
+
+	return err;
+}
+EXPORT_SYMBOL_GPL(mlx4_INIT_PORT);
+
+int mlx4_CLOSE_PORT_wrapper(struct mlx4_dev *dev, int slave,
+			    struct mlx4_vhcr *vhcr,
+			    struct mlx4_cmd_mailbox *inbox,
+			    struct mlx4_cmd_mailbox *outbox,
+			    struct mlx4_cmd_info *cmd)
+{
+	struct mlx4_priv *priv = mlx4_priv(dev);
+	int port = mlx4_slave_convert_port(dev, slave, vhcr->in_modifier);
+	int err;
+
+	if (port < 0)
+		return -EINVAL;
+
+	if (!(priv->mfunc.master.slave_state[slave].init_port_mask &
+	    (1 << port)))
+		return 0;
+
+	if (dev->caps.port_mask[port] != MLX4_PORT_TYPE_IB) {
+		if (priv->mfunc.master.init_port_ref[port] == 1) {
+			err = mlx4_cmd(dev, 0, port, 0, MLX4_CMD_CLOSE_PORT,
+				       MLX4_CMD_TIME_CLASS_A, MLX4_CMD_NATIVE);
+			if (err)
+				return err;
+		}
+		priv->mfunc.master.slave_state[slave].init_port_mask &= ~(1 << port);
+	} else {
+		/* infiniband port */
+		if (slave == mlx4_master_func_num(dev)) {
+			if (!priv->mfunc.master.qp0_state[port].qp0_active &&
+			    priv->mfunc.master.qp0_state[port].port_active) {
+				err = mlx4_cmd(dev, 0, port, 0, MLX4_CMD_CLOSE_PORT,
+					       MLX4_CMD_TIME_CLASS_A, MLX4_CMD_NATIVE);
+				if (err)
+					return err;
+				priv->mfunc.master.slave_state[slave].init_port_mask &= ~(1 << port);
+				priv->mfunc.master.qp0_state[port].port_active = 0;
+			}
+		} else
+			priv->mfunc.master.slave_state[slave].init_port_mask &= ~(1 << port);
+	}
+	--priv->mfunc.master.init_port_ref[port];
+	return 0;
+}
+
+int mlx4_CLOSE_PORT(struct mlx4_dev *dev, int port)
+{
+	return mlx4_cmd(dev, 0, port, 0, MLX4_CMD_CLOSE_PORT,
+			MLX4_CMD_TIME_CLASS_A, MLX4_CMD_WRAPPED);
+}
+EXPORT_SYMBOL_GPL(mlx4_CLOSE_PORT);
+
+int mlx4_CLOSE_HCA(struct mlx4_dev *dev, int panic)
+{
+	return mlx4_cmd(dev, 0, 0, panic, MLX4_CMD_CLOSE_HCA,
+			MLX4_CMD_TIME_CLASS_C, MLX4_CMD_NATIVE);
+}
+
+struct mlx4_config_dev {
+	__be32	update_flags;
+	__be32	rsvd1[3];
+	__be16	vxlan_udp_dport;
+	__be16	rsvd2;
+	__be16  roce_v2_entropy;
+	__be16  roce_v2_udp_dport;
+	__be32	roce_flags;
+	__be32	rsvd4[25];
+	__be16	rsvd5;
+	u8	rsvd6;
+	u8	rx_checksum_val;
+};
+
+#define MLX4_VXLAN_UDP_DPORT (1 << 0)
+#define MLX4_ROCE_V2_UDP_DPORT BIT(3)
+#define MLX4_DISABLE_RX_PORT BIT(18)
+
+static int mlx4_CONFIG_DEV_set(struct mlx4_dev *dev, struct mlx4_config_dev *config_dev)
+{
+	int err;
+	struct mlx4_cmd_mailbox *mailbox;
+
+	mailbox = mlx4_alloc_cmd_mailbox(dev);
+	if (IS_ERR(mailbox))
+		return PTR_ERR(mailbox);
+
+	memcpy(mailbox->buf, config_dev, sizeof(*config_dev));
+
+	err = mlx4_cmd(dev, mailbox->dma, 0, 0, MLX4_CMD_CONFIG_DEV,
+		       MLX4_CMD_TIME_CLASS_B, MLX4_CMD_NATIVE);
+
+	mlx4_free_cmd_mailbox(dev, mailbox);
+	return err;
+}
+
+static int mlx4_CONFIG_DEV_get(struct mlx4_dev *dev, struct mlx4_config_dev *config_dev)
+{
+	int err;
+	struct mlx4_cmd_mailbox *mailbox;
+
+	mailbox = mlx4_alloc_cmd_mailbox(dev);
+	if (IS_ERR(mailbox))
+		return PTR_ERR(mailbox);
+
+	err = mlx4_cmd_box(dev, 0, mailbox->dma, 0, 1, MLX4_CMD_CONFIG_DEV,
+			   MLX4_CMD_TIME_CLASS_A, MLX4_CMD_NATIVE);
+
+	if (!err)
+		memcpy(config_dev, mailbox->buf, sizeof(*config_dev));
+
+	mlx4_free_cmd_mailbox(dev, mailbox);
+	return err;
+}
+
+/* Conversion between the HW values and the actual functionality.
+ * The value represented by the array index,
+ * and the functionality determined by the flags.
+ */
+static const u8 config_dev_csum_flags[] = {
+	[0] =	0,
+	[1] =	MLX4_RX_CSUM_MODE_VAL_NON_TCP_UDP,
+	[2] =	MLX4_RX_CSUM_MODE_VAL_NON_TCP_UDP	|
+		MLX4_RX_CSUM_MODE_L4,
+	[3] =	MLX4_RX_CSUM_MODE_L4			|
+		MLX4_RX_CSUM_MODE_IP_OK_IP_NON_TCP_UDP	|
+		MLX4_RX_CSUM_MODE_MULTI_VLAN
+};
+
+int mlx4_config_dev_retrieval(struct mlx4_dev *dev,
+			      struct mlx4_config_dev_params *params)
+{
+	struct mlx4_config_dev config_dev = {0};
+	int err;
+	u8 csum_mask;
+
+#define CONFIG_DEV_RX_CSUM_MODE_MASK			0x7
+#define CONFIG_DEV_RX_CSUM_MODE_PORT1_BIT_OFFSET	0
+#define CONFIG_DEV_RX_CSUM_MODE_PORT2_BIT_OFFSET	4
+
+	if (!(dev->caps.flags2 & MLX4_DEV_CAP_FLAG2_CONFIG_DEV))
+		return -EOPNOTSUPP;
+
+	err = mlx4_CONFIG_DEV_get(dev, &config_dev);
+	if (err)
+		return err;
+
+	csum_mask = (config_dev.rx_checksum_val >> CONFIG_DEV_RX_CSUM_MODE_PORT1_BIT_OFFSET) &
+			CONFIG_DEV_RX_CSUM_MODE_MASK;
+
+	if (csum_mask >= ARRAY_SIZE(config_dev_csum_flags))
+		return -EINVAL;
+	params->rx_csum_flags_port_1 = config_dev_csum_flags[csum_mask];
+
+	csum_mask = (config_dev.rx_checksum_val >> CONFIG_DEV_RX_CSUM_MODE_PORT2_BIT_OFFSET) &
+			CONFIG_DEV_RX_CSUM_MODE_MASK;
+
+	if (csum_mask >= ARRAY_SIZE(config_dev_csum_flags))
+		return -EINVAL;
+	params->rx_csum_flags_port_2 = config_dev_csum_flags[csum_mask];
+
+	params->vxlan_udp_dport = be16_to_cpu(config_dev.vxlan_udp_dport);
+
+	return 0;
+}
+EXPORT_SYMBOL_GPL(mlx4_config_dev_retrieval);
+
+int mlx4_config_vxlan_port(struct mlx4_dev *dev, __be16 udp_port)
+{
+	struct mlx4_config_dev config_dev;
+
+	memset(&config_dev, 0, sizeof(config_dev));
+	config_dev.update_flags    = cpu_to_be32(MLX4_VXLAN_UDP_DPORT);
+	config_dev.vxlan_udp_dport = udp_port;
+
+	return mlx4_CONFIG_DEV_set(dev, &config_dev);
+}
+EXPORT_SYMBOL_GPL(mlx4_config_vxlan_port);
+
+#define CONFIG_DISABLE_RX_PORT BIT(15)
+int mlx4_disable_rx_port_check(struct mlx4_dev *dev, bool dis)
+{
+	struct mlx4_config_dev config_dev;
+
+	memset(&config_dev, 0, sizeof(config_dev));
+	config_dev.update_flags = cpu_to_be32(MLX4_DISABLE_RX_PORT);
+	if (dis)
+		config_dev.roce_flags =
+			cpu_to_be32(CONFIG_DISABLE_RX_PORT);
+
+	return mlx4_CONFIG_DEV_set(dev, &config_dev);
+}
+
+int mlx4_config_roce_v2_port(struct mlx4_dev *dev, u16 udp_port)
+{
+	struct mlx4_config_dev config_dev;
+
+	memset(&config_dev, 0, sizeof(config_dev));
+	config_dev.update_flags    = cpu_to_be32(MLX4_ROCE_V2_UDP_DPORT);
+	config_dev.roce_v2_udp_dport = cpu_to_be16(udp_port);
+
+	return mlx4_CONFIG_DEV_set(dev, &config_dev);
+}
+EXPORT_SYMBOL_GPL(mlx4_config_roce_v2_port);
+
+int mlx4_virt2phy_port_map(struct mlx4_dev *dev, u32 port1, u32 port2)
+{
+	struct mlx4_cmd_mailbox *mailbox;
+	struct {
+		__be32 v_port1;
+		__be32 v_port2;
+	} *v2p;
+	int err;
+
+	mailbox = mlx4_alloc_cmd_mailbox(dev);
+	if (IS_ERR(mailbox))
+		return -ENOMEM;
+
+	v2p = mailbox->buf;
+	v2p->v_port1 = cpu_to_be32(port1);
+	v2p->v_port2 = cpu_to_be32(port2);
+
+	err = mlx4_cmd(dev, mailbox->dma, 0,
+		       MLX4_SET_PORT_VIRT2PHY, MLX4_CMD_VIRT_PORT_MAP,
+		       MLX4_CMD_TIME_CLASS_B, MLX4_CMD_NATIVE);
+
+	mlx4_free_cmd_mailbox(dev, mailbox);
+	return err;
+}
+
+
+int mlx4_SET_ICM_SIZE(struct mlx4_dev *dev, u64 icm_size, u64 *aux_pages)
+{
+	int ret = mlx4_cmd_imm(dev, icm_size, aux_pages, 0, 0,
+			       MLX4_CMD_SET_ICM_SIZE,
+			       MLX4_CMD_TIME_CLASS_A, MLX4_CMD_NATIVE);
+	if (ret)
+		return ret;
+
+	/*
+	 * Round up number of system pages needed in case
+	 * MLX4_ICM_PAGE_SIZE < PAGE_SIZE.
+	 */
+	*aux_pages = ALIGN(*aux_pages, PAGE_SIZE / MLX4_ICM_PAGE_SIZE) >>
+		(PAGE_SHIFT - MLX4_ICM_PAGE_SHIFT);
+
+	return 0;
+}
+
+int mlx4_NOP(struct mlx4_dev *dev)
+{
+	/* Input modifier of 0x1f means "finish as soon as possible." */
+	return mlx4_cmd(dev, 0, 0x1f, 0, MLX4_CMD_NOP, MLX4_CMD_TIME_CLASS_A,
+			MLX4_CMD_NATIVE);
+}
+
+int mlx4_query_diag_counters(struct mlx4_dev *dev, u8 op_modifier,
+			     const u32 offset[],
+			     u32 value[], size_t array_len, u8 port)
+{
+	struct mlx4_cmd_mailbox *mailbox;
+	u32 *outbox;
+	size_t i;
+	int ret;
+
+	mailbox = mlx4_alloc_cmd_mailbox(dev);
+	if (IS_ERR(mailbox))
+		return PTR_ERR(mailbox);
+
+	outbox = mailbox->buf;
+
+	ret = mlx4_cmd_box(dev, 0, mailbox->dma, port, op_modifier,
+			   MLX4_CMD_DIAG_RPRT, MLX4_CMD_TIME_CLASS_A,
+			   MLX4_CMD_NATIVE);
+	if (ret)
+		goto out;
+
+	for (i = 0; i < array_len; i++) {
+		if (offset[i] > MLX4_MAILBOX_SIZE) {
+			ret = -EINVAL;
+			goto out;
+		}
+
+		MLX4_GET(value[i], outbox, offset[i]);
+	}
+
+out:
+	mlx4_free_cmd_mailbox(dev, mailbox);
+	return ret;
+}
+EXPORT_SYMBOL(mlx4_query_diag_counters);
+
+int mlx4_get_phys_port_id(struct mlx4_dev *dev)
+{
+	u8 port;
+	u32 *outbox;
+	struct mlx4_cmd_mailbox *mailbox;
+	u32 in_mod;
+	u32 guid_hi, guid_lo;
+	int err, ret = 0;
+#define MOD_STAT_CFG_PORT_OFFSET 8
+#define MOD_STAT_CFG_GUID_H	 0X14
+#define MOD_STAT_CFG_GUID_L	 0X1c
+
+	mailbox = mlx4_alloc_cmd_mailbox(dev);
+	if (IS_ERR(mailbox))
+		return PTR_ERR(mailbox);
+	outbox = mailbox->buf;
+
+	for (port = 1; port <= dev->caps.num_ports; port++) {
+		in_mod = port << MOD_STAT_CFG_PORT_OFFSET;
+		err = mlx4_cmd_box(dev, 0, mailbox->dma, in_mod, 0x2,
+				   MLX4_CMD_MOD_STAT_CFG, MLX4_CMD_TIME_CLASS_A,
+				   MLX4_CMD_NATIVE);
+		if (err) {
+			mlx4_err(dev, "Fail to get port %d uplink guid\n",
+				 port);
+			ret = err;
+		} else {
+			MLX4_GET(guid_hi, outbox, MOD_STAT_CFG_GUID_H);
+			MLX4_GET(guid_lo, outbox, MOD_STAT_CFG_GUID_L);
+			dev->caps.phys_port_id[port] = (u64)guid_lo |
+						       (u64)guid_hi << 32;
+		}
+	}
+	mlx4_free_cmd_mailbox(dev, mailbox);
+	return ret;
+}
+
+#define MLX4_WOL_SETUP_MODE (5 << 28)
+int mlx4_wol_read(struct mlx4_dev *dev, u64 *config, int port)
+{
+	u32 in_mod = MLX4_WOL_SETUP_MODE | port << 8;
+
+	return mlx4_cmd_imm(dev, 0, config, in_mod, 0x3,
+			    MLX4_CMD_MOD_STAT_CFG, MLX4_CMD_TIME_CLASS_A,
+			    MLX4_CMD_NATIVE);
+}
+EXPORT_SYMBOL_GPL(mlx4_wol_read);
+
+int mlx4_wol_write(struct mlx4_dev *dev, u64 config, int port)
+{
+	u32 in_mod = MLX4_WOL_SETUP_MODE | port << 8;
+
+	return mlx4_cmd(dev, config, in_mod, 0x1, MLX4_CMD_MOD_STAT_CFG,
+			MLX4_CMD_TIME_CLASS_A, MLX4_CMD_NATIVE);
+}
+EXPORT_SYMBOL_GPL(mlx4_wol_write);
+
+enum {
+	ADD_TO_MCG = 0x26,
+};
+
+
+void mlx4_opreq_action(struct work_struct *work)
+{
+	struct mlx4_priv *priv = container_of(work, struct mlx4_priv,
+					      opreq_task);
+	struct mlx4_dev *dev = &priv->dev;
+	int num_tasks = atomic_read(&priv->opreq_count);
+	struct mlx4_cmd_mailbox *mailbox;
+	struct mlx4_mgm *mgm;
+	u32 *outbox;
+	u32 modifier;
+	u16 token;
+	u16 type;
+	int err;
+	u32 num_qps;
+	struct mlx4_qp qp;
+	int i;
+	u8 rem_mcg;
+	u8 prot;
+
+#define GET_OP_REQ_MODIFIER_OFFSET	0x08
+#define GET_OP_REQ_TOKEN_OFFSET		0x14
+#define GET_OP_REQ_TYPE_OFFSET		0x1a
+#define GET_OP_REQ_DATA_OFFSET		0x20
+
+	mailbox = mlx4_alloc_cmd_mailbox(dev);
+	if (IS_ERR(mailbox)) {
+		mlx4_err(dev, "Failed to allocate mailbox for GET_OP_REQ\n");
+		return;
+	}
+	outbox = mailbox->buf;
+
+	while (num_tasks) {
+		err = mlx4_cmd_box(dev, 0, mailbox->dma, 0, 0,
+				   MLX4_CMD_GET_OP_REQ, MLX4_CMD_TIME_CLASS_A,
+				   MLX4_CMD_NATIVE);
+		if (err) {
+			mlx4_err(dev, "Failed to retrieve required operation: %d\n",
+				 err);
+			goto out;
+		}
+		MLX4_GET(modifier, outbox, GET_OP_REQ_MODIFIER_OFFSET);
+		MLX4_GET(token, outbox, GET_OP_REQ_TOKEN_OFFSET);
+		MLX4_GET(type, outbox, GET_OP_REQ_TYPE_OFFSET);
+		type &= 0xfff;
+
+		switch (type) {
+		case ADD_TO_MCG:
+			if (dev->caps.steering_mode ==
+			    MLX4_STEERING_MODE_DEVICE_MANAGED) {
+				mlx4_warn(dev, "ADD MCG operation is not supported in DEVICE_MANAGED steering mode\n");
+				err = EPERM;
+				break;
+			}
+			mgm = (struct mlx4_mgm *)((u8 *)(outbox) +
+						  GET_OP_REQ_DATA_OFFSET);
+			num_qps = be32_to_cpu(mgm->members_count) &
+				  MGM_QPN_MASK;
+			rem_mcg = ((u8 *)(&mgm->members_count))[0] & 1;
+			prot = ((u8 *)(&mgm->members_count))[0] >> 6;
+
+			for (i = 0; i < num_qps; i++) {
+				qp.qpn = be32_to_cpu(mgm->qp[i]);
+				if (rem_mcg)
+					err = mlx4_multicast_detach(dev, &qp,
+								    mgm->gid,
+								    prot, 0);
+				else
+					err = mlx4_multicast_attach(dev, &qp,
+								    mgm->gid,
+								    mgm->gid[5]
+								    , 0, prot,
+								    NULL);
+				if (err)
+					break;
+			}
+			break;
+		default:
+			mlx4_warn(dev, "Bad type for required operation\n");
+			err = EINVAL;
+			break;
+		}
+		err = mlx4_cmd(dev, 0, ((u32) err |
+					(__force u32)cpu_to_be32(token) << 16),
+			       1, MLX4_CMD_GET_OP_REQ, MLX4_CMD_TIME_CLASS_A,
+			       MLX4_CMD_NATIVE);
+		if (err) {
+			mlx4_err(dev, "Failed to acknowledge required request: %d\n",
+				 err);
+			goto out;
+		}
+		memset(outbox, 0, 0xffc);
+		num_tasks = atomic_dec_return(&priv->opreq_count);
+	}
+
+out:
+	mlx4_free_cmd_mailbox(dev, mailbox);
+}
+
+static int mlx4_check_smp_firewall_active(struct mlx4_dev *dev,
+					  struct mlx4_cmd_mailbox *mailbox)
+{
+#define MLX4_CMD_MAD_DEMUX_SET_ATTR_OFFSET		0x10
+#define MLX4_CMD_MAD_DEMUX_GETRESP_ATTR_OFFSET		0x20
+#define MLX4_CMD_MAD_DEMUX_TRAP_ATTR_OFFSET		0x40
+#define MLX4_CMD_MAD_DEMUX_TRAP_REPRESS_ATTR_OFFSET	0x70
+
+	u32 set_attr_mask, getresp_attr_mask;
+	u32 trap_attr_mask, traprepress_attr_mask;
+
+	MLX4_GET(set_attr_mask, mailbox->buf,
+		 MLX4_CMD_MAD_DEMUX_SET_ATTR_OFFSET);
+	mlx4_dbg(dev, "SMP firewall set_attribute_mask = 0x%x\n",
+		 set_attr_mask);
+
+	MLX4_GET(getresp_attr_mask, mailbox->buf,
+		 MLX4_CMD_MAD_DEMUX_GETRESP_ATTR_OFFSET);
+	mlx4_dbg(dev, "SMP firewall getresp_attribute_mask = 0x%x\n",
+		 getresp_attr_mask);
+
+	MLX4_GET(trap_attr_mask, mailbox->buf,
+		 MLX4_CMD_MAD_DEMUX_TRAP_ATTR_OFFSET);
+	mlx4_dbg(dev, "SMP firewall trap_attribute_mask = 0x%x\n",
+		 trap_attr_mask);
+
+	MLX4_GET(traprepress_attr_mask, mailbox->buf,
+		 MLX4_CMD_MAD_DEMUX_TRAP_REPRESS_ATTR_OFFSET);
+	mlx4_dbg(dev, "SMP firewall traprepress_attribute_mask = 0x%x\n",
+		 traprepress_attr_mask);
+
+	if (set_attr_mask && getresp_attr_mask && trap_attr_mask &&
+	    traprepress_attr_mask)
+		return 1;
+
+	return 0;
+}
+
+int mlx4_config_mad_demux(struct mlx4_dev *dev)
+{
+	struct mlx4_cmd_mailbox *mailbox;
+	int err;
+
+	/* Check if mad_demux is supported */
+	if (!(dev->caps.flags2 & MLX4_DEV_CAP_FLAG2_MAD_DEMUX))
+		return 0;
+
+	mailbox = mlx4_alloc_cmd_mailbox(dev);
+	if (IS_ERR(mailbox)) {
+		mlx4_warn(dev, "Failed to allocate mailbox for cmd MAD_DEMUX");
+		return -ENOMEM;
+	}
+
+	/* Query mad_demux to find out which MADs are handled by internal sma */
+	err = mlx4_cmd_box(dev, 0, mailbox->dma, 0x01 /* subn mgmt class */,
+			   MLX4_CMD_MAD_DEMUX_QUERY_RESTR, MLX4_CMD_MAD_DEMUX,
+			   MLX4_CMD_TIME_CLASS_B, MLX4_CMD_NATIVE);
+	if (err) {
+		mlx4_warn(dev, "MLX4_CMD_MAD_DEMUX: query restrictions failed (%d)\n",
+			  err);
+		goto out;
+	}
+
+	if (mlx4_check_smp_firewall_active(dev, mailbox))
+		dev->flags |= MLX4_FLAG_SECURE_HOST;
+
+	/* Config mad_demux to handle all MADs returned by the query above */
+	err = mlx4_cmd(dev, mailbox->dma, 0x01 /* subn mgmt class */,
+		       MLX4_CMD_MAD_DEMUX_CONFIG, MLX4_CMD_MAD_DEMUX,
+		       MLX4_CMD_TIME_CLASS_B, MLX4_CMD_NATIVE);
+	if (err) {
+		mlx4_warn(dev, "MLX4_CMD_MAD_DEMUX: configure failed (%d)\n", err);
+		goto out;
+	}
+
+	if (dev->flags & MLX4_FLAG_SECURE_HOST)
+		mlx4_warn(dev, "HCA operating in secure-host mode. SMP firewall activated.\n");
+out:
+	mlx4_free_cmd_mailbox(dev, mailbox);
+	return err;
+}
+
+/* Access Reg commands */
+enum mlx4_access_reg_masks {
+	MLX4_ACCESS_REG_STATUS_MASK = 0x7f,
+	MLX4_ACCESS_REG_METHOD_MASK = 0x7f,
+	MLX4_ACCESS_REG_LEN_MASK = 0x7ff
+};
+
+struct mlx4_access_reg {
+	__be16 constant1;
+	u8 status;
+	u8 resrvd1;
+	__be16 reg_id;
+	u8 method;
+	u8 constant2;
+	__be32 resrvd2[2];
+	__be16 len_const;
+	__be16 resrvd3;
+#define MLX4_ACCESS_REG_HEADER_SIZE (20)
+	u8 reg_data[MLX4_MAILBOX_SIZE-MLX4_ACCESS_REG_HEADER_SIZE];
+} __attribute__((__packed__));
+
+/**
+ * mlx4_ACCESS_REG - Generic access reg command.
+ * @dev: mlx4_dev.
+ * @reg_id: register ID to access.
+ * @method: Access method Read/Write.
+ * @reg_len: register length to Read/Write in bytes.
+ * @reg_data: reg_data pointer to Read/Write From/To.
+ *
+ * Access ConnectX registers FW command.
+ * Returns 0 on success and copies outbox mlx4_access_reg data
+ * field into reg_data or a negative error code.
+ */
+static int mlx4_ACCESS_REG(struct mlx4_dev *dev, u16 reg_id,
+			   enum mlx4_access_reg_method method,
+			   u16 reg_len, void *reg_data)
+{
+	struct mlx4_cmd_mailbox *inbox, *outbox;
+	struct mlx4_access_reg *inbuf, *outbuf;
+	int err;
+
+	inbox = mlx4_alloc_cmd_mailbox(dev);
+	if (IS_ERR(inbox))
+		return PTR_ERR(inbox);
+
+	outbox = mlx4_alloc_cmd_mailbox(dev);
+	if (IS_ERR(outbox)) {
+		mlx4_free_cmd_mailbox(dev, inbox);
+		return PTR_ERR(outbox);
+	}
+
+	inbuf = inbox->buf;
+	outbuf = outbox->buf;
+
+	inbuf->constant1 = cpu_to_be16(0x1<<11 | 0x4);
+	inbuf->constant2 = 0x1;
+	inbuf->reg_id = cpu_to_be16(reg_id);
+	inbuf->method = method & MLX4_ACCESS_REG_METHOD_MASK;
+
+	reg_len = min(reg_len, (u16)(sizeof(inbuf->reg_data)));
+	inbuf->len_const =
+		cpu_to_be16(((reg_len/4 + 1) & MLX4_ACCESS_REG_LEN_MASK) |
+			    ((0x3) << 12));
+
+	memcpy(inbuf->reg_data, reg_data, reg_len);
+	err = mlx4_cmd_box(dev, inbox->dma, outbox->dma, 0, 0,
+			   MLX4_CMD_ACCESS_REG, MLX4_CMD_TIME_CLASS_C,
+			   MLX4_CMD_WRAPPED);
+	if (err)
+		goto out;
+
+	if (outbuf->status & MLX4_ACCESS_REG_STATUS_MASK) {
+		err = outbuf->status & MLX4_ACCESS_REG_STATUS_MASK;
+		mlx4_err(dev,
+			 "MLX4_CMD_ACCESS_REG(%x) returned REG status (%x)\n",
+			 reg_id, err);
+		goto out;
+	}
+
+	memcpy(reg_data, outbuf->reg_data, reg_len);
+out:
+	mlx4_free_cmd_mailbox(dev, inbox);
+	mlx4_free_cmd_mailbox(dev, outbox);
+	return err;
+}
+
+/* ConnectX registers IDs */
+enum mlx4_reg_id {
+	MLX4_REG_ID_PTYS = 0x5004,
+};
+
+/**
+ * mlx4_ACCESS_PTYS_REG - Access PTYs (Port Type and Speed)
+ * register
+ * @dev: mlx4_dev.
+ * @method: Access method Read/Write.
+ * @ptys_reg: PTYS register data pointer.
+ *
+ * Access ConnectX PTYS register, to Read/Write Port Type/Speed
+ * configuration
+ * Returns 0 on success or a negative error code.
+ */
+int mlx4_ACCESS_PTYS_REG(struct mlx4_dev *dev,
+			 enum mlx4_access_reg_method method,
+			 struct mlx4_ptys_reg *ptys_reg)
+{
+	return mlx4_ACCESS_REG(dev, MLX4_REG_ID_PTYS,
+			       method, sizeof(*ptys_reg), ptys_reg);
+}
+EXPORT_SYMBOL_GPL(mlx4_ACCESS_PTYS_REG);
+
+int mlx4_ACCESS_REG_wrapper(struct mlx4_dev *dev, int slave,
+			    struct mlx4_vhcr *vhcr,
+			    struct mlx4_cmd_mailbox *inbox,
+			    struct mlx4_cmd_mailbox *outbox,
+			    struct mlx4_cmd_info *cmd)
+{
+	struct mlx4_access_reg *inbuf = inbox->buf;
+	u8 method = inbuf->method & MLX4_ACCESS_REG_METHOD_MASK;
+	u16 reg_id = be16_to_cpu(inbuf->reg_id);
+
+	if (slave != mlx4_master_func_num(dev) &&
+	    method == MLX4_ACCESS_REG_WRITE)
+		return -EPERM;
+
+	if (reg_id == MLX4_REG_ID_PTYS) {
+		struct mlx4_ptys_reg *ptys_reg =
+			(struct mlx4_ptys_reg *)inbuf->reg_data;
+
+		ptys_reg->local_port =
+			mlx4_slave_convert_port(dev, slave,
+						ptys_reg->local_port);
+	}
+
+	return mlx4_cmd_box(dev, inbox->dma, outbox->dma, vhcr->in_modifier,
+			    0, MLX4_CMD_ACCESS_REG, MLX4_CMD_TIME_CLASS_C,
+			    MLX4_CMD_NATIVE);
+}
+
+static int mlx4_SET_PORT_phv_bit(struct mlx4_dev *dev, u8 port, u8 phv_bit)
+{
+#define SET_PORT_GEN_PHV_VALID	0x10
+#define SET_PORT_GEN_PHV_EN	0x80
+
+	struct mlx4_cmd_mailbox *mailbox;
+	struct mlx4_set_port_general_context *context;
+	u32 in_mod;
+	int err;
+
+	mailbox = mlx4_alloc_cmd_mailbox(dev);
+	if (IS_ERR(mailbox))
+		return PTR_ERR(mailbox);
+	context = mailbox->buf;
+
+	context->flags2 |=  SET_PORT_GEN_PHV_VALID;
+	if (phv_bit)
+		context->phv_en |=  SET_PORT_GEN_PHV_EN;
+
+	in_mod = MLX4_SET_PORT_GENERAL << 8 | port;
+	err = mlx4_cmd(dev, mailbox->dma, in_mod, MLX4_SET_PORT_ETH_OPCODE,
+		       MLX4_CMD_SET_PORT, MLX4_CMD_TIME_CLASS_B,
+		       MLX4_CMD_NATIVE);
+
+	mlx4_free_cmd_mailbox(dev, mailbox);
+	return err;
+}
+
+int get_phv_bit(struct mlx4_dev *dev, u8 port, int *phv)
+{
+	int err;
+	struct mlx4_func_cap func_cap;
+
+	memset(&func_cap, 0, sizeof(func_cap));
+	err = mlx4_QUERY_FUNC_CAP(dev, port, &func_cap);
+	if (!err)
+		*phv = func_cap.flags0 & QUERY_FUNC_CAP_PHV_BIT;
+	return err;
+}
+EXPORT_SYMBOL(get_phv_bit);
+
+int set_phv_bit(struct mlx4_dev *dev, u8 port, int new_val)
+{
+	int ret;
+
+	if (mlx4_is_slave(dev))
+		return -EPERM;
+
+	if (dev->caps.flags2 & MLX4_DEV_CAP_FLAG2_PHV_EN &&
+	    !(dev->caps.flags2 & MLX4_DEV_CAP_FLAG2_SKIP_OUTER_VLAN)) {
+		ret = mlx4_SET_PORT_phv_bit(dev, port, new_val);
+		if (!ret)
+			dev->caps.phv_bit[port] = new_val;
+		return ret;
+	}
+
+	return -EOPNOTSUPP;
+}
+EXPORT_SYMBOL(set_phv_bit);
+
+int mlx4_get_is_vlan_offload_disabled(struct mlx4_dev *dev, u8 port,
+				      bool *vlan_offload_disabled)
+{
+	struct mlx4_func_cap func_cap;
+	int err;
+
+	memset(&func_cap, 0, sizeof(func_cap));
+	err = mlx4_QUERY_FUNC_CAP(dev, port, &func_cap);
+	if (!err)
+		*vlan_offload_disabled =
+			!!(func_cap.flags0 &
+			   QUERY_FUNC_CAP_VLAN_OFFLOAD_DISABLE);
+	return err;
+}
+EXPORT_SYMBOL(mlx4_get_is_vlan_offload_disabled);
+
+void mlx4_replace_zero_macs(struct mlx4_dev *dev)
+{
+	int i;
+	u8 mac_addr[ETH_ALEN];
+
+	dev->port_random_macs = 0;
+	for (i = 1; i <= dev->caps.num_ports; ++i)
+		if (!dev->caps.def_mac[i] &&
+		    dev->caps.port_type[i] == MLX4_PORT_TYPE_ETH) {
+			eth_random_addr(mac_addr);
+			dev->port_random_macs |= 1 << i;
+			dev->caps.def_mac[i] = mlx4_mac_to_u64(mac_addr);
+		}
+}
+EXPORT_SYMBOL_GPL(mlx4_replace_zero_macs);
diff --git a/drivers/net/ethernet/mellanox/mlx4/fw.h b/drivers/net/ethernet/mellanox/mlx4/fw.h
new file mode 100644
index 000000000..cf64e54ee
--- /dev/null
+++ b/drivers/net/ethernet/mellanox/mlx4/fw.h
@@ -0,0 +1,257 @@
+/*
+ * Copyright (c) 2004, 2005 Topspin Communications.  All rights reserved.
+ * Copyright (c) 2005, 2006, 2007, 2008 Mellanox Technologies. All rights reserved.
+ * Copyright (c) 2006, 2007 Cisco Systems.  All rights reserved.
+ *
+ * This software is available to you under a choice of one of two
+ * licenses.  You may choose to be licensed under the terms of the GNU
+ * General Public License (GPL) Version 2, available from the file
+ * COPYING in the main directory of this source tree, or the
+ * OpenIB.org BSD license below:
+ *
+ *     Redistribution and use in source and binary forms, with or
+ *     without modification, are permitted provided that the following
+ *     conditions are met:
+ *
+ *      - Redistributions of source code must retain the above
+ *        copyright notice, this list of conditions and the following
+ *        disclaimer.
+ *
+ *      - Redistributions in binary form must reproduce the above
+ *        copyright notice, this list of conditions and the following
+ *        disclaimer in the documentation and/or other materials
+ *        provided with the distribution.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
+ * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
+ * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS
+ * BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN
+ * ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
+ * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+ * SOFTWARE.
+ */
+
+#ifndef MLX4_FW_H
+#define MLX4_FW_H
+
+#include "mlx4.h"
+#include "icm.h"
+
+struct mlx4_mod_stat_cfg {
+	u8 log_pg_sz;
+	u8 log_pg_sz_m;
+};
+
+struct mlx4_port_cap {
+	u8  link_state;
+	u8  supported_port_types;
+	u8  suggested_type;
+	u8  default_sense;
+	u8  log_max_macs;
+	u8  log_max_vlans;
+	int ib_mtu;
+	int max_port_width;
+	int max_vl;
+	int max_tc_eth;
+	int max_gids;
+	int max_pkeys;
+	u64 def_mac;
+	u16 eth_mtu;
+	int trans_type;
+	int vendor_oui;
+	u16 wavelength;
+	u64 trans_code;
+	u8 dmfs_optimized_state;
+};
+
+struct mlx4_dev_cap {
+	int max_srq_sz;
+	int max_qp_sz;
+	int reserved_qps;
+	int max_qps;
+	int reserved_srqs;
+	int max_srqs;
+	int max_cq_sz;
+	int reserved_cqs;
+	int max_cqs;
+	int max_mpts;
+	int reserved_eqs;
+	int max_eqs;
+	int num_sys_eqs;
+	int reserved_mtts;
+	int reserved_mrws;
+	int max_requester_per_qp;
+	int max_responder_per_qp;
+	int max_rdma_global;
+	int local_ca_ack_delay;
+	int num_ports;
+	u32 max_msg_sz;
+	u16 stat_rate_support;
+	int fs_log_max_ucast_qp_range_size;
+	int fs_max_num_qp_per_entry;
+	u64 flags;
+	u64 flags2;
+	int reserved_uars;
+	int uar_size;
+	int min_page_sz;
+	int bf_reg_size;
+	int bf_regs_per_page;
+	int max_sq_sg;
+	int max_sq_desc_sz;
+	int max_rq_sg;
+	int max_rq_desc_sz;
+	int max_qp_per_mcg;
+	int reserved_mgms;
+	int max_mcgs;
+	int reserved_pds;
+	int max_pds;
+	int reserved_xrcds;
+	int max_xrcds;
+	int qpc_entry_sz;
+	int rdmarc_entry_sz;
+	int altc_entry_sz;
+	int aux_entry_sz;
+	int srq_entry_sz;
+	int cqc_entry_sz;
+	int eqc_entry_sz;
+	int dmpt_entry_sz;
+	int cmpt_entry_sz;
+	int mtt_entry_sz;
+	int resize_srq;
+	u32 bmme_flags;
+	u32 reserved_lkey;
+	u64 max_icm_sz;
+	int max_gso_sz;
+	int max_rss_tbl_sz;
+	u32 max_counters;
+	u32 dmfs_high_rate_qpn_base;
+	u32 dmfs_high_rate_qpn_range;
+	struct mlx4_rate_limit_caps rl_caps;
+	u32 health_buffer_addrs;
+	struct mlx4_port_cap port_cap[MLX4_MAX_PORTS + 1];
+	bool wol_port[MLX4_MAX_PORTS + 1];
+	bool map_clock_to_user;
+};
+
+struct mlx4_func_cap {
+	u8	num_ports;
+	u8	flags;
+	u32	pf_context_behaviour;
+	int	qp_quota;
+	int	cq_quota;
+	int	srq_quota;
+	int	mpt_quota;
+	int	mtt_quota;
+	int	max_eq;
+	int	reserved_eq;
+	int	mcg_quota;
+	struct mlx4_spec_qps spec_qps;
+	u32	reserved_lkey;
+	u8	physical_port;
+	u8	flags0;
+	u8	flags1;
+	u64	phys_port_id;
+	u32	extra_flags;
+};
+
+struct mlx4_func {
+	int	bus;
+	int	device;
+	int	function;
+	int	physical_function;
+	int	rsvd_eqs;
+	int	max_eq;
+	int	rsvd_uars;
+};
+
+struct mlx4_adapter {
+	char board_id[MLX4_BOARD_ID_LEN];
+	u8   inta_pin;
+};
+
+struct mlx4_init_hca_param {
+	u64 qpc_base;
+	u64 rdmarc_base;
+	u64 auxc_base;
+	u64 altc_base;
+	u64 srqc_base;
+	u64 cqc_base;
+	u64 eqc_base;
+	u64 mc_base;
+	u64 dmpt_base;
+	u64 cmpt_base;
+	u64 mtt_base;
+	u64 global_caps;
+	u8 log_mc_entry_sz;
+	u8 log_mc_hash_sz;
+	u16 hca_core_clock; /* Internal Clock Frequency (in MHz) */
+	u8  log_num_qps;
+	u8  log_num_srqs;
+	u8  log_num_cqs;
+	u8  log_num_eqs;
+	u16 num_sys_eqs;
+	u8  log_rd_per_qp;
+	u8  log_mc_table_sz;
+	u8  log_mpt_sz;
+	u8  log_uar_sz;
+	u8  mw_enabled;  /* Enable memory windows */
+	u8  uar_page_sz; /* log pg sz in 4k chunks */
+	u8  steering_mode; /* for QUERY_HCA */
+	u8  dmfs_high_steer_mode; /* for QUERY_HCA */
+	u64 dev_cap_enabled;
+	u16 cqe_size; /* For use only when CQE stride feature enabled */
+	u16 eqe_size; /* For use only when EQE stride feature enabled */
+	u8 rss_ip_frags;
+	u8 phv_check_en; /* for QUERY_HCA */
+};
+
+struct mlx4_init_ib_param {
+	int port_width;
+	int vl_cap;
+	int mtu_cap;
+	u16 gid_cap;
+	u16 pkey_cap;
+	int set_guid0;
+	u64 guid0;
+	int set_node_guid;
+	u64 node_guid;
+	int set_si_guid;
+	u64 si_guid;
+};
+
+struct mlx4_set_ib_param {
+	int set_si_guid;
+	int reset_qkey_viol;
+	u64 si_guid;
+	u32 cap_mask;
+};
+
+void mlx4_dev_cap_dump(struct mlx4_dev *dev, struct mlx4_dev_cap *dev_cap);
+int mlx4_QUERY_DEV_CAP(struct mlx4_dev *dev, struct mlx4_dev_cap *dev_cap);
+int mlx4_QUERY_PORT(struct mlx4_dev *dev, int port, struct mlx4_port_cap *port_cap);
+int mlx4_QUERY_FUNC_CAP(struct mlx4_dev *dev, u8 gen_or_port,
+			struct mlx4_func_cap *func_cap);
+int mlx4_QUERY_FUNC_CAP_wrapper(struct mlx4_dev *dev, int slave,
+				struct mlx4_vhcr *vhcr,
+				struct mlx4_cmd_mailbox *inbox,
+				struct mlx4_cmd_mailbox *outbox,
+				struct mlx4_cmd_info *cmd);
+int mlx4_QUERY_FUNC(struct mlx4_dev *dev, struct mlx4_func *func, int slave);
+int mlx4_MAP_FA(struct mlx4_dev *dev, struct mlx4_icm *icm);
+int mlx4_UNMAP_FA(struct mlx4_dev *dev);
+int mlx4_RUN_FW(struct mlx4_dev *dev);
+int mlx4_QUERY_FW(struct mlx4_dev *dev);
+int mlx4_QUERY_ADAPTER(struct mlx4_dev *dev, struct mlx4_adapter *adapter);
+int mlx4_INIT_HCA(struct mlx4_dev *dev, struct mlx4_init_hca_param *param);
+int mlx4_QUERY_HCA(struct mlx4_dev *dev, struct mlx4_init_hca_param *param);
+int mlx4_CLOSE_HCA(struct mlx4_dev *dev, int panic);
+int mlx4_map_cmd(struct mlx4_dev *dev, u16 op, struct mlx4_icm *icm, u64 virt);
+int mlx4_SET_ICM_SIZE(struct mlx4_dev *dev, u64 icm_size, u64 *aux_pages);
+int mlx4_MAP_ICM_AUX(struct mlx4_dev *dev, struct mlx4_icm *icm);
+int mlx4_UNMAP_ICM_AUX(struct mlx4_dev *dev);
+int mlx4_NOP(struct mlx4_dev *dev);
+int mlx4_MOD_STAT_CFG(struct mlx4_dev *dev, struct mlx4_mod_stat_cfg *cfg);
+void mlx4_opreq_action(struct work_struct *work);
+
+#endif /* MLX4_FW_H */
diff --git a/drivers/net/ethernet/mellanox/mlx4/fw_qos.c b/drivers/net/ethernet/mellanox/mlx4/fw_qos.c
new file mode 100644
index 000000000..3a09d7122
--- /dev/null
+++ b/drivers/net/ethernet/mellanox/mlx4/fw_qos.c
@@ -0,0 +1,289 @@
+/*
+ * Copyright (c) 2004, 2005 Topspin Communications.  All rights reserved.
+ * Copyright (c) 2005, 2006, 2007, 2008 Mellanox Technologies.
+ * All rights reserved.
+ *
+ * This software is available to you under a choice of one of two
+ * licenses.  You may choose to be licensed under the terms of the GNU
+ * General Public License (GPL) Version 2, available from the file
+ * COPYING in the main directory of this source tree, or the
+ * OpenIB.org BSD license below:
+ *
+ *     Redistribution and use in source and binary forms, with or
+ *     without modification, are permitted provided that the following
+ *     conditions are met:
+ *
+ *      - Redistributions of source code must retain the above
+ *        copyright notice, this list of conditions and the following
+ *        disclaimer.
+ *
+ *      - Redistributions in binary form must reproduce the above
+ *        copyright notice, this list of conditions and the following
+ *        disclaimer in the documentation and/or other materials
+ *        provided with the distribution.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
+ * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
+ * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS
+ * BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN
+ * ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
+ * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+ * SOFTWARE.
+ */
+
+#include <linux/export.h>
+#include "fw_qos.h"
+#include "fw.h"
+
+enum {
+	/* allocate vpp opcode modifiers */
+	MLX4_ALLOCATE_VPP_ALLOCATE	= 0x0,
+	MLX4_ALLOCATE_VPP_QUERY		= 0x1
+};
+
+enum {
+	/* set vport qos opcode modifiers */
+	MLX4_SET_VPORT_QOS_SET		= 0x0,
+	MLX4_SET_VPORT_QOS_QUERY	= 0x1
+};
+
+struct mlx4_set_port_prio2tc_context {
+	u8 prio2tc[4];
+};
+
+struct mlx4_port_scheduler_tc_cfg_be {
+	__be16 pg;
+	__be16 bw_precentage;
+	__be16 max_bw_units; /* 3-100Mbps, 4-1Gbps, other values - reserved */
+	__be16 max_bw_value;
+};
+
+struct mlx4_set_port_scheduler_context {
+	struct mlx4_port_scheduler_tc_cfg_be tc[MLX4_NUM_TC];
+};
+
+/* Granular Qos (per VF) section */
+struct mlx4_alloc_vpp_param {
+	__be32 available_vpp;
+	__be32 vpp_p_up[MLX4_NUM_UP];
+};
+
+struct mlx4_prio_qos_param {
+	__be32 bw_share;
+	__be32 max_avg_bw;
+	__be32 reserved;
+	__be32 enable;
+	__be32 reserved1[4];
+};
+
+struct mlx4_set_vport_context {
+	__be32 reserved[8];
+	struct mlx4_prio_qos_param qos_p_up[MLX4_NUM_UP];
+};
+
+int mlx4_SET_PORT_PRIO2TC(struct mlx4_dev *dev, u8 port, u8 *prio2tc)
+{
+	struct mlx4_cmd_mailbox *mailbox;
+	struct mlx4_set_port_prio2tc_context *context;
+	int err;
+	u32 in_mod;
+	int i;
+
+	mailbox = mlx4_alloc_cmd_mailbox(dev);
+	if (IS_ERR(mailbox))
+		return PTR_ERR(mailbox);
+
+	context = mailbox->buf;
+
+	for (i = 0; i < MLX4_NUM_UP; i += 2)
+		context->prio2tc[i >> 1] = prio2tc[i] << 4 | prio2tc[i + 1];
+
+	in_mod = MLX4_SET_PORT_PRIO2TC << 8 | port;
+	err = mlx4_cmd(dev, mailbox->dma, in_mod, 1, MLX4_CMD_SET_PORT,
+		       MLX4_CMD_TIME_CLASS_B, MLX4_CMD_NATIVE);
+
+	mlx4_free_cmd_mailbox(dev, mailbox);
+	return err;
+}
+EXPORT_SYMBOL(mlx4_SET_PORT_PRIO2TC);
+
+int mlx4_SET_PORT_SCHEDULER(struct mlx4_dev *dev, u8 port, u8 *tc_tx_bw,
+			    u8 *pg, u16 *ratelimit)
+{
+	struct mlx4_cmd_mailbox *mailbox;
+	struct mlx4_set_port_scheduler_context *context;
+	int err;
+	u32 in_mod;
+	int i;
+
+	mailbox = mlx4_alloc_cmd_mailbox(dev);
+	if (IS_ERR(mailbox))
+		return PTR_ERR(mailbox);
+
+	context = mailbox->buf;
+
+	for (i = 0; i < MLX4_NUM_TC; i++) {
+		struct mlx4_port_scheduler_tc_cfg_be *tc = &context->tc[i];
+		u16 r;
+
+		if (ratelimit && ratelimit[i]) {
+			if (ratelimit[i] <= MLX4_MAX_100M_UNITS_VAL) {
+				r = ratelimit[i];
+				tc->max_bw_units =
+					htons(MLX4_RATELIMIT_100M_UNITS);
+			} else {
+				r = ratelimit[i] / 10;
+				tc->max_bw_units =
+					htons(MLX4_RATELIMIT_1G_UNITS);
+			}
+			tc->max_bw_value = htons(r);
+		} else {
+			tc->max_bw_value = htons(MLX4_RATELIMIT_DEFAULT);
+			tc->max_bw_units = htons(MLX4_RATELIMIT_1G_UNITS);
+		}
+
+		tc->pg = htons(pg[i]);
+		tc->bw_precentage = htons(tc_tx_bw[i]);
+	}
+
+	in_mod = MLX4_SET_PORT_SCHEDULER << 8 | port;
+	err = mlx4_cmd(dev, mailbox->dma, in_mod, 1, MLX4_CMD_SET_PORT,
+		       MLX4_CMD_TIME_CLASS_B, MLX4_CMD_NATIVE);
+
+	mlx4_free_cmd_mailbox(dev, mailbox);
+	return err;
+}
+EXPORT_SYMBOL(mlx4_SET_PORT_SCHEDULER);
+
+int mlx4_ALLOCATE_VPP_get(struct mlx4_dev *dev, u8 port,
+			  u16 *available_vpp, u8 *vpp_p_up)
+{
+	int i;
+	int err;
+	struct mlx4_cmd_mailbox *mailbox;
+	struct mlx4_alloc_vpp_param *out_param;
+
+	mailbox = mlx4_alloc_cmd_mailbox(dev);
+	if (IS_ERR(mailbox))
+		return PTR_ERR(mailbox);
+
+	out_param = mailbox->buf;
+
+	err = mlx4_cmd_box(dev, 0, mailbox->dma, port,
+			   MLX4_ALLOCATE_VPP_QUERY,
+			   MLX4_CMD_ALLOCATE_VPP,
+			   MLX4_CMD_TIME_CLASS_A,
+			   MLX4_CMD_NATIVE);
+	if (err)
+		goto out;
+
+	/* Total number of supported VPPs */
+	*available_vpp = (u16)be32_to_cpu(out_param->available_vpp);
+
+	for (i = 0; i < MLX4_NUM_UP; i++)
+		vpp_p_up[i] = (u8)be32_to_cpu(out_param->vpp_p_up[i]);
+
+out:
+	mlx4_free_cmd_mailbox(dev, mailbox);
+
+	return err;
+}
+EXPORT_SYMBOL(mlx4_ALLOCATE_VPP_get);
+
+int mlx4_ALLOCATE_VPP_set(struct mlx4_dev *dev, u8 port, u8 *vpp_p_up)
+{
+	int i;
+	int err;
+	struct mlx4_cmd_mailbox *mailbox;
+	struct mlx4_alloc_vpp_param *in_param;
+
+	mailbox = mlx4_alloc_cmd_mailbox(dev);
+	if (IS_ERR(mailbox))
+		return PTR_ERR(mailbox);
+
+	in_param = mailbox->buf;
+
+	for (i = 0; i < MLX4_NUM_UP; i++)
+		in_param->vpp_p_up[i] = cpu_to_be32(vpp_p_up[i]);
+
+	err = mlx4_cmd(dev, mailbox->dma, port,
+		       MLX4_ALLOCATE_VPP_ALLOCATE,
+		       MLX4_CMD_ALLOCATE_VPP,
+		       MLX4_CMD_TIME_CLASS_A,
+		       MLX4_CMD_NATIVE);
+
+	mlx4_free_cmd_mailbox(dev, mailbox);
+	return err;
+}
+EXPORT_SYMBOL(mlx4_ALLOCATE_VPP_set);
+
+int mlx4_SET_VPORT_QOS_get(struct mlx4_dev *dev, u8 port, u8 vport,
+			   struct mlx4_vport_qos_param *out_param)
+{
+	int i;
+	int err;
+	struct mlx4_cmd_mailbox *mailbox;
+	struct mlx4_set_vport_context *ctx;
+
+	mailbox = mlx4_alloc_cmd_mailbox(dev);
+	if (IS_ERR(mailbox))
+		return PTR_ERR(mailbox);
+
+	ctx = mailbox->buf;
+
+	err = mlx4_cmd_box(dev, 0, mailbox->dma, (vport << 8) | port,
+			   MLX4_SET_VPORT_QOS_QUERY,
+			   MLX4_CMD_SET_VPORT_QOS,
+			   MLX4_CMD_TIME_CLASS_A,
+			   MLX4_CMD_NATIVE);
+	if (err)
+		goto out;
+
+	for (i = 0; i < MLX4_NUM_UP; i++) {
+		out_param[i].bw_share = be32_to_cpu(ctx->qos_p_up[i].bw_share);
+		out_param[i].max_avg_bw =
+			be32_to_cpu(ctx->qos_p_up[i].max_avg_bw);
+		out_param[i].enable =
+			!!(be32_to_cpu(ctx->qos_p_up[i].enable) & 31);
+	}
+
+out:
+	mlx4_free_cmd_mailbox(dev, mailbox);
+
+	return err;
+}
+EXPORT_SYMBOL(mlx4_SET_VPORT_QOS_get);
+
+int mlx4_SET_VPORT_QOS_set(struct mlx4_dev *dev, u8 port, u8 vport,
+			   struct mlx4_vport_qos_param *in_param)
+{
+	int i;
+	int err;
+	struct mlx4_cmd_mailbox *mailbox;
+	struct mlx4_set_vport_context *ctx;
+
+	mailbox = mlx4_alloc_cmd_mailbox(dev);
+	if (IS_ERR(mailbox))
+		return PTR_ERR(mailbox);
+
+	ctx = mailbox->buf;
+
+	for (i = 0; i < MLX4_NUM_UP; i++) {
+		ctx->qos_p_up[i].bw_share = cpu_to_be32(in_param[i].bw_share);
+		ctx->qos_p_up[i].max_avg_bw =
+				cpu_to_be32(in_param[i].max_avg_bw);
+		ctx->qos_p_up[i].enable =
+				cpu_to_be32(in_param[i].enable << 31);
+	}
+
+	err = mlx4_cmd(dev, mailbox->dma, (vport << 8) | port,
+		       MLX4_SET_VPORT_QOS_SET,
+		       MLX4_CMD_SET_VPORT_QOS,
+		       MLX4_CMD_TIME_CLASS_A,
+		       MLX4_CMD_NATIVE);
+
+	mlx4_free_cmd_mailbox(dev, mailbox);
+	return err;
+}
+EXPORT_SYMBOL(mlx4_SET_VPORT_QOS_set);
diff --git a/drivers/net/ethernet/mellanox/mlx4/fw_qos.h b/drivers/net/ethernet/mellanox/mlx4/fw_qos.h
new file mode 100644
index 000000000..582997577
--- /dev/null
+++ b/drivers/net/ethernet/mellanox/mlx4/fw_qos.h
@@ -0,0 +1,145 @@
+/*
+ * Copyright (c) 2004, 2005 Topspin Communications.  All rights reserved.
+ * Copyright (c) 2005, 2006, 2007, 2008 Mellanox Technologies.
+ * All rights reserved.
+ *
+ * This software is available to you under a choice of one of two
+ * licenses.  You may choose to be licensed under the terms of the GNU
+ * General Public License (GPL) Version 2, available from the file
+ * COPYING in the main directory of this source tree, or the
+ * OpenIB.org BSD license below:
+ *
+ *     Redistribution and use in source and binary forms, with or
+ *     without modification, are permitted provided that the following
+ *     conditions are met:
+ *
+ *      - Redistributions of source code must retain the above
+ *        copyright notice, this list of conditions and the following
+ *        disclaimer.
+ *
+ *      - Redistributions in binary form must reproduce the above
+ *        copyright notice, this list of conditions and the following
+ *        disclaimer in the documentation and/or other materials
+ *        provided with the distribution.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
+ * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
+ * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS
+ * BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN
+ * ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
+ * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+ * SOFTWARE.
+ */
+
+#ifndef MLX4_FW_QOS_H
+#define MLX4_FW_QOS_H
+
+#include <linux/mlx4/cmd.h>
+#include <linux/mlx4/device.h>
+
+#define MLX4_NUM_UP 8
+#define MLX4_NUM_TC 8
+
+/* Default supported priorities for VPP allocation */
+#define MLX4_DEFAULT_QOS_PRIO (0)
+
+/* Derived from FW feature definition, 0 is the default vport fo all QPs */
+#define MLX4_VPP_DEFAULT_VPORT (0)
+
+struct mlx4_vport_qos_param {
+	u32 bw_share;
+	u32 max_avg_bw;
+	u8 enable;
+};
+
+/**
+ * mlx4_SET_PORT_PRIO2TC - This routine maps user priorities to traffic
+ * classes of a given port and device.
+ *
+ * @dev: mlx4_dev.
+ * @port: Physical port number.
+ * @prio2tc: Array of TC associated with each priorities.
+ *
+ * Returns 0 on success or a negative mlx4_core errno code.
+ **/
+int mlx4_SET_PORT_PRIO2TC(struct mlx4_dev *dev, u8 port, u8 *prio2tc);
+
+/**
+ * mlx4_SET_PORT_SCHEDULER - This routine configures the arbitration between
+ * traffic classes (ETS) and configured rate limit for traffic classes.
+ * tc_tx_bw, pg and ratelimit are arrays where each index represents a TC.
+ * The description for those parameters below refers to a single TC.
+ *
+ * @dev: mlx4_dev.
+ * @port: Physical port number.
+ * @tc_tx_bw: The percentage of the bandwidth allocated for traffic class
+ *  within a TC group. The sum of the bw_percentage of all the traffic
+ *  classes within a TC group must equal 100% for correct operation.
+ * @pg: The TC group the traffic class is associated with.
+ * @ratelimit: The maximal bandwidth allowed for the use by this traffic class.
+ *
+ * Returns 0 on success or a negative mlx4_core errno code.
+ **/
+int mlx4_SET_PORT_SCHEDULER(struct mlx4_dev *dev, u8 port, u8 *tc_tx_bw,
+			    u8 *pg, u16 *ratelimit);
+/**
+ * mlx4_ALLOCATE_VPP_get - Query port VPP available resources and allocation.
+ * Before distribution of VPPs to priorities, only available_vpp is returned.
+ * After initialization it returns the distribution of VPPs among priorities.
+ *
+ * @dev: mlx4_dev.
+ * @port: Physical port number.
+ * @available_vpp: Pointer to variable where number of available VPPs is stored
+ * @vpp_p_up: Distribution of VPPs to priorities is stored in this array
+ *
+ * Returns 0 on success or a negative mlx4_core errno code.
+ **/
+int mlx4_ALLOCATE_VPP_get(struct mlx4_dev *dev, u8 port,
+			  u16 *available_vpp, u8 *vpp_p_up);
+/**
+ * mlx4_ALLOCATE_VPP_set - Distribution of VPPs among differnt priorities.
+ * The total number of VPPs assigned to all for a port must not exceed
+ * the value reported by available_vpp in mlx4_ALLOCATE_VPP_get.
+ * VPP allocation is allowed only after the port type has been set,
+ * and while no QPs are open for this port.
+ *
+ * @dev: mlx4_dev.
+ * @port: Physical port number.
+ * @vpp_p_up: Allocation of VPPs to different priorities.
+ *
+ * Returns 0 on success or a negative mlx4_core errno code.
+ **/
+int mlx4_ALLOCATE_VPP_set(struct mlx4_dev *dev, u8 port, u8 *vpp_p_up);
+
+/**
+ * mlx4_SET_VPORT_QOS_get - Query QoS proporties of a Vport.
+ * Each priority allowed for the Vport is assigned with a share of the BW,
+ * and a BW limitation. This commands query the current QoS values.
+ *
+ * @dev: mlx4_dev.
+ * @port: Physical port number.
+ * @vport: Vport id.
+ * @out_param: Array of mlx4_vport_qos_param that will contain the values.
+ *
+ * Returns 0 on success or a negative mlx4_core errno code.
+ **/
+int mlx4_SET_VPORT_QOS_get(struct mlx4_dev *dev, u8 port, u8 vport,
+			   struct mlx4_vport_qos_param *out_param);
+
+/**
+ * mlx4_SET_VPORT_QOS_set - Set QoS proporties of a Vport.
+ * QoS parameters can be modified at any time, but must be initialized
+ * before any QP is associated with the VPort.
+ *
+ * @dev: mlx4_dev.
+ * @port: Physical port number.
+ * @vport: Vport id.
+ * @out_param: Array of mlx4_vport_qos_param which holds the requested values.
+ *
+ * Returns 0 on success or a negative mlx4_core errno code.
+ **/
+int mlx4_SET_VPORT_QOS_set(struct mlx4_dev *dev, u8 port, u8 vport,
+			   struct mlx4_vport_qos_param *in_param);
+
+#endif /* MLX4_FW_QOS_H */
diff --git a/drivers/net/ethernet/mellanox/mlx4/icm.c b/drivers/net/ethernet/mellanox/mlx4/icm.c
new file mode 100644
index 000000000..288fca826
--- /dev/null
+++ b/drivers/net/ethernet/mellanox/mlx4/icm.c
@@ -0,0 +1,494 @@
+/*
+ * Copyright (c) 2005, 2006, 2007, 2008 Mellanox Technologies. All rights reserved.
+ * Copyright (c) 2006, 2007 Cisco Systems, Inc.  All rights reserved.
+ *
+ * This software is available to you under a choice of one of two
+ * licenses.  You may choose to be licensed under the terms of the GNU
+ * General Public License (GPL) Version 2, available from the file
+ * COPYING in the main directory of this source tree, or the
+ * OpenIB.org BSD license below:
+ *
+ *     Redistribution and use in source and binary forms, with or
+ *     without modification, are permitted provided that the following
+ *     conditions are met:
+ *
+ *      - Redistributions of source code must retain the above
+ *        copyright notice, this list of conditions and the following
+ *        disclaimer.
+ *
+ *      - Redistributions in binary form must reproduce the above
+ *        copyright notice, this list of conditions and the following
+ *        disclaimer in the documentation and/or other materials
+ *        provided with the distribution.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
+ * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
+ * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS
+ * BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN
+ * ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
+ * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+ * SOFTWARE.
+ */
+
+#include <linux/errno.h>
+#include <linux/mm.h>
+#include <linux/scatterlist.h>
+#include <linux/slab.h>
+
+#include <linux/mlx4/cmd.h>
+
+#include "mlx4.h"
+#include "icm.h"
+#include "fw.h"
+
+/*
+ * We allocate in as big chunks as we can, up to a maximum of 256 KB
+ * per chunk. Note that the chunks are not necessarily in contiguous
+ * physical memory.
+ */
+enum {
+	MLX4_ICM_ALLOC_SIZE	= 1 << 18,
+	MLX4_TABLE_CHUNK_SIZE	= 1 << 18,
+};
+
+static void mlx4_free_icm_pages(struct mlx4_dev *dev, struct mlx4_icm_chunk *chunk)
+{
+	int i;
+
+	if (chunk->nsg > 0)
+		pci_unmap_sg(dev->persist->pdev, chunk->sg, chunk->npages,
+			     PCI_DMA_BIDIRECTIONAL);
+
+	for (i = 0; i < chunk->npages; ++i)
+		__free_pages(sg_page(&chunk->sg[i]),
+			     get_order(chunk->sg[i].length));
+}
+
+static void mlx4_free_icm_coherent(struct mlx4_dev *dev, struct mlx4_icm_chunk *chunk)
+{
+	int i;
+
+	for (i = 0; i < chunk->npages; ++i)
+		dma_free_coherent(&dev->persist->pdev->dev,
+				  chunk->buf[i].size,
+				  chunk->buf[i].addr,
+				  chunk->buf[i].dma_addr);
+}
+
+void mlx4_free_icm(struct mlx4_dev *dev, struct mlx4_icm *icm, int coherent)
+{
+	struct mlx4_icm_chunk *chunk, *tmp;
+
+	if (!icm)
+		return;
+
+	list_for_each_entry_safe(chunk, tmp, &icm->chunk_list, list) {
+		if (coherent)
+			mlx4_free_icm_coherent(dev, chunk);
+		else
+			mlx4_free_icm_pages(dev, chunk);
+
+		kfree(chunk);
+	}
+
+	kfree(icm);
+}
+
+static int mlx4_alloc_icm_pages(struct scatterlist *mem, int order,
+				gfp_t gfp_mask, int node)
+{
+	struct page *page;
+
+	page = alloc_pages_node(node, gfp_mask, order);
+	if (!page) {
+		page = alloc_pages(gfp_mask, order);
+		if (!page)
+			return -ENOMEM;
+	}
+
+	sg_set_page(mem, page, PAGE_SIZE << order, 0);
+	return 0;
+}
+
+static int mlx4_alloc_icm_coherent(struct device *dev, struct mlx4_icm_buf *buf,
+				   int order, gfp_t gfp_mask)
+{
+	buf->addr = dma_alloc_coherent(dev, PAGE_SIZE << order,
+				       &buf->dma_addr, gfp_mask);
+	if (!buf->addr)
+		return -ENOMEM;
+
+	if (offset_in_page(buf->addr)) {
+		dma_free_coherent(dev, PAGE_SIZE << order, buf->addr,
+				  buf->dma_addr);
+		return -ENOMEM;
+	}
+
+	buf->size = PAGE_SIZE << order;
+	return 0;
+}
+
+struct mlx4_icm *mlx4_alloc_icm(struct mlx4_dev *dev, int npages,
+				gfp_t gfp_mask, int coherent)
+{
+	struct mlx4_icm *icm;
+	struct mlx4_icm_chunk *chunk = NULL;
+	int cur_order;
+	gfp_t mask;
+	int ret;
+
+	/* We use sg_set_buf for coherent allocs, which assumes low memory */
+	BUG_ON(coherent && (gfp_mask & __GFP_HIGHMEM));
+
+	icm = kmalloc_node(sizeof(*icm),
+			   gfp_mask & ~(__GFP_HIGHMEM | __GFP_NOWARN),
+			   dev->numa_node);
+	if (!icm) {
+		icm = kmalloc(sizeof(*icm),
+			      gfp_mask & ~(__GFP_HIGHMEM | __GFP_NOWARN));
+		if (!icm)
+			return NULL;
+	}
+
+	icm->refcount = 0;
+	INIT_LIST_HEAD(&icm->chunk_list);
+
+	cur_order = get_order(MLX4_ICM_ALLOC_SIZE);
+
+	while (npages > 0) {
+		if (!chunk) {
+			chunk = kzalloc_node(sizeof(*chunk),
+					     gfp_mask & ~(__GFP_HIGHMEM |
+							  __GFP_NOWARN),
+					     dev->numa_node);
+			if (!chunk) {
+				chunk = kzalloc(sizeof(*chunk),
+						gfp_mask & ~(__GFP_HIGHMEM |
+							     __GFP_NOWARN));
+				if (!chunk)
+					goto fail;
+			}
+			chunk->coherent = coherent;
+
+			if (!coherent)
+				sg_init_table(chunk->sg, MLX4_ICM_CHUNK_LEN);
+			list_add_tail(&chunk->list, &icm->chunk_list);
+		}
+
+		while (1 << cur_order > npages)
+			--cur_order;
+
+		mask = gfp_mask;
+		if (cur_order)
+			mask &= ~__GFP_DIRECT_RECLAIM;
+
+		if (coherent)
+			ret = mlx4_alloc_icm_coherent(&dev->persist->pdev->dev,
+						&chunk->buf[chunk->npages],
+						cur_order, mask);
+		else
+			ret = mlx4_alloc_icm_pages(&chunk->sg[chunk->npages],
+						   cur_order, mask,
+						   dev->numa_node);
+
+		if (ret) {
+			if (--cur_order < 0)
+				goto fail;
+			else
+				continue;
+		}
+
+		++chunk->npages;
+
+		if (coherent)
+			++chunk->nsg;
+		else if (chunk->npages == MLX4_ICM_CHUNK_LEN) {
+			chunk->nsg = pci_map_sg(dev->persist->pdev, chunk->sg,
+						chunk->npages,
+						PCI_DMA_BIDIRECTIONAL);
+
+			if (chunk->nsg <= 0)
+				goto fail;
+		}
+
+		if (chunk->npages == MLX4_ICM_CHUNK_LEN)
+			chunk = NULL;
+
+		npages -= 1 << cur_order;
+	}
+
+	if (!coherent && chunk) {
+		chunk->nsg = pci_map_sg(dev->persist->pdev, chunk->sg,
+					chunk->npages,
+					PCI_DMA_BIDIRECTIONAL);
+
+		if (chunk->nsg <= 0)
+			goto fail;
+	}
+
+	return icm;
+
+fail:
+	mlx4_free_icm(dev, icm, coherent);
+	return NULL;
+}
+
+static int mlx4_MAP_ICM(struct mlx4_dev *dev, struct mlx4_icm *icm, u64 virt)
+{
+	return mlx4_map_cmd(dev, MLX4_CMD_MAP_ICM, icm, virt);
+}
+
+static int mlx4_UNMAP_ICM(struct mlx4_dev *dev, u64 virt, u32 page_count)
+{
+	return mlx4_cmd(dev, virt, page_count, 0, MLX4_CMD_UNMAP_ICM,
+			MLX4_CMD_TIME_CLASS_B, MLX4_CMD_NATIVE);
+}
+
+int mlx4_MAP_ICM_AUX(struct mlx4_dev *dev, struct mlx4_icm *icm)
+{
+	return mlx4_map_cmd(dev, MLX4_CMD_MAP_ICM_AUX, icm, -1);
+}
+
+int mlx4_UNMAP_ICM_AUX(struct mlx4_dev *dev)
+{
+	return mlx4_cmd(dev, 0, 0, 0, MLX4_CMD_UNMAP_ICM_AUX,
+			MLX4_CMD_TIME_CLASS_B, MLX4_CMD_NATIVE);
+}
+
+int mlx4_table_get(struct mlx4_dev *dev, struct mlx4_icm_table *table, u32 obj)
+{
+	u32 i = (obj & (table->num_obj - 1)) /
+			(MLX4_TABLE_CHUNK_SIZE / table->obj_size);
+	int ret = 0;
+
+	mutex_lock(&table->mutex);
+
+	if (table->icm[i]) {
+		++table->icm[i]->refcount;
+		goto out;
+	}
+
+	table->icm[i] = mlx4_alloc_icm(dev, MLX4_TABLE_CHUNK_SIZE >> PAGE_SHIFT,
+				       (table->lowmem ? GFP_KERNEL : GFP_HIGHUSER) |
+				       __GFP_NOWARN, table->coherent);
+	if (!table->icm[i]) {
+		ret = -ENOMEM;
+		goto out;
+	}
+
+	if (mlx4_MAP_ICM(dev, table->icm[i], table->virt +
+			 (u64) i * MLX4_TABLE_CHUNK_SIZE)) {
+		mlx4_free_icm(dev, table->icm[i], table->coherent);
+		table->icm[i] = NULL;
+		ret = -ENOMEM;
+		goto out;
+	}
+
+	++table->icm[i]->refcount;
+
+out:
+	mutex_unlock(&table->mutex);
+	return ret;
+}
+
+void mlx4_table_put(struct mlx4_dev *dev, struct mlx4_icm_table *table, u32 obj)
+{
+	u32 i;
+	u64 offset;
+
+	i = (obj & (table->num_obj - 1)) / (MLX4_TABLE_CHUNK_SIZE / table->obj_size);
+
+	mutex_lock(&table->mutex);
+
+	if (--table->icm[i]->refcount == 0) {
+		offset = (u64) i * MLX4_TABLE_CHUNK_SIZE;
+		mlx4_UNMAP_ICM(dev, table->virt + offset,
+			       MLX4_TABLE_CHUNK_SIZE / MLX4_ICM_PAGE_SIZE);
+		mlx4_free_icm(dev, table->icm[i], table->coherent);
+		table->icm[i] = NULL;
+	}
+
+	mutex_unlock(&table->mutex);
+}
+
+void *mlx4_table_find(struct mlx4_icm_table *table, u32 obj,
+			dma_addr_t *dma_handle)
+{
+	int offset, dma_offset, i;
+	u64 idx;
+	struct mlx4_icm_chunk *chunk;
+	struct mlx4_icm *icm;
+	void *addr = NULL;
+
+	if (!table->lowmem)
+		return NULL;
+
+	mutex_lock(&table->mutex);
+
+	idx = (u64) (obj & (table->num_obj - 1)) * table->obj_size;
+	icm = table->icm[idx / MLX4_TABLE_CHUNK_SIZE];
+	dma_offset = offset = idx % MLX4_TABLE_CHUNK_SIZE;
+
+	if (!icm)
+		goto out;
+
+	list_for_each_entry(chunk, &icm->chunk_list, list) {
+		for (i = 0; i < chunk->npages; ++i) {
+			dma_addr_t dma_addr;
+			size_t len;
+
+			if (table->coherent) {
+				len = chunk->buf[i].size;
+				dma_addr = chunk->buf[i].dma_addr;
+				addr = chunk->buf[i].addr;
+			} else {
+				struct page *page;
+
+				len = sg_dma_len(&chunk->sg[i]);
+				dma_addr = sg_dma_address(&chunk->sg[i]);
+
+				/* XXX: we should never do this for highmem
+				 * allocation.  This function either needs
+				 * to be split, or the kernel virtual address
+				 * return needs to be made optional.
+				 */
+				page = sg_page(&chunk->sg[i]);
+				addr = lowmem_page_address(page);
+			}
+
+			if (dma_handle && dma_offset >= 0) {
+				if (len > dma_offset)
+					*dma_handle = dma_addr + dma_offset;
+				dma_offset -= len;
+			}
+
+			/*
+			 * DMA mapping can merge pages but not split them,
+			 * so if we found the page, dma_handle has already
+			 * been assigned to.
+			 */
+			if (len > offset)
+				goto out;
+			offset -= len;
+		}
+	}
+
+	addr = NULL;
+out:
+	mutex_unlock(&table->mutex);
+	return addr ? addr + offset : NULL;
+}
+
+int mlx4_table_get_range(struct mlx4_dev *dev, struct mlx4_icm_table *table,
+			 u32 start, u32 end)
+{
+	int inc = MLX4_TABLE_CHUNK_SIZE / table->obj_size;
+	int err;
+	u32 i;
+
+	for (i = start; i <= end; i += inc) {
+		err = mlx4_table_get(dev, table, i);
+		if (err)
+			goto fail;
+	}
+
+	return 0;
+
+fail:
+	while (i > start) {
+		i -= inc;
+		mlx4_table_put(dev, table, i);
+	}
+
+	return err;
+}
+
+void mlx4_table_put_range(struct mlx4_dev *dev, struct mlx4_icm_table *table,
+			  u32 start, u32 end)
+{
+	u32 i;
+
+	for (i = start; i <= end; i += MLX4_TABLE_CHUNK_SIZE / table->obj_size)
+		mlx4_table_put(dev, table, i);
+}
+
+int mlx4_init_icm_table(struct mlx4_dev *dev, struct mlx4_icm_table *table,
+			u64 virt, int obj_size,	u32 nobj, int reserved,
+			int use_lowmem, int use_coherent)
+{
+	int obj_per_chunk;
+	int num_icm;
+	unsigned chunk_size;
+	int i;
+	u64 size;
+
+	obj_per_chunk = MLX4_TABLE_CHUNK_SIZE / obj_size;
+	if (WARN_ON(!obj_per_chunk))
+		return -EINVAL;
+	num_icm = (nobj + obj_per_chunk - 1) / obj_per_chunk;
+
+	table->icm      = kvcalloc(num_icm, sizeof(*table->icm), GFP_KERNEL);
+	if (!table->icm)
+		return -ENOMEM;
+	table->virt     = virt;
+	table->num_icm  = num_icm;
+	table->num_obj  = nobj;
+	table->obj_size = obj_size;
+	table->lowmem   = use_lowmem;
+	table->coherent = use_coherent;
+	mutex_init(&table->mutex);
+
+	size = (u64) nobj * obj_size;
+	for (i = 0; i * MLX4_TABLE_CHUNK_SIZE < reserved * obj_size; ++i) {
+		chunk_size = MLX4_TABLE_CHUNK_SIZE;
+		if ((i + 1) * MLX4_TABLE_CHUNK_SIZE > size)
+			chunk_size = PAGE_ALIGN(size -
+					i * MLX4_TABLE_CHUNK_SIZE);
+
+		table->icm[i] = mlx4_alloc_icm(dev, chunk_size >> PAGE_SHIFT,
+					       (use_lowmem ? GFP_KERNEL : GFP_HIGHUSER) |
+					       __GFP_NOWARN, use_coherent);
+		if (!table->icm[i])
+			goto err;
+		if (mlx4_MAP_ICM(dev, table->icm[i], virt + i * MLX4_TABLE_CHUNK_SIZE)) {
+			mlx4_free_icm(dev, table->icm[i], use_coherent);
+			table->icm[i] = NULL;
+			goto err;
+		}
+
+		/*
+		 * Add a reference to this ICM chunk so that it never
+		 * gets freed (since it contains reserved firmware objects).
+		 */
+		++table->icm[i]->refcount;
+	}
+
+	return 0;
+
+err:
+	for (i = 0; i < num_icm; ++i)
+		if (table->icm[i]) {
+			mlx4_UNMAP_ICM(dev, virt + i * MLX4_TABLE_CHUNK_SIZE,
+				       MLX4_TABLE_CHUNK_SIZE / MLX4_ICM_PAGE_SIZE);
+			mlx4_free_icm(dev, table->icm[i], use_coherent);
+		}
+
+	kvfree(table->icm);
+
+	return -ENOMEM;
+}
+
+void mlx4_cleanup_icm_table(struct mlx4_dev *dev, struct mlx4_icm_table *table)
+{
+	int i;
+
+	for (i = 0; i < table->num_icm; ++i)
+		if (table->icm[i]) {
+			mlx4_UNMAP_ICM(dev, table->virt + i * MLX4_TABLE_CHUNK_SIZE,
+				       MLX4_TABLE_CHUNK_SIZE / MLX4_ICM_PAGE_SIZE);
+			mlx4_free_icm(dev, table->icm[i], table->coherent);
+		}
+
+	kvfree(table->icm);
+}
diff --git a/drivers/net/ethernet/mellanox/mlx4/icm.h b/drivers/net/ethernet/mellanox/mlx4/icm.h
new file mode 100644
index 000000000..d199874b1
--- /dev/null
+++ b/drivers/net/ethernet/mellanox/mlx4/icm.h
@@ -0,0 +1,144 @@
+/*
+ * Copyright (c) 2005, 2006, 2007, 2008 Mellanox Technologies. All rights reserved.
+ * Copyright (c) 2006, 2007 Cisco Systems, Inc.  All rights reserved.
+ *
+ * This software is available to you under a choice of one of two
+ * licenses.  You may choose to be licensed under the terms of the GNU
+ * General Public License (GPL) Version 2, available from the file
+ * COPYING in the main directory of this source tree, or the
+ * OpenIB.org BSD license below:
+ *
+ *     Redistribution and use in source and binary forms, with or
+ *     without modification, are permitted provided that the following
+ *     conditions are met:
+ *
+ *      - Redistributions of source code must retain the above
+ *        copyright notice, this list of conditions and the following
+ *        disclaimer.
+ *
+ *      - Redistributions in binary form must reproduce the above
+ *        copyright notice, this list of conditions and the following
+ *        disclaimer in the documentation and/or other materials
+ *        provided with the distribution.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
+ * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
+ * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS
+ * BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN
+ * ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
+ * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+ * SOFTWARE.
+ */
+
+#ifndef MLX4_ICM_H
+#define MLX4_ICM_H
+
+#include <linux/list.h>
+#include <linux/pci.h>
+#include <linux/mutex.h>
+
+#define MLX4_ICM_CHUNK_LEN						\
+	((256 - sizeof(struct list_head) - 2 * sizeof(int)) /		\
+	 (sizeof(struct scatterlist)))
+
+enum {
+	MLX4_ICM_PAGE_SHIFT	= 12,
+	MLX4_ICM_PAGE_SIZE	= 1 << MLX4_ICM_PAGE_SHIFT,
+};
+
+struct mlx4_icm_buf {
+	void			*addr;
+	size_t			size;
+	dma_addr_t		dma_addr;
+};
+
+struct mlx4_icm_chunk {
+	struct list_head	list;
+	int			npages;
+	int			nsg;
+	bool			coherent;
+	union {
+		struct scatterlist	sg[MLX4_ICM_CHUNK_LEN];
+		struct mlx4_icm_buf	buf[MLX4_ICM_CHUNK_LEN];
+	};
+};
+
+struct mlx4_icm {
+	struct list_head	chunk_list;
+	int			refcount;
+};
+
+struct mlx4_icm_iter {
+	struct mlx4_icm	       *icm;
+	struct mlx4_icm_chunk  *chunk;
+	int			page_idx;
+};
+
+struct mlx4_dev;
+
+struct mlx4_icm *mlx4_alloc_icm(struct mlx4_dev *dev, int npages,
+				gfp_t gfp_mask, int coherent);
+void mlx4_free_icm(struct mlx4_dev *dev, struct mlx4_icm *icm, int coherent);
+
+int mlx4_table_get(struct mlx4_dev *dev, struct mlx4_icm_table *table, u32 obj);
+void mlx4_table_put(struct mlx4_dev *dev, struct mlx4_icm_table *table, u32 obj);
+int mlx4_table_get_range(struct mlx4_dev *dev, struct mlx4_icm_table *table,
+			 u32 start, u32 end);
+void mlx4_table_put_range(struct mlx4_dev *dev, struct mlx4_icm_table *table,
+			  u32 start, u32 end);
+int mlx4_init_icm_table(struct mlx4_dev *dev, struct mlx4_icm_table *table,
+			u64 virt, int obj_size,	u32 nobj, int reserved,
+			int use_lowmem, int use_coherent);
+void mlx4_cleanup_icm_table(struct mlx4_dev *dev, struct mlx4_icm_table *table);
+void *mlx4_table_find(struct mlx4_icm_table *table, u32 obj, dma_addr_t *dma_handle);
+
+static inline void mlx4_icm_first(struct mlx4_icm *icm,
+				  struct mlx4_icm_iter *iter)
+{
+	iter->icm      = icm;
+	iter->chunk    = list_empty(&icm->chunk_list) ?
+		NULL : list_entry(icm->chunk_list.next,
+				  struct mlx4_icm_chunk, list);
+	iter->page_idx = 0;
+}
+
+static inline int mlx4_icm_last(struct mlx4_icm_iter *iter)
+{
+	return !iter->chunk;
+}
+
+static inline void mlx4_icm_next(struct mlx4_icm_iter *iter)
+{
+	if (++iter->page_idx >= iter->chunk->nsg) {
+		if (iter->chunk->list.next == &iter->icm->chunk_list) {
+			iter->chunk = NULL;
+			return;
+		}
+
+		iter->chunk = list_entry(iter->chunk->list.next,
+					 struct mlx4_icm_chunk, list);
+		iter->page_idx = 0;
+	}
+}
+
+static inline dma_addr_t mlx4_icm_addr(struct mlx4_icm_iter *iter)
+{
+	if (iter->chunk->coherent)
+		return iter->chunk->buf[iter->page_idx].dma_addr;
+	else
+		return sg_dma_address(&iter->chunk->sg[iter->page_idx]);
+}
+
+static inline unsigned long mlx4_icm_size(struct mlx4_icm_iter *iter)
+{
+	if (iter->chunk->coherent)
+		return iter->chunk->buf[iter->page_idx].size;
+	else
+		return sg_dma_len(&iter->chunk->sg[iter->page_idx]);
+}
+
+int mlx4_MAP_ICM_AUX(struct mlx4_dev *dev, struct mlx4_icm *icm);
+int mlx4_UNMAP_ICM_AUX(struct mlx4_dev *dev);
+
+#endif /* MLX4_ICM_H */
diff --git a/drivers/net/ethernet/mellanox/mlx4/intf.c b/drivers/net/ethernet/mellanox/mlx4/intf.c
new file mode 100644
index 000000000..65482f004
--- /dev/null
+++ b/drivers/net/ethernet/mellanox/mlx4/intf.c
@@ -0,0 +1,275 @@
+/*
+ * Copyright (c) 2006, 2007 Cisco Systems, Inc. All rights reserved.
+ * Copyright (c) 2007, 2008 Mellanox Technologies. All rights reserved.
+ *
+ * This software is available to you under a choice of one of two
+ * licenses.  You may choose to be licensed under the terms of the GNU
+ * General Public License (GPL) Version 2, available from the file
+ * COPYING in the main directory of this source tree, or the
+ * OpenIB.org BSD license below:
+ *
+ *     Redistribution and use in source and binary forms, with or
+ *     without modification, are permitted provided that the following
+ *     conditions are met:
+ *
+ *      - Redistributions of source code must retain the above
+ *        copyright notice, this list of conditions and the following
+ *        disclaimer.
+ *
+ *      - Redistributions in binary form must reproduce the above
+ *        copyright notice, this list of conditions and the following
+ *        disclaimer in the documentation and/or other materials
+ *        provided with the distribution.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
+ * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
+ * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS
+ * BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN
+ * ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
+ * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+ * SOFTWARE.
+ */
+
+#include <linux/slab.h>
+#include <linux/export.h>
+#include <linux/errno.h>
+#include <net/devlink.h>
+
+#include "mlx4.h"
+
+struct mlx4_device_context {
+	struct list_head	list;
+	struct list_head	bond_list;
+	struct mlx4_interface  *intf;
+	void		       *context;
+};
+
+static LIST_HEAD(intf_list);
+static LIST_HEAD(dev_list);
+static DEFINE_MUTEX(intf_mutex);
+
+static void mlx4_add_device(struct mlx4_interface *intf, struct mlx4_priv *priv)
+{
+	struct mlx4_device_context *dev_ctx;
+
+	dev_ctx = kmalloc(sizeof(*dev_ctx), GFP_KERNEL);
+	if (!dev_ctx)
+		return;
+
+	dev_ctx->intf    = intf;
+	dev_ctx->context = intf->add(&priv->dev);
+
+	if (dev_ctx->context) {
+		spin_lock_irq(&priv->ctx_lock);
+		list_add_tail(&dev_ctx->list, &priv->ctx_list);
+		spin_unlock_irq(&priv->ctx_lock);
+		if (intf->activate)
+			intf->activate(&priv->dev, dev_ctx->context);
+	} else
+		kfree(dev_ctx);
+
+}
+
+static void mlx4_remove_device(struct mlx4_interface *intf, struct mlx4_priv *priv)
+{
+	struct mlx4_device_context *dev_ctx;
+
+	list_for_each_entry(dev_ctx, &priv->ctx_list, list)
+		if (dev_ctx->intf == intf) {
+			spin_lock_irq(&priv->ctx_lock);
+			list_del(&dev_ctx->list);
+			spin_unlock_irq(&priv->ctx_lock);
+
+			intf->remove(&priv->dev, dev_ctx->context);
+			kfree(dev_ctx);
+			return;
+		}
+}
+
+int mlx4_register_interface(struct mlx4_interface *intf)
+{
+	struct mlx4_priv *priv;
+
+	if (!intf->add || !intf->remove)
+		return -EINVAL;
+
+	mutex_lock(&intf_mutex);
+
+	list_add_tail(&intf->list, &intf_list);
+	list_for_each_entry(priv, &dev_list, dev_list) {
+		if (mlx4_is_mfunc(&priv->dev) && (intf->flags & MLX4_INTFF_BONDING)) {
+			mlx4_dbg(&priv->dev,
+				 "SRIOV, disabling HA mode for intf proto %d\n", intf->protocol);
+			intf->flags &= ~MLX4_INTFF_BONDING;
+		}
+		mlx4_add_device(intf, priv);
+	}
+
+	mutex_unlock(&intf_mutex);
+
+	return 0;
+}
+EXPORT_SYMBOL_GPL(mlx4_register_interface);
+
+void mlx4_unregister_interface(struct mlx4_interface *intf)
+{
+	struct mlx4_priv *priv;
+
+	mutex_lock(&intf_mutex);
+
+	list_for_each_entry(priv, &dev_list, dev_list)
+		mlx4_remove_device(intf, priv);
+
+	list_del(&intf->list);
+
+	mutex_unlock(&intf_mutex);
+}
+EXPORT_SYMBOL_GPL(mlx4_unregister_interface);
+
+int mlx4_do_bond(struct mlx4_dev *dev, bool enable)
+{
+	struct mlx4_priv *priv = mlx4_priv(dev);
+	struct mlx4_device_context *dev_ctx = NULL, *temp_dev_ctx;
+	unsigned long flags;
+	int ret;
+	LIST_HEAD(bond_list);
+
+	if (!(dev->caps.flags2 & MLX4_DEV_CAP_FLAG2_PORT_REMAP))
+		return -EOPNOTSUPP;
+
+	ret = mlx4_disable_rx_port_check(dev, enable);
+	if (ret) {
+		mlx4_err(dev, "Fail to %s rx port check\n",
+			 enable ? "enable" : "disable");
+		return ret;
+	}
+	if (enable) {
+		dev->flags |= MLX4_FLAG_BONDED;
+	} else {
+		ret = mlx4_virt2phy_port_map(dev, 1, 2);
+		if (ret) {
+			mlx4_err(dev, "Fail to reset port map\n");
+			return ret;
+		}
+		dev->flags &= ~MLX4_FLAG_BONDED;
+	}
+
+	spin_lock_irqsave(&priv->ctx_lock, flags);
+	list_for_each_entry_safe(dev_ctx, temp_dev_ctx, &priv->ctx_list, list) {
+		if (dev_ctx->intf->flags & MLX4_INTFF_BONDING) {
+			list_add_tail(&dev_ctx->bond_list, &bond_list);
+			list_del(&dev_ctx->list);
+		}
+	}
+	spin_unlock_irqrestore(&priv->ctx_lock, flags);
+
+	list_for_each_entry(dev_ctx, &bond_list, bond_list) {
+		dev_ctx->intf->remove(dev, dev_ctx->context);
+		dev_ctx->context =  dev_ctx->intf->add(dev);
+
+		spin_lock_irqsave(&priv->ctx_lock, flags);
+		list_add_tail(&dev_ctx->list, &priv->ctx_list);
+		spin_unlock_irqrestore(&priv->ctx_lock, flags);
+
+		mlx4_dbg(dev, "Interface for protocol %d restarted with bonded mode %s\n",
+			 dev_ctx->intf->protocol, enable ?
+			 "enabled" : "disabled");
+	}
+	return 0;
+}
+
+void mlx4_dispatch_event(struct mlx4_dev *dev, enum mlx4_dev_event type,
+			 unsigned long param)
+{
+	struct mlx4_priv *priv = mlx4_priv(dev);
+	struct mlx4_device_context *dev_ctx;
+	unsigned long flags;
+
+	spin_lock_irqsave(&priv->ctx_lock, flags);
+
+	list_for_each_entry(dev_ctx, &priv->ctx_list, list)
+		if (dev_ctx->intf->event)
+			dev_ctx->intf->event(dev, dev_ctx->context, type, param);
+
+	spin_unlock_irqrestore(&priv->ctx_lock, flags);
+}
+
+int mlx4_register_device(struct mlx4_dev *dev)
+{
+	struct mlx4_priv *priv = mlx4_priv(dev);
+	struct mlx4_interface *intf;
+
+	mutex_lock(&intf_mutex);
+
+	dev->persist->interface_state |= MLX4_INTERFACE_STATE_UP;
+	list_add_tail(&priv->dev_list, &dev_list);
+	list_for_each_entry(intf, &intf_list, list)
+		mlx4_add_device(intf, priv);
+
+	mutex_unlock(&intf_mutex);
+	mlx4_start_catas_poll(dev);
+
+	return 0;
+}
+
+void mlx4_unregister_device(struct mlx4_dev *dev)
+{
+	struct mlx4_priv *priv = mlx4_priv(dev);
+	struct mlx4_interface *intf;
+
+	if (!(dev->persist->interface_state & MLX4_INTERFACE_STATE_UP))
+		return;
+
+	mlx4_stop_catas_poll(dev);
+	if (dev->persist->interface_state & MLX4_INTERFACE_STATE_DELETION &&
+	    mlx4_is_slave(dev)) {
+		/* In mlx4_remove_one on a VF */
+		u32 slave_read =
+			swab32(readl(&mlx4_priv(dev)->mfunc.comm->slave_read));
+
+		if (mlx4_comm_internal_err(slave_read)) {
+			mlx4_dbg(dev, "%s: comm channel is down, entering error state.\n",
+				 __func__);
+			mlx4_enter_error_state(dev->persist);
+		}
+	}
+	mutex_lock(&intf_mutex);
+
+	list_for_each_entry(intf, &intf_list, list)
+		mlx4_remove_device(intf, priv);
+
+	list_del(&priv->dev_list);
+	dev->persist->interface_state &= ~MLX4_INTERFACE_STATE_UP;
+
+	mutex_unlock(&intf_mutex);
+}
+
+void *mlx4_get_protocol_dev(struct mlx4_dev *dev, enum mlx4_protocol proto, int port)
+{
+	struct mlx4_priv *priv = mlx4_priv(dev);
+	struct mlx4_device_context *dev_ctx;
+	unsigned long flags;
+	void *result = NULL;
+
+	spin_lock_irqsave(&priv->ctx_lock, flags);
+
+	list_for_each_entry(dev_ctx, &priv->ctx_list, list)
+		if (dev_ctx->intf->protocol == proto && dev_ctx->intf->get_dev) {
+			result = dev_ctx->intf->get_dev(dev, dev_ctx->context, port);
+			break;
+		}
+
+	spin_unlock_irqrestore(&priv->ctx_lock, flags);
+
+	return result;
+}
+EXPORT_SYMBOL_GPL(mlx4_get_protocol_dev);
+
+struct devlink_port *mlx4_get_devlink_port(struct mlx4_dev *dev, int port)
+{
+	struct mlx4_port_info *info = &mlx4_priv(dev)->port[port];
+
+	return &info->devlink_port;
+}
+EXPORT_SYMBOL_GPL(mlx4_get_devlink_port);
diff --git a/drivers/net/ethernet/mellanox/mlx4/main.c b/drivers/net/ethernet/mellanox/mlx4/main.c
new file mode 100644
index 000000000..d9707d47f
--- /dev/null
+++ b/drivers/net/ethernet/mellanox/mlx4/main.c
@@ -0,0 +1,4462 @@
+/*
+ * Copyright (c) 2004, 2005 Topspin Communications.  All rights reserved.
+ * Copyright (c) 2005 Sun Microsystems, Inc. All rights reserved.
+ * Copyright (c) 2005, 2006, 2007, 2008 Mellanox Technologies. All rights reserved.
+ * Copyright (c) 2006, 2007 Cisco Systems, Inc. All rights reserved.
+ *
+ * This software is available to you under a choice of one of two
+ * licenses.  You may choose to be licensed under the terms of the GNU
+ * General Public License (GPL) Version 2, available from the file
+ * COPYING in the main directory of this source tree, or the
+ * OpenIB.org BSD license below:
+ *
+ *     Redistribution and use in source and binary forms, with or
+ *     without modification, are permitted provided that the following
+ *     conditions are met:
+ *
+ *      - Redistributions of source code must retain the above
+ *        copyright notice, this list of conditions and the following
+ *        disclaimer.
+ *
+ *      - Redistributions in binary form must reproduce the above
+ *        copyright notice, this list of conditions and the following
+ *        disclaimer in the documentation and/or other materials
+ *        provided with the distribution.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
+ * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
+ * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS
+ * BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN
+ * ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
+ * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+ * SOFTWARE.
+ */
+
+#include <linux/module.h>
+#include <linux/kernel.h>
+#include <linux/init.h>
+#include <linux/errno.h>
+#include <linux/pci.h>
+#include <linux/dma-mapping.h>
+#include <linux/slab.h>
+#include <linux/io-mapping.h>
+#include <linux/delay.h>
+#include <linux/kmod.h>
+#include <linux/etherdevice.h>
+#include <net/devlink.h>
+
+#include <uapi/rdma/mlx4-abi.h>
+#include <linux/mlx4/device.h>
+#include <linux/mlx4/doorbell.h>
+
+#include "mlx4.h"
+#include "fw.h"
+#include "icm.h"
+
+MODULE_AUTHOR("Roland Dreier");
+MODULE_DESCRIPTION("Mellanox ConnectX HCA low-level driver");
+MODULE_LICENSE("Dual BSD/GPL");
+MODULE_VERSION(DRV_VERSION);
+
+struct workqueue_struct *mlx4_wq;
+
+#ifdef CONFIG_MLX4_DEBUG
+
+int mlx4_debug_level = 0;
+module_param_named(debug_level, mlx4_debug_level, int, 0644);
+MODULE_PARM_DESC(debug_level, "Enable debug tracing if > 0");
+
+#endif /* CONFIG_MLX4_DEBUG */
+
+#ifdef CONFIG_PCI_MSI
+
+static int msi_x = 1;
+module_param(msi_x, int, 0444);
+MODULE_PARM_DESC(msi_x, "0 - don't use MSI-X, 1 - use MSI-X, >1 - limit number of MSI-X irqs to msi_x");
+
+#else /* CONFIG_PCI_MSI */
+
+#define msi_x (0)
+
+#endif /* CONFIG_PCI_MSI */
+
+static uint8_t num_vfs[3] = {0, 0, 0};
+static int num_vfs_argc;
+module_param_array(num_vfs, byte , &num_vfs_argc, 0444);
+MODULE_PARM_DESC(num_vfs, "enable #num_vfs functions if num_vfs > 0\n"
+			  "num_vfs=port1,port2,port1+2");
+
+static uint8_t probe_vf[3] = {0, 0, 0};
+static int probe_vfs_argc;
+module_param_array(probe_vf, byte, &probe_vfs_argc, 0444);
+MODULE_PARM_DESC(probe_vf, "number of vfs to probe by pf driver (num_vfs > 0)\n"
+			   "probe_vf=port1,port2,port1+2");
+
+static int mlx4_log_num_mgm_entry_size = MLX4_DEFAULT_MGM_LOG_ENTRY_SIZE;
+module_param_named(log_num_mgm_entry_size,
+			mlx4_log_num_mgm_entry_size, int, 0444);
+MODULE_PARM_DESC(log_num_mgm_entry_size, "log mgm size, that defines the num"
+					 " of qp per mcg, for example:"
+					 " 10 gives 248.range: 7 <="
+					 " log_num_mgm_entry_size <= 12."
+					 " To activate device managed"
+					 " flow steering when available, set to -1");
+
+static bool enable_64b_cqe_eqe = true;
+module_param(enable_64b_cqe_eqe, bool, 0444);
+MODULE_PARM_DESC(enable_64b_cqe_eqe,
+		 "Enable 64 byte CQEs/EQEs when the FW supports this (default: True)");
+
+static bool enable_4k_uar;
+module_param(enable_4k_uar, bool, 0444);
+MODULE_PARM_DESC(enable_4k_uar,
+		 "Enable using 4K UAR. Should not be enabled if have VFs which do not support 4K UARs (default: false)");
+
+#define PF_CONTEXT_BEHAVIOUR_MASK	(MLX4_FUNC_CAP_64B_EQE_CQE | \
+					 MLX4_FUNC_CAP_EQE_CQE_STRIDE | \
+					 MLX4_FUNC_CAP_DMFS_A0_STATIC)
+
+#define RESET_PERSIST_MASK_FLAGS	(MLX4_FLAG_SRIOV)
+
+static char mlx4_version[] =
+	DRV_NAME ": Mellanox ConnectX core driver v"
+	DRV_VERSION "\n";
+
+static const struct mlx4_profile default_profile = {
+	.num_qp		= 1 << 18,
+	.num_srq	= 1 << 16,
+	.rdmarc_per_qp	= 1 << 4,
+	.num_cq		= 1 << 16,
+	.num_mcg	= 1 << 13,
+	.num_mpt	= 1 << 19,
+	.num_mtt	= 1 << 20, /* It is really num mtt segements */
+};
+
+static const struct mlx4_profile low_mem_profile = {
+	.num_qp		= 1 << 17,
+	.num_srq	= 1 << 6,
+	.rdmarc_per_qp	= 1 << 4,
+	.num_cq		= 1 << 8,
+	.num_mcg	= 1 << 8,
+	.num_mpt	= 1 << 9,
+	.num_mtt	= 1 << 7,
+};
+
+static int log_num_mac = 7;
+module_param_named(log_num_mac, log_num_mac, int, 0444);
+MODULE_PARM_DESC(log_num_mac, "Log2 max number of MACs per ETH port (1-7)");
+
+static int log_num_vlan;
+module_param_named(log_num_vlan, log_num_vlan, int, 0444);
+MODULE_PARM_DESC(log_num_vlan, "Log2 max number of VLANs per ETH port (0-7)");
+/* Log2 max number of VLANs per ETH port (0-7) */
+#define MLX4_LOG_NUM_VLANS 7
+#define MLX4_MIN_LOG_NUM_VLANS 0
+#define MLX4_MIN_LOG_NUM_MAC 1
+
+static bool use_prio;
+module_param_named(use_prio, use_prio, bool, 0444);
+MODULE_PARM_DESC(use_prio, "Enable steering by VLAN priority on ETH ports (deprecated)");
+
+int log_mtts_per_seg = ilog2(1);
+module_param_named(log_mtts_per_seg, log_mtts_per_seg, int, 0444);
+MODULE_PARM_DESC(log_mtts_per_seg, "Log2 number of MTT entries per segment "
+		 "(0-7) (default: 0)");
+
+static int port_type_array[2] = {MLX4_PORT_TYPE_NONE, MLX4_PORT_TYPE_NONE};
+static int arr_argc = 2;
+module_param_array(port_type_array, int, &arr_argc, 0444);
+MODULE_PARM_DESC(port_type_array, "Array of port types: HW_DEFAULT (0) is default "
+				"1 for IB, 2 for Ethernet");
+
+struct mlx4_port_config {
+	struct list_head list;
+	enum mlx4_port_type port_type[MLX4_MAX_PORTS + 1];
+	struct pci_dev *pdev;
+};
+
+static atomic_t pf_loading = ATOMIC_INIT(0);
+
+static int mlx4_devlink_ierr_reset_get(struct devlink *devlink, u32 id,
+				       struct devlink_param_gset_ctx *ctx)
+{
+	ctx->val.vbool = !!mlx4_internal_err_reset;
+	return 0;
+}
+
+static int mlx4_devlink_ierr_reset_set(struct devlink *devlink, u32 id,
+				       struct devlink_param_gset_ctx *ctx)
+{
+	mlx4_internal_err_reset = ctx->val.vbool;
+	return 0;
+}
+
+static int mlx4_devlink_crdump_snapshot_get(struct devlink *devlink, u32 id,
+					    struct devlink_param_gset_ctx *ctx)
+{
+	struct mlx4_priv *priv = devlink_priv(devlink);
+	struct mlx4_dev *dev = &priv->dev;
+
+	ctx->val.vbool = dev->persist->crdump.snapshot_enable;
+	return 0;
+}
+
+static int mlx4_devlink_crdump_snapshot_set(struct devlink *devlink, u32 id,
+					    struct devlink_param_gset_ctx *ctx)
+{
+	struct mlx4_priv *priv = devlink_priv(devlink);
+	struct mlx4_dev *dev = &priv->dev;
+
+	dev->persist->crdump.snapshot_enable = ctx->val.vbool;
+	return 0;
+}
+
+static int
+mlx4_devlink_max_macs_validate(struct devlink *devlink, u32 id,
+			       union devlink_param_value val,
+			       struct netlink_ext_ack *extack)
+{
+	u32 value = val.vu32;
+
+	if (value < 1 || value > 128)
+		return -ERANGE;
+
+	if (!is_power_of_2(value)) {
+		NL_SET_ERR_MSG_MOD(extack, "max_macs supported must be power of 2");
+		return -EINVAL;
+	}
+
+	return 0;
+}
+
+enum mlx4_devlink_param_id {
+	MLX4_DEVLINK_PARAM_ID_BASE = DEVLINK_PARAM_GENERIC_ID_MAX,
+	MLX4_DEVLINK_PARAM_ID_ENABLE_64B_CQE_EQE,
+	MLX4_DEVLINK_PARAM_ID_ENABLE_4K_UAR,
+};
+
+static const struct devlink_param mlx4_devlink_params[] = {
+	DEVLINK_PARAM_GENERIC(INT_ERR_RESET,
+			      BIT(DEVLINK_PARAM_CMODE_RUNTIME) |
+			      BIT(DEVLINK_PARAM_CMODE_DRIVERINIT),
+			      mlx4_devlink_ierr_reset_get,
+			      mlx4_devlink_ierr_reset_set, NULL),
+	DEVLINK_PARAM_GENERIC(MAX_MACS,
+			      BIT(DEVLINK_PARAM_CMODE_DRIVERINIT),
+			      NULL, NULL, mlx4_devlink_max_macs_validate),
+	DEVLINK_PARAM_GENERIC(REGION_SNAPSHOT,
+			      BIT(DEVLINK_PARAM_CMODE_RUNTIME) |
+			      BIT(DEVLINK_PARAM_CMODE_DRIVERINIT),
+			      mlx4_devlink_crdump_snapshot_get,
+			      mlx4_devlink_crdump_snapshot_set, NULL),
+	DEVLINK_PARAM_DRIVER(MLX4_DEVLINK_PARAM_ID_ENABLE_64B_CQE_EQE,
+			     "enable_64b_cqe_eqe", DEVLINK_PARAM_TYPE_BOOL,
+			     BIT(DEVLINK_PARAM_CMODE_DRIVERINIT),
+			     NULL, NULL, NULL),
+	DEVLINK_PARAM_DRIVER(MLX4_DEVLINK_PARAM_ID_ENABLE_4K_UAR,
+			     "enable_4k_uar", DEVLINK_PARAM_TYPE_BOOL,
+			     BIT(DEVLINK_PARAM_CMODE_DRIVERINIT),
+			     NULL, NULL, NULL),
+};
+
+static void mlx4_devlink_set_params_init_values(struct devlink *devlink)
+{
+	union devlink_param_value value;
+
+	value.vbool = !!mlx4_internal_err_reset;
+	devlink_param_driverinit_value_set(devlink,
+					   DEVLINK_PARAM_GENERIC_ID_INT_ERR_RESET,
+					   value);
+
+	value.vu32 = 1UL << log_num_mac;
+	devlink_param_driverinit_value_set(devlink,
+					   DEVLINK_PARAM_GENERIC_ID_MAX_MACS,
+					   value);
+
+	value.vbool = enable_64b_cqe_eqe;
+	devlink_param_driverinit_value_set(devlink,
+					   MLX4_DEVLINK_PARAM_ID_ENABLE_64B_CQE_EQE,
+					   value);
+
+	value.vbool = enable_4k_uar;
+	devlink_param_driverinit_value_set(devlink,
+					   MLX4_DEVLINK_PARAM_ID_ENABLE_4K_UAR,
+					   value);
+
+	value.vbool = false;
+	devlink_param_driverinit_value_set(devlink,
+					   DEVLINK_PARAM_GENERIC_ID_REGION_SNAPSHOT,
+					   value);
+}
+
+static inline void mlx4_set_num_reserved_uars(struct mlx4_dev *dev,
+					      struct mlx4_dev_cap *dev_cap)
+{
+	/* The reserved_uars is calculated by system page size unit.
+	 * Therefore, adjustment is added when the uar page size is less
+	 * than the system page size
+	 */
+	dev->caps.reserved_uars	=
+		max_t(int,
+		      mlx4_get_num_reserved_uar(dev),
+		      dev_cap->reserved_uars /
+			(1 << (PAGE_SHIFT - dev->uar_page_shift)));
+}
+
+int mlx4_check_port_params(struct mlx4_dev *dev,
+			   enum mlx4_port_type *port_type)
+{
+	int i;
+
+	if (!(dev->caps.flags & MLX4_DEV_CAP_FLAG_DPDP)) {
+		for (i = 0; i < dev->caps.num_ports - 1; i++) {
+			if (port_type[i] != port_type[i + 1]) {
+				mlx4_err(dev, "Only same port types supported on this HCA, aborting\n");
+				return -EOPNOTSUPP;
+			}
+		}
+	}
+
+	for (i = 0; i < dev->caps.num_ports; i++) {
+		if (!(port_type[i] & dev->caps.supported_type[i+1])) {
+			mlx4_err(dev, "Requested port type for port %d is not supported on this HCA\n",
+				 i + 1);
+			return -EOPNOTSUPP;
+		}
+	}
+	return 0;
+}
+
+static void mlx4_set_port_mask(struct mlx4_dev *dev)
+{
+	int i;
+
+	for (i = 1; i <= dev->caps.num_ports; ++i)
+		dev->caps.port_mask[i] = dev->caps.port_type[i];
+}
+
+enum {
+	MLX4_QUERY_FUNC_NUM_SYS_EQS = 1 << 0,
+};
+
+static int mlx4_query_func(struct mlx4_dev *dev, struct mlx4_dev_cap *dev_cap)
+{
+	int err = 0;
+	struct mlx4_func func;
+
+	if (dev->caps.flags2 & MLX4_DEV_CAP_FLAG2_SYS_EQS) {
+		err = mlx4_QUERY_FUNC(dev, &func, 0);
+		if (err) {
+			mlx4_err(dev, "QUERY_DEV_CAP command failed, aborting.\n");
+			return err;
+		}
+		dev_cap->max_eqs = func.max_eq;
+		dev_cap->reserved_eqs = func.rsvd_eqs;
+		dev_cap->reserved_uars = func.rsvd_uars;
+		err |= MLX4_QUERY_FUNC_NUM_SYS_EQS;
+	}
+	return err;
+}
+
+static void mlx4_enable_cqe_eqe_stride(struct mlx4_dev *dev)
+{
+	struct mlx4_caps *dev_cap = &dev->caps;
+
+	/* FW not supporting or cancelled by user */
+	if (!(dev_cap->flags2 & MLX4_DEV_CAP_FLAG2_EQE_STRIDE) ||
+	    !(dev_cap->flags2 & MLX4_DEV_CAP_FLAG2_CQE_STRIDE))
+		return;
+
+	/* Must have 64B CQE_EQE enabled by FW to use bigger stride
+	 * When FW has NCSI it may decide not to report 64B CQE/EQEs
+	 */
+	if (!(dev_cap->flags & MLX4_DEV_CAP_FLAG_64B_EQE) ||
+	    !(dev_cap->flags & MLX4_DEV_CAP_FLAG_64B_CQE)) {
+		dev_cap->flags2 &= ~MLX4_DEV_CAP_FLAG2_CQE_STRIDE;
+		dev_cap->flags2 &= ~MLX4_DEV_CAP_FLAG2_EQE_STRIDE;
+		return;
+	}
+
+	if (cache_line_size() == 128 || cache_line_size() == 256) {
+		mlx4_dbg(dev, "Enabling CQE stride cacheLine supported\n");
+		/* Changing the real data inside CQE size to 32B */
+		dev_cap->flags &= ~MLX4_DEV_CAP_FLAG_64B_CQE;
+		dev_cap->flags &= ~MLX4_DEV_CAP_FLAG_64B_EQE;
+
+		if (mlx4_is_master(dev))
+			dev_cap->function_caps |= MLX4_FUNC_CAP_EQE_CQE_STRIDE;
+	} else {
+		if (cache_line_size() != 32  && cache_line_size() != 64)
+			mlx4_dbg(dev, "Disabling CQE stride, cacheLine size unsupported\n");
+		dev_cap->flags2 &= ~MLX4_DEV_CAP_FLAG2_CQE_STRIDE;
+		dev_cap->flags2 &= ~MLX4_DEV_CAP_FLAG2_EQE_STRIDE;
+	}
+}
+
+static int _mlx4_dev_port(struct mlx4_dev *dev, int port,
+			  struct mlx4_port_cap *port_cap)
+{
+	dev->caps.vl_cap[port]	    = port_cap->max_vl;
+	dev->caps.ib_mtu_cap[port]	    = port_cap->ib_mtu;
+	dev->phys_caps.gid_phys_table_len[port]  = port_cap->max_gids;
+	dev->phys_caps.pkey_phys_table_len[port] = port_cap->max_pkeys;
+	/* set gid and pkey table operating lengths by default
+	 * to non-sriov values
+	 */
+	dev->caps.gid_table_len[port]  = port_cap->max_gids;
+	dev->caps.pkey_table_len[port] = port_cap->max_pkeys;
+	dev->caps.port_width_cap[port] = port_cap->max_port_width;
+	dev->caps.eth_mtu_cap[port]    = port_cap->eth_mtu;
+	dev->caps.max_tc_eth	       = port_cap->max_tc_eth;
+	dev->caps.def_mac[port]        = port_cap->def_mac;
+	dev->caps.supported_type[port] = port_cap->supported_port_types;
+	dev->caps.suggested_type[port] = port_cap->suggested_type;
+	dev->caps.default_sense[port] = port_cap->default_sense;
+	dev->caps.trans_type[port]	    = port_cap->trans_type;
+	dev->caps.vendor_oui[port]     = port_cap->vendor_oui;
+	dev->caps.wavelength[port]     = port_cap->wavelength;
+	dev->caps.trans_code[port]     = port_cap->trans_code;
+
+	return 0;
+}
+
+static int mlx4_dev_port(struct mlx4_dev *dev, int port,
+			 struct mlx4_port_cap *port_cap)
+{
+	int err = 0;
+
+	err = mlx4_QUERY_PORT(dev, port, port_cap);
+
+	if (err)
+		mlx4_err(dev, "QUERY_PORT command failed.\n");
+
+	return err;
+}
+
+static inline void mlx4_enable_ignore_fcs(struct mlx4_dev *dev)
+{
+	if (!(dev->caps.flags2 & MLX4_DEV_CAP_FLAG2_IGNORE_FCS))
+		return;
+
+	if (mlx4_is_mfunc(dev)) {
+		mlx4_dbg(dev, "SRIOV mode - Disabling Ignore FCS");
+		dev->caps.flags2 &= ~MLX4_DEV_CAP_FLAG2_IGNORE_FCS;
+		return;
+	}
+
+	if (!(dev->caps.flags & MLX4_DEV_CAP_FLAG_FCS_KEEP)) {
+		mlx4_dbg(dev,
+			 "Keep FCS is not supported - Disabling Ignore FCS");
+		dev->caps.flags2 &= ~MLX4_DEV_CAP_FLAG2_IGNORE_FCS;
+		return;
+	}
+}
+
+#define MLX4_A0_STEERING_TABLE_SIZE	256
+static int mlx4_dev_cap(struct mlx4_dev *dev, struct mlx4_dev_cap *dev_cap)
+{
+	int err;
+	int i;
+
+	err = mlx4_QUERY_DEV_CAP(dev, dev_cap);
+	if (err) {
+		mlx4_err(dev, "QUERY_DEV_CAP command failed, aborting\n");
+		return err;
+	}
+	mlx4_dev_cap_dump(dev, dev_cap);
+
+	if (dev_cap->min_page_sz > PAGE_SIZE) {
+		mlx4_err(dev, "HCA minimum page size of %d bigger than kernel PAGE_SIZE of %ld, aborting\n",
+			 dev_cap->min_page_sz, PAGE_SIZE);
+		return -ENODEV;
+	}
+	if (dev_cap->num_ports > MLX4_MAX_PORTS) {
+		mlx4_err(dev, "HCA has %d ports, but we only support %d, aborting\n",
+			 dev_cap->num_ports, MLX4_MAX_PORTS);
+		return -ENODEV;
+	}
+
+	if (dev_cap->uar_size > pci_resource_len(dev->persist->pdev, 2)) {
+		mlx4_err(dev, "HCA reported UAR size of 0x%x bigger than PCI resource 2 size of 0x%llx, aborting\n",
+			 dev_cap->uar_size,
+			 (unsigned long long)
+			 pci_resource_len(dev->persist->pdev, 2));
+		return -ENODEV;
+	}
+
+	dev->caps.num_ports	     = dev_cap->num_ports;
+	dev->caps.num_sys_eqs = dev_cap->num_sys_eqs;
+	dev->phys_caps.num_phys_eqs = dev_cap->flags2 & MLX4_DEV_CAP_FLAG2_SYS_EQS ?
+				      dev->caps.num_sys_eqs :
+				      MLX4_MAX_EQ_NUM;
+	for (i = 1; i <= dev->caps.num_ports; ++i) {
+		err = _mlx4_dev_port(dev, i, dev_cap->port_cap + i);
+		if (err) {
+			mlx4_err(dev, "QUERY_PORT command failed, aborting\n");
+			return err;
+		}
+	}
+
+	dev->caps.map_clock_to_user  = dev_cap->map_clock_to_user;
+	dev->caps.uar_page_size	     = PAGE_SIZE;
+	dev->caps.num_uars	     = dev_cap->uar_size / PAGE_SIZE;
+	dev->caps.local_ca_ack_delay = dev_cap->local_ca_ack_delay;
+	dev->caps.bf_reg_size	     = dev_cap->bf_reg_size;
+	dev->caps.bf_regs_per_page   = dev_cap->bf_regs_per_page;
+	dev->caps.max_sq_sg	     = dev_cap->max_sq_sg;
+	dev->caps.max_rq_sg	     = dev_cap->max_rq_sg;
+	dev->caps.max_wqes	     = dev_cap->max_qp_sz;
+	dev->caps.max_qp_init_rdma   = dev_cap->max_requester_per_qp;
+	dev->caps.max_srq_wqes	     = dev_cap->max_srq_sz;
+	dev->caps.max_srq_sge	     = dev_cap->max_rq_sg - 1;
+	dev->caps.reserved_srqs	     = dev_cap->reserved_srqs;
+	dev->caps.max_sq_desc_sz     = dev_cap->max_sq_desc_sz;
+	dev->caps.max_rq_desc_sz     = dev_cap->max_rq_desc_sz;
+	/*
+	 * Subtract 1 from the limit because we need to allocate a
+	 * spare CQE so the HCA HW can tell the difference between an
+	 * empty CQ and a full CQ.
+	 */
+	dev->caps.max_cqes	     = dev_cap->max_cq_sz - 1;
+	dev->caps.reserved_cqs	     = dev_cap->reserved_cqs;
+	dev->caps.reserved_eqs	     = dev_cap->reserved_eqs;
+	dev->caps.reserved_mtts      = dev_cap->reserved_mtts;
+	dev->caps.reserved_mrws	     = dev_cap->reserved_mrws;
+
+	dev->caps.reserved_pds	     = dev_cap->reserved_pds;
+	dev->caps.reserved_xrcds     = (dev->caps.flags & MLX4_DEV_CAP_FLAG_XRC) ?
+					dev_cap->reserved_xrcds : 0;
+	dev->caps.max_xrcds          = (dev->caps.flags & MLX4_DEV_CAP_FLAG_XRC) ?
+					dev_cap->max_xrcds : 0;
+	dev->caps.mtt_entry_sz       = dev_cap->mtt_entry_sz;
+
+	dev->caps.max_msg_sz         = dev_cap->max_msg_sz;
+	dev->caps.page_size_cap	     = ~(u32) (dev_cap->min_page_sz - 1);
+	dev->caps.flags		     = dev_cap->flags;
+	dev->caps.flags2	     = dev_cap->flags2;
+	dev->caps.bmme_flags	     = dev_cap->bmme_flags;
+	dev->caps.reserved_lkey	     = dev_cap->reserved_lkey;
+	dev->caps.stat_rate_support  = dev_cap->stat_rate_support;
+	dev->caps.max_gso_sz	     = dev_cap->max_gso_sz;
+	dev->caps.max_rss_tbl_sz     = dev_cap->max_rss_tbl_sz;
+	dev->caps.wol_port[1]          = dev_cap->wol_port[1];
+	dev->caps.wol_port[2]          = dev_cap->wol_port[2];
+	dev->caps.health_buffer_addrs  = dev_cap->health_buffer_addrs;
+
+	/* Save uar page shift */
+	if (!mlx4_is_slave(dev)) {
+		/* Virtual PCI function needs to determine UAR page size from
+		 * firmware. Only master PCI function can set the uar page size
+		 */
+		if (enable_4k_uar || !dev->persist->num_vfs)
+			dev->uar_page_shift = DEFAULT_UAR_PAGE_SHIFT;
+		else
+			dev->uar_page_shift = PAGE_SHIFT;
+
+		mlx4_set_num_reserved_uars(dev, dev_cap);
+	}
+
+	if (dev->caps.flags2 & MLX4_DEV_CAP_FLAG2_PHV_EN) {
+		struct mlx4_init_hca_param hca_param;
+
+		memset(&hca_param, 0, sizeof(hca_param));
+		err = mlx4_QUERY_HCA(dev, &hca_param);
+		/* Turn off PHV_EN flag in case phv_check_en is set.
+		 * phv_check_en is a HW check that parse the packet and verify
+		 * phv bit was reported correctly in the wqe. To allow QinQ
+		 * PHV_EN flag should be set and phv_check_en must be cleared
+		 * otherwise QinQ packets will be drop by the HW.
+		 */
+		if (err || hca_param.phv_check_en)
+			dev->caps.flags2 &= ~MLX4_DEV_CAP_FLAG2_PHV_EN;
+	}
+
+	/* Sense port always allowed on supported devices for ConnectX-1 and -2 */
+	if (mlx4_priv(dev)->pci_dev_data & MLX4_PCI_DEV_FORCE_SENSE_PORT)
+		dev->caps.flags |= MLX4_DEV_CAP_FLAG_SENSE_SUPPORT;
+	/* Don't do sense port on multifunction devices (for now at least) */
+	if (mlx4_is_mfunc(dev))
+		dev->caps.flags &= ~MLX4_DEV_CAP_FLAG_SENSE_SUPPORT;
+
+	if (mlx4_low_memory_profile()) {
+		dev->caps.log_num_macs  = MLX4_MIN_LOG_NUM_MAC;
+		dev->caps.log_num_vlans = MLX4_MIN_LOG_NUM_VLANS;
+	} else {
+		dev->caps.log_num_macs  = log_num_mac;
+		dev->caps.log_num_vlans = MLX4_LOG_NUM_VLANS;
+	}
+
+	for (i = 1; i <= dev->caps.num_ports; ++i) {
+		dev->caps.port_type[i] = MLX4_PORT_TYPE_NONE;
+		if (dev->caps.supported_type[i]) {
+			/* if only ETH is supported - assign ETH */
+			if (dev->caps.supported_type[i] == MLX4_PORT_TYPE_ETH)
+				dev->caps.port_type[i] = MLX4_PORT_TYPE_ETH;
+			/* if only IB is supported, assign IB */
+			else if (dev->caps.supported_type[i] ==
+				 MLX4_PORT_TYPE_IB)
+				dev->caps.port_type[i] = MLX4_PORT_TYPE_IB;
+			else {
+				/* if IB and ETH are supported, we set the port
+				 * type according to user selection of port type;
+				 * if user selected none, take the FW hint */
+				if (port_type_array[i - 1] == MLX4_PORT_TYPE_NONE)
+					dev->caps.port_type[i] = dev->caps.suggested_type[i] ?
+						MLX4_PORT_TYPE_ETH : MLX4_PORT_TYPE_IB;
+				else
+					dev->caps.port_type[i] = port_type_array[i - 1];
+			}
+		}
+		/*
+		 * Link sensing is allowed on the port if 3 conditions are true:
+		 * 1. Both protocols are supported on the port.
+		 * 2. Different types are supported on the port
+		 * 3. FW declared that it supports link sensing
+		 */
+		mlx4_priv(dev)->sense.sense_allowed[i] =
+			((dev->caps.supported_type[i] == MLX4_PORT_TYPE_AUTO) &&
+			 (dev->caps.flags & MLX4_DEV_CAP_FLAG_DPDP) &&
+			 (dev->caps.flags & MLX4_DEV_CAP_FLAG_SENSE_SUPPORT));
+
+		/*
+		 * If "default_sense" bit is set, we move the port to "AUTO" mode
+		 * and perform sense_port FW command to try and set the correct
+		 * port type from beginning
+		 */
+		if (mlx4_priv(dev)->sense.sense_allowed[i] && dev->caps.default_sense[i]) {
+			enum mlx4_port_type sensed_port = MLX4_PORT_TYPE_NONE;
+			dev->caps.possible_type[i] = MLX4_PORT_TYPE_AUTO;
+			mlx4_SENSE_PORT(dev, i, &sensed_port);
+			if (sensed_port != MLX4_PORT_TYPE_NONE)
+				dev->caps.port_type[i] = sensed_port;
+		} else {
+			dev->caps.possible_type[i] = dev->caps.port_type[i];
+		}
+
+		if (dev->caps.log_num_macs > dev_cap->port_cap[i].log_max_macs) {
+			dev->caps.log_num_macs = dev_cap->port_cap[i].log_max_macs;
+			mlx4_warn(dev, "Requested number of MACs is too much for port %d, reducing to %d\n",
+				  i, 1 << dev->caps.log_num_macs);
+		}
+		if (dev->caps.log_num_vlans > dev_cap->port_cap[i].log_max_vlans) {
+			dev->caps.log_num_vlans = dev_cap->port_cap[i].log_max_vlans;
+			mlx4_warn(dev, "Requested number of VLANs is too much for port %d, reducing to %d\n",
+				  i, 1 << dev->caps.log_num_vlans);
+		}
+	}
+
+	if (mlx4_is_master(dev) && (dev->caps.num_ports == 2) &&
+	    (port_type_array[0] == MLX4_PORT_TYPE_IB) &&
+	    (port_type_array[1] == MLX4_PORT_TYPE_ETH)) {
+		mlx4_warn(dev,
+			  "Granular QoS per VF not supported with IB/Eth configuration\n");
+		dev->caps.flags2 &= ~MLX4_DEV_CAP_FLAG2_QOS_VPP;
+	}
+
+	dev->caps.max_counters = dev_cap->max_counters;
+
+	dev->caps.reserved_qps_cnt[MLX4_QP_REGION_FW] = dev_cap->reserved_qps;
+	dev->caps.reserved_qps_cnt[MLX4_QP_REGION_ETH_ADDR] =
+		dev->caps.reserved_qps_cnt[MLX4_QP_REGION_FC_ADDR] =
+		(1 << dev->caps.log_num_macs) *
+		(1 << dev->caps.log_num_vlans) *
+		dev->caps.num_ports;
+	dev->caps.reserved_qps_cnt[MLX4_QP_REGION_FC_EXCH] = MLX4_NUM_FEXCH;
+
+	if (dev_cap->dmfs_high_rate_qpn_base > 0 &&
+	    dev->caps.flags2 & MLX4_DEV_CAP_FLAG2_FS_EN)
+		dev->caps.dmfs_high_rate_qpn_base = dev_cap->dmfs_high_rate_qpn_base;
+	else
+		dev->caps.dmfs_high_rate_qpn_base =
+			dev->caps.reserved_qps_cnt[MLX4_QP_REGION_FW];
+
+	if (dev_cap->dmfs_high_rate_qpn_range > 0 &&
+	    dev->caps.flags2 & MLX4_DEV_CAP_FLAG2_FS_EN) {
+		dev->caps.dmfs_high_rate_qpn_range = dev_cap->dmfs_high_rate_qpn_range;
+		dev->caps.dmfs_high_steer_mode = MLX4_STEERING_DMFS_A0_DEFAULT;
+		dev->caps.flags2 |= MLX4_DEV_CAP_FLAG2_FS_A0;
+	} else {
+		dev->caps.dmfs_high_steer_mode = MLX4_STEERING_DMFS_A0_NOT_SUPPORTED;
+		dev->caps.dmfs_high_rate_qpn_base =
+			dev->caps.reserved_qps_cnt[MLX4_QP_REGION_FW];
+		dev->caps.dmfs_high_rate_qpn_range = MLX4_A0_STEERING_TABLE_SIZE;
+	}
+
+	dev->caps.rl_caps = dev_cap->rl_caps;
+
+	dev->caps.reserved_qps_cnt[MLX4_QP_REGION_RSS_RAW_ETH] =
+		dev->caps.dmfs_high_rate_qpn_range;
+
+	dev->caps.reserved_qps = dev->caps.reserved_qps_cnt[MLX4_QP_REGION_FW] +
+		dev->caps.reserved_qps_cnt[MLX4_QP_REGION_ETH_ADDR] +
+		dev->caps.reserved_qps_cnt[MLX4_QP_REGION_FC_ADDR] +
+		dev->caps.reserved_qps_cnt[MLX4_QP_REGION_FC_EXCH];
+
+	dev->caps.sqp_demux = (mlx4_is_master(dev)) ? MLX4_MAX_NUM_SLAVES : 0;
+
+	if (!enable_64b_cqe_eqe && !mlx4_is_slave(dev)) {
+		if (dev_cap->flags &
+		    (MLX4_DEV_CAP_FLAG_64B_CQE | MLX4_DEV_CAP_FLAG_64B_EQE)) {
+			mlx4_warn(dev, "64B EQEs/CQEs supported by the device but not enabled\n");
+			dev->caps.flags &= ~MLX4_DEV_CAP_FLAG_64B_CQE;
+			dev->caps.flags &= ~MLX4_DEV_CAP_FLAG_64B_EQE;
+		}
+
+		if (dev_cap->flags2 &
+		    (MLX4_DEV_CAP_FLAG2_CQE_STRIDE |
+		     MLX4_DEV_CAP_FLAG2_EQE_STRIDE)) {
+			mlx4_warn(dev, "Disabling EQE/CQE stride per user request\n");
+			dev_cap->flags2 &= ~MLX4_DEV_CAP_FLAG2_CQE_STRIDE;
+			dev_cap->flags2 &= ~MLX4_DEV_CAP_FLAG2_EQE_STRIDE;
+		}
+	}
+
+	if ((dev->caps.flags &
+	    (MLX4_DEV_CAP_FLAG_64B_CQE | MLX4_DEV_CAP_FLAG_64B_EQE)) &&
+	    mlx4_is_master(dev))
+		dev->caps.function_caps |= MLX4_FUNC_CAP_64B_EQE_CQE;
+
+	if (!mlx4_is_slave(dev)) {
+		mlx4_enable_cqe_eqe_stride(dev);
+		dev->caps.alloc_res_qp_mask =
+			(dev->caps.bf_reg_size ? MLX4_RESERVE_ETH_BF_QP : 0) |
+			MLX4_RESERVE_A0_QP;
+
+		if (!(dev->caps.flags2 & MLX4_DEV_CAP_FLAG2_ETS_CFG) &&
+		    dev->caps.flags & MLX4_DEV_CAP_FLAG_SET_ETH_SCHED) {
+			mlx4_warn(dev, "Old device ETS support detected\n");
+			mlx4_warn(dev, "Consider upgrading device FW.\n");
+			dev->caps.flags2 |= MLX4_DEV_CAP_FLAG2_ETS_CFG;
+		}
+
+	} else {
+		dev->caps.alloc_res_qp_mask = 0;
+	}
+
+	mlx4_enable_ignore_fcs(dev);
+
+	return 0;
+}
+
+/*The function checks if there are live vf, return the num of them*/
+static int mlx4_how_many_lives_vf(struct mlx4_dev *dev)
+{
+	struct mlx4_priv *priv = mlx4_priv(dev);
+	struct mlx4_slave_state *s_state;
+	int i;
+	int ret = 0;
+
+	for (i = 1/*the ppf is 0*/; i < dev->num_slaves; ++i) {
+		s_state = &priv->mfunc.master.slave_state[i];
+		if (s_state->active && s_state->last_cmd !=
+		    MLX4_COMM_CMD_RESET) {
+			mlx4_warn(dev, "%s: slave: %d is still active\n",
+				  __func__, i);
+			ret++;
+		}
+	}
+	return ret;
+}
+
+int mlx4_get_parav_qkey(struct mlx4_dev *dev, u32 qpn, u32 *qkey)
+{
+	u32 qk = MLX4_RESERVED_QKEY_BASE;
+
+	if (qpn >= dev->phys_caps.base_tunnel_sqpn + 8 * MLX4_MFUNC_MAX ||
+	    qpn < dev->phys_caps.base_proxy_sqpn)
+		return -EINVAL;
+
+	if (qpn >= dev->phys_caps.base_tunnel_sqpn)
+		/* tunnel qp */
+		qk += qpn - dev->phys_caps.base_tunnel_sqpn;
+	else
+		qk += qpn - dev->phys_caps.base_proxy_sqpn;
+	*qkey = qk;
+	return 0;
+}
+EXPORT_SYMBOL(mlx4_get_parav_qkey);
+
+void mlx4_sync_pkey_table(struct mlx4_dev *dev, int slave, int port, int i, int val)
+{
+	struct mlx4_priv *priv = container_of(dev, struct mlx4_priv, dev);
+
+	if (!mlx4_is_master(dev))
+		return;
+
+	priv->virt2phys_pkey[slave][port - 1][i] = val;
+}
+EXPORT_SYMBOL(mlx4_sync_pkey_table);
+
+void mlx4_put_slave_node_guid(struct mlx4_dev *dev, int slave, __be64 guid)
+{
+	struct mlx4_priv *priv = container_of(dev, struct mlx4_priv, dev);
+
+	if (!mlx4_is_master(dev))
+		return;
+
+	priv->slave_node_guids[slave] = guid;
+}
+EXPORT_SYMBOL(mlx4_put_slave_node_guid);
+
+__be64 mlx4_get_slave_node_guid(struct mlx4_dev *dev, int slave)
+{
+	struct mlx4_priv *priv = container_of(dev, struct mlx4_priv, dev);
+
+	if (!mlx4_is_master(dev))
+		return 0;
+
+	return priv->slave_node_guids[slave];
+}
+EXPORT_SYMBOL(mlx4_get_slave_node_guid);
+
+int mlx4_is_slave_active(struct mlx4_dev *dev, int slave)
+{
+	struct mlx4_priv *priv = mlx4_priv(dev);
+	struct mlx4_slave_state *s_slave;
+
+	if (!mlx4_is_master(dev))
+		return 0;
+
+	s_slave = &priv->mfunc.master.slave_state[slave];
+	return !!s_slave->active;
+}
+EXPORT_SYMBOL(mlx4_is_slave_active);
+
+void mlx4_handle_eth_header_mcast_prio(struct mlx4_net_trans_rule_hw_ctrl *ctrl,
+				       struct _rule_hw *eth_header)
+{
+	if (is_multicast_ether_addr(eth_header->eth.dst_mac) ||
+	    is_broadcast_ether_addr(eth_header->eth.dst_mac)) {
+		struct mlx4_net_trans_rule_hw_eth *eth =
+			(struct mlx4_net_trans_rule_hw_eth *)eth_header;
+		struct _rule_hw *next_rule = (struct _rule_hw *)(eth + 1);
+		bool last_rule = next_rule->size == 0 && next_rule->id == 0 &&
+			next_rule->rsvd == 0;
+
+		if (last_rule)
+			ctrl->prio = cpu_to_be16(MLX4_DOMAIN_NIC);
+	}
+}
+EXPORT_SYMBOL(mlx4_handle_eth_header_mcast_prio);
+
+static void slave_adjust_steering_mode(struct mlx4_dev *dev,
+				       struct mlx4_dev_cap *dev_cap,
+				       struct mlx4_init_hca_param *hca_param)
+{
+	dev->caps.steering_mode = hca_param->steering_mode;
+	if (dev->caps.steering_mode == MLX4_STEERING_MODE_DEVICE_MANAGED) {
+		dev->caps.num_qp_per_mgm = dev_cap->fs_max_num_qp_per_entry;
+		dev->caps.fs_log_max_ucast_qp_range_size =
+			dev_cap->fs_log_max_ucast_qp_range_size;
+	} else
+		dev->caps.num_qp_per_mgm =
+			4 * ((1 << hca_param->log_mc_entry_sz)/16 - 2);
+
+	mlx4_dbg(dev, "Steering mode is: %s\n",
+		 mlx4_steering_mode_str(dev->caps.steering_mode));
+}
+
+static void mlx4_slave_destroy_special_qp_cap(struct mlx4_dev *dev)
+{
+	kfree(dev->caps.spec_qps);
+	dev->caps.spec_qps = NULL;
+}
+
+static int mlx4_slave_special_qp_cap(struct mlx4_dev *dev)
+{
+	struct mlx4_func_cap *func_cap = NULL;
+	struct mlx4_caps *caps = &dev->caps;
+	int i, err = 0;
+
+	func_cap = kzalloc(sizeof(*func_cap), GFP_KERNEL);
+	caps->spec_qps = kcalloc(caps->num_ports, sizeof(*caps->spec_qps), GFP_KERNEL);
+
+	if (!func_cap || !caps->spec_qps) {
+		mlx4_err(dev, "Failed to allocate memory for special qps cap\n");
+		err = -ENOMEM;
+		goto err_mem;
+	}
+
+	for (i = 1; i <= caps->num_ports; ++i) {
+		err = mlx4_QUERY_FUNC_CAP(dev, i, func_cap);
+		if (err) {
+			mlx4_err(dev, "QUERY_FUNC_CAP port command failed for port %d, aborting (%d)\n",
+				 i, err);
+			goto err_mem;
+		}
+		caps->spec_qps[i - 1] = func_cap->spec_qps;
+		caps->port_mask[i] = caps->port_type[i];
+		caps->phys_port_id[i] = func_cap->phys_port_id;
+		err = mlx4_get_slave_pkey_gid_tbl_len(dev, i,
+						      &caps->gid_table_len[i],
+						      &caps->pkey_table_len[i]);
+		if (err) {
+			mlx4_err(dev, "QUERY_PORT command failed for port %d, aborting (%d)\n",
+				 i, err);
+			goto err_mem;
+		}
+	}
+
+err_mem:
+	if (err)
+		mlx4_slave_destroy_special_qp_cap(dev);
+	kfree(func_cap);
+	return err;
+}
+
+static int mlx4_slave_cap(struct mlx4_dev *dev)
+{
+	int			   err;
+	u32			   page_size;
+	struct mlx4_dev_cap	   *dev_cap = NULL;
+	struct mlx4_func_cap	   *func_cap = NULL;
+	struct mlx4_init_hca_param *hca_param = NULL;
+
+	hca_param = kzalloc(sizeof(*hca_param), GFP_KERNEL);
+	func_cap = kzalloc(sizeof(*func_cap), GFP_KERNEL);
+	dev_cap = kzalloc(sizeof(*dev_cap), GFP_KERNEL);
+	if (!hca_param || !func_cap || !dev_cap) {
+		mlx4_err(dev, "Failed to allocate memory for slave_cap\n");
+		err = -ENOMEM;
+		goto free_mem;
+	}
+
+	err = mlx4_QUERY_HCA(dev, hca_param);
+	if (err) {
+		mlx4_err(dev, "QUERY_HCA command failed, aborting\n");
+		goto free_mem;
+	}
+
+	/* fail if the hca has an unknown global capability
+	 * at this time global_caps should be always zeroed
+	 */
+	if (hca_param->global_caps) {
+		mlx4_err(dev, "Unknown hca global capabilities\n");
+		err = -EINVAL;
+		goto free_mem;
+	}
+
+	dev->caps.hca_core_clock = hca_param->hca_core_clock;
+
+	dev->caps.max_qp_dest_rdma = 1 << hca_param->log_rd_per_qp;
+	err = mlx4_dev_cap(dev, dev_cap);
+	if (err) {
+		mlx4_err(dev, "QUERY_DEV_CAP command failed, aborting\n");
+		goto free_mem;
+	}
+
+	err = mlx4_QUERY_FW(dev);
+	if (err)
+		mlx4_err(dev, "QUERY_FW command failed: could not get FW version\n");
+
+	page_size = ~dev->caps.page_size_cap + 1;
+	mlx4_warn(dev, "HCA minimum page size:%d\n", page_size);
+	if (page_size > PAGE_SIZE) {
+		mlx4_err(dev, "HCA minimum page size of %d bigger than kernel PAGE_SIZE of %ld, aborting\n",
+			 page_size, PAGE_SIZE);
+		err = -ENODEV;
+		goto free_mem;
+	}
+
+	/* Set uar_page_shift for VF */
+	dev->uar_page_shift = hca_param->uar_page_sz + 12;
+
+	/* Make sure the master uar page size is valid */
+	if (dev->uar_page_shift > PAGE_SHIFT) {
+		mlx4_err(dev,
+			 "Invalid configuration: uar page size is larger than system page size\n");
+		err = -ENODEV;
+		goto free_mem;
+	}
+
+	/* Set reserved_uars based on the uar_page_shift */
+	mlx4_set_num_reserved_uars(dev, dev_cap);
+
+	/* Although uar page size in FW differs from system page size,
+	 * upper software layers (mlx4_ib, mlx4_en and part of mlx4_core)
+	 * still works with assumption that uar page size == system page size
+	 */
+	dev->caps.uar_page_size = PAGE_SIZE;
+
+	err = mlx4_QUERY_FUNC_CAP(dev, 0, func_cap);
+	if (err) {
+		mlx4_err(dev, "QUERY_FUNC_CAP general command failed, aborting (%d)\n",
+			 err);
+		goto free_mem;
+	}
+
+	if ((func_cap->pf_context_behaviour | PF_CONTEXT_BEHAVIOUR_MASK) !=
+	    PF_CONTEXT_BEHAVIOUR_MASK) {
+		mlx4_err(dev, "Unknown pf context behaviour %x known flags %x\n",
+			 func_cap->pf_context_behaviour,
+			 PF_CONTEXT_BEHAVIOUR_MASK);
+		err = -EINVAL;
+		goto free_mem;
+	}
+
+	dev->caps.num_ports		= func_cap->num_ports;
+	dev->quotas.qp			= func_cap->qp_quota;
+	dev->quotas.srq			= func_cap->srq_quota;
+	dev->quotas.cq			= func_cap->cq_quota;
+	dev->quotas.mpt			= func_cap->mpt_quota;
+	dev->quotas.mtt			= func_cap->mtt_quota;
+	dev->caps.num_qps		= 1 << hca_param->log_num_qps;
+	dev->caps.num_srqs		= 1 << hca_param->log_num_srqs;
+	dev->caps.num_cqs		= 1 << hca_param->log_num_cqs;
+	dev->caps.num_mpts		= 1 << hca_param->log_mpt_sz;
+	dev->caps.num_eqs		= func_cap->max_eq;
+	dev->caps.reserved_eqs		= func_cap->reserved_eq;
+	dev->caps.reserved_lkey		= func_cap->reserved_lkey;
+	dev->caps.num_pds               = MLX4_NUM_PDS;
+	dev->caps.num_mgms              = 0;
+	dev->caps.num_amgms             = 0;
+
+	if (dev->caps.num_ports > MLX4_MAX_PORTS) {
+		mlx4_err(dev, "HCA has %d ports, but we only support %d, aborting\n",
+			 dev->caps.num_ports, MLX4_MAX_PORTS);
+		err = -ENODEV;
+		goto free_mem;
+	}
+
+	mlx4_replace_zero_macs(dev);
+
+	err = mlx4_slave_special_qp_cap(dev);
+	if (err) {
+		mlx4_err(dev, "Set special QP caps failed. aborting\n");
+		goto free_mem;
+	}
+
+	if (dev->caps.uar_page_size * (dev->caps.num_uars -
+				       dev->caps.reserved_uars) >
+				       pci_resource_len(dev->persist->pdev,
+							2)) {
+		mlx4_err(dev, "HCA reported UAR region size of 0x%x bigger than PCI resource 2 size of 0x%llx, aborting\n",
+			 dev->caps.uar_page_size * dev->caps.num_uars,
+			 (unsigned long long)
+			 pci_resource_len(dev->persist->pdev, 2));
+		err = -ENOMEM;
+		goto err_mem;
+	}
+
+	if (hca_param->dev_cap_enabled & MLX4_DEV_CAP_64B_EQE_ENABLED) {
+		dev->caps.eqe_size   = 64;
+		dev->caps.eqe_factor = 1;
+	} else {
+		dev->caps.eqe_size   = 32;
+		dev->caps.eqe_factor = 0;
+	}
+
+	if (hca_param->dev_cap_enabled & MLX4_DEV_CAP_64B_CQE_ENABLED) {
+		dev->caps.cqe_size   = 64;
+		dev->caps.userspace_caps |= MLX4_USER_DEV_CAP_LARGE_CQE;
+	} else {
+		dev->caps.cqe_size   = 32;
+	}
+
+	if (hca_param->dev_cap_enabled & MLX4_DEV_CAP_EQE_STRIDE_ENABLED) {
+		dev->caps.eqe_size = hca_param->eqe_size;
+		dev->caps.eqe_factor = 0;
+	}
+
+	if (hca_param->dev_cap_enabled & MLX4_DEV_CAP_CQE_STRIDE_ENABLED) {
+		dev->caps.cqe_size = hca_param->cqe_size;
+		/* User still need to know when CQE > 32B */
+		dev->caps.userspace_caps |= MLX4_USER_DEV_CAP_LARGE_CQE;
+	}
+
+	dev->caps.flags2 &= ~MLX4_DEV_CAP_FLAG2_TS;
+	mlx4_warn(dev, "Timestamping is not supported in slave mode\n");
+
+	dev->caps.flags2 &= ~MLX4_DEV_CAP_FLAG2_USER_MAC_EN;
+	mlx4_dbg(dev, "User MAC FW update is not supported in slave mode\n");
+
+	slave_adjust_steering_mode(dev, dev_cap, hca_param);
+	mlx4_dbg(dev, "RSS support for IP fragments is %s\n",
+		 hca_param->rss_ip_frags ? "on" : "off");
+
+	if (func_cap->extra_flags & MLX4_QUERY_FUNC_FLAGS_BF_RES_QP &&
+	    dev->caps.bf_reg_size)
+		dev->caps.alloc_res_qp_mask |= MLX4_RESERVE_ETH_BF_QP;
+
+	if (func_cap->extra_flags & MLX4_QUERY_FUNC_FLAGS_A0_RES_QP)
+		dev->caps.alloc_res_qp_mask |= MLX4_RESERVE_A0_QP;
+
+err_mem:
+	if (err)
+		mlx4_slave_destroy_special_qp_cap(dev);
+free_mem:
+	kfree(hca_param);
+	kfree(func_cap);
+	kfree(dev_cap);
+	return err;
+}
+
+static void mlx4_request_modules(struct mlx4_dev *dev)
+{
+	int port;
+	int has_ib_port = false;
+	int has_eth_port = false;
+#define EN_DRV_NAME	"mlx4_en"
+#define IB_DRV_NAME	"mlx4_ib"
+
+	for (port = 1; port <= dev->caps.num_ports; port++) {
+		if (dev->caps.port_type[port] == MLX4_PORT_TYPE_IB)
+			has_ib_port = true;
+		else if (dev->caps.port_type[port] == MLX4_PORT_TYPE_ETH)
+			has_eth_port = true;
+	}
+
+	if (has_eth_port)
+		request_module_nowait(EN_DRV_NAME);
+	if (has_ib_port || (dev->caps.flags & MLX4_DEV_CAP_FLAG_IBOE))
+		request_module_nowait(IB_DRV_NAME);
+}
+
+/*
+ * Change the port configuration of the device.
+ * Every user of this function must hold the port mutex.
+ */
+int mlx4_change_port_types(struct mlx4_dev *dev,
+			   enum mlx4_port_type *port_types)
+{
+	int err = 0;
+	int change = 0;
+	int port;
+
+	for (port = 0; port <  dev->caps.num_ports; port++) {
+		/* Change the port type only if the new type is different
+		 * from the current, and not set to Auto */
+		if (port_types[port] != dev->caps.port_type[port + 1])
+			change = 1;
+	}
+	if (change) {
+		mlx4_unregister_device(dev);
+		for (port = 1; port <= dev->caps.num_ports; port++) {
+			mlx4_CLOSE_PORT(dev, port);
+			dev->caps.port_type[port] = port_types[port - 1];
+			err = mlx4_SET_PORT(dev, port, -1);
+			if (err) {
+				mlx4_err(dev, "Failed to set port %d, aborting\n",
+					 port);
+				goto out;
+			}
+		}
+		mlx4_set_port_mask(dev);
+		err = mlx4_register_device(dev);
+		if (err) {
+			mlx4_err(dev, "Failed to register device\n");
+			goto out;
+		}
+		mlx4_request_modules(dev);
+	}
+
+out:
+	return err;
+}
+
+static ssize_t show_port_type(struct device *dev,
+			      struct device_attribute *attr,
+			      char *buf)
+{
+	struct mlx4_port_info *info = container_of(attr, struct mlx4_port_info,
+						   port_attr);
+	struct mlx4_dev *mdev = info->dev;
+	char type[8];
+
+	sprintf(type, "%s",
+		(mdev->caps.port_type[info->port] == MLX4_PORT_TYPE_IB) ?
+		"ib" : "eth");
+	if (mdev->caps.possible_type[info->port] == MLX4_PORT_TYPE_AUTO)
+		sprintf(buf, "auto (%s)\n", type);
+	else
+		sprintf(buf, "%s\n", type);
+
+	return strlen(buf);
+}
+
+static int __set_port_type(struct mlx4_port_info *info,
+			   enum mlx4_port_type port_type)
+{
+	struct mlx4_dev *mdev = info->dev;
+	struct mlx4_priv *priv = mlx4_priv(mdev);
+	enum mlx4_port_type types[MLX4_MAX_PORTS];
+	enum mlx4_port_type new_types[MLX4_MAX_PORTS];
+	int i;
+	int err = 0;
+
+	if ((port_type & mdev->caps.supported_type[info->port]) != port_type) {
+		mlx4_err(mdev,
+			 "Requested port type for port %d is not supported on this HCA\n",
+			 info->port);
+		return -EOPNOTSUPP;
+	}
+
+	mlx4_stop_sense(mdev);
+	mutex_lock(&priv->port_mutex);
+	info->tmp_type = port_type;
+
+	/* Possible type is always the one that was delivered */
+	mdev->caps.possible_type[info->port] = info->tmp_type;
+
+	for (i = 0; i < mdev->caps.num_ports; i++) {
+		types[i] = priv->port[i+1].tmp_type ? priv->port[i+1].tmp_type :
+					mdev->caps.possible_type[i+1];
+		if (types[i] == MLX4_PORT_TYPE_AUTO)
+			types[i] = mdev->caps.port_type[i+1];
+	}
+
+	if (!(mdev->caps.flags & MLX4_DEV_CAP_FLAG_DPDP) &&
+	    !(mdev->caps.flags & MLX4_DEV_CAP_FLAG_SENSE_SUPPORT)) {
+		for (i = 1; i <= mdev->caps.num_ports; i++) {
+			if (mdev->caps.possible_type[i] == MLX4_PORT_TYPE_AUTO) {
+				mdev->caps.possible_type[i] = mdev->caps.port_type[i];
+				err = -EOPNOTSUPP;
+			}
+		}
+	}
+	if (err) {
+		mlx4_err(mdev, "Auto sensing is not supported on this HCA. Set only 'eth' or 'ib' for both ports (should be the same)\n");
+		goto out;
+	}
+
+	mlx4_do_sense_ports(mdev, new_types, types);
+
+	err = mlx4_check_port_params(mdev, new_types);
+	if (err)
+		goto out;
+
+	/* We are about to apply the changes after the configuration
+	 * was verified, no need to remember the temporary types
+	 * any more */
+	for (i = 0; i < mdev->caps.num_ports; i++)
+		priv->port[i + 1].tmp_type = 0;
+
+	err = mlx4_change_port_types(mdev, new_types);
+
+out:
+	mlx4_start_sense(mdev);
+	mutex_unlock(&priv->port_mutex);
+
+	return err;
+}
+
+static ssize_t set_port_type(struct device *dev,
+			     struct device_attribute *attr,
+			     const char *buf, size_t count)
+{
+	struct mlx4_port_info *info = container_of(attr, struct mlx4_port_info,
+						   port_attr);
+	struct mlx4_dev *mdev = info->dev;
+	enum mlx4_port_type port_type;
+	static DEFINE_MUTEX(set_port_type_mutex);
+	int err;
+
+	mutex_lock(&set_port_type_mutex);
+
+	if (!strcmp(buf, "ib\n")) {
+		port_type = MLX4_PORT_TYPE_IB;
+	} else if (!strcmp(buf, "eth\n")) {
+		port_type = MLX4_PORT_TYPE_ETH;
+	} else if (!strcmp(buf, "auto\n")) {
+		port_type = MLX4_PORT_TYPE_AUTO;
+	} else {
+		mlx4_err(mdev, "%s is not supported port type\n", buf);
+		err = -EINVAL;
+		goto err_out;
+	}
+
+	err = __set_port_type(info, port_type);
+
+err_out:
+	mutex_unlock(&set_port_type_mutex);
+
+	return err ? err : count;
+}
+
+enum ibta_mtu {
+	IB_MTU_256  = 1,
+	IB_MTU_512  = 2,
+	IB_MTU_1024 = 3,
+	IB_MTU_2048 = 4,
+	IB_MTU_4096 = 5
+};
+
+static inline int int_to_ibta_mtu(int mtu)
+{
+	switch (mtu) {
+	case 256:  return IB_MTU_256;
+	case 512:  return IB_MTU_512;
+	case 1024: return IB_MTU_1024;
+	case 2048: return IB_MTU_2048;
+	case 4096: return IB_MTU_4096;
+	default: return -1;
+	}
+}
+
+static inline int ibta_mtu_to_int(enum ibta_mtu mtu)
+{
+	switch (mtu) {
+	case IB_MTU_256:  return  256;
+	case IB_MTU_512:  return  512;
+	case IB_MTU_1024: return 1024;
+	case IB_MTU_2048: return 2048;
+	case IB_MTU_4096: return 4096;
+	default: return -1;
+	}
+}
+
+static ssize_t show_port_ib_mtu(struct device *dev,
+			     struct device_attribute *attr,
+			     char *buf)
+{
+	struct mlx4_port_info *info = container_of(attr, struct mlx4_port_info,
+						   port_mtu_attr);
+	struct mlx4_dev *mdev = info->dev;
+
+	if (mdev->caps.port_type[info->port] == MLX4_PORT_TYPE_ETH)
+		mlx4_warn(mdev, "port level mtu is only used for IB ports\n");
+
+	sprintf(buf, "%d\n",
+			ibta_mtu_to_int(mdev->caps.port_ib_mtu[info->port]));
+	return strlen(buf);
+}
+
+static ssize_t set_port_ib_mtu(struct device *dev,
+			     struct device_attribute *attr,
+			     const char *buf, size_t count)
+{
+	struct mlx4_port_info *info = container_of(attr, struct mlx4_port_info,
+						   port_mtu_attr);
+	struct mlx4_dev *mdev = info->dev;
+	struct mlx4_priv *priv = mlx4_priv(mdev);
+	int err, port, mtu, ibta_mtu = -1;
+
+	if (mdev->caps.port_type[info->port] == MLX4_PORT_TYPE_ETH) {
+		mlx4_warn(mdev, "port level mtu is only used for IB ports\n");
+		return -EINVAL;
+	}
+
+	err = kstrtoint(buf, 0, &mtu);
+	if (!err)
+		ibta_mtu = int_to_ibta_mtu(mtu);
+
+	if (err || ibta_mtu < 0) {
+		mlx4_err(mdev, "%s is invalid IBTA mtu\n", buf);
+		return -EINVAL;
+	}
+
+	mdev->caps.port_ib_mtu[info->port] = ibta_mtu;
+
+	mlx4_stop_sense(mdev);
+	mutex_lock(&priv->port_mutex);
+	mlx4_unregister_device(mdev);
+	for (port = 1; port <= mdev->caps.num_ports; port++) {
+		mlx4_CLOSE_PORT(mdev, port);
+		err = mlx4_SET_PORT(mdev, port, -1);
+		if (err) {
+			mlx4_err(mdev, "Failed to set port %d, aborting\n",
+				 port);
+			goto err_set_port;
+		}
+	}
+	err = mlx4_register_device(mdev);
+err_set_port:
+	mutex_unlock(&priv->port_mutex);
+	mlx4_start_sense(mdev);
+	return err ? err : count;
+}
+
+/* bond for multi-function device */
+#define MAX_MF_BOND_ALLOWED_SLAVES 63
+static int mlx4_mf_bond(struct mlx4_dev *dev)
+{
+	int err = 0;
+	int nvfs;
+	struct mlx4_slaves_pport slaves_port1;
+	struct mlx4_slaves_pport slaves_port2;
+	DECLARE_BITMAP(slaves_port_1_2, MLX4_MFUNC_MAX);
+
+	slaves_port1 = mlx4_phys_to_slaves_pport(dev, 1);
+	slaves_port2 = mlx4_phys_to_slaves_pport(dev, 2);
+	bitmap_and(slaves_port_1_2,
+		   slaves_port1.slaves, slaves_port2.slaves,
+		   dev->persist->num_vfs + 1);
+
+	/* only single port vfs are allowed */
+	if (bitmap_weight(slaves_port_1_2, dev->persist->num_vfs + 1) > 1) {
+		mlx4_warn(dev, "HA mode unsupported for dual ported VFs\n");
+		return -EINVAL;
+	}
+
+	/* number of virtual functions is number of total functions minus one
+	 * physical function for each port.
+	 */
+	nvfs = bitmap_weight(slaves_port1.slaves, dev->persist->num_vfs + 1) +
+		bitmap_weight(slaves_port2.slaves, dev->persist->num_vfs + 1) - 2;
+
+	/* limit on maximum allowed VFs */
+	if (nvfs > MAX_MF_BOND_ALLOWED_SLAVES) {
+		mlx4_warn(dev, "HA mode is not supported for %d VFs (max %d are allowed)\n",
+			  nvfs, MAX_MF_BOND_ALLOWED_SLAVES);
+		return -EINVAL;
+	}
+
+	if (dev->caps.steering_mode != MLX4_STEERING_MODE_DEVICE_MANAGED) {
+		mlx4_warn(dev, "HA mode unsupported for NON DMFS steering\n");
+		return -EINVAL;
+	}
+
+	err = mlx4_bond_mac_table(dev);
+	if (err)
+		return err;
+	err = mlx4_bond_vlan_table(dev);
+	if (err)
+		goto err1;
+	err = mlx4_bond_fs_rules(dev);
+	if (err)
+		goto err2;
+
+	return 0;
+err2:
+	(void)mlx4_unbond_vlan_table(dev);
+err1:
+	(void)mlx4_unbond_mac_table(dev);
+	return err;
+}
+
+static int mlx4_mf_unbond(struct mlx4_dev *dev)
+{
+	int ret, ret1;
+
+	ret = mlx4_unbond_fs_rules(dev);
+	if (ret)
+		mlx4_warn(dev, "multifunction unbond for flow rules failed (%d)\n", ret);
+	ret1 = mlx4_unbond_mac_table(dev);
+	if (ret1) {
+		mlx4_warn(dev, "multifunction unbond for MAC table failed (%d)\n", ret1);
+		ret = ret1;
+	}
+	ret1 = mlx4_unbond_vlan_table(dev);
+	if (ret1) {
+		mlx4_warn(dev, "multifunction unbond for VLAN table failed (%d)\n", ret1);
+		ret = ret1;
+	}
+	return ret;
+}
+
+int mlx4_bond(struct mlx4_dev *dev)
+{
+	int ret = 0;
+	struct mlx4_priv *priv = mlx4_priv(dev);
+
+	mutex_lock(&priv->bond_mutex);
+
+	if (!mlx4_is_bonded(dev)) {
+		ret = mlx4_do_bond(dev, true);
+		if (ret)
+			mlx4_err(dev, "Failed to bond device: %d\n", ret);
+		if (!ret && mlx4_is_master(dev)) {
+			ret = mlx4_mf_bond(dev);
+			if (ret) {
+				mlx4_err(dev, "bond for multifunction failed\n");
+				mlx4_do_bond(dev, false);
+			}
+		}
+	}
+
+	mutex_unlock(&priv->bond_mutex);
+	if (!ret)
+		mlx4_dbg(dev, "Device is bonded\n");
+
+	return ret;
+}
+EXPORT_SYMBOL_GPL(mlx4_bond);
+
+int mlx4_unbond(struct mlx4_dev *dev)
+{
+	int ret = 0;
+	struct mlx4_priv *priv = mlx4_priv(dev);
+
+	mutex_lock(&priv->bond_mutex);
+
+	if (mlx4_is_bonded(dev)) {
+		int ret2 = 0;
+
+		ret = mlx4_do_bond(dev, false);
+		if (ret)
+			mlx4_err(dev, "Failed to unbond device: %d\n", ret);
+		if (mlx4_is_master(dev))
+			ret2 = mlx4_mf_unbond(dev);
+		if (ret2) {
+			mlx4_warn(dev, "Failed to unbond device for multifunction (%d)\n", ret2);
+			ret = ret2;
+		}
+	}
+
+	mutex_unlock(&priv->bond_mutex);
+	if (!ret)
+		mlx4_dbg(dev, "Device is unbonded\n");
+
+	return ret;
+}
+EXPORT_SYMBOL_GPL(mlx4_unbond);
+
+
+int mlx4_port_map_set(struct mlx4_dev *dev, struct mlx4_port_map *v2p)
+{
+	u8 port1 = v2p->port1;
+	u8 port2 = v2p->port2;
+	struct mlx4_priv *priv = mlx4_priv(dev);
+	int err;
+
+	if (!(dev->caps.flags2 & MLX4_DEV_CAP_FLAG2_PORT_REMAP))
+		return -EOPNOTSUPP;
+
+	mutex_lock(&priv->bond_mutex);
+
+	/* zero means keep current mapping for this port */
+	if (port1 == 0)
+		port1 = priv->v2p.port1;
+	if (port2 == 0)
+		port2 = priv->v2p.port2;
+
+	if ((port1 < 1) || (port1 > MLX4_MAX_PORTS) ||
+	    (port2 < 1) || (port2 > MLX4_MAX_PORTS) ||
+	    (port1 == 2 && port2 == 1)) {
+		/* besides boundary checks cross mapping makes
+		 * no sense and therefore not allowed */
+		err = -EINVAL;
+	} else if ((port1 == priv->v2p.port1) &&
+		 (port2 == priv->v2p.port2)) {
+		err = 0;
+	} else {
+		err = mlx4_virt2phy_port_map(dev, port1, port2);
+		if (!err) {
+			mlx4_dbg(dev, "port map changed: [%d][%d]\n",
+				 port1, port2);
+			priv->v2p.port1 = port1;
+			priv->v2p.port2 = port2;
+		} else {
+			mlx4_err(dev, "Failed to change port mape: %d\n", err);
+		}
+	}
+
+	mutex_unlock(&priv->bond_mutex);
+	return err;
+}
+EXPORT_SYMBOL_GPL(mlx4_port_map_set);
+
+static int mlx4_load_fw(struct mlx4_dev *dev)
+{
+	struct mlx4_priv *priv = mlx4_priv(dev);
+	int err;
+
+	priv->fw.fw_icm = mlx4_alloc_icm(dev, priv->fw.fw_pages,
+					 GFP_HIGHUSER | __GFP_NOWARN, 0);
+	if (!priv->fw.fw_icm) {
+		mlx4_err(dev, "Couldn't allocate FW area, aborting\n");
+		return -ENOMEM;
+	}
+
+	err = mlx4_MAP_FA(dev, priv->fw.fw_icm);
+	if (err) {
+		mlx4_err(dev, "MAP_FA command failed, aborting\n");
+		goto err_free;
+	}
+
+	err = mlx4_RUN_FW(dev);
+	if (err) {
+		mlx4_err(dev, "RUN_FW command failed, aborting\n");
+		goto err_unmap_fa;
+	}
+
+	return 0;
+
+err_unmap_fa:
+	mlx4_UNMAP_FA(dev);
+
+err_free:
+	mlx4_free_icm(dev, priv->fw.fw_icm, 0);
+	return err;
+}
+
+static int mlx4_init_cmpt_table(struct mlx4_dev *dev, u64 cmpt_base,
+				int cmpt_entry_sz)
+{
+	struct mlx4_priv *priv = mlx4_priv(dev);
+	int err;
+	int num_eqs;
+
+	err = mlx4_init_icm_table(dev, &priv->qp_table.cmpt_table,
+				  cmpt_base +
+				  ((u64) (MLX4_CMPT_TYPE_QP *
+					  cmpt_entry_sz) << MLX4_CMPT_SHIFT),
+				  cmpt_entry_sz, dev->caps.num_qps,
+				  dev->caps.reserved_qps_cnt[MLX4_QP_REGION_FW],
+				  0, 0);
+	if (err)
+		goto err;
+
+	err = mlx4_init_icm_table(dev, &priv->srq_table.cmpt_table,
+				  cmpt_base +
+				  ((u64) (MLX4_CMPT_TYPE_SRQ *
+					  cmpt_entry_sz) << MLX4_CMPT_SHIFT),
+				  cmpt_entry_sz, dev->caps.num_srqs,
+				  dev->caps.reserved_srqs, 0, 0);
+	if (err)
+		goto err_qp;
+
+	err = mlx4_init_icm_table(dev, &priv->cq_table.cmpt_table,
+				  cmpt_base +
+				  ((u64) (MLX4_CMPT_TYPE_CQ *
+					  cmpt_entry_sz) << MLX4_CMPT_SHIFT),
+				  cmpt_entry_sz, dev->caps.num_cqs,
+				  dev->caps.reserved_cqs, 0, 0);
+	if (err)
+		goto err_srq;
+
+	num_eqs = dev->phys_caps.num_phys_eqs;
+	err = mlx4_init_icm_table(dev, &priv->eq_table.cmpt_table,
+				  cmpt_base +
+				  ((u64) (MLX4_CMPT_TYPE_EQ *
+					  cmpt_entry_sz) << MLX4_CMPT_SHIFT),
+				  cmpt_entry_sz, num_eqs, num_eqs, 0, 0);
+	if (err)
+		goto err_cq;
+
+	return 0;
+
+err_cq:
+	mlx4_cleanup_icm_table(dev, &priv->cq_table.cmpt_table);
+
+err_srq:
+	mlx4_cleanup_icm_table(dev, &priv->srq_table.cmpt_table);
+
+err_qp:
+	mlx4_cleanup_icm_table(dev, &priv->qp_table.cmpt_table);
+
+err:
+	return err;
+}
+
+static int mlx4_init_icm(struct mlx4_dev *dev, struct mlx4_dev_cap *dev_cap,
+			 struct mlx4_init_hca_param *init_hca, u64 icm_size)
+{
+	struct mlx4_priv *priv = mlx4_priv(dev);
+	u64 aux_pages;
+	int num_eqs;
+	int err;
+
+	err = mlx4_SET_ICM_SIZE(dev, icm_size, &aux_pages);
+	if (err) {
+		mlx4_err(dev, "SET_ICM_SIZE command failed, aborting\n");
+		return err;
+	}
+
+	mlx4_dbg(dev, "%lld KB of HCA context requires %lld KB aux memory\n",
+		 (unsigned long long) icm_size >> 10,
+		 (unsigned long long) aux_pages << 2);
+
+	priv->fw.aux_icm = mlx4_alloc_icm(dev, aux_pages,
+					  GFP_HIGHUSER | __GFP_NOWARN, 0);
+	if (!priv->fw.aux_icm) {
+		mlx4_err(dev, "Couldn't allocate aux memory, aborting\n");
+		return -ENOMEM;
+	}
+
+	err = mlx4_MAP_ICM_AUX(dev, priv->fw.aux_icm);
+	if (err) {
+		mlx4_err(dev, "MAP_ICM_AUX command failed, aborting\n");
+		goto err_free_aux;
+	}
+
+	err = mlx4_init_cmpt_table(dev, init_hca->cmpt_base, dev_cap->cmpt_entry_sz);
+	if (err) {
+		mlx4_err(dev, "Failed to map cMPT context memory, aborting\n");
+		goto err_unmap_aux;
+	}
+
+
+	num_eqs = dev->phys_caps.num_phys_eqs;
+	err = mlx4_init_icm_table(dev, &priv->eq_table.table,
+				  init_hca->eqc_base, dev_cap->eqc_entry_sz,
+				  num_eqs, num_eqs, 0, 0);
+	if (err) {
+		mlx4_err(dev, "Failed to map EQ context memory, aborting\n");
+		goto err_unmap_cmpt;
+	}
+
+	/*
+	 * Reserved MTT entries must be aligned up to a cacheline
+	 * boundary, since the FW will write to them, while the driver
+	 * writes to all other MTT entries. (The variable
+	 * dev->caps.mtt_entry_sz below is really the MTT segment
+	 * size, not the raw entry size)
+	 */
+	dev->caps.reserved_mtts =
+		ALIGN(dev->caps.reserved_mtts * dev->caps.mtt_entry_sz,
+		      dma_get_cache_alignment()) / dev->caps.mtt_entry_sz;
+
+	err = mlx4_init_icm_table(dev, &priv->mr_table.mtt_table,
+				  init_hca->mtt_base,
+				  dev->caps.mtt_entry_sz,
+				  dev->caps.num_mtts,
+				  dev->caps.reserved_mtts, 1, 0);
+	if (err) {
+		mlx4_err(dev, "Failed to map MTT context memory, aborting\n");
+		goto err_unmap_eq;
+	}
+
+	err = mlx4_init_icm_table(dev, &priv->mr_table.dmpt_table,
+				  init_hca->dmpt_base,
+				  dev_cap->dmpt_entry_sz,
+				  dev->caps.num_mpts,
+				  dev->caps.reserved_mrws, 1, 1);
+	if (err) {
+		mlx4_err(dev, "Failed to map dMPT context memory, aborting\n");
+		goto err_unmap_mtt;
+	}
+
+	err = mlx4_init_icm_table(dev, &priv->qp_table.qp_table,
+				  init_hca->qpc_base,
+				  dev_cap->qpc_entry_sz,
+				  dev->caps.num_qps,
+				  dev->caps.reserved_qps_cnt[MLX4_QP_REGION_FW],
+				  0, 0);
+	if (err) {
+		mlx4_err(dev, "Failed to map QP context memory, aborting\n");
+		goto err_unmap_dmpt;
+	}
+
+	err = mlx4_init_icm_table(dev, &priv->qp_table.auxc_table,
+				  init_hca->auxc_base,
+				  dev_cap->aux_entry_sz,
+				  dev->caps.num_qps,
+				  dev->caps.reserved_qps_cnt[MLX4_QP_REGION_FW],
+				  0, 0);
+	if (err) {
+		mlx4_err(dev, "Failed to map AUXC context memory, aborting\n");
+		goto err_unmap_qp;
+	}
+
+	err = mlx4_init_icm_table(dev, &priv->qp_table.altc_table,
+				  init_hca->altc_base,
+				  dev_cap->altc_entry_sz,
+				  dev->caps.num_qps,
+				  dev->caps.reserved_qps_cnt[MLX4_QP_REGION_FW],
+				  0, 0);
+	if (err) {
+		mlx4_err(dev, "Failed to map ALTC context memory, aborting\n");
+		goto err_unmap_auxc;
+	}
+
+	err = mlx4_init_icm_table(dev, &priv->qp_table.rdmarc_table,
+				  init_hca->rdmarc_base,
+				  dev_cap->rdmarc_entry_sz << priv->qp_table.rdmarc_shift,
+				  dev->caps.num_qps,
+				  dev->caps.reserved_qps_cnt[MLX4_QP_REGION_FW],
+				  0, 0);
+	if (err) {
+		mlx4_err(dev, "Failed to map RDMARC context memory, aborting\n");
+		goto err_unmap_altc;
+	}
+
+	err = mlx4_init_icm_table(dev, &priv->cq_table.table,
+				  init_hca->cqc_base,
+				  dev_cap->cqc_entry_sz,
+				  dev->caps.num_cqs,
+				  dev->caps.reserved_cqs, 0, 0);
+	if (err) {
+		mlx4_err(dev, "Failed to map CQ context memory, aborting\n");
+		goto err_unmap_rdmarc;
+	}
+
+	err = mlx4_init_icm_table(dev, &priv->srq_table.table,
+				  init_hca->srqc_base,
+				  dev_cap->srq_entry_sz,
+				  dev->caps.num_srqs,
+				  dev->caps.reserved_srqs, 0, 0);
+	if (err) {
+		mlx4_err(dev, "Failed to map SRQ context memory, aborting\n");
+		goto err_unmap_cq;
+	}
+
+	/*
+	 * For flow steering device managed mode it is required to use
+	 * mlx4_init_icm_table. For B0 steering mode it's not strictly
+	 * required, but for simplicity just map the whole multicast
+	 * group table now.  The table isn't very big and it's a lot
+	 * easier than trying to track ref counts.
+	 */
+	err = mlx4_init_icm_table(dev, &priv->mcg_table.table,
+				  init_hca->mc_base,
+				  mlx4_get_mgm_entry_size(dev),
+				  dev->caps.num_mgms + dev->caps.num_amgms,
+				  dev->caps.num_mgms + dev->caps.num_amgms,
+				  0, 0);
+	if (err) {
+		mlx4_err(dev, "Failed to map MCG context memory, aborting\n");
+		goto err_unmap_srq;
+	}
+
+	return 0;
+
+err_unmap_srq:
+	mlx4_cleanup_icm_table(dev, &priv->srq_table.table);
+
+err_unmap_cq:
+	mlx4_cleanup_icm_table(dev, &priv->cq_table.table);
+
+err_unmap_rdmarc:
+	mlx4_cleanup_icm_table(dev, &priv->qp_table.rdmarc_table);
+
+err_unmap_altc:
+	mlx4_cleanup_icm_table(dev, &priv->qp_table.altc_table);
+
+err_unmap_auxc:
+	mlx4_cleanup_icm_table(dev, &priv->qp_table.auxc_table);
+
+err_unmap_qp:
+	mlx4_cleanup_icm_table(dev, &priv->qp_table.qp_table);
+
+err_unmap_dmpt:
+	mlx4_cleanup_icm_table(dev, &priv->mr_table.dmpt_table);
+
+err_unmap_mtt:
+	mlx4_cleanup_icm_table(dev, &priv->mr_table.mtt_table);
+
+err_unmap_eq:
+	mlx4_cleanup_icm_table(dev, &priv->eq_table.table);
+
+err_unmap_cmpt:
+	mlx4_cleanup_icm_table(dev, &priv->eq_table.cmpt_table);
+	mlx4_cleanup_icm_table(dev, &priv->cq_table.cmpt_table);
+	mlx4_cleanup_icm_table(dev, &priv->srq_table.cmpt_table);
+	mlx4_cleanup_icm_table(dev, &priv->qp_table.cmpt_table);
+
+err_unmap_aux:
+	mlx4_UNMAP_ICM_AUX(dev);
+
+err_free_aux:
+	mlx4_free_icm(dev, priv->fw.aux_icm, 0);
+
+	return err;
+}
+
+static void mlx4_free_icms(struct mlx4_dev *dev)
+{
+	struct mlx4_priv *priv = mlx4_priv(dev);
+
+	mlx4_cleanup_icm_table(dev, &priv->mcg_table.table);
+	mlx4_cleanup_icm_table(dev, &priv->srq_table.table);
+	mlx4_cleanup_icm_table(dev, &priv->cq_table.table);
+	mlx4_cleanup_icm_table(dev, &priv->qp_table.rdmarc_table);
+	mlx4_cleanup_icm_table(dev, &priv->qp_table.altc_table);
+	mlx4_cleanup_icm_table(dev, &priv->qp_table.auxc_table);
+	mlx4_cleanup_icm_table(dev, &priv->qp_table.qp_table);
+	mlx4_cleanup_icm_table(dev, &priv->mr_table.dmpt_table);
+	mlx4_cleanup_icm_table(dev, &priv->mr_table.mtt_table);
+	mlx4_cleanup_icm_table(dev, &priv->eq_table.table);
+	mlx4_cleanup_icm_table(dev, &priv->eq_table.cmpt_table);
+	mlx4_cleanup_icm_table(dev, &priv->cq_table.cmpt_table);
+	mlx4_cleanup_icm_table(dev, &priv->srq_table.cmpt_table);
+	mlx4_cleanup_icm_table(dev, &priv->qp_table.cmpt_table);
+
+	mlx4_UNMAP_ICM_AUX(dev);
+	mlx4_free_icm(dev, priv->fw.aux_icm, 0);
+}
+
+static void mlx4_slave_exit(struct mlx4_dev *dev)
+{
+	struct mlx4_priv *priv = mlx4_priv(dev);
+
+	mutex_lock(&priv->cmd.slave_cmd_mutex);
+	if (mlx4_comm_cmd(dev, MLX4_COMM_CMD_RESET, 0, MLX4_COMM_CMD_NA_OP,
+			  MLX4_COMM_TIME))
+		mlx4_warn(dev, "Failed to close slave function\n");
+	mutex_unlock(&priv->cmd.slave_cmd_mutex);
+}
+
+static int map_bf_area(struct mlx4_dev *dev)
+{
+	struct mlx4_priv *priv = mlx4_priv(dev);
+	resource_size_t bf_start;
+	resource_size_t bf_len;
+	int err = 0;
+
+	if (!dev->caps.bf_reg_size)
+		return -ENXIO;
+
+	bf_start = pci_resource_start(dev->persist->pdev, 2) +
+			(dev->caps.num_uars << PAGE_SHIFT);
+	bf_len = pci_resource_len(dev->persist->pdev, 2) -
+			(dev->caps.num_uars << PAGE_SHIFT);
+	priv->bf_mapping = io_mapping_create_wc(bf_start, bf_len);
+	if (!priv->bf_mapping)
+		err = -ENOMEM;
+
+	return err;
+}
+
+static void unmap_bf_area(struct mlx4_dev *dev)
+{
+	if (mlx4_priv(dev)->bf_mapping)
+		io_mapping_free(mlx4_priv(dev)->bf_mapping);
+}
+
+u64 mlx4_read_clock(struct mlx4_dev *dev)
+{
+	u32 clockhi, clocklo, clockhi1;
+	u64 cycles;
+	int i;
+	struct mlx4_priv *priv = mlx4_priv(dev);
+
+	for (i = 0; i < 10; i++) {
+		clockhi = swab32(readl(priv->clock_mapping));
+		clocklo = swab32(readl(priv->clock_mapping + 4));
+		clockhi1 = swab32(readl(priv->clock_mapping));
+		if (clockhi == clockhi1)
+			break;
+	}
+
+	cycles = (u64) clockhi << 32 | (u64) clocklo;
+
+	return cycles;
+}
+EXPORT_SYMBOL_GPL(mlx4_read_clock);
+
+
+static int map_internal_clock(struct mlx4_dev *dev)
+{
+	struct mlx4_priv *priv = mlx4_priv(dev);
+
+	priv->clock_mapping =
+		ioremap(pci_resource_start(dev->persist->pdev,
+					   priv->fw.clock_bar) +
+			priv->fw.clock_offset, MLX4_CLOCK_SIZE);
+
+	if (!priv->clock_mapping)
+		return -ENOMEM;
+
+	return 0;
+}
+
+int mlx4_get_internal_clock_params(struct mlx4_dev *dev,
+				   struct mlx4_clock_params *params)
+{
+	struct mlx4_priv *priv = mlx4_priv(dev);
+
+	if (mlx4_is_slave(dev))
+		return -EOPNOTSUPP;
+
+	if (!dev->caps.map_clock_to_user) {
+		mlx4_dbg(dev, "Map clock to user is not supported.\n");
+		return -EOPNOTSUPP;
+	}
+
+	if (!params)
+		return -EINVAL;
+
+	params->bar = priv->fw.clock_bar;
+	params->offset = priv->fw.clock_offset;
+	params->size = MLX4_CLOCK_SIZE;
+
+	return 0;
+}
+EXPORT_SYMBOL_GPL(mlx4_get_internal_clock_params);
+
+static void unmap_internal_clock(struct mlx4_dev *dev)
+{
+	struct mlx4_priv *priv = mlx4_priv(dev);
+
+	if (priv->clock_mapping)
+		iounmap(priv->clock_mapping);
+}
+
+static void mlx4_close_hca(struct mlx4_dev *dev)
+{
+	unmap_internal_clock(dev);
+	unmap_bf_area(dev);
+	if (mlx4_is_slave(dev))
+		mlx4_slave_exit(dev);
+	else {
+		mlx4_CLOSE_HCA(dev, 0);
+		mlx4_free_icms(dev);
+	}
+}
+
+static void mlx4_close_fw(struct mlx4_dev *dev)
+{
+	if (!mlx4_is_slave(dev)) {
+		mlx4_UNMAP_FA(dev);
+		mlx4_free_icm(dev, mlx4_priv(dev)->fw.fw_icm, 0);
+	}
+}
+
+static int mlx4_comm_check_offline(struct mlx4_dev *dev)
+{
+#define COMM_CHAN_OFFLINE_OFFSET 0x09
+
+	u32 comm_flags;
+	u32 offline_bit;
+	unsigned long end;
+	struct mlx4_priv *priv = mlx4_priv(dev);
+
+	end = msecs_to_jiffies(MLX4_COMM_OFFLINE_TIME_OUT) + jiffies;
+	while (time_before(jiffies, end)) {
+		comm_flags = swab32(readl((__iomem char *)priv->mfunc.comm +
+					  MLX4_COMM_CHAN_FLAGS));
+		offline_bit = (comm_flags &
+			       (u32)(1 << COMM_CHAN_OFFLINE_OFFSET));
+		if (!offline_bit)
+			return 0;
+
+		/* If device removal has been requested,
+		 * do not continue retrying.
+		 */
+		if (dev->persist->interface_state &
+		    MLX4_INTERFACE_STATE_NOWAIT)
+			break;
+
+		/* There are cases as part of AER/Reset flow that PF needs
+		 * around 100 msec to load. We therefore sleep for 100 msec
+		 * to allow other tasks to make use of that CPU during this
+		 * time interval.
+		 */
+		msleep(100);
+	}
+	mlx4_err(dev, "Communication channel is offline.\n");
+	return -EIO;
+}
+
+static void mlx4_reset_vf_support(struct mlx4_dev *dev)
+{
+#define COMM_CHAN_RST_OFFSET 0x1e
+
+	struct mlx4_priv *priv = mlx4_priv(dev);
+	u32 comm_rst;
+	u32 comm_caps;
+
+	comm_caps = swab32(readl((__iomem char *)priv->mfunc.comm +
+				 MLX4_COMM_CHAN_CAPS));
+	comm_rst = (comm_caps & (u32)(1 << COMM_CHAN_RST_OFFSET));
+
+	if (comm_rst)
+		dev->caps.vf_caps |= MLX4_VF_CAP_FLAG_RESET;
+}
+
+static int mlx4_init_slave(struct mlx4_dev *dev)
+{
+	struct mlx4_priv *priv = mlx4_priv(dev);
+	u64 dma = (u64) priv->mfunc.vhcr_dma;
+	int ret_from_reset = 0;
+	u32 slave_read;
+	u32 cmd_channel_ver;
+
+	if (atomic_read(&pf_loading)) {
+		mlx4_warn(dev, "PF is not ready - Deferring probe\n");
+		return -EPROBE_DEFER;
+	}
+
+	mutex_lock(&priv->cmd.slave_cmd_mutex);
+	priv->cmd.max_cmds = 1;
+	if (mlx4_comm_check_offline(dev)) {
+		mlx4_err(dev, "PF is not responsive, skipping initialization\n");
+		goto err_offline;
+	}
+
+	mlx4_reset_vf_support(dev);
+	mlx4_warn(dev, "Sending reset\n");
+	ret_from_reset = mlx4_comm_cmd(dev, MLX4_COMM_CMD_RESET, 0,
+				       MLX4_COMM_CMD_NA_OP, MLX4_COMM_TIME);
+	/* if we are in the middle of flr the slave will try
+	 * NUM_OF_RESET_RETRIES times before leaving.*/
+	if (ret_from_reset) {
+		if (MLX4_DELAY_RESET_SLAVE == ret_from_reset) {
+			mlx4_warn(dev, "slave is currently in the middle of FLR - Deferring probe\n");
+			mutex_unlock(&priv->cmd.slave_cmd_mutex);
+			return -EPROBE_DEFER;
+		} else
+			goto err;
+	}
+
+	/* check the driver version - the slave I/F revision
+	 * must match the master's */
+	slave_read = swab32(readl(&priv->mfunc.comm->slave_read));
+	cmd_channel_ver = mlx4_comm_get_version();
+
+	if (MLX4_COMM_GET_IF_REV(cmd_channel_ver) !=
+		MLX4_COMM_GET_IF_REV(slave_read)) {
+		mlx4_err(dev, "slave driver version is not supported by the master\n");
+		goto err;
+	}
+
+	mlx4_warn(dev, "Sending vhcr0\n");
+	if (mlx4_comm_cmd(dev, MLX4_COMM_CMD_VHCR0, dma >> 48,
+			     MLX4_COMM_CMD_NA_OP, MLX4_COMM_TIME))
+		goto err;
+	if (mlx4_comm_cmd(dev, MLX4_COMM_CMD_VHCR1, dma >> 32,
+			     MLX4_COMM_CMD_NA_OP, MLX4_COMM_TIME))
+		goto err;
+	if (mlx4_comm_cmd(dev, MLX4_COMM_CMD_VHCR2, dma >> 16,
+			     MLX4_COMM_CMD_NA_OP, MLX4_COMM_TIME))
+		goto err;
+	if (mlx4_comm_cmd(dev, MLX4_COMM_CMD_VHCR_EN, dma,
+			  MLX4_COMM_CMD_NA_OP, MLX4_COMM_TIME))
+		goto err;
+
+	mutex_unlock(&priv->cmd.slave_cmd_mutex);
+	return 0;
+
+err:
+	mlx4_comm_cmd(dev, MLX4_COMM_CMD_RESET, 0, MLX4_COMM_CMD_NA_OP, 0);
+err_offline:
+	mutex_unlock(&priv->cmd.slave_cmd_mutex);
+	return -EIO;
+}
+
+static void mlx4_parav_master_pf_caps(struct mlx4_dev *dev)
+{
+	int i;
+
+	for (i = 1; i <= dev->caps.num_ports; i++) {
+		if (dev->caps.port_type[i] == MLX4_PORT_TYPE_ETH)
+			dev->caps.gid_table_len[i] =
+				mlx4_get_slave_num_gids(dev, 0, i);
+		else
+			dev->caps.gid_table_len[i] = 1;
+		dev->caps.pkey_table_len[i] =
+			dev->phys_caps.pkey_phys_table_len[i] - 1;
+	}
+}
+
+static int choose_log_fs_mgm_entry_size(int qp_per_entry)
+{
+	int i = MLX4_MIN_MGM_LOG_ENTRY_SIZE;
+
+	for (i = MLX4_MIN_MGM_LOG_ENTRY_SIZE; i <= MLX4_MAX_MGM_LOG_ENTRY_SIZE;
+	      i++) {
+		if (qp_per_entry <= 4 * ((1 << i) / 16 - 2))
+			break;
+	}
+
+	return (i <= MLX4_MAX_MGM_LOG_ENTRY_SIZE) ? i : -1;
+}
+
+static const char *dmfs_high_rate_steering_mode_str(int dmfs_high_steer_mode)
+{
+	switch (dmfs_high_steer_mode) {
+	case MLX4_STEERING_DMFS_A0_DEFAULT:
+		return "default performance";
+
+	case MLX4_STEERING_DMFS_A0_DYNAMIC:
+		return "dynamic hybrid mode";
+
+	case MLX4_STEERING_DMFS_A0_STATIC:
+		return "performance optimized for limited rule configuration (static)";
+
+	case MLX4_STEERING_DMFS_A0_DISABLE:
+		return "disabled performance optimized steering";
+
+	case MLX4_STEERING_DMFS_A0_NOT_SUPPORTED:
+		return "performance optimized steering not supported";
+
+	default:
+		return "Unrecognized mode";
+	}
+}
+
+#define MLX4_DMFS_A0_STEERING			(1UL << 2)
+
+static void choose_steering_mode(struct mlx4_dev *dev,
+				 struct mlx4_dev_cap *dev_cap)
+{
+	if (mlx4_log_num_mgm_entry_size <= 0) {
+		if ((-mlx4_log_num_mgm_entry_size) & MLX4_DMFS_A0_STEERING) {
+			if (dev->caps.dmfs_high_steer_mode ==
+			    MLX4_STEERING_DMFS_A0_NOT_SUPPORTED)
+				mlx4_err(dev, "DMFS high rate mode not supported\n");
+			else
+				dev->caps.dmfs_high_steer_mode =
+					MLX4_STEERING_DMFS_A0_STATIC;
+		}
+	}
+
+	if (mlx4_log_num_mgm_entry_size <= 0 &&
+	    dev_cap->flags2 & MLX4_DEV_CAP_FLAG2_FS_EN &&
+	    (!mlx4_is_mfunc(dev) ||
+	     (dev_cap->fs_max_num_qp_per_entry >=
+	     (dev->persist->num_vfs + 1))) &&
+	    choose_log_fs_mgm_entry_size(dev_cap->fs_max_num_qp_per_entry) >=
+		MLX4_MIN_MGM_LOG_ENTRY_SIZE) {
+		dev->oper_log_mgm_entry_size =
+			choose_log_fs_mgm_entry_size(dev_cap->fs_max_num_qp_per_entry);
+		dev->caps.steering_mode = MLX4_STEERING_MODE_DEVICE_MANAGED;
+		dev->caps.num_qp_per_mgm = dev_cap->fs_max_num_qp_per_entry;
+		dev->caps.fs_log_max_ucast_qp_range_size =
+			dev_cap->fs_log_max_ucast_qp_range_size;
+	} else {
+		if (dev->caps.dmfs_high_steer_mode !=
+		    MLX4_STEERING_DMFS_A0_NOT_SUPPORTED)
+			dev->caps.dmfs_high_steer_mode = MLX4_STEERING_DMFS_A0_DISABLE;
+		if (dev->caps.flags & MLX4_DEV_CAP_FLAG_VEP_UC_STEER &&
+		    dev->caps.flags & MLX4_DEV_CAP_FLAG_VEP_MC_STEER)
+			dev->caps.steering_mode = MLX4_STEERING_MODE_B0;
+		else {
+			dev->caps.steering_mode = MLX4_STEERING_MODE_A0;
+
+			if (dev->caps.flags & MLX4_DEV_CAP_FLAG_VEP_UC_STEER ||
+			    dev->caps.flags & MLX4_DEV_CAP_FLAG_VEP_MC_STEER)
+				mlx4_warn(dev, "Must have both UC_STEER and MC_STEER flags set to use B0 steering - falling back to A0 steering mode\n");
+		}
+		dev->oper_log_mgm_entry_size =
+			mlx4_log_num_mgm_entry_size > 0 ?
+			mlx4_log_num_mgm_entry_size :
+			MLX4_DEFAULT_MGM_LOG_ENTRY_SIZE;
+		dev->caps.num_qp_per_mgm = mlx4_get_qp_per_mgm(dev);
+	}
+	mlx4_dbg(dev, "Steering mode is: %s, oper_log_mgm_entry_size = %d, modparam log_num_mgm_entry_size = %d\n",
+		 mlx4_steering_mode_str(dev->caps.steering_mode),
+		 dev->oper_log_mgm_entry_size,
+		 mlx4_log_num_mgm_entry_size);
+}
+
+static void choose_tunnel_offload_mode(struct mlx4_dev *dev,
+				       struct mlx4_dev_cap *dev_cap)
+{
+	if (dev->caps.steering_mode == MLX4_STEERING_MODE_DEVICE_MANAGED &&
+	    dev_cap->flags2 & MLX4_DEV_CAP_FLAG2_VXLAN_OFFLOADS)
+		dev->caps.tunnel_offload_mode = MLX4_TUNNEL_OFFLOAD_MODE_VXLAN;
+	else
+		dev->caps.tunnel_offload_mode = MLX4_TUNNEL_OFFLOAD_MODE_NONE;
+
+	mlx4_dbg(dev, "Tunneling offload mode is: %s\n",  (dev->caps.tunnel_offload_mode
+		 == MLX4_TUNNEL_OFFLOAD_MODE_VXLAN) ? "vxlan" : "none");
+}
+
+static int mlx4_validate_optimized_steering(struct mlx4_dev *dev)
+{
+	int i;
+	struct mlx4_port_cap port_cap;
+
+	if (dev->caps.dmfs_high_steer_mode == MLX4_STEERING_DMFS_A0_NOT_SUPPORTED)
+		return -EINVAL;
+
+	for (i = 1; i <= dev->caps.num_ports; i++) {
+		if (mlx4_dev_port(dev, i, &port_cap)) {
+			mlx4_err(dev,
+				 "QUERY_DEV_CAP command failed, can't veify DMFS high rate steering.\n");
+		} else if ((dev->caps.dmfs_high_steer_mode !=
+			    MLX4_STEERING_DMFS_A0_DEFAULT) &&
+			   (port_cap.dmfs_optimized_state ==
+			    !!(dev->caps.dmfs_high_steer_mode ==
+			    MLX4_STEERING_DMFS_A0_DISABLE))) {
+			mlx4_err(dev,
+				 "DMFS high rate steer mode differ, driver requested %s but %s in FW.\n",
+				 dmfs_high_rate_steering_mode_str(
+					dev->caps.dmfs_high_steer_mode),
+				 (port_cap.dmfs_optimized_state ?
+					"enabled" : "disabled"));
+		}
+	}
+
+	return 0;
+}
+
+static int mlx4_init_fw(struct mlx4_dev *dev)
+{
+	struct mlx4_mod_stat_cfg   mlx4_cfg;
+	int err = 0;
+
+	if (!mlx4_is_slave(dev)) {
+		err = mlx4_QUERY_FW(dev);
+		if (err) {
+			if (err == -EACCES)
+				mlx4_info(dev, "non-primary physical function, skipping\n");
+			else
+				mlx4_err(dev, "QUERY_FW command failed, aborting\n");
+			return err;
+		}
+
+		err = mlx4_load_fw(dev);
+		if (err) {
+			mlx4_err(dev, "Failed to start FW, aborting\n");
+			return err;
+		}
+
+		mlx4_cfg.log_pg_sz_m = 1;
+		mlx4_cfg.log_pg_sz = 0;
+		err = mlx4_MOD_STAT_CFG(dev, &mlx4_cfg);
+		if (err)
+			mlx4_warn(dev, "Failed to override log_pg_sz parameter\n");
+	}
+
+	return err;
+}
+
+static int mlx4_init_hca(struct mlx4_dev *dev)
+{
+	struct mlx4_priv	  *priv = mlx4_priv(dev);
+	struct mlx4_adapter	   adapter;
+	struct mlx4_dev_cap	   dev_cap;
+	struct mlx4_profile	   profile;
+	struct mlx4_init_hca_param init_hca;
+	u64 icm_size;
+	struct mlx4_config_dev_params params;
+	int err;
+
+	if (!mlx4_is_slave(dev)) {
+		err = mlx4_dev_cap(dev, &dev_cap);
+		if (err) {
+			mlx4_err(dev, "QUERY_DEV_CAP command failed, aborting\n");
+			return err;
+		}
+
+		choose_steering_mode(dev, &dev_cap);
+		choose_tunnel_offload_mode(dev, &dev_cap);
+
+		if (dev->caps.dmfs_high_steer_mode == MLX4_STEERING_DMFS_A0_STATIC &&
+		    mlx4_is_master(dev))
+			dev->caps.function_caps |= MLX4_FUNC_CAP_DMFS_A0_STATIC;
+
+		err = mlx4_get_phys_port_id(dev);
+		if (err)
+			mlx4_err(dev, "Fail to get physical port id\n");
+
+		if (mlx4_is_master(dev))
+			mlx4_parav_master_pf_caps(dev);
+
+		if (mlx4_low_memory_profile()) {
+			mlx4_info(dev, "Running from within kdump kernel. Using low memory profile\n");
+			profile = low_mem_profile;
+		} else {
+			profile = default_profile;
+		}
+		if (dev->caps.steering_mode ==
+		    MLX4_STEERING_MODE_DEVICE_MANAGED)
+			profile.num_mcg = MLX4_FS_NUM_MCG;
+
+		icm_size = mlx4_make_profile(dev, &profile, &dev_cap,
+					     &init_hca);
+		if ((long long) icm_size < 0) {
+			err = icm_size;
+			return err;
+		}
+
+		dev->caps.max_fmr_maps = (1 << (32 - ilog2(dev->caps.num_mpts))) - 1;
+
+		if (enable_4k_uar || !dev->persist->num_vfs) {
+			init_hca.log_uar_sz = ilog2(dev->caps.num_uars) +
+						    PAGE_SHIFT - DEFAULT_UAR_PAGE_SHIFT;
+			init_hca.uar_page_sz = DEFAULT_UAR_PAGE_SHIFT - 12;
+		} else {
+			init_hca.log_uar_sz = ilog2(dev->caps.num_uars);
+			init_hca.uar_page_sz = PAGE_SHIFT - 12;
+		}
+
+		init_hca.mw_enabled = 0;
+		if (dev->caps.flags & MLX4_DEV_CAP_FLAG_MEM_WINDOW ||
+		    dev->caps.bmme_flags & MLX4_BMME_FLAG_TYPE_2_WIN)
+			init_hca.mw_enabled = INIT_HCA_TPT_MW_ENABLE;
+
+		err = mlx4_init_icm(dev, &dev_cap, &init_hca, icm_size);
+		if (err)
+			return err;
+
+		err = mlx4_INIT_HCA(dev, &init_hca);
+		if (err) {
+			mlx4_err(dev, "INIT_HCA command failed, aborting\n");
+			goto err_free_icm;
+		}
+
+		if (dev_cap.flags2 & MLX4_DEV_CAP_FLAG2_SYS_EQS) {
+			err = mlx4_query_func(dev, &dev_cap);
+			if (err < 0) {
+				mlx4_err(dev, "QUERY_FUNC command failed, aborting.\n");
+				goto err_close;
+			} else if (err & MLX4_QUERY_FUNC_NUM_SYS_EQS) {
+				dev->caps.num_eqs = dev_cap.max_eqs;
+				dev->caps.reserved_eqs = dev_cap.reserved_eqs;
+				dev->caps.reserved_uars = dev_cap.reserved_uars;
+			}
+		}
+
+		/*
+		 * If TS is supported by FW
+		 * read HCA frequency by QUERY_HCA command
+		 */
+		if (dev->caps.flags2 & MLX4_DEV_CAP_FLAG2_TS) {
+			memset(&init_hca, 0, sizeof(init_hca));
+			err = mlx4_QUERY_HCA(dev, &init_hca);
+			if (err) {
+				mlx4_err(dev, "QUERY_HCA command failed, disable timestamp\n");
+				dev->caps.flags2 &= ~MLX4_DEV_CAP_FLAG2_TS;
+			} else {
+				dev->caps.hca_core_clock =
+					init_hca.hca_core_clock;
+			}
+
+			/* In case we got HCA frequency 0 - disable timestamping
+			 * to avoid dividing by zero
+			 */
+			if (!dev->caps.hca_core_clock) {
+				dev->caps.flags2 &= ~MLX4_DEV_CAP_FLAG2_TS;
+				mlx4_err(dev,
+					 "HCA frequency is 0 - timestamping is not supported\n");
+			} else if (map_internal_clock(dev)) {
+				/*
+				 * Map internal clock,
+				 * in case of failure disable timestamping
+				 */
+				dev->caps.flags2 &= ~MLX4_DEV_CAP_FLAG2_TS;
+				mlx4_err(dev, "Failed to map internal clock. Timestamping is not supported\n");
+			}
+		}
+
+		if (dev->caps.dmfs_high_steer_mode !=
+		    MLX4_STEERING_DMFS_A0_NOT_SUPPORTED) {
+			if (mlx4_validate_optimized_steering(dev))
+				mlx4_warn(dev, "Optimized steering validation failed\n");
+
+			if (dev->caps.dmfs_high_steer_mode ==
+			    MLX4_STEERING_DMFS_A0_DISABLE) {
+				dev->caps.dmfs_high_rate_qpn_base =
+					dev->caps.reserved_qps_cnt[MLX4_QP_REGION_FW];
+				dev->caps.dmfs_high_rate_qpn_range =
+					MLX4_A0_STEERING_TABLE_SIZE;
+			}
+
+			mlx4_info(dev, "DMFS high rate steer mode is: %s\n",
+				  dmfs_high_rate_steering_mode_str(
+					dev->caps.dmfs_high_steer_mode));
+		}
+	} else {
+		err = mlx4_init_slave(dev);
+		if (err) {
+			if (err != -EPROBE_DEFER)
+				mlx4_err(dev, "Failed to initialize slave\n");
+			return err;
+		}
+
+		err = mlx4_slave_cap(dev);
+		if (err) {
+			mlx4_err(dev, "Failed to obtain slave caps\n");
+			goto err_close;
+		}
+	}
+
+	if (map_bf_area(dev))
+		mlx4_dbg(dev, "Failed to map blue flame area\n");
+
+	/*Only the master set the ports, all the rest got it from it.*/
+	if (!mlx4_is_slave(dev))
+		mlx4_set_port_mask(dev);
+
+	err = mlx4_QUERY_ADAPTER(dev, &adapter);
+	if (err) {
+		mlx4_err(dev, "QUERY_ADAPTER command failed, aborting\n");
+		goto unmap_bf;
+	}
+
+	/* Query CONFIG_DEV parameters */
+	err = mlx4_config_dev_retrieval(dev, &params);
+	if (err && err != -EOPNOTSUPP) {
+		mlx4_err(dev, "Failed to query CONFIG_DEV parameters\n");
+	} else if (!err) {
+		dev->caps.rx_checksum_flags_port[1] = params.rx_csum_flags_port_1;
+		dev->caps.rx_checksum_flags_port[2] = params.rx_csum_flags_port_2;
+	}
+	priv->eq_table.inta_pin = adapter.inta_pin;
+	memcpy(dev->board_id, adapter.board_id, sizeof(dev->board_id));
+
+	return 0;
+
+unmap_bf:
+	unmap_internal_clock(dev);
+	unmap_bf_area(dev);
+
+	if (mlx4_is_slave(dev))
+		mlx4_slave_destroy_special_qp_cap(dev);
+
+err_close:
+	if (mlx4_is_slave(dev))
+		mlx4_slave_exit(dev);
+	else
+		mlx4_CLOSE_HCA(dev, 0);
+
+err_free_icm:
+	if (!mlx4_is_slave(dev))
+		mlx4_free_icms(dev);
+
+	return err;
+}
+
+static int mlx4_init_counters_table(struct mlx4_dev *dev)
+{
+	struct mlx4_priv *priv = mlx4_priv(dev);
+	int nent_pow2;
+
+	if (!(dev->caps.flags & MLX4_DEV_CAP_FLAG_COUNTERS))
+		return -ENOENT;
+
+	if (!dev->caps.max_counters)
+		return -ENOSPC;
+
+	nent_pow2 = roundup_pow_of_two(dev->caps.max_counters);
+	/* reserve last counter index for sink counter */
+	return mlx4_bitmap_init(&priv->counters_bitmap, nent_pow2,
+				nent_pow2 - 1, 0,
+				nent_pow2 - dev->caps.max_counters + 1);
+}
+
+static void mlx4_cleanup_counters_table(struct mlx4_dev *dev)
+{
+	if (!(dev->caps.flags & MLX4_DEV_CAP_FLAG_COUNTERS))
+		return;
+
+	if (!dev->caps.max_counters)
+		return;
+
+	mlx4_bitmap_cleanup(&mlx4_priv(dev)->counters_bitmap);
+}
+
+static void mlx4_cleanup_default_counters(struct mlx4_dev *dev)
+{
+	struct mlx4_priv *priv = mlx4_priv(dev);
+	int port;
+
+	for (port = 0; port < dev->caps.num_ports; port++)
+		if (priv->def_counter[port] != -1)
+			mlx4_counter_free(dev,  priv->def_counter[port]);
+}
+
+static int mlx4_allocate_default_counters(struct mlx4_dev *dev)
+{
+	struct mlx4_priv *priv = mlx4_priv(dev);
+	int port, err = 0;
+	u32 idx;
+
+	for (port = 0; port < dev->caps.num_ports; port++)
+		priv->def_counter[port] = -1;
+
+	for (port = 0; port < dev->caps.num_ports; port++) {
+		err = mlx4_counter_alloc(dev, &idx, MLX4_RES_USAGE_DRIVER);
+
+		if (!err || err == -ENOSPC) {
+			priv->def_counter[port] = idx;
+			err = 0;
+		} else if (err == -ENOENT) {
+			err = 0;
+			continue;
+		} else if (mlx4_is_slave(dev) && err == -EINVAL) {
+			priv->def_counter[port] = MLX4_SINK_COUNTER_INDEX(dev);
+			mlx4_warn(dev, "can't allocate counter from old PF driver, using index %d\n",
+				  MLX4_SINK_COUNTER_INDEX(dev));
+			err = 0;
+		} else {
+			mlx4_err(dev, "%s: failed to allocate default counter port %d err %d\n",
+				 __func__, port + 1, err);
+			mlx4_cleanup_default_counters(dev);
+			return err;
+		}
+
+		mlx4_dbg(dev, "%s: default counter index %d for port %d\n",
+			 __func__, priv->def_counter[port], port + 1);
+	}
+
+	return err;
+}
+
+int __mlx4_counter_alloc(struct mlx4_dev *dev, u32 *idx)
+{
+	struct mlx4_priv *priv = mlx4_priv(dev);
+
+	if (!(dev->caps.flags & MLX4_DEV_CAP_FLAG_COUNTERS))
+		return -ENOENT;
+
+	*idx = mlx4_bitmap_alloc(&priv->counters_bitmap);
+	if (*idx == -1) {
+		*idx = MLX4_SINK_COUNTER_INDEX(dev);
+		return -ENOSPC;
+	}
+
+	return 0;
+}
+
+int mlx4_counter_alloc(struct mlx4_dev *dev, u32 *idx, u8 usage)
+{
+	u32 in_modifier = RES_COUNTER | (((u32)usage & 3) << 30);
+	u64 out_param;
+	int err;
+
+	if (mlx4_is_mfunc(dev)) {
+		err = mlx4_cmd_imm(dev, 0, &out_param, in_modifier,
+				   RES_OP_RESERVE, MLX4_CMD_ALLOC_RES,
+				   MLX4_CMD_TIME_CLASS_A, MLX4_CMD_WRAPPED);
+		if (!err)
+			*idx = get_param_l(&out_param);
+		if (WARN_ON(err == -ENOSPC))
+			err = -EINVAL;
+		return err;
+	}
+	return __mlx4_counter_alloc(dev, idx);
+}
+EXPORT_SYMBOL_GPL(mlx4_counter_alloc);
+
+static int __mlx4_clear_if_stat(struct mlx4_dev *dev,
+				u8 counter_index)
+{
+	struct mlx4_cmd_mailbox *if_stat_mailbox;
+	int err;
+	u32 if_stat_in_mod = (counter_index & 0xff) | MLX4_QUERY_IF_STAT_RESET;
+
+	if_stat_mailbox = mlx4_alloc_cmd_mailbox(dev);
+	if (IS_ERR(if_stat_mailbox))
+		return PTR_ERR(if_stat_mailbox);
+
+	err = mlx4_cmd_box(dev, 0, if_stat_mailbox->dma, if_stat_in_mod, 0,
+			   MLX4_CMD_QUERY_IF_STAT, MLX4_CMD_TIME_CLASS_C,
+			   MLX4_CMD_NATIVE);
+
+	mlx4_free_cmd_mailbox(dev, if_stat_mailbox);
+	return err;
+}
+
+void __mlx4_counter_free(struct mlx4_dev *dev, u32 idx)
+{
+	if (!(dev->caps.flags & MLX4_DEV_CAP_FLAG_COUNTERS))
+		return;
+
+	if (idx == MLX4_SINK_COUNTER_INDEX(dev))
+		return;
+
+	__mlx4_clear_if_stat(dev, idx);
+
+	mlx4_bitmap_free(&mlx4_priv(dev)->counters_bitmap, idx, MLX4_USE_RR);
+	return;
+}
+
+void mlx4_counter_free(struct mlx4_dev *dev, u32 idx)
+{
+	u64 in_param = 0;
+
+	if (mlx4_is_mfunc(dev)) {
+		set_param_l(&in_param, idx);
+		mlx4_cmd(dev, in_param, RES_COUNTER, RES_OP_RESERVE,
+			 MLX4_CMD_FREE_RES, MLX4_CMD_TIME_CLASS_A,
+			 MLX4_CMD_WRAPPED);
+		return;
+	}
+	__mlx4_counter_free(dev, idx);
+}
+EXPORT_SYMBOL_GPL(mlx4_counter_free);
+
+int mlx4_get_default_counter_index(struct mlx4_dev *dev, int port)
+{
+	struct mlx4_priv *priv = mlx4_priv(dev);
+
+	return priv->def_counter[port - 1];
+}
+EXPORT_SYMBOL_GPL(mlx4_get_default_counter_index);
+
+void mlx4_set_admin_guid(struct mlx4_dev *dev, __be64 guid, int entry, int port)
+{
+	struct mlx4_priv *priv = mlx4_priv(dev);
+
+	priv->mfunc.master.vf_admin[entry].vport[port].guid = guid;
+}
+EXPORT_SYMBOL_GPL(mlx4_set_admin_guid);
+
+__be64 mlx4_get_admin_guid(struct mlx4_dev *dev, int entry, int port)
+{
+	struct mlx4_priv *priv = mlx4_priv(dev);
+
+	return priv->mfunc.master.vf_admin[entry].vport[port].guid;
+}
+EXPORT_SYMBOL_GPL(mlx4_get_admin_guid);
+
+void mlx4_set_random_admin_guid(struct mlx4_dev *dev, int entry, int port)
+{
+	struct mlx4_priv *priv = mlx4_priv(dev);
+	__be64 guid;
+
+	/* hw GUID */
+	if (entry == 0)
+		return;
+
+	get_random_bytes((char *)&guid, sizeof(guid));
+	guid &= ~(cpu_to_be64(1ULL << 56));
+	guid |= cpu_to_be64(1ULL << 57);
+	priv->mfunc.master.vf_admin[entry].vport[port].guid = guid;
+}
+
+static int mlx4_setup_hca(struct mlx4_dev *dev)
+{
+	struct mlx4_priv *priv = mlx4_priv(dev);
+	int err;
+	int port;
+	__be32 ib_port_default_caps;
+
+	err = mlx4_init_uar_table(dev);
+	if (err) {
+		mlx4_err(dev, "Failed to initialize user access region table, aborting\n");
+		return err;
+	}
+
+	err = mlx4_uar_alloc(dev, &priv->driver_uar);
+	if (err) {
+		mlx4_err(dev, "Failed to allocate driver access region, aborting\n");
+		goto err_uar_table_free;
+	}
+
+	priv->kar = ioremap((phys_addr_t) priv->driver_uar.pfn << PAGE_SHIFT, PAGE_SIZE);
+	if (!priv->kar) {
+		mlx4_err(dev, "Couldn't map kernel access region, aborting\n");
+		err = -ENOMEM;
+		goto err_uar_free;
+	}
+
+	err = mlx4_init_pd_table(dev);
+	if (err) {
+		mlx4_err(dev, "Failed to initialize protection domain table, aborting\n");
+		goto err_kar_unmap;
+	}
+
+	err = mlx4_init_xrcd_table(dev);
+	if (err) {
+		mlx4_err(dev, "Failed to initialize reliable connection domain table, aborting\n");
+		goto err_pd_table_free;
+	}
+
+	err = mlx4_init_mr_table(dev);
+	if (err) {
+		mlx4_err(dev, "Failed to initialize memory region table, aborting\n");
+		goto err_xrcd_table_free;
+	}
+
+	if (!mlx4_is_slave(dev)) {
+		err = mlx4_init_mcg_table(dev);
+		if (err) {
+			mlx4_err(dev, "Failed to initialize multicast group table, aborting\n");
+			goto err_mr_table_free;
+		}
+		err = mlx4_config_mad_demux(dev);
+		if (err) {
+			mlx4_err(dev, "Failed in config_mad_demux, aborting\n");
+			goto err_mcg_table_free;
+		}
+	}
+
+	err = mlx4_init_eq_table(dev);
+	if (err) {
+		mlx4_err(dev, "Failed to initialize event queue table, aborting\n");
+		goto err_mcg_table_free;
+	}
+
+	err = mlx4_cmd_use_events(dev);
+	if (err) {
+		mlx4_err(dev, "Failed to switch to event-driven firmware commands, aborting\n");
+		goto err_eq_table_free;
+	}
+
+	err = mlx4_NOP(dev);
+	if (err) {
+		if (dev->flags & MLX4_FLAG_MSI_X) {
+			mlx4_warn(dev, "NOP command failed to generate MSI-X interrupt IRQ %d)\n",
+				  priv->eq_table.eq[MLX4_EQ_ASYNC].irq);
+			mlx4_warn(dev, "Trying again without MSI-X\n");
+		} else {
+			mlx4_err(dev, "NOP command failed to generate interrupt (IRQ %d), aborting\n",
+				 priv->eq_table.eq[MLX4_EQ_ASYNC].irq);
+			mlx4_err(dev, "BIOS or ACPI interrupt routing problem?\n");
+		}
+
+		goto err_cmd_poll;
+	}
+
+	mlx4_dbg(dev, "NOP command IRQ test passed\n");
+
+	err = mlx4_init_cq_table(dev);
+	if (err) {
+		mlx4_err(dev, "Failed to initialize completion queue table, aborting\n");
+		goto err_cmd_poll;
+	}
+
+	err = mlx4_init_srq_table(dev);
+	if (err) {
+		mlx4_err(dev, "Failed to initialize shared receive queue table, aborting\n");
+		goto err_cq_table_free;
+	}
+
+	err = mlx4_init_qp_table(dev);
+	if (err) {
+		mlx4_err(dev, "Failed to initialize queue pair table, aborting\n");
+		goto err_srq_table_free;
+	}
+
+	if (!mlx4_is_slave(dev)) {
+		err = mlx4_init_counters_table(dev);
+		if (err && err != -ENOENT) {
+			mlx4_err(dev, "Failed to initialize counters table, aborting\n");
+			goto err_qp_table_free;
+		}
+	}
+
+	err = mlx4_allocate_default_counters(dev);
+	if (err) {
+		mlx4_err(dev, "Failed to allocate default counters, aborting\n");
+		goto err_counters_table_free;
+	}
+
+	if (!mlx4_is_slave(dev)) {
+		for (port = 1; port <= dev->caps.num_ports; port++) {
+			ib_port_default_caps = 0;
+			err = mlx4_get_port_ib_caps(dev, port,
+						    &ib_port_default_caps);
+			if (err)
+				mlx4_warn(dev, "failed to get port %d default ib capabilities (%d). Continuing with caps = 0\n",
+					  port, err);
+			dev->caps.ib_port_def_cap[port] = ib_port_default_caps;
+
+			/* initialize per-slave default ib port capabilities */
+			if (mlx4_is_master(dev)) {
+				int i;
+				for (i = 0; i < dev->num_slaves; i++) {
+					if (i == mlx4_master_func_num(dev))
+						continue;
+					priv->mfunc.master.slave_state[i].ib_cap_mask[port] =
+						ib_port_default_caps;
+				}
+			}
+
+			if (mlx4_is_mfunc(dev))
+				dev->caps.port_ib_mtu[port] = IB_MTU_2048;
+			else
+				dev->caps.port_ib_mtu[port] = IB_MTU_4096;
+
+			err = mlx4_SET_PORT(dev, port, mlx4_is_master(dev) ?
+					    dev->caps.pkey_table_len[port] : -1);
+			if (err) {
+				mlx4_err(dev, "Failed to set port %d, aborting\n",
+					 port);
+				goto err_default_countes_free;
+			}
+		}
+	}
+
+	return 0;
+
+err_default_countes_free:
+	mlx4_cleanup_default_counters(dev);
+
+err_counters_table_free:
+	if (!mlx4_is_slave(dev))
+		mlx4_cleanup_counters_table(dev);
+
+err_qp_table_free:
+	mlx4_cleanup_qp_table(dev);
+
+err_srq_table_free:
+	mlx4_cleanup_srq_table(dev);
+
+err_cq_table_free:
+	mlx4_cleanup_cq_table(dev);
+
+err_cmd_poll:
+	mlx4_cmd_use_polling(dev);
+
+err_eq_table_free:
+	mlx4_cleanup_eq_table(dev);
+
+err_mcg_table_free:
+	if (!mlx4_is_slave(dev))
+		mlx4_cleanup_mcg_table(dev);
+
+err_mr_table_free:
+	mlx4_cleanup_mr_table(dev);
+
+err_xrcd_table_free:
+	mlx4_cleanup_xrcd_table(dev);
+
+err_pd_table_free:
+	mlx4_cleanup_pd_table(dev);
+
+err_kar_unmap:
+	iounmap(priv->kar);
+
+err_uar_free:
+	mlx4_uar_free(dev, &priv->driver_uar);
+
+err_uar_table_free:
+	mlx4_cleanup_uar_table(dev);
+	return err;
+}
+
+static int mlx4_init_affinity_hint(struct mlx4_dev *dev, int port, int eqn)
+{
+	int requested_cpu = 0;
+	struct mlx4_priv *priv = mlx4_priv(dev);
+	struct mlx4_eq *eq;
+	int off = 0;
+	int i;
+
+	if (eqn > dev->caps.num_comp_vectors)
+		return -EINVAL;
+
+	for (i = 1; i < port; i++)
+		off += mlx4_get_eqs_per_port(dev, i);
+
+	requested_cpu = eqn - off - !!(eqn > MLX4_EQ_ASYNC);
+
+	/* Meaning EQs are shared, and this call comes from the second port */
+	if (requested_cpu < 0)
+		return 0;
+
+	eq = &priv->eq_table.eq[eqn];
+
+	if (!zalloc_cpumask_var(&eq->affinity_mask, GFP_KERNEL))
+		return -ENOMEM;
+
+	cpumask_set_cpu(requested_cpu, eq->affinity_mask);
+
+	return 0;
+}
+
+static void mlx4_enable_msi_x(struct mlx4_dev *dev)
+{
+	struct mlx4_priv *priv = mlx4_priv(dev);
+	struct msix_entry *entries;
+	int i;
+	int port = 0;
+
+	if (msi_x) {
+		int nreq = min3(dev->caps.num_ports *
+				(int)num_online_cpus() + 1,
+				dev->caps.num_eqs - dev->caps.reserved_eqs,
+				MAX_MSIX);
+
+		if (msi_x > 1)
+			nreq = min_t(int, nreq, msi_x);
+
+		entries = kcalloc(nreq, sizeof(*entries), GFP_KERNEL);
+		if (!entries)
+			goto no_msi;
+
+		for (i = 0; i < nreq; ++i)
+			entries[i].entry = i;
+
+		nreq = pci_enable_msix_range(dev->persist->pdev, entries, 2,
+					     nreq);
+
+		if (nreq < 0 || nreq < MLX4_EQ_ASYNC) {
+			kfree(entries);
+			goto no_msi;
+		}
+		/* 1 is reserved for events (asyncrounous EQ) */
+		dev->caps.num_comp_vectors = nreq - 1;
+
+		priv->eq_table.eq[MLX4_EQ_ASYNC].irq = entries[0].vector;
+		bitmap_zero(priv->eq_table.eq[MLX4_EQ_ASYNC].actv_ports.ports,
+			    dev->caps.num_ports);
+
+		for (i = 0; i < dev->caps.num_comp_vectors + 1; i++) {
+			if (i == MLX4_EQ_ASYNC)
+				continue;
+
+			priv->eq_table.eq[i].irq =
+				entries[i + 1 - !!(i > MLX4_EQ_ASYNC)].vector;
+
+			if (MLX4_IS_LEGACY_EQ_MODE(dev->caps)) {
+				bitmap_fill(priv->eq_table.eq[i].actv_ports.ports,
+					    dev->caps.num_ports);
+				/* We don't set affinity hint when there
+				 * aren't enough EQs
+				 */
+			} else {
+				set_bit(port,
+					priv->eq_table.eq[i].actv_ports.ports);
+				if (mlx4_init_affinity_hint(dev, port + 1, i))
+					mlx4_warn(dev, "Couldn't init hint cpumask for EQ %d\n",
+						  i);
+			}
+			/* We divide the Eqs evenly between the two ports.
+			 * (dev->caps.num_comp_vectors / dev->caps.num_ports)
+			 * refers to the number of Eqs per port
+			 * (i.e eqs_per_port). Theoretically, we would like to
+			 * write something like (i + 1) % eqs_per_port == 0.
+			 * However, since there's an asynchronous Eq, we have
+			 * to skip over it by comparing this condition to
+			 * !!((i + 1) > MLX4_EQ_ASYNC).
+			 */
+			if ((dev->caps.num_comp_vectors > dev->caps.num_ports) &&
+			    ((i + 1) %
+			     (dev->caps.num_comp_vectors / dev->caps.num_ports)) ==
+			    !!((i + 1) > MLX4_EQ_ASYNC))
+				/* If dev->caps.num_comp_vectors < dev->caps.num_ports,
+				 * everything is shared anyway.
+				 */
+				port++;
+		}
+
+		dev->flags |= MLX4_FLAG_MSI_X;
+
+		kfree(entries);
+		return;
+	}
+
+no_msi:
+	dev->caps.num_comp_vectors = 1;
+
+	BUG_ON(MLX4_EQ_ASYNC >= 2);
+	for (i = 0; i < 2; ++i) {
+		priv->eq_table.eq[i].irq = dev->persist->pdev->irq;
+		if (i != MLX4_EQ_ASYNC) {
+			bitmap_fill(priv->eq_table.eq[i].actv_ports.ports,
+				    dev->caps.num_ports);
+		}
+	}
+}
+
+static int mlx4_init_port_info(struct mlx4_dev *dev, int port)
+{
+	struct devlink *devlink = priv_to_devlink(mlx4_priv(dev));
+	struct mlx4_port_info *info = &mlx4_priv(dev)->port[port];
+	int err;
+
+	err = devlink_port_register(devlink, &info->devlink_port, port);
+	if (err)
+		return err;
+
+	info->dev = dev;
+	info->port = port;
+	if (!mlx4_is_slave(dev)) {
+		mlx4_init_mac_table(dev, &info->mac_table);
+		mlx4_init_vlan_table(dev, &info->vlan_table);
+		mlx4_init_roce_gid_table(dev, &info->gid_table);
+		info->base_qpn = mlx4_get_base_qpn(dev, port);
+	}
+
+	sprintf(info->dev_name, "mlx4_port%d", port);
+	info->port_attr.attr.name = info->dev_name;
+	if (mlx4_is_mfunc(dev)) {
+		info->port_attr.attr.mode = 0444;
+	} else {
+		info->port_attr.attr.mode = 0644;
+		info->port_attr.store     = set_port_type;
+	}
+	info->port_attr.show      = show_port_type;
+	sysfs_attr_init(&info->port_attr.attr);
+
+	err = device_create_file(&dev->persist->pdev->dev, &info->port_attr);
+	if (err) {
+		mlx4_err(dev, "Failed to create file for port %d\n", port);
+		devlink_port_unregister(&info->devlink_port);
+		info->port = -1;
+		return err;
+	}
+
+	sprintf(info->dev_mtu_name, "mlx4_port%d_mtu", port);
+	info->port_mtu_attr.attr.name = info->dev_mtu_name;
+	if (mlx4_is_mfunc(dev)) {
+		info->port_mtu_attr.attr.mode = 0444;
+	} else {
+		info->port_mtu_attr.attr.mode = 0644;
+		info->port_mtu_attr.store     = set_port_ib_mtu;
+	}
+	info->port_mtu_attr.show      = show_port_ib_mtu;
+	sysfs_attr_init(&info->port_mtu_attr.attr);
+
+	err = device_create_file(&dev->persist->pdev->dev,
+				 &info->port_mtu_attr);
+	if (err) {
+		mlx4_err(dev, "Failed to create mtu file for port %d\n", port);
+		device_remove_file(&info->dev->persist->pdev->dev,
+				   &info->port_attr);
+		devlink_port_unregister(&info->devlink_port);
+		info->port = -1;
+		return err;
+	}
+
+	return 0;
+}
+
+static void mlx4_cleanup_port_info(struct mlx4_port_info *info)
+{
+	if (info->port < 0)
+		return;
+
+	device_remove_file(&info->dev->persist->pdev->dev, &info->port_attr);
+	device_remove_file(&info->dev->persist->pdev->dev,
+			   &info->port_mtu_attr);
+	devlink_port_unregister(&info->devlink_port);
+
+#ifdef CONFIG_RFS_ACCEL
+	free_irq_cpu_rmap(info->rmap);
+	info->rmap = NULL;
+#endif
+}
+
+static int mlx4_init_steering(struct mlx4_dev *dev)
+{
+	struct mlx4_priv *priv = mlx4_priv(dev);
+	int num_entries = dev->caps.num_ports;
+	int i, j;
+
+	priv->steer = kcalloc(num_entries, sizeof(struct mlx4_steer),
+			      GFP_KERNEL);
+	if (!priv->steer)
+		return -ENOMEM;
+
+	for (i = 0; i < num_entries; i++)
+		for (j = 0; j < MLX4_NUM_STEERS; j++) {
+			INIT_LIST_HEAD(&priv->steer[i].promisc_qps[j]);
+			INIT_LIST_HEAD(&priv->steer[i].steer_entries[j]);
+		}
+	return 0;
+}
+
+static void mlx4_clear_steering(struct mlx4_dev *dev)
+{
+	struct mlx4_priv *priv = mlx4_priv(dev);
+	struct mlx4_steer_index *entry, *tmp_entry;
+	struct mlx4_promisc_qp *pqp, *tmp_pqp;
+	int num_entries = dev->caps.num_ports;
+	int i, j;
+
+	for (i = 0; i < num_entries; i++) {
+		for (j = 0; j < MLX4_NUM_STEERS; j++) {
+			list_for_each_entry_safe(pqp, tmp_pqp,
+						 &priv->steer[i].promisc_qps[j],
+						 list) {
+				list_del(&pqp->list);
+				kfree(pqp);
+			}
+			list_for_each_entry_safe(entry, tmp_entry,
+						 &priv->steer[i].steer_entries[j],
+						 list) {
+				list_del(&entry->list);
+				list_for_each_entry_safe(pqp, tmp_pqp,
+							 &entry->duplicates,
+							 list) {
+					list_del(&pqp->list);
+					kfree(pqp);
+				}
+				kfree(entry);
+			}
+		}
+	}
+	kfree(priv->steer);
+}
+
+static int extended_func_num(struct pci_dev *pdev)
+{
+	return PCI_SLOT(pdev->devfn) * 8 + PCI_FUNC(pdev->devfn);
+}
+
+#define MLX4_OWNER_BASE	0x8069c
+#define MLX4_OWNER_SIZE	4
+
+static int mlx4_get_ownership(struct mlx4_dev *dev)
+{
+	void __iomem *owner;
+	u32 ret;
+
+	if (pci_channel_offline(dev->persist->pdev))
+		return -EIO;
+
+	owner = ioremap(pci_resource_start(dev->persist->pdev, 0) +
+			MLX4_OWNER_BASE,
+			MLX4_OWNER_SIZE);
+	if (!owner) {
+		mlx4_err(dev, "Failed to obtain ownership bit\n");
+		return -ENOMEM;
+	}
+
+	ret = readl(owner);
+	iounmap(owner);
+	return (int) !!ret;
+}
+
+static void mlx4_free_ownership(struct mlx4_dev *dev)
+{
+	void __iomem *owner;
+
+	if (pci_channel_offline(dev->persist->pdev))
+		return;
+
+	owner = ioremap(pci_resource_start(dev->persist->pdev, 0) +
+			MLX4_OWNER_BASE,
+			MLX4_OWNER_SIZE);
+	if (!owner) {
+		mlx4_err(dev, "Failed to obtain ownership bit\n");
+		return;
+	}
+	writel(0, owner);
+	msleep(1000);
+	iounmap(owner);
+}
+
+#define SRIOV_VALID_STATE(flags) (!!((flags) & MLX4_FLAG_SRIOV)	==\
+				  !!((flags) & MLX4_FLAG_MASTER))
+
+static u64 mlx4_enable_sriov(struct mlx4_dev *dev, struct pci_dev *pdev,
+			     u8 total_vfs, int existing_vfs, int reset_flow)
+{
+	u64 dev_flags = dev->flags;
+	int err = 0;
+	int fw_enabled_sriov_vfs = min(pci_sriov_get_totalvfs(pdev),
+					MLX4_MAX_NUM_VF);
+
+	if (reset_flow) {
+		dev->dev_vfs = kcalloc(total_vfs, sizeof(*dev->dev_vfs),
+				       GFP_KERNEL);
+		if (!dev->dev_vfs)
+			goto free_mem;
+		return dev_flags;
+	}
+
+	atomic_inc(&pf_loading);
+	if (dev->flags &  MLX4_FLAG_SRIOV) {
+		if (existing_vfs != total_vfs) {
+			mlx4_err(dev, "SR-IOV was already enabled, but with num_vfs (%d) different than requested (%d)\n",
+				 existing_vfs, total_vfs);
+			total_vfs = existing_vfs;
+		}
+	}
+
+	dev->dev_vfs = kcalloc(total_vfs, sizeof(*dev->dev_vfs), GFP_KERNEL);
+	if (NULL == dev->dev_vfs) {
+		mlx4_err(dev, "Failed to allocate memory for VFs\n");
+		goto disable_sriov;
+	}
+
+	if (!(dev->flags &  MLX4_FLAG_SRIOV)) {
+		if (total_vfs > fw_enabled_sriov_vfs) {
+			mlx4_err(dev, "requested vfs (%d) > available vfs (%d). Continuing without SR_IOV\n",
+				 total_vfs, fw_enabled_sriov_vfs);
+			err = -ENOMEM;
+			goto disable_sriov;
+		}
+		mlx4_warn(dev, "Enabling SR-IOV with %d VFs\n", total_vfs);
+		err = pci_enable_sriov(pdev, total_vfs);
+	}
+	if (err) {
+		mlx4_err(dev, "Failed to enable SR-IOV, continuing without SR-IOV (err = %d)\n",
+			 err);
+		goto disable_sriov;
+	} else {
+		mlx4_warn(dev, "Running in master mode\n");
+		dev_flags |= MLX4_FLAG_SRIOV |
+			MLX4_FLAG_MASTER;
+		dev_flags &= ~MLX4_FLAG_SLAVE;
+		dev->persist->num_vfs = total_vfs;
+	}
+	return dev_flags;
+
+disable_sriov:
+	atomic_dec(&pf_loading);
+free_mem:
+	dev->persist->num_vfs = 0;
+	kfree(dev->dev_vfs);
+        dev->dev_vfs = NULL;
+	return dev_flags & ~MLX4_FLAG_MASTER;
+}
+
+enum {
+	MLX4_DEV_CAP_CHECK_NUM_VFS_ABOVE_64 = -1,
+};
+
+static int mlx4_check_dev_cap(struct mlx4_dev *dev, struct mlx4_dev_cap *dev_cap,
+			      int *nvfs)
+{
+	int requested_vfs = nvfs[0] + nvfs[1] + nvfs[2];
+	/* Checking for 64 VFs as a limitation of CX2 */
+	if (!(dev_cap->flags2 & MLX4_DEV_CAP_FLAG2_80_VFS) &&
+	    requested_vfs >= 64) {
+		mlx4_err(dev, "Requested %d VFs, but FW does not support more than 64\n",
+			 requested_vfs);
+		return MLX4_DEV_CAP_CHECK_NUM_VFS_ABOVE_64;
+	}
+	return 0;
+}
+
+static int mlx4_pci_enable_device(struct mlx4_dev *dev)
+{
+	struct pci_dev *pdev = dev->persist->pdev;
+	int err = 0;
+
+	mutex_lock(&dev->persist->pci_status_mutex);
+	if (dev->persist->pci_status == MLX4_PCI_STATUS_DISABLED) {
+		err = pci_enable_device(pdev);
+		if (!err)
+			dev->persist->pci_status = MLX4_PCI_STATUS_ENABLED;
+	}
+	mutex_unlock(&dev->persist->pci_status_mutex);
+
+	return err;
+}
+
+static void mlx4_pci_disable_device(struct mlx4_dev *dev)
+{
+	struct pci_dev *pdev = dev->persist->pdev;
+
+	mutex_lock(&dev->persist->pci_status_mutex);
+	if (dev->persist->pci_status == MLX4_PCI_STATUS_ENABLED) {
+		pci_disable_device(pdev);
+		dev->persist->pci_status = MLX4_PCI_STATUS_DISABLED;
+	}
+	mutex_unlock(&dev->persist->pci_status_mutex);
+}
+
+static int mlx4_load_one(struct pci_dev *pdev, int pci_dev_data,
+			 int total_vfs, int *nvfs, struct mlx4_priv *priv,
+			 int reset_flow)
+{
+	struct mlx4_dev *dev;
+	unsigned sum = 0;
+	int err;
+	int port;
+	int i;
+	struct mlx4_dev_cap *dev_cap = NULL;
+	int existing_vfs = 0;
+
+	dev = &priv->dev;
+
+	INIT_LIST_HEAD(&priv->ctx_list);
+	spin_lock_init(&priv->ctx_lock);
+
+	mutex_init(&priv->port_mutex);
+	mutex_init(&priv->bond_mutex);
+
+	INIT_LIST_HEAD(&priv->pgdir_list);
+	mutex_init(&priv->pgdir_mutex);
+	spin_lock_init(&priv->cmd.context_lock);
+
+	INIT_LIST_HEAD(&priv->bf_list);
+	mutex_init(&priv->bf_mutex);
+
+	dev->rev_id = pdev->revision;
+	dev->numa_node = dev_to_node(&pdev->dev);
+
+	/* Detect if this device is a virtual function */
+	if (pci_dev_data & MLX4_PCI_DEV_IS_VF) {
+		mlx4_warn(dev, "Detected virtual function - running in slave mode\n");
+		dev->flags |= MLX4_FLAG_SLAVE;
+	} else {
+		/* We reset the device and enable SRIOV only for physical
+		 * devices.  Try to claim ownership on the device;
+		 * if already taken, skip -- do not allow multiple PFs */
+		err = mlx4_get_ownership(dev);
+		if (err) {
+			if (err < 0)
+				return err;
+			else {
+				mlx4_warn(dev, "Multiple PFs not yet supported - Skipping PF\n");
+				return -EINVAL;
+			}
+		}
+
+		atomic_set(&priv->opreq_count, 0);
+		INIT_WORK(&priv->opreq_task, mlx4_opreq_action);
+
+		/*
+		 * Now reset the HCA before we touch the PCI capabilities or
+		 * attempt a firmware command, since a boot ROM may have left
+		 * the HCA in an undefined state.
+		 */
+		err = mlx4_reset(dev);
+		if (err) {
+			mlx4_err(dev, "Failed to reset HCA, aborting\n");
+			goto err_sriov;
+		}
+
+		if (total_vfs) {
+			dev->flags = MLX4_FLAG_MASTER;
+			existing_vfs = pci_num_vf(pdev);
+			if (existing_vfs)
+				dev->flags |= MLX4_FLAG_SRIOV;
+			dev->persist->num_vfs = total_vfs;
+		}
+	}
+
+	/* on load remove any previous indication of internal error,
+	 * device is up.
+	 */
+	dev->persist->state = MLX4_DEVICE_STATE_UP;
+
+slave_start:
+	err = mlx4_cmd_init(dev);
+	if (err) {
+		mlx4_err(dev, "Failed to init command interface, aborting\n");
+		goto err_sriov;
+	}
+
+	/* In slave functions, the communication channel must be initialized
+	 * before posting commands. Also, init num_slaves before calling
+	 * mlx4_init_hca */
+	if (mlx4_is_mfunc(dev)) {
+		if (mlx4_is_master(dev)) {
+			dev->num_slaves = MLX4_MAX_NUM_SLAVES;
+
+		} else {
+			dev->num_slaves = 0;
+			err = mlx4_multi_func_init(dev);
+			if (err) {
+				mlx4_err(dev, "Failed to init slave mfunc interface, aborting\n");
+				goto err_cmd;
+			}
+		}
+	}
+
+	err = mlx4_init_fw(dev);
+	if (err) {
+		mlx4_err(dev, "Failed to init fw, aborting.\n");
+		goto err_mfunc;
+	}
+
+	if (mlx4_is_master(dev)) {
+		/* when we hit the goto slave_start below, dev_cap already initialized */
+		if (!dev_cap) {
+			dev_cap = kzalloc(sizeof(*dev_cap), GFP_KERNEL);
+
+			if (!dev_cap) {
+				err = -ENOMEM;
+				goto err_fw;
+			}
+
+			err = mlx4_QUERY_DEV_CAP(dev, dev_cap);
+			if (err) {
+				mlx4_err(dev, "QUERY_DEV_CAP command failed, aborting.\n");
+				goto err_fw;
+			}
+
+			if (mlx4_check_dev_cap(dev, dev_cap, nvfs))
+				goto err_fw;
+
+			if (!(dev_cap->flags2 & MLX4_DEV_CAP_FLAG2_SYS_EQS)) {
+				u64 dev_flags = mlx4_enable_sriov(dev, pdev,
+								  total_vfs,
+								  existing_vfs,
+								  reset_flow);
+
+				mlx4_close_fw(dev);
+				mlx4_cmd_cleanup(dev, MLX4_CMD_CLEANUP_ALL);
+				dev->flags = dev_flags;
+				if (!SRIOV_VALID_STATE(dev->flags)) {
+					mlx4_err(dev, "Invalid SRIOV state\n");
+					goto err_sriov;
+				}
+				err = mlx4_reset(dev);
+				if (err) {
+					mlx4_err(dev, "Failed to reset HCA, aborting.\n");
+					goto err_sriov;
+				}
+				goto slave_start;
+			}
+		} else {
+			/* Legacy mode FW requires SRIOV to be enabled before
+			 * doing QUERY_DEV_CAP, since max_eq's value is different if
+			 * SRIOV is enabled.
+			 */
+			memset(dev_cap, 0, sizeof(*dev_cap));
+			err = mlx4_QUERY_DEV_CAP(dev, dev_cap);
+			if (err) {
+				mlx4_err(dev, "QUERY_DEV_CAP command failed, aborting.\n");
+				goto err_fw;
+			}
+
+			if (mlx4_check_dev_cap(dev, dev_cap, nvfs))
+				goto err_fw;
+		}
+	}
+
+	err = mlx4_init_hca(dev);
+	if (err) {
+		if (err == -EACCES) {
+			/* Not primary Physical function
+			 * Running in slave mode */
+			mlx4_cmd_cleanup(dev, MLX4_CMD_CLEANUP_ALL);
+			/* We're not a PF */
+			if (dev->flags & MLX4_FLAG_SRIOV) {
+				if (!existing_vfs)
+					pci_disable_sriov(pdev);
+				if (mlx4_is_master(dev) && !reset_flow)
+					atomic_dec(&pf_loading);
+				dev->flags &= ~MLX4_FLAG_SRIOV;
+			}
+			if (!mlx4_is_slave(dev))
+				mlx4_free_ownership(dev);
+			dev->flags |= MLX4_FLAG_SLAVE;
+			dev->flags &= ~MLX4_FLAG_MASTER;
+			goto slave_start;
+		} else
+			goto err_fw;
+	}
+
+	if (mlx4_is_master(dev) && (dev_cap->flags2 & MLX4_DEV_CAP_FLAG2_SYS_EQS)) {
+		u64 dev_flags = mlx4_enable_sriov(dev, pdev, total_vfs,
+						  existing_vfs, reset_flow);
+
+		if ((dev->flags ^ dev_flags) & (MLX4_FLAG_MASTER | MLX4_FLAG_SLAVE)) {
+			mlx4_cmd_cleanup(dev, MLX4_CMD_CLEANUP_VHCR);
+			dev->flags = dev_flags;
+			err = mlx4_cmd_init(dev);
+			if (err) {
+				/* Only VHCR is cleaned up, so could still
+				 * send FW commands
+				 */
+				mlx4_err(dev, "Failed to init VHCR command interface, aborting\n");
+				goto err_close;
+			}
+		} else {
+			dev->flags = dev_flags;
+		}
+
+		if (!SRIOV_VALID_STATE(dev->flags)) {
+			mlx4_err(dev, "Invalid SRIOV state\n");
+			err = -EINVAL;
+			goto err_close;
+		}
+	}
+
+	/* check if the device is functioning at its maximum possible speed.
+	 * No return code for this call, just warn the user in case of PCI
+	 * express device capabilities are under-satisfied by the bus.
+	 */
+	if (!mlx4_is_slave(dev))
+		pcie_print_link_status(dev->persist->pdev);
+
+	/* In master functions, the communication channel must be initialized
+	 * after obtaining its address from fw */
+	if (mlx4_is_master(dev)) {
+		if (dev->caps.num_ports < 2 &&
+		    num_vfs_argc > 1) {
+			err = -EINVAL;
+			mlx4_err(dev,
+				 "Error: Trying to configure VFs on port 2, but HCA has only %d physical ports\n",
+				 dev->caps.num_ports);
+			goto err_close;
+		}
+		memcpy(dev->persist->nvfs, nvfs, sizeof(dev->persist->nvfs));
+
+		for (i = 0;
+		     i < sizeof(dev->persist->nvfs)/
+		     sizeof(dev->persist->nvfs[0]); i++) {
+			unsigned j;
+
+			for (j = 0; j < dev->persist->nvfs[i]; ++sum, ++j) {
+				dev->dev_vfs[sum].min_port = i < 2 ? i + 1 : 1;
+				dev->dev_vfs[sum].n_ports = i < 2 ? 1 :
+					dev->caps.num_ports;
+			}
+		}
+
+		/* In master functions, the communication channel
+		 * must be initialized after obtaining its address from fw
+		 */
+		err = mlx4_multi_func_init(dev);
+		if (err) {
+			mlx4_err(dev, "Failed to init master mfunc interface, aborting.\n");
+			goto err_close;
+		}
+	}
+
+	err = mlx4_alloc_eq_table(dev);
+	if (err)
+		goto err_master_mfunc;
+
+	bitmap_zero(priv->msix_ctl.pool_bm, MAX_MSIX);
+	mutex_init(&priv->msix_ctl.pool_lock);
+
+	mlx4_enable_msi_x(dev);
+	if ((mlx4_is_mfunc(dev)) &&
+	    !(dev->flags & MLX4_FLAG_MSI_X)) {
+		err = -EOPNOTSUPP;
+		mlx4_err(dev, "INTx is not supported in multi-function mode, aborting\n");
+		goto err_free_eq;
+	}
+
+	if (!mlx4_is_slave(dev)) {
+		err = mlx4_init_steering(dev);
+		if (err)
+			goto err_disable_msix;
+	}
+
+	mlx4_init_quotas(dev);
+
+	err = mlx4_setup_hca(dev);
+	if (err == -EBUSY && (dev->flags & MLX4_FLAG_MSI_X) &&
+	    !mlx4_is_mfunc(dev)) {
+		dev->flags &= ~MLX4_FLAG_MSI_X;
+		dev->caps.num_comp_vectors = 1;
+		pci_disable_msix(pdev);
+		err = mlx4_setup_hca(dev);
+	}
+
+	if (err)
+		goto err_steer;
+
+	/* When PF resources are ready arm its comm channel to enable
+	 * getting commands
+	 */
+	if (mlx4_is_master(dev)) {
+		err = mlx4_ARM_COMM_CHANNEL(dev);
+		if (err) {
+			mlx4_err(dev, " Failed to arm comm channel eq: %x\n",
+				 err);
+			goto err_steer;
+		}
+	}
+
+	for (port = 1; port <= dev->caps.num_ports; port++) {
+		err = mlx4_init_port_info(dev, port);
+		if (err)
+			goto err_port;
+	}
+
+	priv->v2p.port1 = 1;
+	priv->v2p.port2 = 2;
+
+	err = mlx4_register_device(dev);
+	if (err)
+		goto err_port;
+
+	mlx4_request_modules(dev);
+
+	mlx4_sense_init(dev);
+	mlx4_start_sense(dev);
+
+	priv->removed = 0;
+
+	if (mlx4_is_master(dev) && dev->persist->num_vfs && !reset_flow)
+		atomic_dec(&pf_loading);
+
+	kfree(dev_cap);
+	return 0;
+
+err_port:
+	for (--port; port >= 1; --port)
+		mlx4_cleanup_port_info(&priv->port[port]);
+
+	mlx4_cleanup_default_counters(dev);
+	if (!mlx4_is_slave(dev))
+		mlx4_cleanup_counters_table(dev);
+	mlx4_cleanup_qp_table(dev);
+	mlx4_cleanup_srq_table(dev);
+	mlx4_cleanup_cq_table(dev);
+	mlx4_cmd_use_polling(dev);
+	mlx4_cleanup_eq_table(dev);
+	mlx4_cleanup_mcg_table(dev);
+	mlx4_cleanup_mr_table(dev);
+	mlx4_cleanup_xrcd_table(dev);
+	mlx4_cleanup_pd_table(dev);
+	mlx4_cleanup_uar_table(dev);
+
+err_steer:
+	if (!mlx4_is_slave(dev))
+		mlx4_clear_steering(dev);
+
+err_disable_msix:
+	if (dev->flags & MLX4_FLAG_MSI_X)
+		pci_disable_msix(pdev);
+
+err_free_eq:
+	mlx4_free_eq_table(dev);
+
+err_master_mfunc:
+	if (mlx4_is_master(dev)) {
+		mlx4_free_resource_tracker(dev, RES_TR_FREE_STRUCTS_ONLY);
+		mlx4_multi_func_cleanup(dev);
+	}
+
+	if (mlx4_is_slave(dev))
+		mlx4_slave_destroy_special_qp_cap(dev);
+
+err_close:
+	mlx4_close_hca(dev);
+
+err_fw:
+	mlx4_close_fw(dev);
+
+err_mfunc:
+	if (mlx4_is_slave(dev))
+		mlx4_multi_func_cleanup(dev);
+
+err_cmd:
+	mlx4_cmd_cleanup(dev, MLX4_CMD_CLEANUP_ALL);
+
+err_sriov:
+	if (dev->flags & MLX4_FLAG_SRIOV && !existing_vfs) {
+		pci_disable_sriov(pdev);
+		dev->flags &= ~MLX4_FLAG_SRIOV;
+	}
+
+	if (mlx4_is_master(dev) && dev->persist->num_vfs && !reset_flow)
+		atomic_dec(&pf_loading);
+
+	kfree(priv->dev.dev_vfs);
+
+	if (!mlx4_is_slave(dev))
+		mlx4_free_ownership(dev);
+
+	kfree(dev_cap);
+	return err;
+}
+
+static int __mlx4_init_one(struct pci_dev *pdev, int pci_dev_data,
+			   struct mlx4_priv *priv)
+{
+	int err;
+	int nvfs[MLX4_MAX_PORTS + 1] = {0, 0, 0};
+	int prb_vf[MLX4_MAX_PORTS + 1] = {0, 0, 0};
+	const int param_map[MLX4_MAX_PORTS + 1][MLX4_MAX_PORTS + 1] = {
+		{2, 0, 0}, {0, 1, 2}, {0, 1, 2} };
+	unsigned total_vfs = 0;
+	unsigned int i;
+
+	pr_info(DRV_NAME ": Initializing %s\n", pci_name(pdev));
+
+	err = mlx4_pci_enable_device(&priv->dev);
+	if (err) {
+		dev_err(&pdev->dev, "Cannot enable PCI device, aborting\n");
+		return err;
+	}
+
+	/* Due to requirement that all VFs and the PF are *guaranteed* 2 MACS
+	 * per port, we must limit the number of VFs to 63 (since their are
+	 * 128 MACs)
+	 */
+	for (i = 0; i < ARRAY_SIZE(nvfs) && i < num_vfs_argc;
+	     total_vfs += nvfs[param_map[num_vfs_argc - 1][i]], i++) {
+		nvfs[param_map[num_vfs_argc - 1][i]] = num_vfs[i];
+		if (nvfs[i] < 0) {
+			dev_err(&pdev->dev, "num_vfs module parameter cannot be negative\n");
+			err = -EINVAL;
+			goto err_disable_pdev;
+		}
+	}
+	for (i = 0; i < ARRAY_SIZE(prb_vf) && i < probe_vfs_argc;
+	     i++) {
+		prb_vf[param_map[probe_vfs_argc - 1][i]] = probe_vf[i];
+		if (prb_vf[i] < 0 || prb_vf[i] > nvfs[i]) {
+			dev_err(&pdev->dev, "probe_vf module parameter cannot be negative or greater than num_vfs\n");
+			err = -EINVAL;
+			goto err_disable_pdev;
+		}
+	}
+	if (total_vfs > MLX4_MAX_NUM_VF) {
+		dev_err(&pdev->dev,
+			"Requested more VF's (%d) than allowed by hw (%d)\n",
+			total_vfs, MLX4_MAX_NUM_VF);
+		err = -EINVAL;
+		goto err_disable_pdev;
+	}
+
+	for (i = 0; i < MLX4_MAX_PORTS; i++) {
+		if (nvfs[i] + nvfs[2] > MLX4_MAX_NUM_VF_P_PORT) {
+			dev_err(&pdev->dev,
+				"Requested more VF's (%d) for port (%d) than allowed by driver (%d)\n",
+				nvfs[i] + nvfs[2], i + 1,
+				MLX4_MAX_NUM_VF_P_PORT);
+			err = -EINVAL;
+			goto err_disable_pdev;
+		}
+	}
+
+	/* Check for BARs. */
+	if (!(pci_dev_data & MLX4_PCI_DEV_IS_VF) &&
+	    !(pci_resource_flags(pdev, 0) & IORESOURCE_MEM)) {
+		dev_err(&pdev->dev, "Missing DCS, aborting (driver_data: 0x%x, pci_resource_flags(pdev, 0):0x%lx)\n",
+			pci_dev_data, pci_resource_flags(pdev, 0));
+		err = -ENODEV;
+		goto err_disable_pdev;
+	}
+	if (!(pci_resource_flags(pdev, 2) & IORESOURCE_MEM)) {
+		dev_err(&pdev->dev, "Missing UAR, aborting\n");
+		err = -ENODEV;
+		goto err_disable_pdev;
+	}
+
+	err = pci_request_regions(pdev, DRV_NAME);
+	if (err) {
+		dev_err(&pdev->dev, "Couldn't get PCI resources, aborting\n");
+		goto err_disable_pdev;
+	}
+
+	pci_set_master(pdev);
+
+	err = pci_set_dma_mask(pdev, DMA_BIT_MASK(64));
+	if (err) {
+		dev_warn(&pdev->dev, "Warning: couldn't set 64-bit PCI DMA mask\n");
+		err = pci_set_dma_mask(pdev, DMA_BIT_MASK(32));
+		if (err) {
+			dev_err(&pdev->dev, "Can't set PCI DMA mask, aborting\n");
+			goto err_release_regions;
+		}
+	}
+	err = pci_set_consistent_dma_mask(pdev, DMA_BIT_MASK(64));
+	if (err) {
+		dev_warn(&pdev->dev, "Warning: couldn't set 64-bit consistent PCI DMA mask\n");
+		err = pci_set_consistent_dma_mask(pdev, DMA_BIT_MASK(32));
+		if (err) {
+			dev_err(&pdev->dev, "Can't set consistent PCI DMA mask, aborting\n");
+			goto err_release_regions;
+		}
+	}
+
+	/* Allow large DMA segments, up to the firmware limit of 1 GB */
+	dma_set_max_seg_size(&pdev->dev, 1024 * 1024 * 1024);
+	/* Detect if this device is a virtual function */
+	if (pci_dev_data & MLX4_PCI_DEV_IS_VF) {
+		/* When acting as pf, we normally skip vfs unless explicitly
+		 * requested to probe them.
+		 */
+		if (total_vfs) {
+			unsigned vfs_offset = 0;
+
+			for (i = 0; i < ARRAY_SIZE(nvfs) &&
+			     vfs_offset + nvfs[i] < extended_func_num(pdev);
+			     vfs_offset += nvfs[i], i++)
+				;
+			if (i == ARRAY_SIZE(nvfs)) {
+				err = -ENODEV;
+				goto err_release_regions;
+			}
+			if ((extended_func_num(pdev) - vfs_offset)
+			    > prb_vf[i]) {
+				dev_warn(&pdev->dev, "Skipping virtual function:%d\n",
+					 extended_func_num(pdev));
+				err = -ENODEV;
+				goto err_release_regions;
+			}
+		}
+	}
+
+	err = mlx4_crdump_init(&priv->dev);
+	if (err)
+		goto err_release_regions;
+
+	err = mlx4_catas_init(&priv->dev);
+	if (err)
+		goto err_crdump;
+
+	err = mlx4_load_one(pdev, pci_dev_data, total_vfs, nvfs, priv, 0);
+	if (err)
+		goto err_catas;
+
+	return 0;
+
+err_catas:
+	mlx4_catas_end(&priv->dev);
+
+err_crdump:
+	mlx4_crdump_end(&priv->dev);
+
+err_release_regions:
+	pci_release_regions(pdev);
+
+err_disable_pdev:
+	mlx4_pci_disable_device(&priv->dev);
+	return err;
+}
+
+static int mlx4_devlink_port_type_set(struct devlink_port *devlink_port,
+				      enum devlink_port_type port_type)
+{
+	struct mlx4_port_info *info = container_of(devlink_port,
+						   struct mlx4_port_info,
+						   devlink_port);
+	enum mlx4_port_type mlx4_port_type;
+
+	switch (port_type) {
+	case DEVLINK_PORT_TYPE_AUTO:
+		mlx4_port_type = MLX4_PORT_TYPE_AUTO;
+		break;
+	case DEVLINK_PORT_TYPE_ETH:
+		mlx4_port_type = MLX4_PORT_TYPE_ETH;
+		break;
+	case DEVLINK_PORT_TYPE_IB:
+		mlx4_port_type = MLX4_PORT_TYPE_IB;
+		break;
+	default:
+		return -EOPNOTSUPP;
+	}
+
+	return __set_port_type(info, mlx4_port_type);
+}
+
+static void mlx4_devlink_param_load_driverinit_values(struct devlink *devlink)
+{
+	struct mlx4_priv *priv = devlink_priv(devlink);
+	struct mlx4_dev *dev = &priv->dev;
+	struct mlx4_fw_crdump *crdump = &dev->persist->crdump;
+	union devlink_param_value saved_value;
+	int err;
+
+	err = devlink_param_driverinit_value_get(devlink,
+						 DEVLINK_PARAM_GENERIC_ID_INT_ERR_RESET,
+						 &saved_value);
+	if (!err && mlx4_internal_err_reset != saved_value.vbool) {
+		mlx4_internal_err_reset = saved_value.vbool;
+		/* Notify on value changed on runtime configuration mode */
+		devlink_param_value_changed(devlink,
+					    DEVLINK_PARAM_GENERIC_ID_INT_ERR_RESET);
+	}
+	err = devlink_param_driverinit_value_get(devlink,
+						 DEVLINK_PARAM_GENERIC_ID_MAX_MACS,
+						 &saved_value);
+	if (!err)
+		log_num_mac = order_base_2(saved_value.vu32);
+	err = devlink_param_driverinit_value_get(devlink,
+						 MLX4_DEVLINK_PARAM_ID_ENABLE_64B_CQE_EQE,
+						 &saved_value);
+	if (!err)
+		enable_64b_cqe_eqe = saved_value.vbool;
+	err = devlink_param_driverinit_value_get(devlink,
+						 MLX4_DEVLINK_PARAM_ID_ENABLE_4K_UAR,
+						 &saved_value);
+	if (!err)
+		enable_4k_uar = saved_value.vbool;
+	err = devlink_param_driverinit_value_get(devlink,
+						 DEVLINK_PARAM_GENERIC_ID_REGION_SNAPSHOT,
+						 &saved_value);
+	if (!err && crdump->snapshot_enable != saved_value.vbool) {
+		crdump->snapshot_enable = saved_value.vbool;
+		devlink_param_value_changed(devlink,
+					    DEVLINK_PARAM_GENERIC_ID_REGION_SNAPSHOT);
+	}
+}
+
+static int mlx4_devlink_reload(struct devlink *devlink,
+			       struct netlink_ext_ack *extack)
+{
+	struct mlx4_priv *priv = devlink_priv(devlink);
+	struct mlx4_dev *dev = &priv->dev;
+	struct mlx4_dev_persistent *persist = dev->persist;
+	int err;
+
+	if (persist->num_vfs)
+		mlx4_warn(persist->dev, "Reload performed on PF, will cause reset on operating Virtual Functions\n");
+	err = mlx4_restart_one(persist->pdev, true, devlink);
+	if (err)
+		mlx4_err(persist->dev, "mlx4_restart_one failed, ret=%d\n", err);
+
+	return err;
+}
+
+static const struct devlink_ops mlx4_devlink_ops = {
+	.port_type_set	= mlx4_devlink_port_type_set,
+	.reload		= mlx4_devlink_reload,
+};
+
+static int mlx4_init_one(struct pci_dev *pdev, const struct pci_device_id *id)
+{
+	struct devlink *devlink;
+	struct mlx4_priv *priv;
+	struct mlx4_dev *dev;
+	int ret;
+
+	printk_once(KERN_INFO "%s", mlx4_version);
+
+	devlink = devlink_alloc(&mlx4_devlink_ops, sizeof(*priv));
+	if (!devlink)
+		return -ENOMEM;
+	priv = devlink_priv(devlink);
+
+	dev       = &priv->dev;
+	dev->persist = kzalloc(sizeof(*dev->persist), GFP_KERNEL);
+	if (!dev->persist) {
+		ret = -ENOMEM;
+		goto err_devlink_free;
+	}
+	dev->persist->pdev = pdev;
+	dev->persist->dev = dev;
+	pci_set_drvdata(pdev, dev->persist);
+	priv->pci_dev_data = id->driver_data;
+	mutex_init(&dev->persist->device_state_mutex);
+	mutex_init(&dev->persist->interface_state_mutex);
+	mutex_init(&dev->persist->pci_status_mutex);
+
+	ret = devlink_register(devlink, &pdev->dev);
+	if (ret)
+		goto err_persist_free;
+	ret = devlink_params_register(devlink, mlx4_devlink_params,
+				      ARRAY_SIZE(mlx4_devlink_params));
+	if (ret)
+		goto err_devlink_unregister;
+	mlx4_devlink_set_params_init_values(devlink);
+	ret =  __mlx4_init_one(pdev, id->driver_data, priv);
+	if (ret)
+		goto err_params_unregister;
+
+	pci_save_state(pdev);
+	return 0;
+
+err_params_unregister:
+	devlink_params_unregister(devlink, mlx4_devlink_params,
+				  ARRAY_SIZE(mlx4_devlink_params));
+err_devlink_unregister:
+	devlink_unregister(devlink);
+err_persist_free:
+	kfree(dev->persist);
+err_devlink_free:
+	devlink_free(devlink);
+	return ret;
+}
+
+static void mlx4_clean_dev(struct mlx4_dev *dev)
+{
+	struct mlx4_dev_persistent *persist = dev->persist;
+	struct mlx4_priv *priv = mlx4_priv(dev);
+	unsigned long	flags = (dev->flags & RESET_PERSIST_MASK_FLAGS);
+
+	memset(priv, 0, sizeof(*priv));
+	priv->dev.persist = persist;
+	priv->dev.flags = flags;
+}
+
+static void mlx4_unload_one(struct pci_dev *pdev)
+{
+	struct mlx4_dev_persistent *persist = pci_get_drvdata(pdev);
+	struct mlx4_dev  *dev  = persist->dev;
+	struct mlx4_priv *priv = mlx4_priv(dev);
+	int               pci_dev_data;
+	int p, i;
+
+	if (priv->removed)
+		return;
+
+	/* saving current ports type for further use */
+	for (i = 0; i < dev->caps.num_ports; i++) {
+		dev->persist->curr_port_type[i] = dev->caps.port_type[i + 1];
+		dev->persist->curr_port_poss_type[i] = dev->caps.
+						       possible_type[i + 1];
+	}
+
+	pci_dev_data = priv->pci_dev_data;
+
+	mlx4_stop_sense(dev);
+	mlx4_unregister_device(dev);
+
+	for (p = 1; p <= dev->caps.num_ports; p++) {
+		mlx4_cleanup_port_info(&priv->port[p]);
+		mlx4_CLOSE_PORT(dev, p);
+	}
+
+	if (mlx4_is_master(dev))
+		mlx4_free_resource_tracker(dev,
+					   RES_TR_FREE_SLAVES_ONLY);
+
+	mlx4_cleanup_default_counters(dev);
+	if (!mlx4_is_slave(dev))
+		mlx4_cleanup_counters_table(dev);
+	mlx4_cleanup_qp_table(dev);
+	mlx4_cleanup_srq_table(dev);
+	mlx4_cleanup_cq_table(dev);
+	mlx4_cmd_use_polling(dev);
+	mlx4_cleanup_eq_table(dev);
+	mlx4_cleanup_mcg_table(dev);
+	mlx4_cleanup_mr_table(dev);
+	mlx4_cleanup_xrcd_table(dev);
+	mlx4_cleanup_pd_table(dev);
+
+	if (mlx4_is_master(dev))
+		mlx4_free_resource_tracker(dev,
+					   RES_TR_FREE_STRUCTS_ONLY);
+
+	iounmap(priv->kar);
+	mlx4_uar_free(dev, &priv->driver_uar);
+	mlx4_cleanup_uar_table(dev);
+	if (!mlx4_is_slave(dev))
+		mlx4_clear_steering(dev);
+	mlx4_free_eq_table(dev);
+	if (mlx4_is_master(dev))
+		mlx4_multi_func_cleanup(dev);
+	mlx4_close_hca(dev);
+	mlx4_close_fw(dev);
+	if (mlx4_is_slave(dev))
+		mlx4_multi_func_cleanup(dev);
+	mlx4_cmd_cleanup(dev, MLX4_CMD_CLEANUP_ALL);
+
+	if (dev->flags & MLX4_FLAG_MSI_X)
+		pci_disable_msix(pdev);
+
+	if (!mlx4_is_slave(dev))
+		mlx4_free_ownership(dev);
+
+	mlx4_slave_destroy_special_qp_cap(dev);
+	kfree(dev->dev_vfs);
+
+	mlx4_clean_dev(dev);
+	priv->pci_dev_data = pci_dev_data;
+	priv->removed = 1;
+}
+
+static void mlx4_remove_one(struct pci_dev *pdev)
+{
+	struct mlx4_dev_persistent *persist = pci_get_drvdata(pdev);
+	struct mlx4_dev  *dev  = persist->dev;
+	struct mlx4_priv *priv = mlx4_priv(dev);
+	struct devlink *devlink = priv_to_devlink(priv);
+	int active_vfs = 0;
+
+	if (mlx4_is_slave(dev))
+		persist->interface_state |= MLX4_INTERFACE_STATE_NOWAIT;
+
+	mutex_lock(&persist->interface_state_mutex);
+	persist->interface_state |= MLX4_INTERFACE_STATE_DELETION;
+	mutex_unlock(&persist->interface_state_mutex);
+
+	/* Disabling SR-IOV is not allowed while there are active vf's */
+	if (mlx4_is_master(dev) && dev->flags & MLX4_FLAG_SRIOV) {
+		active_vfs = mlx4_how_many_lives_vf(dev);
+		if (active_vfs) {
+			pr_warn("Removing PF when there are active VF's !!\n");
+			pr_warn("Will not disable SR-IOV.\n");
+		}
+	}
+
+	/* device marked to be under deletion running now without the lock
+	 * letting other tasks to be terminated
+	 */
+	if (persist->interface_state & MLX4_INTERFACE_STATE_UP)
+		mlx4_unload_one(pdev);
+	else
+		mlx4_info(dev, "%s: interface is down\n", __func__);
+	mlx4_catas_end(dev);
+	mlx4_crdump_end(dev);
+	if (dev->flags & MLX4_FLAG_SRIOV && !active_vfs) {
+		mlx4_warn(dev, "Disabling SR-IOV\n");
+		pci_disable_sriov(pdev);
+	}
+
+	pci_release_regions(pdev);
+	mlx4_pci_disable_device(dev);
+	devlink_params_unregister(devlink, mlx4_devlink_params,
+				  ARRAY_SIZE(mlx4_devlink_params));
+	devlink_unregister(devlink);
+	kfree(dev->persist);
+	devlink_free(devlink);
+}
+
+static int restore_current_port_types(struct mlx4_dev *dev,
+				      enum mlx4_port_type *types,
+				      enum mlx4_port_type *poss_types)
+{
+	struct mlx4_priv *priv = mlx4_priv(dev);
+	int err, i;
+
+	mlx4_stop_sense(dev);
+
+	mutex_lock(&priv->port_mutex);
+	for (i = 0; i < dev->caps.num_ports; i++)
+		dev->caps.possible_type[i + 1] = poss_types[i];
+	err = mlx4_change_port_types(dev, types);
+	mlx4_start_sense(dev);
+	mutex_unlock(&priv->port_mutex);
+
+	return err;
+}
+
+int mlx4_restart_one(struct pci_dev *pdev, bool reload, struct devlink *devlink)
+{
+	struct mlx4_dev_persistent *persist = pci_get_drvdata(pdev);
+	struct mlx4_dev	 *dev  = persist->dev;
+	struct mlx4_priv *priv = mlx4_priv(dev);
+	int nvfs[MLX4_MAX_PORTS + 1] = {0, 0, 0};
+	int pci_dev_data, err, total_vfs;
+
+	pci_dev_data = priv->pci_dev_data;
+	total_vfs = dev->persist->num_vfs;
+	memcpy(nvfs, dev->persist->nvfs, sizeof(dev->persist->nvfs));
+
+	mlx4_unload_one(pdev);
+	if (reload)
+		mlx4_devlink_param_load_driverinit_values(devlink);
+	err = mlx4_load_one(pdev, pci_dev_data, total_vfs, nvfs, priv, 1);
+	if (err) {
+		mlx4_err(dev, "%s: ERROR: mlx4_load_one failed, pci_name=%s, err=%d\n",
+			 __func__, pci_name(pdev), err);
+		return err;
+	}
+
+	err = restore_current_port_types(dev, dev->persist->curr_port_type,
+					 dev->persist->curr_port_poss_type);
+	if (err)
+		mlx4_err(dev, "could not restore original port types (%d)\n",
+			 err);
+
+	return err;
+}
+
+#define MLX_SP(id) { PCI_VDEVICE(MELLANOX, id), MLX4_PCI_DEV_FORCE_SENSE_PORT }
+#define MLX_VF(id) { PCI_VDEVICE(MELLANOX, id), MLX4_PCI_DEV_IS_VF }
+#define MLX_GN(id) { PCI_VDEVICE(MELLANOX, id), 0 }
+
+static const struct pci_device_id mlx4_pci_table[] = {
+#ifdef CONFIG_MLX4_CORE_GEN2
+	/* MT25408 "Hermon" */
+	MLX_SP(PCI_DEVICE_ID_MELLANOX_HERMON_SDR),	/* SDR */
+	MLX_SP(PCI_DEVICE_ID_MELLANOX_HERMON_DDR),	/* DDR */
+	MLX_SP(PCI_DEVICE_ID_MELLANOX_HERMON_QDR),	/* QDR */
+	MLX_SP(PCI_DEVICE_ID_MELLANOX_HERMON_DDR_GEN2), /* DDR Gen2 */
+	MLX_SP(PCI_DEVICE_ID_MELLANOX_HERMON_QDR_GEN2),	/* QDR Gen2 */
+	MLX_SP(PCI_DEVICE_ID_MELLANOX_HERMON_EN),	/* EN 10GigE */
+	MLX_SP(PCI_DEVICE_ID_MELLANOX_HERMON_EN_GEN2),  /* EN 10GigE Gen2 */
+	/* MT25458 ConnectX EN 10GBASE-T */
+	MLX_SP(PCI_DEVICE_ID_MELLANOX_CONNECTX_EN),
+	MLX_SP(PCI_DEVICE_ID_MELLANOX_CONNECTX_EN_T_GEN2),	/* Gen2 */
+	/* MT26468 ConnectX EN 10GigE PCIe Gen2*/
+	MLX_SP(PCI_DEVICE_ID_MELLANOX_CONNECTX_EN_GEN2),
+	/* MT26438 ConnectX EN 40GigE PCIe Gen2 5GT/s */
+	MLX_SP(PCI_DEVICE_ID_MELLANOX_CONNECTX_EN_5_GEN2),
+	/* MT26478 ConnectX2 40GigE PCIe Gen2 */
+	MLX_SP(PCI_DEVICE_ID_MELLANOX_CONNECTX2),
+	/* MT25400 Family [ConnectX-2] */
+	MLX_VF(0x1002),					/* Virtual Function */
+#endif /* CONFIG_MLX4_CORE_GEN2 */
+	/* MT27500 Family [ConnectX-3] */
+	MLX_GN(PCI_DEVICE_ID_MELLANOX_CONNECTX3),
+	MLX_VF(0x1004),					/* Virtual Function */
+	MLX_GN(0x1005),					/* MT27510 Family */
+	MLX_GN(0x1006),					/* MT27511 Family */
+	MLX_GN(PCI_DEVICE_ID_MELLANOX_CONNECTX3_PRO),	/* MT27520 Family */
+	MLX_GN(0x1008),					/* MT27521 Family */
+	MLX_GN(0x1009),					/* MT27530 Family */
+	MLX_GN(0x100a),					/* MT27531 Family */
+	MLX_GN(0x100b),					/* MT27540 Family */
+	MLX_GN(0x100c),					/* MT27541 Family */
+	MLX_GN(0x100d),					/* MT27550 Family */
+	MLX_GN(0x100e),					/* MT27551 Family */
+	MLX_GN(0x100f),					/* MT27560 Family */
+	MLX_GN(0x1010),					/* MT27561 Family */
+
+	/*
+	 * See the mellanox_check_broken_intx_masking() quirk when
+	 * adding devices
+	 */
+
+	{ 0, }
+};
+
+MODULE_DEVICE_TABLE(pci, mlx4_pci_table);
+
+static pci_ers_result_t mlx4_pci_err_detected(struct pci_dev *pdev,
+					      pci_channel_state_t state)
+{
+	struct mlx4_dev_persistent *persist = pci_get_drvdata(pdev);
+
+	mlx4_err(persist->dev, "mlx4_pci_err_detected was called\n");
+	mlx4_enter_error_state(persist);
+
+	mutex_lock(&persist->interface_state_mutex);
+	if (persist->interface_state & MLX4_INTERFACE_STATE_UP)
+		mlx4_unload_one(pdev);
+
+	mutex_unlock(&persist->interface_state_mutex);
+	if (state == pci_channel_io_perm_failure)
+		return PCI_ERS_RESULT_DISCONNECT;
+
+	mlx4_pci_disable_device(persist->dev);
+	return PCI_ERS_RESULT_NEED_RESET;
+}
+
+static pci_ers_result_t mlx4_pci_slot_reset(struct pci_dev *pdev)
+{
+	struct mlx4_dev_persistent *persist = pci_get_drvdata(pdev);
+	struct mlx4_dev	 *dev  = persist->dev;
+	int err;
+
+	mlx4_err(dev, "mlx4_pci_slot_reset was called\n");
+	err = mlx4_pci_enable_device(dev);
+	if (err) {
+		mlx4_err(dev, "Can not re-enable device, err=%d\n", err);
+		return PCI_ERS_RESULT_DISCONNECT;
+	}
+
+	pci_set_master(pdev);
+	pci_restore_state(pdev);
+	pci_save_state(pdev);
+	return PCI_ERS_RESULT_RECOVERED;
+}
+
+static void mlx4_pci_resume(struct pci_dev *pdev)
+{
+	struct mlx4_dev_persistent *persist = pci_get_drvdata(pdev);
+	struct mlx4_dev	 *dev  = persist->dev;
+	struct mlx4_priv *priv = mlx4_priv(dev);
+	int nvfs[MLX4_MAX_PORTS + 1] = {0, 0, 0};
+	int total_vfs;
+	int err;
+
+	mlx4_err(dev, "%s was called\n", __func__);
+	total_vfs = dev->persist->num_vfs;
+	memcpy(nvfs, dev->persist->nvfs, sizeof(dev->persist->nvfs));
+
+	mutex_lock(&persist->interface_state_mutex);
+	if (!(persist->interface_state & MLX4_INTERFACE_STATE_UP)) {
+		err = mlx4_load_one(pdev, priv->pci_dev_data, total_vfs, nvfs,
+				    priv, 1);
+		if (err) {
+			mlx4_err(dev, "%s: mlx4_load_one failed, err=%d\n",
+				 __func__,  err);
+			goto end;
+		}
+
+		err = restore_current_port_types(dev, dev->persist->
+						 curr_port_type, dev->persist->
+						 curr_port_poss_type);
+		if (err)
+			mlx4_err(dev, "could not restore original port types (%d)\n", err);
+	}
+end:
+	mutex_unlock(&persist->interface_state_mutex);
+
+}
+
+static void mlx4_shutdown(struct pci_dev *pdev)
+{
+	struct mlx4_dev_persistent *persist = pci_get_drvdata(pdev);
+	struct mlx4_dev *dev = persist->dev;
+
+	mlx4_info(persist->dev, "mlx4_shutdown was called\n");
+	mutex_lock(&persist->interface_state_mutex);
+	if (persist->interface_state & MLX4_INTERFACE_STATE_UP)
+		mlx4_unload_one(pdev);
+	mutex_unlock(&persist->interface_state_mutex);
+	mlx4_pci_disable_device(dev);
+}
+
+static const struct pci_error_handlers mlx4_err_handler = {
+	.error_detected = mlx4_pci_err_detected,
+	.slot_reset     = mlx4_pci_slot_reset,
+	.resume		= mlx4_pci_resume,
+};
+
+static int mlx4_suspend(struct pci_dev *pdev, pm_message_t state)
+{
+	struct mlx4_dev_persistent *persist = pci_get_drvdata(pdev);
+	struct mlx4_dev	*dev = persist->dev;
+
+	mlx4_err(dev, "suspend was called\n");
+	mutex_lock(&persist->interface_state_mutex);
+	if (persist->interface_state & MLX4_INTERFACE_STATE_UP)
+		mlx4_unload_one(pdev);
+	mutex_unlock(&persist->interface_state_mutex);
+
+	return 0;
+}
+
+static int mlx4_resume(struct pci_dev *pdev)
+{
+	struct mlx4_dev_persistent *persist = pci_get_drvdata(pdev);
+	struct mlx4_dev	*dev = persist->dev;
+	struct mlx4_priv *priv = mlx4_priv(dev);
+	int nvfs[MLX4_MAX_PORTS + 1] = {0, 0, 0};
+	int total_vfs;
+	int ret = 0;
+
+	mlx4_err(dev, "resume was called\n");
+	total_vfs = dev->persist->num_vfs;
+	memcpy(nvfs, dev->persist->nvfs, sizeof(dev->persist->nvfs));
+
+	mutex_lock(&persist->interface_state_mutex);
+	if (!(persist->interface_state & MLX4_INTERFACE_STATE_UP)) {
+		ret = mlx4_load_one(pdev, priv->pci_dev_data, total_vfs,
+				    nvfs, priv, 1);
+		if (!ret) {
+			ret = restore_current_port_types(dev,
+					dev->persist->curr_port_type,
+					dev->persist->curr_port_poss_type);
+			if (ret)
+				mlx4_err(dev, "resume: could not restore original port types (%d)\n", ret);
+		}
+	}
+	mutex_unlock(&persist->interface_state_mutex);
+
+	return ret;
+}
+
+static struct pci_driver mlx4_driver = {
+	.name		= DRV_NAME,
+	.id_table	= mlx4_pci_table,
+	.probe		= mlx4_init_one,
+	.shutdown	= mlx4_shutdown,
+	.remove		= mlx4_remove_one,
+	.suspend	= mlx4_suspend,
+	.resume		= mlx4_resume,
+	.err_handler    = &mlx4_err_handler,
+};
+
+static int __init mlx4_verify_params(void)
+{
+	if (msi_x < 0) {
+		pr_warn("mlx4_core: bad msi_x: %d\n", msi_x);
+		return -1;
+	}
+
+	if ((log_num_mac < 0) || (log_num_mac > 7)) {
+		pr_warn("mlx4_core: bad num_mac: %d\n", log_num_mac);
+		return -1;
+	}
+
+	if (log_num_vlan != 0)
+		pr_warn("mlx4_core: log_num_vlan - obsolete module param, using %d\n",
+			MLX4_LOG_NUM_VLANS);
+
+	if (use_prio != 0)
+		pr_warn("mlx4_core: use_prio - obsolete module param, ignored\n");
+
+	if ((log_mtts_per_seg < 0) || (log_mtts_per_seg > 7)) {
+		pr_warn("mlx4_core: bad log_mtts_per_seg: %d\n",
+			log_mtts_per_seg);
+		return -1;
+	}
+
+	/* Check if module param for ports type has legal combination */
+	if (port_type_array[0] == false && port_type_array[1] == true) {
+		pr_warn("Module parameter configuration ETH/IB is not supported. Switching to default configuration IB/IB\n");
+		port_type_array[0] = true;
+	}
+
+	if (mlx4_log_num_mgm_entry_size < -7 ||
+	    (mlx4_log_num_mgm_entry_size > 0 &&
+	     (mlx4_log_num_mgm_entry_size < MLX4_MIN_MGM_LOG_ENTRY_SIZE ||
+	      mlx4_log_num_mgm_entry_size > MLX4_MAX_MGM_LOG_ENTRY_SIZE))) {
+		pr_warn("mlx4_core: mlx4_log_num_mgm_entry_size (%d) not in legal range (-7..0 or %d..%d)\n",
+			mlx4_log_num_mgm_entry_size,
+			MLX4_MIN_MGM_LOG_ENTRY_SIZE,
+			MLX4_MAX_MGM_LOG_ENTRY_SIZE);
+		return -1;
+	}
+
+	return 0;
+}
+
+static int __init mlx4_init(void)
+{
+	int ret;
+
+	if (mlx4_verify_params())
+		return -EINVAL;
+
+
+	mlx4_wq = create_singlethread_workqueue("mlx4");
+	if (!mlx4_wq)
+		return -ENOMEM;
+
+	ret = pci_register_driver(&mlx4_driver);
+	if (ret < 0)
+		destroy_workqueue(mlx4_wq);
+	return ret < 0 ? ret : 0;
+}
+
+static void __exit mlx4_cleanup(void)
+{
+	pci_unregister_driver(&mlx4_driver);
+	destroy_workqueue(mlx4_wq);
+}
+
+module_init(mlx4_init);
+module_exit(mlx4_cleanup);
diff --git a/drivers/net/ethernet/mellanox/mlx4/mcg.c b/drivers/net/ethernet/mellanox/mlx4/mcg.c
new file mode 100644
index 000000000..9c481823b
--- /dev/null
+++ b/drivers/net/ethernet/mellanox/mlx4/mcg.c
@@ -0,0 +1,1649 @@
+/*
+ * Copyright (c) 2006, 2007 Cisco Systems, Inc.  All rights reserved.
+ * Copyright (c) 2007, 2008 Mellanox Technologies. All rights reserved.
+ *
+ * This software is available to you under a choice of one of two
+ * licenses.  You may choose to be licensed under the terms of the GNU
+ * General Public License (GPL) Version 2, available from the file
+ * COPYING in the main directory of this source tree, or the
+ * OpenIB.org BSD license below:
+ *
+ *     Redistribution and use in source and binary forms, with or
+ *     without modification, are permitted provided that the following
+ *     conditions are met:
+ *
+ *      - Redistributions of source code must retain the above
+ *        copyright notice, this list of conditions and the following
+ *        disclaimer.
+ *
+ *      - Redistributions in binary form must reproduce the above
+ *        copyright notice, this list of conditions and the following
+ *        disclaimer in the documentation and/or other materials
+ *        provided with the distribution.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
+ * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
+ * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS
+ * BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN
+ * ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
+ * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+ * SOFTWARE.
+ */
+
+#include <linux/string.h>
+#include <linux/etherdevice.h>
+
+#include <linux/mlx4/cmd.h>
+#include <linux/mlx4/qp.h>
+#include <linux/export.h>
+
+#include "mlx4.h"
+
+int mlx4_get_mgm_entry_size(struct mlx4_dev *dev)
+{
+	return 1 << dev->oper_log_mgm_entry_size;
+}
+
+int mlx4_get_qp_per_mgm(struct mlx4_dev *dev)
+{
+	return 4 * (mlx4_get_mgm_entry_size(dev) / 16 - 2);
+}
+
+static int mlx4_QP_FLOW_STEERING_ATTACH(struct mlx4_dev *dev,
+					struct mlx4_cmd_mailbox *mailbox,
+					u32 size,
+					u64 *reg_id)
+{
+	u64 imm;
+	int err = 0;
+
+	err = mlx4_cmd_imm(dev, mailbox->dma, &imm, size, 0,
+			   MLX4_QP_FLOW_STEERING_ATTACH, MLX4_CMD_TIME_CLASS_A,
+			   MLX4_CMD_NATIVE);
+	if (err)
+		return err;
+	*reg_id = imm;
+
+	return err;
+}
+
+static int mlx4_QP_FLOW_STEERING_DETACH(struct mlx4_dev *dev, u64 regid)
+{
+	int err = 0;
+
+	err = mlx4_cmd(dev, regid, 0, 0,
+		       MLX4_QP_FLOW_STEERING_DETACH, MLX4_CMD_TIME_CLASS_A,
+		       MLX4_CMD_NATIVE);
+
+	return err;
+}
+
+static int mlx4_READ_ENTRY(struct mlx4_dev *dev, int index,
+			   struct mlx4_cmd_mailbox *mailbox)
+{
+	return mlx4_cmd_box(dev, 0, mailbox->dma, index, 0, MLX4_CMD_READ_MCG,
+			    MLX4_CMD_TIME_CLASS_A, MLX4_CMD_NATIVE);
+}
+
+static int mlx4_WRITE_ENTRY(struct mlx4_dev *dev, int index,
+			    struct mlx4_cmd_mailbox *mailbox)
+{
+	return mlx4_cmd(dev, mailbox->dma, index, 0, MLX4_CMD_WRITE_MCG,
+			MLX4_CMD_TIME_CLASS_A, MLX4_CMD_NATIVE);
+}
+
+static int mlx4_WRITE_PROMISC(struct mlx4_dev *dev, u8 port, u8 steer,
+			      struct mlx4_cmd_mailbox *mailbox)
+{
+	u32 in_mod;
+
+	in_mod = (u32) port << 16 | steer << 1;
+	return mlx4_cmd(dev, mailbox->dma, in_mod, 0x1,
+			MLX4_CMD_WRITE_MCG, MLX4_CMD_TIME_CLASS_A,
+			MLX4_CMD_NATIVE);
+}
+
+static int mlx4_GID_HASH(struct mlx4_dev *dev, struct mlx4_cmd_mailbox *mailbox,
+			 u16 *hash, u8 op_mod)
+{
+	u64 imm;
+	int err;
+
+	err = mlx4_cmd_imm(dev, mailbox->dma, &imm, 0, op_mod,
+			   MLX4_CMD_MGID_HASH, MLX4_CMD_TIME_CLASS_A,
+			   MLX4_CMD_NATIVE);
+
+	if (!err)
+		*hash = imm;
+
+	return err;
+}
+
+static struct mlx4_promisc_qp *get_promisc_qp(struct mlx4_dev *dev, u8 port,
+					      enum mlx4_steer_type steer,
+					      u32 qpn)
+{
+	struct mlx4_steer *s_steer;
+	struct mlx4_promisc_qp *pqp;
+
+	if (port < 1 || port > dev->caps.num_ports)
+		return NULL;
+
+	s_steer = &mlx4_priv(dev)->steer[port - 1];
+
+	list_for_each_entry(pqp, &s_steer->promisc_qps[steer], list) {
+		if (pqp->qpn == qpn)
+			return pqp;
+	}
+	/* not found */
+	return NULL;
+}
+
+/*
+ * Add new entry to steering data structure.
+ * All promisc QPs should be added as well
+ */
+static int new_steering_entry(struct mlx4_dev *dev, u8 port,
+			      enum mlx4_steer_type steer,
+			      unsigned int index, u32 qpn)
+{
+	struct mlx4_steer *s_steer;
+	struct mlx4_cmd_mailbox *mailbox;
+	struct mlx4_mgm *mgm;
+	u32 members_count;
+	struct mlx4_steer_index *new_entry;
+	struct mlx4_promisc_qp *pqp;
+	struct mlx4_promisc_qp *dqp = NULL;
+	u32 prot;
+	int err;
+
+	if (port < 1 || port > dev->caps.num_ports)
+		return -EINVAL;
+
+	s_steer = &mlx4_priv(dev)->steer[port - 1];
+	new_entry = kzalloc(sizeof(*new_entry), GFP_KERNEL);
+	if (!new_entry)
+		return -ENOMEM;
+
+	INIT_LIST_HEAD(&new_entry->duplicates);
+	new_entry->index = index;
+	list_add_tail(&new_entry->list, &s_steer->steer_entries[steer]);
+
+	/* If the given qpn is also a promisc qp,
+	 * it should be inserted to duplicates list
+	 */
+	pqp = get_promisc_qp(dev, port, steer, qpn);
+	if (pqp) {
+		dqp = kmalloc(sizeof(*dqp), GFP_KERNEL);
+		if (!dqp) {
+			err = -ENOMEM;
+			goto out_alloc;
+		}
+		dqp->qpn = qpn;
+		list_add_tail(&dqp->list, &new_entry->duplicates);
+	}
+
+	/* if no promisc qps for this vep, we are done */
+	if (list_empty(&s_steer->promisc_qps[steer]))
+		return 0;
+
+	/* now need to add all the promisc qps to the new
+	 * steering entry, as they should also receive the packets
+	 * destined to this address */
+	mailbox = mlx4_alloc_cmd_mailbox(dev);
+	if (IS_ERR(mailbox)) {
+		err = -ENOMEM;
+		goto out_alloc;
+	}
+	mgm = mailbox->buf;
+
+	err = mlx4_READ_ENTRY(dev, index, mailbox);
+	if (err)
+		goto out_mailbox;
+
+	members_count = be32_to_cpu(mgm->members_count) & 0xffffff;
+	prot = be32_to_cpu(mgm->members_count) >> 30;
+	list_for_each_entry(pqp, &s_steer->promisc_qps[steer], list) {
+		/* don't add already existing qpn */
+		if (pqp->qpn == qpn)
+			continue;
+		if (members_count == dev->caps.num_qp_per_mgm) {
+			/* out of space */
+			err = -ENOMEM;
+			goto out_mailbox;
+		}
+
+		/* add the qpn */
+		mgm->qp[members_count++] = cpu_to_be32(pqp->qpn & MGM_QPN_MASK);
+	}
+	/* update the qps count and update the entry with all the promisc qps*/
+	mgm->members_count = cpu_to_be32(members_count | (prot << 30));
+	err = mlx4_WRITE_ENTRY(dev, index, mailbox);
+
+out_mailbox:
+	mlx4_free_cmd_mailbox(dev, mailbox);
+	if (!err)
+		return 0;
+out_alloc:
+	if (dqp) {
+		list_del(&dqp->list);
+		kfree(dqp);
+	}
+	list_del(&new_entry->list);
+	kfree(new_entry);
+	return err;
+}
+
+/* update the data structures with existing steering entry */
+static int existing_steering_entry(struct mlx4_dev *dev, u8 port,
+				   enum mlx4_steer_type steer,
+				   unsigned int index, u32 qpn)
+{
+	struct mlx4_steer *s_steer;
+	struct mlx4_steer_index *tmp_entry, *entry = NULL;
+	struct mlx4_promisc_qp *pqp;
+	struct mlx4_promisc_qp *dqp;
+
+	if (port < 1 || port > dev->caps.num_ports)
+		return -EINVAL;
+
+	s_steer = &mlx4_priv(dev)->steer[port - 1];
+
+	pqp = get_promisc_qp(dev, port, steer, qpn);
+	if (!pqp)
+		return 0; /* nothing to do */
+
+	list_for_each_entry(tmp_entry, &s_steer->steer_entries[steer], list) {
+		if (tmp_entry->index == index) {
+			entry = tmp_entry;
+			break;
+		}
+	}
+	if (unlikely(!entry)) {
+		mlx4_warn(dev, "Steering entry at index %x is not registered\n", index);
+		return -EINVAL;
+	}
+
+	/* the given qpn is listed as a promisc qpn
+	 * we need to add it as a duplicate to this entry
+	 * for future references */
+	list_for_each_entry(dqp, &entry->duplicates, list) {
+		if (qpn == dqp->qpn)
+			return 0; /* qp is already duplicated */
+	}
+
+	/* add the qp as a duplicate on this index */
+	dqp = kmalloc(sizeof(*dqp), GFP_KERNEL);
+	if (!dqp)
+		return -ENOMEM;
+	dqp->qpn = qpn;
+	list_add_tail(&dqp->list, &entry->duplicates);
+
+	return 0;
+}
+
+/* Check whether a qpn is a duplicate on steering entry
+ * If so, it should not be removed from mgm */
+static bool check_duplicate_entry(struct mlx4_dev *dev, u8 port,
+				  enum mlx4_steer_type steer,
+				  unsigned int index, u32 qpn)
+{
+	struct mlx4_steer *s_steer;
+	struct mlx4_steer_index *tmp_entry, *entry = NULL;
+	struct mlx4_promisc_qp *dqp, *tmp_dqp;
+
+	if (port < 1 || port > dev->caps.num_ports)
+		return NULL;
+
+	s_steer = &mlx4_priv(dev)->steer[port - 1];
+
+	/* if qp is not promisc, it cannot be duplicated */
+	if (!get_promisc_qp(dev, port, steer, qpn))
+		return false;
+
+	/* The qp is promisc qp so it is a duplicate on this index
+	 * Find the index entry, and remove the duplicate */
+	list_for_each_entry(tmp_entry, &s_steer->steer_entries[steer], list) {
+		if (tmp_entry->index == index) {
+			entry = tmp_entry;
+			break;
+		}
+	}
+	if (unlikely(!entry)) {
+		mlx4_warn(dev, "Steering entry for index %x is not registered\n", index);
+		return false;
+	}
+	list_for_each_entry_safe(dqp, tmp_dqp, &entry->duplicates, list) {
+		if (dqp->qpn == qpn) {
+			list_del(&dqp->list);
+			kfree(dqp);
+		}
+	}
+	return true;
+}
+
+/* Returns true if all the QPs != tqpn contained in this entry
+ * are Promisc QPs. Returns false otherwise.
+ */
+static bool promisc_steering_entry(struct mlx4_dev *dev, u8 port,
+				   enum mlx4_steer_type steer,
+				   unsigned int index, u32 tqpn,
+				   u32 *members_count)
+{
+	struct mlx4_cmd_mailbox *mailbox;
+	struct mlx4_mgm *mgm;
+	u32 m_count;
+	bool ret = false;
+	int i;
+
+	if (port < 1 || port > dev->caps.num_ports)
+		return false;
+
+	mailbox = mlx4_alloc_cmd_mailbox(dev);
+	if (IS_ERR(mailbox))
+		return false;
+	mgm = mailbox->buf;
+
+	if (mlx4_READ_ENTRY(dev, index, mailbox))
+		goto out;
+	m_count = be32_to_cpu(mgm->members_count) & 0xffffff;
+	if (members_count)
+		*members_count = m_count;
+
+	for (i = 0;  i < m_count; i++) {
+		u32 qpn = be32_to_cpu(mgm->qp[i]) & MGM_QPN_MASK;
+		if (!get_promisc_qp(dev, port, steer, qpn) && qpn != tqpn) {
+			/* the qp is not promisc, the entry can't be removed */
+			goto out;
+		}
+	}
+	ret = true;
+out:
+	mlx4_free_cmd_mailbox(dev, mailbox);
+	return ret;
+}
+
+/* IF a steering entry contains only promisc QPs, it can be removed. */
+static bool can_remove_steering_entry(struct mlx4_dev *dev, u8 port,
+				      enum mlx4_steer_type steer,
+				      unsigned int index, u32 tqpn)
+{
+	struct mlx4_steer *s_steer;
+	struct mlx4_steer_index *entry = NULL, *tmp_entry;
+	u32 members_count;
+	bool ret = false;
+
+	if (port < 1 || port > dev->caps.num_ports)
+		return NULL;
+
+	s_steer = &mlx4_priv(dev)->steer[port - 1];
+
+	if (!promisc_steering_entry(dev, port, steer, index,
+				    tqpn, &members_count))
+		goto out;
+
+	/* All the qps currently registered for this entry are promiscuous,
+	  * Checking for duplicates */
+	ret = true;
+	list_for_each_entry_safe(entry, tmp_entry, &s_steer->steer_entries[steer], list) {
+		if (entry->index == index) {
+			if (list_empty(&entry->duplicates) ||
+			    members_count == 1) {
+				struct mlx4_promisc_qp *pqp, *tmp_pqp;
+				/* If there is only 1 entry in duplicates then
+				 * this is the QP we want to delete, going over
+				 * the list and deleting the entry.
+				 */
+				list_del(&entry->list);
+				list_for_each_entry_safe(pqp, tmp_pqp,
+							 &entry->duplicates,
+							 list) {
+					list_del(&pqp->list);
+					kfree(pqp);
+				}
+				kfree(entry);
+			} else {
+				/* This entry contains duplicates so it shouldn't be removed */
+				ret = false;
+				goto out;
+			}
+		}
+	}
+
+out:
+	return ret;
+}
+
+static int add_promisc_qp(struct mlx4_dev *dev, u8 port,
+			  enum mlx4_steer_type steer, u32 qpn)
+{
+	struct mlx4_steer *s_steer;
+	struct mlx4_cmd_mailbox *mailbox;
+	struct mlx4_mgm *mgm;
+	struct mlx4_steer_index *entry;
+	struct mlx4_promisc_qp *pqp;
+	struct mlx4_promisc_qp *dqp;
+	u32 members_count;
+	u32 prot;
+	int i;
+	bool found;
+	int err;
+	struct mlx4_priv *priv = mlx4_priv(dev);
+
+	if (port < 1 || port > dev->caps.num_ports)
+		return -EINVAL;
+
+	s_steer = &mlx4_priv(dev)->steer[port - 1];
+
+	mutex_lock(&priv->mcg_table.mutex);
+
+	if (get_promisc_qp(dev, port, steer, qpn)) {
+		err = 0;  /* Noting to do, already exists */
+		goto out_mutex;
+	}
+
+	pqp = kmalloc(sizeof(*pqp), GFP_KERNEL);
+	if (!pqp) {
+		err = -ENOMEM;
+		goto out_mutex;
+	}
+	pqp->qpn = qpn;
+
+	mailbox = mlx4_alloc_cmd_mailbox(dev);
+	if (IS_ERR(mailbox)) {
+		err = -ENOMEM;
+		goto out_alloc;
+	}
+	mgm = mailbox->buf;
+
+	if (!(mlx4_is_mfunc(dev) && steer == MLX4_UC_STEER)) {
+		/* The promisc QP needs to be added for each one of the steering
+		 * entries. If it already exists, needs to be added as
+		 * a duplicate for this entry.
+		 */
+		list_for_each_entry(entry,
+				    &s_steer->steer_entries[steer],
+				    list) {
+			err = mlx4_READ_ENTRY(dev, entry->index, mailbox);
+			if (err)
+				goto out_mailbox;
+
+			members_count = be32_to_cpu(mgm->members_count) &
+					0xffffff;
+			prot = be32_to_cpu(mgm->members_count) >> 30;
+			found = false;
+			for (i = 0; i < members_count; i++) {
+				if ((be32_to_cpu(mgm->qp[i]) &
+				     MGM_QPN_MASK) == qpn) {
+					/* Entry already exists.
+					 * Add to duplicates.
+					 */
+					dqp = kmalloc(sizeof(*dqp), GFP_KERNEL);
+					if (!dqp) {
+						err = -ENOMEM;
+						goto out_mailbox;
+					}
+					dqp->qpn = qpn;
+					list_add_tail(&dqp->list,
+						      &entry->duplicates);
+					found = true;
+				}
+			}
+			if (!found) {
+				/* Need to add the qpn to mgm */
+				if (members_count ==
+				    dev->caps.num_qp_per_mgm) {
+					/* entry is full */
+					err = -ENOMEM;
+					goto out_mailbox;
+				}
+				mgm->qp[members_count++] =
+					cpu_to_be32(qpn & MGM_QPN_MASK);
+				mgm->members_count =
+					cpu_to_be32(members_count |
+						    (prot << 30));
+				err = mlx4_WRITE_ENTRY(dev, entry->index,
+						       mailbox);
+				if (err)
+					goto out_mailbox;
+			}
+		}
+	}
+
+	/* add the new qpn to list of promisc qps */
+	list_add_tail(&pqp->list, &s_steer->promisc_qps[steer]);
+	/* now need to add all the promisc qps to default entry */
+	memset(mgm, 0, sizeof(*mgm));
+	members_count = 0;
+	list_for_each_entry(dqp, &s_steer->promisc_qps[steer], list) {
+		if (members_count == dev->caps.num_qp_per_mgm) {
+			/* entry is full */
+			err = -ENOMEM;
+			goto out_list;
+		}
+		mgm->qp[members_count++] = cpu_to_be32(dqp->qpn & MGM_QPN_MASK);
+	}
+	mgm->members_count = cpu_to_be32(members_count | MLX4_PROT_ETH << 30);
+
+	err = mlx4_WRITE_PROMISC(dev, port, steer, mailbox);
+	if (err)
+		goto out_list;
+
+	mlx4_free_cmd_mailbox(dev, mailbox);
+	mutex_unlock(&priv->mcg_table.mutex);
+	return 0;
+
+out_list:
+	list_del(&pqp->list);
+out_mailbox:
+	mlx4_free_cmd_mailbox(dev, mailbox);
+out_alloc:
+	kfree(pqp);
+out_mutex:
+	mutex_unlock(&priv->mcg_table.mutex);
+	return err;
+}
+
+static int remove_promisc_qp(struct mlx4_dev *dev, u8 port,
+			     enum mlx4_steer_type steer, u32 qpn)
+{
+	struct mlx4_priv *priv = mlx4_priv(dev);
+	struct mlx4_steer *s_steer;
+	struct mlx4_cmd_mailbox *mailbox;
+	struct mlx4_mgm *mgm;
+	struct mlx4_steer_index *entry, *tmp_entry;
+	struct mlx4_promisc_qp *pqp;
+	struct mlx4_promisc_qp *dqp;
+	u32 members_count;
+	bool found;
+	bool back_to_list = false;
+	int i;
+	int err;
+
+	if (port < 1 || port > dev->caps.num_ports)
+		return -EINVAL;
+
+	s_steer = &mlx4_priv(dev)->steer[port - 1];
+	mutex_lock(&priv->mcg_table.mutex);
+
+	pqp = get_promisc_qp(dev, port, steer, qpn);
+	if (unlikely(!pqp)) {
+		mlx4_warn(dev, "QP %x is not promiscuous QP\n", qpn);
+		/* nothing to do */
+		err = 0;
+		goto out_mutex;
+	}
+
+	/*remove from list of promisc qps */
+	list_del(&pqp->list);
+
+	/* set the default entry not to include the removed one */
+	mailbox = mlx4_alloc_cmd_mailbox(dev);
+	if (IS_ERR(mailbox)) {
+		err = -ENOMEM;
+		back_to_list = true;
+		goto out_list;
+	}
+	mgm = mailbox->buf;
+	members_count = 0;
+	list_for_each_entry(dqp, &s_steer->promisc_qps[steer], list)
+		mgm->qp[members_count++] = cpu_to_be32(dqp->qpn & MGM_QPN_MASK);
+	mgm->members_count = cpu_to_be32(members_count | MLX4_PROT_ETH << 30);
+
+	err = mlx4_WRITE_PROMISC(dev, port, steer, mailbox);
+	if (err)
+		goto out_mailbox;
+
+	if (!(mlx4_is_mfunc(dev) && steer == MLX4_UC_STEER)) {
+		/* Remove the QP from all the steering entries */
+		list_for_each_entry_safe(entry, tmp_entry,
+					 &s_steer->steer_entries[steer],
+					 list) {
+			found = false;
+			list_for_each_entry(dqp, &entry->duplicates, list) {
+				if (dqp->qpn == qpn) {
+					found = true;
+					break;
+				}
+			}
+			if (found) {
+				/* A duplicate, no need to change the MGM,
+				 * only update the duplicates list
+				 */
+				list_del(&dqp->list);
+				kfree(dqp);
+			} else {
+				int loc = -1;
+
+				err = mlx4_READ_ENTRY(dev,
+						      entry->index,
+						      mailbox);
+				if (err)
+					goto out_mailbox;
+				members_count =
+					be32_to_cpu(mgm->members_count) &
+					0xffffff;
+				if (!members_count) {
+					mlx4_warn(dev, "QP %06x wasn't found in entry %x mcount=0. deleting entry...\n",
+						  qpn, entry->index);
+					list_del(&entry->list);
+					kfree(entry);
+					continue;
+				}
+
+				for (i = 0; i < members_count; ++i)
+					if ((be32_to_cpu(mgm->qp[i]) &
+					     MGM_QPN_MASK) == qpn) {
+						loc = i;
+						break;
+					}
+
+				if (loc < 0) {
+					mlx4_err(dev, "QP %06x wasn't found in entry %d\n",
+						 qpn, entry->index);
+					err = -EINVAL;
+					goto out_mailbox;
+				}
+
+				/* Copy the last QP in this MGM
+				 * over removed QP
+				 */
+				mgm->qp[loc] = mgm->qp[members_count - 1];
+				mgm->qp[members_count - 1] = 0;
+				mgm->members_count =
+					cpu_to_be32(--members_count |
+						    (MLX4_PROT_ETH << 30));
+
+				err = mlx4_WRITE_ENTRY(dev,
+						       entry->index,
+						       mailbox);
+				if (err)
+					goto out_mailbox;
+			}
+		}
+	}
+
+out_mailbox:
+	mlx4_free_cmd_mailbox(dev, mailbox);
+out_list:
+	if (back_to_list)
+		list_add_tail(&pqp->list, &s_steer->promisc_qps[steer]);
+	else
+		kfree(pqp);
+out_mutex:
+	mutex_unlock(&priv->mcg_table.mutex);
+	return err;
+}
+
+/*
+ * Caller must hold MCG table semaphore.  gid and mgm parameters must
+ * be properly aligned for command interface.
+ *
+ *  Returns 0 unless a firmware command error occurs.
+ *
+ * If GID is found in MGM or MGM is empty, *index = *hash, *prev = -1
+ * and *mgm holds MGM entry.
+ *
+ * if GID is found in AMGM, *index = index in AMGM, *prev = index of
+ * previous entry in hash chain and *mgm holds AMGM entry.
+ *
+ * If no AMGM exists for given gid, *index = -1, *prev = index of last
+ * entry in hash chain and *mgm holds end of hash chain.
+ */
+static int find_entry(struct mlx4_dev *dev, u8 port,
+		      u8 *gid, enum mlx4_protocol prot,
+		      struct mlx4_cmd_mailbox *mgm_mailbox,
+		      int *prev, int *index)
+{
+	struct mlx4_cmd_mailbox *mailbox;
+	struct mlx4_mgm *mgm = mgm_mailbox->buf;
+	u8 *mgid;
+	int err;
+	u16 hash;
+	u8 op_mod = (prot == MLX4_PROT_ETH) ?
+		!!(dev->caps.flags & MLX4_DEV_CAP_FLAG_VEP_MC_STEER) : 0;
+
+	mailbox = mlx4_alloc_cmd_mailbox(dev);
+	if (IS_ERR(mailbox))
+		return -ENOMEM;
+	mgid = mailbox->buf;
+
+	memcpy(mgid, gid, 16);
+
+	err = mlx4_GID_HASH(dev, mailbox, &hash, op_mod);
+	mlx4_free_cmd_mailbox(dev, mailbox);
+	if (err)
+		return err;
+
+	if (0)
+		mlx4_dbg(dev, "Hash for %pI6 is %04x\n", gid, hash);
+
+	*index = hash;
+	*prev  = -1;
+
+	do {
+		err = mlx4_READ_ENTRY(dev, *index, mgm_mailbox);
+		if (err)
+			return err;
+
+		if (!(be32_to_cpu(mgm->members_count) & 0xffffff)) {
+			if (*index != hash) {
+				mlx4_err(dev, "Found zero MGID in AMGM\n");
+				err = -EINVAL;
+			}
+			return err;
+		}
+
+		if (!memcmp(mgm->gid, gid, 16) &&
+		    be32_to_cpu(mgm->members_count) >> 30 == prot)
+			return err;
+
+		*prev = *index;
+		*index = be32_to_cpu(mgm->next_gid_index) >> 6;
+	} while (*index);
+
+	*index = -1;
+	return err;
+}
+
+static const u8 __promisc_mode[] = {
+	[MLX4_FS_REGULAR]   = 0x0,
+	[MLX4_FS_ALL_DEFAULT] = 0x1,
+	[MLX4_FS_MC_DEFAULT] = 0x3,
+	[MLX4_FS_MIRROR_RX_PORT] = 0x4,
+	[MLX4_FS_MIRROR_SX_PORT] = 0x5,
+	[MLX4_FS_UC_SNIFFER] = 0x6,
+	[MLX4_FS_MC_SNIFFER] = 0x7,
+};
+
+int mlx4_map_sw_to_hw_steering_mode(struct mlx4_dev *dev,
+				    enum mlx4_net_trans_promisc_mode flow_type)
+{
+	if (flow_type >= MLX4_FS_MODE_NUM) {
+		mlx4_err(dev, "Invalid flow type. type = %d\n", flow_type);
+		return -EINVAL;
+	}
+	return __promisc_mode[flow_type];
+}
+EXPORT_SYMBOL_GPL(mlx4_map_sw_to_hw_steering_mode);
+
+static void trans_rule_ctrl_to_hw(struct mlx4_net_trans_rule *ctrl,
+				  struct mlx4_net_trans_rule_hw_ctrl *hw)
+{
+	u8 flags = 0;
+
+	flags = ctrl->queue_mode == MLX4_NET_TRANS_Q_LIFO ? 1 : 0;
+	flags |= ctrl->exclusive ? (1 << 2) : 0;
+	flags |= ctrl->allow_loopback ? (1 << 3) : 0;
+
+	hw->flags = flags;
+	hw->type = __promisc_mode[ctrl->promisc_mode];
+	hw->prio = cpu_to_be16(ctrl->priority);
+	hw->port = ctrl->port;
+	hw->qpn = cpu_to_be32(ctrl->qpn);
+}
+
+const u16 __sw_id_hw[] = {
+	[MLX4_NET_TRANS_RULE_ID_ETH]     = 0xE001,
+	[MLX4_NET_TRANS_RULE_ID_IB]      = 0xE005,
+	[MLX4_NET_TRANS_RULE_ID_IPV6]    = 0xE003,
+	[MLX4_NET_TRANS_RULE_ID_IPV4]    = 0xE002,
+	[MLX4_NET_TRANS_RULE_ID_TCP]     = 0xE004,
+	[MLX4_NET_TRANS_RULE_ID_UDP]     = 0xE006,
+	[MLX4_NET_TRANS_RULE_ID_VXLAN]	 = 0xE008
+};
+
+int mlx4_map_sw_to_hw_steering_id(struct mlx4_dev *dev,
+				  enum mlx4_net_trans_rule_id id)
+{
+	if (id >= MLX4_NET_TRANS_RULE_NUM) {
+		mlx4_err(dev, "Invalid network rule id. id = %d\n", id);
+		return -EINVAL;
+	}
+	return __sw_id_hw[id];
+}
+EXPORT_SYMBOL_GPL(mlx4_map_sw_to_hw_steering_id);
+
+static const int __rule_hw_sz[] = {
+	[MLX4_NET_TRANS_RULE_ID_ETH] =
+		sizeof(struct mlx4_net_trans_rule_hw_eth),
+	[MLX4_NET_TRANS_RULE_ID_IB] =
+		sizeof(struct mlx4_net_trans_rule_hw_ib),
+	[MLX4_NET_TRANS_RULE_ID_IPV6] = 0,
+	[MLX4_NET_TRANS_RULE_ID_IPV4] =
+		sizeof(struct mlx4_net_trans_rule_hw_ipv4),
+	[MLX4_NET_TRANS_RULE_ID_TCP] =
+		sizeof(struct mlx4_net_trans_rule_hw_tcp_udp),
+	[MLX4_NET_TRANS_RULE_ID_UDP] =
+		sizeof(struct mlx4_net_trans_rule_hw_tcp_udp),
+	[MLX4_NET_TRANS_RULE_ID_VXLAN] =
+		sizeof(struct mlx4_net_trans_rule_hw_vxlan)
+};
+
+int mlx4_hw_rule_sz(struct mlx4_dev *dev,
+	       enum mlx4_net_trans_rule_id id)
+{
+	if (id >= MLX4_NET_TRANS_RULE_NUM) {
+		mlx4_err(dev, "Invalid network rule id. id = %d\n", id);
+		return -EINVAL;
+	}
+
+	return __rule_hw_sz[id];
+}
+EXPORT_SYMBOL_GPL(mlx4_hw_rule_sz);
+
+static int parse_trans_rule(struct mlx4_dev *dev, struct mlx4_spec_list *spec,
+			    struct _rule_hw *rule_hw)
+{
+	if (mlx4_hw_rule_sz(dev, spec->id) < 0)
+		return -EINVAL;
+	memset(rule_hw, 0, mlx4_hw_rule_sz(dev, spec->id));
+	rule_hw->id = cpu_to_be16(__sw_id_hw[spec->id]);
+	rule_hw->size = mlx4_hw_rule_sz(dev, spec->id) >> 2;
+
+	switch (spec->id) {
+	case MLX4_NET_TRANS_RULE_ID_ETH:
+		memcpy(rule_hw->eth.dst_mac, spec->eth.dst_mac, ETH_ALEN);
+		memcpy(rule_hw->eth.dst_mac_msk, spec->eth.dst_mac_msk,
+		       ETH_ALEN);
+		memcpy(rule_hw->eth.src_mac, spec->eth.src_mac, ETH_ALEN);
+		memcpy(rule_hw->eth.src_mac_msk, spec->eth.src_mac_msk,
+		       ETH_ALEN);
+		if (spec->eth.ether_type_enable) {
+			rule_hw->eth.ether_type_enable = 1;
+			rule_hw->eth.ether_type = spec->eth.ether_type;
+		}
+		rule_hw->eth.vlan_tag = spec->eth.vlan_id;
+		rule_hw->eth.vlan_tag_msk = spec->eth.vlan_id_msk;
+		break;
+
+	case MLX4_NET_TRANS_RULE_ID_IB:
+		rule_hw->ib.l3_qpn = spec->ib.l3_qpn;
+		rule_hw->ib.qpn_mask = spec->ib.qpn_msk;
+		memcpy(&rule_hw->ib.dst_gid, &spec->ib.dst_gid, 16);
+		memcpy(&rule_hw->ib.dst_gid_msk, &spec->ib.dst_gid_msk, 16);
+		break;
+
+	case MLX4_NET_TRANS_RULE_ID_IPV6:
+		return -EOPNOTSUPP;
+
+	case MLX4_NET_TRANS_RULE_ID_IPV4:
+		rule_hw->ipv4.src_ip = spec->ipv4.src_ip;
+		rule_hw->ipv4.src_ip_msk = spec->ipv4.src_ip_msk;
+		rule_hw->ipv4.dst_ip = spec->ipv4.dst_ip;
+		rule_hw->ipv4.dst_ip_msk = spec->ipv4.dst_ip_msk;
+		break;
+
+	case MLX4_NET_TRANS_RULE_ID_TCP:
+	case MLX4_NET_TRANS_RULE_ID_UDP:
+		rule_hw->tcp_udp.dst_port = spec->tcp_udp.dst_port;
+		rule_hw->tcp_udp.dst_port_msk = spec->tcp_udp.dst_port_msk;
+		rule_hw->tcp_udp.src_port = spec->tcp_udp.src_port;
+		rule_hw->tcp_udp.src_port_msk = spec->tcp_udp.src_port_msk;
+		break;
+
+	case MLX4_NET_TRANS_RULE_ID_VXLAN:
+		rule_hw->vxlan.vni =
+			cpu_to_be32(be32_to_cpu(spec->vxlan.vni) << 8);
+		rule_hw->vxlan.vni_mask =
+			cpu_to_be32(be32_to_cpu(spec->vxlan.vni_mask) << 8);
+		break;
+
+	default:
+		return -EINVAL;
+	}
+
+	return __rule_hw_sz[spec->id];
+}
+
+static void mlx4_err_rule(struct mlx4_dev *dev, char *str,
+			  struct mlx4_net_trans_rule *rule)
+{
+#define BUF_SIZE 256
+	struct mlx4_spec_list *cur;
+	char buf[BUF_SIZE];
+	int len = 0;
+
+	mlx4_err(dev, "%s", str);
+	len += snprintf(buf + len, BUF_SIZE - len,
+			"port = %d prio = 0x%x qp = 0x%x ",
+			rule->port, rule->priority, rule->qpn);
+
+	list_for_each_entry(cur, &rule->list, list) {
+		switch (cur->id) {
+		case MLX4_NET_TRANS_RULE_ID_ETH:
+			len += snprintf(buf + len, BUF_SIZE - len,
+					"dmac = %pM ", &cur->eth.dst_mac);
+			if (cur->eth.ether_type)
+				len += snprintf(buf + len, BUF_SIZE - len,
+						"ethertype = 0x%x ",
+						be16_to_cpu(cur->eth.ether_type));
+			if (cur->eth.vlan_id)
+				len += snprintf(buf + len, BUF_SIZE - len,
+						"vlan-id = %d ",
+						be16_to_cpu(cur->eth.vlan_id));
+			break;
+
+		case MLX4_NET_TRANS_RULE_ID_IPV4:
+			if (cur->ipv4.src_ip)
+				len += snprintf(buf + len, BUF_SIZE - len,
+						"src-ip = %pI4 ",
+						&cur->ipv4.src_ip);
+			if (cur->ipv4.dst_ip)
+				len += snprintf(buf + len, BUF_SIZE - len,
+						"dst-ip = %pI4 ",
+						&cur->ipv4.dst_ip);
+			break;
+
+		case MLX4_NET_TRANS_RULE_ID_TCP:
+		case MLX4_NET_TRANS_RULE_ID_UDP:
+			if (cur->tcp_udp.src_port)
+				len += snprintf(buf + len, BUF_SIZE - len,
+						"src-port = %d ",
+						be16_to_cpu(cur->tcp_udp.src_port));
+			if (cur->tcp_udp.dst_port)
+				len += snprintf(buf + len, BUF_SIZE - len,
+						"dst-port = %d ",
+						be16_to_cpu(cur->tcp_udp.dst_port));
+			break;
+
+		case MLX4_NET_TRANS_RULE_ID_IB:
+			len += snprintf(buf + len, BUF_SIZE - len,
+					"dst-gid = %pI6\n", cur->ib.dst_gid);
+			len += snprintf(buf + len, BUF_SIZE - len,
+					"dst-gid-mask = %pI6\n",
+					cur->ib.dst_gid_msk);
+			break;
+
+		case MLX4_NET_TRANS_RULE_ID_VXLAN:
+			len += snprintf(buf + len, BUF_SIZE - len,
+					"VNID = %d ", be32_to_cpu(cur->vxlan.vni));
+			break;
+		case MLX4_NET_TRANS_RULE_ID_IPV6:
+			break;
+
+		default:
+			break;
+		}
+	}
+	len += snprintf(buf + len, BUF_SIZE - len, "\n");
+	mlx4_err(dev, "%s", buf);
+
+	if (len >= BUF_SIZE)
+		mlx4_err(dev, "Network rule error message was truncated, print buffer is too small\n");
+}
+
+int mlx4_flow_attach(struct mlx4_dev *dev,
+		     struct mlx4_net_trans_rule *rule, u64 *reg_id)
+{
+	struct mlx4_cmd_mailbox *mailbox;
+	struct mlx4_spec_list *cur;
+	u32 size = 0;
+	int ret;
+
+	mailbox = mlx4_alloc_cmd_mailbox(dev);
+	if (IS_ERR(mailbox))
+		return PTR_ERR(mailbox);
+
+	if (!mlx4_qp_lookup(dev, rule->qpn)) {
+		mlx4_err_rule(dev, "QP doesn't exist\n", rule);
+		ret = -EINVAL;
+		goto out;
+	}
+
+	trans_rule_ctrl_to_hw(rule, mailbox->buf);
+
+	size += sizeof(struct mlx4_net_trans_rule_hw_ctrl);
+
+	list_for_each_entry(cur, &rule->list, list) {
+		ret = parse_trans_rule(dev, cur, mailbox->buf + size);
+		if (ret < 0)
+			goto out;
+
+		size += ret;
+	}
+
+	ret = mlx4_QP_FLOW_STEERING_ATTACH(dev, mailbox, size >> 2, reg_id);
+	if (ret == -ENOMEM) {
+		mlx4_err_rule(dev,
+			      "mcg table is full. Fail to register network rule\n",
+			      rule);
+	} else if (ret) {
+		if (ret == -ENXIO) {
+			if (dev->caps.steering_mode != MLX4_STEERING_MODE_DEVICE_MANAGED)
+				mlx4_err_rule(dev,
+					      "DMFS is not enabled, "
+					      "failed to register network rule.\n",
+					      rule);
+			else
+				mlx4_err_rule(dev,
+					      "Rule exceeds the dmfs_high_rate_mode limitations, "
+					      "failed to register network rule.\n",
+					      rule);
+
+		} else {
+			mlx4_err_rule(dev, "Fail to register network rule.\n", rule);
+		}
+	}
+
+out:
+	mlx4_free_cmd_mailbox(dev, mailbox);
+
+	return ret;
+}
+EXPORT_SYMBOL_GPL(mlx4_flow_attach);
+
+int mlx4_flow_detach(struct mlx4_dev *dev, u64 reg_id)
+{
+	int err;
+
+	err = mlx4_QP_FLOW_STEERING_DETACH(dev, reg_id);
+	if (err)
+		mlx4_err(dev, "Fail to detach network rule. registration id = 0x%llx\n",
+			 reg_id);
+	return err;
+}
+EXPORT_SYMBOL_GPL(mlx4_flow_detach);
+
+int mlx4_tunnel_steer_add(struct mlx4_dev *dev, unsigned char *addr,
+			  int port, int qpn, u16 prio, u64 *reg_id)
+{
+	int err;
+	struct mlx4_spec_list spec_eth_outer = { {NULL} };
+	struct mlx4_spec_list spec_vxlan     = { {NULL} };
+	struct mlx4_spec_list spec_eth_inner = { {NULL} };
+
+	struct mlx4_net_trans_rule rule = {
+		.queue_mode = MLX4_NET_TRANS_Q_FIFO,
+		.exclusive = 0,
+		.allow_loopback = 1,
+		.promisc_mode = MLX4_FS_REGULAR,
+	};
+
+	__be64 mac_mask = cpu_to_be64(MLX4_MAC_MASK << 16);
+
+	rule.port = port;
+	rule.qpn = qpn;
+	rule.priority = prio;
+	INIT_LIST_HEAD(&rule.list);
+
+	spec_eth_outer.id = MLX4_NET_TRANS_RULE_ID_ETH;
+	memcpy(spec_eth_outer.eth.dst_mac, addr, ETH_ALEN);
+	memcpy(spec_eth_outer.eth.dst_mac_msk, &mac_mask, ETH_ALEN);
+
+	spec_vxlan.id = MLX4_NET_TRANS_RULE_ID_VXLAN;    /* any vxlan header */
+	spec_eth_inner.id = MLX4_NET_TRANS_RULE_ID_ETH;	 /* any inner eth header */
+
+	list_add_tail(&spec_eth_outer.list, &rule.list);
+	list_add_tail(&spec_vxlan.list,     &rule.list);
+	list_add_tail(&spec_eth_inner.list, &rule.list);
+
+	err = mlx4_flow_attach(dev, &rule, reg_id);
+	return err;
+}
+EXPORT_SYMBOL(mlx4_tunnel_steer_add);
+
+int mlx4_FLOW_STEERING_IB_UC_QP_RANGE(struct mlx4_dev *dev, u32 min_range_qpn,
+				      u32 max_range_qpn)
+{
+	int err;
+	u64 in_param;
+
+	in_param = ((u64) min_range_qpn) << 32;
+	in_param |= ((u64) max_range_qpn) & 0xFFFFFFFF;
+
+	err = mlx4_cmd(dev, in_param, 0, 0,
+			MLX4_FLOW_STEERING_IB_UC_QP_RANGE,
+			MLX4_CMD_TIME_CLASS_A, MLX4_CMD_NATIVE);
+
+	return err;
+}
+EXPORT_SYMBOL_GPL(mlx4_FLOW_STEERING_IB_UC_QP_RANGE);
+
+int mlx4_qp_attach_common(struct mlx4_dev *dev, struct mlx4_qp *qp, u8 gid[16],
+			  int block_mcast_loopback, enum mlx4_protocol prot,
+			  enum mlx4_steer_type steer)
+{
+	struct mlx4_priv *priv = mlx4_priv(dev);
+	struct mlx4_cmd_mailbox *mailbox;
+	struct mlx4_mgm *mgm;
+	u32 members_count;
+	int index = -1, prev;
+	int link = 0;
+	int i;
+	int err;
+	u8 port = gid[5];
+	u8 new_entry = 0;
+
+	mailbox = mlx4_alloc_cmd_mailbox(dev);
+	if (IS_ERR(mailbox))
+		return PTR_ERR(mailbox);
+	mgm = mailbox->buf;
+
+	mutex_lock(&priv->mcg_table.mutex);
+	err = find_entry(dev, port, gid, prot,
+			 mailbox, &prev, &index);
+	if (err)
+		goto out;
+
+	if (index != -1) {
+		if (!(be32_to_cpu(mgm->members_count) & 0xffffff)) {
+			new_entry = 1;
+			memcpy(mgm->gid, gid, 16);
+		}
+	} else {
+		link = 1;
+
+		index = mlx4_bitmap_alloc(&priv->mcg_table.bitmap);
+		if (index == -1) {
+			mlx4_err(dev, "No AMGM entries left\n");
+			err = -ENOMEM;
+			goto out;
+		}
+		index += dev->caps.num_mgms;
+
+		new_entry = 1;
+		memset(mgm, 0, sizeof(*mgm));
+		memcpy(mgm->gid, gid, 16);
+	}
+
+	members_count = be32_to_cpu(mgm->members_count) & 0xffffff;
+	if (members_count == dev->caps.num_qp_per_mgm) {
+		mlx4_err(dev, "MGM at index %x is full\n", index);
+		err = -ENOMEM;
+		goto out;
+	}
+
+	for (i = 0; i < members_count; ++i)
+		if ((be32_to_cpu(mgm->qp[i]) & MGM_QPN_MASK) == qp->qpn) {
+			mlx4_dbg(dev, "QP %06x already a member of MGM\n", qp->qpn);
+			err = 0;
+			goto out;
+		}
+
+	if (block_mcast_loopback)
+		mgm->qp[members_count++] = cpu_to_be32((qp->qpn & MGM_QPN_MASK) |
+						       (1U << MGM_BLCK_LB_BIT));
+	else
+		mgm->qp[members_count++] = cpu_to_be32(qp->qpn & MGM_QPN_MASK);
+
+	mgm->members_count = cpu_to_be32(members_count | (u32) prot << 30);
+
+	err = mlx4_WRITE_ENTRY(dev, index, mailbox);
+	if (err)
+		goto out;
+
+	if (!link)
+		goto out;
+
+	err = mlx4_READ_ENTRY(dev, prev, mailbox);
+	if (err)
+		goto out;
+
+	mgm->next_gid_index = cpu_to_be32(index << 6);
+
+	err = mlx4_WRITE_ENTRY(dev, prev, mailbox);
+	if (err)
+		goto out;
+
+out:
+	if (prot == MLX4_PROT_ETH && index != -1) {
+		/* manage the steering entry for promisc mode */
+		if (new_entry)
+			err = new_steering_entry(dev, port, steer,
+						 index, qp->qpn);
+		else
+			err = existing_steering_entry(dev, port, steer,
+						      index, qp->qpn);
+	}
+	if (err && link && index != -1) {
+		if (index < dev->caps.num_mgms)
+			mlx4_warn(dev, "Got AMGM index %d < %d\n",
+				  index, dev->caps.num_mgms);
+		else
+			mlx4_bitmap_free(&priv->mcg_table.bitmap,
+					 index - dev->caps.num_mgms, MLX4_USE_RR);
+	}
+	mutex_unlock(&priv->mcg_table.mutex);
+
+	mlx4_free_cmd_mailbox(dev, mailbox);
+	return err;
+}
+
+int mlx4_qp_detach_common(struct mlx4_dev *dev, struct mlx4_qp *qp, u8 gid[16],
+			  enum mlx4_protocol prot, enum mlx4_steer_type steer)
+{
+	struct mlx4_priv *priv = mlx4_priv(dev);
+	struct mlx4_cmd_mailbox *mailbox;
+	struct mlx4_mgm *mgm;
+	u32 members_count;
+	int prev, index;
+	int i, loc = -1;
+	int err;
+	u8 port = gid[5];
+	bool removed_entry = false;
+
+	mailbox = mlx4_alloc_cmd_mailbox(dev);
+	if (IS_ERR(mailbox))
+		return PTR_ERR(mailbox);
+	mgm = mailbox->buf;
+
+	mutex_lock(&priv->mcg_table.mutex);
+
+	err = find_entry(dev, port, gid, prot,
+			 mailbox, &prev, &index);
+	if (err)
+		goto out;
+
+	if (index == -1) {
+		mlx4_err(dev, "MGID %pI6 not found\n", gid);
+		err = -EINVAL;
+		goto out;
+	}
+
+	/* If this QP is also a promisc QP, it shouldn't be removed only if
+	 * at least one none promisc QP is also attached to this MCG
+	 */
+	if (prot == MLX4_PROT_ETH &&
+	    check_duplicate_entry(dev, port, steer, index, qp->qpn) &&
+	    !promisc_steering_entry(dev, port, steer, index, qp->qpn, NULL))
+			goto out;
+
+	members_count = be32_to_cpu(mgm->members_count) & 0xffffff;
+	for (i = 0; i < members_count; ++i)
+		if ((be32_to_cpu(mgm->qp[i]) & MGM_QPN_MASK) == qp->qpn) {
+			loc = i;
+			break;
+		}
+
+	if (loc == -1) {
+		mlx4_err(dev, "QP %06x not found in MGM\n", qp->qpn);
+		err = -EINVAL;
+		goto out;
+	}
+
+	/* copy the last QP in this MGM over removed QP */
+	mgm->qp[loc] = mgm->qp[members_count - 1];
+	mgm->qp[members_count - 1] = 0;
+	mgm->members_count = cpu_to_be32(--members_count | (u32) prot << 30);
+
+	if (prot == MLX4_PROT_ETH)
+		removed_entry = can_remove_steering_entry(dev, port, steer,
+								index, qp->qpn);
+	if (members_count && (prot != MLX4_PROT_ETH || !removed_entry)) {
+		err = mlx4_WRITE_ENTRY(dev, index, mailbox);
+		goto out;
+	}
+
+	/* We are going to delete the entry, members count should be 0 */
+	mgm->members_count = cpu_to_be32((u32) prot << 30);
+
+	if (prev == -1) {
+		/* Remove entry from MGM */
+		int amgm_index = be32_to_cpu(mgm->next_gid_index) >> 6;
+		if (amgm_index) {
+			err = mlx4_READ_ENTRY(dev, amgm_index, mailbox);
+			if (err)
+				goto out;
+		} else
+			memset(mgm->gid, 0, 16);
+
+		err = mlx4_WRITE_ENTRY(dev, index, mailbox);
+		if (err)
+			goto out;
+
+		if (amgm_index) {
+			if (amgm_index < dev->caps.num_mgms)
+				mlx4_warn(dev, "MGM entry %d had AMGM index %d < %d\n",
+					  index, amgm_index, dev->caps.num_mgms);
+			else
+				mlx4_bitmap_free(&priv->mcg_table.bitmap,
+						 amgm_index - dev->caps.num_mgms, MLX4_USE_RR);
+		}
+	} else {
+		/* Remove entry from AMGM */
+		int cur_next_index = be32_to_cpu(mgm->next_gid_index) >> 6;
+		err = mlx4_READ_ENTRY(dev, prev, mailbox);
+		if (err)
+			goto out;
+
+		mgm->next_gid_index = cpu_to_be32(cur_next_index << 6);
+
+		err = mlx4_WRITE_ENTRY(dev, prev, mailbox);
+		if (err)
+			goto out;
+
+		if (index < dev->caps.num_mgms)
+			mlx4_warn(dev, "entry %d had next AMGM index %d < %d\n",
+				  prev, index, dev->caps.num_mgms);
+		else
+			mlx4_bitmap_free(&priv->mcg_table.bitmap,
+					 index - dev->caps.num_mgms, MLX4_USE_RR);
+	}
+
+out:
+	mutex_unlock(&priv->mcg_table.mutex);
+
+	mlx4_free_cmd_mailbox(dev, mailbox);
+	if (err && dev->persist->state & MLX4_DEVICE_STATE_INTERNAL_ERROR)
+		/* In case device is under an error, return success as a closing command */
+		err = 0;
+	return err;
+}
+
+static int mlx4_QP_ATTACH(struct mlx4_dev *dev, struct mlx4_qp *qp,
+			  u8 gid[16], u8 attach, u8 block_loopback,
+			  enum mlx4_protocol prot)
+{
+	struct mlx4_cmd_mailbox *mailbox;
+	int err = 0;
+	int qpn;
+
+	if (!mlx4_is_mfunc(dev))
+		return -EBADF;
+
+	mailbox = mlx4_alloc_cmd_mailbox(dev);
+	if (IS_ERR(mailbox))
+		return PTR_ERR(mailbox);
+
+	memcpy(mailbox->buf, gid, 16);
+	qpn = qp->qpn;
+	qpn |= (prot << 28);
+	if (attach && block_loopback)
+		qpn |= (1 << 31);
+
+	err = mlx4_cmd(dev, mailbox->dma, qpn, attach,
+		       MLX4_CMD_QP_ATTACH, MLX4_CMD_TIME_CLASS_A,
+		       MLX4_CMD_WRAPPED);
+
+	mlx4_free_cmd_mailbox(dev, mailbox);
+	if (err && !attach &&
+	    dev->persist->state & MLX4_DEVICE_STATE_INTERNAL_ERROR)
+		err = 0;
+	return err;
+}
+
+int mlx4_trans_to_dmfs_attach(struct mlx4_dev *dev, struct mlx4_qp *qp,
+			      u8 gid[16], u8 port,
+			      int block_mcast_loopback,
+			      enum mlx4_protocol prot, u64 *reg_id)
+{
+		struct mlx4_spec_list spec = { {NULL} };
+		__be64 mac_mask = cpu_to_be64(MLX4_MAC_MASK << 16);
+
+		struct mlx4_net_trans_rule rule = {
+			.queue_mode = MLX4_NET_TRANS_Q_FIFO,
+			.exclusive = 0,
+			.promisc_mode = MLX4_FS_REGULAR,
+			.priority = MLX4_DOMAIN_NIC,
+		};
+
+		rule.allow_loopback = !block_mcast_loopback;
+		rule.port = port;
+		rule.qpn = qp->qpn;
+		INIT_LIST_HEAD(&rule.list);
+
+		switch (prot) {
+		case MLX4_PROT_ETH:
+			spec.id = MLX4_NET_TRANS_RULE_ID_ETH;
+			memcpy(spec.eth.dst_mac, &gid[10], ETH_ALEN);
+			memcpy(spec.eth.dst_mac_msk, &mac_mask, ETH_ALEN);
+			break;
+
+		case MLX4_PROT_IB_IPV6:
+			spec.id = MLX4_NET_TRANS_RULE_ID_IB;
+			memcpy(spec.ib.dst_gid, gid, 16);
+			memset(&spec.ib.dst_gid_msk, 0xff, 16);
+			break;
+		default:
+			return -EINVAL;
+		}
+		list_add_tail(&spec.list, &rule.list);
+
+		return mlx4_flow_attach(dev, &rule, reg_id);
+}
+
+int mlx4_multicast_attach(struct mlx4_dev *dev, struct mlx4_qp *qp, u8 gid[16],
+			  u8 port, int block_mcast_loopback,
+			  enum mlx4_protocol prot, u64 *reg_id)
+{
+	switch (dev->caps.steering_mode) {
+	case MLX4_STEERING_MODE_A0:
+		if (prot == MLX4_PROT_ETH)
+			return 0;
+		/* fall through */
+
+	case MLX4_STEERING_MODE_B0:
+		if (prot == MLX4_PROT_ETH)
+			gid[7] |= (MLX4_MC_STEER << 1);
+
+		if (mlx4_is_mfunc(dev))
+			return mlx4_QP_ATTACH(dev, qp, gid, 1,
+					      block_mcast_loopback, prot);
+		return mlx4_qp_attach_common(dev, qp, gid,
+					     block_mcast_loopback, prot,
+					     MLX4_MC_STEER);
+
+	case MLX4_STEERING_MODE_DEVICE_MANAGED:
+		return mlx4_trans_to_dmfs_attach(dev, qp, gid, port,
+						 block_mcast_loopback,
+						 prot, reg_id);
+	default:
+		return -EINVAL;
+	}
+}
+EXPORT_SYMBOL_GPL(mlx4_multicast_attach);
+
+int mlx4_multicast_detach(struct mlx4_dev *dev, struct mlx4_qp *qp, u8 gid[16],
+			  enum mlx4_protocol prot, u64 reg_id)
+{
+	switch (dev->caps.steering_mode) {
+	case MLX4_STEERING_MODE_A0:
+		if (prot == MLX4_PROT_ETH)
+			return 0;
+		/* fall through */
+
+	case MLX4_STEERING_MODE_B0:
+		if (prot == MLX4_PROT_ETH)
+			gid[7] |= (MLX4_MC_STEER << 1);
+
+		if (mlx4_is_mfunc(dev))
+			return mlx4_QP_ATTACH(dev, qp, gid, 0, 0, prot);
+
+		return mlx4_qp_detach_common(dev, qp, gid, prot,
+					     MLX4_MC_STEER);
+
+	case MLX4_STEERING_MODE_DEVICE_MANAGED:
+		return mlx4_flow_detach(dev, reg_id);
+
+	default:
+		return -EINVAL;
+	}
+}
+EXPORT_SYMBOL_GPL(mlx4_multicast_detach);
+
+int mlx4_flow_steer_promisc_add(struct mlx4_dev *dev, u8 port,
+				u32 qpn, enum mlx4_net_trans_promisc_mode mode)
+{
+	struct mlx4_net_trans_rule rule = {
+		.queue_mode = MLX4_NET_TRANS_Q_FIFO,
+		.exclusive = 0,
+		.allow_loopback = 1,
+	};
+
+	u64 *regid_p;
+
+	switch (mode) {
+	case MLX4_FS_ALL_DEFAULT:
+		regid_p = &dev->regid_promisc_array[port];
+		break;
+	case MLX4_FS_MC_DEFAULT:
+		regid_p = &dev->regid_allmulti_array[port];
+		break;
+	default:
+		return -1;
+	}
+
+	if (*regid_p != 0)
+		return -1;
+
+	rule.promisc_mode = mode;
+	rule.port = port;
+	rule.qpn = qpn;
+	INIT_LIST_HEAD(&rule.list);
+	mlx4_info(dev, "going promisc on %x\n", port);
+
+	return  mlx4_flow_attach(dev, &rule, regid_p);
+}
+EXPORT_SYMBOL_GPL(mlx4_flow_steer_promisc_add);
+
+int mlx4_flow_steer_promisc_remove(struct mlx4_dev *dev, u8 port,
+				   enum mlx4_net_trans_promisc_mode mode)
+{
+	int ret;
+	u64 *regid_p;
+
+	switch (mode) {
+	case MLX4_FS_ALL_DEFAULT:
+		regid_p = &dev->regid_promisc_array[port];
+		break;
+	case MLX4_FS_MC_DEFAULT:
+		regid_p = &dev->regid_allmulti_array[port];
+		break;
+	default:
+		return -1;
+	}
+
+	if (*regid_p == 0)
+		return -1;
+
+	ret =  mlx4_flow_detach(dev, *regid_p);
+	if (ret == 0)
+		*regid_p = 0;
+
+	return ret;
+}
+EXPORT_SYMBOL_GPL(mlx4_flow_steer_promisc_remove);
+
+int mlx4_unicast_attach(struct mlx4_dev *dev,
+			struct mlx4_qp *qp, u8 gid[16],
+			int block_mcast_loopback, enum mlx4_protocol prot)
+{
+	if (prot == MLX4_PROT_ETH)
+		gid[7] |= (MLX4_UC_STEER << 1);
+
+	if (mlx4_is_mfunc(dev))
+		return mlx4_QP_ATTACH(dev, qp, gid, 1,
+					block_mcast_loopback, prot);
+
+	return mlx4_qp_attach_common(dev, qp, gid, block_mcast_loopback,
+					prot, MLX4_UC_STEER);
+}
+EXPORT_SYMBOL_GPL(mlx4_unicast_attach);
+
+int mlx4_unicast_detach(struct mlx4_dev *dev, struct mlx4_qp *qp,
+			       u8 gid[16], enum mlx4_protocol prot)
+{
+	if (prot == MLX4_PROT_ETH)
+		gid[7] |= (MLX4_UC_STEER << 1);
+
+	if (mlx4_is_mfunc(dev))
+		return mlx4_QP_ATTACH(dev, qp, gid, 0, 0, prot);
+
+	return mlx4_qp_detach_common(dev, qp, gid, prot, MLX4_UC_STEER);
+}
+EXPORT_SYMBOL_GPL(mlx4_unicast_detach);
+
+int mlx4_PROMISC_wrapper(struct mlx4_dev *dev, int slave,
+			 struct mlx4_vhcr *vhcr,
+			 struct mlx4_cmd_mailbox *inbox,
+			 struct mlx4_cmd_mailbox *outbox,
+			 struct mlx4_cmd_info *cmd)
+{
+	u32 qpn = (u32) vhcr->in_param & 0xffffffff;
+	int port = mlx4_slave_convert_port(dev, slave, vhcr->in_param >> 62);
+	enum mlx4_steer_type steer = vhcr->in_modifier;
+
+	if (port < 0)
+		return -EINVAL;
+
+	/* Promiscuous unicast is not allowed in mfunc */
+	if (mlx4_is_mfunc(dev) && steer == MLX4_UC_STEER)
+		return 0;
+
+	if (vhcr->op_modifier)
+		return add_promisc_qp(dev, port, steer, qpn);
+	else
+		return remove_promisc_qp(dev, port, steer, qpn);
+}
+
+static int mlx4_PROMISC(struct mlx4_dev *dev, u32 qpn,
+			enum mlx4_steer_type steer, u8 add, u8 port)
+{
+	return mlx4_cmd(dev, (u64) qpn | (u64) port << 62, (u32) steer, add,
+			MLX4_CMD_PROMISC, MLX4_CMD_TIME_CLASS_A,
+			MLX4_CMD_WRAPPED);
+}
+
+int mlx4_multicast_promisc_add(struct mlx4_dev *dev, u32 qpn, u8 port)
+{
+	if (mlx4_is_mfunc(dev))
+		return mlx4_PROMISC(dev, qpn, MLX4_MC_STEER, 1, port);
+
+	return add_promisc_qp(dev, port, MLX4_MC_STEER, qpn);
+}
+EXPORT_SYMBOL_GPL(mlx4_multicast_promisc_add);
+
+int mlx4_multicast_promisc_remove(struct mlx4_dev *dev, u32 qpn, u8 port)
+{
+	if (mlx4_is_mfunc(dev))
+		return mlx4_PROMISC(dev, qpn, MLX4_MC_STEER, 0, port);
+
+	return remove_promisc_qp(dev, port, MLX4_MC_STEER, qpn);
+}
+EXPORT_SYMBOL_GPL(mlx4_multicast_promisc_remove);
+
+int mlx4_unicast_promisc_add(struct mlx4_dev *dev, u32 qpn, u8 port)
+{
+	if (mlx4_is_mfunc(dev))
+		return mlx4_PROMISC(dev, qpn, MLX4_UC_STEER, 1, port);
+
+	return add_promisc_qp(dev, port, MLX4_UC_STEER, qpn);
+}
+EXPORT_SYMBOL_GPL(mlx4_unicast_promisc_add);
+
+int mlx4_unicast_promisc_remove(struct mlx4_dev *dev, u32 qpn, u8 port)
+{
+	if (mlx4_is_mfunc(dev))
+		return mlx4_PROMISC(dev, qpn, MLX4_UC_STEER, 0, port);
+
+	return remove_promisc_qp(dev, port, MLX4_UC_STEER, qpn);
+}
+EXPORT_SYMBOL_GPL(mlx4_unicast_promisc_remove);
+
+int mlx4_init_mcg_table(struct mlx4_dev *dev)
+{
+	struct mlx4_priv *priv = mlx4_priv(dev);
+	int err;
+
+	/* No need for mcg_table when fw managed the mcg table*/
+	if (dev->caps.steering_mode ==
+	    MLX4_STEERING_MODE_DEVICE_MANAGED)
+		return 0;
+	err = mlx4_bitmap_init(&priv->mcg_table.bitmap, dev->caps.num_amgms,
+			       dev->caps.num_amgms - 1, 0, 0);
+	if (err)
+		return err;
+
+	mutex_init(&priv->mcg_table.mutex);
+
+	return 0;
+}
+
+void mlx4_cleanup_mcg_table(struct mlx4_dev *dev)
+{
+	if (dev->caps.steering_mode !=
+	    MLX4_STEERING_MODE_DEVICE_MANAGED)
+		mlx4_bitmap_cleanup(&mlx4_priv(dev)->mcg_table.bitmap);
+}
diff --git a/drivers/net/ethernet/mellanox/mlx4/mlx4.h b/drivers/net/ethernet/mellanox/mlx4/mlx4.h
new file mode 100644
index 000000000..23f1b5b51
--- /dev/null
+++ b/drivers/net/ethernet/mellanox/mlx4/mlx4.h
@@ -0,0 +1,1486 @@
+/*
+ * Copyright (c) 2004, 2005 Topspin Communications.  All rights reserved.
+ * Copyright (c) 2005 Sun Microsystems, Inc. All rights reserved.
+ * Copyright (c) 2005, 2006, 2007 Cisco Systems.  All rights reserved.
+ * Copyright (c) 2005, 2006, 2007, 2008 Mellanox Technologies. All rights reserved.
+ * Copyright (c) 2004 Voltaire, Inc. All rights reserved.
+ *
+ * This software is available to you under a choice of one of two
+ * licenses.  You may choose to be licensed under the terms of the GNU
+ * General Public License (GPL) Version 2, available from the file
+ * COPYING in the main directory of this source tree, or the
+ * OpenIB.org BSD license below:
+ *
+ *     Redistribution and use in source and binary forms, with or
+ *     without modification, are permitted provided that the following
+ *     conditions are met:
+ *
+ *      - Redistributions of source code must retain the above
+ *        copyright notice, this list of conditions and the following
+ *        disclaimer.
+ *
+ *      - Redistributions in binary form must reproduce the above
+ *        copyright notice, this list of conditions and the following
+ *        disclaimer in the documentation and/or other materials
+ *        provided with the distribution.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
+ * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
+ * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS
+ * BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN
+ * ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
+ * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+ * SOFTWARE.
+ */
+
+#ifndef MLX4_H
+#define MLX4_H
+
+#include <linux/mutex.h>
+#include <linux/radix-tree.h>
+#include <linux/rbtree.h>
+#include <linux/timer.h>
+#include <linux/semaphore.h>
+#include <linux/workqueue.h>
+#include <linux/interrupt.h>
+#include <linux/spinlock.h>
+#include <net/devlink.h>
+#include <linux/rwsem.h>
+
+#include <linux/mlx4/device.h>
+#include <linux/mlx4/driver.h>
+#include <linux/mlx4/doorbell.h>
+#include <linux/mlx4/cmd.h>
+#include "fw_qos.h"
+
+#define DRV_NAME	"mlx4_core"
+#define DRV_VERSION	"4.0-0"
+#define DRV_NAME_FOR_FW		"Linux," DRV_NAME "," DRV_VERSION
+
+#define MLX4_FS_UDP_UC_EN		(1 << 1)
+#define MLX4_FS_TCP_UC_EN		(1 << 2)
+#define MLX4_FS_NUM_OF_L2_ADDR		8
+#define MLX4_FS_MGM_LOG_ENTRY_SIZE	7
+#define MLX4_FS_NUM_MCG			(1 << 17)
+
+#define INIT_HCA_TPT_MW_ENABLE          (1 << 7)
+
+#define MLX4_QUERY_IF_STAT_RESET	BIT(31)
+
+enum {
+	MLX4_HCR_BASE		= 0x80680,
+	MLX4_HCR_SIZE		= 0x0001c,
+	MLX4_CLR_INT_SIZE	= 0x00008,
+	MLX4_SLAVE_COMM_BASE	= 0x0,
+	MLX4_COMM_PAGESIZE	= 0x1000,
+	MLX4_CLOCK_SIZE		= 0x00008,
+	MLX4_COMM_CHAN_CAPS	= 0x8,
+	MLX4_COMM_CHAN_FLAGS	= 0xc
+};
+
+enum {
+	MLX4_DEFAULT_MGM_LOG_ENTRY_SIZE = 10,
+	MLX4_MIN_MGM_LOG_ENTRY_SIZE = 7,
+	MLX4_MAX_MGM_LOG_ENTRY_SIZE = 12,
+	MLX4_MAX_QP_PER_MGM = 4 * ((1 << MLX4_MAX_MGM_LOG_ENTRY_SIZE) / 16 - 2),
+};
+
+enum {
+	MLX4_NUM_PDS		= 1 << 15
+};
+
+enum {
+	MLX4_CMPT_TYPE_QP	= 0,
+	MLX4_CMPT_TYPE_SRQ	= 1,
+	MLX4_CMPT_TYPE_CQ	= 2,
+	MLX4_CMPT_TYPE_EQ	= 3,
+	MLX4_CMPT_NUM_TYPE
+};
+
+enum {
+	MLX4_CMPT_SHIFT		= 24,
+	MLX4_NUM_CMPTS		= MLX4_CMPT_NUM_TYPE << MLX4_CMPT_SHIFT
+};
+
+enum mlx4_mpt_state {
+	MLX4_MPT_DISABLED = 0,
+	MLX4_MPT_EN_HW,
+	MLX4_MPT_EN_SW
+};
+
+#define MLX4_COMM_TIME		10000
+#define MLX4_COMM_OFFLINE_TIME_OUT 30000
+#define MLX4_COMM_CMD_NA_OP    0x0
+
+
+enum {
+	MLX4_COMM_CMD_RESET,
+	MLX4_COMM_CMD_VHCR0,
+	MLX4_COMM_CMD_VHCR1,
+	MLX4_COMM_CMD_VHCR2,
+	MLX4_COMM_CMD_VHCR_EN,
+	MLX4_COMM_CMD_VHCR_POST,
+	MLX4_COMM_CMD_FLR = 254
+};
+
+enum {
+	MLX4_VF_SMI_DISABLED,
+	MLX4_VF_SMI_ENABLED
+};
+
+/*The flag indicates that the slave should delay the RESET cmd*/
+#define MLX4_DELAY_RESET_SLAVE 0xbbbbbbb
+/*indicates how many retries will be done if we are in the middle of FLR*/
+#define NUM_OF_RESET_RETRIES	10
+#define SLEEP_TIME_IN_RESET	(2 * 1000)
+enum mlx4_resource {
+	RES_QP,
+	RES_CQ,
+	RES_SRQ,
+	RES_XRCD,
+	RES_MPT,
+	RES_MTT,
+	RES_MAC,
+	RES_VLAN,
+	RES_NPORT_ID,
+	RES_COUNTER,
+	RES_FS_RULE,
+	RES_EQ,
+	MLX4_NUM_OF_RESOURCE_TYPE
+};
+
+enum mlx4_alloc_mode {
+	RES_OP_RESERVE,
+	RES_OP_RESERVE_AND_MAP,
+	RES_OP_MAP_ICM,
+};
+
+enum mlx4_res_tracker_free_type {
+	RES_TR_FREE_ALL,
+	RES_TR_FREE_SLAVES_ONLY,
+	RES_TR_FREE_STRUCTS_ONLY,
+};
+
+/*
+ *Virtual HCR structures.
+ * mlx4_vhcr is the sw representation, in machine endianness
+ *
+ * mlx4_vhcr_cmd is the formalized structure, the one that is passed
+ * to FW to go through communication channel.
+ * It is big endian, and has the same structure as the physical HCR
+ * used by command interface
+ */
+struct mlx4_vhcr {
+	u64	in_param;
+	u64	out_param;
+	u32	in_modifier;
+	u32	errno;
+	u16	op;
+	u16	token;
+	u8	op_modifier;
+	u8	e_bit;
+};
+
+struct mlx4_vhcr_cmd {
+	__be64 in_param;
+	__be32 in_modifier;
+	u32 reserved1;
+	__be64 out_param;
+	__be16 token;
+	u16 reserved;
+	u8 status;
+	u8 flags;
+	__be16 opcode;
+};
+
+struct mlx4_cmd_info {
+	u16 opcode;
+	bool has_inbox;
+	bool has_outbox;
+	bool out_is_imm;
+	bool encode_slave_id;
+	int (*verify)(struct mlx4_dev *dev, int slave, struct mlx4_vhcr *vhcr,
+		      struct mlx4_cmd_mailbox *inbox);
+	int (*wrapper)(struct mlx4_dev *dev, int slave, struct mlx4_vhcr *vhcr,
+		       struct mlx4_cmd_mailbox *inbox,
+		       struct mlx4_cmd_mailbox *outbox,
+		       struct mlx4_cmd_info *cmd);
+};
+
+#ifdef CONFIG_MLX4_DEBUG
+extern int mlx4_debug_level;
+#else /* CONFIG_MLX4_DEBUG */
+#define mlx4_debug_level	(0)
+#endif /* CONFIG_MLX4_DEBUG */
+
+#define mlx4_dbg(mdev, format, ...)					\
+do {									\
+	if (mlx4_debug_level)						\
+		dev_printk(KERN_DEBUG,					\
+			   &(mdev)->persist->pdev->dev, format,		\
+			   ##__VA_ARGS__);				\
+} while (0)
+
+#define mlx4_err(mdev, format, ...)					\
+	dev_err(&(mdev)->persist->pdev->dev, format, ##__VA_ARGS__)
+#define mlx4_info(mdev, format, ...)					\
+	dev_info(&(mdev)->persist->pdev->dev, format, ##__VA_ARGS__)
+#define mlx4_warn(mdev, format, ...)					\
+	dev_warn(&(mdev)->persist->pdev->dev, format, ##__VA_ARGS__)
+
+extern int log_mtts_per_seg;
+extern int mlx4_internal_err_reset;
+
+#define MLX4_MAX_NUM_SLAVES	(min(MLX4_MAX_NUM_PF + MLX4_MAX_NUM_VF, \
+				     MLX4_MFUNC_MAX))
+#define ALL_SLAVES 0xff
+
+struct mlx4_bitmap {
+	u32			last;
+	u32			top;
+	u32			max;
+	u32                     reserved_top;
+	u32			mask;
+	u32			avail;
+	u32			effective_len;
+	spinlock_t		lock;
+	unsigned long	       *table;
+};
+
+struct mlx4_buddy {
+	unsigned long	      **bits;
+	unsigned int	       *num_free;
+	u32			max_order;
+	spinlock_t		lock;
+};
+
+struct mlx4_icm;
+
+struct mlx4_icm_table {
+	u64			virt;
+	int			num_icm;
+	u32			num_obj;
+	int			obj_size;
+	int			lowmem;
+	int			coherent;
+	struct mutex		mutex;
+	struct mlx4_icm	      **icm;
+};
+
+#define MLX4_MPT_FLAG_SW_OWNS	    (0xfUL << 28)
+#define MLX4_MPT_FLAG_FREE	    (0x3UL << 28)
+#define MLX4_MPT_FLAG_MIO	    (1 << 17)
+#define MLX4_MPT_FLAG_BIND_ENABLE   (1 << 15)
+#define MLX4_MPT_FLAG_PHYSICAL	    (1 <<  9)
+#define MLX4_MPT_FLAG_REGION	    (1 <<  8)
+
+#define MLX4_MPT_PD_MASK	    (0x1FFFFUL)
+#define MLX4_MPT_PD_VF_MASK	    (0xFE0000UL)
+#define MLX4_MPT_PD_FLAG_FAST_REG   (1 << 27)
+#define MLX4_MPT_PD_FLAG_RAE	    (1 << 28)
+#define MLX4_MPT_PD_FLAG_EN_INV	    (3 << 24)
+
+#define MLX4_MPT_QP_FLAG_BOUND_QP   (1 << 7)
+
+#define MLX4_MPT_STATUS_SW		0xF0
+#define MLX4_MPT_STATUS_HW		0x00
+
+#define MLX4_CQE_SIZE_MASK_STRIDE	0x3
+#define MLX4_EQE_SIZE_MASK_STRIDE	0x30
+
+#define MLX4_EQ_ASYNC			0
+#define MLX4_EQ_TO_CQ_VECTOR(vector)	((vector) - \
+					 !!((int)(vector) >= MLX4_EQ_ASYNC))
+#define MLX4_CQ_TO_EQ_VECTOR(vector)	((vector) + \
+					 !!((int)(vector) >= MLX4_EQ_ASYNC))
+
+/*
+ * Must be packed because mtt_seg is 64 bits but only aligned to 32 bits.
+ */
+struct mlx4_mpt_entry {
+	__be32 flags;
+	__be32 qpn;
+	__be32 key;
+	__be32 pd_flags;
+	__be64 start;
+	__be64 length;
+	__be32 lkey;
+	__be32 win_cnt;
+	u8	reserved1[3];
+	u8	mtt_rep;
+	__be64 mtt_addr;
+	__be32 mtt_sz;
+	__be32 entity_size;
+	__be32 first_byte_offset;
+} __packed;
+
+/*
+ * Must be packed because start is 64 bits but only aligned to 32 bits.
+ */
+struct mlx4_eq_context {
+	__be32			flags;
+	u16			reserved1[3];
+	__be16			page_offset;
+	u8			log_eq_size;
+	u8			reserved2[4];
+	u8			eq_period;
+	u8			reserved3;
+	u8			eq_max_count;
+	u8			reserved4[3];
+	u8			intr;
+	u8			log_page_size;
+	u8			reserved5[2];
+	u8			mtt_base_addr_h;
+	__be32			mtt_base_addr_l;
+	u32			reserved6[2];
+	__be32			consumer_index;
+	__be32			producer_index;
+	u32			reserved7[4];
+};
+
+struct mlx4_cq_context {
+	__be32			flags;
+	u16			reserved1[3];
+	__be16			page_offset;
+	__be32			logsize_usrpage;
+	__be16			cq_period;
+	__be16			cq_max_count;
+	u8			reserved2[3];
+	u8			comp_eqn;
+	u8			log_page_size;
+	u8			reserved3[2];
+	u8			mtt_base_addr_h;
+	__be32			mtt_base_addr_l;
+	__be32			last_notified_index;
+	__be32			solicit_producer_index;
+	__be32			consumer_index;
+	__be32			producer_index;
+	u32			reserved4[2];
+	__be64			db_rec_addr;
+};
+
+struct mlx4_srq_context {
+	__be32			state_logsize_srqn;
+	u8			logstride;
+	u8			reserved1;
+	__be16			xrcd;
+	__be32			pg_offset_cqn;
+	u32			reserved2;
+	u8			log_page_size;
+	u8			reserved3[2];
+	u8			mtt_base_addr_h;
+	__be32			mtt_base_addr_l;
+	__be32			pd;
+	__be16			limit_watermark;
+	__be16			wqe_cnt;
+	u16			reserved4;
+	__be16			wqe_counter;
+	u32			reserved5;
+	__be64			db_rec_addr;
+};
+
+struct mlx4_eq_tasklet {
+	struct list_head list;
+	struct list_head process_list;
+	struct tasklet_struct task;
+	/* lock on completion tasklet list */
+	spinlock_t lock;
+};
+
+struct mlx4_eq {
+	struct mlx4_dev	       *dev;
+	void __iomem	       *doorbell;
+	int			eqn;
+	u32			cons_index;
+	u16			irq;
+	u16			have_irq;
+	int			nent;
+	struct mlx4_buf_list   *page_list;
+	struct mlx4_mtt		mtt;
+	struct mlx4_eq_tasklet	tasklet_ctx;
+	struct mlx4_active_ports actv_ports;
+	u32			ref_count;
+	cpumask_var_t		affinity_mask;
+};
+
+struct mlx4_slave_eqe {
+	u8 type;
+	u8 port;
+	u32 param;
+};
+
+struct mlx4_slave_event_eq_info {
+	int eqn;
+	u16 token;
+};
+
+struct mlx4_profile {
+	int			num_qp;
+	int			rdmarc_per_qp;
+	int			num_srq;
+	int			num_cq;
+	int			num_mcg;
+	int			num_mpt;
+	unsigned		num_mtt;
+};
+
+struct mlx4_fw {
+	u64			clr_int_base;
+	u64			catas_offset;
+	u64			comm_base;
+	u64			clock_offset;
+	struct mlx4_icm	       *fw_icm;
+	struct mlx4_icm	       *aux_icm;
+	u32			catas_size;
+	u16			fw_pages;
+	u8			clr_int_bar;
+	u8			catas_bar;
+	u8			comm_bar;
+	u8			clock_bar;
+};
+
+struct mlx4_comm {
+	u32			slave_write;
+	u32			slave_read;
+};
+
+enum {
+	MLX4_MCAST_CONFIG       = 0,
+	MLX4_MCAST_DISABLE      = 1,
+	MLX4_MCAST_ENABLE       = 2,
+};
+
+#define VLAN_FLTR_SIZE	128
+
+struct mlx4_vlan_fltr {
+	__be32 entry[VLAN_FLTR_SIZE];
+};
+
+struct mlx4_mcast_entry {
+	struct list_head list;
+	u64 addr;
+};
+
+struct mlx4_promisc_qp {
+	struct list_head list;
+	u32 qpn;
+};
+
+struct mlx4_steer_index {
+	struct list_head list;
+	unsigned int index;
+	struct list_head duplicates;
+};
+
+#define MLX4_EVENT_TYPES_NUM 64
+
+struct mlx4_slave_state {
+	u8 comm_toggle;
+	u8 last_cmd;
+	u8 init_port_mask;
+	bool active;
+	bool old_vlan_api;
+	bool vst_qinq_supported;
+	u8 function;
+	dma_addr_t vhcr_dma;
+	u16 user_mtu[MLX4_MAX_PORTS + 1];
+	u16 mtu[MLX4_MAX_PORTS + 1];
+	__be32 ib_cap_mask[MLX4_MAX_PORTS + 1];
+	struct mlx4_slave_eqe eq[MLX4_MFUNC_MAX_EQES];
+	struct list_head mcast_filters[MLX4_MAX_PORTS + 1];
+	struct mlx4_vlan_fltr *vlan_filter[MLX4_MAX_PORTS + 1];
+	/* event type to eq number lookup */
+	struct mlx4_slave_event_eq_info event_eq[MLX4_EVENT_TYPES_NUM];
+	u16 eq_pi;
+	u16 eq_ci;
+	spinlock_t lock;
+	/*initialized via the kzalloc*/
+	u8 is_slave_going_down;
+	u32 cookie;
+	enum slave_port_state port_state[MLX4_MAX_PORTS + 1];
+};
+
+#define MLX4_VGT 4095
+#define NO_INDX  (-1)
+
+struct mlx4_vport_state {
+	u64 mac;
+	u16 default_vlan;
+	u8  default_qos;
+	__be16 vlan_proto;
+	u32 tx_rate;
+	bool spoofchk;
+	u32 link_state;
+	u8 qos_vport;
+	__be64 guid;
+};
+
+struct mlx4_vf_admin_state {
+	struct mlx4_vport_state vport[MLX4_MAX_PORTS + 1];
+	u8 enable_smi[MLX4_MAX_PORTS + 1];
+};
+
+struct mlx4_vport_oper_state {
+	struct mlx4_vport_state state;
+	int mac_idx;
+	int vlan_idx;
+};
+
+struct mlx4_vf_oper_state {
+	struct mlx4_vport_oper_state vport[MLX4_MAX_PORTS + 1];
+	u8 smi_enabled[MLX4_MAX_PORTS + 1];
+};
+
+struct slave_list {
+	struct mutex mutex;
+	struct list_head res_list[MLX4_NUM_OF_RESOURCE_TYPE];
+};
+
+struct resource_allocator {
+	spinlock_t alloc_lock; /* protect quotas */
+	union {
+		unsigned int res_reserved;
+		unsigned int res_port_rsvd[MLX4_MAX_PORTS];
+	};
+	union {
+		int res_free;
+		int res_port_free[MLX4_MAX_PORTS];
+	};
+	int *quota;
+	int *allocated;
+	int *guaranteed;
+};
+
+struct mlx4_resource_tracker {
+	spinlock_t lock;
+	/* tree for each resources */
+	struct rb_root res_tree[MLX4_NUM_OF_RESOURCE_TYPE];
+	/* num_of_slave's lists, one per slave */
+	struct slave_list *slave_list;
+	struct resource_allocator res_alloc[MLX4_NUM_OF_RESOURCE_TYPE];
+};
+
+#define SLAVE_EVENT_EQ_SIZE	128
+struct mlx4_slave_event_eq {
+	u32 eqn;
+	u32 cons;
+	u32 prod;
+	spinlock_t event_lock;
+	struct mlx4_eqe event_eqe[SLAVE_EVENT_EQ_SIZE];
+};
+
+struct mlx4_qos_manager {
+	int num_of_qos_vfs;
+	DECLARE_BITMAP(priority_bm, MLX4_NUM_UP);
+};
+
+struct mlx4_master_qp0_state {
+	int proxy_qp0_active;
+	int qp0_active;
+	int port_active;
+};
+
+struct mlx4_mfunc_master_ctx {
+	struct mlx4_slave_state *slave_state;
+	struct mlx4_vf_admin_state *vf_admin;
+	struct mlx4_vf_oper_state *vf_oper;
+	struct mlx4_master_qp0_state qp0_state[MLX4_MAX_PORTS + 1];
+	int			init_port_ref[MLX4_MAX_PORTS + 1];
+	u16			max_mtu[MLX4_MAX_PORTS + 1];
+	u16			max_user_mtu[MLX4_MAX_PORTS + 1];
+	u8			pptx;
+	u8			pprx;
+	int			disable_mcast_ref[MLX4_MAX_PORTS + 1];
+	struct mlx4_resource_tracker res_tracker;
+	struct workqueue_struct *comm_wq;
+	struct work_struct	comm_work;
+	struct work_struct	slave_event_work;
+	struct work_struct	slave_flr_event_work;
+	spinlock_t		slave_state_lock;
+	__be32			comm_arm_bit_vector[4];
+	struct mlx4_eqe		cmd_eqe;
+	struct mlx4_slave_event_eq slave_eq;
+	struct mutex		gen_eqe_mutex[MLX4_MFUNC_MAX];
+	struct mlx4_qos_manager qos_ctl[MLX4_MAX_PORTS + 1];
+};
+
+struct mlx4_mfunc {
+	struct mlx4_comm __iomem       *comm;
+	struct mlx4_vhcr_cmd	       *vhcr;
+	dma_addr_t			vhcr_dma;
+
+	struct mlx4_mfunc_master_ctx	master;
+};
+
+#define MGM_QPN_MASK       0x00FFFFFF
+#define MGM_BLCK_LB_BIT    30
+
+struct mlx4_mgm {
+	__be32			next_gid_index;
+	__be32			members_count;
+	u32			reserved[2];
+	u8			gid[16];
+	__be32			qp[MLX4_MAX_QP_PER_MGM];
+};
+
+struct mlx4_cmd {
+	struct dma_pool	       *pool;
+	void __iomem	       *hcr;
+	struct mutex		slave_cmd_mutex;
+	struct semaphore	poll_sem;
+	struct semaphore	event_sem;
+	struct rw_semaphore	switch_sem;
+	int			max_cmds;
+	spinlock_t		context_lock;
+	int			free_head;
+	struct mlx4_cmd_context *context;
+	u16			token_mask;
+	u8			use_events;
+	u8			toggle;
+	u8			comm_toggle;
+	u8			initialized;
+};
+
+enum {
+	MLX4_VF_IMMED_VLAN_FLAG_VLAN = 1 << 0,
+	MLX4_VF_IMMED_VLAN_FLAG_QOS = 1 << 1,
+	MLX4_VF_IMMED_VLAN_FLAG_LINK_DISABLE = 1 << 2,
+};
+struct mlx4_vf_immed_vlan_work {
+	struct work_struct	work;
+	struct mlx4_priv	*priv;
+	int			flags;
+	int			slave;
+	int			vlan_ix;
+	int			orig_vlan_ix;
+	u8			port;
+	u8			qos;
+	u8                      qos_vport;
+	u16			vlan_id;
+	u16			orig_vlan_id;
+	__be16			vlan_proto;
+};
+
+
+struct mlx4_uar_table {
+	struct mlx4_bitmap	bitmap;
+};
+
+struct mlx4_mr_table {
+	struct mlx4_bitmap	mpt_bitmap;
+	struct mlx4_buddy	mtt_buddy;
+	u64			mtt_base;
+	u64			mpt_base;
+	struct mlx4_icm_table	mtt_table;
+	struct mlx4_icm_table	dmpt_table;
+};
+
+struct mlx4_cq_table {
+	struct mlx4_bitmap	bitmap;
+	spinlock_t		lock;
+	struct radix_tree_root	tree;
+	struct mlx4_icm_table	table;
+	struct mlx4_icm_table	cmpt_table;
+};
+
+struct mlx4_eq_table {
+	struct mlx4_bitmap	bitmap;
+	char		       *irq_names;
+	void __iomem	       *clr_int;
+	void __iomem	      **uar_map;
+	u32			clr_mask;
+	struct mlx4_eq	       *eq;
+	struct mlx4_icm_table	table;
+	struct mlx4_icm_table	cmpt_table;
+	int			have_irq;
+	u8			inta_pin;
+};
+
+struct mlx4_srq_table {
+	struct mlx4_bitmap	bitmap;
+	spinlock_t		lock;
+	struct radix_tree_root	tree;
+	struct mlx4_icm_table	table;
+	struct mlx4_icm_table	cmpt_table;
+};
+
+enum mlx4_qp_table_zones {
+	MLX4_QP_TABLE_ZONE_GENERAL,
+	MLX4_QP_TABLE_ZONE_RSS,
+	MLX4_QP_TABLE_ZONE_RAW_ETH,
+	MLX4_QP_TABLE_ZONE_NUM
+};
+
+struct mlx4_qp_table {
+	struct mlx4_bitmap	*bitmap_gen;
+	struct mlx4_zone_allocator *zones;
+	u32			zones_uids[MLX4_QP_TABLE_ZONE_NUM];
+	u32			rdmarc_base;
+	int			rdmarc_shift;
+	spinlock_t		lock;
+	struct mlx4_icm_table	qp_table;
+	struct mlx4_icm_table	auxc_table;
+	struct mlx4_icm_table	altc_table;
+	struct mlx4_icm_table	rdmarc_table;
+	struct mlx4_icm_table	cmpt_table;
+};
+
+struct mlx4_mcg_table {
+	struct mutex		mutex;
+	struct mlx4_bitmap	bitmap;
+	struct mlx4_icm_table	table;
+};
+
+struct mlx4_catas_err {
+	u32 __iomem	       *map;
+	struct timer_list	timer;
+	struct list_head	list;
+};
+
+#define MLX4_MAX_MAC_NUM	128
+#define MLX4_MAC_TABLE_SIZE	(MLX4_MAX_MAC_NUM << 3)
+
+struct mlx4_mac_table {
+	__be64			entries[MLX4_MAX_MAC_NUM];
+	int			refs[MLX4_MAX_MAC_NUM];
+	bool			is_dup[MLX4_MAX_MAC_NUM];
+	struct mutex		mutex;
+	int			total;
+	int			max;
+};
+
+#define MLX4_ROCE_GID_ENTRY_SIZE	16
+
+struct mlx4_roce_gid_entry {
+	u8 raw[MLX4_ROCE_GID_ENTRY_SIZE];
+};
+
+struct mlx4_roce_gid_table {
+	struct mlx4_roce_gid_entry	roce_gids[MLX4_ROCE_MAX_GIDS];
+	struct mutex			mutex;
+};
+
+#define MLX4_MAX_VLAN_NUM	128
+#define MLX4_VLAN_TABLE_SIZE	(MLX4_MAX_VLAN_NUM << 2)
+
+struct mlx4_vlan_table {
+	__be32			entries[MLX4_MAX_VLAN_NUM];
+	int			refs[MLX4_MAX_VLAN_NUM];
+	int			is_dup[MLX4_MAX_VLAN_NUM];
+	struct mutex		mutex;
+	int			total;
+	int			max;
+};
+
+#define SET_PORT_GEN_ALL_VALID	(MLX4_FLAG_V_MTU_MASK	| \
+				 MLX4_FLAG_V_PPRX_MASK	| \
+				 MLX4_FLAG_V_PPTX_MASK)
+#define SET_PORT_PROMISC_SHIFT		31
+#define SET_PORT_MC_PROMISC_SHIFT	30
+
+enum {
+	MCAST_DIRECT_ONLY	= 0,
+	MCAST_DIRECT		= 1,
+	MCAST_DEFAULT		= 2
+};
+
+
+struct mlx4_set_port_general_context {
+	u16 reserved1;
+	u8 flags2;
+	u8 flags;
+	union {
+		u8 ignore_fcs;
+		u8 roce_mode;
+	};
+	u8 reserved2;
+	__be16 mtu;
+	u8 pptx;
+	u8 pfctx;
+	u16 reserved3;
+	u8 pprx;
+	u8 pfcrx;
+	u16 reserved4;
+	u32 reserved5;
+	u8 phv_en;
+	u8 reserved6[5];
+	__be16 user_mtu;
+	u16 reserved7;
+	u8 user_mac[6];
+};
+
+struct mlx4_set_port_rqp_calc_context {
+	__be32 base_qpn;
+	u8 rererved;
+	u8 n_mac;
+	u8 n_vlan;
+	u8 n_prio;
+	u8 reserved2[3];
+	u8 mac_miss;
+	u8 intra_no_vlan;
+	u8 no_vlan;
+	u8 intra_vlan_miss;
+	u8 vlan_miss;
+	u8 reserved3[3];
+	u8 no_vlan_prio;
+	__be32 promisc;
+	__be32 mcast;
+};
+
+struct mlx4_port_info {
+	struct mlx4_dev	       *dev;
+	int			port;
+	char			dev_name[16];
+	struct device_attribute port_attr;
+	enum mlx4_port_type	tmp_type;
+	char			dev_mtu_name[16];
+	struct device_attribute port_mtu_attr;
+	struct mlx4_mac_table	mac_table;
+	struct mlx4_vlan_table	vlan_table;
+	struct mlx4_roce_gid_table gid_table;
+	int			base_qpn;
+	struct cpu_rmap		*rmap;
+	struct devlink_port	devlink_port;
+};
+
+struct mlx4_sense {
+	struct mlx4_dev		*dev;
+	u8			do_sense_port[MLX4_MAX_PORTS + 1];
+	u8			sense_allowed[MLX4_MAX_PORTS + 1];
+	struct delayed_work	sense_poll;
+};
+
+struct mlx4_msix_ctl {
+	DECLARE_BITMAP(pool_bm, MAX_MSIX);
+	struct mutex	pool_lock;
+};
+
+struct mlx4_steer {
+	struct list_head promisc_qps[MLX4_NUM_STEERS];
+	struct list_head steer_entries[MLX4_NUM_STEERS];
+};
+
+enum {
+	MLX4_PCI_DEV_IS_VF		= 1 << 0,
+	MLX4_PCI_DEV_FORCE_SENSE_PORT	= 1 << 1,
+};
+
+enum {
+	MLX4_NO_RR	= 0,
+	MLX4_USE_RR	= 1,
+};
+
+struct mlx4_priv {
+	struct mlx4_dev		dev;
+
+	struct list_head	dev_list;
+	struct list_head	ctx_list;
+	spinlock_t		ctx_lock;
+
+	int			pci_dev_data;
+	int                     removed;
+
+	struct list_head        pgdir_list;
+	struct mutex            pgdir_mutex;
+
+	struct mlx4_fw		fw;
+	struct mlx4_cmd		cmd;
+	struct mlx4_mfunc	mfunc;
+
+	struct mlx4_bitmap	pd_bitmap;
+	struct mlx4_bitmap	xrcd_bitmap;
+	struct mlx4_uar_table	uar_table;
+	struct mlx4_mr_table	mr_table;
+	struct mlx4_cq_table	cq_table;
+	struct mlx4_eq_table	eq_table;
+	struct mlx4_srq_table	srq_table;
+	struct mlx4_qp_table	qp_table;
+	struct mlx4_mcg_table	mcg_table;
+	struct mlx4_bitmap	counters_bitmap;
+	int			def_counter[MLX4_MAX_PORTS];
+
+	struct mlx4_catas_err	catas_err;
+
+	void __iomem	       *clr_base;
+
+	struct mlx4_uar		driver_uar;
+	void __iomem	       *kar;
+	struct mlx4_port_info	port[MLX4_MAX_PORTS + 1];
+	struct mlx4_sense       sense;
+	struct mutex		port_mutex;
+	struct mlx4_msix_ctl	msix_ctl;
+	struct mlx4_steer	*steer;
+	struct list_head	bf_list;
+	struct mutex		bf_mutex;
+	struct io_mapping	*bf_mapping;
+	void __iomem            *clock_mapping;
+	int			reserved_mtts;
+	int			fs_hash_mode;
+	u8 virt2phys_pkey[MLX4_MFUNC_MAX][MLX4_MAX_PORTS][MLX4_MAX_PORT_PKEYS];
+	struct mlx4_port_map	v2p; /* cached port mapping configuration */
+	struct mutex		bond_mutex; /* for bond mode */
+	__be64			slave_node_guids[MLX4_MFUNC_MAX];
+
+	atomic_t		opreq_count;
+	struct work_struct	opreq_task;
+};
+
+static inline struct mlx4_priv *mlx4_priv(struct mlx4_dev *dev)
+{
+	return container_of(dev, struct mlx4_priv, dev);
+}
+
+#define MLX4_SENSE_RANGE	(HZ * 3)
+
+extern struct workqueue_struct *mlx4_wq;
+
+u32 mlx4_bitmap_alloc(struct mlx4_bitmap *bitmap);
+void mlx4_bitmap_free(struct mlx4_bitmap *bitmap, u32 obj, int use_rr);
+u32 mlx4_bitmap_alloc_range(struct mlx4_bitmap *bitmap, int cnt,
+			    int align, u32 skip_mask);
+void mlx4_bitmap_free_range(struct mlx4_bitmap *bitmap, u32 obj, int cnt,
+			    int use_rr);
+u32 mlx4_bitmap_avail(struct mlx4_bitmap *bitmap);
+int mlx4_bitmap_init(struct mlx4_bitmap *bitmap, u32 num, u32 mask,
+		     u32 reserved_bot, u32 resetrved_top);
+void mlx4_bitmap_cleanup(struct mlx4_bitmap *bitmap);
+
+int mlx4_reset(struct mlx4_dev *dev);
+
+int mlx4_alloc_eq_table(struct mlx4_dev *dev);
+void mlx4_free_eq_table(struct mlx4_dev *dev);
+
+int mlx4_init_pd_table(struct mlx4_dev *dev);
+int mlx4_init_xrcd_table(struct mlx4_dev *dev);
+int mlx4_init_uar_table(struct mlx4_dev *dev);
+int mlx4_init_mr_table(struct mlx4_dev *dev);
+int mlx4_init_eq_table(struct mlx4_dev *dev);
+int mlx4_init_cq_table(struct mlx4_dev *dev);
+int mlx4_init_qp_table(struct mlx4_dev *dev);
+int mlx4_init_srq_table(struct mlx4_dev *dev);
+int mlx4_init_mcg_table(struct mlx4_dev *dev);
+
+void mlx4_cleanup_pd_table(struct mlx4_dev *dev);
+void mlx4_cleanup_xrcd_table(struct mlx4_dev *dev);
+void mlx4_cleanup_uar_table(struct mlx4_dev *dev);
+void mlx4_cleanup_mr_table(struct mlx4_dev *dev);
+void mlx4_cleanup_eq_table(struct mlx4_dev *dev);
+void mlx4_cleanup_cq_table(struct mlx4_dev *dev);
+void mlx4_cleanup_qp_table(struct mlx4_dev *dev);
+void mlx4_cleanup_srq_table(struct mlx4_dev *dev);
+void mlx4_cleanup_mcg_table(struct mlx4_dev *dev);
+int __mlx4_qp_alloc_icm(struct mlx4_dev *dev, int qpn);
+void __mlx4_qp_free_icm(struct mlx4_dev *dev, int qpn);
+int __mlx4_cq_alloc_icm(struct mlx4_dev *dev, int *cqn);
+void __mlx4_cq_free_icm(struct mlx4_dev *dev, int cqn);
+int __mlx4_srq_alloc_icm(struct mlx4_dev *dev, int *srqn);
+void __mlx4_srq_free_icm(struct mlx4_dev *dev, int srqn);
+int __mlx4_mpt_reserve(struct mlx4_dev *dev);
+void __mlx4_mpt_release(struct mlx4_dev *dev, u32 index);
+int __mlx4_mpt_alloc_icm(struct mlx4_dev *dev, u32 index);
+void __mlx4_mpt_free_icm(struct mlx4_dev *dev, u32 index);
+u32 __mlx4_alloc_mtt_range(struct mlx4_dev *dev, int order);
+void __mlx4_free_mtt_range(struct mlx4_dev *dev, u32 first_seg, int order);
+
+int mlx4_WRITE_MTT_wrapper(struct mlx4_dev *dev, int slave,
+			   struct mlx4_vhcr *vhcr,
+			   struct mlx4_cmd_mailbox *inbox,
+			   struct mlx4_cmd_mailbox *outbox,
+			   struct mlx4_cmd_info *cmd);
+int mlx4_SYNC_TPT_wrapper(struct mlx4_dev *dev, int slave,
+			   struct mlx4_vhcr *vhcr,
+			   struct mlx4_cmd_mailbox *inbox,
+			   struct mlx4_cmd_mailbox *outbox,
+			   struct mlx4_cmd_info *cmd);
+int mlx4_SW2HW_MPT_wrapper(struct mlx4_dev *dev, int slave,
+			   struct mlx4_vhcr *vhcr,
+			   struct mlx4_cmd_mailbox *inbox,
+			   struct mlx4_cmd_mailbox *outbox,
+			   struct mlx4_cmd_info *cmd);
+int mlx4_HW2SW_MPT_wrapper(struct mlx4_dev *dev, int slave,
+			   struct mlx4_vhcr *vhcr,
+			   struct mlx4_cmd_mailbox *inbox,
+			   struct mlx4_cmd_mailbox *outbox,
+			   struct mlx4_cmd_info *cmd);
+int mlx4_QUERY_MPT_wrapper(struct mlx4_dev *dev, int slave,
+			   struct mlx4_vhcr *vhcr,
+			   struct mlx4_cmd_mailbox *inbox,
+			   struct mlx4_cmd_mailbox *outbox,
+			   struct mlx4_cmd_info *cmd);
+int mlx4_SW2HW_EQ_wrapper(struct mlx4_dev *dev, int slave,
+			  struct mlx4_vhcr *vhcr,
+			  struct mlx4_cmd_mailbox *inbox,
+			  struct mlx4_cmd_mailbox *outbox,
+			  struct mlx4_cmd_info *cmd);
+int mlx4_CONFIG_DEV_wrapper(struct mlx4_dev *dev, int slave,
+			    struct mlx4_vhcr *vhcr,
+			    struct mlx4_cmd_mailbox *inbox,
+			    struct mlx4_cmd_mailbox *outbox,
+			    struct mlx4_cmd_info *cmd);
+int mlx4_DMA_wrapper(struct mlx4_dev *dev, int slave,
+		     struct mlx4_vhcr *vhcr,
+		     struct mlx4_cmd_mailbox *inbox,
+		     struct mlx4_cmd_mailbox *outbox,
+		     struct mlx4_cmd_info *cmd);
+int __mlx4_qp_reserve_range(struct mlx4_dev *dev, int cnt, int align,
+			    int *base, u8 flags);
+void __mlx4_qp_release_range(struct mlx4_dev *dev, int base_qpn, int cnt);
+int __mlx4_register_mac(struct mlx4_dev *dev, u8 port, u64 mac);
+void __mlx4_unregister_mac(struct mlx4_dev *dev, u8 port, u64 mac);
+int __mlx4_write_mtt(struct mlx4_dev *dev, struct mlx4_mtt *mtt,
+		     int start_index, int npages, u64 *page_list);
+int __mlx4_counter_alloc(struct mlx4_dev *dev, u32 *idx);
+void __mlx4_counter_free(struct mlx4_dev *dev, u32 idx);
+int mlx4_calc_vf_counters(struct mlx4_dev *dev, int slave, int port,
+			  struct mlx4_counter *data);
+int __mlx4_xrcd_alloc(struct mlx4_dev *dev, u32 *xrcdn);
+void __mlx4_xrcd_free(struct mlx4_dev *dev, u32 xrcdn);
+
+void mlx4_start_catas_poll(struct mlx4_dev *dev);
+void mlx4_stop_catas_poll(struct mlx4_dev *dev);
+int mlx4_catas_init(struct mlx4_dev *dev);
+void mlx4_catas_end(struct mlx4_dev *dev);
+int mlx4_crdump_init(struct mlx4_dev *dev);
+void mlx4_crdump_end(struct mlx4_dev *dev);
+int mlx4_restart_one(struct pci_dev *pdev, bool reload,
+		     struct devlink *devlink);
+int mlx4_register_device(struct mlx4_dev *dev);
+void mlx4_unregister_device(struct mlx4_dev *dev);
+void mlx4_dispatch_event(struct mlx4_dev *dev, enum mlx4_dev_event type,
+			 unsigned long param);
+
+struct mlx4_dev_cap;
+struct mlx4_init_hca_param;
+
+u64 mlx4_make_profile(struct mlx4_dev *dev,
+		      struct mlx4_profile *request,
+		      struct mlx4_dev_cap *dev_cap,
+		      struct mlx4_init_hca_param *init_hca);
+void mlx4_master_comm_channel(struct work_struct *work);
+void mlx4_gen_slave_eqe(struct work_struct *work);
+void mlx4_master_handle_slave_flr(struct work_struct *work);
+
+int mlx4_ALLOC_RES_wrapper(struct mlx4_dev *dev, int slave,
+			   struct mlx4_vhcr *vhcr,
+			   struct mlx4_cmd_mailbox *inbox,
+			   struct mlx4_cmd_mailbox *outbox,
+			   struct mlx4_cmd_info *cmd);
+int mlx4_FREE_RES_wrapper(struct mlx4_dev *dev, int slave,
+			  struct mlx4_vhcr *vhcr,
+			  struct mlx4_cmd_mailbox *inbox,
+			  struct mlx4_cmd_mailbox *outbox,
+			  struct mlx4_cmd_info *cmd);
+int mlx4_MAP_EQ_wrapper(struct mlx4_dev *dev, int slave,
+			struct mlx4_vhcr *vhcr, struct mlx4_cmd_mailbox *inbox,
+			struct mlx4_cmd_mailbox *outbox,
+			struct mlx4_cmd_info *cmd);
+int mlx4_COMM_INT_wrapper(struct mlx4_dev *dev, int slave,
+			  struct mlx4_vhcr *vhcr,
+			  struct mlx4_cmd_mailbox *inbox,
+			  struct mlx4_cmd_mailbox *outbox,
+			  struct mlx4_cmd_info *cmd);
+int mlx4_HW2SW_EQ_wrapper(struct mlx4_dev *dev, int slave,
+			    struct mlx4_vhcr *vhcr,
+			    struct mlx4_cmd_mailbox *inbox,
+			    struct mlx4_cmd_mailbox *outbox,
+			  struct mlx4_cmd_info *cmd);
+int mlx4_QUERY_EQ_wrapper(struct mlx4_dev *dev, int slave,
+			  struct mlx4_vhcr *vhcr,
+			  struct mlx4_cmd_mailbox *inbox,
+			  struct mlx4_cmd_mailbox *outbox,
+			  struct mlx4_cmd_info *cmd);
+int mlx4_SW2HW_CQ_wrapper(struct mlx4_dev *dev, int slave,
+			  struct mlx4_vhcr *vhcr,
+			  struct mlx4_cmd_mailbox *inbox,
+			  struct mlx4_cmd_mailbox *outbox,
+			  struct mlx4_cmd_info *cmd);
+int mlx4_HW2SW_CQ_wrapper(struct mlx4_dev *dev, int slave,
+			  struct mlx4_vhcr *vhcr,
+			  struct mlx4_cmd_mailbox *inbox,
+			  struct mlx4_cmd_mailbox *outbox,
+			  struct mlx4_cmd_info *cmd);
+int mlx4_QUERY_CQ_wrapper(struct mlx4_dev *dev, int slave,
+			  struct mlx4_vhcr *vhcr,
+			  struct mlx4_cmd_mailbox *inbox,
+			  struct mlx4_cmd_mailbox *outbox,
+			  struct mlx4_cmd_info *cmd);
+int mlx4_MODIFY_CQ_wrapper(struct mlx4_dev *dev, int slave,
+			  struct mlx4_vhcr *vhcr,
+			  struct mlx4_cmd_mailbox *inbox,
+			  struct mlx4_cmd_mailbox *outbox,
+			   struct mlx4_cmd_info *cmd);
+int mlx4_SW2HW_SRQ_wrapper(struct mlx4_dev *dev, int slave,
+			   struct mlx4_vhcr *vhcr,
+			   struct mlx4_cmd_mailbox *inbox,
+			   struct mlx4_cmd_mailbox *outbox,
+			   struct mlx4_cmd_info *cmd);
+int mlx4_HW2SW_SRQ_wrapper(struct mlx4_dev *dev, int slave,
+			   struct mlx4_vhcr *vhcr,
+			   struct mlx4_cmd_mailbox *inbox,
+			   struct mlx4_cmd_mailbox *outbox,
+			   struct mlx4_cmd_info *cmd);
+int mlx4_QUERY_SRQ_wrapper(struct mlx4_dev *dev, int slave,
+			   struct mlx4_vhcr *vhcr,
+			   struct mlx4_cmd_mailbox *inbox,
+			   struct mlx4_cmd_mailbox *outbox,
+			   struct mlx4_cmd_info *cmd);
+int mlx4_ARM_SRQ_wrapper(struct mlx4_dev *dev, int slave,
+			 struct mlx4_vhcr *vhcr,
+			 struct mlx4_cmd_mailbox *inbox,
+			 struct mlx4_cmd_mailbox *outbox,
+			 struct mlx4_cmd_info *cmd);
+int mlx4_GEN_QP_wrapper(struct mlx4_dev *dev, int slave,
+			struct mlx4_vhcr *vhcr,
+			struct mlx4_cmd_mailbox *inbox,
+			struct mlx4_cmd_mailbox *outbox,
+			struct mlx4_cmd_info *cmd);
+int mlx4_RST2INIT_QP_wrapper(struct mlx4_dev *dev, int slave,
+			     struct mlx4_vhcr *vhcr,
+			     struct mlx4_cmd_mailbox *inbox,
+			     struct mlx4_cmd_mailbox *outbox,
+			     struct mlx4_cmd_info *cmd);
+int mlx4_INIT2INIT_QP_wrapper(struct mlx4_dev *dev, int slave,
+			      struct mlx4_vhcr *vhcr,
+			      struct mlx4_cmd_mailbox *inbox,
+			      struct mlx4_cmd_mailbox *outbox,
+			      struct mlx4_cmd_info *cmd);
+int mlx4_INIT2RTR_QP_wrapper(struct mlx4_dev *dev, int slave,
+			     struct mlx4_vhcr *vhcr,
+			     struct mlx4_cmd_mailbox *inbox,
+			     struct mlx4_cmd_mailbox *outbox,
+			     struct mlx4_cmd_info *cmd);
+int mlx4_RTR2RTS_QP_wrapper(struct mlx4_dev *dev, int slave,
+			    struct mlx4_vhcr *vhcr,
+			    struct mlx4_cmd_mailbox *inbox,
+			    struct mlx4_cmd_mailbox *outbox,
+			    struct mlx4_cmd_info *cmd);
+int mlx4_RTS2RTS_QP_wrapper(struct mlx4_dev *dev, int slave,
+			    struct mlx4_vhcr *vhcr,
+			    struct mlx4_cmd_mailbox *inbox,
+			    struct mlx4_cmd_mailbox *outbox,
+			    struct mlx4_cmd_info *cmd);
+int mlx4_SQERR2RTS_QP_wrapper(struct mlx4_dev *dev, int slave,
+			      struct mlx4_vhcr *vhcr,
+			      struct mlx4_cmd_mailbox *inbox,
+			      struct mlx4_cmd_mailbox *outbox,
+			      struct mlx4_cmd_info *cmd);
+int mlx4_2ERR_QP_wrapper(struct mlx4_dev *dev, int slave,
+			 struct mlx4_vhcr *vhcr,
+			 struct mlx4_cmd_mailbox *inbox,
+			 struct mlx4_cmd_mailbox *outbox,
+			 struct mlx4_cmd_info *cmd);
+int mlx4_RTS2SQD_QP_wrapper(struct mlx4_dev *dev, int slave,
+			    struct mlx4_vhcr *vhcr,
+			    struct mlx4_cmd_mailbox *inbox,
+			    struct mlx4_cmd_mailbox *outbox,
+			    struct mlx4_cmd_info *cmd);
+int mlx4_SQD2SQD_QP_wrapper(struct mlx4_dev *dev, int slave,
+			    struct mlx4_vhcr *vhcr,
+			    struct mlx4_cmd_mailbox *inbox,
+			    struct mlx4_cmd_mailbox *outbox,
+			    struct mlx4_cmd_info *cmd);
+int mlx4_SQD2RTS_QP_wrapper(struct mlx4_dev *dev, int slave,
+			    struct mlx4_vhcr *vhcr,
+			    struct mlx4_cmd_mailbox *inbox,
+			    struct mlx4_cmd_mailbox *outbox,
+			    struct mlx4_cmd_info *cmd);
+int mlx4_2RST_QP_wrapper(struct mlx4_dev *dev, int slave,
+			 struct mlx4_vhcr *vhcr,
+			 struct mlx4_cmd_mailbox *inbox,
+			 struct mlx4_cmd_mailbox *outbox,
+			 struct mlx4_cmd_info *cmd);
+int mlx4_QUERY_QP_wrapper(struct mlx4_dev *dev, int slave,
+			  struct mlx4_vhcr *vhcr,
+			  struct mlx4_cmd_mailbox *inbox,
+			  struct mlx4_cmd_mailbox *outbox,
+			  struct mlx4_cmd_info *cmd);
+
+int mlx4_GEN_EQE(struct mlx4_dev *dev, int slave, struct mlx4_eqe *eqe);
+
+enum {
+	MLX4_CMD_CLEANUP_STRUCT = 1UL << 0,
+	MLX4_CMD_CLEANUP_POOL	= 1UL << 1,
+	MLX4_CMD_CLEANUP_HCR	= 1UL << 2,
+	MLX4_CMD_CLEANUP_VHCR	= 1UL << 3,
+	MLX4_CMD_CLEANUP_ALL	= (MLX4_CMD_CLEANUP_VHCR << 1) - 1
+};
+
+int mlx4_cmd_init(struct mlx4_dev *dev);
+void mlx4_cmd_cleanup(struct mlx4_dev *dev, int cleanup_mask);
+int mlx4_multi_func_init(struct mlx4_dev *dev);
+int mlx4_ARM_COMM_CHANNEL(struct mlx4_dev *dev);
+void mlx4_multi_func_cleanup(struct mlx4_dev *dev);
+void mlx4_cmd_event(struct mlx4_dev *dev, u16 token, u8 status, u64 out_param);
+int mlx4_cmd_use_events(struct mlx4_dev *dev);
+void mlx4_cmd_use_polling(struct mlx4_dev *dev);
+
+int mlx4_comm_cmd(struct mlx4_dev *dev, u8 cmd, u16 param,
+		  u16 op, unsigned long timeout);
+
+void mlx4_cq_tasklet_cb(unsigned long data);
+void mlx4_cq_completion(struct mlx4_dev *dev, u32 cqn);
+void mlx4_cq_event(struct mlx4_dev *dev, u32 cqn, int event_type);
+
+void mlx4_qp_event(struct mlx4_dev *dev, u32 qpn, int event_type);
+
+void mlx4_srq_event(struct mlx4_dev *dev, u32 srqn, int event_type);
+
+void mlx4_enter_error_state(struct mlx4_dev_persistent *persist);
+int mlx4_comm_internal_err(u32 slave_read);
+
+int mlx4_crdump_collect(struct mlx4_dev *dev);
+
+int mlx4_SENSE_PORT(struct mlx4_dev *dev, int port,
+		    enum mlx4_port_type *type);
+void mlx4_do_sense_ports(struct mlx4_dev *dev,
+			 enum mlx4_port_type *stype,
+			 enum mlx4_port_type *defaults);
+void mlx4_start_sense(struct mlx4_dev *dev);
+void mlx4_stop_sense(struct mlx4_dev *dev);
+void mlx4_sense_init(struct mlx4_dev *dev);
+int mlx4_check_port_params(struct mlx4_dev *dev,
+			   enum mlx4_port_type *port_type);
+int mlx4_change_port_types(struct mlx4_dev *dev,
+			   enum mlx4_port_type *port_types);
+
+void mlx4_init_mac_table(struct mlx4_dev *dev, struct mlx4_mac_table *table);
+void mlx4_init_vlan_table(struct mlx4_dev *dev, struct mlx4_vlan_table *table);
+void mlx4_init_roce_gid_table(struct mlx4_dev *dev,
+			      struct mlx4_roce_gid_table *table);
+void __mlx4_unregister_vlan(struct mlx4_dev *dev, u8 port, u16 vlan);
+int __mlx4_register_vlan(struct mlx4_dev *dev, u8 port, u16 vlan, int *index);
+int mlx4_bond_vlan_table(struct mlx4_dev *dev);
+int mlx4_unbond_vlan_table(struct mlx4_dev *dev);
+int mlx4_bond_mac_table(struct mlx4_dev *dev);
+int mlx4_unbond_mac_table(struct mlx4_dev *dev);
+
+int mlx4_SET_PORT(struct mlx4_dev *dev, u8 port, int pkey_tbl_sz);
+/* resource tracker functions*/
+int mlx4_get_slave_from_resource_id(struct mlx4_dev *dev,
+				    enum mlx4_resource resource_type,
+				    u64 resource_id, int *slave);
+void mlx4_delete_all_resources_for_slave(struct mlx4_dev *dev, int slave_id);
+void mlx4_reset_roce_gids(struct mlx4_dev *dev, int slave);
+int mlx4_init_resource_tracker(struct mlx4_dev *dev);
+
+void mlx4_free_resource_tracker(struct mlx4_dev *dev,
+				enum mlx4_res_tracker_free_type type);
+
+int mlx4_QUERY_FW_wrapper(struct mlx4_dev *dev, int slave,
+			  struct mlx4_vhcr *vhcr,
+			  struct mlx4_cmd_mailbox *inbox,
+			  struct mlx4_cmd_mailbox *outbox,
+			  struct mlx4_cmd_info *cmd);
+int mlx4_SET_PORT_wrapper(struct mlx4_dev *dev, int slave,
+			  struct mlx4_vhcr *vhcr,
+			  struct mlx4_cmd_mailbox *inbox,
+			  struct mlx4_cmd_mailbox *outbox,
+			  struct mlx4_cmd_info *cmd);
+int mlx4_INIT_PORT_wrapper(struct mlx4_dev *dev, int slave,
+			   struct mlx4_vhcr *vhcr,
+			   struct mlx4_cmd_mailbox *inbox,
+			   struct mlx4_cmd_mailbox *outbox,
+			   struct mlx4_cmd_info *cmd);
+int mlx4_CLOSE_PORT_wrapper(struct mlx4_dev *dev, int slave,
+			    struct mlx4_vhcr *vhcr,
+			    struct mlx4_cmd_mailbox *inbox,
+			    struct mlx4_cmd_mailbox *outbox,
+			    struct mlx4_cmd_info *cmd);
+int mlx4_QUERY_DEV_CAP_wrapper(struct mlx4_dev *dev, int slave,
+			       struct mlx4_vhcr *vhcr,
+			       struct mlx4_cmd_mailbox *inbox,
+			       struct mlx4_cmd_mailbox *outbox,
+			       struct mlx4_cmd_info *cmd);
+int mlx4_QUERY_PORT_wrapper(struct mlx4_dev *dev, int slave,
+			    struct mlx4_vhcr *vhcr,
+			    struct mlx4_cmd_mailbox *inbox,
+			    struct mlx4_cmd_mailbox *outbox,
+			    struct mlx4_cmd_info *cmd);
+int mlx4_get_port_ib_caps(struct mlx4_dev *dev, u8 port, __be32 *caps);
+
+int mlx4_get_slave_pkey_gid_tbl_len(struct mlx4_dev *dev, u8 port,
+				    int *gid_tbl_len, int *pkey_tbl_len);
+
+int mlx4_QP_ATTACH_wrapper(struct mlx4_dev *dev, int slave,
+			   struct mlx4_vhcr *vhcr,
+			   struct mlx4_cmd_mailbox *inbox,
+			   struct mlx4_cmd_mailbox *outbox,
+			   struct mlx4_cmd_info *cmd);
+
+int mlx4_UPDATE_QP_wrapper(struct mlx4_dev *dev, int slave,
+			   struct mlx4_vhcr *vhcr,
+			   struct mlx4_cmd_mailbox *inbox,
+			   struct mlx4_cmd_mailbox *outbox,
+			   struct mlx4_cmd_info *cmd);
+
+int mlx4_PROMISC_wrapper(struct mlx4_dev *dev, int slave,
+			 struct mlx4_vhcr *vhcr,
+			 struct mlx4_cmd_mailbox *inbox,
+			 struct mlx4_cmd_mailbox *outbox,
+			 struct mlx4_cmd_info *cmd);
+int mlx4_qp_detach_common(struct mlx4_dev *dev, struct mlx4_qp *qp, u8 gid[16],
+			  enum mlx4_protocol prot, enum mlx4_steer_type steer);
+int mlx4_qp_attach_common(struct mlx4_dev *dev, struct mlx4_qp *qp, u8 gid[16],
+			  int block_mcast_loopback, enum mlx4_protocol prot,
+			  enum mlx4_steer_type steer);
+int mlx4_trans_to_dmfs_attach(struct mlx4_dev *dev, struct mlx4_qp *qp,
+			      u8 gid[16], u8 port,
+			      int block_mcast_loopback,
+			      enum mlx4_protocol prot, u64 *reg_id);
+int mlx4_SET_MCAST_FLTR_wrapper(struct mlx4_dev *dev, int slave,
+				struct mlx4_vhcr *vhcr,
+				struct mlx4_cmd_mailbox *inbox,
+				struct mlx4_cmd_mailbox *outbox,
+				struct mlx4_cmd_info *cmd);
+int mlx4_SET_VLAN_FLTR_wrapper(struct mlx4_dev *dev, int slave,
+			       struct mlx4_vhcr *vhcr,
+			       struct mlx4_cmd_mailbox *inbox,
+			       struct mlx4_cmd_mailbox *outbox,
+			       struct mlx4_cmd_info *cmd);
+int mlx4_common_set_vlan_fltr(struct mlx4_dev *dev, int function,
+				     int port, void *buf);
+int mlx4_DUMP_ETH_STATS_wrapper(struct mlx4_dev *dev, int slave,
+				   struct mlx4_vhcr *vhcr,
+				   struct mlx4_cmd_mailbox *inbox,
+				   struct mlx4_cmd_mailbox *outbox,
+				struct mlx4_cmd_info *cmd);
+int mlx4_PKEY_TABLE_wrapper(struct mlx4_dev *dev, int slave,
+			    struct mlx4_vhcr *vhcr,
+			    struct mlx4_cmd_mailbox *inbox,
+			    struct mlx4_cmd_mailbox *outbox,
+			    struct mlx4_cmd_info *cmd);
+int mlx4_QUERY_IF_STAT_wrapper(struct mlx4_dev *dev, int slave,
+			       struct mlx4_vhcr *vhcr,
+			       struct mlx4_cmd_mailbox *inbox,
+			       struct mlx4_cmd_mailbox *outbox,
+			       struct mlx4_cmd_info *cmd);
+int mlx4_QP_FLOW_STEERING_ATTACH_wrapper(struct mlx4_dev *dev, int slave,
+					 struct mlx4_vhcr *vhcr,
+					 struct mlx4_cmd_mailbox *inbox,
+					 struct mlx4_cmd_mailbox *outbox,
+					 struct mlx4_cmd_info *cmd);
+int mlx4_QP_FLOW_STEERING_DETACH_wrapper(struct mlx4_dev *dev, int slave,
+					 struct mlx4_vhcr *vhcr,
+					 struct mlx4_cmd_mailbox *inbox,
+					 struct mlx4_cmd_mailbox *outbox,
+					 struct mlx4_cmd_info *cmd);
+int mlx4_ACCESS_REG_wrapper(struct mlx4_dev *dev, int slave,
+			    struct mlx4_vhcr *vhcr,
+			    struct mlx4_cmd_mailbox *inbox,
+			    struct mlx4_cmd_mailbox *outbox,
+			    struct mlx4_cmd_info *cmd);
+
+int mlx4_get_mgm_entry_size(struct mlx4_dev *dev);
+int mlx4_get_qp_per_mgm(struct mlx4_dev *dev);
+
+static inline void set_param_l(u64 *arg, u32 val)
+{
+	*arg = (*arg & 0xffffffff00000000ULL) | (u64) val;
+}
+
+static inline void set_param_h(u64 *arg, u32 val)
+{
+	*arg = (*arg & 0xffffffff) | ((u64) val << 32);
+}
+
+static inline u32 get_param_l(u64 *arg)
+{
+	return (u32) (*arg & 0xffffffff);
+}
+
+static inline u32 get_param_h(u64 *arg)
+{
+	return (u32)(*arg >> 32);
+}
+
+static inline spinlock_t *mlx4_tlock(struct mlx4_dev *dev)
+{
+	return &mlx4_priv(dev)->mfunc.master.res_tracker.lock;
+}
+
+#define NOT_MASKED_PD_BITS 17
+
+void mlx4_vf_immed_vlan_work_handler(struct work_struct *_work);
+
+void mlx4_init_quotas(struct mlx4_dev *dev);
+
+/* for VFs, replace zero MACs with randomly-generated MACs at driver start */
+void mlx4_replace_zero_macs(struct mlx4_dev *dev);
+int mlx4_get_slave_num_gids(struct mlx4_dev *dev, int slave, int port);
+/* Returns the VF index of slave */
+int mlx4_get_vf_indx(struct mlx4_dev *dev, int slave);
+int mlx4_config_mad_demux(struct mlx4_dev *dev);
+int mlx4_do_bond(struct mlx4_dev *dev, bool enable);
+int mlx4_bond_fs_rules(struct mlx4_dev *dev);
+int mlx4_unbond_fs_rules(struct mlx4_dev *dev);
+
+enum mlx4_zone_flags {
+	MLX4_ZONE_ALLOW_ALLOC_FROM_LOWER_PRIO	= 1UL << 0,
+	MLX4_ZONE_ALLOW_ALLOC_FROM_EQ_PRIO	= 1UL << 1,
+	MLX4_ZONE_FALLBACK_TO_HIGHER_PRIO	= 1UL << 2,
+	MLX4_ZONE_USE_RR			= 1UL << 3,
+};
+
+enum mlx4_zone_alloc_flags {
+	/* No two objects could overlap between zones. UID
+	 * could be left unused. If this flag is given and
+	 * two overlapped zones are used, an object will be free'd
+	 * from the smallest possible matching zone.
+	 */
+	MLX4_ZONE_ALLOC_FLAGS_NO_OVERLAP	= 1UL << 0,
+};
+
+struct mlx4_zone_allocator;
+
+/* Create a new zone allocator */
+struct mlx4_zone_allocator *mlx4_zone_allocator_create(enum mlx4_zone_alloc_flags flags);
+
+/* Attach a mlx4_bitmap <bitmap> of priority <priority> to the zone allocator
+ * <zone_alloc>. Allocating an object from this zone adds an offset <offset>.
+ * Similarly, when searching for an object to free, this offset it taken into
+ * account. The use_rr mlx4_ib parameter for allocating objects from this <bitmap>
+ * is given through the MLX4_ZONE_USE_RR flag in <flags>.
+ * When an allocation fails, <zone_alloc> tries to allocate from other zones
+ * according to the policy set by <flags>. <puid> is the unique identifier
+ * received to this zone.
+ */
+int mlx4_zone_add_one(struct mlx4_zone_allocator *zone_alloc,
+		      struct mlx4_bitmap *bitmap,
+		      u32 flags,
+		      int priority,
+		      int offset,
+		      u32 *puid);
+
+/* Remove bitmap indicated by <uid> from <zone_alloc> */
+int mlx4_zone_remove_one(struct mlx4_zone_allocator *zone_alloc, u32 uid);
+
+/* Delete the zone allocator <zone_alloc. This function doesn't destroy
+ * the attached bitmaps.
+ */
+void mlx4_zone_allocator_destroy(struct mlx4_zone_allocator *zone_alloc);
+
+/* Allocate <count> objects with align <align> and skip_mask <skip_mask>
+ * from the mlx4_bitmap whose uid is <uid>. The bitmap which we actually
+ * allocated from is returned in <puid>. If the allocation fails, a negative
+ * number is returned. Otherwise, the offset of the first object is returned.
+ */
+u32 mlx4_zone_alloc_entries(struct mlx4_zone_allocator *zones, u32 uid, int count,
+			    int align, u32 skip_mask, u32 *puid);
+
+/* Free <count> objects, start from <obj> of the uid <uid> from zone_allocator
+ * <zones>.
+ */
+u32 mlx4_zone_free_entries(struct mlx4_zone_allocator *zones,
+			   u32 uid, u32 obj, u32 count);
+
+/* If <zones> was allocated with MLX4_ZONE_ALLOC_FLAGS_NO_OVERLAP, instead of
+ * specifying the uid when freeing an object, zone allocator could figure it by
+ * itself. Other parameters are similar to mlx4_zone_free.
+ */
+u32 mlx4_zone_free_entries_unique(struct mlx4_zone_allocator *zones, u32 obj, u32 count);
+
+/* Returns a pointer to mlx4_bitmap that was attached to <zones> with <uid> */
+struct mlx4_bitmap *mlx4_zone_get_bitmap(struct mlx4_zone_allocator *zones, u32 uid);
+
+#endif /* MLX4_H */
diff --git a/drivers/net/ethernet/mellanox/mlx4/mlx4_en.h b/drivers/net/ethernet/mellanox/mlx4/mlx4_en.h
new file mode 100644
index 000000000..3d5597d5b
--- /dev/null
+++ b/drivers/net/ethernet/mellanox/mlx4/mlx4_en.h
@@ -0,0 +1,858 @@
+/*
+ * Copyright (c) 2007 Mellanox Technologies. All rights reserved.
+ *
+ * This software is available to you under a choice of one of two
+ * licenses.  You may choose to be licensed under the terms of the GNU
+ * General Public License (GPL) Version 2, available from the file
+ * COPYING in the main directory of this source tree, or the
+ * OpenIB.org BSD license below:
+ *
+ *     Redistribution and use in source and binary forms, with or
+ *     without modification, are permitted provided that the following
+ *     conditions are met:
+ *
+ *      - Redistributions of source code must retain the above
+ *        copyright notice, this list of conditions and the following
+ *        disclaimer.
+ *
+ *      - Redistributions in binary form must reproduce the above
+ *        copyright notice, this list of conditions and the following
+ *        disclaimer in the documentation and/or other materials
+ *        provided with the distribution.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
+ * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
+ * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS
+ * BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN
+ * ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
+ * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+ * SOFTWARE.
+ *
+ */
+
+#ifndef _MLX4_EN_H_
+#define _MLX4_EN_H_
+
+#include <linux/bitops.h>
+#include <linux/compiler.h>
+#include <linux/list.h>
+#include <linux/mutex.h>
+#include <linux/netdevice.h>
+#include <linux/if_vlan.h>
+#include <linux/net_tstamp.h>
+#ifdef CONFIG_MLX4_EN_DCB
+#include <linux/dcbnl.h>
+#endif
+#include <linux/cpu_rmap.h>
+#include <linux/ptp_clock_kernel.h>
+#include <net/xdp.h>
+
+#include <linux/mlx4/device.h>
+#include <linux/mlx4/qp.h>
+#include <linux/mlx4/cq.h>
+#include <linux/mlx4/srq.h>
+#include <linux/mlx4/doorbell.h>
+#include <linux/mlx4/cmd.h>
+
+#include "en_port.h"
+#include "mlx4_stats.h"
+
+#define DRV_NAME	"mlx4_en"
+#define DRV_VERSION	"4.0-0"
+
+#define MLX4_EN_MSG_LEVEL	(NETIF_MSG_LINK | NETIF_MSG_IFDOWN)
+
+/*
+ * Device constants
+ */
+
+
+#define MLX4_EN_PAGE_SHIFT	12
+#define MLX4_EN_PAGE_SIZE	(1 << MLX4_EN_PAGE_SHIFT)
+#define DEF_RX_RINGS		16
+#define MAX_RX_RINGS		128
+#define MIN_RX_RINGS		4
+#define LOG_TXBB_SIZE		6
+#define TXBB_SIZE		BIT(LOG_TXBB_SIZE)
+#define HEADROOM		(2048 / TXBB_SIZE + 1)
+#define STAMP_STRIDE		64
+#define STAMP_DWORDS		(STAMP_STRIDE / 4)
+#define STAMP_SHIFT		31
+#define STAMP_VAL		0x7fffffff
+#define STATS_DELAY		(HZ / 4)
+#define SERVICE_TASK_DELAY	(HZ / 4)
+#define MAX_NUM_OF_FS_RULES	256
+
+#define MLX4_EN_FILTER_HASH_SHIFT 4
+#define MLX4_EN_FILTER_EXPIRY_QUOTA 60
+
+/* Typical TSO descriptor with 16 gather entries is 352 bytes... */
+#define MAX_DESC_SIZE		512
+#define MAX_DESC_TXBBS		(MAX_DESC_SIZE / TXBB_SIZE)
+
+/*
+ * OS related constants and tunables
+ */
+
+#define MLX4_EN_PRIV_FLAGS_BLUEFLAME 1
+#define MLX4_EN_PRIV_FLAGS_PHV	     2
+
+#define MLX4_EN_WATCHDOG_TIMEOUT	(15 * HZ)
+
+/* Use the maximum between 16384 and a single page */
+#define MLX4_EN_ALLOC_SIZE	PAGE_ALIGN(16384)
+
+#define MLX4_EN_MAX_RX_FRAGS	4
+
+/* Maximum ring sizes */
+#define MLX4_EN_MAX_TX_SIZE	8192
+#define MLX4_EN_MAX_RX_SIZE	8192
+
+/* Minimum ring size for our page-allocation scheme to work */
+#define MLX4_EN_MIN_RX_SIZE	(MLX4_EN_ALLOC_SIZE / SMP_CACHE_BYTES)
+#define MLX4_EN_MIN_TX_SIZE	(4096 / TXBB_SIZE)
+
+#define MLX4_EN_SMALL_PKT_SIZE		64
+#define MLX4_EN_MIN_TX_RING_P_UP	1
+#define MLX4_EN_MAX_TX_RING_P_UP	32
+#define MLX4_EN_NUM_UP_LOW		1
+#define MLX4_EN_NUM_UP_HIGH		8
+#define MLX4_EN_DEF_RX_RING_SIZE  	1024
+#define MLX4_EN_DEF_TX_RING_SIZE	MLX4_EN_DEF_RX_RING_SIZE
+#define MAX_TX_RINGS			(MLX4_EN_MAX_TX_RING_P_UP * \
+					 MLX4_EN_NUM_UP_HIGH)
+
+#define MLX4_EN_DEFAULT_TX_WORK		256
+
+/* Target number of packets to coalesce with interrupt moderation */
+#define MLX4_EN_RX_COAL_TARGET	44
+#define MLX4_EN_RX_COAL_TIME	0x10
+
+#define MLX4_EN_TX_COAL_PKTS	16
+#define MLX4_EN_TX_COAL_TIME	0x10
+
+#define MLX4_EN_MAX_COAL_PKTS	U16_MAX
+#define MLX4_EN_MAX_COAL_TIME	U16_MAX
+
+#define MLX4_EN_RX_RATE_LOW		400000
+#define MLX4_EN_RX_COAL_TIME_LOW	0
+#define MLX4_EN_RX_RATE_HIGH		450000
+#define MLX4_EN_RX_COAL_TIME_HIGH	128
+#define MLX4_EN_RX_SIZE_THRESH		1024
+#define MLX4_EN_RX_RATE_THRESH		(1000000 / MLX4_EN_RX_COAL_TIME_HIGH)
+#define MLX4_EN_SAMPLE_INTERVAL		0
+#define MLX4_EN_AVG_PKT_SMALL		256
+
+#define MLX4_EN_AUTO_CONF	0xffff
+
+#define MLX4_EN_DEF_RX_PAUSE	1
+#define MLX4_EN_DEF_TX_PAUSE	1
+
+/* Interval between successive polls in the Tx routine when polling is used
+   instead of interrupts (in per-core Tx rings) - should be power of 2 */
+#define MLX4_EN_TX_POLL_MODER	16
+#define MLX4_EN_TX_POLL_TIMEOUT	(HZ / 4)
+
+#define SMALL_PACKET_SIZE      (256 - NET_IP_ALIGN)
+#define HEADER_COPY_SIZE       (128 - NET_IP_ALIGN)
+#define MLX4_LOOPBACK_TEST_PAYLOAD (HEADER_COPY_SIZE - ETH_HLEN)
+#define PREAMBLE_LEN           8
+#define MLX4_SELFTEST_LB_MIN_MTU (MLX4_LOOPBACK_TEST_PAYLOAD + NET_IP_ALIGN + \
+				  ETH_HLEN + PREAMBLE_LEN)
+
+/* VLAN_HLEN is added twice,to support skb vlan tagged with multiple
+ * headers. (For example: ETH_P_8021Q and ETH_P_8021AD).
+ */
+#define MLX4_EN_EFF_MTU(mtu)	((mtu) + ETH_HLEN + (2 * VLAN_HLEN))
+#define ETH_BCAST		0xffffffffffffULL
+
+#define MLX4_EN_LOOPBACK_RETRIES	5
+#define MLX4_EN_LOOPBACK_TIMEOUT	100
+
+#ifdef MLX4_EN_PERF_STAT
+/* Number of samples to 'average' */
+#define AVG_SIZE			128
+#define AVG_FACTOR			1024
+
+#define INC_PERF_COUNTER(cnt)		(++(cnt))
+#define ADD_PERF_COUNTER(cnt, add)	((cnt) += (add))
+#define AVG_PERF_COUNTER(cnt, sample) \
+	((cnt) = ((cnt) * (AVG_SIZE - 1) + (sample) * AVG_FACTOR) / AVG_SIZE)
+#define GET_PERF_COUNTER(cnt)		(cnt)
+#define GET_AVG_PERF_COUNTER(cnt)	((cnt) / AVG_FACTOR)
+
+#else
+
+#define INC_PERF_COUNTER(cnt)		do {} while (0)
+#define ADD_PERF_COUNTER(cnt, add)	do {} while (0)
+#define AVG_PERF_COUNTER(cnt, sample)	do {} while (0)
+#define GET_PERF_COUNTER(cnt)		(0)
+#define GET_AVG_PERF_COUNTER(cnt)	(0)
+#endif /* MLX4_EN_PERF_STAT */
+
+/* Constants for TX flow */
+enum {
+	MAX_INLINE = 104, /* 128 - 16 - 4 - 4 */
+	MAX_BF = 256,
+	MIN_PKT_LEN = 17,
+};
+
+/*
+ * Configurables
+ */
+
+enum cq_type {
+	/* keep tx types first */
+	TX,
+	TX_XDP,
+#define MLX4_EN_NUM_TX_TYPES (TX_XDP + 1)
+	RX,
+};
+
+
+/*
+ * Useful macros
+ */
+#define ROUNDUP_LOG2(x)		ilog2(roundup_pow_of_two(x))
+#define XNOR(x, y)		(!(x) == !(y))
+
+
+struct mlx4_en_tx_info {
+	union {
+		struct sk_buff *skb;
+		struct page *page;
+	};
+	dma_addr_t	map0_dma;
+	u32		map0_byte_count;
+	u32		nr_txbb;
+	u32		nr_bytes;
+	u8		linear;
+	u8		data_offset;
+	u8		inl;
+	u8		ts_requested;
+	u8		nr_maps;
+} ____cacheline_aligned_in_smp;
+
+
+#define MLX4_EN_BIT_DESC_OWN	0x80000000
+#define CTRL_SIZE	sizeof(struct mlx4_wqe_ctrl_seg)
+#define MLX4_EN_MEMTYPE_PAD	0x100
+#define DS_SIZE		sizeof(struct mlx4_wqe_data_seg)
+
+
+struct mlx4_en_tx_desc {
+	struct mlx4_wqe_ctrl_seg ctrl;
+	union {
+		struct mlx4_wqe_data_seg data; /* at least one data segment */
+		struct mlx4_wqe_lso_seg lso;
+		struct mlx4_wqe_inline_seg inl;
+	};
+};
+
+#define MLX4_EN_USE_SRQ		0x01000000
+
+#define MLX4_EN_CX3_LOW_ID	0x1000
+#define MLX4_EN_CX3_HIGH_ID	0x1005
+
+struct mlx4_en_rx_alloc {
+	struct page	*page;
+	dma_addr_t	dma;
+	u32		page_offset;
+};
+
+#define MLX4_EN_CACHE_SIZE (2 * NAPI_POLL_WEIGHT)
+
+struct mlx4_en_page_cache {
+	u32 index;
+	struct {
+		struct page	*page;
+		dma_addr_t	dma;
+	} buf[MLX4_EN_CACHE_SIZE];
+};
+
+enum {
+	MLX4_EN_TX_RING_STATE_RECOVERING,
+};
+
+struct mlx4_en_priv;
+
+struct mlx4_en_tx_ring {
+	/* cache line used and dirtied in tx completion
+	 * (mlx4_en_free_tx_buf())
+	 */
+	u32			last_nr_txbb;
+	u32			cons;
+	unsigned long		wake_queue;
+	struct netdev_queue	*tx_queue;
+	u32			(*free_tx_desc)(struct mlx4_en_priv *priv,
+						struct mlx4_en_tx_ring *ring,
+						int index,
+						u64 timestamp, int napi_mode);
+	struct mlx4_en_rx_ring	*recycle_ring;
+
+	/* cache line used and dirtied in mlx4_en_xmit() */
+	u32			prod ____cacheline_aligned_in_smp;
+	unsigned int		tx_dropped;
+	unsigned long		bytes;
+	unsigned long		packets;
+	unsigned long		tx_csum;
+	unsigned long		tso_packets;
+	unsigned long		xmit_more;
+	struct mlx4_bf		bf;
+
+	/* Following part should be mostly read */
+	__be32			doorbell_qpn;
+	__be32			mr_key;
+	u32			size; /* number of TXBBs */
+	u32			size_mask;
+	u32			full_size;
+	u32			buf_size;
+	void			*buf;
+	struct mlx4_en_tx_info	*tx_info;
+	int			qpn;
+	u8			queue_index;
+	bool			bf_enabled;
+	bool			bf_alloced;
+	u8			hwtstamp_tx_type;
+	u8			*bounce_buf;
+
+	/* Not used in fast path
+	 * Only queue_stopped might be used if BQL is not properly working.
+	 */
+	unsigned long		queue_stopped;
+	unsigned long		state;
+	struct mlx4_hwq_resources sp_wqres;
+	struct mlx4_qp		sp_qp;
+	struct mlx4_qp_context	sp_context;
+	cpumask_t		sp_affinity_mask;
+	enum mlx4_qp_state	sp_qp_state;
+	u16			sp_stride;
+	u16			sp_cqn;	/* index of port CQ associated with this ring */
+} ____cacheline_aligned_in_smp;
+
+struct mlx4_en_rx_desc {
+	/* actual number of entries depends on rx ring stride */
+	struct mlx4_wqe_data_seg data[0];
+};
+
+struct mlx4_en_rx_ring {
+	struct mlx4_hwq_resources wqres;
+	u32 size ;	/* number of Rx descs*/
+	u32 actual_size;
+	u32 size_mask;
+	u16 stride;
+	u16 log_stride;
+	u16 cqn;	/* index of port CQ associated with this ring */
+	u32 prod;
+	u32 cons;
+	u32 buf_size;
+	u8  fcs_del;
+	void *buf;
+	void *rx_info;
+	struct bpf_prog __rcu *xdp_prog;
+	struct mlx4_en_page_cache page_cache;
+	unsigned long bytes;
+	unsigned long packets;
+	unsigned long csum_ok;
+	unsigned long csum_none;
+	unsigned long csum_complete;
+	unsigned long rx_alloc_pages;
+	unsigned long xdp_drop;
+	unsigned long xdp_tx;
+	unsigned long xdp_tx_full;
+	unsigned long dropped;
+	int hwtstamp_rx_filter;
+	cpumask_var_t affinity_mask;
+	struct xdp_rxq_info xdp_rxq;
+};
+
+struct mlx4_en_cq {
+	struct mlx4_cq          mcq;
+	struct mlx4_hwq_resources wqres;
+	int                     ring;
+	struct net_device      *dev;
+	union {
+		struct napi_struct napi;
+		bool               xdp_busy;
+	};
+	int size;
+	int buf_size;
+	int vector;
+	enum cq_type type;
+	u16 moder_time;
+	u16 moder_cnt;
+	struct mlx4_cqe *buf;
+#define MLX4_EN_OPCODE_ERROR	0x1e
+
+	struct irq_desc *irq_desc;
+};
+
+struct mlx4_en_port_profile {
+	u32 flags;
+	u32 tx_ring_num[MLX4_EN_NUM_TX_TYPES];
+	u32 rx_ring_num;
+	u32 tx_ring_size;
+	u32 rx_ring_size;
+	u8 num_tx_rings_p_up;
+	u8 rx_pause;
+	u8 rx_ppp;
+	u8 tx_pause;
+	u8 tx_ppp;
+	u8 num_up;
+	int rss_rings;
+	int inline_thold;
+	struct hwtstamp_config hwtstamp_config;
+};
+
+struct mlx4_en_profile {
+	int udp_rss;
+	u8 rss_mask;
+	u32 active_ports;
+	u32 small_pkt_int;
+	u8 no_reset;
+	u8 max_num_tx_rings_p_up;
+	struct mlx4_en_port_profile prof[MLX4_MAX_PORTS + 1];
+};
+
+struct mlx4_en_dev {
+	struct mlx4_dev         *dev;
+	struct pci_dev		*pdev;
+	struct mutex		state_lock;
+	struct net_device       *pndev[MLX4_MAX_PORTS + 1];
+	struct net_device       *upper[MLX4_MAX_PORTS + 1];
+	u32                     port_cnt;
+	bool			device_up;
+	struct mlx4_en_profile  profile;
+	u32			LSO_support;
+	struct workqueue_struct *workqueue;
+	struct device           *dma_device;
+	void __iomem            *uar_map;
+	struct mlx4_uar         priv_uar;
+	struct mlx4_mr		mr;
+	u32                     priv_pdn;
+	spinlock_t              uar_lock;
+	u8			mac_removed[MLX4_MAX_PORTS + 1];
+	u32			nominal_c_mult;
+	struct cyclecounter	cycles;
+	seqlock_t		clock_lock;
+	struct timecounter	clock;
+	unsigned long		last_overflow_check;
+	struct ptp_clock	*ptp_clock;
+	struct ptp_clock_info	ptp_clock_info;
+	struct notifier_block	nb;
+};
+
+
+struct mlx4_en_rss_map {
+	int base_qpn;
+	struct mlx4_qp qps[MAX_RX_RINGS];
+	enum mlx4_qp_state state[MAX_RX_RINGS];
+	struct mlx4_qp *indir_qp;
+	enum mlx4_qp_state indir_state;
+};
+
+enum mlx4_en_port_flag {
+	MLX4_EN_PORT_ANC = 1<<0, /* Auto-negotiation complete */
+	MLX4_EN_PORT_ANE = 1<<1, /* Auto-negotiation enabled */
+};
+
+struct mlx4_en_port_state {
+	int link_state;
+	int link_speed;
+	int transceiver;
+	u32 flags;
+};
+
+enum mlx4_en_mclist_act {
+	MCLIST_NONE,
+	MCLIST_REM,
+	MCLIST_ADD,
+};
+
+struct mlx4_en_mc_list {
+	struct list_head	list;
+	enum mlx4_en_mclist_act	action;
+	u8			addr[ETH_ALEN];
+	u64			reg_id;
+	u64			tunnel_reg_id;
+};
+
+struct mlx4_en_frag_info {
+	u16 frag_size;
+	u32 frag_stride;
+};
+
+#ifdef CONFIG_MLX4_EN_DCB
+/* Minimal TC BW - setting to 0 will block traffic */
+#define MLX4_EN_BW_MIN 1
+#define MLX4_EN_BW_MAX 100 /* Utilize 100% of the line */
+
+#define MLX4_EN_TC_VENDOR 0
+#define MLX4_EN_TC_ETS 7
+
+enum dcb_pfc_type {
+	pfc_disabled = 0,
+	pfc_enabled_full,
+	pfc_enabled_tx,
+	pfc_enabled_rx
+};
+
+struct mlx4_en_cee_config {
+	bool	pfc_state;
+	enum	dcb_pfc_type dcb_pfc[MLX4_EN_NUM_UP_HIGH];
+};
+#endif
+
+struct ethtool_flow_id {
+	struct list_head list;
+	struct ethtool_rx_flow_spec flow_spec;
+	u64 id;
+};
+
+enum {
+	MLX4_EN_FLAG_PROMISC		= (1 << 0),
+	MLX4_EN_FLAG_MC_PROMISC		= (1 << 1),
+	/* whether we need to enable hardware loopback by putting dmac
+	 * in Tx WQE
+	 */
+	MLX4_EN_FLAG_ENABLE_HW_LOOPBACK	= (1 << 2),
+	/* whether we need to drop packets that hardware loopback-ed */
+	MLX4_EN_FLAG_RX_FILTER_NEEDED	= (1 << 3),
+	MLX4_EN_FLAG_FORCE_PROMISC	= (1 << 4),
+	MLX4_EN_FLAG_RX_CSUM_NON_TCP_UDP	= (1 << 5),
+#ifdef CONFIG_MLX4_EN_DCB
+	MLX4_EN_FLAG_DCB_ENABLED        = (1 << 6),
+#endif
+};
+
+#define PORT_BEACON_MAX_LIMIT (65535)
+#define MLX4_EN_MAC_HASH_SIZE (1 << BITS_PER_BYTE)
+#define MLX4_EN_MAC_HASH_IDX 5
+
+struct mlx4_en_stats_bitmap {
+	DECLARE_BITMAP(bitmap, NUM_ALL_STATS);
+	struct mutex mutex; /* for mutual access to stats bitmap */
+};
+
+enum {
+	MLX4_EN_STATE_FLAG_RESTARTING,
+};
+
+struct mlx4_en_priv {
+	struct mlx4_en_dev *mdev;
+	struct mlx4_en_port_profile *prof;
+	struct net_device *dev;
+	unsigned long active_vlans[BITS_TO_LONGS(VLAN_N_VID)];
+	struct mlx4_en_port_state port_state;
+	spinlock_t stats_lock;
+	struct ethtool_flow_id ethtool_rules[MAX_NUM_OF_FS_RULES];
+	/* To allow rules removal while port is going down */
+	struct list_head ethtool_list;
+
+	unsigned long last_moder_packets[MAX_RX_RINGS];
+	unsigned long last_moder_tx_packets;
+	unsigned long last_moder_bytes[MAX_RX_RINGS];
+	unsigned long last_moder_jiffies;
+	int last_moder_time[MAX_RX_RINGS];
+	u16 rx_usecs;
+	u16 rx_frames;
+	u16 tx_usecs;
+	u16 tx_frames;
+	u32 pkt_rate_low;
+	u16 rx_usecs_low;
+	u32 pkt_rate_high;
+	u16 rx_usecs_high;
+	u32 sample_interval;
+	u32 adaptive_rx_coal;
+	u32 msg_enable;
+	u32 loopback_ok;
+	u32 validate_loopback;
+
+	struct mlx4_hwq_resources res;
+	int link_state;
+	int last_link_state;
+	bool port_up;
+	int port;
+	int registered;
+	int allocated;
+	int stride;
+	unsigned char current_mac[ETH_ALEN + 2];
+	int mac_index;
+	unsigned max_mtu;
+	int base_qpn;
+	int cqe_factor;
+	int cqe_size;
+
+	struct mlx4_en_rss_map rss_map;
+	__be32 ctrl_flags;
+	u32 flags;
+	u8 num_tx_rings_p_up;
+	u32 tx_work_limit;
+	u32 tx_ring_num[MLX4_EN_NUM_TX_TYPES];
+	u32 rx_ring_num;
+	u32 rx_skb_size;
+	struct mlx4_en_frag_info frag_info[MLX4_EN_MAX_RX_FRAGS];
+	u8 num_frags;
+	u8 log_rx_info;
+	u8 dma_dir;
+	u16 rx_headroom;
+
+	struct mlx4_en_tx_ring **tx_ring[MLX4_EN_NUM_TX_TYPES];
+	struct mlx4_en_rx_ring *rx_ring[MAX_RX_RINGS];
+	struct mlx4_en_cq **tx_cq[MLX4_EN_NUM_TX_TYPES];
+	struct mlx4_en_cq *rx_cq[MAX_RX_RINGS];
+	struct mlx4_qp drop_qp;
+	struct work_struct rx_mode_task;
+	struct work_struct restart_task;
+	struct work_struct linkstate_task;
+	struct delayed_work stats_task;
+	struct delayed_work service_task;
+	struct work_struct vxlan_add_task;
+	struct work_struct vxlan_del_task;
+	struct mlx4_en_perf_stats pstats;
+	struct mlx4_en_pkt_stats pkstats;
+	struct mlx4_en_counter_stats pf_stats;
+	struct mlx4_en_flow_stats_rx rx_priority_flowstats[MLX4_NUM_PRIORITIES];
+	struct mlx4_en_flow_stats_tx tx_priority_flowstats[MLX4_NUM_PRIORITIES];
+	struct mlx4_en_flow_stats_rx rx_flowstats;
+	struct mlx4_en_flow_stats_tx tx_flowstats;
+	struct mlx4_en_port_stats port_stats;
+	struct mlx4_en_xdp_stats xdp_stats;
+	struct mlx4_en_phy_stats phy_stats;
+	struct mlx4_en_stats_bitmap stats_bitmap;
+	struct list_head mc_list;
+	struct list_head curr_list;
+	u64 broadcast_id;
+	struct mlx4_en_stat_out_mbox hw_stats;
+	int vids[128];
+	bool wol;
+	struct device *ddev;
+	struct hlist_head mac_hash[MLX4_EN_MAC_HASH_SIZE];
+	struct hwtstamp_config hwtstamp_config;
+	u32 counter_index;
+
+#ifdef CONFIG_MLX4_EN_DCB
+#define MLX4_EN_DCB_ENABLED	0x3
+	struct ieee_ets ets;
+	u16 maxrate[IEEE_8021QAZ_MAX_TCS];
+	enum dcbnl_cndd_states cndd_state[IEEE_8021QAZ_MAX_TCS];
+	struct mlx4_en_cee_config cee_config;
+	u8 dcbx_cap;
+#endif
+#ifdef CONFIG_RFS_ACCEL
+	spinlock_t filters_lock;
+	int last_filter_id;
+	struct list_head filters;
+	struct hlist_head filter_hash[1 << MLX4_EN_FILTER_HASH_SHIFT];
+#endif
+	u64 tunnel_reg_id;
+	__be16 vxlan_port;
+
+	u32 pflags;
+	u8 rss_key[MLX4_EN_RSS_KEY_SIZE];
+	u8 rss_hash_fn;
+	unsigned long state;
+};
+
+enum mlx4_en_wol {
+	MLX4_EN_WOL_MAGIC = (1ULL << 61),
+	MLX4_EN_WOL_ENABLED = (1ULL << 62),
+};
+
+struct mlx4_mac_entry {
+	struct hlist_node hlist;
+	unsigned char mac[ETH_ALEN + 2];
+	u64 reg_id;
+	struct rcu_head rcu;
+};
+
+static inline struct mlx4_cqe *mlx4_en_get_cqe(void *buf, int idx, int cqe_sz)
+{
+	return buf + idx * cqe_sz;
+}
+
+#define MLX4_EN_WOL_DO_MODIFY (1ULL << 63)
+
+void mlx4_en_init_ptys2ethtool_map(void);
+void mlx4_en_update_loopback_state(struct net_device *dev,
+				   netdev_features_t features);
+
+void mlx4_en_destroy_netdev(struct net_device *dev);
+int mlx4_en_init_netdev(struct mlx4_en_dev *mdev, int port,
+			struct mlx4_en_port_profile *prof);
+
+int mlx4_en_start_port(struct net_device *dev);
+void mlx4_en_stop_port(struct net_device *dev, int detach);
+
+void mlx4_en_set_stats_bitmap(struct mlx4_dev *dev,
+			      struct mlx4_en_stats_bitmap *stats_bitmap,
+			      u8 rx_ppp, u8 rx_pause,
+			      u8 tx_ppp, u8 tx_pause);
+
+int mlx4_en_try_alloc_resources(struct mlx4_en_priv *priv,
+				struct mlx4_en_priv *tmp,
+				struct mlx4_en_port_profile *prof,
+				bool carry_xdp_prog);
+void mlx4_en_safe_replace_resources(struct mlx4_en_priv *priv,
+				    struct mlx4_en_priv *tmp);
+
+int mlx4_en_create_cq(struct mlx4_en_priv *priv, struct mlx4_en_cq **pcq,
+		      int entries, int ring, enum cq_type mode, int node);
+void mlx4_en_destroy_cq(struct mlx4_en_priv *priv, struct mlx4_en_cq **pcq);
+int mlx4_en_activate_cq(struct mlx4_en_priv *priv, struct mlx4_en_cq *cq,
+			int cq_idx);
+void mlx4_en_deactivate_cq(struct mlx4_en_priv *priv, struct mlx4_en_cq *cq);
+int mlx4_en_set_cq_moder(struct mlx4_en_priv *priv, struct mlx4_en_cq *cq);
+void mlx4_en_arm_cq(struct mlx4_en_priv *priv, struct mlx4_en_cq *cq);
+
+void mlx4_en_tx_irq(struct mlx4_cq *mcq);
+u16 mlx4_en_select_queue(struct net_device *dev, struct sk_buff *skb,
+			 struct net_device *sb_dev,
+			 select_queue_fallback_t fallback);
+netdev_tx_t mlx4_en_xmit(struct sk_buff *skb, struct net_device *dev);
+netdev_tx_t mlx4_en_xmit_frame(struct mlx4_en_rx_ring *rx_ring,
+			       struct mlx4_en_rx_alloc *frame,
+			       struct mlx4_en_priv *priv, unsigned int length,
+			       int tx_ind, bool *doorbell_pending);
+void mlx4_en_xmit_doorbell(struct mlx4_en_tx_ring *ring);
+bool mlx4_en_rx_recycle(struct mlx4_en_rx_ring *ring,
+			struct mlx4_en_rx_alloc *frame);
+
+int mlx4_en_create_tx_ring(struct mlx4_en_priv *priv,
+			   struct mlx4_en_tx_ring **pring,
+			   u32 size, u16 stride,
+			   int node, int queue_index);
+void mlx4_en_destroy_tx_ring(struct mlx4_en_priv *priv,
+			     struct mlx4_en_tx_ring **pring);
+void mlx4_en_init_tx_xdp_ring_descs(struct mlx4_en_priv *priv,
+				    struct mlx4_en_tx_ring *ring);
+int mlx4_en_activate_tx_ring(struct mlx4_en_priv *priv,
+			     struct mlx4_en_tx_ring *ring,
+			     int cq, int user_prio);
+void mlx4_en_deactivate_tx_ring(struct mlx4_en_priv *priv,
+				struct mlx4_en_tx_ring *ring);
+void mlx4_en_set_num_rx_rings(struct mlx4_en_dev *mdev);
+void mlx4_en_recover_from_oom(struct mlx4_en_priv *priv);
+int mlx4_en_create_rx_ring(struct mlx4_en_priv *priv,
+			   struct mlx4_en_rx_ring **pring,
+			   u32 size, u16 stride, int node, int queue_index);
+void mlx4_en_destroy_rx_ring(struct mlx4_en_priv *priv,
+			     struct mlx4_en_rx_ring **pring,
+			     u32 size, u16 stride);
+int mlx4_en_activate_rx_rings(struct mlx4_en_priv *priv);
+void mlx4_en_deactivate_rx_ring(struct mlx4_en_priv *priv,
+				struct mlx4_en_rx_ring *ring);
+int mlx4_en_process_rx_cq(struct net_device *dev,
+			  struct mlx4_en_cq *cq,
+			  int budget);
+int mlx4_en_poll_rx_cq(struct napi_struct *napi, int budget);
+int mlx4_en_poll_tx_cq(struct napi_struct *napi, int budget);
+bool mlx4_en_process_tx_cq(struct net_device *dev,
+			   struct mlx4_en_cq *cq, int napi_budget);
+u32 mlx4_en_free_tx_desc(struct mlx4_en_priv *priv,
+			 struct mlx4_en_tx_ring *ring,
+			 int index, u64 timestamp,
+			 int napi_mode);
+u32 mlx4_en_recycle_tx_desc(struct mlx4_en_priv *priv,
+			    struct mlx4_en_tx_ring *ring,
+			    int index, u64 timestamp,
+			    int napi_mode);
+void mlx4_en_fill_qp_context(struct mlx4_en_priv *priv, int size, int stride,
+		int is_tx, int rss, int qpn, int cqn, int user_prio,
+		struct mlx4_qp_context *context);
+void mlx4_en_sqp_event(struct mlx4_qp *qp, enum mlx4_event event);
+int mlx4_en_change_mcast_lb(struct mlx4_en_priv *priv, struct mlx4_qp *qp,
+			    int loopback);
+void mlx4_en_calc_rx_buf(struct net_device *dev);
+int mlx4_en_config_rss_steer(struct mlx4_en_priv *priv);
+void mlx4_en_release_rss_steer(struct mlx4_en_priv *priv);
+int mlx4_en_create_drop_qp(struct mlx4_en_priv *priv);
+void mlx4_en_destroy_drop_qp(struct mlx4_en_priv *priv);
+int mlx4_en_free_tx_buf(struct net_device *dev, struct mlx4_en_tx_ring *ring);
+void mlx4_en_rx_irq(struct mlx4_cq *mcq);
+
+int mlx4_SET_MCAST_FLTR(struct mlx4_dev *dev, u8 port, u64 mac, u64 clear, u8 mode);
+int mlx4_SET_VLAN_FLTR(struct mlx4_dev *dev, struct mlx4_en_priv *priv);
+
+void mlx4_en_fold_software_stats(struct net_device *dev);
+int mlx4_en_DUMP_ETH_STATS(struct mlx4_en_dev *mdev, u8 port, u8 reset);
+int mlx4_en_QUERY_PORT(struct mlx4_en_dev *mdev, u8 port);
+
+#ifdef CONFIG_MLX4_EN_DCB
+extern const struct dcbnl_rtnl_ops mlx4_en_dcbnl_ops;
+extern const struct dcbnl_rtnl_ops mlx4_en_dcbnl_pfc_ops;
+#endif
+
+int mlx4_en_setup_tc(struct net_device *dev, u8 up);
+int mlx4_en_alloc_tx_queue_per_tc(struct net_device *dev, u8 tc);
+
+#ifdef CONFIG_RFS_ACCEL
+void mlx4_en_cleanup_filters(struct mlx4_en_priv *priv);
+#endif
+
+#define MLX4_EN_NUM_SELF_TEST	5
+void mlx4_en_ex_selftest(struct net_device *dev, u32 *flags, u64 *buf);
+void mlx4_en_ptp_overflow_check(struct mlx4_en_dev *mdev);
+
+#define DEV_FEATURE_CHANGED(dev, new_features, feature) \
+	((dev->features & feature) ^ (new_features & feature))
+
+int mlx4_en_moderation_update(struct mlx4_en_priv *priv);
+int mlx4_en_reset_config(struct net_device *dev,
+			 struct hwtstamp_config ts_config,
+			 netdev_features_t new_features);
+void mlx4_en_update_pfc_stats_bitmap(struct mlx4_dev *dev,
+				     struct mlx4_en_stats_bitmap *stats_bitmap,
+				     u8 rx_ppp, u8 rx_pause,
+				     u8 tx_ppp, u8 tx_pause);
+int mlx4_en_netdev_event(struct notifier_block *this,
+			 unsigned long event, void *ptr);
+
+/*
+ * Functions for time stamping
+ */
+u64 mlx4_en_get_cqe_ts(struct mlx4_cqe *cqe);
+void mlx4_en_fill_hwtstamps(struct mlx4_en_dev *mdev,
+			    struct skb_shared_hwtstamps *hwts,
+			    u64 timestamp);
+void mlx4_en_init_timestamp(struct mlx4_en_dev *mdev);
+void mlx4_en_remove_timestamp(struct mlx4_en_dev *mdev);
+
+/* Globals
+ */
+extern const struct ethtool_ops mlx4_en_ethtool_ops;
+
+
+
+/*
+ * printk / logging functions
+ */
+
+__printf(3, 4)
+void en_print(const char *level, const struct mlx4_en_priv *priv,
+	      const char *format, ...);
+
+#define en_dbg(mlevel, priv, format, ...)				\
+do {									\
+	if (NETIF_MSG_##mlevel & (priv)->msg_enable)			\
+		en_print(KERN_DEBUG, priv, format, ##__VA_ARGS__);	\
+} while (0)
+#define en_warn(priv, format, ...)					\
+	en_print(KERN_WARNING, priv, format, ##__VA_ARGS__)
+#define en_err(priv, format, ...)					\
+	en_print(KERN_ERR, priv, format, ##__VA_ARGS__)
+#define en_info(priv, format, ...)					\
+	en_print(KERN_INFO, priv, format, ##__VA_ARGS__)
+
+#define mlx4_err(mdev, format, ...)					\
+	pr_err(DRV_NAME " %s: " format,					\
+	       dev_name(&(mdev)->pdev->dev), ##__VA_ARGS__)
+#define mlx4_info(mdev, format, ...)					\
+	pr_info(DRV_NAME " %s: " format,				\
+		dev_name(&(mdev)->pdev->dev), ##__VA_ARGS__)
+#define mlx4_warn(mdev, format, ...)					\
+	pr_warn(DRV_NAME " %s: " format,				\
+		dev_name(&(mdev)->pdev->dev), ##__VA_ARGS__)
+
+#endif
diff --git a/drivers/net/ethernet/mellanox/mlx4/mlx4_stats.h b/drivers/net/ethernet/mellanox/mlx4/mlx4_stats.h
new file mode 100644
index 000000000..86b6051da
--- /dev/null
+++ b/drivers/net/ethernet/mellanox/mlx4/mlx4_stats.h
@@ -0,0 +1,132 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+#ifndef _MLX4_STATS_
+#define _MLX4_STATS_
+
+#ifdef MLX4_EN_PERF_STAT
+#define NUM_PERF_STATS			NUM_PERF_COUNTERS
+#else
+#define NUM_PERF_STATS			0
+#endif
+
+#define NUM_PRIORITIES	9
+#define NUM_PRIORITY_STATS 2
+
+struct mlx4_en_pkt_stats {
+	unsigned long rx_multicast_packets;
+	unsigned long rx_broadcast_packets;
+	unsigned long rx_jabbers;
+	unsigned long rx_in_range_length_error;
+	unsigned long rx_out_range_length_error;
+	unsigned long tx_multicast_packets;
+	unsigned long tx_broadcast_packets;
+	unsigned long rx_prio[NUM_PRIORITIES][NUM_PRIORITY_STATS];
+	unsigned long tx_prio[NUM_PRIORITIES][NUM_PRIORITY_STATS];
+#define NUM_PKT_STATS		43
+};
+
+struct mlx4_en_counter_stats {
+	unsigned long rx_packets;
+	unsigned long rx_bytes;
+	unsigned long tx_packets;
+	unsigned long tx_bytes;
+#define NUM_PF_STATS      4
+};
+
+struct mlx4_en_port_stats {
+	unsigned long tso_packets;
+	unsigned long xmit_more;
+	unsigned long queue_stopped;
+	unsigned long wake_queue;
+	unsigned long tx_timeout;
+	unsigned long rx_alloc_pages;
+	unsigned long rx_chksum_good;
+	unsigned long rx_chksum_none;
+	unsigned long rx_chksum_complete;
+	unsigned long tx_chksum_offload;
+#define NUM_PORT_STATS		10
+};
+
+struct mlx4_en_perf_stats {
+	u32 tx_poll;
+	u64 tx_pktsz_avg;
+	u32 inflight_avg;
+	u16 tx_coal_avg;
+	u16 rx_coal_avg;
+	u32 napi_quota;
+#define NUM_PERF_COUNTERS		6
+};
+
+struct mlx4_en_xdp_stats {
+	unsigned long rx_xdp_drop;
+	unsigned long rx_xdp_tx;
+	unsigned long rx_xdp_tx_full;
+#define NUM_XDP_STATS		3
+};
+
+struct mlx4_en_phy_stats {
+	unsigned long rx_packets_phy;
+	unsigned long rx_bytes_phy;
+	unsigned long tx_packets_phy;
+	unsigned long tx_bytes_phy;
+#define NUM_PHY_STATS		4
+};
+
+#define NUM_MAIN_STATS	21
+
+#define MLX4_NUM_PRIORITIES	8
+
+struct mlx4_en_flow_stats_rx {
+	u64 rx_pause;
+	u64 rx_pause_duration;
+	u64 rx_pause_transition;
+#define NUM_FLOW_STATS_RX	3
+#define NUM_FLOW_PRIORITY_STATS_RX	(NUM_FLOW_STATS_RX * \
+					 MLX4_NUM_PRIORITIES)
+};
+
+struct mlx4_en_flow_stats_tx {
+	u64 tx_pause;
+	u64 tx_pause_duration;
+	u64 tx_pause_transition;
+#define NUM_FLOW_STATS_TX	3
+#define NUM_FLOW_PRIORITY_STATS_TX	(NUM_FLOW_STATS_TX * \
+					 MLX4_NUM_PRIORITIES)
+};
+
+#define NUM_FLOW_STATS (NUM_FLOW_STATS_RX + NUM_FLOW_STATS_TX + \
+			NUM_FLOW_PRIORITY_STATS_TX + \
+			NUM_FLOW_PRIORITY_STATS_RX)
+
+struct mlx4_en_stat_out_flow_control_mbox {
+	/* Total number of PAUSE frames received from the far-end port */
+	__be64 rx_pause;
+	/* Total number of microseconds that far-end port requested to pause
+	* transmission of packets
+	*/
+	__be64 rx_pause_duration;
+	/* Number of received transmission from XOFF state to XON state */
+	__be64 rx_pause_transition;
+	/* Total number of PAUSE frames sent from the far-end port */
+	__be64 tx_pause;
+	/* Total time in microseconds that transmission of packets has been
+	* paused
+	*/
+	__be64 tx_pause_duration;
+	/* Number of transmitter transitions from XOFF state to XON state */
+	__be64 tx_pause_transition;
+	/* Reserverd */
+	__be64 reserved[2];
+};
+
+enum {
+	MLX4_DUMP_ETH_STATS_FLOW_CONTROL = 1 << 12
+};
+
+#define NUM_ALL_STATS	(NUM_MAIN_STATS + NUM_PORT_STATS + NUM_PKT_STATS + \
+			 NUM_FLOW_STATS + NUM_PERF_STATS + NUM_PF_STATS + \
+			 NUM_XDP_STATS + NUM_PHY_STATS)
+
+#define MLX4_FIND_NETDEV_STAT(n) (offsetof(struct net_device_stats, n) / \
+				  sizeof(((struct net_device_stats *)0)->n))
+
+#endif
diff --git a/drivers/net/ethernet/mellanox/mlx4/mr.c b/drivers/net/ethernet/mellanox/mlx4/mr.c
new file mode 100644
index 000000000..cfa0bba39
--- /dev/null
+++ b/drivers/net/ethernet/mellanox/mlx4/mr.c
@@ -0,0 +1,1157 @@
+/*
+ * Copyright (c) 2004 Topspin Communications.  All rights reserved.
+ * Copyright (c) 2005, 2006, 2007, 2008 Mellanox Technologies. All rights reserved.
+ * Copyright (c) 2006, 2007 Cisco Systems, Inc.  All rights reserved.
+ *
+ * This software is available to you under a choice of one of two
+ * licenses.  You may choose to be licensed under the terms of the GNU
+ * General Public License (GPL) Version 2, available from the file
+ * COPYING in the main directory of this source tree, or the
+ * OpenIB.org BSD license below:
+ *
+ *     Redistribution and use in source and binary forms, with or
+ *     without modification, are permitted provided that the following
+ *     conditions are met:
+ *
+ *      - Redistributions of source code must retain the above
+ *        copyright notice, this list of conditions and the following
+ *        disclaimer.
+ *
+ *      - Redistributions in binary form must reproduce the above
+ *        copyright notice, this list of conditions and the following
+ *        disclaimer in the documentation and/or other materials
+ *        provided with the distribution.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
+ * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
+ * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS
+ * BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN
+ * ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
+ * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+ * SOFTWARE.
+ */
+
+#include <linux/errno.h>
+#include <linux/export.h>
+#include <linux/slab.h>
+#include <linux/kernel.h>
+#include <linux/vmalloc.h>
+
+#include <linux/mlx4/cmd.h>
+
+#include "mlx4.h"
+#include "icm.h"
+
+static u32 mlx4_buddy_alloc(struct mlx4_buddy *buddy, int order)
+{
+	int o;
+	int m;
+	u32 seg;
+
+	spin_lock(&buddy->lock);
+
+	for (o = order; o <= buddy->max_order; ++o)
+		if (buddy->num_free[o]) {
+			m = 1 << (buddy->max_order - o);
+			seg = find_first_bit(buddy->bits[o], m);
+			if (seg < m)
+				goto found;
+		}
+
+	spin_unlock(&buddy->lock);
+	return -1;
+
+ found:
+	clear_bit(seg, buddy->bits[o]);
+	--buddy->num_free[o];
+
+	while (o > order) {
+		--o;
+		seg <<= 1;
+		set_bit(seg ^ 1, buddy->bits[o]);
+		++buddy->num_free[o];
+	}
+
+	spin_unlock(&buddy->lock);
+
+	seg <<= order;
+
+	return seg;
+}
+
+static void mlx4_buddy_free(struct mlx4_buddy *buddy, u32 seg, int order)
+{
+	seg >>= order;
+
+	spin_lock(&buddy->lock);
+
+	while (test_bit(seg ^ 1, buddy->bits[order])) {
+		clear_bit(seg ^ 1, buddy->bits[order]);
+		--buddy->num_free[order];
+		seg >>= 1;
+		++order;
+	}
+
+	set_bit(seg, buddy->bits[order]);
+	++buddy->num_free[order];
+
+	spin_unlock(&buddy->lock);
+}
+
+static int mlx4_buddy_init(struct mlx4_buddy *buddy, int max_order)
+{
+	int i, s;
+
+	buddy->max_order = max_order;
+	spin_lock_init(&buddy->lock);
+
+	buddy->bits = kcalloc(buddy->max_order + 1, sizeof(long *),
+			      GFP_KERNEL);
+	buddy->num_free = kcalloc(buddy->max_order + 1, sizeof(*buddy->num_free),
+				  GFP_KERNEL);
+	if (!buddy->bits || !buddy->num_free)
+		goto err_out;
+
+	for (i = 0; i <= buddy->max_order; ++i) {
+		s = BITS_TO_LONGS(1UL << (buddy->max_order - i));
+		buddy->bits[i] = kvmalloc_array(s, sizeof(long), GFP_KERNEL | __GFP_ZERO);
+		if (!buddy->bits[i])
+			goto err_out_free;
+	}
+
+	set_bit(0, buddy->bits[buddy->max_order]);
+	buddy->num_free[buddy->max_order] = 1;
+
+	return 0;
+
+err_out_free:
+	for (i = 0; i <= buddy->max_order; ++i)
+		kvfree(buddy->bits[i]);
+
+err_out:
+	kfree(buddy->bits);
+	kfree(buddy->num_free);
+
+	return -ENOMEM;
+}
+
+static void mlx4_buddy_cleanup(struct mlx4_buddy *buddy)
+{
+	int i;
+
+	for (i = 0; i <= buddy->max_order; ++i)
+		kvfree(buddy->bits[i]);
+
+	kfree(buddy->bits);
+	kfree(buddy->num_free);
+}
+
+u32 __mlx4_alloc_mtt_range(struct mlx4_dev *dev, int order)
+{
+	struct mlx4_mr_table *mr_table = &mlx4_priv(dev)->mr_table;
+	u32 seg;
+	int seg_order;
+	u32 offset;
+
+	seg_order = max_t(int, order - log_mtts_per_seg, 0);
+
+	seg = mlx4_buddy_alloc(&mr_table->mtt_buddy, seg_order);
+	if (seg == -1)
+		return -1;
+
+	offset = seg * (1 << log_mtts_per_seg);
+
+	if (mlx4_table_get_range(dev, &mr_table->mtt_table, offset,
+				 offset + (1 << order) - 1)) {
+		mlx4_buddy_free(&mr_table->mtt_buddy, seg, seg_order);
+		return -1;
+	}
+
+	return offset;
+}
+
+static u32 mlx4_alloc_mtt_range(struct mlx4_dev *dev, int order)
+{
+	u64 in_param = 0;
+	u64 out_param;
+	int err;
+
+	if (mlx4_is_mfunc(dev)) {
+		set_param_l(&in_param, order);
+		err = mlx4_cmd_imm(dev, in_param, &out_param, RES_MTT,
+						       RES_OP_RESERVE_AND_MAP,
+						       MLX4_CMD_ALLOC_RES,
+						       MLX4_CMD_TIME_CLASS_A,
+						       MLX4_CMD_WRAPPED);
+		if (err)
+			return -1;
+		return get_param_l(&out_param);
+	}
+	return __mlx4_alloc_mtt_range(dev, order);
+}
+
+int mlx4_mtt_init(struct mlx4_dev *dev, int npages, int page_shift,
+		  struct mlx4_mtt *mtt)
+{
+	int i;
+
+	if (!npages) {
+		mtt->order      = -1;
+		mtt->page_shift = MLX4_ICM_PAGE_SHIFT;
+		return 0;
+	} else
+		mtt->page_shift = page_shift;
+
+	for (mtt->order = 0, i = 1; i < npages; i <<= 1)
+		++mtt->order;
+
+	mtt->offset = mlx4_alloc_mtt_range(dev, mtt->order);
+	if (mtt->offset == -1)
+		return -ENOMEM;
+
+	return 0;
+}
+EXPORT_SYMBOL_GPL(mlx4_mtt_init);
+
+void __mlx4_free_mtt_range(struct mlx4_dev *dev, u32 offset, int order)
+{
+	u32 first_seg;
+	int seg_order;
+	struct mlx4_mr_table *mr_table = &mlx4_priv(dev)->mr_table;
+
+	seg_order = max_t(int, order - log_mtts_per_seg, 0);
+	first_seg = offset / (1 << log_mtts_per_seg);
+
+	mlx4_buddy_free(&mr_table->mtt_buddy, first_seg, seg_order);
+	mlx4_table_put_range(dev, &mr_table->mtt_table, offset,
+			     offset + (1 << order) - 1);
+}
+
+static void mlx4_free_mtt_range(struct mlx4_dev *dev, u32 offset, int order)
+{
+	u64 in_param = 0;
+	int err;
+
+	if (mlx4_is_mfunc(dev)) {
+		set_param_l(&in_param, offset);
+		set_param_h(&in_param, order);
+		err = mlx4_cmd(dev, in_param, RES_MTT, RES_OP_RESERVE_AND_MAP,
+						       MLX4_CMD_FREE_RES,
+						       MLX4_CMD_TIME_CLASS_A,
+						       MLX4_CMD_WRAPPED);
+		if (err)
+			mlx4_warn(dev, "Failed to free mtt range at:%d order:%d\n",
+				  offset, order);
+		return;
+	}
+	__mlx4_free_mtt_range(dev, offset, order);
+}
+
+void mlx4_mtt_cleanup(struct mlx4_dev *dev, struct mlx4_mtt *mtt)
+{
+	if (mtt->order < 0)
+		return;
+
+	mlx4_free_mtt_range(dev, mtt->offset, mtt->order);
+}
+EXPORT_SYMBOL_GPL(mlx4_mtt_cleanup);
+
+u64 mlx4_mtt_addr(struct mlx4_dev *dev, struct mlx4_mtt *mtt)
+{
+	return (u64) mtt->offset * dev->caps.mtt_entry_sz;
+}
+EXPORT_SYMBOL_GPL(mlx4_mtt_addr);
+
+static u32 hw_index_to_key(u32 ind)
+{
+	return (ind >> 24) | (ind << 8);
+}
+
+static u32 key_to_hw_index(u32 key)
+{
+	return (key << 24) | (key >> 8);
+}
+
+static int mlx4_SW2HW_MPT(struct mlx4_dev *dev, struct mlx4_cmd_mailbox *mailbox,
+			  int mpt_index)
+{
+	return mlx4_cmd(dev, mailbox->dma, mpt_index,
+			0, MLX4_CMD_SW2HW_MPT, MLX4_CMD_TIME_CLASS_B,
+			MLX4_CMD_WRAPPED);
+}
+
+static int mlx4_HW2SW_MPT(struct mlx4_dev *dev, struct mlx4_cmd_mailbox *mailbox,
+			  int mpt_index)
+{
+	return mlx4_cmd_box(dev, 0, mailbox ? mailbox->dma : 0, mpt_index,
+			    !mailbox, MLX4_CMD_HW2SW_MPT,
+			    MLX4_CMD_TIME_CLASS_B, MLX4_CMD_WRAPPED);
+}
+
+/* Must protect against concurrent access */
+int mlx4_mr_hw_get_mpt(struct mlx4_dev *dev, struct mlx4_mr *mmr,
+		       struct mlx4_mpt_entry ***mpt_entry)
+{
+	int err;
+	int key = key_to_hw_index(mmr->key) & (dev->caps.num_mpts - 1);
+	struct mlx4_cmd_mailbox *mailbox = NULL;
+
+	if (mmr->enabled != MLX4_MPT_EN_HW)
+		return -EINVAL;
+
+	err = mlx4_HW2SW_MPT(dev, NULL, key);
+	if (err) {
+		mlx4_warn(dev, "HW2SW_MPT failed (%d).", err);
+		mlx4_warn(dev, "Most likely the MR has MWs bound to it.\n");
+		return err;
+	}
+
+	mmr->enabled = MLX4_MPT_EN_SW;
+
+	if (!mlx4_is_mfunc(dev)) {
+		**mpt_entry = mlx4_table_find(
+				&mlx4_priv(dev)->mr_table.dmpt_table,
+				key, NULL);
+	} else {
+		mailbox = mlx4_alloc_cmd_mailbox(dev);
+		if (IS_ERR(mailbox))
+			return PTR_ERR(mailbox);
+
+		err = mlx4_cmd_box(dev, 0, mailbox->dma, key,
+				   0, MLX4_CMD_QUERY_MPT,
+				   MLX4_CMD_TIME_CLASS_B,
+				   MLX4_CMD_WRAPPED);
+		if (err)
+			goto free_mailbox;
+
+		*mpt_entry = (struct mlx4_mpt_entry **)&mailbox->buf;
+	}
+
+	if (!(*mpt_entry) || !(**mpt_entry)) {
+		err = -ENOMEM;
+		goto free_mailbox;
+	}
+
+	return 0;
+
+free_mailbox:
+	mlx4_free_cmd_mailbox(dev, mailbox);
+	return err;
+}
+EXPORT_SYMBOL_GPL(mlx4_mr_hw_get_mpt);
+
+int mlx4_mr_hw_write_mpt(struct mlx4_dev *dev, struct mlx4_mr *mmr,
+			 struct mlx4_mpt_entry **mpt_entry)
+{
+	int err;
+
+	if (!mlx4_is_mfunc(dev)) {
+		/* Make sure any changes to this entry are flushed */
+		wmb();
+
+		*(u8 *)(*mpt_entry) = MLX4_MPT_STATUS_HW;
+
+		/* Make sure the new status is written */
+		wmb();
+
+		err = mlx4_SYNC_TPT(dev);
+	} else {
+		int key = key_to_hw_index(mmr->key) & (dev->caps.num_mpts - 1);
+
+		struct mlx4_cmd_mailbox *mailbox =
+			container_of((void *)mpt_entry, struct mlx4_cmd_mailbox,
+				     buf);
+
+		(*mpt_entry)->lkey = 0;
+		err = mlx4_SW2HW_MPT(dev, mailbox, key);
+	}
+
+	if (!err) {
+		mmr->pd = be32_to_cpu((*mpt_entry)->pd_flags) & MLX4_MPT_PD_MASK;
+		mmr->enabled = MLX4_MPT_EN_HW;
+	}
+	return err;
+}
+EXPORT_SYMBOL_GPL(mlx4_mr_hw_write_mpt);
+
+void mlx4_mr_hw_put_mpt(struct mlx4_dev *dev,
+			struct mlx4_mpt_entry **mpt_entry)
+{
+	if (mlx4_is_mfunc(dev)) {
+		struct mlx4_cmd_mailbox *mailbox =
+			container_of((void *)mpt_entry, struct mlx4_cmd_mailbox,
+				     buf);
+		mlx4_free_cmd_mailbox(dev, mailbox);
+	}
+}
+EXPORT_SYMBOL_GPL(mlx4_mr_hw_put_mpt);
+
+int mlx4_mr_hw_change_pd(struct mlx4_dev *dev, struct mlx4_mpt_entry *mpt_entry,
+			 u32 pdn)
+{
+	u32 pd_flags = be32_to_cpu(mpt_entry->pd_flags) & ~MLX4_MPT_PD_MASK;
+	/* The wrapper function will put the slave's id here */
+	if (mlx4_is_mfunc(dev))
+		pd_flags &= ~MLX4_MPT_PD_VF_MASK;
+
+	mpt_entry->pd_flags = cpu_to_be32(pd_flags |
+					  (pdn & MLX4_MPT_PD_MASK)
+					  | MLX4_MPT_PD_FLAG_EN_INV);
+	return 0;
+}
+EXPORT_SYMBOL_GPL(mlx4_mr_hw_change_pd);
+
+int mlx4_mr_hw_change_access(struct mlx4_dev *dev,
+			     struct mlx4_mpt_entry *mpt_entry,
+			     u32 access)
+{
+	u32 flags = (be32_to_cpu(mpt_entry->flags) & ~MLX4_PERM_MASK) |
+		    (access & MLX4_PERM_MASK);
+
+	mpt_entry->flags = cpu_to_be32(flags);
+	return 0;
+}
+EXPORT_SYMBOL_GPL(mlx4_mr_hw_change_access);
+
+static int mlx4_mr_alloc_reserved(struct mlx4_dev *dev, u32 mridx, u32 pd,
+			   u64 iova, u64 size, u32 access, int npages,
+			   int page_shift, struct mlx4_mr *mr)
+{
+	mr->iova       = iova;
+	mr->size       = size;
+	mr->pd	       = pd;
+	mr->access     = access;
+	mr->enabled    = MLX4_MPT_DISABLED;
+	mr->key	       = hw_index_to_key(mridx);
+
+	return mlx4_mtt_init(dev, npages, page_shift, &mr->mtt);
+}
+
+static int mlx4_WRITE_MTT(struct mlx4_dev *dev,
+			  struct mlx4_cmd_mailbox *mailbox,
+			  int num_entries)
+{
+	return mlx4_cmd(dev, mailbox->dma, num_entries, 0, MLX4_CMD_WRITE_MTT,
+			MLX4_CMD_TIME_CLASS_A,  MLX4_CMD_WRAPPED);
+}
+
+int __mlx4_mpt_reserve(struct mlx4_dev *dev)
+{
+	struct mlx4_priv *priv = mlx4_priv(dev);
+
+	return mlx4_bitmap_alloc(&priv->mr_table.mpt_bitmap);
+}
+
+static int mlx4_mpt_reserve(struct mlx4_dev *dev)
+{
+	u64 out_param;
+
+	if (mlx4_is_mfunc(dev)) {
+		if (mlx4_cmd_imm(dev, 0, &out_param, RES_MPT, RES_OP_RESERVE,
+				   MLX4_CMD_ALLOC_RES,
+				   MLX4_CMD_TIME_CLASS_A, MLX4_CMD_WRAPPED))
+			return -1;
+		return get_param_l(&out_param);
+	}
+	return  __mlx4_mpt_reserve(dev);
+}
+
+void __mlx4_mpt_release(struct mlx4_dev *dev, u32 index)
+{
+	struct mlx4_priv *priv = mlx4_priv(dev);
+
+	mlx4_bitmap_free(&priv->mr_table.mpt_bitmap, index, MLX4_NO_RR);
+}
+
+static void mlx4_mpt_release(struct mlx4_dev *dev, u32 index)
+{
+	u64 in_param = 0;
+
+	if (mlx4_is_mfunc(dev)) {
+		set_param_l(&in_param, index);
+		if (mlx4_cmd(dev, in_param, RES_MPT, RES_OP_RESERVE,
+			       MLX4_CMD_FREE_RES,
+			       MLX4_CMD_TIME_CLASS_A, MLX4_CMD_WRAPPED))
+			mlx4_warn(dev, "Failed to release mr index:%d\n",
+				  index);
+		return;
+	}
+	__mlx4_mpt_release(dev, index);
+}
+
+int __mlx4_mpt_alloc_icm(struct mlx4_dev *dev, u32 index)
+{
+	struct mlx4_mr_table *mr_table = &mlx4_priv(dev)->mr_table;
+
+	return mlx4_table_get(dev, &mr_table->dmpt_table, index);
+}
+
+static int mlx4_mpt_alloc_icm(struct mlx4_dev *dev, u32 index)
+{
+	u64 param = 0;
+
+	if (mlx4_is_mfunc(dev)) {
+		set_param_l(&param, index);
+		return mlx4_cmd_imm(dev, param, &param, RES_MPT, RES_OP_MAP_ICM,
+							MLX4_CMD_ALLOC_RES,
+							MLX4_CMD_TIME_CLASS_A,
+							MLX4_CMD_WRAPPED);
+	}
+	return __mlx4_mpt_alloc_icm(dev, index);
+}
+
+void __mlx4_mpt_free_icm(struct mlx4_dev *dev, u32 index)
+{
+	struct mlx4_mr_table *mr_table = &mlx4_priv(dev)->mr_table;
+
+	mlx4_table_put(dev, &mr_table->dmpt_table, index);
+}
+
+static void mlx4_mpt_free_icm(struct mlx4_dev *dev, u32 index)
+{
+	u64 in_param = 0;
+
+	if (mlx4_is_mfunc(dev)) {
+		set_param_l(&in_param, index);
+		if (mlx4_cmd(dev, in_param, RES_MPT, RES_OP_MAP_ICM,
+			     MLX4_CMD_FREE_RES, MLX4_CMD_TIME_CLASS_A,
+			     MLX4_CMD_WRAPPED))
+			mlx4_warn(dev, "Failed to free icm of mr index:%d\n",
+				  index);
+		return;
+	}
+	return __mlx4_mpt_free_icm(dev, index);
+}
+
+int mlx4_mr_alloc(struct mlx4_dev *dev, u32 pd, u64 iova, u64 size, u32 access,
+		  int npages, int page_shift, struct mlx4_mr *mr)
+{
+	u32 index;
+	int err;
+
+	index = mlx4_mpt_reserve(dev);
+	if (index == -1)
+		return -ENOMEM;
+
+	err = mlx4_mr_alloc_reserved(dev, index, pd, iova, size,
+				     access, npages, page_shift, mr);
+	if (err)
+		mlx4_mpt_release(dev, index);
+
+	return err;
+}
+EXPORT_SYMBOL_GPL(mlx4_mr_alloc);
+
+static int mlx4_mr_free_reserved(struct mlx4_dev *dev, struct mlx4_mr *mr)
+{
+	int err;
+
+	if (mr->enabled == MLX4_MPT_EN_HW) {
+		err = mlx4_HW2SW_MPT(dev, NULL,
+				     key_to_hw_index(mr->key) &
+				     (dev->caps.num_mpts - 1));
+		if (err) {
+			mlx4_warn(dev, "HW2SW_MPT failed (%d), MR has MWs bound to it\n",
+				  err);
+			return err;
+		}
+
+		mr->enabled = MLX4_MPT_EN_SW;
+	}
+	mlx4_mtt_cleanup(dev, &mr->mtt);
+
+	return 0;
+}
+
+int mlx4_mr_free(struct mlx4_dev *dev, struct mlx4_mr *mr)
+{
+	int ret;
+
+	ret = mlx4_mr_free_reserved(dev, mr);
+	if (ret)
+		return ret;
+	if (mr->enabled)
+		mlx4_mpt_free_icm(dev, key_to_hw_index(mr->key));
+	mlx4_mpt_release(dev, key_to_hw_index(mr->key));
+
+	return 0;
+}
+EXPORT_SYMBOL_GPL(mlx4_mr_free);
+
+void mlx4_mr_rereg_mem_cleanup(struct mlx4_dev *dev, struct mlx4_mr *mr)
+{
+	mlx4_mtt_cleanup(dev, &mr->mtt);
+	mr->mtt.order = -1;
+}
+EXPORT_SYMBOL_GPL(mlx4_mr_rereg_mem_cleanup);
+
+int mlx4_mr_rereg_mem_write(struct mlx4_dev *dev, struct mlx4_mr *mr,
+			    u64 iova, u64 size, int npages,
+			    int page_shift, struct mlx4_mpt_entry *mpt_entry)
+{
+	int err;
+
+	err = mlx4_mtt_init(dev, npages, page_shift, &mr->mtt);
+	if (err)
+		return err;
+
+	mpt_entry->start       = cpu_to_be64(iova);
+	mpt_entry->length      = cpu_to_be64(size);
+	mpt_entry->entity_size = cpu_to_be32(page_shift);
+	mpt_entry->flags    &= ~(cpu_to_be32(MLX4_MPT_FLAG_FREE |
+					   MLX4_MPT_FLAG_SW_OWNS));
+	if (mr->mtt.order < 0) {
+		mpt_entry->flags |= cpu_to_be32(MLX4_MPT_FLAG_PHYSICAL);
+		mpt_entry->mtt_addr = 0;
+	} else {
+		mpt_entry->mtt_addr = cpu_to_be64(mlx4_mtt_addr(dev,
+						  &mr->mtt));
+		if (mr->mtt.page_shift == 0)
+			mpt_entry->mtt_sz    = cpu_to_be32(1 << mr->mtt.order);
+	}
+	if (mr->mtt.order >= 0 && mr->mtt.page_shift == 0) {
+		/* fast register MR in free state */
+		mpt_entry->flags    |= cpu_to_be32(MLX4_MPT_FLAG_FREE);
+		mpt_entry->pd_flags |= cpu_to_be32(MLX4_MPT_PD_FLAG_FAST_REG |
+						   MLX4_MPT_PD_FLAG_RAE);
+	} else {
+		mpt_entry->flags    |= cpu_to_be32(MLX4_MPT_FLAG_SW_OWNS);
+	}
+	mr->enabled = MLX4_MPT_EN_SW;
+
+	return 0;
+}
+EXPORT_SYMBOL_GPL(mlx4_mr_rereg_mem_write);
+
+int mlx4_mr_enable(struct mlx4_dev *dev, struct mlx4_mr *mr)
+{
+	struct mlx4_cmd_mailbox *mailbox;
+	struct mlx4_mpt_entry *mpt_entry;
+	int err;
+
+	err = mlx4_mpt_alloc_icm(dev, key_to_hw_index(mr->key));
+	if (err)
+		return err;
+
+	mailbox = mlx4_alloc_cmd_mailbox(dev);
+	if (IS_ERR(mailbox)) {
+		err = PTR_ERR(mailbox);
+		goto err_table;
+	}
+	mpt_entry = mailbox->buf;
+	mpt_entry->flags = cpu_to_be32(MLX4_MPT_FLAG_MIO	 |
+				       MLX4_MPT_FLAG_REGION	 |
+				       mr->access);
+
+	mpt_entry->key	       = cpu_to_be32(key_to_hw_index(mr->key));
+	mpt_entry->pd_flags    = cpu_to_be32(mr->pd | MLX4_MPT_PD_FLAG_EN_INV);
+	mpt_entry->start       = cpu_to_be64(mr->iova);
+	mpt_entry->length      = cpu_to_be64(mr->size);
+	mpt_entry->entity_size = cpu_to_be32(mr->mtt.page_shift);
+
+	if (mr->mtt.order < 0) {
+		mpt_entry->flags |= cpu_to_be32(MLX4_MPT_FLAG_PHYSICAL);
+		mpt_entry->mtt_addr = 0;
+	} else {
+		mpt_entry->mtt_addr = cpu_to_be64(mlx4_mtt_addr(dev,
+						  &mr->mtt));
+	}
+
+	if (mr->mtt.order >= 0 && mr->mtt.page_shift == 0) {
+		/* fast register MR in free state */
+		mpt_entry->flags    |= cpu_to_be32(MLX4_MPT_FLAG_FREE);
+		mpt_entry->pd_flags |= cpu_to_be32(MLX4_MPT_PD_FLAG_FAST_REG |
+						   MLX4_MPT_PD_FLAG_RAE);
+		mpt_entry->mtt_sz    = cpu_to_be32(1 << mr->mtt.order);
+	} else {
+		mpt_entry->flags    |= cpu_to_be32(MLX4_MPT_FLAG_SW_OWNS);
+	}
+
+	err = mlx4_SW2HW_MPT(dev, mailbox,
+			     key_to_hw_index(mr->key) & (dev->caps.num_mpts - 1));
+	if (err) {
+		mlx4_warn(dev, "SW2HW_MPT failed (%d)\n", err);
+		goto err_cmd;
+	}
+	mr->enabled = MLX4_MPT_EN_HW;
+
+	mlx4_free_cmd_mailbox(dev, mailbox);
+
+	return 0;
+
+err_cmd:
+	mlx4_free_cmd_mailbox(dev, mailbox);
+
+err_table:
+	mlx4_mpt_free_icm(dev, key_to_hw_index(mr->key));
+	return err;
+}
+EXPORT_SYMBOL_GPL(mlx4_mr_enable);
+
+static int mlx4_write_mtt_chunk(struct mlx4_dev *dev, struct mlx4_mtt *mtt,
+				int start_index, int npages, u64 *page_list)
+{
+	struct mlx4_priv *priv = mlx4_priv(dev);
+	__be64 *mtts;
+	dma_addr_t dma_handle;
+	int i;
+
+	mtts = mlx4_table_find(&priv->mr_table.mtt_table, mtt->offset +
+			       start_index, &dma_handle);
+
+	if (!mtts)
+		return -ENOMEM;
+
+	dma_sync_single_for_cpu(&dev->persist->pdev->dev, dma_handle,
+				npages * sizeof(u64), DMA_TO_DEVICE);
+
+	for (i = 0; i < npages; ++i)
+		mtts[i] = cpu_to_be64(page_list[i] | MLX4_MTT_FLAG_PRESENT);
+
+	dma_sync_single_for_device(&dev->persist->pdev->dev, dma_handle,
+				   npages * sizeof(u64), DMA_TO_DEVICE);
+
+	return 0;
+}
+
+int __mlx4_write_mtt(struct mlx4_dev *dev, struct mlx4_mtt *mtt,
+		     int start_index, int npages, u64 *page_list)
+{
+	int err = 0;
+	int chunk;
+	int mtts_per_page;
+	int max_mtts_first_page;
+
+	/* compute how may mtts fit in the first page */
+	mtts_per_page = PAGE_SIZE / sizeof(u64);
+	max_mtts_first_page = mtts_per_page - (mtt->offset + start_index)
+			      % mtts_per_page;
+
+	chunk = min_t(int, max_mtts_first_page, npages);
+
+	while (npages > 0) {
+		err = mlx4_write_mtt_chunk(dev, mtt, start_index, chunk, page_list);
+		if (err)
+			return err;
+		npages      -= chunk;
+		start_index += chunk;
+		page_list   += chunk;
+
+		chunk = min_t(int, mtts_per_page, npages);
+	}
+	return err;
+}
+
+int mlx4_write_mtt(struct mlx4_dev *dev, struct mlx4_mtt *mtt,
+		   int start_index, int npages, u64 *page_list)
+{
+	struct mlx4_cmd_mailbox *mailbox = NULL;
+	__be64 *inbox = NULL;
+	int chunk;
+	int err = 0;
+	int i;
+
+	if (mtt->order < 0)
+		return -EINVAL;
+
+	if (mlx4_is_mfunc(dev)) {
+		mailbox = mlx4_alloc_cmd_mailbox(dev);
+		if (IS_ERR(mailbox))
+			return PTR_ERR(mailbox);
+		inbox = mailbox->buf;
+
+		while (npages > 0) {
+			chunk = min_t(int, MLX4_MAILBOX_SIZE / sizeof(u64) - 2,
+				      npages);
+			inbox[0] = cpu_to_be64(mtt->offset + start_index);
+			inbox[1] = 0;
+			for (i = 0; i < chunk; ++i)
+				inbox[i + 2] = cpu_to_be64(page_list[i] |
+					       MLX4_MTT_FLAG_PRESENT);
+			err = mlx4_WRITE_MTT(dev, mailbox, chunk);
+			if (err) {
+				mlx4_free_cmd_mailbox(dev, mailbox);
+				return err;
+			}
+
+			npages      -= chunk;
+			start_index += chunk;
+			page_list   += chunk;
+		}
+		mlx4_free_cmd_mailbox(dev, mailbox);
+		return err;
+	}
+
+	return __mlx4_write_mtt(dev, mtt, start_index, npages, page_list);
+}
+EXPORT_SYMBOL_GPL(mlx4_write_mtt);
+
+int mlx4_buf_write_mtt(struct mlx4_dev *dev, struct mlx4_mtt *mtt,
+		       struct mlx4_buf *buf)
+{
+	u64 *page_list;
+	int err;
+	int i;
+
+	page_list = kcalloc(buf->npages, sizeof(*page_list), GFP_KERNEL);
+	if (!page_list)
+		return -ENOMEM;
+
+	for (i = 0; i < buf->npages; ++i)
+		if (buf->nbufs == 1)
+			page_list[i] = buf->direct.map + (i << buf->page_shift);
+		else
+			page_list[i] = buf->page_list[i].map;
+
+	err = mlx4_write_mtt(dev, mtt, 0, buf->npages, page_list);
+
+	kfree(page_list);
+	return err;
+}
+EXPORT_SYMBOL_GPL(mlx4_buf_write_mtt);
+
+int mlx4_mw_alloc(struct mlx4_dev *dev, u32 pd, enum mlx4_mw_type type,
+		  struct mlx4_mw *mw)
+{
+	u32 index;
+
+	if ((type == MLX4_MW_TYPE_1 &&
+	     !(dev->caps.flags & MLX4_DEV_CAP_FLAG_MEM_WINDOW)) ||
+	     (type == MLX4_MW_TYPE_2 &&
+	     !(dev->caps.bmme_flags & MLX4_BMME_FLAG_TYPE_2_WIN)))
+		return -EOPNOTSUPP;
+
+	index = mlx4_mpt_reserve(dev);
+	if (index == -1)
+		return -ENOMEM;
+
+	mw->key	    = hw_index_to_key(index);
+	mw->pd      = pd;
+	mw->type    = type;
+	mw->enabled = MLX4_MPT_DISABLED;
+
+	return 0;
+}
+EXPORT_SYMBOL_GPL(mlx4_mw_alloc);
+
+int mlx4_mw_enable(struct mlx4_dev *dev, struct mlx4_mw *mw)
+{
+	struct mlx4_cmd_mailbox *mailbox;
+	struct mlx4_mpt_entry *mpt_entry;
+	int err;
+
+	err = mlx4_mpt_alloc_icm(dev, key_to_hw_index(mw->key));
+	if (err)
+		return err;
+
+	mailbox = mlx4_alloc_cmd_mailbox(dev);
+	if (IS_ERR(mailbox)) {
+		err = PTR_ERR(mailbox);
+		goto err_table;
+	}
+	mpt_entry = mailbox->buf;
+
+	/* Note that the MLX4_MPT_FLAG_REGION bit in mpt_entry->flags is turned
+	 * off, thus creating a memory window and not a memory region.
+	 */
+	mpt_entry->key	       = cpu_to_be32(key_to_hw_index(mw->key));
+	mpt_entry->pd_flags    = cpu_to_be32(mw->pd);
+	if (mw->type == MLX4_MW_TYPE_2) {
+		mpt_entry->flags    |= cpu_to_be32(MLX4_MPT_FLAG_FREE);
+		mpt_entry->qpn       = cpu_to_be32(MLX4_MPT_QP_FLAG_BOUND_QP);
+		mpt_entry->pd_flags |= cpu_to_be32(MLX4_MPT_PD_FLAG_EN_INV);
+	}
+
+	err = mlx4_SW2HW_MPT(dev, mailbox,
+			     key_to_hw_index(mw->key) &
+			     (dev->caps.num_mpts - 1));
+	if (err) {
+		mlx4_warn(dev, "SW2HW_MPT failed (%d)\n", err);
+		goto err_cmd;
+	}
+	mw->enabled = MLX4_MPT_EN_HW;
+
+	mlx4_free_cmd_mailbox(dev, mailbox);
+
+	return 0;
+
+err_cmd:
+	mlx4_free_cmd_mailbox(dev, mailbox);
+
+err_table:
+	mlx4_mpt_free_icm(dev, key_to_hw_index(mw->key));
+	return err;
+}
+EXPORT_SYMBOL_GPL(mlx4_mw_enable);
+
+void mlx4_mw_free(struct mlx4_dev *dev, struct mlx4_mw *mw)
+{
+	int err;
+
+	if (mw->enabled == MLX4_MPT_EN_HW) {
+		err = mlx4_HW2SW_MPT(dev, NULL,
+				     key_to_hw_index(mw->key) &
+				     (dev->caps.num_mpts - 1));
+		if (err)
+			mlx4_warn(dev, "xxx HW2SW_MPT failed (%d)\n", err);
+
+		mw->enabled = MLX4_MPT_EN_SW;
+	}
+	if (mw->enabled)
+		mlx4_mpt_free_icm(dev, key_to_hw_index(mw->key));
+	mlx4_mpt_release(dev, key_to_hw_index(mw->key));
+}
+EXPORT_SYMBOL_GPL(mlx4_mw_free);
+
+int mlx4_init_mr_table(struct mlx4_dev *dev)
+{
+	struct mlx4_priv *priv = mlx4_priv(dev);
+	struct mlx4_mr_table *mr_table = &priv->mr_table;
+	int err;
+
+	/* Nothing to do for slaves - all MR handling is forwarded
+	* to the master */
+	if (mlx4_is_slave(dev))
+		return 0;
+
+	if (!is_power_of_2(dev->caps.num_mpts))
+		return -EINVAL;
+
+	err = mlx4_bitmap_init(&mr_table->mpt_bitmap, dev->caps.num_mpts,
+			       ~0, dev->caps.reserved_mrws, 0);
+	if (err)
+		return err;
+
+	err = mlx4_buddy_init(&mr_table->mtt_buddy,
+			      ilog2((u32)dev->caps.num_mtts /
+			      (1 << log_mtts_per_seg)));
+	if (err)
+		goto err_buddy;
+
+	if (dev->caps.reserved_mtts) {
+		priv->reserved_mtts =
+			mlx4_alloc_mtt_range(dev,
+					     fls(dev->caps.reserved_mtts - 1));
+		if (priv->reserved_mtts < 0) {
+			mlx4_warn(dev, "MTT table of order %u is too small\n",
+				  mr_table->mtt_buddy.max_order);
+			err = -ENOMEM;
+			goto err_reserve_mtts;
+		}
+	}
+
+	return 0;
+
+err_reserve_mtts:
+	mlx4_buddy_cleanup(&mr_table->mtt_buddy);
+
+err_buddy:
+	mlx4_bitmap_cleanup(&mr_table->mpt_bitmap);
+
+	return err;
+}
+
+void mlx4_cleanup_mr_table(struct mlx4_dev *dev)
+{
+	struct mlx4_priv *priv = mlx4_priv(dev);
+	struct mlx4_mr_table *mr_table = &priv->mr_table;
+
+	if (mlx4_is_slave(dev))
+		return;
+	if (priv->reserved_mtts >= 0)
+		mlx4_free_mtt_range(dev, priv->reserved_mtts,
+				    fls(dev->caps.reserved_mtts - 1));
+	mlx4_buddy_cleanup(&mr_table->mtt_buddy);
+	mlx4_bitmap_cleanup(&mr_table->mpt_bitmap);
+}
+
+static inline int mlx4_check_fmr(struct mlx4_fmr *fmr, u64 *page_list,
+				  int npages, u64 iova)
+{
+	int i, page_mask;
+
+	if (npages > fmr->max_pages)
+		return -EINVAL;
+
+	page_mask = (1 << fmr->page_shift) - 1;
+
+	/* We are getting page lists, so va must be page aligned. */
+	if (iova & page_mask)
+		return -EINVAL;
+
+	/* Trust the user not to pass misaligned data in page_list */
+	if (0)
+		for (i = 0; i < npages; ++i) {
+			if (page_list[i] & ~page_mask)
+				return -EINVAL;
+		}
+
+	if (fmr->maps >= fmr->max_maps)
+		return -EINVAL;
+
+	return 0;
+}
+
+int mlx4_map_phys_fmr(struct mlx4_dev *dev, struct mlx4_fmr *fmr, u64 *page_list,
+		      int npages, u64 iova, u32 *lkey, u32 *rkey)
+{
+	u32 key;
+	int i, err;
+
+	err = mlx4_check_fmr(fmr, page_list, npages, iova);
+	if (err)
+		return err;
+
+	++fmr->maps;
+
+	key = key_to_hw_index(fmr->mr.key);
+	key += dev->caps.num_mpts;
+	*lkey = *rkey = fmr->mr.key = hw_index_to_key(key);
+
+	*(u8 *) fmr->mpt = MLX4_MPT_STATUS_SW;
+
+	/* Make sure MPT status is visible before writing MTT entries */
+	wmb();
+
+	dma_sync_single_for_cpu(&dev->persist->pdev->dev, fmr->dma_handle,
+				npages * sizeof(u64), DMA_TO_DEVICE);
+
+	for (i = 0; i < npages; ++i)
+		fmr->mtts[i] = cpu_to_be64(page_list[i] | MLX4_MTT_FLAG_PRESENT);
+
+	dma_sync_single_for_device(&dev->persist->pdev->dev, fmr->dma_handle,
+				   npages * sizeof(u64), DMA_TO_DEVICE);
+
+	fmr->mpt->key    = cpu_to_be32(key);
+	fmr->mpt->lkey   = cpu_to_be32(key);
+	fmr->mpt->length = cpu_to_be64(npages * (1ull << fmr->page_shift));
+	fmr->mpt->start  = cpu_to_be64(iova);
+
+	/* Make MTT entries are visible before setting MPT status */
+	wmb();
+
+	*(u8 *) fmr->mpt = MLX4_MPT_STATUS_HW;
+
+	/* Make sure MPT status is visible before consumer can use FMR */
+	wmb();
+
+	return 0;
+}
+EXPORT_SYMBOL_GPL(mlx4_map_phys_fmr);
+
+int mlx4_fmr_alloc(struct mlx4_dev *dev, u32 pd, u32 access, int max_pages,
+		   int max_maps, u8 page_shift, struct mlx4_fmr *fmr)
+{
+	struct mlx4_priv *priv = mlx4_priv(dev);
+	int err = -ENOMEM;
+
+	if (max_maps > dev->caps.max_fmr_maps)
+		return -EINVAL;
+
+	if (page_shift < (ffs(dev->caps.page_size_cap) - 1) || page_shift >= 32)
+		return -EINVAL;
+
+	/* All MTTs must fit in the same page */
+	if (max_pages * sizeof(*fmr->mtts) > PAGE_SIZE)
+		return -EINVAL;
+
+	fmr->page_shift = page_shift;
+	fmr->max_pages  = max_pages;
+	fmr->max_maps   = max_maps;
+	fmr->maps = 0;
+
+	err = mlx4_mr_alloc(dev, pd, 0, 0, access, max_pages,
+			    page_shift, &fmr->mr);
+	if (err)
+		return err;
+
+	fmr->mtts = mlx4_table_find(&priv->mr_table.mtt_table,
+				    fmr->mr.mtt.offset,
+				    &fmr->dma_handle);
+
+	if (!fmr->mtts) {
+		err = -ENOMEM;
+		goto err_free;
+	}
+
+	return 0;
+
+err_free:
+	(void) mlx4_mr_free(dev, &fmr->mr);
+	return err;
+}
+EXPORT_SYMBOL_GPL(mlx4_fmr_alloc);
+
+int mlx4_fmr_enable(struct mlx4_dev *dev, struct mlx4_fmr *fmr)
+{
+	struct mlx4_priv *priv = mlx4_priv(dev);
+	int err;
+
+	err = mlx4_mr_enable(dev, &fmr->mr);
+	if (err)
+		return err;
+
+	fmr->mpt = mlx4_table_find(&priv->mr_table.dmpt_table,
+				    key_to_hw_index(fmr->mr.key), NULL);
+	if (!fmr->mpt)
+		return -ENOMEM;
+
+	return 0;
+}
+EXPORT_SYMBOL_GPL(mlx4_fmr_enable);
+
+void mlx4_fmr_unmap(struct mlx4_dev *dev, struct mlx4_fmr *fmr,
+		    u32 *lkey, u32 *rkey)
+{
+	if (!fmr->maps)
+		return;
+
+	/* To unmap: it is sufficient to take back ownership from HW */
+	*(u8 *)fmr->mpt = MLX4_MPT_STATUS_SW;
+
+	/* Make sure MPT status is visible */
+	wmb();
+
+	fmr->maps = 0;
+}
+EXPORT_SYMBOL_GPL(mlx4_fmr_unmap);
+
+int mlx4_fmr_free(struct mlx4_dev *dev, struct mlx4_fmr *fmr)
+{
+	int ret;
+
+	if (fmr->maps)
+		return -EBUSY;
+	if (fmr->mr.enabled == MLX4_MPT_EN_HW) {
+		/* In case of FMR was enabled and unmapped
+		 * make sure to give ownership of MPT back to HW
+		 * so HW2SW_MPT command will success.
+		 */
+		*(u8 *)fmr->mpt = MLX4_MPT_STATUS_SW;
+		/* Make sure MPT status is visible before changing MPT fields */
+		wmb();
+		fmr->mpt->length = 0;
+		fmr->mpt->start  = 0;
+		/* Make sure MPT data is visible after changing MPT status */
+		wmb();
+		*(u8 *)fmr->mpt = MLX4_MPT_STATUS_HW;
+		/* make sure MPT status is visible */
+		wmb();
+	}
+
+	ret = mlx4_mr_free(dev, &fmr->mr);
+	if (ret)
+		return ret;
+	fmr->mr.enabled = MLX4_MPT_DISABLED;
+
+	return 0;
+}
+EXPORT_SYMBOL_GPL(mlx4_fmr_free);
+
+int mlx4_SYNC_TPT(struct mlx4_dev *dev)
+{
+	return mlx4_cmd(dev, 0, 0, 0, MLX4_CMD_SYNC_TPT,
+			MLX4_CMD_TIME_CLASS_A, MLX4_CMD_NATIVE);
+}
+EXPORT_SYMBOL_GPL(mlx4_SYNC_TPT);
diff --git a/drivers/net/ethernet/mellanox/mlx4/pd.c b/drivers/net/ethernet/mellanox/mlx4/pd.c
new file mode 100644
index 000000000..6fc156a39
--- /dev/null
+++ b/drivers/net/ethernet/mellanox/mlx4/pd.c
@@ -0,0 +1,295 @@
+/*
+ * Copyright (c) 2006, 2007 Cisco Systems, Inc.  All rights reserved.
+ * Copyright (c) 2005 Mellanox Technologies. All rights reserved.
+ *
+ * This software is available to you under a choice of one of two
+ * licenses.  You may choose to be licensed under the terms of the GNU
+ * General Public License (GPL) Version 2, available from the file
+ * COPYING in the main directory of this source tree, or the
+ * OpenIB.org BSD license below:
+ *
+ *     Redistribution and use in source and binary forms, with or
+ *     without modification, are permitted provided that the following
+ *     conditions are met:
+ *
+ *      - Redistributions of source code must retain the above
+ *        copyright notice, this list of conditions and the following
+ *        disclaimer.
+ *
+ *      - Redistributions in binary form must reproduce the above
+ *        copyright notice, this list of conditions and the following
+ *        disclaimer in the documentation and/or other materials
+ *        provided with the distribution.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
+ * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
+ * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS
+ * BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN
+ * ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
+ * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+ * SOFTWARE.
+ */
+
+#include <linux/errno.h>
+#include <linux/export.h>
+#include <linux/io-mapping.h>
+
+#include <asm/page.h>
+
+#include "mlx4.h"
+#include "icm.h"
+
+enum {
+	MLX4_NUM_RESERVED_UARS = 8
+};
+
+int mlx4_pd_alloc(struct mlx4_dev *dev, u32 *pdn)
+{
+	struct mlx4_priv *priv = mlx4_priv(dev);
+
+	*pdn = mlx4_bitmap_alloc(&priv->pd_bitmap);
+	if (*pdn == -1)
+		return -ENOMEM;
+
+	return 0;
+}
+EXPORT_SYMBOL_GPL(mlx4_pd_alloc);
+
+void mlx4_pd_free(struct mlx4_dev *dev, u32 pdn)
+{
+	mlx4_bitmap_free(&mlx4_priv(dev)->pd_bitmap, pdn, MLX4_USE_RR);
+}
+EXPORT_SYMBOL_GPL(mlx4_pd_free);
+
+int __mlx4_xrcd_alloc(struct mlx4_dev *dev, u32 *xrcdn)
+{
+	struct mlx4_priv *priv = mlx4_priv(dev);
+
+	*xrcdn = mlx4_bitmap_alloc(&priv->xrcd_bitmap);
+	if (*xrcdn == -1)
+		return -ENOMEM;
+
+	return 0;
+}
+
+int mlx4_xrcd_alloc(struct mlx4_dev *dev, u32 *xrcdn)
+{
+	u64 out_param;
+	int err;
+
+	if (mlx4_is_mfunc(dev)) {
+		err = mlx4_cmd_imm(dev, 0, &out_param,
+				   RES_XRCD, RES_OP_RESERVE,
+				   MLX4_CMD_ALLOC_RES,
+				   MLX4_CMD_TIME_CLASS_A, MLX4_CMD_WRAPPED);
+		if (err)
+			return err;
+
+		*xrcdn = get_param_l(&out_param);
+		return 0;
+	}
+	return __mlx4_xrcd_alloc(dev, xrcdn);
+}
+EXPORT_SYMBOL_GPL(mlx4_xrcd_alloc);
+
+void __mlx4_xrcd_free(struct mlx4_dev *dev, u32 xrcdn)
+{
+	mlx4_bitmap_free(&mlx4_priv(dev)->xrcd_bitmap, xrcdn, MLX4_USE_RR);
+}
+
+void mlx4_xrcd_free(struct mlx4_dev *dev, u32 xrcdn)
+{
+	u64 in_param = 0;
+	int err;
+
+	if (mlx4_is_mfunc(dev)) {
+		set_param_l(&in_param, xrcdn);
+		err = mlx4_cmd(dev, in_param, RES_XRCD,
+			       RES_OP_RESERVE, MLX4_CMD_FREE_RES,
+			       MLX4_CMD_TIME_CLASS_A, MLX4_CMD_WRAPPED);
+		if (err)
+			mlx4_warn(dev, "Failed to release xrcdn %d\n", xrcdn);
+	} else
+		__mlx4_xrcd_free(dev, xrcdn);
+}
+EXPORT_SYMBOL_GPL(mlx4_xrcd_free);
+
+int mlx4_init_pd_table(struct mlx4_dev *dev)
+{
+	struct mlx4_priv *priv = mlx4_priv(dev);
+
+	return mlx4_bitmap_init(&priv->pd_bitmap, dev->caps.num_pds,
+				(1 << NOT_MASKED_PD_BITS) - 1,
+				 dev->caps.reserved_pds, 0);
+}
+
+void mlx4_cleanup_pd_table(struct mlx4_dev *dev)
+{
+	mlx4_bitmap_cleanup(&mlx4_priv(dev)->pd_bitmap);
+}
+
+int mlx4_init_xrcd_table(struct mlx4_dev *dev)
+{
+	struct mlx4_priv *priv = mlx4_priv(dev);
+
+	return mlx4_bitmap_init(&priv->xrcd_bitmap, (1 << 16),
+				(1 << 16) - 1, dev->caps.reserved_xrcds + 1, 0);
+}
+
+void mlx4_cleanup_xrcd_table(struct mlx4_dev *dev)
+{
+	mlx4_bitmap_cleanup(&mlx4_priv(dev)->xrcd_bitmap);
+}
+
+int mlx4_uar_alloc(struct mlx4_dev *dev, struct mlx4_uar *uar)
+{
+	int offset;
+
+	uar->index = mlx4_bitmap_alloc(&mlx4_priv(dev)->uar_table.bitmap);
+	if (uar->index == -1)
+		return -ENOMEM;
+
+	if (mlx4_is_slave(dev))
+		offset = uar->index % ((int)pci_resource_len(dev->persist->pdev,
+							     2) /
+				       dev->caps.uar_page_size);
+	else
+		offset = uar->index;
+	uar->pfn = (pci_resource_start(dev->persist->pdev, 2) >> PAGE_SHIFT)
+		    + offset;
+	uar->map = NULL;
+	return 0;
+}
+EXPORT_SYMBOL_GPL(mlx4_uar_alloc);
+
+void mlx4_uar_free(struct mlx4_dev *dev, struct mlx4_uar *uar)
+{
+	mlx4_bitmap_free(&mlx4_priv(dev)->uar_table.bitmap, uar->index, MLX4_USE_RR);
+}
+EXPORT_SYMBOL_GPL(mlx4_uar_free);
+
+int mlx4_bf_alloc(struct mlx4_dev *dev, struct mlx4_bf *bf, int node)
+{
+	struct mlx4_priv *priv = mlx4_priv(dev);
+	struct mlx4_uar *uar;
+	int err = 0;
+	int idx;
+
+	if (!priv->bf_mapping)
+		return -ENOMEM;
+
+	mutex_lock(&priv->bf_mutex);
+	if (!list_empty(&priv->bf_list))
+		uar = list_entry(priv->bf_list.next, struct mlx4_uar, bf_list);
+	else {
+		if (mlx4_bitmap_avail(&priv->uar_table.bitmap) < MLX4_NUM_RESERVED_UARS) {
+			err = -ENOMEM;
+			goto out;
+		}
+		uar = kmalloc_node(sizeof(*uar), GFP_KERNEL, node);
+		if (!uar) {
+			uar = kmalloc(sizeof(*uar), GFP_KERNEL);
+			if (!uar) {
+				err = -ENOMEM;
+				goto out;
+			}
+		}
+		err = mlx4_uar_alloc(dev, uar);
+		if (err)
+			goto free_kmalloc;
+
+		uar->map = ioremap(uar->pfn << PAGE_SHIFT, PAGE_SIZE);
+		if (!uar->map) {
+			err = -ENOMEM;
+			goto free_uar;
+		}
+
+		uar->bf_map = io_mapping_map_wc(priv->bf_mapping,
+						uar->index << PAGE_SHIFT,
+						PAGE_SIZE);
+		if (!uar->bf_map) {
+			err = -ENOMEM;
+			goto unamp_uar;
+		}
+		uar->free_bf_bmap = 0;
+		list_add(&uar->bf_list, &priv->bf_list);
+	}
+
+	idx = ffz(uar->free_bf_bmap);
+	uar->free_bf_bmap |= 1 << idx;
+	bf->uar = uar;
+	bf->offset = 0;
+	bf->buf_size = dev->caps.bf_reg_size / 2;
+	bf->reg = uar->bf_map + idx * dev->caps.bf_reg_size;
+	if (uar->free_bf_bmap == (1 << dev->caps.bf_regs_per_page) - 1)
+		list_del_init(&uar->bf_list);
+
+	goto out;
+
+unamp_uar:
+	bf->uar = NULL;
+	iounmap(uar->map);
+
+free_uar:
+	mlx4_uar_free(dev, uar);
+
+free_kmalloc:
+	kfree(uar);
+
+out:
+	mutex_unlock(&priv->bf_mutex);
+	return err;
+}
+EXPORT_SYMBOL_GPL(mlx4_bf_alloc);
+
+void mlx4_bf_free(struct mlx4_dev *dev, struct mlx4_bf *bf)
+{
+	struct mlx4_priv *priv = mlx4_priv(dev);
+	int idx;
+
+	if (!bf->uar || !bf->uar->bf_map)
+		return;
+
+	mutex_lock(&priv->bf_mutex);
+	idx = (bf->reg - bf->uar->bf_map) / dev->caps.bf_reg_size;
+	bf->uar->free_bf_bmap &= ~(1 << idx);
+	if (!bf->uar->free_bf_bmap) {
+		if (!list_empty(&bf->uar->bf_list))
+			list_del(&bf->uar->bf_list);
+
+		io_mapping_unmap(bf->uar->bf_map);
+		iounmap(bf->uar->map);
+		mlx4_uar_free(dev, bf->uar);
+		kfree(bf->uar);
+	} else if (list_empty(&bf->uar->bf_list))
+		list_add(&bf->uar->bf_list, &priv->bf_list);
+
+	mutex_unlock(&priv->bf_mutex);
+}
+EXPORT_SYMBOL_GPL(mlx4_bf_free);
+
+int mlx4_init_uar_table(struct mlx4_dev *dev)
+{
+	int num_reserved_uar = mlx4_get_num_reserved_uar(dev);
+
+	mlx4_dbg(dev, "uar_page_shift = %d", dev->uar_page_shift);
+	mlx4_dbg(dev, "Effective reserved_uars=%d", dev->caps.reserved_uars);
+
+	if (dev->caps.num_uars <= num_reserved_uar) {
+		mlx4_err(
+			dev, "Only %d UAR pages (need more than %d)\n",
+			dev->caps.num_uars, num_reserved_uar);
+		mlx4_err(dev, "Increase firmware log2_uar_bar_megabytes?\n");
+		return -ENODEV;
+	}
+
+	return mlx4_bitmap_init(&mlx4_priv(dev)->uar_table.bitmap,
+				dev->caps.num_uars, dev->caps.num_uars - 1,
+				dev->caps.reserved_uars, 0);
+}
+
+void mlx4_cleanup_uar_table(struct mlx4_dev *dev)
+{
+	mlx4_bitmap_cleanup(&mlx4_priv(dev)->uar_table.bitmap);
+}
diff --git a/drivers/net/ethernet/mellanox/mlx4/port.c b/drivers/net/ethernet/mellanox/mlx4/port.c
new file mode 100644
index 000000000..256a06b3c
--- /dev/null
+++ b/drivers/net/ethernet/mellanox/mlx4/port.c
@@ -0,0 +1,2229 @@
+/*
+ * Copyright (c) 2007 Mellanox Technologies. All rights reserved.
+ *
+ * This software is available to you under a choice of one of two
+ * licenses.  You may choose to be licensed under the terms of the GNU
+ * General Public License (GPL) Version 2, available from the file
+ * COPYING in the main directory of this source tree, or the
+ * OpenIB.org BSD license below:
+ *
+ *     Redistribution and use in source and binary forms, with or
+ *     without modification, are permitted provided that the following
+ *     conditions are met:
+ *
+ *      - Redistributions of source code must retain the above
+ *        copyright notice, this list of conditions and the following
+ *        disclaimer.
+ *
+ *      - Redistributions in binary form must reproduce the above
+ *        copyright notice, this list of conditions and the following
+ *        disclaimer in the documentation and/or other materials
+ *        provided with the distribution.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
+ * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
+ * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS
+ * BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN
+ * ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
+ * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+ * SOFTWARE.
+ */
+
+#include <linux/errno.h>
+#include <linux/if_ether.h>
+#include <linux/if_vlan.h>
+#include <linux/export.h>
+
+#include <linux/mlx4/cmd.h>
+
+#include "mlx4.h"
+#include "mlx4_stats.h"
+
+#define MLX4_MAC_VALID		(1ull << 63)
+
+#define MLX4_VLAN_VALID		(1u << 31)
+#define MLX4_VLAN_MASK		0xfff
+
+#define MLX4_STATS_TRAFFIC_COUNTERS_MASK	0xfULL
+#define MLX4_STATS_TRAFFIC_DROPS_MASK		0xc0ULL
+#define MLX4_STATS_ERROR_COUNTERS_MASK		0x1ffc30ULL
+#define MLX4_STATS_PORT_COUNTERS_MASK		0x1fe00000ULL
+
+#define MLX4_FLAG2_V_IGNORE_FCS_MASK		BIT(1)
+#define MLX4_FLAG2_V_USER_MTU_MASK		BIT(5)
+#define MLX4_FLAG2_V_USER_MAC_MASK		BIT(6)
+#define MLX4_FLAG_V_MTU_MASK			BIT(0)
+#define MLX4_FLAG_V_PPRX_MASK			BIT(1)
+#define MLX4_FLAG_V_PPTX_MASK			BIT(2)
+#define MLX4_IGNORE_FCS_MASK			0x1
+#define MLX4_TC_MAX_NUMBER			8
+
+void mlx4_init_mac_table(struct mlx4_dev *dev, struct mlx4_mac_table *table)
+{
+	int i;
+
+	mutex_init(&table->mutex);
+	for (i = 0; i < MLX4_MAX_MAC_NUM; i++) {
+		table->entries[i] = 0;
+		table->refs[i]	 = 0;
+		table->is_dup[i] = false;
+	}
+	table->max   = 1 << dev->caps.log_num_macs;
+	table->total = 0;
+}
+
+void mlx4_init_vlan_table(struct mlx4_dev *dev, struct mlx4_vlan_table *table)
+{
+	int i;
+
+	mutex_init(&table->mutex);
+	for (i = 0; i < MLX4_MAX_VLAN_NUM; i++) {
+		table->entries[i] = 0;
+		table->refs[i]	 = 0;
+		table->is_dup[i] = false;
+	}
+	table->max   = (1 << dev->caps.log_num_vlans) - MLX4_VLAN_REGULAR;
+	table->total = 0;
+}
+
+void mlx4_init_roce_gid_table(struct mlx4_dev *dev,
+			      struct mlx4_roce_gid_table *table)
+{
+	int i;
+
+	mutex_init(&table->mutex);
+	for (i = 0; i < MLX4_ROCE_MAX_GIDS; i++)
+		memset(table->roce_gids[i].raw, 0, MLX4_ROCE_GID_ENTRY_SIZE);
+}
+
+static int validate_index(struct mlx4_dev *dev,
+			  struct mlx4_mac_table *table, int index)
+{
+	int err = 0;
+
+	if (index < 0 || index >= table->max || !table->entries[index]) {
+		mlx4_warn(dev, "No valid Mac entry for the given index\n");
+		err = -EINVAL;
+	}
+	return err;
+}
+
+static int find_index(struct mlx4_dev *dev,
+		      struct mlx4_mac_table *table, u64 mac)
+{
+	int i;
+
+	for (i = 0; i < MLX4_MAX_MAC_NUM; i++) {
+		if (table->refs[i] &&
+		    (MLX4_MAC_MASK & mac) ==
+		    (MLX4_MAC_MASK & be64_to_cpu(table->entries[i])))
+			return i;
+	}
+	/* Mac not found */
+	return -EINVAL;
+}
+
+static int mlx4_set_port_mac_table(struct mlx4_dev *dev, u8 port,
+				   __be64 *entries)
+{
+	struct mlx4_cmd_mailbox *mailbox;
+	u32 in_mod;
+	int err;
+
+	mailbox = mlx4_alloc_cmd_mailbox(dev);
+	if (IS_ERR(mailbox))
+		return PTR_ERR(mailbox);
+
+	memcpy(mailbox->buf, entries, MLX4_MAC_TABLE_SIZE);
+
+	in_mod = MLX4_SET_PORT_MAC_TABLE << 8 | port;
+
+	err = mlx4_cmd(dev, mailbox->dma, in_mod, MLX4_SET_PORT_ETH_OPCODE,
+		       MLX4_CMD_SET_PORT, MLX4_CMD_TIME_CLASS_B,
+		       MLX4_CMD_NATIVE);
+
+	mlx4_free_cmd_mailbox(dev, mailbox);
+	return err;
+}
+
+int mlx4_find_cached_mac(struct mlx4_dev *dev, u8 port, u64 mac, int *idx)
+{
+	struct mlx4_port_info *info = &mlx4_priv(dev)->port[port];
+	struct mlx4_mac_table *table = &info->mac_table;
+	int i;
+
+	for (i = 0; i < MLX4_MAX_MAC_NUM; i++) {
+		if (!table->refs[i])
+			continue;
+
+		if (mac == (MLX4_MAC_MASK & be64_to_cpu(table->entries[i]))) {
+			*idx = i;
+			return 0;
+		}
+	}
+
+	return -ENOENT;
+}
+EXPORT_SYMBOL_GPL(mlx4_find_cached_mac);
+
+static bool mlx4_need_mf_bond(struct mlx4_dev *dev)
+{
+	int i, num_eth_ports = 0;
+
+	if (!mlx4_is_mfunc(dev))
+		return false;
+	mlx4_foreach_port(i, dev, MLX4_PORT_TYPE_ETH)
+		++num_eth_ports;
+
+	return (num_eth_ports ==  2) ? true : false;
+}
+
+int __mlx4_register_mac(struct mlx4_dev *dev, u8 port, u64 mac)
+{
+	struct mlx4_port_info *info = &mlx4_priv(dev)->port[port];
+	struct mlx4_mac_table *table = &info->mac_table;
+	int i, err = 0;
+	int free = -1;
+	int free_for_dup = -1;
+	bool dup = mlx4_is_mf_bonded(dev);
+	u8 dup_port = (port == 1) ? 2 : 1;
+	struct mlx4_mac_table *dup_table = &mlx4_priv(dev)->port[dup_port].mac_table;
+	bool need_mf_bond = mlx4_need_mf_bond(dev);
+	bool can_mf_bond = true;
+
+	mlx4_dbg(dev, "Registering MAC: 0x%llx for port %d %s duplicate\n",
+		 (unsigned long long)mac, port,
+		 dup ? "with" : "without");
+
+	if (need_mf_bond) {
+		if (port == 1) {
+			mutex_lock(&table->mutex);
+			mutex_lock_nested(&dup_table->mutex, SINGLE_DEPTH_NESTING);
+		} else {
+			mutex_lock(&dup_table->mutex);
+			mutex_lock_nested(&table->mutex, SINGLE_DEPTH_NESTING);
+		}
+	} else {
+		mutex_lock(&table->mutex);
+	}
+
+	if (need_mf_bond) {
+		int index_at_port = -1;
+		int index_at_dup_port = -1;
+
+		for (i = 0; i < MLX4_MAX_MAC_NUM; i++) {
+			if (((MLX4_MAC_MASK & mac) == (MLX4_MAC_MASK & be64_to_cpu(table->entries[i]))))
+				index_at_port = i;
+			if (((MLX4_MAC_MASK & mac) == (MLX4_MAC_MASK & be64_to_cpu(dup_table->entries[i]))))
+				index_at_dup_port = i;
+		}
+
+		/* check that same mac is not in the tables at different indices */
+		if ((index_at_port != index_at_dup_port) &&
+		    (index_at_port >= 0) &&
+		    (index_at_dup_port >= 0))
+			can_mf_bond = false;
+
+		/* If the mac is already in the primary table, the slot must be
+		 * available in the duplicate table as well.
+		 */
+		if (index_at_port >= 0 && index_at_dup_port < 0 &&
+		    dup_table->refs[index_at_port]) {
+			can_mf_bond = false;
+		}
+		/* If the mac is already in the duplicate table, check that the
+		 * corresponding index is not occupied in the primary table, or
+		 * the primary table already contains the mac at the same index.
+		 * Otherwise, you cannot bond (primary contains a different mac
+		 * at that index).
+		 */
+		if (index_at_dup_port >= 0) {
+			if (!table->refs[index_at_dup_port] ||
+			    ((MLX4_MAC_MASK & mac) == (MLX4_MAC_MASK & be64_to_cpu(table->entries[index_at_dup_port]))))
+				free_for_dup = index_at_dup_port;
+			else
+				can_mf_bond = false;
+		}
+	}
+
+	for (i = 0; i < MLX4_MAX_MAC_NUM; i++) {
+		if (!table->refs[i]) {
+			if (free < 0)
+				free = i;
+			if (free_for_dup < 0 && need_mf_bond && can_mf_bond) {
+				if (!dup_table->refs[i])
+					free_for_dup = i;
+			}
+			continue;
+		}
+
+		if ((MLX4_MAC_MASK & mac) ==
+		     (MLX4_MAC_MASK & be64_to_cpu(table->entries[i]))) {
+			/* MAC already registered, increment ref count */
+			err = i;
+			++table->refs[i];
+			if (dup) {
+				u64 dup_mac = MLX4_MAC_MASK & be64_to_cpu(dup_table->entries[i]);
+
+				if (dup_mac != mac || !dup_table->is_dup[i]) {
+					mlx4_warn(dev, "register mac: expect duplicate mac 0x%llx on port %d index %d\n",
+						  mac, dup_port, i);
+				}
+			}
+			goto out;
+		}
+	}
+
+	if (need_mf_bond && (free_for_dup < 0)) {
+		if (dup) {
+			mlx4_warn(dev, "Fail to allocate duplicate MAC table entry\n");
+			mlx4_warn(dev, "High Availability for virtual functions may not work as expected\n");
+			dup = false;
+		}
+		can_mf_bond = false;
+	}
+
+	if (need_mf_bond && can_mf_bond)
+		free = free_for_dup;
+
+	mlx4_dbg(dev, "Free MAC index is %d\n", free);
+
+	if (table->total == table->max) {
+		/* No free mac entries */
+		err = -ENOSPC;
+		goto out;
+	}
+
+	/* Register new MAC */
+	table->entries[free] = cpu_to_be64(mac | MLX4_MAC_VALID);
+
+	err = mlx4_set_port_mac_table(dev, port, table->entries);
+	if (unlikely(err)) {
+		mlx4_err(dev, "Failed adding MAC: 0x%llx\n",
+			 (unsigned long long) mac);
+		table->entries[free] = 0;
+		goto out;
+	}
+	table->refs[free] = 1;
+	table->is_dup[free] = false;
+	++table->total;
+	if (dup) {
+		dup_table->refs[free] = 0;
+		dup_table->is_dup[free] = true;
+		dup_table->entries[free] = cpu_to_be64(mac | MLX4_MAC_VALID);
+
+		err = mlx4_set_port_mac_table(dev, dup_port, dup_table->entries);
+		if (unlikely(err)) {
+			mlx4_warn(dev, "Failed adding duplicate mac: 0x%llx\n", mac);
+			dup_table->is_dup[free] = false;
+			dup_table->entries[free] = 0;
+			goto out;
+		}
+		++dup_table->total;
+	}
+	err = free;
+out:
+	if (need_mf_bond) {
+		if (port == 2) {
+			mutex_unlock(&table->mutex);
+			mutex_unlock(&dup_table->mutex);
+		} else {
+			mutex_unlock(&dup_table->mutex);
+			mutex_unlock(&table->mutex);
+		}
+	} else {
+		mutex_unlock(&table->mutex);
+	}
+	return err;
+}
+EXPORT_SYMBOL_GPL(__mlx4_register_mac);
+
+int mlx4_register_mac(struct mlx4_dev *dev, u8 port, u64 mac)
+{
+	u64 out_param = 0;
+	int err = -EINVAL;
+
+	if (mlx4_is_mfunc(dev)) {
+		if (!(dev->flags & MLX4_FLAG_OLD_REG_MAC)) {
+			err = mlx4_cmd_imm(dev, mac, &out_param,
+					   ((u32) port) << 8 | (u32) RES_MAC,
+					   RES_OP_RESERVE_AND_MAP, MLX4_CMD_ALLOC_RES,
+					   MLX4_CMD_TIME_CLASS_A, MLX4_CMD_WRAPPED);
+		}
+		if (err && err == -EINVAL && mlx4_is_slave(dev)) {
+			/* retry using old REG_MAC format */
+			set_param_l(&out_param, port);
+			err = mlx4_cmd_imm(dev, mac, &out_param, RES_MAC,
+					   RES_OP_RESERVE_AND_MAP, MLX4_CMD_ALLOC_RES,
+					   MLX4_CMD_TIME_CLASS_A, MLX4_CMD_WRAPPED);
+			if (!err)
+				dev->flags |= MLX4_FLAG_OLD_REG_MAC;
+		}
+		if (err)
+			return err;
+
+		return get_param_l(&out_param);
+	}
+	return __mlx4_register_mac(dev, port, mac);
+}
+EXPORT_SYMBOL_GPL(mlx4_register_mac);
+
+int mlx4_get_base_qpn(struct mlx4_dev *dev, u8 port)
+{
+	return dev->caps.reserved_qps_base[MLX4_QP_REGION_ETH_ADDR] +
+			(port - 1) * (1 << dev->caps.log_num_macs);
+}
+EXPORT_SYMBOL_GPL(mlx4_get_base_qpn);
+
+void __mlx4_unregister_mac(struct mlx4_dev *dev, u8 port, u64 mac)
+{
+	struct mlx4_port_info *info;
+	struct mlx4_mac_table *table;
+	int index;
+	bool dup = mlx4_is_mf_bonded(dev);
+	u8 dup_port = (port == 1) ? 2 : 1;
+	struct mlx4_mac_table *dup_table = &mlx4_priv(dev)->port[dup_port].mac_table;
+
+	if (port < 1 || port > dev->caps.num_ports) {
+		mlx4_warn(dev, "invalid port number (%d), aborting...\n", port);
+		return;
+	}
+	info = &mlx4_priv(dev)->port[port];
+	table = &info->mac_table;
+
+	if (dup) {
+		if (port == 1) {
+			mutex_lock(&table->mutex);
+			mutex_lock_nested(&dup_table->mutex, SINGLE_DEPTH_NESTING);
+		} else {
+			mutex_lock(&dup_table->mutex);
+			mutex_lock_nested(&table->mutex, SINGLE_DEPTH_NESTING);
+		}
+	} else {
+		mutex_lock(&table->mutex);
+	}
+
+	index = find_index(dev, table, mac);
+
+	if (validate_index(dev, table, index))
+		goto out;
+
+	if (--table->refs[index] || table->is_dup[index]) {
+		mlx4_dbg(dev, "Have more references for index %d, no need to modify mac table\n",
+			 index);
+		if (!table->refs[index])
+			dup_table->is_dup[index] = false;
+		goto out;
+	}
+
+	table->entries[index] = 0;
+	if (mlx4_set_port_mac_table(dev, port, table->entries))
+		mlx4_warn(dev, "Fail to set mac in port %d during unregister\n", port);
+	--table->total;
+
+	if (dup) {
+		dup_table->is_dup[index] = false;
+		if (dup_table->refs[index])
+			goto out;
+		dup_table->entries[index] = 0;
+		if (mlx4_set_port_mac_table(dev, dup_port, dup_table->entries))
+			mlx4_warn(dev, "Fail to set mac in duplicate port %d during unregister\n", dup_port);
+
+		--table->total;
+	}
+out:
+	if (dup) {
+		if (port == 2) {
+			mutex_unlock(&table->mutex);
+			mutex_unlock(&dup_table->mutex);
+		} else {
+			mutex_unlock(&dup_table->mutex);
+			mutex_unlock(&table->mutex);
+		}
+	} else {
+		mutex_unlock(&table->mutex);
+	}
+}
+EXPORT_SYMBOL_GPL(__mlx4_unregister_mac);
+
+void mlx4_unregister_mac(struct mlx4_dev *dev, u8 port, u64 mac)
+{
+	u64 out_param = 0;
+
+	if (mlx4_is_mfunc(dev)) {
+		if (!(dev->flags & MLX4_FLAG_OLD_REG_MAC)) {
+			(void) mlx4_cmd_imm(dev, mac, &out_param,
+					    ((u32) port) << 8 | (u32) RES_MAC,
+					    RES_OP_RESERVE_AND_MAP, MLX4_CMD_FREE_RES,
+					    MLX4_CMD_TIME_CLASS_A, MLX4_CMD_WRAPPED);
+		} else {
+			/* use old unregister mac format */
+			set_param_l(&out_param, port);
+			(void) mlx4_cmd_imm(dev, mac, &out_param, RES_MAC,
+					    RES_OP_RESERVE_AND_MAP, MLX4_CMD_FREE_RES,
+					    MLX4_CMD_TIME_CLASS_A, MLX4_CMD_WRAPPED);
+		}
+		return;
+	}
+	__mlx4_unregister_mac(dev, port, mac);
+	return;
+}
+EXPORT_SYMBOL_GPL(mlx4_unregister_mac);
+
+int __mlx4_replace_mac(struct mlx4_dev *dev, u8 port, int qpn, u64 new_mac)
+{
+	struct mlx4_port_info *info = &mlx4_priv(dev)->port[port];
+	struct mlx4_mac_table *table = &info->mac_table;
+	int index = qpn - info->base_qpn;
+	int err = 0;
+	bool dup = mlx4_is_mf_bonded(dev);
+	u8 dup_port = (port == 1) ? 2 : 1;
+	struct mlx4_mac_table *dup_table = &mlx4_priv(dev)->port[dup_port].mac_table;
+
+	/* CX1 doesn't support multi-functions */
+	if (dup) {
+		if (port == 1) {
+			mutex_lock(&table->mutex);
+			mutex_lock_nested(&dup_table->mutex, SINGLE_DEPTH_NESTING);
+		} else {
+			mutex_lock(&dup_table->mutex);
+			mutex_lock_nested(&table->mutex, SINGLE_DEPTH_NESTING);
+		}
+	} else {
+		mutex_lock(&table->mutex);
+	}
+
+	err = validate_index(dev, table, index);
+	if (err)
+		goto out;
+
+	table->entries[index] = cpu_to_be64(new_mac | MLX4_MAC_VALID);
+
+	err = mlx4_set_port_mac_table(dev, port, table->entries);
+	if (unlikely(err)) {
+		mlx4_err(dev, "Failed adding MAC: 0x%llx\n",
+			 (unsigned long long) new_mac);
+		table->entries[index] = 0;
+	} else {
+		if (dup) {
+			dup_table->entries[index] = cpu_to_be64(new_mac | MLX4_MAC_VALID);
+
+			err = mlx4_set_port_mac_table(dev, dup_port, dup_table->entries);
+			if (unlikely(err)) {
+				mlx4_err(dev, "Failed adding duplicate MAC: 0x%llx\n",
+					 (unsigned long long)new_mac);
+				dup_table->entries[index] = 0;
+			}
+		}
+	}
+out:
+	if (dup) {
+		if (port == 2) {
+			mutex_unlock(&table->mutex);
+			mutex_unlock(&dup_table->mutex);
+		} else {
+			mutex_unlock(&dup_table->mutex);
+			mutex_unlock(&table->mutex);
+		}
+	} else {
+		mutex_unlock(&table->mutex);
+	}
+	return err;
+}
+EXPORT_SYMBOL_GPL(__mlx4_replace_mac);
+
+static int mlx4_set_port_vlan_table(struct mlx4_dev *dev, u8 port,
+				    __be32 *entries)
+{
+	struct mlx4_cmd_mailbox *mailbox;
+	u32 in_mod;
+	int err;
+
+	mailbox = mlx4_alloc_cmd_mailbox(dev);
+	if (IS_ERR(mailbox))
+		return PTR_ERR(mailbox);
+
+	memcpy(mailbox->buf, entries, MLX4_VLAN_TABLE_SIZE);
+	in_mod = MLX4_SET_PORT_VLAN_TABLE << 8 | port;
+	err = mlx4_cmd(dev, mailbox->dma, in_mod, MLX4_SET_PORT_ETH_OPCODE,
+		       MLX4_CMD_SET_PORT, MLX4_CMD_TIME_CLASS_B,
+		       MLX4_CMD_NATIVE);
+
+	mlx4_free_cmd_mailbox(dev, mailbox);
+
+	return err;
+}
+
+int mlx4_find_cached_vlan(struct mlx4_dev *dev, u8 port, u16 vid, int *idx)
+{
+	struct mlx4_vlan_table *table = &mlx4_priv(dev)->port[port].vlan_table;
+	int i;
+
+	for (i = 0; i < MLX4_MAX_VLAN_NUM; ++i) {
+		if (table->refs[i] &&
+		    (vid == (MLX4_VLAN_MASK &
+			      be32_to_cpu(table->entries[i])))) {
+			/* VLAN already registered, increase reference count */
+			*idx = i;
+			return 0;
+		}
+	}
+
+	return -ENOENT;
+}
+EXPORT_SYMBOL_GPL(mlx4_find_cached_vlan);
+
+int __mlx4_register_vlan(struct mlx4_dev *dev, u8 port, u16 vlan,
+				int *index)
+{
+	struct mlx4_vlan_table *table = &mlx4_priv(dev)->port[port].vlan_table;
+	int i, err = 0;
+	int free = -1;
+	int free_for_dup = -1;
+	bool dup = mlx4_is_mf_bonded(dev);
+	u8 dup_port = (port == 1) ? 2 : 1;
+	struct mlx4_vlan_table *dup_table = &mlx4_priv(dev)->port[dup_port].vlan_table;
+	bool need_mf_bond = mlx4_need_mf_bond(dev);
+	bool can_mf_bond = true;
+
+	mlx4_dbg(dev, "Registering VLAN: %d for port %d %s duplicate\n",
+		 vlan, port,
+		 dup ? "with" : "without");
+
+	if (need_mf_bond) {
+		if (port == 1) {
+			mutex_lock(&table->mutex);
+			mutex_lock_nested(&dup_table->mutex, SINGLE_DEPTH_NESTING);
+		} else {
+			mutex_lock(&dup_table->mutex);
+			mutex_lock_nested(&table->mutex, SINGLE_DEPTH_NESTING);
+		}
+	} else {
+		mutex_lock(&table->mutex);
+	}
+
+	if (table->total == table->max) {
+		/* No free vlan entries */
+		err = -ENOSPC;
+		goto out;
+	}
+
+	if (need_mf_bond) {
+		int index_at_port = -1;
+		int index_at_dup_port = -1;
+
+		for (i = MLX4_VLAN_REGULAR; i < MLX4_MAX_VLAN_NUM; i++) {
+			if (vlan == (MLX4_VLAN_MASK & be32_to_cpu(table->entries[i])))
+				index_at_port = i;
+			if (vlan == (MLX4_VLAN_MASK & be32_to_cpu(dup_table->entries[i])))
+				index_at_dup_port = i;
+		}
+		/* check that same vlan is not in the tables at different indices */
+		if ((index_at_port != index_at_dup_port) &&
+		    (index_at_port >= 0) &&
+		    (index_at_dup_port >= 0))
+			can_mf_bond = false;
+
+		/* If the vlan is already in the primary table, the slot must be
+		 * available in the duplicate table as well.
+		 */
+		if (index_at_port >= 0 && index_at_dup_port < 0 &&
+		    dup_table->refs[index_at_port]) {
+			can_mf_bond = false;
+		}
+		/* If the vlan is already in the duplicate table, check that the
+		 * corresponding index is not occupied in the primary table, or
+		 * the primary table already contains the vlan at the same index.
+		 * Otherwise, you cannot bond (primary contains a different vlan
+		 * at that index).
+		 */
+		if (index_at_dup_port >= 0) {
+			if (!table->refs[index_at_dup_port] ||
+			    (vlan == (MLX4_VLAN_MASK & be32_to_cpu(dup_table->entries[index_at_dup_port]))))
+				free_for_dup = index_at_dup_port;
+			else
+				can_mf_bond = false;
+		}
+	}
+
+	for (i = MLX4_VLAN_REGULAR; i < MLX4_MAX_VLAN_NUM; i++) {
+		if (!table->refs[i]) {
+			if (free < 0)
+				free = i;
+			if (free_for_dup < 0 && need_mf_bond && can_mf_bond) {
+				if (!dup_table->refs[i])
+					free_for_dup = i;
+			}
+		}
+
+		if ((table->refs[i] || table->is_dup[i]) &&
+		    (vlan == (MLX4_VLAN_MASK &
+			      be32_to_cpu(table->entries[i])))) {
+			/* Vlan already registered, increase references count */
+			mlx4_dbg(dev, "vlan %u is already registered.\n", vlan);
+			*index = i;
+			++table->refs[i];
+			if (dup) {
+				u16 dup_vlan = MLX4_VLAN_MASK & be32_to_cpu(dup_table->entries[i]);
+
+				if (dup_vlan != vlan || !dup_table->is_dup[i]) {
+					mlx4_warn(dev, "register vlan: expected duplicate vlan %u on port %d index %d\n",
+						  vlan, dup_port, i);
+				}
+			}
+			goto out;
+		}
+	}
+
+	if (need_mf_bond && (free_for_dup < 0)) {
+		if (dup) {
+			mlx4_warn(dev, "Fail to allocate duplicate VLAN table entry\n");
+			mlx4_warn(dev, "High Availability for virtual functions may not work as expected\n");
+			dup = false;
+		}
+		can_mf_bond = false;
+	}
+
+	if (need_mf_bond && can_mf_bond)
+		free = free_for_dup;
+
+	if (free < 0) {
+		err = -ENOMEM;
+		goto out;
+	}
+
+	/* Register new VLAN */
+	table->refs[free] = 1;
+	table->is_dup[free] = false;
+	table->entries[free] = cpu_to_be32(vlan | MLX4_VLAN_VALID);
+
+	err = mlx4_set_port_vlan_table(dev, port, table->entries);
+	if (unlikely(err)) {
+		mlx4_warn(dev, "Failed adding vlan: %u\n", vlan);
+		table->refs[free] = 0;
+		table->entries[free] = 0;
+		goto out;
+	}
+	++table->total;
+	if (dup) {
+		dup_table->refs[free] = 0;
+		dup_table->is_dup[free] = true;
+		dup_table->entries[free] = cpu_to_be32(vlan | MLX4_VLAN_VALID);
+
+		err = mlx4_set_port_vlan_table(dev, dup_port, dup_table->entries);
+		if (unlikely(err)) {
+			mlx4_warn(dev, "Failed adding duplicate vlan: %u\n", vlan);
+			dup_table->is_dup[free] = false;
+			dup_table->entries[free] = 0;
+			goto out;
+		}
+		++dup_table->total;
+	}
+
+	*index = free;
+out:
+	if (need_mf_bond) {
+		if (port == 2) {
+			mutex_unlock(&table->mutex);
+			mutex_unlock(&dup_table->mutex);
+		} else {
+			mutex_unlock(&dup_table->mutex);
+			mutex_unlock(&table->mutex);
+		}
+	} else {
+		mutex_unlock(&table->mutex);
+	}
+	return err;
+}
+
+int mlx4_register_vlan(struct mlx4_dev *dev, u8 port, u16 vlan, int *index)
+{
+	u64 out_param = 0;
+	int err;
+
+	if (vlan > 4095)
+		return -EINVAL;
+
+	if (mlx4_is_mfunc(dev)) {
+		err = mlx4_cmd_imm(dev, vlan, &out_param,
+				   ((u32) port) << 8 | (u32) RES_VLAN,
+				   RES_OP_RESERVE_AND_MAP, MLX4_CMD_ALLOC_RES,
+				   MLX4_CMD_TIME_CLASS_A, MLX4_CMD_WRAPPED);
+		if (!err)
+			*index = get_param_l(&out_param);
+
+		return err;
+	}
+	return __mlx4_register_vlan(dev, port, vlan, index);
+}
+EXPORT_SYMBOL_GPL(mlx4_register_vlan);
+
+void __mlx4_unregister_vlan(struct mlx4_dev *dev, u8 port, u16 vlan)
+{
+	struct mlx4_vlan_table *table = &mlx4_priv(dev)->port[port].vlan_table;
+	int index;
+	bool dup = mlx4_is_mf_bonded(dev);
+	u8 dup_port = (port == 1) ? 2 : 1;
+	struct mlx4_vlan_table *dup_table = &mlx4_priv(dev)->port[dup_port].vlan_table;
+
+	if (dup) {
+		if (port == 1) {
+			mutex_lock(&table->mutex);
+			mutex_lock_nested(&dup_table->mutex, SINGLE_DEPTH_NESTING);
+		} else {
+			mutex_lock(&dup_table->mutex);
+			mutex_lock_nested(&table->mutex, SINGLE_DEPTH_NESTING);
+		}
+	} else {
+		mutex_lock(&table->mutex);
+	}
+
+	if (mlx4_find_cached_vlan(dev, port, vlan, &index)) {
+		mlx4_warn(dev, "vlan 0x%x is not in the vlan table\n", vlan);
+		goto out;
+	}
+
+	if (index < MLX4_VLAN_REGULAR) {
+		mlx4_warn(dev, "Trying to free special vlan index %d\n", index);
+		goto out;
+	}
+
+	if (--table->refs[index] || table->is_dup[index]) {
+		mlx4_dbg(dev, "Have %d more references for index %d, no need to modify vlan table\n",
+			 table->refs[index], index);
+		if (!table->refs[index])
+			dup_table->is_dup[index] = false;
+		goto out;
+	}
+	table->entries[index] = 0;
+	if (mlx4_set_port_vlan_table(dev, port, table->entries))
+		mlx4_warn(dev, "Fail to set vlan in port %d during unregister\n", port);
+	--table->total;
+	if (dup) {
+		dup_table->is_dup[index] = false;
+		if (dup_table->refs[index])
+			goto out;
+		dup_table->entries[index] = 0;
+		if (mlx4_set_port_vlan_table(dev, dup_port, dup_table->entries))
+			mlx4_warn(dev, "Fail to set vlan in duplicate port %d during unregister\n", dup_port);
+		--dup_table->total;
+	}
+out:
+	if (dup) {
+		if (port == 2) {
+			mutex_unlock(&table->mutex);
+			mutex_unlock(&dup_table->mutex);
+		} else {
+			mutex_unlock(&dup_table->mutex);
+			mutex_unlock(&table->mutex);
+		}
+	} else {
+		mutex_unlock(&table->mutex);
+	}
+}
+
+void mlx4_unregister_vlan(struct mlx4_dev *dev, u8 port, u16 vlan)
+{
+	u64 out_param = 0;
+
+	if (mlx4_is_mfunc(dev)) {
+		(void) mlx4_cmd_imm(dev, vlan, &out_param,
+				    ((u32) port) << 8 | (u32) RES_VLAN,
+				    RES_OP_RESERVE_AND_MAP,
+				    MLX4_CMD_FREE_RES, MLX4_CMD_TIME_CLASS_A,
+				    MLX4_CMD_WRAPPED);
+		return;
+	}
+	__mlx4_unregister_vlan(dev, port, vlan);
+}
+EXPORT_SYMBOL_GPL(mlx4_unregister_vlan);
+
+int mlx4_bond_mac_table(struct mlx4_dev *dev)
+{
+	struct mlx4_mac_table *t1 = &mlx4_priv(dev)->port[1].mac_table;
+	struct mlx4_mac_table *t2 = &mlx4_priv(dev)->port[2].mac_table;
+	int ret = 0;
+	int i;
+	bool update1 = false;
+	bool update2 = false;
+
+	mutex_lock(&t1->mutex);
+	mutex_lock(&t2->mutex);
+	for (i = 0; i < MLX4_MAX_MAC_NUM; i++) {
+		if ((t1->entries[i] != t2->entries[i]) &&
+		    t1->entries[i] && t2->entries[i]) {
+			mlx4_warn(dev, "can't duplicate entry %d in mac table\n", i);
+			ret = -EINVAL;
+			goto unlock;
+		}
+	}
+
+	for (i = 0; i < MLX4_MAX_MAC_NUM; i++) {
+		if (t1->entries[i] && !t2->entries[i]) {
+			t2->entries[i] = t1->entries[i];
+			t2->is_dup[i] = true;
+			update2 = true;
+		} else if (!t1->entries[i] && t2->entries[i]) {
+			t1->entries[i] = t2->entries[i];
+			t1->is_dup[i] = true;
+			update1 = true;
+		} else if (t1->entries[i] && t2->entries[i]) {
+			t1->is_dup[i] = true;
+			t2->is_dup[i] = true;
+		}
+	}
+
+	if (update1) {
+		ret = mlx4_set_port_mac_table(dev, 1, t1->entries);
+		if (ret)
+			mlx4_warn(dev, "failed to set MAC table for port 1 (%d)\n", ret);
+	}
+	if (!ret && update2) {
+		ret = mlx4_set_port_mac_table(dev, 2, t2->entries);
+		if (ret)
+			mlx4_warn(dev, "failed to set MAC table for port 2 (%d)\n", ret);
+	}
+
+	if (ret)
+		mlx4_warn(dev, "failed to create mirror MAC tables\n");
+unlock:
+	mutex_unlock(&t2->mutex);
+	mutex_unlock(&t1->mutex);
+	return ret;
+}
+
+int mlx4_unbond_mac_table(struct mlx4_dev *dev)
+{
+	struct mlx4_mac_table *t1 = &mlx4_priv(dev)->port[1].mac_table;
+	struct mlx4_mac_table *t2 = &mlx4_priv(dev)->port[2].mac_table;
+	int ret = 0;
+	int ret1;
+	int i;
+	bool update1 = false;
+	bool update2 = false;
+
+	mutex_lock(&t1->mutex);
+	mutex_lock(&t2->mutex);
+	for (i = 0; i < MLX4_MAX_MAC_NUM; i++) {
+		if (t1->entries[i] != t2->entries[i]) {
+			mlx4_warn(dev, "mac table is in an unexpected state when trying to unbond\n");
+			ret = -EINVAL;
+			goto unlock;
+		}
+	}
+
+	for (i = 0; i < MLX4_MAX_MAC_NUM; i++) {
+		if (!t1->entries[i])
+			continue;
+		t1->is_dup[i] = false;
+		if (!t1->refs[i]) {
+			t1->entries[i] = 0;
+			update1 = true;
+		}
+		t2->is_dup[i] = false;
+		if (!t2->refs[i]) {
+			t2->entries[i] = 0;
+			update2 = true;
+		}
+	}
+
+	if (update1) {
+		ret = mlx4_set_port_mac_table(dev, 1, t1->entries);
+		if (ret)
+			mlx4_warn(dev, "failed to unmirror MAC tables for port 1(%d)\n", ret);
+	}
+	if (update2) {
+		ret1 = mlx4_set_port_mac_table(dev, 2, t2->entries);
+		if (ret1) {
+			mlx4_warn(dev, "failed to unmirror MAC tables for port 2(%d)\n", ret1);
+			ret = ret1;
+		}
+	}
+unlock:
+	mutex_unlock(&t2->mutex);
+	mutex_unlock(&t1->mutex);
+	return ret;
+}
+
+int mlx4_bond_vlan_table(struct mlx4_dev *dev)
+{
+	struct mlx4_vlan_table *t1 = &mlx4_priv(dev)->port[1].vlan_table;
+	struct mlx4_vlan_table *t2 = &mlx4_priv(dev)->port[2].vlan_table;
+	int ret = 0;
+	int i;
+	bool update1 = false;
+	bool update2 = false;
+
+	mutex_lock(&t1->mutex);
+	mutex_lock(&t2->mutex);
+	for (i = 0; i < MLX4_MAX_VLAN_NUM; i++) {
+		if ((t1->entries[i] != t2->entries[i]) &&
+		    t1->entries[i] && t2->entries[i]) {
+			mlx4_warn(dev, "can't duplicate entry %d in vlan table\n", i);
+			ret = -EINVAL;
+			goto unlock;
+		}
+	}
+
+	for (i = 0; i < MLX4_MAX_VLAN_NUM; i++) {
+		if (t1->entries[i] && !t2->entries[i]) {
+			t2->entries[i] = t1->entries[i];
+			t2->is_dup[i] = true;
+			update2 = true;
+		} else if (!t1->entries[i] && t2->entries[i]) {
+			t1->entries[i] = t2->entries[i];
+			t1->is_dup[i] = true;
+			update1 = true;
+		} else if (t1->entries[i] && t2->entries[i]) {
+			t1->is_dup[i] = true;
+			t2->is_dup[i] = true;
+		}
+	}
+
+	if (update1) {
+		ret = mlx4_set_port_vlan_table(dev, 1, t1->entries);
+		if (ret)
+			mlx4_warn(dev, "failed to set VLAN table for port 1 (%d)\n", ret);
+	}
+	if (!ret && update2) {
+		ret = mlx4_set_port_vlan_table(dev, 2, t2->entries);
+		if (ret)
+			mlx4_warn(dev, "failed to set VLAN table for port 2 (%d)\n", ret);
+	}
+
+	if (ret)
+		mlx4_warn(dev, "failed to create mirror VLAN tables\n");
+unlock:
+	mutex_unlock(&t2->mutex);
+	mutex_unlock(&t1->mutex);
+	return ret;
+}
+
+int mlx4_unbond_vlan_table(struct mlx4_dev *dev)
+{
+	struct mlx4_vlan_table *t1 = &mlx4_priv(dev)->port[1].vlan_table;
+	struct mlx4_vlan_table *t2 = &mlx4_priv(dev)->port[2].vlan_table;
+	int ret = 0;
+	int ret1;
+	int i;
+	bool update1 = false;
+	bool update2 = false;
+
+	mutex_lock(&t1->mutex);
+	mutex_lock(&t2->mutex);
+	for (i = 0; i < MLX4_MAX_VLAN_NUM; i++) {
+		if (t1->entries[i] != t2->entries[i]) {
+			mlx4_warn(dev, "vlan table is in an unexpected state when trying to unbond\n");
+			ret = -EINVAL;
+			goto unlock;
+		}
+	}
+
+	for (i = 0; i < MLX4_MAX_VLAN_NUM; i++) {
+		if (!t1->entries[i])
+			continue;
+		t1->is_dup[i] = false;
+		if (!t1->refs[i]) {
+			t1->entries[i] = 0;
+			update1 = true;
+		}
+		t2->is_dup[i] = false;
+		if (!t2->refs[i]) {
+			t2->entries[i] = 0;
+			update2 = true;
+		}
+	}
+
+	if (update1) {
+		ret = mlx4_set_port_vlan_table(dev, 1, t1->entries);
+		if (ret)
+			mlx4_warn(dev, "failed to unmirror VLAN tables for port 1(%d)\n", ret);
+	}
+	if (update2) {
+		ret1 = mlx4_set_port_vlan_table(dev, 2, t2->entries);
+		if (ret1) {
+			mlx4_warn(dev, "failed to unmirror VLAN tables for port 2(%d)\n", ret1);
+			ret = ret1;
+		}
+	}
+unlock:
+	mutex_unlock(&t2->mutex);
+	mutex_unlock(&t1->mutex);
+	return ret;
+}
+
+int mlx4_get_port_ib_caps(struct mlx4_dev *dev, u8 port, __be32 *caps)
+{
+	struct mlx4_cmd_mailbox *inmailbox, *outmailbox;
+	u8 *inbuf, *outbuf;
+	int err;
+
+	inmailbox = mlx4_alloc_cmd_mailbox(dev);
+	if (IS_ERR(inmailbox))
+		return PTR_ERR(inmailbox);
+
+	outmailbox = mlx4_alloc_cmd_mailbox(dev);
+	if (IS_ERR(outmailbox)) {
+		mlx4_free_cmd_mailbox(dev, inmailbox);
+		return PTR_ERR(outmailbox);
+	}
+
+	inbuf = inmailbox->buf;
+	outbuf = outmailbox->buf;
+	inbuf[0] = 1;
+	inbuf[1] = 1;
+	inbuf[2] = 1;
+	inbuf[3] = 1;
+	*(__be16 *) (&inbuf[16]) = cpu_to_be16(0x0015);
+	*(__be32 *) (&inbuf[20]) = cpu_to_be32(port);
+
+	err = mlx4_cmd_box(dev, inmailbox->dma, outmailbox->dma, port, 3,
+			   MLX4_CMD_MAD_IFC, MLX4_CMD_TIME_CLASS_C,
+			   MLX4_CMD_NATIVE);
+	if (!err)
+		*caps = *(__be32 *) (outbuf + 84);
+	mlx4_free_cmd_mailbox(dev, inmailbox);
+	mlx4_free_cmd_mailbox(dev, outmailbox);
+	return err;
+}
+static struct mlx4_roce_gid_entry zgid_entry;
+
+int mlx4_get_slave_num_gids(struct mlx4_dev *dev, int slave, int port)
+{
+	int vfs;
+	int slave_gid = slave;
+	unsigned i;
+	struct mlx4_slaves_pport slaves_pport;
+	struct mlx4_active_ports actv_ports;
+	unsigned max_port_p_one;
+
+	if (slave == 0)
+		return MLX4_ROCE_PF_GIDS;
+
+	/* Slave is a VF */
+	slaves_pport = mlx4_phys_to_slaves_pport(dev, port);
+	actv_ports = mlx4_get_active_ports(dev, slave);
+	max_port_p_one = find_first_bit(actv_ports.ports, dev->caps.num_ports) +
+		bitmap_weight(actv_ports.ports, dev->caps.num_ports) + 1;
+
+	for (i = 1; i < max_port_p_one; i++) {
+		struct mlx4_active_ports exclusive_ports;
+		struct mlx4_slaves_pport slaves_pport_actv;
+		bitmap_zero(exclusive_ports.ports, dev->caps.num_ports);
+		set_bit(i - 1, exclusive_ports.ports);
+		if (i == port)
+			continue;
+		slaves_pport_actv = mlx4_phys_to_slaves_pport_actv(
+				    dev, &exclusive_ports);
+		slave_gid -= bitmap_weight(slaves_pport_actv.slaves,
+					   dev->persist->num_vfs + 1);
+	}
+	vfs = bitmap_weight(slaves_pport.slaves, dev->persist->num_vfs + 1) - 1;
+	if (slave_gid <= ((MLX4_ROCE_MAX_GIDS - MLX4_ROCE_PF_GIDS) % vfs))
+		return ((MLX4_ROCE_MAX_GIDS - MLX4_ROCE_PF_GIDS) / vfs) + 1;
+	return (MLX4_ROCE_MAX_GIDS - MLX4_ROCE_PF_GIDS) / vfs;
+}
+
+int mlx4_get_base_gid_ix(struct mlx4_dev *dev, int slave, int port)
+{
+	int gids;
+	unsigned i;
+	int slave_gid = slave;
+	int vfs;
+
+	struct mlx4_slaves_pport slaves_pport;
+	struct mlx4_active_ports actv_ports;
+	unsigned max_port_p_one;
+
+	if (slave == 0)
+		return 0;
+
+	slaves_pport = mlx4_phys_to_slaves_pport(dev, port);
+	actv_ports = mlx4_get_active_ports(dev, slave);
+	max_port_p_one = find_first_bit(actv_ports.ports, dev->caps.num_ports) +
+		bitmap_weight(actv_ports.ports, dev->caps.num_ports) + 1;
+
+	for (i = 1; i < max_port_p_one; i++) {
+		struct mlx4_active_ports exclusive_ports;
+		struct mlx4_slaves_pport slaves_pport_actv;
+		bitmap_zero(exclusive_ports.ports, dev->caps.num_ports);
+		set_bit(i - 1, exclusive_ports.ports);
+		if (i == port)
+			continue;
+		slaves_pport_actv = mlx4_phys_to_slaves_pport_actv(
+				    dev, &exclusive_ports);
+		slave_gid -= bitmap_weight(slaves_pport_actv.slaves,
+					   dev->persist->num_vfs + 1);
+	}
+	gids = MLX4_ROCE_MAX_GIDS - MLX4_ROCE_PF_GIDS;
+	vfs = bitmap_weight(slaves_pport.slaves, dev->persist->num_vfs + 1) - 1;
+	if (slave_gid <= gids % vfs)
+		return MLX4_ROCE_PF_GIDS + ((gids / vfs) + 1) * (slave_gid - 1);
+
+	return MLX4_ROCE_PF_GIDS + (gids % vfs) +
+		((gids / vfs) * (slave_gid - 1));
+}
+EXPORT_SYMBOL_GPL(mlx4_get_base_gid_ix);
+
+static int mlx4_reset_roce_port_gids(struct mlx4_dev *dev, int slave,
+				     int port, struct mlx4_cmd_mailbox *mailbox)
+{
+	struct mlx4_roce_gid_entry *gid_entry_mbox;
+	struct mlx4_priv *priv = mlx4_priv(dev);
+	int num_gids, base, offset;
+	int i, err;
+
+	num_gids = mlx4_get_slave_num_gids(dev, slave, port);
+	base = mlx4_get_base_gid_ix(dev, slave, port);
+
+	memset(mailbox->buf, 0, MLX4_MAILBOX_SIZE);
+
+	mutex_lock(&(priv->port[port].gid_table.mutex));
+	/* Zero-out gids belonging to that slave in the port GID table */
+	for (i = 0, offset = base; i < num_gids; offset++, i++)
+		memcpy(priv->port[port].gid_table.roce_gids[offset].raw,
+		       zgid_entry.raw, MLX4_ROCE_GID_ENTRY_SIZE);
+
+	/* Now, copy roce port gids table to mailbox for passing to FW */
+	gid_entry_mbox = (struct mlx4_roce_gid_entry *)mailbox->buf;
+	for (i = 0; i < MLX4_ROCE_MAX_GIDS; gid_entry_mbox++, i++)
+		memcpy(gid_entry_mbox->raw,
+		       priv->port[port].gid_table.roce_gids[i].raw,
+		       MLX4_ROCE_GID_ENTRY_SIZE);
+
+	err = mlx4_cmd(dev, mailbox->dma,
+		       ((u32)port) | (MLX4_SET_PORT_GID_TABLE << 8),
+		       MLX4_SET_PORT_ETH_OPCODE, MLX4_CMD_SET_PORT,
+		       MLX4_CMD_TIME_CLASS_B, MLX4_CMD_NATIVE);
+	mutex_unlock(&(priv->port[port].gid_table.mutex));
+	return err;
+}
+
+
+void mlx4_reset_roce_gids(struct mlx4_dev *dev, int slave)
+{
+	struct mlx4_active_ports actv_ports;
+	struct mlx4_cmd_mailbox *mailbox;
+	int num_eth_ports, err;
+	int i;
+
+	if (slave < 0 || slave > dev->persist->num_vfs)
+		return;
+
+	actv_ports = mlx4_get_active_ports(dev, slave);
+
+	for (i = 0, num_eth_ports = 0; i < dev->caps.num_ports; i++) {
+		if (test_bit(i, actv_ports.ports)) {
+			if (dev->caps.port_type[i + 1] != MLX4_PORT_TYPE_ETH)
+				continue;
+			num_eth_ports++;
+		}
+	}
+
+	if (!num_eth_ports)
+		return;
+
+	/* have ETH ports.  Alloc mailbox for SET_PORT command */
+	mailbox = mlx4_alloc_cmd_mailbox(dev);
+	if (IS_ERR(mailbox))
+		return;
+
+	for (i = 0; i < dev->caps.num_ports; i++) {
+		if (test_bit(i, actv_ports.ports)) {
+			if (dev->caps.port_type[i + 1] != MLX4_PORT_TYPE_ETH)
+				continue;
+			err = mlx4_reset_roce_port_gids(dev, slave, i + 1, mailbox);
+			if (err)
+				mlx4_warn(dev, "Could not reset ETH port GID table for slave %d, port %d (%d)\n",
+					  slave, i + 1, err);
+		}
+	}
+
+	mlx4_free_cmd_mailbox(dev, mailbox);
+	return;
+}
+
+static void
+mlx4_en_set_port_mtu(struct mlx4_dev *dev, int slave, int port,
+		     struct mlx4_set_port_general_context *gen_context)
+{
+	struct mlx4_priv *priv = mlx4_priv(dev);
+	struct mlx4_mfunc_master_ctx *master = &priv->mfunc.master;
+	struct mlx4_slave_state *slave_st = &master->slave_state[slave];
+	u16 mtu, prev_mtu;
+
+	/* Mtu is configured as the max USER_MTU among all
+	 * the functions on the port.
+	 */
+	mtu = be16_to_cpu(gen_context->mtu);
+	mtu = min_t(int, mtu, dev->caps.eth_mtu_cap[port] +
+		    ETH_HLEN + VLAN_HLEN + ETH_FCS_LEN);
+	prev_mtu = slave_st->mtu[port];
+	slave_st->mtu[port] = mtu;
+	if (mtu > master->max_mtu[port])
+		master->max_mtu[port] = mtu;
+	if (mtu < prev_mtu && prev_mtu == master->max_mtu[port]) {
+		int i;
+
+		slave_st->mtu[port] = mtu;
+		master->max_mtu[port] = mtu;
+		for (i = 0; i < dev->num_slaves; i++)
+			master->max_mtu[port] =
+				max_t(u16, master->max_mtu[port],
+				      master->slave_state[i].mtu[port]);
+	}
+	gen_context->mtu = cpu_to_be16(master->max_mtu[port]);
+}
+
+static void
+mlx4_en_set_port_user_mtu(struct mlx4_dev *dev, int slave, int port,
+			  struct mlx4_set_port_general_context *gen_context)
+{
+	struct mlx4_priv *priv = mlx4_priv(dev);
+	struct mlx4_mfunc_master_ctx *master = &priv->mfunc.master;
+	struct mlx4_slave_state *slave_st = &master->slave_state[slave];
+	u16 user_mtu, prev_user_mtu;
+
+	/* User Mtu is configured as the max USER_MTU among all
+	 * the functions on the port.
+	 */
+	user_mtu = be16_to_cpu(gen_context->user_mtu);
+	user_mtu = min_t(int, user_mtu, dev->caps.eth_mtu_cap[port]);
+	prev_user_mtu = slave_st->user_mtu[port];
+	slave_st->user_mtu[port] = user_mtu;
+	if (user_mtu > master->max_user_mtu[port])
+		master->max_user_mtu[port] = user_mtu;
+	if (user_mtu < prev_user_mtu &&
+	    prev_user_mtu == master->max_user_mtu[port]) {
+		int i;
+
+		slave_st->user_mtu[port] = user_mtu;
+		master->max_user_mtu[port] = user_mtu;
+		for (i = 0; i < dev->num_slaves; i++)
+			master->max_user_mtu[port] =
+				max_t(u16, master->max_user_mtu[port],
+				      master->slave_state[i].user_mtu[port]);
+	}
+	gen_context->user_mtu = cpu_to_be16(master->max_user_mtu[port]);
+}
+
+static void
+mlx4_en_set_port_global_pause(struct mlx4_dev *dev, int slave,
+			      struct mlx4_set_port_general_context *gen_context)
+{
+	struct mlx4_priv *priv = mlx4_priv(dev);
+	struct mlx4_mfunc_master_ctx *master = &priv->mfunc.master;
+
+	/* Slave cannot change Global Pause configuration */
+	if (slave != mlx4_master_func_num(dev) &&
+	    (gen_context->pptx != master->pptx ||
+	     gen_context->pprx != master->pprx)) {
+		gen_context->pptx = master->pptx;
+		gen_context->pprx = master->pprx;
+		mlx4_warn(dev, "denying Global Pause change for slave:%d\n",
+			  slave);
+	} else {
+		master->pptx = gen_context->pptx;
+		master->pprx = gen_context->pprx;
+	}
+}
+
+static int mlx4_common_set_port(struct mlx4_dev *dev, int slave, u32 in_mod,
+				u8 op_mod, struct mlx4_cmd_mailbox *inbox)
+{
+	struct mlx4_priv *priv = mlx4_priv(dev);
+	struct mlx4_port_info *port_info;
+	struct mlx4_set_port_rqp_calc_context *qpn_context;
+	struct mlx4_set_port_general_context *gen_context;
+	struct mlx4_roce_gid_entry *gid_entry_tbl, *gid_entry_mbox, *gid_entry_mb1;
+	int reset_qkey_viols;
+	int port;
+	int is_eth;
+	int num_gids;
+	int base;
+	u32 in_modifier;
+	u32 promisc;
+	int err;
+	int i, j;
+	int offset;
+	__be32 agg_cap_mask;
+	__be32 slave_cap_mask;
+	__be32 new_cap_mask;
+
+	port = in_mod & 0xff;
+	in_modifier = in_mod >> 8;
+	is_eth = op_mod;
+	port_info = &priv->port[port];
+
+	/* Slaves cannot perform SET_PORT operations,
+	 * except for changing MTU and USER_MTU.
+	 */
+	if (is_eth) {
+		if (slave != dev->caps.function &&
+		    in_modifier != MLX4_SET_PORT_GENERAL &&
+		    in_modifier != MLX4_SET_PORT_GID_TABLE) {
+			mlx4_warn(dev, "denying SET_PORT for slave:%d\n",
+					slave);
+			return -EINVAL;
+		}
+		switch (in_modifier) {
+		case MLX4_SET_PORT_RQP_CALC:
+			qpn_context = inbox->buf;
+			qpn_context->base_qpn =
+				cpu_to_be32(port_info->base_qpn);
+			qpn_context->n_mac = 0x7;
+			promisc = be32_to_cpu(qpn_context->promisc) >>
+				SET_PORT_PROMISC_SHIFT;
+			qpn_context->promisc = cpu_to_be32(
+				promisc << SET_PORT_PROMISC_SHIFT |
+				port_info->base_qpn);
+			promisc = be32_to_cpu(qpn_context->mcast) >>
+				SET_PORT_MC_PROMISC_SHIFT;
+			qpn_context->mcast = cpu_to_be32(
+				promisc << SET_PORT_MC_PROMISC_SHIFT |
+				port_info->base_qpn);
+			break;
+		case MLX4_SET_PORT_GENERAL:
+			gen_context = inbox->buf;
+
+			if (gen_context->flags & MLX4_FLAG_V_MTU_MASK)
+				mlx4_en_set_port_mtu(dev, slave, port,
+						     gen_context);
+
+			if (gen_context->flags2 & MLX4_FLAG2_V_USER_MTU_MASK)
+				mlx4_en_set_port_user_mtu(dev, slave, port,
+							  gen_context);
+
+			if (gen_context->flags &
+			    (MLX4_FLAG_V_PPRX_MASK | MLX4_FLAG_V_PPTX_MASK))
+				mlx4_en_set_port_global_pause(dev, slave,
+							      gen_context);
+
+			break;
+		case MLX4_SET_PORT_GID_TABLE:
+			/* change to MULTIPLE entries: number of guest's gids
+			 * need a FOR-loop here over number of gids the guest has.
+			 * 1. Check no duplicates in gids passed by slave
+			 */
+			num_gids = mlx4_get_slave_num_gids(dev, slave, port);
+			base = mlx4_get_base_gid_ix(dev, slave, port);
+			gid_entry_mbox = (struct mlx4_roce_gid_entry *)(inbox->buf);
+			for (i = 0; i < num_gids; gid_entry_mbox++, i++) {
+				if (!memcmp(gid_entry_mbox->raw, zgid_entry.raw,
+					    sizeof(zgid_entry)))
+					continue;
+				gid_entry_mb1 = gid_entry_mbox + 1;
+				for (j = i + 1; j < num_gids; gid_entry_mb1++, j++) {
+					if (!memcmp(gid_entry_mb1->raw,
+						    zgid_entry.raw, sizeof(zgid_entry)))
+						continue;
+					if (!memcmp(gid_entry_mb1->raw, gid_entry_mbox->raw,
+						    sizeof(gid_entry_mbox->raw))) {
+						/* found duplicate */
+						return -EINVAL;
+					}
+				}
+			}
+
+			/* 2. Check that do not have duplicates in OTHER
+			 *    entries in the port GID table
+			 */
+
+			mutex_lock(&(priv->port[port].gid_table.mutex));
+			for (i = 0; i < MLX4_ROCE_MAX_GIDS; i++) {
+				if (i >= base && i < base + num_gids)
+					continue; /* don't compare to slave's current gids */
+				gid_entry_tbl = &priv->port[port].gid_table.roce_gids[i];
+				if (!memcmp(gid_entry_tbl->raw, zgid_entry.raw, sizeof(zgid_entry)))
+					continue;
+				gid_entry_mbox = (struct mlx4_roce_gid_entry *)(inbox->buf);
+				for (j = 0; j < num_gids; gid_entry_mbox++, j++) {
+					if (!memcmp(gid_entry_mbox->raw, zgid_entry.raw,
+						    sizeof(zgid_entry)))
+						continue;
+					if (!memcmp(gid_entry_mbox->raw, gid_entry_tbl->raw,
+						    sizeof(gid_entry_tbl->raw))) {
+						/* found duplicate */
+						mlx4_warn(dev, "requested gid entry for slave:%d is a duplicate of gid at index %d\n",
+							  slave, i);
+						mutex_unlock(&(priv->port[port].gid_table.mutex));
+						return -EINVAL;
+					}
+				}
+			}
+
+			/* insert slave GIDs with memcpy, starting at slave's base index */
+			gid_entry_mbox = (struct mlx4_roce_gid_entry *)(inbox->buf);
+			for (i = 0, offset = base; i < num_gids; gid_entry_mbox++, offset++, i++)
+				memcpy(priv->port[port].gid_table.roce_gids[offset].raw,
+				       gid_entry_mbox->raw, MLX4_ROCE_GID_ENTRY_SIZE);
+
+			/* Now, copy roce port gids table to current mailbox for passing to FW */
+			gid_entry_mbox = (struct mlx4_roce_gid_entry *)(inbox->buf);
+			for (i = 0; i < MLX4_ROCE_MAX_GIDS; gid_entry_mbox++, i++)
+				memcpy(gid_entry_mbox->raw,
+				       priv->port[port].gid_table.roce_gids[i].raw,
+				       MLX4_ROCE_GID_ENTRY_SIZE);
+
+			err = mlx4_cmd(dev, inbox->dma, in_mod & 0xffff, op_mod,
+				       MLX4_CMD_SET_PORT, MLX4_CMD_TIME_CLASS_B,
+				       MLX4_CMD_NATIVE);
+			mutex_unlock(&(priv->port[port].gid_table.mutex));
+			return err;
+		}
+
+		return mlx4_cmd(dev, inbox->dma, in_mod & 0xffff, op_mod,
+				MLX4_CMD_SET_PORT, MLX4_CMD_TIME_CLASS_B,
+				MLX4_CMD_NATIVE);
+	}
+
+	/* Slaves are not allowed to SET_PORT beacon (LED) blink */
+	if (op_mod == MLX4_SET_PORT_BEACON_OPCODE) {
+		mlx4_warn(dev, "denying SET_PORT Beacon slave:%d\n", slave);
+		return -EPERM;
+	}
+
+	/* For IB, we only consider:
+	 * - The capability mask, which is set to the aggregate of all
+	 *   slave function capabilities
+	 * - The QKey violatin counter - reset according to each request.
+	 */
+
+	if (dev->flags & MLX4_FLAG_OLD_PORT_CMDS) {
+		reset_qkey_viols = (*(u8 *) inbox->buf) & 0x40;
+		new_cap_mask = ((__be32 *) inbox->buf)[2];
+	} else {
+		reset_qkey_viols = ((u8 *) inbox->buf)[3] & 0x1;
+		new_cap_mask = ((__be32 *) inbox->buf)[1];
+	}
+
+	/* slave may not set the IS_SM capability for the port */
+	if (slave != mlx4_master_func_num(dev) &&
+	    (be32_to_cpu(new_cap_mask) & MLX4_PORT_CAP_IS_SM))
+		return -EINVAL;
+
+	/* No DEV_MGMT in multifunc mode */
+	if (mlx4_is_mfunc(dev) &&
+	    (be32_to_cpu(new_cap_mask) & MLX4_PORT_CAP_DEV_MGMT_SUP))
+		return -EINVAL;
+
+	agg_cap_mask = 0;
+	slave_cap_mask =
+		priv->mfunc.master.slave_state[slave].ib_cap_mask[port];
+	priv->mfunc.master.slave_state[slave].ib_cap_mask[port] = new_cap_mask;
+	for (i = 0; i < dev->num_slaves; i++)
+		agg_cap_mask |=
+			priv->mfunc.master.slave_state[i].ib_cap_mask[port];
+
+	/* only clear mailbox for guests.  Master may be setting
+	* MTU or PKEY table size
+	*/
+	if (slave != dev->caps.function)
+		memset(inbox->buf, 0, 256);
+	if (dev->flags & MLX4_FLAG_OLD_PORT_CMDS) {
+		*(u8 *) inbox->buf	   |= !!reset_qkey_viols << 6;
+		((__be32 *) inbox->buf)[2] = agg_cap_mask;
+	} else {
+		((u8 *) inbox->buf)[3]     |= !!reset_qkey_viols;
+		((__be32 *) inbox->buf)[1] = agg_cap_mask;
+	}
+
+	err = mlx4_cmd(dev, inbox->dma, port, is_eth, MLX4_CMD_SET_PORT,
+		       MLX4_CMD_TIME_CLASS_B, MLX4_CMD_NATIVE);
+	if (err)
+		priv->mfunc.master.slave_state[slave].ib_cap_mask[port] =
+			slave_cap_mask;
+	return err;
+}
+
+int mlx4_SET_PORT_wrapper(struct mlx4_dev *dev, int slave,
+			  struct mlx4_vhcr *vhcr,
+			  struct mlx4_cmd_mailbox *inbox,
+			  struct mlx4_cmd_mailbox *outbox,
+			  struct mlx4_cmd_info *cmd)
+{
+	int port = mlx4_slave_convert_port(
+			dev, slave, vhcr->in_modifier & 0xFF);
+
+	if (port < 0)
+		return -EINVAL;
+
+	vhcr->in_modifier = (vhcr->in_modifier & ~0xFF) |
+			    (port & 0xFF);
+
+	return mlx4_common_set_port(dev, slave, vhcr->in_modifier,
+				    vhcr->op_modifier, inbox);
+}
+
+/* bit locations for set port command with zero op modifier */
+enum {
+	MLX4_SET_PORT_VL_CAP	 = 4, /* bits 7:4 */
+	MLX4_SET_PORT_MTU_CAP	 = 12, /* bits 15:12 */
+	MLX4_CHANGE_PORT_PKEY_TBL_SZ = 20,
+	MLX4_CHANGE_PORT_VL_CAP	 = 21,
+	MLX4_CHANGE_PORT_MTU_CAP = 22,
+};
+
+int mlx4_SET_PORT(struct mlx4_dev *dev, u8 port, int pkey_tbl_sz)
+{
+	struct mlx4_cmd_mailbox *mailbox;
+	int err, vl_cap, pkey_tbl_flag = 0;
+
+	if (dev->caps.port_type[port] == MLX4_PORT_TYPE_ETH)
+		return 0;
+
+	mailbox = mlx4_alloc_cmd_mailbox(dev);
+	if (IS_ERR(mailbox))
+		return PTR_ERR(mailbox);
+
+	((__be32 *) mailbox->buf)[1] = dev->caps.ib_port_def_cap[port];
+
+	if (pkey_tbl_sz >= 0 && mlx4_is_master(dev)) {
+		pkey_tbl_flag = 1;
+		((__be16 *) mailbox->buf)[20] = cpu_to_be16(pkey_tbl_sz);
+	}
+
+	/* IB VL CAP enum isn't used by the firmware, just numerical values */
+	for (vl_cap = 8; vl_cap >= 1; vl_cap >>= 1) {
+		((__be32 *) mailbox->buf)[0] = cpu_to_be32(
+			(1 << MLX4_CHANGE_PORT_MTU_CAP) |
+			(1 << MLX4_CHANGE_PORT_VL_CAP)  |
+			(pkey_tbl_flag << MLX4_CHANGE_PORT_PKEY_TBL_SZ) |
+			(dev->caps.port_ib_mtu[port] << MLX4_SET_PORT_MTU_CAP) |
+			(vl_cap << MLX4_SET_PORT_VL_CAP));
+		err = mlx4_cmd(dev, mailbox->dma, port,
+			       MLX4_SET_PORT_IB_OPCODE, MLX4_CMD_SET_PORT,
+			       MLX4_CMD_TIME_CLASS_B, MLX4_CMD_WRAPPED);
+		if (err != -ENOMEM)
+			break;
+	}
+
+	mlx4_free_cmd_mailbox(dev, mailbox);
+	return err;
+}
+
+#define SET_PORT_ROCE_2_FLAGS          0x10
+#define MLX4_SET_PORT_ROCE_V1_V2       0x2
+int mlx4_SET_PORT_general(struct mlx4_dev *dev, u8 port, int mtu,
+			  u8 pptx, u8 pfctx, u8 pprx, u8 pfcrx)
+{
+	struct mlx4_cmd_mailbox *mailbox;
+	struct mlx4_set_port_general_context *context;
+	int err;
+	u32 in_mod;
+
+	mailbox = mlx4_alloc_cmd_mailbox(dev);
+	if (IS_ERR(mailbox))
+		return PTR_ERR(mailbox);
+	context = mailbox->buf;
+	context->flags = SET_PORT_GEN_ALL_VALID;
+	context->mtu = cpu_to_be16(mtu);
+	context->pptx = (pptx * (!pfctx)) << 7;
+	context->pfctx = pfctx;
+	context->pprx = (pprx * (!pfcrx)) << 7;
+	context->pfcrx = pfcrx;
+
+	if (dev->caps.flags2 & MLX4_DEV_CAP_FLAG2_ROCE_V1_V2) {
+		context->flags |= SET_PORT_ROCE_2_FLAGS;
+		context->roce_mode |=
+			MLX4_SET_PORT_ROCE_V1_V2 << 4;
+	}
+	in_mod = MLX4_SET_PORT_GENERAL << 8 | port;
+	err = mlx4_cmd(dev, mailbox->dma, in_mod, MLX4_SET_PORT_ETH_OPCODE,
+		       MLX4_CMD_SET_PORT, MLX4_CMD_TIME_CLASS_B,
+		       MLX4_CMD_WRAPPED);
+
+	mlx4_free_cmd_mailbox(dev, mailbox);
+	return err;
+}
+EXPORT_SYMBOL(mlx4_SET_PORT_general);
+
+int mlx4_SET_PORT_qpn_calc(struct mlx4_dev *dev, u8 port, u32 base_qpn,
+			   u8 promisc)
+{
+	struct mlx4_cmd_mailbox *mailbox;
+	struct mlx4_set_port_rqp_calc_context *context;
+	int err;
+	u32 in_mod;
+	u32 m_promisc = (dev->caps.flags & MLX4_DEV_CAP_FLAG_VEP_MC_STEER) ?
+		MCAST_DIRECT : MCAST_DEFAULT;
+
+	if (dev->caps.steering_mode != MLX4_STEERING_MODE_A0)
+		return 0;
+
+	mailbox = mlx4_alloc_cmd_mailbox(dev);
+	if (IS_ERR(mailbox))
+		return PTR_ERR(mailbox);
+	context = mailbox->buf;
+	context->base_qpn = cpu_to_be32(base_qpn);
+	context->n_mac = dev->caps.log_num_macs;
+	context->promisc = cpu_to_be32(promisc << SET_PORT_PROMISC_SHIFT |
+				       base_qpn);
+	context->mcast = cpu_to_be32(m_promisc << SET_PORT_MC_PROMISC_SHIFT |
+				     base_qpn);
+	context->intra_no_vlan = 0;
+	context->no_vlan = MLX4_NO_VLAN_IDX;
+	context->intra_vlan_miss = 0;
+	context->vlan_miss = MLX4_VLAN_MISS_IDX;
+
+	in_mod = MLX4_SET_PORT_RQP_CALC << 8 | port;
+	err = mlx4_cmd(dev, mailbox->dma, in_mod, MLX4_SET_PORT_ETH_OPCODE,
+		       MLX4_CMD_SET_PORT, MLX4_CMD_TIME_CLASS_B,
+		       MLX4_CMD_WRAPPED);
+
+	mlx4_free_cmd_mailbox(dev, mailbox);
+	return err;
+}
+EXPORT_SYMBOL(mlx4_SET_PORT_qpn_calc);
+
+int mlx4_SET_PORT_user_mtu(struct mlx4_dev *dev, u8 port, u16 user_mtu)
+{
+	struct mlx4_cmd_mailbox *mailbox;
+	struct mlx4_set_port_general_context *context;
+	u32 in_mod;
+	int err;
+
+	mailbox = mlx4_alloc_cmd_mailbox(dev);
+	if (IS_ERR(mailbox))
+		return PTR_ERR(mailbox);
+	context = mailbox->buf;
+	context->flags2 |= MLX4_FLAG2_V_USER_MTU_MASK;
+	context->user_mtu = cpu_to_be16(user_mtu);
+
+	in_mod = MLX4_SET_PORT_GENERAL << 8 | port;
+	err = mlx4_cmd(dev, mailbox->dma, in_mod, MLX4_SET_PORT_ETH_OPCODE,
+		       MLX4_CMD_SET_PORT, MLX4_CMD_TIME_CLASS_B,
+		       MLX4_CMD_WRAPPED);
+
+	mlx4_free_cmd_mailbox(dev, mailbox);
+	return err;
+}
+EXPORT_SYMBOL(mlx4_SET_PORT_user_mtu);
+
+int mlx4_SET_PORT_user_mac(struct mlx4_dev *dev, u8 port, u8 *user_mac)
+{
+	struct mlx4_cmd_mailbox *mailbox;
+	struct mlx4_set_port_general_context *context;
+	u32 in_mod;
+	int err;
+
+	mailbox = mlx4_alloc_cmd_mailbox(dev);
+	if (IS_ERR(mailbox))
+		return PTR_ERR(mailbox);
+	context = mailbox->buf;
+	context->flags2 |= MLX4_FLAG2_V_USER_MAC_MASK;
+	memcpy(context->user_mac, user_mac, sizeof(context->user_mac));
+
+	in_mod = MLX4_SET_PORT_GENERAL << 8 | port;
+	err = mlx4_cmd(dev, mailbox->dma, in_mod, MLX4_SET_PORT_ETH_OPCODE,
+		       MLX4_CMD_SET_PORT, MLX4_CMD_TIME_CLASS_B,
+		       MLX4_CMD_NATIVE);
+
+	mlx4_free_cmd_mailbox(dev, mailbox);
+	return err;
+}
+EXPORT_SYMBOL(mlx4_SET_PORT_user_mac);
+
+int mlx4_SET_PORT_fcs_check(struct mlx4_dev *dev, u8 port, u8 ignore_fcs_value)
+{
+	struct mlx4_cmd_mailbox *mailbox;
+	struct mlx4_set_port_general_context *context;
+	u32 in_mod;
+	int err;
+
+	mailbox = mlx4_alloc_cmd_mailbox(dev);
+	if (IS_ERR(mailbox))
+		return PTR_ERR(mailbox);
+	context = mailbox->buf;
+	context->flags2 |= MLX4_FLAG2_V_IGNORE_FCS_MASK;
+	if (ignore_fcs_value)
+		context->ignore_fcs |= MLX4_IGNORE_FCS_MASK;
+	else
+		context->ignore_fcs &= ~MLX4_IGNORE_FCS_MASK;
+
+	in_mod = MLX4_SET_PORT_GENERAL << 8 | port;
+	err = mlx4_cmd(dev, mailbox->dma, in_mod, 1, MLX4_CMD_SET_PORT,
+		       MLX4_CMD_TIME_CLASS_B, MLX4_CMD_NATIVE);
+
+	mlx4_free_cmd_mailbox(dev, mailbox);
+	return err;
+}
+EXPORT_SYMBOL(mlx4_SET_PORT_fcs_check);
+
+enum {
+	VXLAN_ENABLE_MODIFY	= 1 << 7,
+	VXLAN_STEERING_MODIFY	= 1 << 6,
+
+	VXLAN_ENABLE		= 1 << 7,
+};
+
+struct mlx4_set_port_vxlan_context {
+	u32	reserved1;
+	u8	modify_flags;
+	u8	reserved2;
+	u8	enable_flags;
+	u8	steering;
+};
+
+int mlx4_SET_PORT_VXLAN(struct mlx4_dev *dev, u8 port, u8 steering, int enable)
+{
+	int err;
+	u32 in_mod;
+	struct mlx4_cmd_mailbox *mailbox;
+	struct mlx4_set_port_vxlan_context  *context;
+
+	mailbox = mlx4_alloc_cmd_mailbox(dev);
+	if (IS_ERR(mailbox))
+		return PTR_ERR(mailbox);
+	context = mailbox->buf;
+	memset(context, 0, sizeof(*context));
+
+	context->modify_flags = VXLAN_ENABLE_MODIFY | VXLAN_STEERING_MODIFY;
+	if (enable)
+		context->enable_flags = VXLAN_ENABLE;
+	context->steering  = steering;
+
+	in_mod = MLX4_SET_PORT_VXLAN << 8 | port;
+	err = mlx4_cmd(dev, mailbox->dma, in_mod, MLX4_SET_PORT_ETH_OPCODE,
+		       MLX4_CMD_SET_PORT, MLX4_CMD_TIME_CLASS_B,
+		       MLX4_CMD_NATIVE);
+
+	mlx4_free_cmd_mailbox(dev, mailbox);
+	return err;
+}
+EXPORT_SYMBOL(mlx4_SET_PORT_VXLAN);
+
+int mlx4_SET_PORT_BEACON(struct mlx4_dev *dev, u8 port, u16 time)
+{
+	int err;
+	struct mlx4_cmd_mailbox *mailbox;
+
+	mailbox = mlx4_alloc_cmd_mailbox(dev);
+	if (IS_ERR(mailbox))
+		return PTR_ERR(mailbox);
+
+	*((__be32 *)mailbox->buf) = cpu_to_be32(time);
+
+	err = mlx4_cmd(dev, mailbox->dma, port, MLX4_SET_PORT_BEACON_OPCODE,
+		       MLX4_CMD_SET_PORT, MLX4_CMD_TIME_CLASS_B,
+		       MLX4_CMD_NATIVE);
+
+	mlx4_free_cmd_mailbox(dev, mailbox);
+	return err;
+}
+EXPORT_SYMBOL(mlx4_SET_PORT_BEACON);
+
+int mlx4_SET_MCAST_FLTR_wrapper(struct mlx4_dev *dev, int slave,
+				struct mlx4_vhcr *vhcr,
+				struct mlx4_cmd_mailbox *inbox,
+				struct mlx4_cmd_mailbox *outbox,
+				struct mlx4_cmd_info *cmd)
+{
+	int err = 0;
+
+	return err;
+}
+
+int mlx4_SET_MCAST_FLTR(struct mlx4_dev *dev, u8 port,
+			u64 mac, u64 clear, u8 mode)
+{
+	return mlx4_cmd(dev, (mac | (clear << 63)), port, mode,
+			MLX4_CMD_SET_MCAST_FLTR, MLX4_CMD_TIME_CLASS_B,
+			MLX4_CMD_WRAPPED);
+}
+EXPORT_SYMBOL(mlx4_SET_MCAST_FLTR);
+
+int mlx4_SET_VLAN_FLTR_wrapper(struct mlx4_dev *dev, int slave,
+			       struct mlx4_vhcr *vhcr,
+			       struct mlx4_cmd_mailbox *inbox,
+			       struct mlx4_cmd_mailbox *outbox,
+			       struct mlx4_cmd_info *cmd)
+{
+	int err = 0;
+
+	return err;
+}
+
+int mlx4_DUMP_ETH_STATS_wrapper(struct mlx4_dev *dev, int slave,
+				struct mlx4_vhcr *vhcr,
+				struct mlx4_cmd_mailbox *inbox,
+				struct mlx4_cmd_mailbox *outbox,
+				struct mlx4_cmd_info *cmd)
+{
+	return 0;
+}
+
+int mlx4_get_slave_from_roce_gid(struct mlx4_dev *dev, int port, u8 *gid,
+				 int *slave_id)
+{
+	struct mlx4_priv *priv = mlx4_priv(dev);
+	int i, found_ix = -1;
+	int vf_gids = MLX4_ROCE_MAX_GIDS - MLX4_ROCE_PF_GIDS;
+	struct mlx4_slaves_pport slaves_pport;
+	unsigned num_vfs;
+	int slave_gid;
+
+	if (!mlx4_is_mfunc(dev))
+		return -EINVAL;
+
+	slaves_pport = mlx4_phys_to_slaves_pport(dev, port);
+	num_vfs = bitmap_weight(slaves_pport.slaves,
+				dev->persist->num_vfs + 1) - 1;
+
+	for (i = 0; i < MLX4_ROCE_MAX_GIDS; i++) {
+		if (!memcmp(priv->port[port].gid_table.roce_gids[i].raw, gid,
+			    MLX4_ROCE_GID_ENTRY_SIZE)) {
+			found_ix = i;
+			break;
+		}
+	}
+
+	if (found_ix >= 0) {
+		/* Calculate a slave_gid which is the slave number in the gid
+		 * table and not a globally unique slave number.
+		 */
+		if (found_ix < MLX4_ROCE_PF_GIDS)
+			slave_gid = 0;
+		else if (found_ix < MLX4_ROCE_PF_GIDS + (vf_gids % num_vfs) *
+			 (vf_gids / num_vfs + 1))
+			slave_gid = ((found_ix - MLX4_ROCE_PF_GIDS) /
+				     (vf_gids / num_vfs + 1)) + 1;
+		else
+			slave_gid =
+			((found_ix - MLX4_ROCE_PF_GIDS -
+			  ((vf_gids % num_vfs) * ((vf_gids / num_vfs + 1)))) /
+			 (vf_gids / num_vfs)) + vf_gids % num_vfs + 1;
+
+		/* Calculate the globally unique slave id */
+		if (slave_gid) {
+			struct mlx4_active_ports exclusive_ports;
+			struct mlx4_active_ports actv_ports;
+			struct mlx4_slaves_pport slaves_pport_actv;
+			unsigned max_port_p_one;
+			int num_vfs_before = 0;
+			int candidate_slave_gid;
+
+			/* Calculate how many VFs are on the previous port, if exists */
+			for (i = 1; i < port; i++) {
+				bitmap_zero(exclusive_ports.ports, dev->caps.num_ports);
+				set_bit(i - 1, exclusive_ports.ports);
+				slaves_pport_actv =
+					mlx4_phys_to_slaves_pport_actv(
+							dev, &exclusive_ports);
+				num_vfs_before += bitmap_weight(
+						slaves_pport_actv.slaves,
+						dev->persist->num_vfs + 1);
+			}
+
+			/* candidate_slave_gid isn't necessarily the correct slave, but
+			 * it has the same number of ports and is assigned to the same
+			 * ports as the real slave we're looking for. On dual port VF,
+			 * slave_gid = [single port VFs on port <port>] +
+			 * [offset of the current slave from the first dual port VF] +
+			 * 1 (for the PF).
+			 */
+			candidate_slave_gid = slave_gid + num_vfs_before;
+
+			actv_ports = mlx4_get_active_ports(dev, candidate_slave_gid);
+			max_port_p_one = find_first_bit(
+				actv_ports.ports, dev->caps.num_ports) +
+				bitmap_weight(actv_ports.ports,
+					      dev->caps.num_ports) + 1;
+
+			/* Calculate the real slave number */
+			for (i = 1; i < max_port_p_one; i++) {
+				if (i == port)
+					continue;
+				bitmap_zero(exclusive_ports.ports,
+					    dev->caps.num_ports);
+				set_bit(i - 1, exclusive_ports.ports);
+				slaves_pport_actv =
+					mlx4_phys_to_slaves_pport_actv(
+						dev, &exclusive_ports);
+				slave_gid += bitmap_weight(
+						slaves_pport_actv.slaves,
+						dev->persist->num_vfs + 1);
+			}
+		}
+		*slave_id = slave_gid;
+	}
+
+	return (found_ix >= 0) ? 0 : -EINVAL;
+}
+EXPORT_SYMBOL(mlx4_get_slave_from_roce_gid);
+
+int mlx4_get_roce_gid_from_slave(struct mlx4_dev *dev, int port, int slave_id,
+				 u8 *gid)
+{
+	struct mlx4_priv *priv = mlx4_priv(dev);
+
+	if (!mlx4_is_master(dev))
+		return -EINVAL;
+
+	memcpy(gid, priv->port[port].gid_table.roce_gids[slave_id].raw,
+	       MLX4_ROCE_GID_ENTRY_SIZE);
+	return 0;
+}
+EXPORT_SYMBOL(mlx4_get_roce_gid_from_slave);
+
+/* Cable Module Info */
+#define MODULE_INFO_MAX_READ 48
+
+#define I2C_ADDR_LOW  0x50
+#define I2C_ADDR_HIGH 0x51
+#define I2C_PAGE_SIZE 256
+#define I2C_HIGH_PAGE_SIZE 128
+
+/* Module Info Data */
+struct mlx4_cable_info {
+	u8	i2c_addr;
+	u8	page_num;
+	__be16	dev_mem_address;
+	__be16	reserved1;
+	__be16	size;
+	__be32	reserved2[2];
+	u8	data[MODULE_INFO_MAX_READ];
+};
+
+enum cable_info_err {
+	 CABLE_INF_INV_PORT      = 0x1,
+	 CABLE_INF_OP_NOSUP      = 0x2,
+	 CABLE_INF_NOT_CONN      = 0x3,
+	 CABLE_INF_NO_EEPRM      = 0x4,
+	 CABLE_INF_PAGE_ERR      = 0x5,
+	 CABLE_INF_INV_ADDR      = 0x6,
+	 CABLE_INF_I2C_ADDR      = 0x7,
+	 CABLE_INF_QSFP_VIO      = 0x8,
+	 CABLE_INF_I2C_BUSY      = 0x9,
+};
+
+#define MAD_STATUS_2_CABLE_ERR(mad_status) ((mad_status >> 8) & 0xFF)
+
+static inline const char *cable_info_mad_err_str(u16 mad_status)
+{
+	u8 err = MAD_STATUS_2_CABLE_ERR(mad_status);
+
+	switch (err) {
+	case CABLE_INF_INV_PORT:
+		return "invalid port selected";
+	case CABLE_INF_OP_NOSUP:
+		return "operation not supported for this port (the port is of type CX4 or internal)";
+	case CABLE_INF_NOT_CONN:
+		return "cable is not connected";
+	case CABLE_INF_NO_EEPRM:
+		return "the connected cable has no EPROM (passive copper cable)";
+	case CABLE_INF_PAGE_ERR:
+		return "page number is greater than 15";
+	case CABLE_INF_INV_ADDR:
+		return "invalid device_address or size (that is, size equals 0 or address+size is greater than 256)";
+	case CABLE_INF_I2C_ADDR:
+		return "invalid I2C slave address";
+	case CABLE_INF_QSFP_VIO:
+		return "at least one cable violates the QSFP specification and ignores the modsel signal";
+	case CABLE_INF_I2C_BUSY:
+		return "I2C bus is constantly busy";
+	}
+	return "Unknown Error";
+}
+
+static int mlx4_get_module_id(struct mlx4_dev *dev, u8 port, u8 *module_id)
+{
+	struct mlx4_cmd_mailbox *inbox, *outbox;
+	struct mlx4_mad_ifc *inmad, *outmad;
+	struct mlx4_cable_info *cable_info;
+	int ret;
+
+	inbox = mlx4_alloc_cmd_mailbox(dev);
+	if (IS_ERR(inbox))
+		return PTR_ERR(inbox);
+
+	outbox = mlx4_alloc_cmd_mailbox(dev);
+	if (IS_ERR(outbox)) {
+		mlx4_free_cmd_mailbox(dev, inbox);
+		return PTR_ERR(outbox);
+	}
+
+	inmad = (struct mlx4_mad_ifc *)(inbox->buf);
+	outmad = (struct mlx4_mad_ifc *)(outbox->buf);
+
+	inmad->method = 0x1; /* Get */
+	inmad->class_version = 0x1;
+	inmad->mgmt_class = 0x1;
+	inmad->base_version = 0x1;
+	inmad->attr_id = cpu_to_be16(0xFF60); /* Module Info */
+
+	cable_info = (struct mlx4_cable_info *)inmad->data;
+	cable_info->dev_mem_address = 0;
+	cable_info->page_num = 0;
+	cable_info->i2c_addr = I2C_ADDR_LOW;
+	cable_info->size = cpu_to_be16(1);
+
+	ret = mlx4_cmd_box(dev, inbox->dma, outbox->dma, port, 3,
+			   MLX4_CMD_MAD_IFC, MLX4_CMD_TIME_CLASS_C,
+			   MLX4_CMD_NATIVE);
+	if (ret)
+		goto out;
+
+	if (be16_to_cpu(outmad->status)) {
+		/* Mad returned with bad status */
+		ret = be16_to_cpu(outmad->status);
+		mlx4_warn(dev,
+			  "MLX4_CMD_MAD_IFC Get Module ID attr(%x) port(%d) i2c_addr(%x) offset(%d) size(%d): Response Mad Status(%x) - %s\n",
+			  0xFF60, port, I2C_ADDR_LOW, 0, 1, ret,
+			  cable_info_mad_err_str(ret));
+		ret = -ret;
+		goto out;
+	}
+	cable_info = (struct mlx4_cable_info *)outmad->data;
+	*module_id = cable_info->data[0];
+out:
+	mlx4_free_cmd_mailbox(dev, inbox);
+	mlx4_free_cmd_mailbox(dev, outbox);
+	return ret;
+}
+
+static void mlx4_sfp_eeprom_params_set(u8 *i2c_addr, u8 *page_num, u16 *offset)
+{
+	*i2c_addr = I2C_ADDR_LOW;
+	*page_num = 0;
+
+	if (*offset < I2C_PAGE_SIZE)
+		return;
+
+	*i2c_addr = I2C_ADDR_HIGH;
+	*offset -= I2C_PAGE_SIZE;
+}
+
+static void mlx4_qsfp_eeprom_params_set(u8 *i2c_addr, u8 *page_num, u16 *offset)
+{
+	/* Offsets 0-255 belong to page 0.
+	 * Offsets 256-639 belong to pages 01, 02, 03.
+	 * For example, offset 400 is page 02: 1 + (400 - 256) / 128 = 2
+	 */
+	if (*offset < I2C_PAGE_SIZE)
+		*page_num = 0;
+	else
+		*page_num = 1 + (*offset - I2C_PAGE_SIZE) / I2C_HIGH_PAGE_SIZE;
+	*i2c_addr = I2C_ADDR_LOW;
+	*offset -= *page_num * I2C_HIGH_PAGE_SIZE;
+}
+
+/**
+ * mlx4_get_module_info - Read cable module eeprom data
+ * @dev: mlx4_dev.
+ * @port: port number.
+ * @offset: byte offset in eeprom to start reading data from.
+ * @size: num of bytes to read.
+ * @data: output buffer to put the requested data into.
+ *
+ * Reads cable module eeprom data, puts the outcome data into
+ * data pointer paramer.
+ * Returns num of read bytes on success or a negative error
+ * code.
+ */
+int mlx4_get_module_info(struct mlx4_dev *dev, u8 port,
+			 u16 offset, u16 size, u8 *data)
+{
+	struct mlx4_cmd_mailbox *inbox, *outbox;
+	struct mlx4_mad_ifc *inmad, *outmad;
+	struct mlx4_cable_info *cable_info;
+	u8 module_id, i2c_addr, page_num;
+	int ret;
+
+	if (size > MODULE_INFO_MAX_READ)
+		size = MODULE_INFO_MAX_READ;
+
+	ret = mlx4_get_module_id(dev, port, &module_id);
+	if (ret)
+		return ret;
+
+	switch (module_id) {
+	case MLX4_MODULE_ID_SFP:
+		mlx4_sfp_eeprom_params_set(&i2c_addr, &page_num, &offset);
+		break;
+	case MLX4_MODULE_ID_QSFP:
+	case MLX4_MODULE_ID_QSFP_PLUS:
+	case MLX4_MODULE_ID_QSFP28:
+		mlx4_qsfp_eeprom_params_set(&i2c_addr, &page_num, &offset);
+		break;
+	default:
+		mlx4_err(dev, "Module ID not recognized: %#x\n", module_id);
+		return -EINVAL;
+	}
+
+	inbox = mlx4_alloc_cmd_mailbox(dev);
+	if (IS_ERR(inbox))
+		return PTR_ERR(inbox);
+
+	outbox = mlx4_alloc_cmd_mailbox(dev);
+	if (IS_ERR(outbox)) {
+		mlx4_free_cmd_mailbox(dev, inbox);
+		return PTR_ERR(outbox);
+	}
+
+	inmad = (struct mlx4_mad_ifc *)(inbox->buf);
+	outmad = (struct mlx4_mad_ifc *)(outbox->buf);
+
+	inmad->method = 0x1; /* Get */
+	inmad->class_version = 0x1;
+	inmad->mgmt_class = 0x1;
+	inmad->base_version = 0x1;
+	inmad->attr_id = cpu_to_be16(0xFF60); /* Module Info */
+
+	if (offset < I2C_PAGE_SIZE && offset + size > I2C_PAGE_SIZE)
+		/* Cross pages reads are not allowed
+		 * read until offset 256 in low page
+		 */
+		size -= offset + size - I2C_PAGE_SIZE;
+
+	cable_info = (struct mlx4_cable_info *)inmad->data;
+	cable_info->dev_mem_address = cpu_to_be16(offset);
+	cable_info->page_num = page_num;
+	cable_info->i2c_addr = i2c_addr;
+	cable_info->size = cpu_to_be16(size);
+
+	ret = mlx4_cmd_box(dev, inbox->dma, outbox->dma, port, 3,
+			   MLX4_CMD_MAD_IFC, MLX4_CMD_TIME_CLASS_C,
+			   MLX4_CMD_NATIVE);
+	if (ret)
+		goto out;
+
+	if (be16_to_cpu(outmad->status)) {
+		/* Mad returned with bad status */
+		ret = be16_to_cpu(outmad->status);
+		mlx4_warn(dev,
+			  "MLX4_CMD_MAD_IFC Get Module info attr(%x) port(%d) i2c_addr(%x) offset(%d) size(%d): Response Mad Status(%x) - %s\n",
+			  0xFF60, port, i2c_addr, offset, size,
+			  ret, cable_info_mad_err_str(ret));
+
+		if (i2c_addr == I2C_ADDR_HIGH &&
+		    MAD_STATUS_2_CABLE_ERR(ret) == CABLE_INF_I2C_ADDR)
+			/* Some SFP cables do not support i2c slave
+			 * address 0x51 (high page), abort silently.
+			 */
+			ret = 0;
+		else
+			ret = -ret;
+		goto out;
+	}
+	cable_info = (struct mlx4_cable_info *)outmad->data;
+	memcpy(data, cable_info->data, size);
+	ret = size;
+out:
+	mlx4_free_cmd_mailbox(dev, inbox);
+	mlx4_free_cmd_mailbox(dev, outbox);
+	return ret;
+}
+EXPORT_SYMBOL(mlx4_get_module_info);
+
+int mlx4_max_tc(struct mlx4_dev *dev)
+{
+	u8 num_tc = dev->caps.max_tc_eth;
+
+	if (!num_tc)
+		num_tc = MLX4_TC_MAX_NUMBER;
+
+	return num_tc;
+}
+EXPORT_SYMBOL(mlx4_max_tc);
diff --git a/drivers/net/ethernet/mellanox/mlx4/profile.c b/drivers/net/ethernet/mellanox/mlx4/profile.c
new file mode 100644
index 000000000..ba361c5fb
--- /dev/null
+++ b/drivers/net/ethernet/mellanox/mlx4/profile.c
@@ -0,0 +1,270 @@
+/*
+ * Copyright (c) 2004, 2005 Topspin Communications.  All rights reserved.
+ * Copyright (c) 2005 Mellanox Technologies. All rights reserved.
+ * Copyright (c) 2006, 2007 Cisco Systems, Inc. All rights reserved.
+ *
+ * This software is available to you under a choice of one of two
+ * licenses.  You may choose to be licensed under the terms of the GNU
+ * General Public License (GPL) Version 2, available from the file
+ * COPYING in the main directory of this source tree, or the
+ * OpenIB.org BSD license below:
+ *
+ *     Redistribution and use in source and binary forms, with or
+ *     without modification, are permitted provided that the following
+ *     conditions are met:
+ *
+ *      - Redistributions of source code must retain the above
+ *        copyright notice, this list of conditions and the following
+ *        disclaimer.
+ *
+ *      - Redistributions in binary form must reproduce the above
+ *        copyright notice, this list of conditions and the following
+ *        disclaimer in the documentation and/or other materials
+ *        provided with the distribution.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
+ * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
+ * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS
+ * BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN
+ * ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
+ * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+ * SOFTWARE.
+ */
+
+#include <linux/slab.h>
+
+#include "mlx4.h"
+#include "fw.h"
+
+enum {
+	MLX4_RES_QP,
+	MLX4_RES_RDMARC,
+	MLX4_RES_ALTC,
+	MLX4_RES_AUXC,
+	MLX4_RES_SRQ,
+	MLX4_RES_CQ,
+	MLX4_RES_EQ,
+	MLX4_RES_DMPT,
+	MLX4_RES_CMPT,
+	MLX4_RES_MTT,
+	MLX4_RES_MCG,
+	MLX4_RES_NUM
+};
+
+static const char *res_name[] = {
+	[MLX4_RES_QP]		= "QP",
+	[MLX4_RES_RDMARC]	= "RDMARC",
+	[MLX4_RES_ALTC]		= "ALTC",
+	[MLX4_RES_AUXC]		= "AUXC",
+	[MLX4_RES_SRQ]		= "SRQ",
+	[MLX4_RES_CQ]		= "CQ",
+	[MLX4_RES_EQ]		= "EQ",
+	[MLX4_RES_DMPT]		= "DMPT",
+	[MLX4_RES_CMPT]		= "CMPT",
+	[MLX4_RES_MTT]		= "MTT",
+	[MLX4_RES_MCG]		= "MCG",
+};
+
+u64 mlx4_make_profile(struct mlx4_dev *dev,
+		      struct mlx4_profile *request,
+		      struct mlx4_dev_cap *dev_cap,
+		      struct mlx4_init_hca_param *init_hca)
+{
+	struct mlx4_priv *priv = mlx4_priv(dev);
+	struct mlx4_resource {
+		u64 size;
+		u64 start;
+		int type;
+		u32 num;
+		int log_num;
+	};
+
+	u64 total_size = 0;
+	struct mlx4_resource *profile;
+	struct sysinfo si;
+	int i, j;
+
+	profile = kcalloc(MLX4_RES_NUM, sizeof(*profile), GFP_KERNEL);
+	if (!profile)
+		return -ENOMEM;
+
+	/*
+	 * We want to scale the number of MTTs with the size of the
+	 * system memory, since it makes sense to register a lot of
+	 * memory on a system with a lot of memory.  As a heuristic,
+	 * make sure we have enough MTTs to cover twice the system
+	 * memory (with PAGE_SIZE entries).
+	 *
+	 * This number has to be a power of two and fit into 32 bits
+	 * due to device limitations, so cap this at 2^31 as well.
+	 * That limits us to 8TB of memory registration per HCA with
+	 * 4KB pages, which is probably OK for the next few months.
+	 */
+	si_meminfo(&si);
+	request->num_mtt =
+		roundup_pow_of_two(max_t(unsigned, request->num_mtt,
+					 min(1UL << (31 - log_mtts_per_seg),
+					     (si.totalram << 1) >> log_mtts_per_seg)));
+
+
+	profile[MLX4_RES_QP].size     = dev_cap->qpc_entry_sz;
+	profile[MLX4_RES_RDMARC].size = dev_cap->rdmarc_entry_sz;
+	profile[MLX4_RES_ALTC].size   = dev_cap->altc_entry_sz;
+	profile[MLX4_RES_AUXC].size   = dev_cap->aux_entry_sz;
+	profile[MLX4_RES_SRQ].size    = dev_cap->srq_entry_sz;
+	profile[MLX4_RES_CQ].size     = dev_cap->cqc_entry_sz;
+	profile[MLX4_RES_EQ].size     = dev_cap->eqc_entry_sz;
+	profile[MLX4_RES_DMPT].size   = dev_cap->dmpt_entry_sz;
+	profile[MLX4_RES_CMPT].size   = dev_cap->cmpt_entry_sz;
+	profile[MLX4_RES_MTT].size    = dev_cap->mtt_entry_sz;
+	profile[MLX4_RES_MCG].size    = mlx4_get_mgm_entry_size(dev);
+
+	profile[MLX4_RES_QP].num      = request->num_qp;
+	profile[MLX4_RES_RDMARC].num  = request->num_qp * request->rdmarc_per_qp;
+	profile[MLX4_RES_ALTC].num    = request->num_qp;
+	profile[MLX4_RES_AUXC].num    = request->num_qp;
+	profile[MLX4_RES_SRQ].num     = request->num_srq;
+	profile[MLX4_RES_CQ].num      = request->num_cq;
+	profile[MLX4_RES_EQ].num = mlx4_is_mfunc(dev) ? dev->phys_caps.num_phys_eqs :
+					min_t(unsigned, dev_cap->max_eqs, MAX_MSIX);
+	profile[MLX4_RES_DMPT].num    = request->num_mpt;
+	profile[MLX4_RES_CMPT].num    = MLX4_NUM_CMPTS;
+	profile[MLX4_RES_MTT].num     = request->num_mtt * (1 << log_mtts_per_seg);
+	profile[MLX4_RES_MCG].num     = request->num_mcg;
+
+	for (i = 0; i < MLX4_RES_NUM; ++i) {
+		profile[i].type     = i;
+		profile[i].num      = roundup_pow_of_two(profile[i].num);
+		profile[i].log_num  = ilog2(profile[i].num);
+		profile[i].size    *= profile[i].num;
+		profile[i].size     = max(profile[i].size, (u64) PAGE_SIZE);
+	}
+
+	/*
+	 * Sort the resources in decreasing order of size.  Since they
+	 * all have sizes that are powers of 2, we'll be able to keep
+	 * resources aligned to their size and pack them without gaps
+	 * using the sorted order.
+	 */
+	for (i = MLX4_RES_NUM; i > 0; --i)
+		for (j = 1; j < i; ++j) {
+			if (profile[j].size > profile[j - 1].size)
+				swap(profile[j], profile[j - 1]);
+		}
+
+	for (i = 0; i < MLX4_RES_NUM; ++i) {
+		if (profile[i].size) {
+			profile[i].start = total_size;
+			total_size	+= profile[i].size;
+		}
+
+		if (total_size > dev_cap->max_icm_sz) {
+			mlx4_err(dev, "Profile requires 0x%llx bytes; won't fit in 0x%llx bytes of context memory\n",
+				 (unsigned long long) total_size,
+				 (unsigned long long) dev_cap->max_icm_sz);
+			kfree(profile);
+			return -ENOMEM;
+		}
+
+		if (profile[i].size)
+			mlx4_dbg(dev, "  profile[%2d] (%6s): 2^%02d entries @ 0x%10llx, size 0x%10llx\n",
+				 i, res_name[profile[i].type],
+				 profile[i].log_num,
+				 (unsigned long long) profile[i].start,
+				 (unsigned long long) profile[i].size);
+	}
+
+	mlx4_dbg(dev, "HCA context memory: reserving %d KB\n",
+		 (int) (total_size >> 10));
+
+	for (i = 0; i < MLX4_RES_NUM; ++i) {
+		switch (profile[i].type) {
+		case MLX4_RES_QP:
+			dev->caps.num_qps     = profile[i].num;
+			init_hca->qpc_base    = profile[i].start;
+			init_hca->log_num_qps = profile[i].log_num;
+			break;
+		case MLX4_RES_RDMARC:
+			for (priv->qp_table.rdmarc_shift = 0;
+			     request->num_qp << priv->qp_table.rdmarc_shift < profile[i].num;
+			     ++priv->qp_table.rdmarc_shift)
+				; /* nothing */
+			dev->caps.max_qp_dest_rdma = 1 << priv->qp_table.rdmarc_shift;
+			priv->qp_table.rdmarc_base   = (u32) profile[i].start;
+			init_hca->rdmarc_base	     = profile[i].start;
+			init_hca->log_rd_per_qp	     = priv->qp_table.rdmarc_shift;
+			break;
+		case MLX4_RES_ALTC:
+			init_hca->altc_base = profile[i].start;
+			break;
+		case MLX4_RES_AUXC:
+			init_hca->auxc_base = profile[i].start;
+			break;
+		case MLX4_RES_SRQ:
+			dev->caps.num_srqs     = profile[i].num;
+			init_hca->srqc_base    = profile[i].start;
+			init_hca->log_num_srqs = profile[i].log_num;
+			break;
+		case MLX4_RES_CQ:
+			dev->caps.num_cqs     = profile[i].num;
+			init_hca->cqc_base    = profile[i].start;
+			init_hca->log_num_cqs = profile[i].log_num;
+			break;
+		case MLX4_RES_EQ:
+			if (dev_cap->flags2 & MLX4_DEV_CAP_FLAG2_SYS_EQS) {
+				init_hca->log_num_eqs = 0x1f;
+				init_hca->eqc_base    = profile[i].start;
+				init_hca->num_sys_eqs = dev_cap->num_sys_eqs;
+			} else {
+				dev->caps.num_eqs     = roundup_pow_of_two(
+								min_t(unsigned,
+								      dev_cap->max_eqs,
+								      MAX_MSIX));
+				init_hca->eqc_base    = profile[i].start;
+				init_hca->log_num_eqs = ilog2(dev->caps.num_eqs);
+			}
+			break;
+		case MLX4_RES_DMPT:
+			dev->caps.num_mpts	= profile[i].num;
+			priv->mr_table.mpt_base = profile[i].start;
+			init_hca->dmpt_base	= profile[i].start;
+			init_hca->log_mpt_sz	= profile[i].log_num;
+			break;
+		case MLX4_RES_CMPT:
+			init_hca->cmpt_base	 = profile[i].start;
+			break;
+		case MLX4_RES_MTT:
+			dev->caps.num_mtts	 = profile[i].num;
+			priv->mr_table.mtt_base	 = profile[i].start;
+			init_hca->mtt_base	 = profile[i].start;
+			break;
+		case MLX4_RES_MCG:
+			init_hca->mc_base	  = profile[i].start;
+			init_hca->log_mc_entry_sz =
+					ilog2(mlx4_get_mgm_entry_size(dev));
+			init_hca->log_mc_table_sz = profile[i].log_num;
+			if (dev->caps.steering_mode ==
+			    MLX4_STEERING_MODE_DEVICE_MANAGED) {
+				dev->caps.num_mgms = profile[i].num;
+			} else {
+				init_hca->log_mc_hash_sz =
+						profile[i].log_num - 1;
+				dev->caps.num_mgms = profile[i].num >> 1;
+				dev->caps.num_amgms = profile[i].num >> 1;
+			}
+			break;
+		default:
+			break;
+		}
+	}
+
+	/*
+	 * PDs don't take any HCA memory, but we assign them as part
+	 * of the HCA profile anyway.
+	 */
+	dev->caps.num_pds = MLX4_NUM_PDS;
+
+	kfree(profile);
+	return total_size;
+}
diff --git a/drivers/net/ethernet/mellanox/mlx4/qp.c b/drivers/net/ethernet/mellanox/mlx4/qp.c
new file mode 100644
index 000000000..427e7a318
--- /dev/null
+++ b/drivers/net/ethernet/mellanox/mlx4/qp.c
@@ -0,0 +1,965 @@
+/*
+ * Copyright (c) 2004 Topspin Communications.  All rights reserved.
+ * Copyright (c) 2005, 2006, 2007 Cisco Systems, Inc. All rights reserved.
+ * Copyright (c) 2005, 2006, 2007, 2008 Mellanox Technologies. All rights reserved.
+ * Copyright (c) 2004 Voltaire, Inc. All rights reserved.
+ *
+ * This software is available to you under a choice of one of two
+ * licenses.  You may choose to be licensed under the terms of the GNU
+ * General Public License (GPL) Version 2, available from the file
+ * COPYING in the main directory of this source tree, or the
+ * OpenIB.org BSD license below:
+ *
+ *     Redistribution and use in source and binary forms, with or
+ *     without modification, are permitted provided that the following
+ *     conditions are met:
+ *
+ *      - Redistributions of source code must retain the above
+ *        copyright notice, this list of conditions and the following
+ *        disclaimer.
+ *
+ *      - Redistributions in binary form must reproduce the above
+ *        copyright notice, this list of conditions and the following
+ *        disclaimer in the documentation and/or other materials
+ *        provided with the distribution.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
+ * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
+ * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS
+ * BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN
+ * ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
+ * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+ * SOFTWARE.
+ */
+
+#include <linux/gfp.h>
+#include <linux/export.h>
+
+#include <linux/mlx4/cmd.h>
+#include <linux/mlx4/qp.h>
+
+#include "mlx4.h"
+#include "icm.h"
+
+/* QP to support BF should have bits 6,7 cleared */
+#define MLX4_BF_QP_SKIP_MASK	0xc0
+#define MLX4_MAX_BF_QP_RANGE	0x40
+
+void mlx4_qp_event(struct mlx4_dev *dev, u32 qpn, int event_type)
+{
+	struct mlx4_qp_table *qp_table = &mlx4_priv(dev)->qp_table;
+	struct mlx4_qp *qp;
+
+	spin_lock(&qp_table->lock);
+
+	qp = __mlx4_qp_lookup(dev, qpn);
+	if (qp)
+		refcount_inc(&qp->refcount);
+
+	spin_unlock(&qp_table->lock);
+
+	if (!qp) {
+		mlx4_dbg(dev, "Async event for none existent QP %08x\n", qpn);
+		return;
+	}
+
+	qp->event(qp, event_type);
+
+	if (refcount_dec_and_test(&qp->refcount))
+		complete(&qp->free);
+}
+
+/* used for INIT/CLOSE port logic */
+static int is_master_qp0(struct mlx4_dev *dev, struct mlx4_qp *qp, int *real_qp0, int *proxy_qp0)
+{
+	/* this procedure is called after we already know we are on the master */
+	/* qp0 is either the proxy qp0, or the real qp0 */
+	u32 pf_proxy_offset = dev->phys_caps.base_proxy_sqpn + 8 * mlx4_master_func_num(dev);
+	*proxy_qp0 = qp->qpn >= pf_proxy_offset && qp->qpn <= pf_proxy_offset + 1;
+
+	*real_qp0 = qp->qpn >= dev->phys_caps.base_sqpn &&
+		qp->qpn <= dev->phys_caps.base_sqpn + 1;
+
+	return *real_qp0 || *proxy_qp0;
+}
+
+static int __mlx4_qp_modify(struct mlx4_dev *dev, struct mlx4_mtt *mtt,
+		     enum mlx4_qp_state cur_state, enum mlx4_qp_state new_state,
+		     struct mlx4_qp_context *context,
+		     enum mlx4_qp_optpar optpar,
+		     int sqd_event, struct mlx4_qp *qp, int native)
+{
+	static const u16 op[MLX4_QP_NUM_STATE][MLX4_QP_NUM_STATE] = {
+		[MLX4_QP_STATE_RST] = {
+			[MLX4_QP_STATE_RST]	= MLX4_CMD_2RST_QP,
+			[MLX4_QP_STATE_ERR]	= MLX4_CMD_2ERR_QP,
+			[MLX4_QP_STATE_INIT]	= MLX4_CMD_RST2INIT_QP,
+		},
+		[MLX4_QP_STATE_INIT]  = {
+			[MLX4_QP_STATE_RST]	= MLX4_CMD_2RST_QP,
+			[MLX4_QP_STATE_ERR]	= MLX4_CMD_2ERR_QP,
+			[MLX4_QP_STATE_INIT]	= MLX4_CMD_INIT2INIT_QP,
+			[MLX4_QP_STATE_RTR]	= MLX4_CMD_INIT2RTR_QP,
+		},
+		[MLX4_QP_STATE_RTR]   = {
+			[MLX4_QP_STATE_RST]	= MLX4_CMD_2RST_QP,
+			[MLX4_QP_STATE_ERR]	= MLX4_CMD_2ERR_QP,
+			[MLX4_QP_STATE_RTS]	= MLX4_CMD_RTR2RTS_QP,
+		},
+		[MLX4_QP_STATE_RTS]   = {
+			[MLX4_QP_STATE_RST]	= MLX4_CMD_2RST_QP,
+			[MLX4_QP_STATE_ERR]	= MLX4_CMD_2ERR_QP,
+			[MLX4_QP_STATE_RTS]	= MLX4_CMD_RTS2RTS_QP,
+			[MLX4_QP_STATE_SQD]	= MLX4_CMD_RTS2SQD_QP,
+		},
+		[MLX4_QP_STATE_SQD] = {
+			[MLX4_QP_STATE_RST]	= MLX4_CMD_2RST_QP,
+			[MLX4_QP_STATE_ERR]	= MLX4_CMD_2ERR_QP,
+			[MLX4_QP_STATE_RTS]	= MLX4_CMD_SQD2RTS_QP,
+			[MLX4_QP_STATE_SQD]	= MLX4_CMD_SQD2SQD_QP,
+		},
+		[MLX4_QP_STATE_SQER] = {
+			[MLX4_QP_STATE_RST]	= MLX4_CMD_2RST_QP,
+			[MLX4_QP_STATE_ERR]	= MLX4_CMD_2ERR_QP,
+			[MLX4_QP_STATE_RTS]	= MLX4_CMD_SQERR2RTS_QP,
+		},
+		[MLX4_QP_STATE_ERR] = {
+			[MLX4_QP_STATE_RST]	= MLX4_CMD_2RST_QP,
+			[MLX4_QP_STATE_ERR]	= MLX4_CMD_2ERR_QP,
+		}
+	};
+
+	struct mlx4_priv *priv = mlx4_priv(dev);
+	struct mlx4_cmd_mailbox *mailbox;
+	int ret = 0;
+	int real_qp0 = 0;
+	int proxy_qp0 = 0;
+	u8 port;
+
+	if (cur_state >= MLX4_QP_NUM_STATE || new_state >= MLX4_QP_NUM_STATE ||
+	    !op[cur_state][new_state])
+		return -EINVAL;
+
+	if (op[cur_state][new_state] == MLX4_CMD_2RST_QP) {
+		ret = mlx4_cmd(dev, 0, qp->qpn, 2,
+			MLX4_CMD_2RST_QP, MLX4_CMD_TIME_CLASS_A, native);
+		if (mlx4_is_master(dev) && cur_state != MLX4_QP_STATE_ERR &&
+		    cur_state != MLX4_QP_STATE_RST &&
+		    is_master_qp0(dev, qp, &real_qp0, &proxy_qp0)) {
+			port = (qp->qpn & 1) + 1;
+			if (proxy_qp0)
+				priv->mfunc.master.qp0_state[port].proxy_qp0_active = 0;
+			else
+				priv->mfunc.master.qp0_state[port].qp0_active = 0;
+		}
+		return ret;
+	}
+
+	mailbox = mlx4_alloc_cmd_mailbox(dev);
+	if (IS_ERR(mailbox))
+		return PTR_ERR(mailbox);
+
+	if (cur_state == MLX4_QP_STATE_RST && new_state == MLX4_QP_STATE_INIT) {
+		u64 mtt_addr = mlx4_mtt_addr(dev, mtt);
+		context->mtt_base_addr_h = mtt_addr >> 32;
+		context->mtt_base_addr_l = cpu_to_be32(mtt_addr & 0xffffffff);
+		context->log_page_size   = mtt->page_shift - MLX4_ICM_PAGE_SHIFT;
+	}
+
+	if ((cur_state == MLX4_QP_STATE_RTR) &&
+	    (new_state == MLX4_QP_STATE_RTS) &&
+	    dev->caps.flags2 & MLX4_DEV_CAP_FLAG2_ROCE_V1_V2)
+		context->roce_entropy =
+			cpu_to_be16(mlx4_qp_roce_entropy(dev, qp->qpn));
+
+	*(__be32 *) mailbox->buf = cpu_to_be32(optpar);
+	memcpy(mailbox->buf + 8, context, sizeof(*context));
+
+	((struct mlx4_qp_context *) (mailbox->buf + 8))->local_qpn =
+		cpu_to_be32(qp->qpn);
+
+	ret = mlx4_cmd(dev, mailbox->dma,
+		       qp->qpn | (!!sqd_event << 31),
+		       new_state == MLX4_QP_STATE_RST ? 2 : 0,
+		       op[cur_state][new_state], MLX4_CMD_TIME_CLASS_C, native);
+
+	if (mlx4_is_master(dev) && is_master_qp0(dev, qp, &real_qp0, &proxy_qp0)) {
+		port = (qp->qpn & 1) + 1;
+		if (cur_state != MLX4_QP_STATE_ERR &&
+		    cur_state != MLX4_QP_STATE_RST &&
+		    new_state == MLX4_QP_STATE_ERR) {
+			if (proxy_qp0)
+				priv->mfunc.master.qp0_state[port].proxy_qp0_active = 0;
+			else
+				priv->mfunc.master.qp0_state[port].qp0_active = 0;
+		} else if (new_state == MLX4_QP_STATE_RTR) {
+			if (proxy_qp0)
+				priv->mfunc.master.qp0_state[port].proxy_qp0_active = 1;
+			else
+				priv->mfunc.master.qp0_state[port].qp0_active = 1;
+		}
+	}
+
+	mlx4_free_cmd_mailbox(dev, mailbox);
+	return ret;
+}
+
+int mlx4_qp_modify(struct mlx4_dev *dev, struct mlx4_mtt *mtt,
+		   enum mlx4_qp_state cur_state, enum mlx4_qp_state new_state,
+		   struct mlx4_qp_context *context,
+		   enum mlx4_qp_optpar optpar,
+		   int sqd_event, struct mlx4_qp *qp)
+{
+	return __mlx4_qp_modify(dev, mtt, cur_state, new_state, context,
+				optpar, sqd_event, qp, 0);
+}
+EXPORT_SYMBOL_GPL(mlx4_qp_modify);
+
+int __mlx4_qp_reserve_range(struct mlx4_dev *dev, int cnt, int align,
+			    int *base, u8 flags)
+{
+	u32 uid;
+	int bf_qp = !!(flags & (u8)MLX4_RESERVE_ETH_BF_QP);
+
+	struct mlx4_priv *priv = mlx4_priv(dev);
+	struct mlx4_qp_table *qp_table = &priv->qp_table;
+
+	if (cnt > MLX4_MAX_BF_QP_RANGE && bf_qp)
+		return -ENOMEM;
+
+	uid = MLX4_QP_TABLE_ZONE_GENERAL;
+	if (flags & (u8)MLX4_RESERVE_A0_QP) {
+		if (bf_qp)
+			uid = MLX4_QP_TABLE_ZONE_RAW_ETH;
+		else
+			uid = MLX4_QP_TABLE_ZONE_RSS;
+	}
+
+	*base = mlx4_zone_alloc_entries(qp_table->zones, uid, cnt, align,
+					bf_qp ? MLX4_BF_QP_SKIP_MASK : 0, NULL);
+	if (*base == -1)
+		return -ENOMEM;
+
+	return 0;
+}
+
+int mlx4_qp_reserve_range(struct mlx4_dev *dev, int cnt, int align,
+			  int *base, u8 flags, u8 usage)
+{
+	u32 in_modifier = RES_QP | (((u32)usage & 3) << 30);
+	u64 in_param = 0;
+	u64 out_param;
+	int err;
+
+	/* Turn off all unsupported QP allocation flags */
+	flags &= dev->caps.alloc_res_qp_mask;
+
+	if (mlx4_is_mfunc(dev)) {
+		set_param_l(&in_param, (((u32)flags) << 24) | (u32)cnt);
+		set_param_h(&in_param, align);
+		err = mlx4_cmd_imm(dev, in_param, &out_param,
+				   in_modifier, RES_OP_RESERVE,
+				   MLX4_CMD_ALLOC_RES,
+				   MLX4_CMD_TIME_CLASS_A, MLX4_CMD_WRAPPED);
+		if (err)
+			return err;
+
+		*base = get_param_l(&out_param);
+		return 0;
+	}
+	return __mlx4_qp_reserve_range(dev, cnt, align, base, flags);
+}
+EXPORT_SYMBOL_GPL(mlx4_qp_reserve_range);
+
+void __mlx4_qp_release_range(struct mlx4_dev *dev, int base_qpn, int cnt)
+{
+	struct mlx4_priv *priv = mlx4_priv(dev);
+	struct mlx4_qp_table *qp_table = &priv->qp_table;
+
+	if (mlx4_is_qp_reserved(dev, (u32) base_qpn))
+		return;
+	mlx4_zone_free_entries_unique(qp_table->zones, base_qpn, cnt);
+}
+
+void mlx4_qp_release_range(struct mlx4_dev *dev, int base_qpn, int cnt)
+{
+	u64 in_param = 0;
+	int err;
+
+	if (!cnt)
+		return;
+
+	if (mlx4_is_mfunc(dev)) {
+		set_param_l(&in_param, base_qpn);
+		set_param_h(&in_param, cnt);
+		err = mlx4_cmd(dev, in_param, RES_QP, RES_OP_RESERVE,
+			       MLX4_CMD_FREE_RES,
+			       MLX4_CMD_TIME_CLASS_A, MLX4_CMD_WRAPPED);
+		if (err) {
+			mlx4_warn(dev, "Failed to release qp range base:%d cnt:%d\n",
+				  base_qpn, cnt);
+		}
+	} else
+		 __mlx4_qp_release_range(dev, base_qpn, cnt);
+}
+EXPORT_SYMBOL_GPL(mlx4_qp_release_range);
+
+int __mlx4_qp_alloc_icm(struct mlx4_dev *dev, int qpn)
+{
+	struct mlx4_priv *priv = mlx4_priv(dev);
+	struct mlx4_qp_table *qp_table = &priv->qp_table;
+	int err;
+
+	err = mlx4_table_get(dev, &qp_table->qp_table, qpn);
+	if (err)
+		goto err_out;
+
+	err = mlx4_table_get(dev, &qp_table->auxc_table, qpn);
+	if (err)
+		goto err_put_qp;
+
+	err = mlx4_table_get(dev, &qp_table->altc_table, qpn);
+	if (err)
+		goto err_put_auxc;
+
+	err = mlx4_table_get(dev, &qp_table->rdmarc_table, qpn);
+	if (err)
+		goto err_put_altc;
+
+	err = mlx4_table_get(dev, &qp_table->cmpt_table, qpn);
+	if (err)
+		goto err_put_rdmarc;
+
+	return 0;
+
+err_put_rdmarc:
+	mlx4_table_put(dev, &qp_table->rdmarc_table, qpn);
+
+err_put_altc:
+	mlx4_table_put(dev, &qp_table->altc_table, qpn);
+
+err_put_auxc:
+	mlx4_table_put(dev, &qp_table->auxc_table, qpn);
+
+err_put_qp:
+	mlx4_table_put(dev, &qp_table->qp_table, qpn);
+
+err_out:
+	return err;
+}
+
+static int mlx4_qp_alloc_icm(struct mlx4_dev *dev, int qpn)
+{
+	u64 param = 0;
+
+	if (mlx4_is_mfunc(dev)) {
+		set_param_l(&param, qpn);
+		return mlx4_cmd_imm(dev, param, &param, RES_QP, RES_OP_MAP_ICM,
+				    MLX4_CMD_ALLOC_RES, MLX4_CMD_TIME_CLASS_A,
+				    MLX4_CMD_WRAPPED);
+	}
+	return __mlx4_qp_alloc_icm(dev, qpn);
+}
+
+void __mlx4_qp_free_icm(struct mlx4_dev *dev, int qpn)
+{
+	struct mlx4_priv *priv = mlx4_priv(dev);
+	struct mlx4_qp_table *qp_table = &priv->qp_table;
+
+	mlx4_table_put(dev, &qp_table->cmpt_table, qpn);
+	mlx4_table_put(dev, &qp_table->rdmarc_table, qpn);
+	mlx4_table_put(dev, &qp_table->altc_table, qpn);
+	mlx4_table_put(dev, &qp_table->auxc_table, qpn);
+	mlx4_table_put(dev, &qp_table->qp_table, qpn);
+}
+
+static void mlx4_qp_free_icm(struct mlx4_dev *dev, int qpn)
+{
+	u64 in_param = 0;
+
+	if (mlx4_is_mfunc(dev)) {
+		set_param_l(&in_param, qpn);
+		if (mlx4_cmd(dev, in_param, RES_QP, RES_OP_MAP_ICM,
+			     MLX4_CMD_FREE_RES, MLX4_CMD_TIME_CLASS_A,
+			     MLX4_CMD_WRAPPED))
+			mlx4_warn(dev, "Failed to free icm of qp:%d\n", qpn);
+	} else
+		__mlx4_qp_free_icm(dev, qpn);
+}
+
+struct mlx4_qp *mlx4_qp_lookup(struct mlx4_dev *dev, u32 qpn)
+{
+	struct mlx4_qp_table *qp_table = &mlx4_priv(dev)->qp_table;
+	struct mlx4_qp *qp;
+
+	spin_lock_irq(&qp_table->lock);
+
+	qp = __mlx4_qp_lookup(dev, qpn);
+
+	spin_unlock_irq(&qp_table->lock);
+	return qp;
+}
+
+int mlx4_qp_alloc(struct mlx4_dev *dev, int qpn, struct mlx4_qp *qp)
+{
+	struct mlx4_priv *priv = mlx4_priv(dev);
+	struct mlx4_qp_table *qp_table = &priv->qp_table;
+	int err;
+
+	if (!qpn)
+		return -EINVAL;
+
+	qp->qpn = qpn;
+
+	err = mlx4_qp_alloc_icm(dev, qpn);
+	if (err)
+		return err;
+
+	spin_lock_irq(&qp_table->lock);
+	err = radix_tree_insert(&dev->qp_table_tree, qp->qpn &
+				(dev->caps.num_qps - 1), qp);
+	spin_unlock_irq(&qp_table->lock);
+	if (err)
+		goto err_icm;
+
+	refcount_set(&qp->refcount, 1);
+	init_completion(&qp->free);
+
+	return 0;
+
+err_icm:
+	mlx4_qp_free_icm(dev, qpn);
+	return err;
+}
+
+EXPORT_SYMBOL_GPL(mlx4_qp_alloc);
+
+int mlx4_update_qp(struct mlx4_dev *dev, u32 qpn,
+		   enum mlx4_update_qp_attr attr,
+		   struct mlx4_update_qp_params *params)
+{
+	struct mlx4_cmd_mailbox *mailbox;
+	struct mlx4_update_qp_context *cmd;
+	u64 pri_addr_path_mask = 0;
+	u64 qp_mask = 0;
+	int err = 0;
+
+	if (!attr || (attr & ~MLX4_UPDATE_QP_SUPPORTED_ATTRS))
+		return -EINVAL;
+
+	mailbox = mlx4_alloc_cmd_mailbox(dev);
+	if (IS_ERR(mailbox))
+		return PTR_ERR(mailbox);
+
+	cmd = (struct mlx4_update_qp_context *)mailbox->buf;
+
+	if (attr & MLX4_UPDATE_QP_SMAC) {
+		pri_addr_path_mask |= 1ULL << MLX4_UPD_QP_PATH_MASK_MAC_INDEX;
+		cmd->qp_context.pri_path.grh_mylmc = params->smac_index;
+	}
+
+	if (attr & MLX4_UPDATE_QP_ETH_SRC_CHECK_MC_LB) {
+		if (!(dev->caps.flags2
+		      & MLX4_DEV_CAP_FLAG2_UPDATE_QP_SRC_CHECK_LB)) {
+			mlx4_warn(dev,
+				  "Trying to set src check LB, but it isn't supported\n");
+			err = -EOPNOTSUPP;
+			goto out;
+		}
+		pri_addr_path_mask |=
+			1ULL << MLX4_UPD_QP_PATH_MASK_ETH_SRC_CHECK_MC_LB;
+		if (params->flags &
+		    MLX4_UPDATE_QP_PARAMS_FLAGS_ETH_CHECK_MC_LB) {
+			cmd->qp_context.pri_path.fl |=
+				MLX4_FL_ETH_SRC_CHECK_MC_LB;
+		}
+	}
+
+	if (attr & MLX4_UPDATE_QP_VSD) {
+		qp_mask |= 1ULL << MLX4_UPD_QP_MASK_VSD;
+		if (params->flags & MLX4_UPDATE_QP_PARAMS_FLAGS_VSD_ENABLE)
+			cmd->qp_context.param3 |= cpu_to_be32(MLX4_STRIP_VLAN);
+	}
+
+	if (attr & MLX4_UPDATE_QP_RATE_LIMIT) {
+		qp_mask |= 1ULL << MLX4_UPD_QP_MASK_RATE_LIMIT;
+		cmd->qp_context.rate_limit_params = cpu_to_be16((params->rate_unit << 14) | params->rate_val);
+	}
+
+	if (attr & MLX4_UPDATE_QP_QOS_VPORT) {
+		if (!(dev->caps.flags2 & MLX4_DEV_CAP_FLAG2_QOS_VPP)) {
+			mlx4_warn(dev, "Granular QoS per VF is not enabled\n");
+			err = -EOPNOTSUPP;
+			goto out;
+		}
+
+		qp_mask |= 1ULL << MLX4_UPD_QP_MASK_QOS_VPP;
+		cmd->qp_context.qos_vport = params->qos_vport;
+	}
+
+	cmd->primary_addr_path_mask = cpu_to_be64(pri_addr_path_mask);
+	cmd->qp_mask = cpu_to_be64(qp_mask);
+
+	err = mlx4_cmd(dev, mailbox->dma, qpn & 0xffffff, 0,
+		       MLX4_CMD_UPDATE_QP, MLX4_CMD_TIME_CLASS_A,
+		       MLX4_CMD_NATIVE);
+out:
+	mlx4_free_cmd_mailbox(dev, mailbox);
+	return err;
+}
+EXPORT_SYMBOL_GPL(mlx4_update_qp);
+
+void mlx4_qp_remove(struct mlx4_dev *dev, struct mlx4_qp *qp)
+{
+	struct mlx4_qp_table *qp_table = &mlx4_priv(dev)->qp_table;
+	unsigned long flags;
+
+	spin_lock_irqsave(&qp_table->lock, flags);
+	radix_tree_delete(&dev->qp_table_tree, qp->qpn & (dev->caps.num_qps - 1));
+	spin_unlock_irqrestore(&qp_table->lock, flags);
+}
+EXPORT_SYMBOL_GPL(mlx4_qp_remove);
+
+void mlx4_qp_free(struct mlx4_dev *dev, struct mlx4_qp *qp)
+{
+	if (refcount_dec_and_test(&qp->refcount))
+		complete(&qp->free);
+	wait_for_completion(&qp->free);
+
+	mlx4_qp_free_icm(dev, qp->qpn);
+}
+EXPORT_SYMBOL_GPL(mlx4_qp_free);
+
+static int mlx4_CONF_SPECIAL_QP(struct mlx4_dev *dev, u32 base_qpn)
+{
+	return mlx4_cmd(dev, 0, base_qpn, 0, MLX4_CMD_CONF_SPECIAL_QP,
+			MLX4_CMD_TIME_CLASS_B, MLX4_CMD_NATIVE);
+}
+
+#define MLX4_QP_TABLE_RSS_ETH_PRIORITY 2
+#define MLX4_QP_TABLE_RAW_ETH_PRIORITY 1
+#define MLX4_QP_TABLE_RAW_ETH_SIZE     256
+
+static int mlx4_create_zones(struct mlx4_dev *dev,
+			     u32 reserved_bottom_general,
+			     u32 reserved_top_general,
+			     u32 reserved_bottom_rss,
+			     u32 start_offset_rss,
+			     u32 max_table_offset)
+{
+	struct mlx4_qp_table *qp_table = &mlx4_priv(dev)->qp_table;
+	struct mlx4_bitmap (*bitmap)[MLX4_QP_TABLE_ZONE_NUM] = NULL;
+	int bitmap_initialized = 0;
+	u32 last_offset;
+	int k;
+	int err;
+
+	qp_table->zones = mlx4_zone_allocator_create(MLX4_ZONE_ALLOC_FLAGS_NO_OVERLAP);
+
+	if (NULL == qp_table->zones)
+		return -ENOMEM;
+
+	bitmap = kmalloc(sizeof(*bitmap), GFP_KERNEL);
+
+	if (NULL == bitmap) {
+		err = -ENOMEM;
+		goto free_zone;
+	}
+
+	err = mlx4_bitmap_init(*bitmap + MLX4_QP_TABLE_ZONE_GENERAL, dev->caps.num_qps,
+			       (1 << 23) - 1, reserved_bottom_general,
+			       reserved_top_general);
+
+	if (err)
+		goto free_bitmap;
+
+	++bitmap_initialized;
+
+	err = mlx4_zone_add_one(qp_table->zones, *bitmap + MLX4_QP_TABLE_ZONE_GENERAL,
+				MLX4_ZONE_FALLBACK_TO_HIGHER_PRIO |
+				MLX4_ZONE_USE_RR, 0,
+				0, qp_table->zones_uids + MLX4_QP_TABLE_ZONE_GENERAL);
+
+	if (err)
+		goto free_bitmap;
+
+	err = mlx4_bitmap_init(*bitmap + MLX4_QP_TABLE_ZONE_RSS,
+			       reserved_bottom_rss,
+			       reserved_bottom_rss - 1,
+			       dev->caps.reserved_qps_cnt[MLX4_QP_REGION_FW],
+			       reserved_bottom_rss - start_offset_rss);
+
+	if (err)
+		goto free_bitmap;
+
+	++bitmap_initialized;
+
+	err = mlx4_zone_add_one(qp_table->zones, *bitmap + MLX4_QP_TABLE_ZONE_RSS,
+				MLX4_ZONE_ALLOW_ALLOC_FROM_LOWER_PRIO |
+				MLX4_ZONE_ALLOW_ALLOC_FROM_EQ_PRIO |
+				MLX4_ZONE_USE_RR, MLX4_QP_TABLE_RSS_ETH_PRIORITY,
+				0, qp_table->zones_uids + MLX4_QP_TABLE_ZONE_RSS);
+
+	if (err)
+		goto free_bitmap;
+
+	last_offset = dev->caps.reserved_qps_cnt[MLX4_QP_REGION_FW];
+	/*  We have a single zone for the A0 steering QPs area of the FW. This area
+	 *  needs to be split into subareas. One set of subareas is for RSS QPs
+	 *  (in which qp number bits 6 and/or 7 are set); the other set of subareas
+	 *  is for RAW_ETH QPs, which require that both bits 6 and 7 are zero.
+	 *  Currently, the values returned by the FW (A0 steering area starting qp number
+	 *  and A0 steering area size) are such that there are only two subareas -- one
+	 *  for RSS and one for RAW_ETH.
+	 */
+	for (k = MLX4_QP_TABLE_ZONE_RSS + 1; k < sizeof(*bitmap)/sizeof((*bitmap)[0]);
+	     k++) {
+		int size;
+		u32 offset = start_offset_rss;
+		u32 bf_mask;
+		u32 requested_size;
+
+		/* Assuming MLX4_BF_QP_SKIP_MASK is consecutive ones, this calculates
+		 * a mask of all LSB bits set until (and not including) the first
+		 * set bit of  MLX4_BF_QP_SKIP_MASK. For example, if MLX4_BF_QP_SKIP_MASK
+		 * is 0xc0, bf_mask will be 0x3f.
+		 */
+		bf_mask = (MLX4_BF_QP_SKIP_MASK & ~(MLX4_BF_QP_SKIP_MASK - 1)) - 1;
+		requested_size = min((u32)MLX4_QP_TABLE_RAW_ETH_SIZE, bf_mask + 1);
+
+		if (((last_offset & MLX4_BF_QP_SKIP_MASK) &&
+		     ((int)(max_table_offset - last_offset)) >=
+		     roundup_pow_of_two(MLX4_BF_QP_SKIP_MASK)) ||
+		    (!(last_offset & MLX4_BF_QP_SKIP_MASK) &&
+		     !((last_offset + requested_size - 1) &
+		       MLX4_BF_QP_SKIP_MASK)))
+			size = requested_size;
+		else {
+			u32 candidate_offset =
+				(last_offset | MLX4_BF_QP_SKIP_MASK | bf_mask) + 1;
+
+			if (last_offset & MLX4_BF_QP_SKIP_MASK)
+				last_offset = candidate_offset;
+
+			/* From this point, the BF bits are 0 */
+
+			if (last_offset > max_table_offset) {
+				/* need to skip */
+				size = -1;
+			} else {
+				size = min3(max_table_offset - last_offset,
+					    bf_mask - (last_offset & bf_mask),
+					    requested_size);
+				if (size < requested_size) {
+					int candidate_size;
+
+					candidate_size = min3(
+						max_table_offset - candidate_offset,
+						bf_mask - (last_offset & bf_mask),
+						requested_size);
+
+					/*  We will not take this path if last_offset was
+					 *  already set above to candidate_offset
+					 */
+					if (candidate_size > size) {
+						last_offset = candidate_offset;
+						size = candidate_size;
+					}
+				}
+			}
+		}
+
+		if (size > 0) {
+			/* mlx4_bitmap_alloc_range will find a contiguous range of "size"
+			 * QPs in which both bits 6 and 7 are zero, because we pass it the
+			 * MLX4_BF_SKIP_MASK).
+			 */
+			offset = mlx4_bitmap_alloc_range(
+					*bitmap + MLX4_QP_TABLE_ZONE_RSS,
+					size, 1,
+					MLX4_BF_QP_SKIP_MASK);
+
+			if (offset == (u32)-1) {
+				err = -ENOMEM;
+				break;
+			}
+
+			last_offset = offset + size;
+
+			err = mlx4_bitmap_init(*bitmap + k, roundup_pow_of_two(size),
+					       roundup_pow_of_two(size) - 1, 0,
+					       roundup_pow_of_two(size) - size);
+		} else {
+			/* Add an empty bitmap, we'll allocate from different zones (since
+			 * at least one is reserved)
+			 */
+			err = mlx4_bitmap_init(*bitmap + k, 1,
+					       MLX4_QP_TABLE_RAW_ETH_SIZE - 1, 0,
+					       0);
+			mlx4_bitmap_alloc_range(*bitmap + k, 1, 1, 0);
+		}
+
+		if (err)
+			break;
+
+		++bitmap_initialized;
+
+		err = mlx4_zone_add_one(qp_table->zones, *bitmap + k,
+					MLX4_ZONE_ALLOW_ALLOC_FROM_LOWER_PRIO |
+					MLX4_ZONE_ALLOW_ALLOC_FROM_EQ_PRIO |
+					MLX4_ZONE_USE_RR, MLX4_QP_TABLE_RAW_ETH_PRIORITY,
+					offset, qp_table->zones_uids + k);
+
+		if (err)
+			break;
+	}
+
+	if (err)
+		goto free_bitmap;
+
+	qp_table->bitmap_gen = *bitmap;
+
+	return err;
+
+free_bitmap:
+	for (k = 0; k < bitmap_initialized; k++)
+		mlx4_bitmap_cleanup(*bitmap + k);
+	kfree(bitmap);
+free_zone:
+	mlx4_zone_allocator_destroy(qp_table->zones);
+	return err;
+}
+
+static void mlx4_cleanup_qp_zones(struct mlx4_dev *dev)
+{
+	struct mlx4_qp_table *qp_table = &mlx4_priv(dev)->qp_table;
+
+	if (qp_table->zones) {
+		int i;
+
+		for (i = 0;
+		     i < sizeof(qp_table->zones_uids)/sizeof(qp_table->zones_uids[0]);
+		     i++) {
+			struct mlx4_bitmap *bitmap =
+				mlx4_zone_get_bitmap(qp_table->zones,
+						     qp_table->zones_uids[i]);
+
+			mlx4_zone_remove_one(qp_table->zones, qp_table->zones_uids[i]);
+			if (NULL == bitmap)
+				continue;
+
+			mlx4_bitmap_cleanup(bitmap);
+		}
+		mlx4_zone_allocator_destroy(qp_table->zones);
+		kfree(qp_table->bitmap_gen);
+		qp_table->bitmap_gen = NULL;
+		qp_table->zones = NULL;
+	}
+}
+
+int mlx4_init_qp_table(struct mlx4_dev *dev)
+{
+	struct mlx4_qp_table *qp_table = &mlx4_priv(dev)->qp_table;
+	int err;
+	int reserved_from_top = 0;
+	int reserved_from_bot;
+	int k;
+	int fixed_reserved_from_bot_rv = 0;
+	int bottom_reserved_for_rss_bitmap;
+	u32 max_table_offset = dev->caps.dmfs_high_rate_qpn_base +
+			dev->caps.dmfs_high_rate_qpn_range;
+
+	spin_lock_init(&qp_table->lock);
+	INIT_RADIX_TREE(&dev->qp_table_tree, GFP_ATOMIC);
+	if (mlx4_is_slave(dev))
+		return 0;
+
+	/* We reserve 2 extra QPs per port for the special QPs.  The
+	 * block of special QPs must be aligned to a multiple of 8, so
+	 * round up.
+	 *
+	 * We also reserve the MSB of the 24-bit QP number to indicate
+	 * that a QP is an XRC QP.
+	 */
+	for (k = 0; k <= MLX4_QP_REGION_BOTTOM; k++)
+		fixed_reserved_from_bot_rv += dev->caps.reserved_qps_cnt[k];
+
+	if (fixed_reserved_from_bot_rv < max_table_offset)
+		fixed_reserved_from_bot_rv = max_table_offset;
+
+	/* We reserve at least 1 extra for bitmaps that we don't have enough space for*/
+	bottom_reserved_for_rss_bitmap =
+		roundup_pow_of_two(fixed_reserved_from_bot_rv + 1);
+	dev->phys_caps.base_sqpn = ALIGN(bottom_reserved_for_rss_bitmap, 8);
+
+	{
+		int sort[MLX4_NUM_QP_REGION];
+		int i, j;
+		int last_base = dev->caps.num_qps;
+
+		for (i = 1; i < MLX4_NUM_QP_REGION; ++i)
+			sort[i] = i;
+
+		for (i = MLX4_NUM_QP_REGION; i > MLX4_QP_REGION_BOTTOM; --i) {
+			for (j = MLX4_QP_REGION_BOTTOM + 2; j < i; ++j) {
+				if (dev->caps.reserved_qps_cnt[sort[j]] >
+				    dev->caps.reserved_qps_cnt[sort[j - 1]])
+					swap(sort[j], sort[j - 1]);
+			}
+		}
+
+		for (i = MLX4_QP_REGION_BOTTOM + 1; i < MLX4_NUM_QP_REGION; ++i) {
+			last_base -= dev->caps.reserved_qps_cnt[sort[i]];
+			dev->caps.reserved_qps_base[sort[i]] = last_base;
+			reserved_from_top +=
+				dev->caps.reserved_qps_cnt[sort[i]];
+		}
+	}
+
+       /* Reserve 8 real SQPs in both native and SRIOV modes.
+	* In addition, in SRIOV mode, reserve 8 proxy SQPs per function
+	* (for all PFs and VFs), and 8 corresponding tunnel QPs.
+	* Each proxy SQP works opposite its own tunnel QP.
+	*
+	* The QPs are arranged as follows:
+	* a. 8 real SQPs
+	* b. All the proxy SQPs (8 per function)
+	* c. All the tunnel QPs (8 per function)
+	*/
+	reserved_from_bot = mlx4_num_reserved_sqps(dev);
+	if (reserved_from_bot + reserved_from_top > dev->caps.num_qps) {
+		mlx4_err(dev, "Number of reserved QPs is higher than number of QPs\n");
+		return -EINVAL;
+	}
+
+	err = mlx4_create_zones(dev, reserved_from_bot, reserved_from_bot,
+				bottom_reserved_for_rss_bitmap,
+				fixed_reserved_from_bot_rv,
+				max_table_offset);
+
+	if (err)
+		return err;
+
+	if (mlx4_is_mfunc(dev)) {
+		/* for PPF use */
+		dev->phys_caps.base_proxy_sqpn = dev->phys_caps.base_sqpn + 8;
+		dev->phys_caps.base_tunnel_sqpn = dev->phys_caps.base_sqpn + 8 + 8 * MLX4_MFUNC_MAX;
+
+		/* In mfunc, calculate proxy and tunnel qp offsets for the PF here,
+		 * since the PF does not call mlx4_slave_caps */
+		dev->caps.spec_qps = kcalloc(dev->caps.num_ports,
+					     sizeof(*dev->caps.spec_qps),
+					     GFP_KERNEL);
+		if (!dev->caps.spec_qps) {
+			err = -ENOMEM;
+			goto err_mem;
+		}
+
+		for (k = 0; k < dev->caps.num_ports; k++) {
+			dev->caps.spec_qps[k].qp0_proxy = dev->phys_caps.base_proxy_sqpn +
+				8 * mlx4_master_func_num(dev) + k;
+			dev->caps.spec_qps[k].qp0_tunnel = dev->caps.spec_qps[k].qp0_proxy + 8 * MLX4_MFUNC_MAX;
+			dev->caps.spec_qps[k].qp1_proxy = dev->phys_caps.base_proxy_sqpn +
+				8 * mlx4_master_func_num(dev) + MLX4_MAX_PORTS + k;
+			dev->caps.spec_qps[k].qp1_tunnel = dev->caps.spec_qps[k].qp1_proxy + 8 * MLX4_MFUNC_MAX;
+		}
+	}
+
+
+	err = mlx4_CONF_SPECIAL_QP(dev, dev->phys_caps.base_sqpn);
+	if (err)
+		goto err_mem;
+
+	return err;
+
+err_mem:
+	kfree(dev->caps.spec_qps);
+	dev->caps.spec_qps = NULL;
+	mlx4_cleanup_qp_zones(dev);
+	return err;
+}
+
+void mlx4_cleanup_qp_table(struct mlx4_dev *dev)
+{
+	if (mlx4_is_slave(dev))
+		return;
+
+	mlx4_CONF_SPECIAL_QP(dev, 0);
+
+	mlx4_cleanup_qp_zones(dev);
+}
+
+int mlx4_qp_query(struct mlx4_dev *dev, struct mlx4_qp *qp,
+		  struct mlx4_qp_context *context)
+{
+	struct mlx4_cmd_mailbox *mailbox;
+	int err;
+
+	mailbox = mlx4_alloc_cmd_mailbox(dev);
+	if (IS_ERR(mailbox))
+		return PTR_ERR(mailbox);
+
+	err = mlx4_cmd_box(dev, 0, mailbox->dma, qp->qpn, 0,
+			   MLX4_CMD_QUERY_QP, MLX4_CMD_TIME_CLASS_A,
+			   MLX4_CMD_WRAPPED);
+	if (!err)
+		memcpy(context, mailbox->buf + 8, sizeof(*context));
+
+	mlx4_free_cmd_mailbox(dev, mailbox);
+	return err;
+}
+EXPORT_SYMBOL_GPL(mlx4_qp_query);
+
+int mlx4_qp_to_ready(struct mlx4_dev *dev, struct mlx4_mtt *mtt,
+		     struct mlx4_qp_context *context,
+		     struct mlx4_qp *qp, enum mlx4_qp_state *qp_state)
+{
+	int err;
+	int i;
+	enum mlx4_qp_state states[] = {
+		MLX4_QP_STATE_RST,
+		MLX4_QP_STATE_INIT,
+		MLX4_QP_STATE_RTR,
+		MLX4_QP_STATE_RTS
+	};
+
+	for (i = 0; i < ARRAY_SIZE(states) - 1; i++) {
+		context->flags &= cpu_to_be32(~(0xf << 28));
+		context->flags |= cpu_to_be32(states[i + 1] << 28);
+		if (states[i + 1] != MLX4_QP_STATE_RTR)
+			context->params2 &= ~cpu_to_be32(MLX4_QP_BIT_FPP);
+		err = mlx4_qp_modify(dev, mtt, states[i], states[i + 1],
+				     context, 0, 0, qp);
+		if (err) {
+			mlx4_err(dev, "Failed to bring QP to state: %d with error: %d\n",
+				 states[i + 1], err);
+			return err;
+		}
+
+		*qp_state = states[i + 1];
+	}
+
+	return 0;
+}
+EXPORT_SYMBOL_GPL(mlx4_qp_to_ready);
+
+u16 mlx4_qp_roce_entropy(struct mlx4_dev *dev, u32 qpn)
+{
+	struct mlx4_qp_context context;
+	struct mlx4_qp qp;
+	int err;
+
+	qp.qpn = qpn;
+	err = mlx4_qp_query(dev, &qp, &context);
+	if (!err) {
+		u32 dest_qpn = be32_to_cpu(context.remote_qpn) & 0xffffff;
+		u16 folded_dst = folded_qp(dest_qpn);
+		u16 folded_src = folded_qp(qpn);
+
+		return (dest_qpn != qpn) ?
+			((folded_dst ^ folded_src) | 0xC000) :
+			folded_src | 0xC000;
+	}
+	return 0xdead;
+}
diff --git a/drivers/net/ethernet/mellanox/mlx4/reset.c b/drivers/net/ethernet/mellanox/mlx4/reset.c
new file mode 100644
index 000000000..0076d8858
--- /dev/null
+++ b/drivers/net/ethernet/mellanox/mlx4/reset.c
@@ -0,0 +1,184 @@
+/*
+ * Copyright (c) 2006, 2007 Cisco Systems, Inc.  All rights reserved.
+ * Copyright (c) 2007, 2008 Mellanox Technologies. All rights reserved.
+ *
+ * This software is available to you under a choice of one of two
+ * licenses.  You may choose to be licensed under the terms of the GNU
+ * General Public License (GPL) Version 2, available from the file
+ * COPYING in the main directory of this source tree, or the
+ * OpenIB.org BSD license below:
+ *
+ *     Redistribution and use in source and binary forms, with or
+ *     without modification, are permitted provided that the following
+ *     conditions are met:
+ *
+ *      - Redistributions of source code must retain the above
+ *        copyright notice, this list of conditions and the following
+ *        disclaimer.
+ *
+ *      - Redistributions in binary form must reproduce the above
+ *        copyright notice, this list of conditions and the following
+ *        disclaimer in the documentation and/or other materials
+ *        provided with the distribution.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
+ * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
+ * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS
+ * BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN
+ * ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
+ * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+ * SOFTWARE.
+ */
+
+#include <linux/errno.h>
+#include <linux/pci.h>
+#include <linux/delay.h>
+#include <linux/slab.h>
+#include <linux/jiffies.h>
+
+#include "mlx4.h"
+
+int mlx4_reset(struct mlx4_dev *dev)
+{
+	void __iomem *reset;
+	u32 *hca_header = NULL;
+	int pcie_cap;
+	u16 devctl;
+	u16 linkctl;
+	u16 vendor;
+	unsigned long end;
+	u32 sem;
+	int i;
+	int err = 0;
+
+#define MLX4_RESET_BASE		0xf0000
+#define MLX4_RESET_SIZE		  0x400
+#define MLX4_SEM_OFFSET		  0x3fc
+#define MLX4_RESET_OFFSET	   0x10
+#define MLX4_RESET_VALUE	swab32(1)
+
+#define MLX4_SEM_TIMEOUT_JIFFIES	(10 * HZ)
+#define MLX4_RESET_TIMEOUT_JIFFIES	(2 * HZ)
+
+	/*
+	 * Reset the chip.  This is somewhat ugly because we have to
+	 * save off the PCI header before reset and then restore it
+	 * after the chip reboots.  We skip config space offsets 22
+	 * and 23 since those have a special meaning.
+	 */
+
+	/* Do we need to save off the full 4K PCI Express header?? */
+	hca_header = kmalloc(256, GFP_KERNEL);
+	if (!hca_header) {
+		err = -ENOMEM;
+		mlx4_err(dev, "Couldn't allocate memory to save HCA PCI header, aborting\n");
+		goto out;
+	}
+
+	pcie_cap = pci_pcie_cap(dev->persist->pdev);
+
+	for (i = 0; i < 64; ++i) {
+		if (i == 22 || i == 23)
+			continue;
+		if (pci_read_config_dword(dev->persist->pdev, i * 4,
+					  hca_header + i)) {
+			err = -ENODEV;
+			mlx4_err(dev, "Couldn't save HCA PCI header, aborting\n");
+			goto out;
+		}
+	}
+
+	reset = ioremap(pci_resource_start(dev->persist->pdev, 0) +
+			MLX4_RESET_BASE,
+			MLX4_RESET_SIZE);
+	if (!reset) {
+		err = -ENOMEM;
+		mlx4_err(dev, "Couldn't map HCA reset register, aborting\n");
+		goto out;
+	}
+
+	/* grab HW semaphore to lock out flash updates */
+	end = jiffies + MLX4_SEM_TIMEOUT_JIFFIES;
+	do {
+		sem = readl(reset + MLX4_SEM_OFFSET);
+		if (!sem)
+			break;
+
+		msleep(1);
+	} while (time_before(jiffies, end));
+
+	if (sem) {
+		mlx4_err(dev, "Failed to obtain HW semaphore, aborting\n");
+		err = -EAGAIN;
+		iounmap(reset);
+		goto out;
+	}
+
+	/* actually hit reset */
+	writel(MLX4_RESET_VALUE, reset + MLX4_RESET_OFFSET);
+	iounmap(reset);
+
+	/* Docs say to wait one second before accessing device */
+	msleep(1000);
+
+	end = jiffies + MLX4_RESET_TIMEOUT_JIFFIES;
+	do {
+		if (!pci_read_config_word(dev->persist->pdev, PCI_VENDOR_ID,
+					  &vendor) && vendor != 0xffff)
+			break;
+
+		msleep(1);
+	} while (time_before(jiffies, end));
+
+	if (vendor == 0xffff) {
+		err = -ENODEV;
+		mlx4_err(dev, "PCI device did not come back after reset, aborting\n");
+		goto out;
+	}
+
+	/* Now restore the PCI headers */
+	if (pcie_cap) {
+		devctl = hca_header[(pcie_cap + PCI_EXP_DEVCTL) / 4];
+		if (pcie_capability_write_word(dev->persist->pdev,
+					       PCI_EXP_DEVCTL,
+					       devctl)) {
+			err = -ENODEV;
+			mlx4_err(dev, "Couldn't restore HCA PCI Express Device Control register, aborting\n");
+			goto out;
+		}
+		linkctl = hca_header[(pcie_cap + PCI_EXP_LNKCTL) / 4];
+		if (pcie_capability_write_word(dev->persist->pdev,
+					       PCI_EXP_LNKCTL,
+					       linkctl)) {
+			err = -ENODEV;
+			mlx4_err(dev, "Couldn't restore HCA PCI Express Link control register, aborting\n");
+			goto out;
+		}
+	}
+
+	for (i = 0; i < 16; ++i) {
+		if (i * 4 == PCI_COMMAND)
+			continue;
+
+		if (pci_write_config_dword(dev->persist->pdev, i * 4,
+					   hca_header[i])) {
+			err = -ENODEV;
+			mlx4_err(dev, "Couldn't restore HCA reg %x, aborting\n",
+				 i);
+			goto out;
+		}
+	}
+
+	if (pci_write_config_dword(dev->persist->pdev, PCI_COMMAND,
+				   hca_header[PCI_COMMAND / 4])) {
+		err = -ENODEV;
+		mlx4_err(dev, "Couldn't restore HCA COMMAND, aborting\n");
+		goto out;
+	}
+
+out:
+	kfree(hca_header);
+
+	return err;
+}
diff --git a/drivers/net/ethernet/mellanox/mlx4/resource_tracker.c b/drivers/net/ethernet/mellanox/mlx4/resource_tracker.c
new file mode 100644
index 000000000..550e48932
--- /dev/null
+++ b/drivers/net/ethernet/mellanox/mlx4/resource_tracker.c
@@ -0,0 +1,5427 @@
+/*
+ * Copyright (c) 2004, 2005 Topspin Communications.  All rights reserved.
+ * Copyright (c) 2005, 2006, 2007, 2008 Mellanox Technologies.
+ * All rights reserved.
+ * Copyright (c) 2005, 2006, 2007 Cisco Systems, Inc.  All rights reserved.
+ *
+ * This software is available to you under a choice of one of two
+ * licenses.  You may choose to be licensed under the terms of the GNU
+ * General Public License (GPL) Version 2, available from the file
+ * COPYING in the main directory of this source tree, or the
+ * OpenIB.org BSD license below:
+ *
+ *     Redistribution and use in source and binary forms, with or
+ *     without modification, are permitted provided that the following
+ *     conditions are met:
+ *
+ *      - Redistributions of source code must retain the above
+ *        copyright notice, this list of conditions and the following
+ *        disclaimer.
+ *
+ *      - Redistributions in binary form must reproduce the above
+ *        copyright notice, this list of conditions and the following
+ *        disclaimer in the documentation and/or other materials
+ *        provided with the distribution.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
+ * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
+ * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS
+ * BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN
+ * ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
+ * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+ * SOFTWARE.
+ */
+
+#include <linux/sched.h>
+#include <linux/pci.h>
+#include <linux/errno.h>
+#include <linux/kernel.h>
+#include <linux/io.h>
+#include <linux/slab.h>
+#include <linux/mlx4/cmd.h>
+#include <linux/mlx4/qp.h>
+#include <linux/if_ether.h>
+#include <linux/etherdevice.h>
+
+#include "mlx4.h"
+#include "fw.h"
+#include "mlx4_stats.h"
+
+#define MLX4_MAC_VALID		(1ull << 63)
+#define MLX4_PF_COUNTERS_PER_PORT	2
+#define MLX4_VF_COUNTERS_PER_PORT	1
+
+struct mac_res {
+	struct list_head list;
+	u64 mac;
+	int ref_count;
+	u8 smac_index;
+	u8 port;
+};
+
+struct vlan_res {
+	struct list_head list;
+	u16 vlan;
+	int ref_count;
+	int vlan_index;
+	u8 port;
+};
+
+struct res_common {
+	struct list_head	list;
+	struct rb_node		node;
+	u64		        res_id;
+	int			owner;
+	int			state;
+	int			from_state;
+	int			to_state;
+	int			removing;
+	const char		*func_name;
+};
+
+enum {
+	RES_ANY_BUSY = 1
+};
+
+struct res_gid {
+	struct list_head	list;
+	u8			gid[16];
+	enum mlx4_protocol	prot;
+	enum mlx4_steer_type	steer;
+	u64			reg_id;
+};
+
+enum res_qp_states {
+	RES_QP_BUSY = RES_ANY_BUSY,
+
+	/* QP number was allocated */
+	RES_QP_RESERVED,
+
+	/* ICM memory for QP context was mapped */
+	RES_QP_MAPPED,
+
+	/* QP is in hw ownership */
+	RES_QP_HW
+};
+
+struct res_qp {
+	struct res_common	com;
+	struct res_mtt	       *mtt;
+	struct res_cq	       *rcq;
+	struct res_cq	       *scq;
+	struct res_srq	       *srq;
+	struct list_head	mcg_list;
+	spinlock_t		mcg_spl;
+	int			local_qpn;
+	atomic_t		ref_count;
+	u32			qpc_flags;
+	/* saved qp params before VST enforcement in order to restore on VGT */
+	u8			sched_queue;
+	__be32			param3;
+	u8			vlan_control;
+	u8			fvl_rx;
+	u8			pri_path_fl;
+	u8			vlan_index;
+	u8			feup;
+};
+
+enum res_mtt_states {
+	RES_MTT_BUSY = RES_ANY_BUSY,
+	RES_MTT_ALLOCATED,
+};
+
+static inline const char *mtt_states_str(enum res_mtt_states state)
+{
+	switch (state) {
+	case RES_MTT_BUSY: return "RES_MTT_BUSY";
+	case RES_MTT_ALLOCATED: return "RES_MTT_ALLOCATED";
+	default: return "Unknown";
+	}
+}
+
+struct res_mtt {
+	struct res_common	com;
+	int			order;
+	atomic_t		ref_count;
+};
+
+enum res_mpt_states {
+	RES_MPT_BUSY = RES_ANY_BUSY,
+	RES_MPT_RESERVED,
+	RES_MPT_MAPPED,
+	RES_MPT_HW,
+};
+
+struct res_mpt {
+	struct res_common	com;
+	struct res_mtt	       *mtt;
+	int			key;
+};
+
+enum res_eq_states {
+	RES_EQ_BUSY = RES_ANY_BUSY,
+	RES_EQ_RESERVED,
+	RES_EQ_HW,
+};
+
+struct res_eq {
+	struct res_common	com;
+	struct res_mtt	       *mtt;
+};
+
+enum res_cq_states {
+	RES_CQ_BUSY = RES_ANY_BUSY,
+	RES_CQ_ALLOCATED,
+	RES_CQ_HW,
+};
+
+struct res_cq {
+	struct res_common	com;
+	struct res_mtt	       *mtt;
+	atomic_t		ref_count;
+};
+
+enum res_srq_states {
+	RES_SRQ_BUSY = RES_ANY_BUSY,
+	RES_SRQ_ALLOCATED,
+	RES_SRQ_HW,
+};
+
+struct res_srq {
+	struct res_common	com;
+	struct res_mtt	       *mtt;
+	struct res_cq	       *cq;
+	atomic_t		ref_count;
+};
+
+enum res_counter_states {
+	RES_COUNTER_BUSY = RES_ANY_BUSY,
+	RES_COUNTER_ALLOCATED,
+};
+
+struct res_counter {
+	struct res_common	com;
+	int			port;
+};
+
+enum res_xrcdn_states {
+	RES_XRCD_BUSY = RES_ANY_BUSY,
+	RES_XRCD_ALLOCATED,
+};
+
+struct res_xrcdn {
+	struct res_common	com;
+	int			port;
+};
+
+enum res_fs_rule_states {
+	RES_FS_RULE_BUSY = RES_ANY_BUSY,
+	RES_FS_RULE_ALLOCATED,
+};
+
+struct res_fs_rule {
+	struct res_common	com;
+	int			qpn;
+	/* VF DMFS mbox with port flipped */
+	void			*mirr_mbox;
+	/* > 0 --> apply mirror when getting into HA mode      */
+	/* = 0 --> un-apply mirror when getting out of HA mode */
+	u32			mirr_mbox_size;
+	struct list_head	mirr_list;
+	u64			mirr_rule_id;
+};
+
+static void *res_tracker_lookup(struct rb_root *root, u64 res_id)
+{
+	struct rb_node *node = root->rb_node;
+
+	while (node) {
+		struct res_common *res = rb_entry(node, struct res_common,
+						  node);
+
+		if (res_id < res->res_id)
+			node = node->rb_left;
+		else if (res_id > res->res_id)
+			node = node->rb_right;
+		else
+			return res;
+	}
+	return NULL;
+}
+
+static int res_tracker_insert(struct rb_root *root, struct res_common *res)
+{
+	struct rb_node **new = &(root->rb_node), *parent = NULL;
+
+	/* Figure out where to put new node */
+	while (*new) {
+		struct res_common *this = rb_entry(*new, struct res_common,
+						   node);
+
+		parent = *new;
+		if (res->res_id < this->res_id)
+			new = &((*new)->rb_left);
+		else if (res->res_id > this->res_id)
+			new = &((*new)->rb_right);
+		else
+			return -EEXIST;
+	}
+
+	/* Add new node and rebalance tree. */
+	rb_link_node(&res->node, parent, new);
+	rb_insert_color(&res->node, root);
+
+	return 0;
+}
+
+enum qp_transition {
+	QP_TRANS_INIT2RTR,
+	QP_TRANS_RTR2RTS,
+	QP_TRANS_RTS2RTS,
+	QP_TRANS_SQERR2RTS,
+	QP_TRANS_SQD2SQD,
+	QP_TRANS_SQD2RTS
+};
+
+/* For Debug uses */
+static const char *resource_str(enum mlx4_resource rt)
+{
+	switch (rt) {
+	case RES_QP: return "RES_QP";
+	case RES_CQ: return "RES_CQ";
+	case RES_SRQ: return "RES_SRQ";
+	case RES_MPT: return "RES_MPT";
+	case RES_MTT: return "RES_MTT";
+	case RES_MAC: return  "RES_MAC";
+	case RES_VLAN: return  "RES_VLAN";
+	case RES_EQ: return "RES_EQ";
+	case RES_COUNTER: return "RES_COUNTER";
+	case RES_FS_RULE: return "RES_FS_RULE";
+	case RES_XRCD: return "RES_XRCD";
+	default: return "Unknown resource type !!!";
+	};
+}
+
+static void rem_slave_vlans(struct mlx4_dev *dev, int slave);
+static inline int mlx4_grant_resource(struct mlx4_dev *dev, int slave,
+				      enum mlx4_resource res_type, int count,
+				      int port)
+{
+	struct mlx4_priv *priv = mlx4_priv(dev);
+	struct resource_allocator *res_alloc =
+		&priv->mfunc.master.res_tracker.res_alloc[res_type];
+	int err = -EDQUOT;
+	int allocated, free, reserved, guaranteed, from_free;
+	int from_rsvd;
+
+	if (slave > dev->persist->num_vfs)
+		return -EINVAL;
+
+	spin_lock(&res_alloc->alloc_lock);
+	allocated = (port > 0) ?
+		res_alloc->allocated[(port - 1) *
+		(dev->persist->num_vfs + 1) + slave] :
+		res_alloc->allocated[slave];
+	free = (port > 0) ? res_alloc->res_port_free[port - 1] :
+		res_alloc->res_free;
+	reserved = (port > 0) ? res_alloc->res_port_rsvd[port - 1] :
+		res_alloc->res_reserved;
+	guaranteed = res_alloc->guaranteed[slave];
+
+	if (allocated + count > res_alloc->quota[slave]) {
+		mlx4_warn(dev, "VF %d port %d res %s: quota exceeded, count %d alloc %d quota %d\n",
+			  slave, port, resource_str(res_type), count,
+			  allocated, res_alloc->quota[slave]);
+		goto out;
+	}
+
+	if (allocated + count <= guaranteed) {
+		err = 0;
+		from_rsvd = count;
+	} else {
+		/* portion may need to be obtained from free area */
+		if (guaranteed - allocated > 0)
+			from_free = count - (guaranteed - allocated);
+		else
+			from_free = count;
+
+		from_rsvd = count - from_free;
+
+		if (free - from_free >= reserved)
+			err = 0;
+		else
+			mlx4_warn(dev, "VF %d port %d res %s: free pool empty, free %d from_free %d rsvd %d\n",
+				  slave, port, resource_str(res_type), free,
+				  from_free, reserved);
+	}
+
+	if (!err) {
+		/* grant the request */
+		if (port > 0) {
+			res_alloc->allocated[(port - 1) *
+			(dev->persist->num_vfs + 1) + slave] += count;
+			res_alloc->res_port_free[port - 1] -= count;
+			res_alloc->res_port_rsvd[port - 1] -= from_rsvd;
+		} else {
+			res_alloc->allocated[slave] += count;
+			res_alloc->res_free -= count;
+			res_alloc->res_reserved -= from_rsvd;
+		}
+	}
+
+out:
+	spin_unlock(&res_alloc->alloc_lock);
+	return err;
+}
+
+static inline void mlx4_release_resource(struct mlx4_dev *dev, int slave,
+				    enum mlx4_resource res_type, int count,
+				    int port)
+{
+	struct mlx4_priv *priv = mlx4_priv(dev);
+	struct resource_allocator *res_alloc =
+		&priv->mfunc.master.res_tracker.res_alloc[res_type];
+	int allocated, guaranteed, from_rsvd;
+
+	if (slave > dev->persist->num_vfs)
+		return;
+
+	spin_lock(&res_alloc->alloc_lock);
+
+	allocated = (port > 0) ?
+		res_alloc->allocated[(port - 1) *
+		(dev->persist->num_vfs + 1) + slave] :
+		res_alloc->allocated[slave];
+	guaranteed = res_alloc->guaranteed[slave];
+
+	if (allocated - count >= guaranteed) {
+		from_rsvd = 0;
+	} else {
+		/* portion may need to be returned to reserved area */
+		if (allocated - guaranteed > 0)
+			from_rsvd = count - (allocated - guaranteed);
+		else
+			from_rsvd = count;
+	}
+
+	if (port > 0) {
+		res_alloc->allocated[(port - 1) *
+		(dev->persist->num_vfs + 1) + slave] -= count;
+		res_alloc->res_port_free[port - 1] += count;
+		res_alloc->res_port_rsvd[port - 1] += from_rsvd;
+	} else {
+		res_alloc->allocated[slave] -= count;
+		res_alloc->res_free += count;
+		res_alloc->res_reserved += from_rsvd;
+	}
+
+	spin_unlock(&res_alloc->alloc_lock);
+	return;
+}
+
+static inline void initialize_res_quotas(struct mlx4_dev *dev,
+					 struct resource_allocator *res_alloc,
+					 enum mlx4_resource res_type,
+					 int vf, int num_instances)
+{
+	res_alloc->guaranteed[vf] = num_instances /
+				    (2 * (dev->persist->num_vfs + 1));
+	res_alloc->quota[vf] = (num_instances / 2) + res_alloc->guaranteed[vf];
+	if (vf == mlx4_master_func_num(dev)) {
+		res_alloc->res_free = num_instances;
+		if (res_type == RES_MTT) {
+			/* reserved mtts will be taken out of the PF allocation */
+			res_alloc->res_free += dev->caps.reserved_mtts;
+			res_alloc->guaranteed[vf] += dev->caps.reserved_mtts;
+			res_alloc->quota[vf] += dev->caps.reserved_mtts;
+		}
+	}
+}
+
+void mlx4_init_quotas(struct mlx4_dev *dev)
+{
+	struct mlx4_priv *priv = mlx4_priv(dev);
+	int pf;
+
+	/* quotas for VFs are initialized in mlx4_slave_cap */
+	if (mlx4_is_slave(dev))
+		return;
+
+	if (!mlx4_is_mfunc(dev)) {
+		dev->quotas.qp = dev->caps.num_qps - dev->caps.reserved_qps -
+			mlx4_num_reserved_sqps(dev);
+		dev->quotas.cq = dev->caps.num_cqs - dev->caps.reserved_cqs;
+		dev->quotas.srq = dev->caps.num_srqs - dev->caps.reserved_srqs;
+		dev->quotas.mtt = dev->caps.num_mtts - dev->caps.reserved_mtts;
+		dev->quotas.mpt = dev->caps.num_mpts - dev->caps.reserved_mrws;
+		return;
+	}
+
+	pf = mlx4_master_func_num(dev);
+	dev->quotas.qp =
+		priv->mfunc.master.res_tracker.res_alloc[RES_QP].quota[pf];
+	dev->quotas.cq =
+		priv->mfunc.master.res_tracker.res_alloc[RES_CQ].quota[pf];
+	dev->quotas.srq =
+		priv->mfunc.master.res_tracker.res_alloc[RES_SRQ].quota[pf];
+	dev->quotas.mtt =
+		priv->mfunc.master.res_tracker.res_alloc[RES_MTT].quota[pf];
+	dev->quotas.mpt =
+		priv->mfunc.master.res_tracker.res_alloc[RES_MPT].quota[pf];
+}
+
+static int
+mlx4_calc_res_counter_guaranteed(struct mlx4_dev *dev,
+				 struct resource_allocator *res_alloc,
+				 int vf)
+{
+	struct mlx4_active_ports actv_ports;
+	int ports, counters_guaranteed;
+
+	/* For master, only allocate according to the number of phys ports */
+	if (vf == mlx4_master_func_num(dev))
+		return MLX4_PF_COUNTERS_PER_PORT * dev->caps.num_ports;
+
+	/* calculate real number of ports for the VF */
+	actv_ports = mlx4_get_active_ports(dev, vf);
+	ports = bitmap_weight(actv_ports.ports, dev->caps.num_ports);
+	counters_guaranteed = ports * MLX4_VF_COUNTERS_PER_PORT;
+
+	/* If we do not have enough counters for this VF, do not
+	 * allocate any for it. '-1' to reduce the sink counter.
+	 */
+	if ((res_alloc->res_reserved + counters_guaranteed) >
+	    (dev->caps.max_counters - 1))
+		return 0;
+
+	return counters_guaranteed;
+}
+
+int mlx4_init_resource_tracker(struct mlx4_dev *dev)
+{
+	struct mlx4_priv *priv = mlx4_priv(dev);
+	int i, j;
+	int t;
+
+	priv->mfunc.master.res_tracker.slave_list =
+		kcalloc(dev->num_slaves, sizeof(struct slave_list),
+			GFP_KERNEL);
+	if (!priv->mfunc.master.res_tracker.slave_list)
+		return -ENOMEM;
+
+	for (i = 0 ; i < dev->num_slaves; i++) {
+		for (t = 0; t < MLX4_NUM_OF_RESOURCE_TYPE; ++t)
+			INIT_LIST_HEAD(&priv->mfunc.master.res_tracker.
+				       slave_list[i].res_list[t]);
+		mutex_init(&priv->mfunc.master.res_tracker.slave_list[i].mutex);
+	}
+
+	mlx4_dbg(dev, "Started init_resource_tracker: %ld slaves\n",
+		 dev->num_slaves);
+	for (i = 0 ; i < MLX4_NUM_OF_RESOURCE_TYPE; i++)
+		priv->mfunc.master.res_tracker.res_tree[i] = RB_ROOT;
+
+	for (i = 0; i < MLX4_NUM_OF_RESOURCE_TYPE; i++) {
+		struct resource_allocator *res_alloc =
+			&priv->mfunc.master.res_tracker.res_alloc[i];
+		res_alloc->quota = kmalloc_array(dev->persist->num_vfs + 1,
+						 sizeof(int),
+						 GFP_KERNEL);
+		res_alloc->guaranteed = kmalloc_array(dev->persist->num_vfs + 1,
+						      sizeof(int),
+						      GFP_KERNEL);
+		if (i == RES_MAC || i == RES_VLAN)
+			res_alloc->allocated =
+				kcalloc(MLX4_MAX_PORTS *
+						(dev->persist->num_vfs + 1),
+					sizeof(int), GFP_KERNEL);
+		else
+			res_alloc->allocated =
+				kcalloc(dev->persist->num_vfs + 1,
+					sizeof(int), GFP_KERNEL);
+		/* Reduce the sink counter */
+		if (i == RES_COUNTER)
+			res_alloc->res_free = dev->caps.max_counters - 1;
+
+		if (!res_alloc->quota || !res_alloc->guaranteed ||
+		    !res_alloc->allocated)
+			goto no_mem_err;
+
+		spin_lock_init(&res_alloc->alloc_lock);
+		for (t = 0; t < dev->persist->num_vfs + 1; t++) {
+			struct mlx4_active_ports actv_ports =
+				mlx4_get_active_ports(dev, t);
+			switch (i) {
+			case RES_QP:
+				initialize_res_quotas(dev, res_alloc, RES_QP,
+						      t, dev->caps.num_qps -
+						      dev->caps.reserved_qps -
+						      mlx4_num_reserved_sqps(dev));
+				break;
+			case RES_CQ:
+				initialize_res_quotas(dev, res_alloc, RES_CQ,
+						      t, dev->caps.num_cqs -
+						      dev->caps.reserved_cqs);
+				break;
+			case RES_SRQ:
+				initialize_res_quotas(dev, res_alloc, RES_SRQ,
+						      t, dev->caps.num_srqs -
+						      dev->caps.reserved_srqs);
+				break;
+			case RES_MPT:
+				initialize_res_quotas(dev, res_alloc, RES_MPT,
+						      t, dev->caps.num_mpts -
+						      dev->caps.reserved_mrws);
+				break;
+			case RES_MTT:
+				initialize_res_quotas(dev, res_alloc, RES_MTT,
+						      t, dev->caps.num_mtts -
+						      dev->caps.reserved_mtts);
+				break;
+			case RES_MAC:
+				if (t == mlx4_master_func_num(dev)) {
+					int max_vfs_pport = 0;
+					/* Calculate the max vfs per port for */
+					/* both ports.			      */
+					for (j = 0; j < dev->caps.num_ports;
+					     j++) {
+						struct mlx4_slaves_pport slaves_pport =
+							mlx4_phys_to_slaves_pport(dev, j + 1);
+						unsigned current_slaves =
+							bitmap_weight(slaves_pport.slaves,
+								      dev->caps.num_ports) - 1;
+						if (max_vfs_pport < current_slaves)
+							max_vfs_pport =
+								current_slaves;
+					}
+					res_alloc->quota[t] =
+						MLX4_MAX_MAC_NUM -
+						2 * max_vfs_pport;
+					res_alloc->guaranteed[t] = 2;
+					for (j = 0; j < MLX4_MAX_PORTS; j++)
+						res_alloc->res_port_free[j] =
+							MLX4_MAX_MAC_NUM;
+				} else {
+					res_alloc->quota[t] = MLX4_MAX_MAC_NUM;
+					res_alloc->guaranteed[t] = 2;
+				}
+				break;
+			case RES_VLAN:
+				if (t == mlx4_master_func_num(dev)) {
+					res_alloc->quota[t] = MLX4_MAX_VLAN_NUM;
+					res_alloc->guaranteed[t] = MLX4_MAX_VLAN_NUM / 2;
+					for (j = 0; j < MLX4_MAX_PORTS; j++)
+						res_alloc->res_port_free[j] =
+							res_alloc->quota[t];
+				} else {
+					res_alloc->quota[t] = MLX4_MAX_VLAN_NUM / 2;
+					res_alloc->guaranteed[t] = 0;
+				}
+				break;
+			case RES_COUNTER:
+				res_alloc->quota[t] = dev->caps.max_counters;
+				res_alloc->guaranteed[t] =
+					mlx4_calc_res_counter_guaranteed(dev, res_alloc, t);
+				break;
+			default:
+				break;
+			}
+			if (i == RES_MAC || i == RES_VLAN) {
+				for (j = 0; j < dev->caps.num_ports; j++)
+					if (test_bit(j, actv_ports.ports))
+						res_alloc->res_port_rsvd[j] +=
+							res_alloc->guaranteed[t];
+			} else {
+				res_alloc->res_reserved += res_alloc->guaranteed[t];
+			}
+		}
+	}
+	spin_lock_init(&priv->mfunc.master.res_tracker.lock);
+	return 0;
+
+no_mem_err:
+	for (i = 0; i < MLX4_NUM_OF_RESOURCE_TYPE; i++) {
+		kfree(priv->mfunc.master.res_tracker.res_alloc[i].allocated);
+		priv->mfunc.master.res_tracker.res_alloc[i].allocated = NULL;
+		kfree(priv->mfunc.master.res_tracker.res_alloc[i].guaranteed);
+		priv->mfunc.master.res_tracker.res_alloc[i].guaranteed = NULL;
+		kfree(priv->mfunc.master.res_tracker.res_alloc[i].quota);
+		priv->mfunc.master.res_tracker.res_alloc[i].quota = NULL;
+	}
+	return -ENOMEM;
+}
+
+void mlx4_free_resource_tracker(struct mlx4_dev *dev,
+				enum mlx4_res_tracker_free_type type)
+{
+	struct mlx4_priv *priv = mlx4_priv(dev);
+	int i;
+
+	if (priv->mfunc.master.res_tracker.slave_list) {
+		if (type != RES_TR_FREE_STRUCTS_ONLY) {
+			for (i = 0; i < dev->num_slaves; i++) {
+				if (type == RES_TR_FREE_ALL ||
+				    dev->caps.function != i)
+					mlx4_delete_all_resources_for_slave(dev, i);
+			}
+			/* free master's vlans */
+			i = dev->caps.function;
+			mlx4_reset_roce_gids(dev, i);
+			mutex_lock(&priv->mfunc.master.res_tracker.slave_list[i].mutex);
+			rem_slave_vlans(dev, i);
+			mutex_unlock(&priv->mfunc.master.res_tracker.slave_list[i].mutex);
+		}
+
+		if (type != RES_TR_FREE_SLAVES_ONLY) {
+			for (i = 0; i < MLX4_NUM_OF_RESOURCE_TYPE; i++) {
+				kfree(priv->mfunc.master.res_tracker.res_alloc[i].allocated);
+				priv->mfunc.master.res_tracker.res_alloc[i].allocated = NULL;
+				kfree(priv->mfunc.master.res_tracker.res_alloc[i].guaranteed);
+				priv->mfunc.master.res_tracker.res_alloc[i].guaranteed = NULL;
+				kfree(priv->mfunc.master.res_tracker.res_alloc[i].quota);
+				priv->mfunc.master.res_tracker.res_alloc[i].quota = NULL;
+			}
+			kfree(priv->mfunc.master.res_tracker.slave_list);
+			priv->mfunc.master.res_tracker.slave_list = NULL;
+		}
+	}
+}
+
+static void update_pkey_index(struct mlx4_dev *dev, int slave,
+			      struct mlx4_cmd_mailbox *inbox)
+{
+	u8 sched = *(u8 *)(inbox->buf + 64);
+	u8 orig_index = *(u8 *)(inbox->buf + 35);
+	u8 new_index;
+	struct mlx4_priv *priv = mlx4_priv(dev);
+	int port;
+
+	port = (sched >> 6 & 1) + 1;
+
+	new_index = priv->virt2phys_pkey[slave][port - 1][orig_index];
+	*(u8 *)(inbox->buf + 35) = new_index;
+}
+
+static void update_gid(struct mlx4_dev *dev, struct mlx4_cmd_mailbox *inbox,
+		       u8 slave)
+{
+	struct mlx4_qp_context	*qp_ctx = inbox->buf + 8;
+	enum mlx4_qp_optpar	optpar = be32_to_cpu(*(__be32 *) inbox->buf);
+	u32			ts = (be32_to_cpu(qp_ctx->flags) >> 16) & 0xff;
+	int port;
+
+	if (MLX4_QP_ST_UD == ts) {
+		port = (qp_ctx->pri_path.sched_queue >> 6 & 1) + 1;
+		if (mlx4_is_eth(dev, port))
+			qp_ctx->pri_path.mgid_index =
+				mlx4_get_base_gid_ix(dev, slave, port) | 0x80;
+		else
+			qp_ctx->pri_path.mgid_index = slave | 0x80;
+
+	} else if (MLX4_QP_ST_RC == ts || MLX4_QP_ST_XRC == ts || MLX4_QP_ST_UC == ts) {
+		if (optpar & MLX4_QP_OPTPAR_PRIMARY_ADDR_PATH) {
+			port = (qp_ctx->pri_path.sched_queue >> 6 & 1) + 1;
+			if (mlx4_is_eth(dev, port)) {
+				qp_ctx->pri_path.mgid_index +=
+					mlx4_get_base_gid_ix(dev, slave, port);
+				qp_ctx->pri_path.mgid_index &= 0x7f;
+			} else {
+				qp_ctx->pri_path.mgid_index = slave & 0x7F;
+			}
+		}
+		if (optpar & MLX4_QP_OPTPAR_ALT_ADDR_PATH) {
+			port = (qp_ctx->alt_path.sched_queue >> 6 & 1) + 1;
+			if (mlx4_is_eth(dev, port)) {
+				qp_ctx->alt_path.mgid_index +=
+					mlx4_get_base_gid_ix(dev, slave, port);
+				qp_ctx->alt_path.mgid_index &= 0x7f;
+			} else {
+				qp_ctx->alt_path.mgid_index = slave & 0x7F;
+			}
+		}
+	}
+}
+
+static int handle_counter(struct mlx4_dev *dev, struct mlx4_qp_context *qpc,
+			  u8 slave, int port);
+
+static int update_vport_qp_param(struct mlx4_dev *dev,
+				 struct mlx4_cmd_mailbox *inbox,
+				 u8 slave, u32 qpn)
+{
+	struct mlx4_qp_context	*qpc = inbox->buf + 8;
+	struct mlx4_vport_oper_state *vp_oper;
+	struct mlx4_priv *priv;
+	u32 qp_type;
+	int port, err = 0;
+
+	port = (qpc->pri_path.sched_queue & 0x40) ? 2 : 1;
+	priv = mlx4_priv(dev);
+	vp_oper = &priv->mfunc.master.vf_oper[slave].vport[port];
+	qp_type	= (be32_to_cpu(qpc->flags) >> 16) & 0xff;
+
+	err = handle_counter(dev, qpc, slave, port);
+	if (err)
+		goto out;
+
+	if (MLX4_VGT != vp_oper->state.default_vlan) {
+		/* the reserved QPs (special, proxy, tunnel)
+		 * do not operate over vlans
+		 */
+		if (mlx4_is_qp_reserved(dev, qpn))
+			return 0;
+
+		/* force strip vlan by clear vsd, MLX QP refers to Raw Ethernet */
+		if (qp_type == MLX4_QP_ST_UD ||
+		    (qp_type == MLX4_QP_ST_MLX && mlx4_is_eth(dev, port))) {
+			if (dev->caps.bmme_flags & MLX4_BMME_FLAG_VSD_INIT2RTR) {
+				*(__be32 *)inbox->buf =
+					cpu_to_be32(be32_to_cpu(*(__be32 *)inbox->buf) |
+					MLX4_QP_OPTPAR_VLAN_STRIPPING);
+				qpc->param3 &= ~cpu_to_be32(MLX4_STRIP_VLAN);
+			} else {
+				struct mlx4_update_qp_params params = {.flags = 0};
+
+				err = mlx4_update_qp(dev, qpn, MLX4_UPDATE_QP_VSD, &params);
+				if (err)
+					goto out;
+			}
+		}
+
+		/* preserve IF_COUNTER flag */
+		qpc->pri_path.vlan_control &=
+			MLX4_CTRL_ETH_SRC_CHECK_IF_COUNTER;
+		if (vp_oper->state.link_state == IFLA_VF_LINK_STATE_DISABLE &&
+		    dev->caps.flags2 & MLX4_DEV_CAP_FLAG2_UPDATE_QP) {
+			qpc->pri_path.vlan_control |=
+				MLX4_VLAN_CTRL_ETH_TX_BLOCK_TAGGED |
+				MLX4_VLAN_CTRL_ETH_TX_BLOCK_PRIO_TAGGED |
+				MLX4_VLAN_CTRL_ETH_TX_BLOCK_UNTAGGED |
+				MLX4_VLAN_CTRL_ETH_RX_BLOCK_PRIO_TAGGED |
+				MLX4_VLAN_CTRL_ETH_RX_BLOCK_UNTAGGED |
+				MLX4_VLAN_CTRL_ETH_RX_BLOCK_TAGGED;
+		} else if (0 != vp_oper->state.default_vlan) {
+			if (vp_oper->state.vlan_proto == htons(ETH_P_8021AD)) {
+				/* vst QinQ should block untagged on TX,
+				 * but cvlan is in payload and phv is set so
+				 * hw see it as untagged. Block tagged instead.
+				 */
+				qpc->pri_path.vlan_control |=
+					MLX4_VLAN_CTRL_ETH_TX_BLOCK_PRIO_TAGGED |
+					MLX4_VLAN_CTRL_ETH_TX_BLOCK_TAGGED |
+					MLX4_VLAN_CTRL_ETH_RX_BLOCK_PRIO_TAGGED |
+					MLX4_VLAN_CTRL_ETH_RX_BLOCK_UNTAGGED;
+			} else { /* vst 802.1Q */
+				qpc->pri_path.vlan_control |=
+					MLX4_VLAN_CTRL_ETH_TX_BLOCK_TAGGED |
+					MLX4_VLAN_CTRL_ETH_RX_BLOCK_PRIO_TAGGED |
+					MLX4_VLAN_CTRL_ETH_RX_BLOCK_UNTAGGED;
+			}
+		} else { /* priority tagged */
+			qpc->pri_path.vlan_control |=
+				MLX4_VLAN_CTRL_ETH_TX_BLOCK_TAGGED |
+				MLX4_VLAN_CTRL_ETH_RX_BLOCK_TAGGED;
+		}
+
+		qpc->pri_path.fvl_rx |= MLX4_FVL_RX_FORCE_ETH_VLAN;
+		qpc->pri_path.vlan_index = vp_oper->vlan_idx;
+		qpc->pri_path.fl |= MLX4_FL_ETH_HIDE_CQE_VLAN;
+		if (vp_oper->state.vlan_proto == htons(ETH_P_8021AD))
+			qpc->pri_path.fl |= MLX4_FL_SV;
+		else
+			qpc->pri_path.fl |= MLX4_FL_CV;
+		qpc->pri_path.feup |= MLX4_FEUP_FORCE_ETH_UP | MLX4_FVL_FORCE_ETH_VLAN;
+		qpc->pri_path.sched_queue &= 0xC7;
+		qpc->pri_path.sched_queue |= (vp_oper->state.default_qos) << 3;
+		qpc->qos_vport = vp_oper->state.qos_vport;
+	}
+	if (vp_oper->state.spoofchk) {
+		qpc->pri_path.feup |= MLX4_FSM_FORCE_ETH_SRC_MAC;
+		qpc->pri_path.grh_mylmc = (0x80 & qpc->pri_path.grh_mylmc) + vp_oper->mac_idx;
+	}
+out:
+	return err;
+}
+
+static int mpt_mask(struct mlx4_dev *dev)
+{
+	return dev->caps.num_mpts - 1;
+}
+
+static const char *mlx4_resource_type_to_str(enum mlx4_resource t)
+{
+	switch (t) {
+	case RES_QP:
+		return "QP";
+	case RES_CQ:
+		return "CQ";
+	case RES_SRQ:
+		return "SRQ";
+	case RES_XRCD:
+		return "XRCD";
+	case RES_MPT:
+		return "MPT";
+	case RES_MTT:
+		return "MTT";
+	case RES_MAC:
+		return "MAC";
+	case RES_VLAN:
+		return "VLAN";
+	case RES_COUNTER:
+		return "COUNTER";
+	case RES_FS_RULE:
+		return "FS_RULE";
+	case RES_EQ:
+		return "EQ";
+	default:
+		return "INVALID RESOURCE";
+	}
+}
+
+static void *find_res(struct mlx4_dev *dev, u64 res_id,
+		      enum mlx4_resource type)
+{
+	struct mlx4_priv *priv = mlx4_priv(dev);
+
+	return res_tracker_lookup(&priv->mfunc.master.res_tracker.res_tree[type],
+				  res_id);
+}
+
+static int _get_res(struct mlx4_dev *dev, int slave, u64 res_id,
+		    enum mlx4_resource type,
+		    void *res, const char *func_name)
+{
+	struct res_common *r;
+	int err = 0;
+
+	spin_lock_irq(mlx4_tlock(dev));
+	r = find_res(dev, res_id, type);
+	if (!r) {
+		err = -ENONET;
+		goto exit;
+	}
+
+	if (r->state == RES_ANY_BUSY) {
+		mlx4_warn(dev,
+			  "%s(%d) trying to get resource %llx of type %s, but it's already taken by %s\n",
+			  func_name, slave, res_id, mlx4_resource_type_to_str(type),
+			  r->func_name);
+		err = -EBUSY;
+		goto exit;
+	}
+
+	if (r->owner != slave) {
+		err = -EPERM;
+		goto exit;
+	}
+
+	r->from_state = r->state;
+	r->state = RES_ANY_BUSY;
+	r->func_name = func_name;
+
+	if (res)
+		*((struct res_common **)res) = r;
+
+exit:
+	spin_unlock_irq(mlx4_tlock(dev));
+	return err;
+}
+
+#define get_res(dev, slave, res_id, type, res) \
+	_get_res((dev), (slave), (res_id), (type), (res), __func__)
+
+int mlx4_get_slave_from_resource_id(struct mlx4_dev *dev,
+				    enum mlx4_resource type,
+				    u64 res_id, int *slave)
+{
+
+	struct res_common *r;
+	int err = -ENOENT;
+	int id = res_id;
+
+	if (type == RES_QP)
+		id &= 0x7fffff;
+	spin_lock(mlx4_tlock(dev));
+
+	r = find_res(dev, id, type);
+	if (r) {
+		*slave = r->owner;
+		err = 0;
+	}
+	spin_unlock(mlx4_tlock(dev));
+
+	return err;
+}
+
+static void put_res(struct mlx4_dev *dev, int slave, u64 res_id,
+		    enum mlx4_resource type)
+{
+	struct res_common *r;
+
+	spin_lock_irq(mlx4_tlock(dev));
+	r = find_res(dev, res_id, type);
+	if (r) {
+		r->state = r->from_state;
+		r->func_name = "";
+	}
+	spin_unlock_irq(mlx4_tlock(dev));
+}
+
+static int counter_alloc_res(struct mlx4_dev *dev, int slave, int op, int cmd,
+			     u64 in_param, u64 *out_param, int port);
+
+static int handle_existing_counter(struct mlx4_dev *dev, u8 slave, int port,
+				   int counter_index)
+{
+	struct res_common *r;
+	struct res_counter *counter;
+	int ret = 0;
+
+	if (counter_index == MLX4_SINK_COUNTER_INDEX(dev))
+		return ret;
+
+	spin_lock_irq(mlx4_tlock(dev));
+	r = find_res(dev, counter_index, RES_COUNTER);
+	if (!r || r->owner != slave) {
+		ret = -EINVAL;
+	} else {
+		counter = container_of(r, struct res_counter, com);
+		if (!counter->port)
+			counter->port = port;
+	}
+
+	spin_unlock_irq(mlx4_tlock(dev));
+	return ret;
+}
+
+static int handle_unexisting_counter(struct mlx4_dev *dev,
+				     struct mlx4_qp_context *qpc, u8 slave,
+				     int port)
+{
+	struct mlx4_priv *priv = mlx4_priv(dev);
+	struct mlx4_resource_tracker *tracker = &priv->mfunc.master.res_tracker;
+	struct res_common *tmp;
+	struct res_counter *counter;
+	u64 counter_idx = MLX4_SINK_COUNTER_INDEX(dev);
+	int err = 0;
+
+	spin_lock_irq(mlx4_tlock(dev));
+	list_for_each_entry(tmp,
+			    &tracker->slave_list[slave].res_list[RES_COUNTER],
+			    list) {
+		counter = container_of(tmp, struct res_counter, com);
+		if (port == counter->port) {
+			qpc->pri_path.counter_index  = counter->com.res_id;
+			spin_unlock_irq(mlx4_tlock(dev));
+			return 0;
+		}
+	}
+	spin_unlock_irq(mlx4_tlock(dev));
+
+	/* No existing counter, need to allocate a new counter */
+	err = counter_alloc_res(dev, slave, RES_OP_RESERVE, 0, 0, &counter_idx,
+				port);
+	if (err == -ENOENT) {
+		err = 0;
+	} else if (err && err != -ENOSPC) {
+		mlx4_err(dev, "%s: failed to create new counter for slave %d err %d\n",
+			 __func__, slave, err);
+	} else {
+		qpc->pri_path.counter_index = counter_idx;
+		mlx4_dbg(dev, "%s: alloc new counter for slave %d index %d\n",
+			 __func__, slave, qpc->pri_path.counter_index);
+		err = 0;
+	}
+
+	return err;
+}
+
+static int handle_counter(struct mlx4_dev *dev, struct mlx4_qp_context *qpc,
+			  u8 slave, int port)
+{
+	if (qpc->pri_path.counter_index != MLX4_SINK_COUNTER_INDEX(dev))
+		return handle_existing_counter(dev, slave, port,
+					       qpc->pri_path.counter_index);
+
+	return handle_unexisting_counter(dev, qpc, slave, port);
+}
+
+static struct res_common *alloc_qp_tr(int id)
+{
+	struct res_qp *ret;
+
+	ret = kzalloc(sizeof(*ret), GFP_KERNEL);
+	if (!ret)
+		return NULL;
+
+	ret->com.res_id = id;
+	ret->com.state = RES_QP_RESERVED;
+	ret->local_qpn = id;
+	INIT_LIST_HEAD(&ret->mcg_list);
+	spin_lock_init(&ret->mcg_spl);
+	atomic_set(&ret->ref_count, 0);
+
+	return &ret->com;
+}
+
+static struct res_common *alloc_mtt_tr(int id, int order)
+{
+	struct res_mtt *ret;
+
+	ret = kzalloc(sizeof(*ret), GFP_KERNEL);
+	if (!ret)
+		return NULL;
+
+	ret->com.res_id = id;
+	ret->order = order;
+	ret->com.state = RES_MTT_ALLOCATED;
+	atomic_set(&ret->ref_count, 0);
+
+	return &ret->com;
+}
+
+static struct res_common *alloc_mpt_tr(int id, int key)
+{
+	struct res_mpt *ret;
+
+	ret = kzalloc(sizeof(*ret), GFP_KERNEL);
+	if (!ret)
+		return NULL;
+
+	ret->com.res_id = id;
+	ret->com.state = RES_MPT_RESERVED;
+	ret->key = key;
+
+	return &ret->com;
+}
+
+static struct res_common *alloc_eq_tr(int id)
+{
+	struct res_eq *ret;
+
+	ret = kzalloc(sizeof(*ret), GFP_KERNEL);
+	if (!ret)
+		return NULL;
+
+	ret->com.res_id = id;
+	ret->com.state = RES_EQ_RESERVED;
+
+	return &ret->com;
+}
+
+static struct res_common *alloc_cq_tr(int id)
+{
+	struct res_cq *ret;
+
+	ret = kzalloc(sizeof(*ret), GFP_KERNEL);
+	if (!ret)
+		return NULL;
+
+	ret->com.res_id = id;
+	ret->com.state = RES_CQ_ALLOCATED;
+	atomic_set(&ret->ref_count, 0);
+
+	return &ret->com;
+}
+
+static struct res_common *alloc_srq_tr(int id)
+{
+	struct res_srq *ret;
+
+	ret = kzalloc(sizeof(*ret), GFP_KERNEL);
+	if (!ret)
+		return NULL;
+
+	ret->com.res_id = id;
+	ret->com.state = RES_SRQ_ALLOCATED;
+	atomic_set(&ret->ref_count, 0);
+
+	return &ret->com;
+}
+
+static struct res_common *alloc_counter_tr(int id, int port)
+{
+	struct res_counter *ret;
+
+	ret = kzalloc(sizeof(*ret), GFP_KERNEL);
+	if (!ret)
+		return NULL;
+
+	ret->com.res_id = id;
+	ret->com.state = RES_COUNTER_ALLOCATED;
+	ret->port = port;
+
+	return &ret->com;
+}
+
+static struct res_common *alloc_xrcdn_tr(int id)
+{
+	struct res_xrcdn *ret;
+
+	ret = kzalloc(sizeof(*ret), GFP_KERNEL);
+	if (!ret)
+		return NULL;
+
+	ret->com.res_id = id;
+	ret->com.state = RES_XRCD_ALLOCATED;
+
+	return &ret->com;
+}
+
+static struct res_common *alloc_fs_rule_tr(u64 id, int qpn)
+{
+	struct res_fs_rule *ret;
+
+	ret = kzalloc(sizeof(*ret), GFP_KERNEL);
+	if (!ret)
+		return NULL;
+
+	ret->com.res_id = id;
+	ret->com.state = RES_FS_RULE_ALLOCATED;
+	ret->qpn = qpn;
+	return &ret->com;
+}
+
+static struct res_common *alloc_tr(u64 id, enum mlx4_resource type, int slave,
+				   int extra)
+{
+	struct res_common *ret;
+
+	switch (type) {
+	case RES_QP:
+		ret = alloc_qp_tr(id);
+		break;
+	case RES_MPT:
+		ret = alloc_mpt_tr(id, extra);
+		break;
+	case RES_MTT:
+		ret = alloc_mtt_tr(id, extra);
+		break;
+	case RES_EQ:
+		ret = alloc_eq_tr(id);
+		break;
+	case RES_CQ:
+		ret = alloc_cq_tr(id);
+		break;
+	case RES_SRQ:
+		ret = alloc_srq_tr(id);
+		break;
+	case RES_MAC:
+		pr_err("implementation missing\n");
+		return NULL;
+	case RES_COUNTER:
+		ret = alloc_counter_tr(id, extra);
+		break;
+	case RES_XRCD:
+		ret = alloc_xrcdn_tr(id);
+		break;
+	case RES_FS_RULE:
+		ret = alloc_fs_rule_tr(id, extra);
+		break;
+	default:
+		return NULL;
+	}
+	if (ret)
+		ret->owner = slave;
+
+	return ret;
+}
+
+int mlx4_calc_vf_counters(struct mlx4_dev *dev, int slave, int port,
+			  struct mlx4_counter *data)
+{
+	struct mlx4_priv *priv = mlx4_priv(dev);
+	struct mlx4_resource_tracker *tracker = &priv->mfunc.master.res_tracker;
+	struct res_common *tmp;
+	struct res_counter *counter;
+	int *counters_arr;
+	int i = 0, err = 0;
+
+	memset(data, 0, sizeof(*data));
+
+	counters_arr = kmalloc_array(dev->caps.max_counters,
+				     sizeof(*counters_arr), GFP_KERNEL);
+	if (!counters_arr)
+		return -ENOMEM;
+
+	spin_lock_irq(mlx4_tlock(dev));
+	list_for_each_entry(tmp,
+			    &tracker->slave_list[slave].res_list[RES_COUNTER],
+			    list) {
+		counter = container_of(tmp, struct res_counter, com);
+		if (counter->port == port) {
+			counters_arr[i] = (int)tmp->res_id;
+			i++;
+		}
+	}
+	spin_unlock_irq(mlx4_tlock(dev));
+	counters_arr[i] = -1;
+
+	i = 0;
+
+	while (counters_arr[i] != -1) {
+		err = mlx4_get_counter_stats(dev, counters_arr[i], data,
+					     0);
+		if (err) {
+			memset(data, 0, sizeof(*data));
+			goto table_changed;
+		}
+		i++;
+	}
+
+table_changed:
+	kfree(counters_arr);
+	return 0;
+}
+
+static int add_res_range(struct mlx4_dev *dev, int slave, u64 base, int count,
+			 enum mlx4_resource type, int extra)
+{
+	int i;
+	int err;
+	struct mlx4_priv *priv = mlx4_priv(dev);
+	struct res_common **res_arr;
+	struct mlx4_resource_tracker *tracker = &priv->mfunc.master.res_tracker;
+	struct rb_root *root = &tracker->res_tree[type];
+
+	res_arr = kcalloc(count, sizeof(*res_arr), GFP_KERNEL);
+	if (!res_arr)
+		return -ENOMEM;
+
+	for (i = 0; i < count; ++i) {
+		res_arr[i] = alloc_tr(base + i, type, slave, extra);
+		if (!res_arr[i]) {
+			for (--i; i >= 0; --i)
+				kfree(res_arr[i]);
+
+			kfree(res_arr);
+			return -ENOMEM;
+		}
+	}
+
+	spin_lock_irq(mlx4_tlock(dev));
+	for (i = 0; i < count; ++i) {
+		if (find_res(dev, base + i, type)) {
+			err = -EEXIST;
+			goto undo;
+		}
+		err = res_tracker_insert(root, res_arr[i]);
+		if (err)
+			goto undo;
+		list_add_tail(&res_arr[i]->list,
+			      &tracker->slave_list[slave].res_list[type]);
+	}
+	spin_unlock_irq(mlx4_tlock(dev));
+	kfree(res_arr);
+
+	return 0;
+
+undo:
+	for (--i; i >= 0; --i) {
+		rb_erase(&res_arr[i]->node, root);
+		list_del_init(&res_arr[i]->list);
+	}
+
+	spin_unlock_irq(mlx4_tlock(dev));
+
+	for (i = 0; i < count; ++i)
+		kfree(res_arr[i]);
+
+	kfree(res_arr);
+
+	return err;
+}
+
+static int remove_qp_ok(struct res_qp *res)
+{
+	if (res->com.state == RES_QP_BUSY || atomic_read(&res->ref_count) ||
+	    !list_empty(&res->mcg_list)) {
+		pr_err("resource tracker: fail to remove qp, state %d, ref_count %d\n",
+		       res->com.state, atomic_read(&res->ref_count));
+		return -EBUSY;
+	} else if (res->com.state != RES_QP_RESERVED) {
+		return -EPERM;
+	}
+
+	return 0;
+}
+
+static int remove_mtt_ok(struct res_mtt *res, int order)
+{
+	if (res->com.state == RES_MTT_BUSY ||
+	    atomic_read(&res->ref_count)) {
+		pr_devel("%s-%d: state %s, ref_count %d\n",
+			 __func__, __LINE__,
+			 mtt_states_str(res->com.state),
+			 atomic_read(&res->ref_count));
+		return -EBUSY;
+	} else if (res->com.state != RES_MTT_ALLOCATED)
+		return -EPERM;
+	else if (res->order != order)
+		return -EINVAL;
+
+	return 0;
+}
+
+static int remove_mpt_ok(struct res_mpt *res)
+{
+	if (res->com.state == RES_MPT_BUSY)
+		return -EBUSY;
+	else if (res->com.state != RES_MPT_RESERVED)
+		return -EPERM;
+
+	return 0;
+}
+
+static int remove_eq_ok(struct res_eq *res)
+{
+	if (res->com.state == RES_MPT_BUSY)
+		return -EBUSY;
+	else if (res->com.state != RES_MPT_RESERVED)
+		return -EPERM;
+
+	return 0;
+}
+
+static int remove_counter_ok(struct res_counter *res)
+{
+	if (res->com.state == RES_COUNTER_BUSY)
+		return -EBUSY;
+	else if (res->com.state != RES_COUNTER_ALLOCATED)
+		return -EPERM;
+
+	return 0;
+}
+
+static int remove_xrcdn_ok(struct res_xrcdn *res)
+{
+	if (res->com.state == RES_XRCD_BUSY)
+		return -EBUSY;
+	else if (res->com.state != RES_XRCD_ALLOCATED)
+		return -EPERM;
+
+	return 0;
+}
+
+static int remove_fs_rule_ok(struct res_fs_rule *res)
+{
+	if (res->com.state == RES_FS_RULE_BUSY)
+		return -EBUSY;
+	else if (res->com.state != RES_FS_RULE_ALLOCATED)
+		return -EPERM;
+
+	return 0;
+}
+
+static int remove_cq_ok(struct res_cq *res)
+{
+	if (res->com.state == RES_CQ_BUSY)
+		return -EBUSY;
+	else if (res->com.state != RES_CQ_ALLOCATED)
+		return -EPERM;
+
+	return 0;
+}
+
+static int remove_srq_ok(struct res_srq *res)
+{
+	if (res->com.state == RES_SRQ_BUSY)
+		return -EBUSY;
+	else if (res->com.state != RES_SRQ_ALLOCATED)
+		return -EPERM;
+
+	return 0;
+}
+
+static int remove_ok(struct res_common *res, enum mlx4_resource type, int extra)
+{
+	switch (type) {
+	case RES_QP:
+		return remove_qp_ok((struct res_qp *)res);
+	case RES_CQ:
+		return remove_cq_ok((struct res_cq *)res);
+	case RES_SRQ:
+		return remove_srq_ok((struct res_srq *)res);
+	case RES_MPT:
+		return remove_mpt_ok((struct res_mpt *)res);
+	case RES_MTT:
+		return remove_mtt_ok((struct res_mtt *)res, extra);
+	case RES_MAC:
+		return -EOPNOTSUPP;
+	case RES_EQ:
+		return remove_eq_ok((struct res_eq *)res);
+	case RES_COUNTER:
+		return remove_counter_ok((struct res_counter *)res);
+	case RES_XRCD:
+		return remove_xrcdn_ok((struct res_xrcdn *)res);
+	case RES_FS_RULE:
+		return remove_fs_rule_ok((struct res_fs_rule *)res);
+	default:
+		return -EINVAL;
+	}
+}
+
+static int rem_res_range(struct mlx4_dev *dev, int slave, u64 base, int count,
+			 enum mlx4_resource type, int extra)
+{
+	u64 i;
+	int err;
+	struct mlx4_priv *priv = mlx4_priv(dev);
+	struct mlx4_resource_tracker *tracker = &priv->mfunc.master.res_tracker;
+	struct res_common *r;
+
+	spin_lock_irq(mlx4_tlock(dev));
+	for (i = base; i < base + count; ++i) {
+		r = res_tracker_lookup(&tracker->res_tree[type], i);
+		if (!r) {
+			err = -ENOENT;
+			goto out;
+		}
+		if (r->owner != slave) {
+			err = -EPERM;
+			goto out;
+		}
+		err = remove_ok(r, type, extra);
+		if (err)
+			goto out;
+	}
+
+	for (i = base; i < base + count; ++i) {
+		r = res_tracker_lookup(&tracker->res_tree[type], i);
+		rb_erase(&r->node, &tracker->res_tree[type]);
+		list_del(&r->list);
+		kfree(r);
+	}
+	err = 0;
+
+out:
+	spin_unlock_irq(mlx4_tlock(dev));
+
+	return err;
+}
+
+static int qp_res_start_move_to(struct mlx4_dev *dev, int slave, int qpn,
+				enum res_qp_states state, struct res_qp **qp,
+				int alloc)
+{
+	struct mlx4_priv *priv = mlx4_priv(dev);
+	struct mlx4_resource_tracker *tracker = &priv->mfunc.master.res_tracker;
+	struct res_qp *r;
+	int err = 0;
+
+	spin_lock_irq(mlx4_tlock(dev));
+	r = res_tracker_lookup(&tracker->res_tree[RES_QP], qpn);
+	if (!r)
+		err = -ENOENT;
+	else if (r->com.owner != slave)
+		err = -EPERM;
+	else {
+		switch (state) {
+		case RES_QP_BUSY:
+			mlx4_dbg(dev, "%s: failed RES_QP, 0x%llx\n",
+				 __func__, r->com.res_id);
+			err = -EBUSY;
+			break;
+
+		case RES_QP_RESERVED:
+			if (r->com.state == RES_QP_MAPPED && !alloc)
+				break;
+
+			mlx4_dbg(dev, "failed RES_QP, 0x%llx\n", r->com.res_id);
+			err = -EINVAL;
+			break;
+
+		case RES_QP_MAPPED:
+			if ((r->com.state == RES_QP_RESERVED && alloc) ||
+			    r->com.state == RES_QP_HW)
+				break;
+			else {
+				mlx4_dbg(dev, "failed RES_QP, 0x%llx\n",
+					  r->com.res_id);
+				err = -EINVAL;
+			}
+
+			break;
+
+		case RES_QP_HW:
+			if (r->com.state != RES_QP_MAPPED)
+				err = -EINVAL;
+			break;
+		default:
+			err = -EINVAL;
+		}
+
+		if (!err) {
+			r->com.from_state = r->com.state;
+			r->com.to_state = state;
+			r->com.state = RES_QP_BUSY;
+			if (qp)
+				*qp = r;
+		}
+	}
+
+	spin_unlock_irq(mlx4_tlock(dev));
+
+	return err;
+}
+
+static int mr_res_start_move_to(struct mlx4_dev *dev, int slave, int index,
+				enum res_mpt_states state, struct res_mpt **mpt)
+{
+	struct mlx4_priv *priv = mlx4_priv(dev);
+	struct mlx4_resource_tracker *tracker = &priv->mfunc.master.res_tracker;
+	struct res_mpt *r;
+	int err = 0;
+
+	spin_lock_irq(mlx4_tlock(dev));
+	r = res_tracker_lookup(&tracker->res_tree[RES_MPT], index);
+	if (!r)
+		err = -ENOENT;
+	else if (r->com.owner != slave)
+		err = -EPERM;
+	else {
+		switch (state) {
+		case RES_MPT_BUSY:
+			err = -EINVAL;
+			break;
+
+		case RES_MPT_RESERVED:
+			if (r->com.state != RES_MPT_MAPPED)
+				err = -EINVAL;
+			break;
+
+		case RES_MPT_MAPPED:
+			if (r->com.state != RES_MPT_RESERVED &&
+			    r->com.state != RES_MPT_HW)
+				err = -EINVAL;
+			break;
+
+		case RES_MPT_HW:
+			if (r->com.state != RES_MPT_MAPPED)
+				err = -EINVAL;
+			break;
+		default:
+			err = -EINVAL;
+		}
+
+		if (!err) {
+			r->com.from_state = r->com.state;
+			r->com.to_state = state;
+			r->com.state = RES_MPT_BUSY;
+			if (mpt)
+				*mpt = r;
+		}
+	}
+
+	spin_unlock_irq(mlx4_tlock(dev));
+
+	return err;
+}
+
+static int eq_res_start_move_to(struct mlx4_dev *dev, int slave, int index,
+				enum res_eq_states state, struct res_eq **eq)
+{
+	struct mlx4_priv *priv = mlx4_priv(dev);
+	struct mlx4_resource_tracker *tracker = &priv->mfunc.master.res_tracker;
+	struct res_eq *r;
+	int err = 0;
+
+	spin_lock_irq(mlx4_tlock(dev));
+	r = res_tracker_lookup(&tracker->res_tree[RES_EQ], index);
+	if (!r)
+		err = -ENOENT;
+	else if (r->com.owner != slave)
+		err = -EPERM;
+	else {
+		switch (state) {
+		case RES_EQ_BUSY:
+			err = -EINVAL;
+			break;
+
+		case RES_EQ_RESERVED:
+			if (r->com.state != RES_EQ_HW)
+				err = -EINVAL;
+			break;
+
+		case RES_EQ_HW:
+			if (r->com.state != RES_EQ_RESERVED)
+				err = -EINVAL;
+			break;
+
+		default:
+			err = -EINVAL;
+		}
+
+		if (!err) {
+			r->com.from_state = r->com.state;
+			r->com.to_state = state;
+			r->com.state = RES_EQ_BUSY;
+		}
+	}
+
+	spin_unlock_irq(mlx4_tlock(dev));
+
+	if (!err && eq)
+		*eq = r;
+
+	return err;
+}
+
+static int cq_res_start_move_to(struct mlx4_dev *dev, int slave, int cqn,
+				enum res_cq_states state, struct res_cq **cq)
+{
+	struct mlx4_priv *priv = mlx4_priv(dev);
+	struct mlx4_resource_tracker *tracker = &priv->mfunc.master.res_tracker;
+	struct res_cq *r;
+	int err;
+
+	spin_lock_irq(mlx4_tlock(dev));
+	r = res_tracker_lookup(&tracker->res_tree[RES_CQ], cqn);
+	if (!r) {
+		err = -ENOENT;
+	} else if (r->com.owner != slave) {
+		err = -EPERM;
+	} else if (state == RES_CQ_ALLOCATED) {
+		if (r->com.state != RES_CQ_HW)
+			err = -EINVAL;
+		else if (atomic_read(&r->ref_count))
+			err = -EBUSY;
+		else
+			err = 0;
+	} else if (state != RES_CQ_HW || r->com.state != RES_CQ_ALLOCATED) {
+		err = -EINVAL;
+	} else {
+		err = 0;
+	}
+
+	if (!err) {
+		r->com.from_state = r->com.state;
+		r->com.to_state = state;
+		r->com.state = RES_CQ_BUSY;
+		if (cq)
+			*cq = r;
+	}
+
+	spin_unlock_irq(mlx4_tlock(dev));
+
+	return err;
+}
+
+static int srq_res_start_move_to(struct mlx4_dev *dev, int slave, int index,
+				 enum res_srq_states state, struct res_srq **srq)
+{
+	struct mlx4_priv *priv = mlx4_priv(dev);
+	struct mlx4_resource_tracker *tracker = &priv->mfunc.master.res_tracker;
+	struct res_srq *r;
+	int err = 0;
+
+	spin_lock_irq(mlx4_tlock(dev));
+	r = res_tracker_lookup(&tracker->res_tree[RES_SRQ], index);
+	if (!r) {
+		err = -ENOENT;
+	} else if (r->com.owner != slave) {
+		err = -EPERM;
+	} else if (state == RES_SRQ_ALLOCATED) {
+		if (r->com.state != RES_SRQ_HW)
+			err = -EINVAL;
+		else if (atomic_read(&r->ref_count))
+			err = -EBUSY;
+	} else if (state != RES_SRQ_HW || r->com.state != RES_SRQ_ALLOCATED) {
+		err = -EINVAL;
+	}
+
+	if (!err) {
+		r->com.from_state = r->com.state;
+		r->com.to_state = state;
+		r->com.state = RES_SRQ_BUSY;
+		if (srq)
+			*srq = r;
+	}
+
+	spin_unlock_irq(mlx4_tlock(dev));
+
+	return err;
+}
+
+static void res_abort_move(struct mlx4_dev *dev, int slave,
+			   enum mlx4_resource type, int id)
+{
+	struct mlx4_priv *priv = mlx4_priv(dev);
+	struct mlx4_resource_tracker *tracker = &priv->mfunc.master.res_tracker;
+	struct res_common *r;
+
+	spin_lock_irq(mlx4_tlock(dev));
+	r = res_tracker_lookup(&tracker->res_tree[type], id);
+	if (r && (r->owner == slave))
+		r->state = r->from_state;
+	spin_unlock_irq(mlx4_tlock(dev));
+}
+
+static void res_end_move(struct mlx4_dev *dev, int slave,
+			 enum mlx4_resource type, int id)
+{
+	struct mlx4_priv *priv = mlx4_priv(dev);
+	struct mlx4_resource_tracker *tracker = &priv->mfunc.master.res_tracker;
+	struct res_common *r;
+
+	spin_lock_irq(mlx4_tlock(dev));
+	r = res_tracker_lookup(&tracker->res_tree[type], id);
+	if (r && (r->owner == slave))
+		r->state = r->to_state;
+	spin_unlock_irq(mlx4_tlock(dev));
+}
+
+static int valid_reserved(struct mlx4_dev *dev, int slave, int qpn)
+{
+	return mlx4_is_qp_reserved(dev, qpn) &&
+		(mlx4_is_master(dev) || mlx4_is_guest_proxy(dev, slave, qpn));
+}
+
+static int fw_reserved(struct mlx4_dev *dev, int qpn)
+{
+	return qpn < dev->caps.reserved_qps_cnt[MLX4_QP_REGION_FW];
+}
+
+static int qp_alloc_res(struct mlx4_dev *dev, int slave, int op, int cmd,
+			u64 in_param, u64 *out_param)
+{
+	int err;
+	int count;
+	int align;
+	int base;
+	int qpn;
+	u8 flags;
+
+	switch (op) {
+	case RES_OP_RESERVE:
+		count = get_param_l(&in_param) & 0xffffff;
+		/* Turn off all unsupported QP allocation flags that the
+		 * slave tries to set.
+		 */
+		flags = (get_param_l(&in_param) >> 24) & dev->caps.alloc_res_qp_mask;
+		align = get_param_h(&in_param);
+		err = mlx4_grant_resource(dev, slave, RES_QP, count, 0);
+		if (err)
+			return err;
+
+		err = __mlx4_qp_reserve_range(dev, count, align, &base, flags);
+		if (err) {
+			mlx4_release_resource(dev, slave, RES_QP, count, 0);
+			return err;
+		}
+
+		err = add_res_range(dev, slave, base, count, RES_QP, 0);
+		if (err) {
+			mlx4_release_resource(dev, slave, RES_QP, count, 0);
+			__mlx4_qp_release_range(dev, base, count);
+			return err;
+		}
+		set_param_l(out_param, base);
+		break;
+	case RES_OP_MAP_ICM:
+		qpn = get_param_l(&in_param) & 0x7fffff;
+		if (valid_reserved(dev, slave, qpn)) {
+			err = add_res_range(dev, slave, qpn, 1, RES_QP, 0);
+			if (err)
+				return err;
+		}
+
+		err = qp_res_start_move_to(dev, slave, qpn, RES_QP_MAPPED,
+					   NULL, 1);
+		if (err)
+			return err;
+
+		if (!fw_reserved(dev, qpn)) {
+			err = __mlx4_qp_alloc_icm(dev, qpn);
+			if (err) {
+				res_abort_move(dev, slave, RES_QP, qpn);
+				return err;
+			}
+		}
+
+		res_end_move(dev, slave, RES_QP, qpn);
+		break;
+
+	default:
+		err = -EINVAL;
+		break;
+	}
+	return err;
+}
+
+static int mtt_alloc_res(struct mlx4_dev *dev, int slave, int op, int cmd,
+			 u64 in_param, u64 *out_param)
+{
+	int err = -EINVAL;
+	int base;
+	int order;
+
+	if (op != RES_OP_RESERVE_AND_MAP)
+		return err;
+
+	order = get_param_l(&in_param);
+
+	err = mlx4_grant_resource(dev, slave, RES_MTT, 1 << order, 0);
+	if (err)
+		return err;
+
+	base = __mlx4_alloc_mtt_range(dev, order);
+	if (base == -1) {
+		mlx4_release_resource(dev, slave, RES_MTT, 1 << order, 0);
+		return -ENOMEM;
+	}
+
+	err = add_res_range(dev, slave, base, 1, RES_MTT, order);
+	if (err) {
+		mlx4_release_resource(dev, slave, RES_MTT, 1 << order, 0);
+		__mlx4_free_mtt_range(dev, base, order);
+	} else {
+		set_param_l(out_param, base);
+	}
+
+	return err;
+}
+
+static int mpt_alloc_res(struct mlx4_dev *dev, int slave, int op, int cmd,
+			 u64 in_param, u64 *out_param)
+{
+	int err = -EINVAL;
+	int index;
+	int id;
+	struct res_mpt *mpt;
+
+	switch (op) {
+	case RES_OP_RESERVE:
+		err = mlx4_grant_resource(dev, slave, RES_MPT, 1, 0);
+		if (err)
+			break;
+
+		index = __mlx4_mpt_reserve(dev);
+		if (index == -1) {
+			mlx4_release_resource(dev, slave, RES_MPT, 1, 0);
+			break;
+		}
+		id = index & mpt_mask(dev);
+
+		err = add_res_range(dev, slave, id, 1, RES_MPT, index);
+		if (err) {
+			mlx4_release_resource(dev, slave, RES_MPT, 1, 0);
+			__mlx4_mpt_release(dev, index);
+			break;
+		}
+		set_param_l(out_param, index);
+		break;
+	case RES_OP_MAP_ICM:
+		index = get_param_l(&in_param);
+		id = index & mpt_mask(dev);
+		err = mr_res_start_move_to(dev, slave, id,
+					   RES_MPT_MAPPED, &mpt);
+		if (err)
+			return err;
+
+		err = __mlx4_mpt_alloc_icm(dev, mpt->key);
+		if (err) {
+			res_abort_move(dev, slave, RES_MPT, id);
+			return err;
+		}
+
+		res_end_move(dev, slave, RES_MPT, id);
+		break;
+	}
+	return err;
+}
+
+static int cq_alloc_res(struct mlx4_dev *dev, int slave, int op, int cmd,
+			u64 in_param, u64 *out_param)
+{
+	int cqn;
+	int err;
+
+	switch (op) {
+	case RES_OP_RESERVE_AND_MAP:
+		err = mlx4_grant_resource(dev, slave, RES_CQ, 1, 0);
+		if (err)
+			break;
+
+		err = __mlx4_cq_alloc_icm(dev, &cqn);
+		if (err) {
+			mlx4_release_resource(dev, slave, RES_CQ, 1, 0);
+			break;
+		}
+
+		err = add_res_range(dev, slave, cqn, 1, RES_CQ, 0);
+		if (err) {
+			mlx4_release_resource(dev, slave, RES_CQ, 1, 0);
+			__mlx4_cq_free_icm(dev, cqn);
+			break;
+		}
+
+		set_param_l(out_param, cqn);
+		break;
+
+	default:
+		err = -EINVAL;
+	}
+
+	return err;
+}
+
+static int srq_alloc_res(struct mlx4_dev *dev, int slave, int op, int cmd,
+			 u64 in_param, u64 *out_param)
+{
+	int srqn;
+	int err;
+
+	switch (op) {
+	case RES_OP_RESERVE_AND_MAP:
+		err = mlx4_grant_resource(dev, slave, RES_SRQ, 1, 0);
+		if (err)
+			break;
+
+		err = __mlx4_srq_alloc_icm(dev, &srqn);
+		if (err) {
+			mlx4_release_resource(dev, slave, RES_SRQ, 1, 0);
+			break;
+		}
+
+		err = add_res_range(dev, slave, srqn, 1, RES_SRQ, 0);
+		if (err) {
+			mlx4_release_resource(dev, slave, RES_SRQ, 1, 0);
+			__mlx4_srq_free_icm(dev, srqn);
+			break;
+		}
+
+		set_param_l(out_param, srqn);
+		break;
+
+	default:
+		err = -EINVAL;
+	}
+
+	return err;
+}
+
+static int mac_find_smac_ix_in_slave(struct mlx4_dev *dev, int slave, int port,
+				     u8 smac_index, u64 *mac)
+{
+	struct mlx4_priv *priv = mlx4_priv(dev);
+	struct mlx4_resource_tracker *tracker = &priv->mfunc.master.res_tracker;
+	struct list_head *mac_list =
+		&tracker->slave_list[slave].res_list[RES_MAC];
+	struct mac_res *res, *tmp;
+
+	list_for_each_entry_safe(res, tmp, mac_list, list) {
+		if (res->smac_index == smac_index && res->port == (u8) port) {
+			*mac = res->mac;
+			return 0;
+		}
+	}
+	return -ENOENT;
+}
+
+static int mac_add_to_slave(struct mlx4_dev *dev, int slave, u64 mac, int port, u8 smac_index)
+{
+	struct mlx4_priv *priv = mlx4_priv(dev);
+	struct mlx4_resource_tracker *tracker = &priv->mfunc.master.res_tracker;
+	struct list_head *mac_list =
+		&tracker->slave_list[slave].res_list[RES_MAC];
+	struct mac_res *res, *tmp;
+
+	list_for_each_entry_safe(res, tmp, mac_list, list) {
+		if (res->mac == mac && res->port == (u8) port) {
+			/* mac found. update ref count */
+			++res->ref_count;
+			return 0;
+		}
+	}
+
+	if (mlx4_grant_resource(dev, slave, RES_MAC, 1, port))
+		return -EINVAL;
+	res = kzalloc(sizeof(*res), GFP_KERNEL);
+	if (!res) {
+		mlx4_release_resource(dev, slave, RES_MAC, 1, port);
+		return -ENOMEM;
+	}
+	res->mac = mac;
+	res->port = (u8) port;
+	res->smac_index = smac_index;
+	res->ref_count = 1;
+	list_add_tail(&res->list,
+		      &tracker->slave_list[slave].res_list[RES_MAC]);
+	return 0;
+}
+
+static void mac_del_from_slave(struct mlx4_dev *dev, int slave, u64 mac,
+			       int port)
+{
+	struct mlx4_priv *priv = mlx4_priv(dev);
+	struct mlx4_resource_tracker *tracker = &priv->mfunc.master.res_tracker;
+	struct list_head *mac_list =
+		&tracker->slave_list[slave].res_list[RES_MAC];
+	struct mac_res *res, *tmp;
+
+	list_for_each_entry_safe(res, tmp, mac_list, list) {
+		if (res->mac == mac && res->port == (u8) port) {
+			if (!--res->ref_count) {
+				list_del(&res->list);
+				mlx4_release_resource(dev, slave, RES_MAC, 1, port);
+				kfree(res);
+			}
+			break;
+		}
+	}
+}
+
+static void rem_slave_macs(struct mlx4_dev *dev, int slave)
+{
+	struct mlx4_priv *priv = mlx4_priv(dev);
+	struct mlx4_resource_tracker *tracker = &priv->mfunc.master.res_tracker;
+	struct list_head *mac_list =
+		&tracker->slave_list[slave].res_list[RES_MAC];
+	struct mac_res *res, *tmp;
+	int i;
+
+	list_for_each_entry_safe(res, tmp, mac_list, list) {
+		list_del(&res->list);
+		/* dereference the mac the num times the slave referenced it */
+		for (i = 0; i < res->ref_count; i++)
+			__mlx4_unregister_mac(dev, res->port, res->mac);
+		mlx4_release_resource(dev, slave, RES_MAC, 1, res->port);
+		kfree(res);
+	}
+}
+
+static int mac_alloc_res(struct mlx4_dev *dev, int slave, int op, int cmd,
+			 u64 in_param, u64 *out_param, int in_port)
+{
+	int err = -EINVAL;
+	int port;
+	u64 mac;
+	u8 smac_index;
+
+	if (op != RES_OP_RESERVE_AND_MAP)
+		return err;
+
+	port = !in_port ? get_param_l(out_param) : in_port;
+	port = mlx4_slave_convert_port(
+			dev, slave, port);
+
+	if (port < 0)
+		return -EINVAL;
+	mac = in_param;
+
+	err = __mlx4_register_mac(dev, port, mac);
+	if (err >= 0) {
+		smac_index = err;
+		set_param_l(out_param, err);
+		err = 0;
+	}
+
+	if (!err) {
+		err = mac_add_to_slave(dev, slave, mac, port, smac_index);
+		if (err)
+			__mlx4_unregister_mac(dev, port, mac);
+	}
+	return err;
+}
+
+static int vlan_add_to_slave(struct mlx4_dev *dev, int slave, u16 vlan,
+			     int port, int vlan_index)
+{
+	struct mlx4_priv *priv = mlx4_priv(dev);
+	struct mlx4_resource_tracker *tracker = &priv->mfunc.master.res_tracker;
+	struct list_head *vlan_list =
+		&tracker->slave_list[slave].res_list[RES_VLAN];
+	struct vlan_res *res, *tmp;
+
+	list_for_each_entry_safe(res, tmp, vlan_list, list) {
+		if (res->vlan == vlan && res->port == (u8) port) {
+			/* vlan found. update ref count */
+			++res->ref_count;
+			return 0;
+		}
+	}
+
+	if (mlx4_grant_resource(dev, slave, RES_VLAN, 1, port))
+		return -EINVAL;
+	res = kzalloc(sizeof(*res), GFP_KERNEL);
+	if (!res) {
+		mlx4_release_resource(dev, slave, RES_VLAN, 1, port);
+		return -ENOMEM;
+	}
+	res->vlan = vlan;
+	res->port = (u8) port;
+	res->vlan_index = vlan_index;
+	res->ref_count = 1;
+	list_add_tail(&res->list,
+		      &tracker->slave_list[slave].res_list[RES_VLAN]);
+	return 0;
+}
+
+
+static void vlan_del_from_slave(struct mlx4_dev *dev, int slave, u16 vlan,
+				int port)
+{
+	struct mlx4_priv *priv = mlx4_priv(dev);
+	struct mlx4_resource_tracker *tracker = &priv->mfunc.master.res_tracker;
+	struct list_head *vlan_list =
+		&tracker->slave_list[slave].res_list[RES_VLAN];
+	struct vlan_res *res, *tmp;
+
+	list_for_each_entry_safe(res, tmp, vlan_list, list) {
+		if (res->vlan == vlan && res->port == (u8) port) {
+			if (!--res->ref_count) {
+				list_del(&res->list);
+				mlx4_release_resource(dev, slave, RES_VLAN,
+						      1, port);
+				kfree(res);
+			}
+			break;
+		}
+	}
+}
+
+static void rem_slave_vlans(struct mlx4_dev *dev, int slave)
+{
+	struct mlx4_priv *priv = mlx4_priv(dev);
+	struct mlx4_resource_tracker *tracker = &priv->mfunc.master.res_tracker;
+	struct list_head *vlan_list =
+		&tracker->slave_list[slave].res_list[RES_VLAN];
+	struct vlan_res *res, *tmp;
+	int i;
+
+	list_for_each_entry_safe(res, tmp, vlan_list, list) {
+		list_del(&res->list);
+		/* dereference the vlan the num times the slave referenced it */
+		for (i = 0; i < res->ref_count; i++)
+			__mlx4_unregister_vlan(dev, res->port, res->vlan);
+		mlx4_release_resource(dev, slave, RES_VLAN, 1, res->port);
+		kfree(res);
+	}
+}
+
+static int vlan_alloc_res(struct mlx4_dev *dev, int slave, int op, int cmd,
+			  u64 in_param, u64 *out_param, int in_port)
+{
+	struct mlx4_priv *priv = mlx4_priv(dev);
+	struct mlx4_slave_state *slave_state = priv->mfunc.master.slave_state;
+	int err;
+	u16 vlan;
+	int vlan_index;
+	int port;
+
+	port = !in_port ? get_param_l(out_param) : in_port;
+
+	if (!port || op != RES_OP_RESERVE_AND_MAP)
+		return -EINVAL;
+
+	port = mlx4_slave_convert_port(
+			dev, slave, port);
+
+	if (port < 0)
+		return -EINVAL;
+	/* upstream kernels had NOP for reg/unreg vlan. Continue this. */
+	if (!in_port && port > 0 && port <= dev->caps.num_ports) {
+		slave_state[slave].old_vlan_api = true;
+		return 0;
+	}
+
+	vlan = (u16) in_param;
+
+	err = __mlx4_register_vlan(dev, port, vlan, &vlan_index);
+	if (!err) {
+		set_param_l(out_param, (u32) vlan_index);
+		err = vlan_add_to_slave(dev, slave, vlan, port, vlan_index);
+		if (err)
+			__mlx4_unregister_vlan(dev, port, vlan);
+	}
+	return err;
+}
+
+static int counter_alloc_res(struct mlx4_dev *dev, int slave, int op, int cmd,
+			     u64 in_param, u64 *out_param, int port)
+{
+	u32 index;
+	int err;
+
+	if (op != RES_OP_RESERVE)
+		return -EINVAL;
+
+	err = mlx4_grant_resource(dev, slave, RES_COUNTER, 1, 0);
+	if (err)
+		return err;
+
+	err = __mlx4_counter_alloc(dev, &index);
+	if (err) {
+		mlx4_release_resource(dev, slave, RES_COUNTER, 1, 0);
+		return err;
+	}
+
+	err = add_res_range(dev, slave, index, 1, RES_COUNTER, port);
+	if (err) {
+		__mlx4_counter_free(dev, index);
+		mlx4_release_resource(dev, slave, RES_COUNTER, 1, 0);
+	} else {
+		set_param_l(out_param, index);
+	}
+
+	return err;
+}
+
+static int xrcdn_alloc_res(struct mlx4_dev *dev, int slave, int op, int cmd,
+			   u64 in_param, u64 *out_param)
+{
+	u32 xrcdn;
+	int err;
+
+	if (op != RES_OP_RESERVE)
+		return -EINVAL;
+
+	err = __mlx4_xrcd_alloc(dev, &xrcdn);
+	if (err)
+		return err;
+
+	err = add_res_range(dev, slave, xrcdn, 1, RES_XRCD, 0);
+	if (err)
+		__mlx4_xrcd_free(dev, xrcdn);
+	else
+		set_param_l(out_param, xrcdn);
+
+	return err;
+}
+
+int mlx4_ALLOC_RES_wrapper(struct mlx4_dev *dev, int slave,
+			   struct mlx4_vhcr *vhcr,
+			   struct mlx4_cmd_mailbox *inbox,
+			   struct mlx4_cmd_mailbox *outbox,
+			   struct mlx4_cmd_info *cmd)
+{
+	int err;
+	int alop = vhcr->op_modifier;
+
+	switch (vhcr->in_modifier & 0xFF) {
+	case RES_QP:
+		err = qp_alloc_res(dev, slave, vhcr->op_modifier, alop,
+				   vhcr->in_param, &vhcr->out_param);
+		break;
+
+	case RES_MTT:
+		err = mtt_alloc_res(dev, slave, vhcr->op_modifier, alop,
+				    vhcr->in_param, &vhcr->out_param);
+		break;
+
+	case RES_MPT:
+		err = mpt_alloc_res(dev, slave, vhcr->op_modifier, alop,
+				    vhcr->in_param, &vhcr->out_param);
+		break;
+
+	case RES_CQ:
+		err = cq_alloc_res(dev, slave, vhcr->op_modifier, alop,
+				   vhcr->in_param, &vhcr->out_param);
+		break;
+
+	case RES_SRQ:
+		err = srq_alloc_res(dev, slave, vhcr->op_modifier, alop,
+				    vhcr->in_param, &vhcr->out_param);
+		break;
+
+	case RES_MAC:
+		err = mac_alloc_res(dev, slave, vhcr->op_modifier, alop,
+				    vhcr->in_param, &vhcr->out_param,
+				    (vhcr->in_modifier >> 8) & 0xFF);
+		break;
+
+	case RES_VLAN:
+		err = vlan_alloc_res(dev, slave, vhcr->op_modifier, alop,
+				     vhcr->in_param, &vhcr->out_param,
+				     (vhcr->in_modifier >> 8) & 0xFF);
+		break;
+
+	case RES_COUNTER:
+		err = counter_alloc_res(dev, slave, vhcr->op_modifier, alop,
+					vhcr->in_param, &vhcr->out_param, 0);
+		break;
+
+	case RES_XRCD:
+		err = xrcdn_alloc_res(dev, slave, vhcr->op_modifier, alop,
+				      vhcr->in_param, &vhcr->out_param);
+		break;
+
+	default:
+		err = -EINVAL;
+		break;
+	}
+
+	return err;
+}
+
+static int qp_free_res(struct mlx4_dev *dev, int slave, int op, int cmd,
+		       u64 in_param)
+{
+	int err;
+	int count;
+	int base;
+	int qpn;
+
+	switch (op) {
+	case RES_OP_RESERVE:
+		base = get_param_l(&in_param) & 0x7fffff;
+		count = get_param_h(&in_param);
+		err = rem_res_range(dev, slave, base, count, RES_QP, 0);
+		if (err)
+			break;
+		mlx4_release_resource(dev, slave, RES_QP, count, 0);
+		__mlx4_qp_release_range(dev, base, count);
+		break;
+	case RES_OP_MAP_ICM:
+		qpn = get_param_l(&in_param) & 0x7fffff;
+		err = qp_res_start_move_to(dev, slave, qpn, RES_QP_RESERVED,
+					   NULL, 0);
+		if (err)
+			return err;
+
+		if (!fw_reserved(dev, qpn))
+			__mlx4_qp_free_icm(dev, qpn);
+
+		res_end_move(dev, slave, RES_QP, qpn);
+
+		if (valid_reserved(dev, slave, qpn))
+			err = rem_res_range(dev, slave, qpn, 1, RES_QP, 0);
+		break;
+	default:
+		err = -EINVAL;
+		break;
+	}
+	return err;
+}
+
+static int mtt_free_res(struct mlx4_dev *dev, int slave, int op, int cmd,
+			u64 in_param, u64 *out_param)
+{
+	int err = -EINVAL;
+	int base;
+	int order;
+
+	if (op != RES_OP_RESERVE_AND_MAP)
+		return err;
+
+	base = get_param_l(&in_param);
+	order = get_param_h(&in_param);
+	err = rem_res_range(dev, slave, base, 1, RES_MTT, order);
+	if (!err) {
+		mlx4_release_resource(dev, slave, RES_MTT, 1 << order, 0);
+		__mlx4_free_mtt_range(dev, base, order);
+	}
+	return err;
+}
+
+static int mpt_free_res(struct mlx4_dev *dev, int slave, int op, int cmd,
+			u64 in_param)
+{
+	int err = -EINVAL;
+	int index;
+	int id;
+	struct res_mpt *mpt;
+
+	switch (op) {
+	case RES_OP_RESERVE:
+		index = get_param_l(&in_param);
+		id = index & mpt_mask(dev);
+		err = get_res(dev, slave, id, RES_MPT, &mpt);
+		if (err)
+			break;
+		index = mpt->key;
+		put_res(dev, slave, id, RES_MPT);
+
+		err = rem_res_range(dev, slave, id, 1, RES_MPT, 0);
+		if (err)
+			break;
+		mlx4_release_resource(dev, slave, RES_MPT, 1, 0);
+		__mlx4_mpt_release(dev, index);
+		break;
+	case RES_OP_MAP_ICM:
+		index = get_param_l(&in_param);
+		id = index & mpt_mask(dev);
+		err = mr_res_start_move_to(dev, slave, id,
+					   RES_MPT_RESERVED, &mpt);
+		if (err)
+			return err;
+
+		__mlx4_mpt_free_icm(dev, mpt->key);
+		res_end_move(dev, slave, RES_MPT, id);
+		break;
+	default:
+		err = -EINVAL;
+		break;
+	}
+	return err;
+}
+
+static int cq_free_res(struct mlx4_dev *dev, int slave, int op, int cmd,
+		       u64 in_param, u64 *out_param)
+{
+	int cqn;
+	int err;
+
+	switch (op) {
+	case RES_OP_RESERVE_AND_MAP:
+		cqn = get_param_l(&in_param);
+		err = rem_res_range(dev, slave, cqn, 1, RES_CQ, 0);
+		if (err)
+			break;
+
+		mlx4_release_resource(dev, slave, RES_CQ, 1, 0);
+		__mlx4_cq_free_icm(dev, cqn);
+		break;
+
+	default:
+		err = -EINVAL;
+		break;
+	}
+
+	return err;
+}
+
+static int srq_free_res(struct mlx4_dev *dev, int slave, int op, int cmd,
+			u64 in_param, u64 *out_param)
+{
+	int srqn;
+	int err;
+
+	switch (op) {
+	case RES_OP_RESERVE_AND_MAP:
+		srqn = get_param_l(&in_param);
+		err = rem_res_range(dev, slave, srqn, 1, RES_SRQ, 0);
+		if (err)
+			break;
+
+		mlx4_release_resource(dev, slave, RES_SRQ, 1, 0);
+		__mlx4_srq_free_icm(dev, srqn);
+		break;
+
+	default:
+		err = -EINVAL;
+		break;
+	}
+
+	return err;
+}
+
+static int mac_free_res(struct mlx4_dev *dev, int slave, int op, int cmd,
+			    u64 in_param, u64 *out_param, int in_port)
+{
+	int port;
+	int err = 0;
+
+	switch (op) {
+	case RES_OP_RESERVE_AND_MAP:
+		port = !in_port ? get_param_l(out_param) : in_port;
+		port = mlx4_slave_convert_port(
+				dev, slave, port);
+
+		if (port < 0)
+			return -EINVAL;
+		mac_del_from_slave(dev, slave, in_param, port);
+		__mlx4_unregister_mac(dev, port, in_param);
+		break;
+	default:
+		err = -EINVAL;
+		break;
+	}
+
+	return err;
+
+}
+
+static int vlan_free_res(struct mlx4_dev *dev, int slave, int op, int cmd,
+			    u64 in_param, u64 *out_param, int port)
+{
+	struct mlx4_priv *priv = mlx4_priv(dev);
+	struct mlx4_slave_state *slave_state = priv->mfunc.master.slave_state;
+	int err = 0;
+
+	port = mlx4_slave_convert_port(
+			dev, slave, port);
+
+	if (port < 0)
+		return -EINVAL;
+	switch (op) {
+	case RES_OP_RESERVE_AND_MAP:
+		if (slave_state[slave].old_vlan_api)
+			return 0;
+		if (!port)
+			return -EINVAL;
+		vlan_del_from_slave(dev, slave, in_param, port);
+		__mlx4_unregister_vlan(dev, port, in_param);
+		break;
+	default:
+		err = -EINVAL;
+		break;
+	}
+
+	return err;
+}
+
+static int counter_free_res(struct mlx4_dev *dev, int slave, int op, int cmd,
+			    u64 in_param, u64 *out_param)
+{
+	int index;
+	int err;
+
+	if (op != RES_OP_RESERVE)
+		return -EINVAL;
+
+	index = get_param_l(&in_param);
+	if (index == MLX4_SINK_COUNTER_INDEX(dev))
+		return 0;
+
+	err = rem_res_range(dev, slave, index, 1, RES_COUNTER, 0);
+	if (err)
+		return err;
+
+	__mlx4_counter_free(dev, index);
+	mlx4_release_resource(dev, slave, RES_COUNTER, 1, 0);
+
+	return err;
+}
+
+static int xrcdn_free_res(struct mlx4_dev *dev, int slave, int op, int cmd,
+			  u64 in_param, u64 *out_param)
+{
+	int xrcdn;
+	int err;
+
+	if (op != RES_OP_RESERVE)
+		return -EINVAL;
+
+	xrcdn = get_param_l(&in_param);
+	err = rem_res_range(dev, slave, xrcdn, 1, RES_XRCD, 0);
+	if (err)
+		return err;
+
+	__mlx4_xrcd_free(dev, xrcdn);
+
+	return err;
+}
+
+int mlx4_FREE_RES_wrapper(struct mlx4_dev *dev, int slave,
+			  struct mlx4_vhcr *vhcr,
+			  struct mlx4_cmd_mailbox *inbox,
+			  struct mlx4_cmd_mailbox *outbox,
+			  struct mlx4_cmd_info *cmd)
+{
+	int err = -EINVAL;
+	int alop = vhcr->op_modifier;
+
+	switch (vhcr->in_modifier & 0xFF) {
+	case RES_QP:
+		err = qp_free_res(dev, slave, vhcr->op_modifier, alop,
+				  vhcr->in_param);
+		break;
+
+	case RES_MTT:
+		err = mtt_free_res(dev, slave, vhcr->op_modifier, alop,
+				   vhcr->in_param, &vhcr->out_param);
+		break;
+
+	case RES_MPT:
+		err = mpt_free_res(dev, slave, vhcr->op_modifier, alop,
+				   vhcr->in_param);
+		break;
+
+	case RES_CQ:
+		err = cq_free_res(dev, slave, vhcr->op_modifier, alop,
+				  vhcr->in_param, &vhcr->out_param);
+		break;
+
+	case RES_SRQ:
+		err = srq_free_res(dev, slave, vhcr->op_modifier, alop,
+				   vhcr->in_param, &vhcr->out_param);
+		break;
+
+	case RES_MAC:
+		err = mac_free_res(dev, slave, vhcr->op_modifier, alop,
+				   vhcr->in_param, &vhcr->out_param,
+				   (vhcr->in_modifier >> 8) & 0xFF);
+		break;
+
+	case RES_VLAN:
+		err = vlan_free_res(dev, slave, vhcr->op_modifier, alop,
+				    vhcr->in_param, &vhcr->out_param,
+				    (vhcr->in_modifier >> 8) & 0xFF);
+		break;
+
+	case RES_COUNTER:
+		err = counter_free_res(dev, slave, vhcr->op_modifier, alop,
+				       vhcr->in_param, &vhcr->out_param);
+		break;
+
+	case RES_XRCD:
+		err = xrcdn_free_res(dev, slave, vhcr->op_modifier, alop,
+				     vhcr->in_param, &vhcr->out_param);
+
+	default:
+		break;
+	}
+	return err;
+}
+
+/* ugly but other choices are uglier */
+static int mr_phys_mpt(struct mlx4_mpt_entry *mpt)
+{
+	return (be32_to_cpu(mpt->flags) >> 9) & 1;
+}
+
+static int mr_get_mtt_addr(struct mlx4_mpt_entry *mpt)
+{
+	return (int)be64_to_cpu(mpt->mtt_addr) & 0xfffffff8;
+}
+
+static int mr_get_mtt_size(struct mlx4_mpt_entry *mpt)
+{
+	return be32_to_cpu(mpt->mtt_sz);
+}
+
+static u32 mr_get_pd(struct mlx4_mpt_entry *mpt)
+{
+	return be32_to_cpu(mpt->pd_flags) & 0x00ffffff;
+}
+
+static int mr_is_fmr(struct mlx4_mpt_entry *mpt)
+{
+	return be32_to_cpu(mpt->pd_flags) & MLX4_MPT_PD_FLAG_FAST_REG;
+}
+
+static int mr_is_bind_enabled(struct mlx4_mpt_entry *mpt)
+{
+	return be32_to_cpu(mpt->flags) & MLX4_MPT_FLAG_BIND_ENABLE;
+}
+
+static int mr_is_region(struct mlx4_mpt_entry *mpt)
+{
+	return be32_to_cpu(mpt->flags) & MLX4_MPT_FLAG_REGION;
+}
+
+static int qp_get_mtt_addr(struct mlx4_qp_context *qpc)
+{
+	return be32_to_cpu(qpc->mtt_base_addr_l) & 0xfffffff8;
+}
+
+static int srq_get_mtt_addr(struct mlx4_srq_context *srqc)
+{
+	return be32_to_cpu(srqc->mtt_base_addr_l) & 0xfffffff8;
+}
+
+static int qp_get_mtt_size(struct mlx4_qp_context *qpc)
+{
+	int page_shift = (qpc->log_page_size & 0x3f) + 12;
+	int log_sq_size = (qpc->sq_size_stride >> 3) & 0xf;
+	int log_sq_sride = qpc->sq_size_stride & 7;
+	int log_rq_size = (qpc->rq_size_stride >> 3) & 0xf;
+	int log_rq_stride = qpc->rq_size_stride & 7;
+	int srq = (be32_to_cpu(qpc->srqn) >> 24) & 1;
+	int rss = (be32_to_cpu(qpc->flags) >> 13) & 1;
+	u32 ts = (be32_to_cpu(qpc->flags) >> 16) & 0xff;
+	int xrc = (ts == MLX4_QP_ST_XRC) ? 1 : 0;
+	int sq_size;
+	int rq_size;
+	int total_pages;
+	int total_mem;
+	int page_offset = (be32_to_cpu(qpc->params2) >> 6) & 0x3f;
+	int tot;
+
+	sq_size = 1 << (log_sq_size + log_sq_sride + 4);
+	rq_size = (srq|rss|xrc) ? 0 : (1 << (log_rq_size + log_rq_stride + 4));
+	total_mem = sq_size + rq_size;
+	tot = (total_mem + (page_offset << 6)) >> page_shift;
+	total_pages = !tot ? 1 : roundup_pow_of_two(tot);
+
+	return total_pages;
+}
+
+static int check_mtt_range(struct mlx4_dev *dev, int slave, int start,
+			   int size, struct res_mtt *mtt)
+{
+	int res_start = mtt->com.res_id;
+	int res_size = (1 << mtt->order);
+
+	if (start < res_start || start + size > res_start + res_size)
+		return -EPERM;
+	return 0;
+}
+
+int mlx4_SW2HW_MPT_wrapper(struct mlx4_dev *dev, int slave,
+			   struct mlx4_vhcr *vhcr,
+			   struct mlx4_cmd_mailbox *inbox,
+			   struct mlx4_cmd_mailbox *outbox,
+			   struct mlx4_cmd_info *cmd)
+{
+	int err;
+	int index = vhcr->in_modifier;
+	struct res_mtt *mtt;
+	struct res_mpt *mpt = NULL;
+	int mtt_base = mr_get_mtt_addr(inbox->buf) / dev->caps.mtt_entry_sz;
+	int phys;
+	int id;
+	u32 pd;
+	int pd_slave;
+
+	id = index & mpt_mask(dev);
+	err = mr_res_start_move_to(dev, slave, id, RES_MPT_HW, &mpt);
+	if (err)
+		return err;
+
+	/* Disable memory windows for VFs. */
+	if (!mr_is_region(inbox->buf)) {
+		err = -EPERM;
+		goto ex_abort;
+	}
+
+	/* Make sure that the PD bits related to the slave id are zeros. */
+	pd = mr_get_pd(inbox->buf);
+	pd_slave = (pd >> 17) & 0x7f;
+	if (pd_slave != 0 && --pd_slave != slave) {
+		err = -EPERM;
+		goto ex_abort;
+	}
+
+	if (mr_is_fmr(inbox->buf)) {
+		/* FMR and Bind Enable are forbidden in slave devices. */
+		if (mr_is_bind_enabled(inbox->buf)) {
+			err = -EPERM;
+			goto ex_abort;
+		}
+		/* FMR and Memory Windows are also forbidden. */
+		if (!mr_is_region(inbox->buf)) {
+			err = -EPERM;
+			goto ex_abort;
+		}
+	}
+
+	phys = mr_phys_mpt(inbox->buf);
+	if (!phys) {
+		err = get_res(dev, slave, mtt_base, RES_MTT, &mtt);
+		if (err)
+			goto ex_abort;
+
+		err = check_mtt_range(dev, slave, mtt_base,
+				      mr_get_mtt_size(inbox->buf), mtt);
+		if (err)
+			goto ex_put;
+
+		mpt->mtt = mtt;
+	}
+
+	err = mlx4_DMA_wrapper(dev, slave, vhcr, inbox, outbox, cmd);
+	if (err)
+		goto ex_put;
+
+	if (!phys) {
+		atomic_inc(&mtt->ref_count);
+		put_res(dev, slave, mtt->com.res_id, RES_MTT);
+	}
+
+	res_end_move(dev, slave, RES_MPT, id);
+	return 0;
+
+ex_put:
+	if (!phys)
+		put_res(dev, slave, mtt->com.res_id, RES_MTT);
+ex_abort:
+	res_abort_move(dev, slave, RES_MPT, id);
+
+	return err;
+}
+
+int mlx4_HW2SW_MPT_wrapper(struct mlx4_dev *dev, int slave,
+			   struct mlx4_vhcr *vhcr,
+			   struct mlx4_cmd_mailbox *inbox,
+			   struct mlx4_cmd_mailbox *outbox,
+			   struct mlx4_cmd_info *cmd)
+{
+	int err;
+	int index = vhcr->in_modifier;
+	struct res_mpt *mpt;
+	int id;
+
+	id = index & mpt_mask(dev);
+	err = mr_res_start_move_to(dev, slave, id, RES_MPT_MAPPED, &mpt);
+	if (err)
+		return err;
+
+	err = mlx4_DMA_wrapper(dev, slave, vhcr, inbox, outbox, cmd);
+	if (err)
+		goto ex_abort;
+
+	if (mpt->mtt)
+		atomic_dec(&mpt->mtt->ref_count);
+
+	res_end_move(dev, slave, RES_MPT, id);
+	return 0;
+
+ex_abort:
+	res_abort_move(dev, slave, RES_MPT, id);
+
+	return err;
+}
+
+int mlx4_QUERY_MPT_wrapper(struct mlx4_dev *dev, int slave,
+			   struct mlx4_vhcr *vhcr,
+			   struct mlx4_cmd_mailbox *inbox,
+			   struct mlx4_cmd_mailbox *outbox,
+			   struct mlx4_cmd_info *cmd)
+{
+	int err;
+	int index = vhcr->in_modifier;
+	struct res_mpt *mpt;
+	int id;
+
+	id = index & mpt_mask(dev);
+	err = get_res(dev, slave, id, RES_MPT, &mpt);
+	if (err)
+		return err;
+
+	if (mpt->com.from_state == RES_MPT_MAPPED) {
+		/* In order to allow rereg in SRIOV, we need to alter the MPT entry. To do
+		 * that, the VF must read the MPT. But since the MPT entry memory is not
+		 * in the VF's virtual memory space, it must use QUERY_MPT to obtain the
+		 * entry contents. To guarantee that the MPT cannot be changed, the driver
+		 * must perform HW2SW_MPT before this query and return the MPT entry to HW
+		 * ownership fofollowing the change. The change here allows the VF to
+		 * perform QUERY_MPT also when the entry is in SW ownership.
+		 */
+		struct mlx4_mpt_entry *mpt_entry = mlx4_table_find(
+					&mlx4_priv(dev)->mr_table.dmpt_table,
+					mpt->key, NULL);
+
+		if (NULL == mpt_entry || NULL == outbox->buf) {
+			err = -EINVAL;
+			goto out;
+		}
+
+		memcpy(outbox->buf, mpt_entry, sizeof(*mpt_entry));
+
+		err = 0;
+	} else if (mpt->com.from_state == RES_MPT_HW) {
+		err = mlx4_DMA_wrapper(dev, slave, vhcr, inbox, outbox, cmd);
+	} else {
+		err = -EBUSY;
+		goto out;
+	}
+
+
+out:
+	put_res(dev, slave, id, RES_MPT);
+	return err;
+}
+
+static int qp_get_rcqn(struct mlx4_qp_context *qpc)
+{
+	return be32_to_cpu(qpc->cqn_recv) & 0xffffff;
+}
+
+static int qp_get_scqn(struct mlx4_qp_context *qpc)
+{
+	return be32_to_cpu(qpc->cqn_send) & 0xffffff;
+}
+
+static u32 qp_get_srqn(struct mlx4_qp_context *qpc)
+{
+	return be32_to_cpu(qpc->srqn) & 0x1ffffff;
+}
+
+static void adjust_proxy_tun_qkey(struct mlx4_dev *dev, struct mlx4_vhcr *vhcr,
+				  struct mlx4_qp_context *context)
+{
+	u32 qpn = vhcr->in_modifier & 0xffffff;
+	u32 qkey = 0;
+
+	if (mlx4_get_parav_qkey(dev, qpn, &qkey))
+		return;
+
+	/* adjust qkey in qp context */
+	context->qkey = cpu_to_be32(qkey);
+}
+
+static int adjust_qp_sched_queue(struct mlx4_dev *dev, int slave,
+				 struct mlx4_qp_context *qpc,
+				 struct mlx4_cmd_mailbox *inbox);
+
+int mlx4_RST2INIT_QP_wrapper(struct mlx4_dev *dev, int slave,
+			     struct mlx4_vhcr *vhcr,
+			     struct mlx4_cmd_mailbox *inbox,
+			     struct mlx4_cmd_mailbox *outbox,
+			     struct mlx4_cmd_info *cmd)
+{
+	int err;
+	int qpn = vhcr->in_modifier & 0x7fffff;
+	struct res_mtt *mtt;
+	struct res_qp *qp;
+	struct mlx4_qp_context *qpc = inbox->buf + 8;
+	int mtt_base = qp_get_mtt_addr(qpc) / dev->caps.mtt_entry_sz;
+	int mtt_size = qp_get_mtt_size(qpc);
+	struct res_cq *rcq;
+	struct res_cq *scq;
+	int rcqn = qp_get_rcqn(qpc);
+	int scqn = qp_get_scqn(qpc);
+	u32 srqn = qp_get_srqn(qpc) & 0xffffff;
+	int use_srq = (qp_get_srqn(qpc) >> 24) & 1;
+	struct res_srq *srq;
+	int local_qpn = vhcr->in_modifier & 0xffffff;
+
+	err = adjust_qp_sched_queue(dev, slave, qpc, inbox);
+	if (err)
+		return err;
+
+	err = qp_res_start_move_to(dev, slave, qpn, RES_QP_HW, &qp, 0);
+	if (err)
+		return err;
+	qp->local_qpn = local_qpn;
+	qp->sched_queue = 0;
+	qp->param3 = 0;
+	qp->vlan_control = 0;
+	qp->fvl_rx = 0;
+	qp->pri_path_fl = 0;
+	qp->vlan_index = 0;
+	qp->feup = 0;
+	qp->qpc_flags = be32_to_cpu(qpc->flags);
+
+	err = get_res(dev, slave, mtt_base, RES_MTT, &mtt);
+	if (err)
+		goto ex_abort;
+
+	err = check_mtt_range(dev, slave, mtt_base, mtt_size, mtt);
+	if (err)
+		goto ex_put_mtt;
+
+	err = get_res(dev, slave, rcqn, RES_CQ, &rcq);
+	if (err)
+		goto ex_put_mtt;
+
+	if (scqn != rcqn) {
+		err = get_res(dev, slave, scqn, RES_CQ, &scq);
+		if (err)
+			goto ex_put_rcq;
+	} else
+		scq = rcq;
+
+	if (use_srq) {
+		err = get_res(dev, slave, srqn, RES_SRQ, &srq);
+		if (err)
+			goto ex_put_scq;
+	}
+
+	adjust_proxy_tun_qkey(dev, vhcr, qpc);
+	update_pkey_index(dev, slave, inbox);
+	err = mlx4_DMA_wrapper(dev, slave, vhcr, inbox, outbox, cmd);
+	if (err)
+		goto ex_put_srq;
+	atomic_inc(&mtt->ref_count);
+	qp->mtt = mtt;
+	atomic_inc(&rcq->ref_count);
+	qp->rcq = rcq;
+	atomic_inc(&scq->ref_count);
+	qp->scq = scq;
+
+	if (scqn != rcqn)
+		put_res(dev, slave, scqn, RES_CQ);
+
+	if (use_srq) {
+		atomic_inc(&srq->ref_count);
+		put_res(dev, slave, srqn, RES_SRQ);
+		qp->srq = srq;
+	}
+
+	/* Save param3 for dynamic changes from VST back to VGT */
+	qp->param3 = qpc->param3;
+	put_res(dev, slave, rcqn, RES_CQ);
+	put_res(dev, slave, mtt_base, RES_MTT);
+	res_end_move(dev, slave, RES_QP, qpn);
+
+	return 0;
+
+ex_put_srq:
+	if (use_srq)
+		put_res(dev, slave, srqn, RES_SRQ);
+ex_put_scq:
+	if (scqn != rcqn)
+		put_res(dev, slave, scqn, RES_CQ);
+ex_put_rcq:
+	put_res(dev, slave, rcqn, RES_CQ);
+ex_put_mtt:
+	put_res(dev, slave, mtt_base, RES_MTT);
+ex_abort:
+	res_abort_move(dev, slave, RES_QP, qpn);
+
+	return err;
+}
+
+static int eq_get_mtt_addr(struct mlx4_eq_context *eqc)
+{
+	return be32_to_cpu(eqc->mtt_base_addr_l) & 0xfffffff8;
+}
+
+static int eq_get_mtt_size(struct mlx4_eq_context *eqc)
+{
+	int log_eq_size = eqc->log_eq_size & 0x1f;
+	int page_shift = (eqc->log_page_size & 0x3f) + 12;
+
+	if (log_eq_size + 5 < page_shift)
+		return 1;
+
+	return 1 << (log_eq_size + 5 - page_shift);
+}
+
+static int cq_get_mtt_addr(struct mlx4_cq_context *cqc)
+{
+	return be32_to_cpu(cqc->mtt_base_addr_l) & 0xfffffff8;
+}
+
+static int cq_get_mtt_size(struct mlx4_cq_context *cqc)
+{
+	int log_cq_size = (be32_to_cpu(cqc->logsize_usrpage) >> 24) & 0x1f;
+	int page_shift = (cqc->log_page_size & 0x3f) + 12;
+
+	if (log_cq_size + 5 < page_shift)
+		return 1;
+
+	return 1 << (log_cq_size + 5 - page_shift);
+}
+
+int mlx4_SW2HW_EQ_wrapper(struct mlx4_dev *dev, int slave,
+			  struct mlx4_vhcr *vhcr,
+			  struct mlx4_cmd_mailbox *inbox,
+			  struct mlx4_cmd_mailbox *outbox,
+			  struct mlx4_cmd_info *cmd)
+{
+	int err;
+	int eqn = vhcr->in_modifier;
+	int res_id = (slave << 10) | eqn;
+	struct mlx4_eq_context *eqc = inbox->buf;
+	int mtt_base = eq_get_mtt_addr(eqc) / dev->caps.mtt_entry_sz;
+	int mtt_size = eq_get_mtt_size(eqc);
+	struct res_eq *eq;
+	struct res_mtt *mtt;
+
+	err = add_res_range(dev, slave, res_id, 1, RES_EQ, 0);
+	if (err)
+		return err;
+	err = eq_res_start_move_to(dev, slave, res_id, RES_EQ_HW, &eq);
+	if (err)
+		goto out_add;
+
+	err = get_res(dev, slave, mtt_base, RES_MTT, &mtt);
+	if (err)
+		goto out_move;
+
+	err = check_mtt_range(dev, slave, mtt_base, mtt_size, mtt);
+	if (err)
+		goto out_put;
+
+	err = mlx4_DMA_wrapper(dev, slave, vhcr, inbox, outbox, cmd);
+	if (err)
+		goto out_put;
+
+	atomic_inc(&mtt->ref_count);
+	eq->mtt = mtt;
+	put_res(dev, slave, mtt->com.res_id, RES_MTT);
+	res_end_move(dev, slave, RES_EQ, res_id);
+	return 0;
+
+out_put:
+	put_res(dev, slave, mtt->com.res_id, RES_MTT);
+out_move:
+	res_abort_move(dev, slave, RES_EQ, res_id);
+out_add:
+	rem_res_range(dev, slave, res_id, 1, RES_EQ, 0);
+	return err;
+}
+
+int mlx4_CONFIG_DEV_wrapper(struct mlx4_dev *dev, int slave,
+			    struct mlx4_vhcr *vhcr,
+			    struct mlx4_cmd_mailbox *inbox,
+			    struct mlx4_cmd_mailbox *outbox,
+			    struct mlx4_cmd_info *cmd)
+{
+	int err;
+	u8 get = vhcr->op_modifier;
+
+	if (get != 1)
+		return -EPERM;
+
+	err = mlx4_DMA_wrapper(dev, slave, vhcr, inbox, outbox, cmd);
+
+	return err;
+}
+
+static int get_containing_mtt(struct mlx4_dev *dev, int slave, int start,
+			      int len, struct res_mtt **res)
+{
+	struct mlx4_priv *priv = mlx4_priv(dev);
+	struct mlx4_resource_tracker *tracker = &priv->mfunc.master.res_tracker;
+	struct res_mtt *mtt;
+	int err = -EINVAL;
+
+	spin_lock_irq(mlx4_tlock(dev));
+	list_for_each_entry(mtt, &tracker->slave_list[slave].res_list[RES_MTT],
+			    com.list) {
+		if (!check_mtt_range(dev, slave, start, len, mtt)) {
+			*res = mtt;
+			mtt->com.from_state = mtt->com.state;
+			mtt->com.state = RES_MTT_BUSY;
+			err = 0;
+			break;
+		}
+	}
+	spin_unlock_irq(mlx4_tlock(dev));
+
+	return err;
+}
+
+static int verify_qp_parameters(struct mlx4_dev *dev,
+				struct mlx4_vhcr *vhcr,
+				struct mlx4_cmd_mailbox *inbox,
+				enum qp_transition transition, u8 slave)
+{
+	u32			qp_type;
+	u32			qpn;
+	struct mlx4_qp_context	*qp_ctx;
+	enum mlx4_qp_optpar	optpar;
+	int port;
+	int num_gids;
+
+	qp_ctx  = inbox->buf + 8;
+	qp_type	= (be32_to_cpu(qp_ctx->flags) >> 16) & 0xff;
+	optpar	= be32_to_cpu(*(__be32 *) inbox->buf);
+
+	if (slave != mlx4_master_func_num(dev)) {
+		qp_ctx->params2 &= ~cpu_to_be32(MLX4_QP_BIT_FPP);
+		/* setting QP rate-limit is disallowed for VFs */
+		if (qp_ctx->rate_limit_params)
+			return -EPERM;
+	}
+
+	switch (qp_type) {
+	case MLX4_QP_ST_RC:
+	case MLX4_QP_ST_XRC:
+	case MLX4_QP_ST_UC:
+		switch (transition) {
+		case QP_TRANS_INIT2RTR:
+		case QP_TRANS_RTR2RTS:
+		case QP_TRANS_RTS2RTS:
+		case QP_TRANS_SQD2SQD:
+		case QP_TRANS_SQD2RTS:
+			if (slave != mlx4_master_func_num(dev)) {
+				if (optpar & MLX4_QP_OPTPAR_PRIMARY_ADDR_PATH) {
+					port = (qp_ctx->pri_path.sched_queue >> 6 & 1) + 1;
+					if (dev->caps.port_mask[port] != MLX4_PORT_TYPE_IB)
+						num_gids = mlx4_get_slave_num_gids(dev, slave, port);
+					else
+						num_gids = 1;
+					if (qp_ctx->pri_path.mgid_index >= num_gids)
+						return -EINVAL;
+				}
+				if (optpar & MLX4_QP_OPTPAR_ALT_ADDR_PATH) {
+					port = (qp_ctx->alt_path.sched_queue >> 6 & 1) + 1;
+					if (dev->caps.port_mask[port] != MLX4_PORT_TYPE_IB)
+						num_gids = mlx4_get_slave_num_gids(dev, slave, port);
+					else
+						num_gids = 1;
+					if (qp_ctx->alt_path.mgid_index >= num_gids)
+						return -EINVAL;
+				}
+			}
+			break;
+		default:
+			break;
+		}
+		break;
+
+	case MLX4_QP_ST_MLX:
+		qpn = vhcr->in_modifier & 0x7fffff;
+		port = (qp_ctx->pri_path.sched_queue >> 6 & 1) + 1;
+		if (transition == QP_TRANS_INIT2RTR &&
+		    slave != mlx4_master_func_num(dev) &&
+		    mlx4_is_qp_reserved(dev, qpn) &&
+		    !mlx4_vf_smi_enabled(dev, slave, port)) {
+			/* only enabled VFs may create MLX proxy QPs */
+			mlx4_err(dev, "%s: unprivileged slave %d attempting to create an MLX proxy special QP on port %d\n",
+				 __func__, slave, port);
+			return -EPERM;
+		}
+		break;
+
+	default:
+		break;
+	}
+
+	return 0;
+}
+
+int mlx4_WRITE_MTT_wrapper(struct mlx4_dev *dev, int slave,
+			   struct mlx4_vhcr *vhcr,
+			   struct mlx4_cmd_mailbox *inbox,
+			   struct mlx4_cmd_mailbox *outbox,
+			   struct mlx4_cmd_info *cmd)
+{
+	struct mlx4_mtt mtt;
+	__be64 *page_list = inbox->buf;
+	u64 *pg_list = (u64 *)page_list;
+	int i;
+	struct res_mtt *rmtt = NULL;
+	int start = be64_to_cpu(page_list[0]);
+	int npages = vhcr->in_modifier;
+	int err;
+
+	err = get_containing_mtt(dev, slave, start, npages, &rmtt);
+	if (err)
+		return err;
+
+	/* Call the SW implementation of write_mtt:
+	 * - Prepare a dummy mtt struct
+	 * - Translate inbox contents to simple addresses in host endianness */
+	mtt.offset = 0;  /* TBD this is broken but I don't handle it since
+			    we don't really use it */
+	mtt.order = 0;
+	mtt.page_shift = 0;
+	for (i = 0; i < npages; ++i)
+		pg_list[i + 2] = (be64_to_cpu(page_list[i + 2]) & ~1ULL);
+
+	err = __mlx4_write_mtt(dev, &mtt, be64_to_cpu(page_list[0]), npages,
+			       ((u64 *)page_list + 2));
+
+	if (rmtt)
+		put_res(dev, slave, rmtt->com.res_id, RES_MTT);
+
+	return err;
+}
+
+int mlx4_HW2SW_EQ_wrapper(struct mlx4_dev *dev, int slave,
+			  struct mlx4_vhcr *vhcr,
+			  struct mlx4_cmd_mailbox *inbox,
+			  struct mlx4_cmd_mailbox *outbox,
+			  struct mlx4_cmd_info *cmd)
+{
+	int eqn = vhcr->in_modifier;
+	int res_id = eqn | (slave << 10);
+	struct res_eq *eq;
+	int err;
+
+	err = eq_res_start_move_to(dev, slave, res_id, RES_EQ_RESERVED, &eq);
+	if (err)
+		return err;
+
+	err = get_res(dev, slave, eq->mtt->com.res_id, RES_MTT, NULL);
+	if (err)
+		goto ex_abort;
+
+	err = mlx4_DMA_wrapper(dev, slave, vhcr, inbox, outbox, cmd);
+	if (err)
+		goto ex_put;
+
+	atomic_dec(&eq->mtt->ref_count);
+	put_res(dev, slave, eq->mtt->com.res_id, RES_MTT);
+	res_end_move(dev, slave, RES_EQ, res_id);
+	rem_res_range(dev, slave, res_id, 1, RES_EQ, 0);
+
+	return 0;
+
+ex_put:
+	put_res(dev, slave, eq->mtt->com.res_id, RES_MTT);
+ex_abort:
+	res_abort_move(dev, slave, RES_EQ, res_id);
+
+	return err;
+}
+
+int mlx4_GEN_EQE(struct mlx4_dev *dev, int slave, struct mlx4_eqe *eqe)
+{
+	struct mlx4_priv *priv = mlx4_priv(dev);
+	struct mlx4_slave_event_eq_info *event_eq;
+	struct mlx4_cmd_mailbox *mailbox;
+	u32 in_modifier = 0;
+	int err;
+	int res_id;
+	struct res_eq *req;
+
+	if (!priv->mfunc.master.slave_state)
+		return -EINVAL;
+
+	/* check for slave valid, slave not PF, and slave active */
+	if (slave < 0 || slave > dev->persist->num_vfs ||
+	    slave == dev->caps.function ||
+	    !priv->mfunc.master.slave_state[slave].active)
+		return 0;
+
+	event_eq = &priv->mfunc.master.slave_state[slave].event_eq[eqe->type];
+
+	/* Create the event only if the slave is registered */
+	if (event_eq->eqn < 0)
+		return 0;
+
+	mutex_lock(&priv->mfunc.master.gen_eqe_mutex[slave]);
+	res_id = (slave << 10) | event_eq->eqn;
+	err = get_res(dev, slave, res_id, RES_EQ, &req);
+	if (err)
+		goto unlock;
+
+	if (req->com.from_state != RES_EQ_HW) {
+		err = -EINVAL;
+		goto put;
+	}
+
+	mailbox = mlx4_alloc_cmd_mailbox(dev);
+	if (IS_ERR(mailbox)) {
+		err = PTR_ERR(mailbox);
+		goto put;
+	}
+
+	if (eqe->type == MLX4_EVENT_TYPE_CMD) {
+		++event_eq->token;
+		eqe->event.cmd.token = cpu_to_be16(event_eq->token);
+	}
+
+	memcpy(mailbox->buf, (u8 *) eqe, 28);
+
+	in_modifier = (slave & 0xff) | ((event_eq->eqn & 0x3ff) << 16);
+
+	err = mlx4_cmd(dev, mailbox->dma, in_modifier, 0,
+		       MLX4_CMD_GEN_EQE, MLX4_CMD_TIME_CLASS_B,
+		       MLX4_CMD_NATIVE);
+
+	put_res(dev, slave, res_id, RES_EQ);
+	mutex_unlock(&priv->mfunc.master.gen_eqe_mutex[slave]);
+	mlx4_free_cmd_mailbox(dev, mailbox);
+	return err;
+
+put:
+	put_res(dev, slave, res_id, RES_EQ);
+
+unlock:
+	mutex_unlock(&priv->mfunc.master.gen_eqe_mutex[slave]);
+	return err;
+}
+
+int mlx4_QUERY_EQ_wrapper(struct mlx4_dev *dev, int slave,
+			  struct mlx4_vhcr *vhcr,
+			  struct mlx4_cmd_mailbox *inbox,
+			  struct mlx4_cmd_mailbox *outbox,
+			  struct mlx4_cmd_info *cmd)
+{
+	int eqn = vhcr->in_modifier;
+	int res_id = eqn | (slave << 10);
+	struct res_eq *eq;
+	int err;
+
+	err = get_res(dev, slave, res_id, RES_EQ, &eq);
+	if (err)
+		return err;
+
+	if (eq->com.from_state != RES_EQ_HW) {
+		err = -EINVAL;
+		goto ex_put;
+	}
+
+	err = mlx4_DMA_wrapper(dev, slave, vhcr, inbox, outbox, cmd);
+
+ex_put:
+	put_res(dev, slave, res_id, RES_EQ);
+	return err;
+}
+
+int mlx4_SW2HW_CQ_wrapper(struct mlx4_dev *dev, int slave,
+			  struct mlx4_vhcr *vhcr,
+			  struct mlx4_cmd_mailbox *inbox,
+			  struct mlx4_cmd_mailbox *outbox,
+			  struct mlx4_cmd_info *cmd)
+{
+	int err;
+	int cqn = vhcr->in_modifier;
+	struct mlx4_cq_context *cqc = inbox->buf;
+	int mtt_base = cq_get_mtt_addr(cqc) / dev->caps.mtt_entry_sz;
+	struct res_cq *cq = NULL;
+	struct res_mtt *mtt;
+
+	err = cq_res_start_move_to(dev, slave, cqn, RES_CQ_HW, &cq);
+	if (err)
+		return err;
+	err = get_res(dev, slave, mtt_base, RES_MTT, &mtt);
+	if (err)
+		goto out_move;
+	err = check_mtt_range(dev, slave, mtt_base, cq_get_mtt_size(cqc), mtt);
+	if (err)
+		goto out_put;
+	err = mlx4_DMA_wrapper(dev, slave, vhcr, inbox, outbox, cmd);
+	if (err)
+		goto out_put;
+	atomic_inc(&mtt->ref_count);
+	cq->mtt = mtt;
+	put_res(dev, slave, mtt->com.res_id, RES_MTT);
+	res_end_move(dev, slave, RES_CQ, cqn);
+	return 0;
+
+out_put:
+	put_res(dev, slave, mtt->com.res_id, RES_MTT);
+out_move:
+	res_abort_move(dev, slave, RES_CQ, cqn);
+	return err;
+}
+
+int mlx4_HW2SW_CQ_wrapper(struct mlx4_dev *dev, int slave,
+			  struct mlx4_vhcr *vhcr,
+			  struct mlx4_cmd_mailbox *inbox,
+			  struct mlx4_cmd_mailbox *outbox,
+			  struct mlx4_cmd_info *cmd)
+{
+	int err;
+	int cqn = vhcr->in_modifier;
+	struct res_cq *cq = NULL;
+
+	err = cq_res_start_move_to(dev, slave, cqn, RES_CQ_ALLOCATED, &cq);
+	if (err)
+		return err;
+	err = mlx4_DMA_wrapper(dev, slave, vhcr, inbox, outbox, cmd);
+	if (err)
+		goto out_move;
+	atomic_dec(&cq->mtt->ref_count);
+	res_end_move(dev, slave, RES_CQ, cqn);
+	return 0;
+
+out_move:
+	res_abort_move(dev, slave, RES_CQ, cqn);
+	return err;
+}
+
+int mlx4_QUERY_CQ_wrapper(struct mlx4_dev *dev, int slave,
+			  struct mlx4_vhcr *vhcr,
+			  struct mlx4_cmd_mailbox *inbox,
+			  struct mlx4_cmd_mailbox *outbox,
+			  struct mlx4_cmd_info *cmd)
+{
+	int cqn = vhcr->in_modifier;
+	struct res_cq *cq;
+	int err;
+
+	err = get_res(dev, slave, cqn, RES_CQ, &cq);
+	if (err)
+		return err;
+
+	if (cq->com.from_state != RES_CQ_HW)
+		goto ex_put;
+
+	err = mlx4_DMA_wrapper(dev, slave, vhcr, inbox, outbox, cmd);
+ex_put:
+	put_res(dev, slave, cqn, RES_CQ);
+
+	return err;
+}
+
+static int handle_resize(struct mlx4_dev *dev, int slave,
+			 struct mlx4_vhcr *vhcr,
+			 struct mlx4_cmd_mailbox *inbox,
+			 struct mlx4_cmd_mailbox *outbox,
+			 struct mlx4_cmd_info *cmd,
+			 struct res_cq *cq)
+{
+	int err;
+	struct res_mtt *orig_mtt;
+	struct res_mtt *mtt;
+	struct mlx4_cq_context *cqc = inbox->buf;
+	int mtt_base = cq_get_mtt_addr(cqc) / dev->caps.mtt_entry_sz;
+
+	err = get_res(dev, slave, cq->mtt->com.res_id, RES_MTT, &orig_mtt);
+	if (err)
+		return err;
+
+	if (orig_mtt != cq->mtt) {
+		err = -EINVAL;
+		goto ex_put;
+	}
+
+	err = get_res(dev, slave, mtt_base, RES_MTT, &mtt);
+	if (err)
+		goto ex_put;
+
+	err = check_mtt_range(dev, slave, mtt_base, cq_get_mtt_size(cqc), mtt);
+	if (err)
+		goto ex_put1;
+	err = mlx4_DMA_wrapper(dev, slave, vhcr, inbox, outbox, cmd);
+	if (err)
+		goto ex_put1;
+	atomic_dec(&orig_mtt->ref_count);
+	put_res(dev, slave, orig_mtt->com.res_id, RES_MTT);
+	atomic_inc(&mtt->ref_count);
+	cq->mtt = mtt;
+	put_res(dev, slave, mtt->com.res_id, RES_MTT);
+	return 0;
+
+ex_put1:
+	put_res(dev, slave, mtt->com.res_id, RES_MTT);
+ex_put:
+	put_res(dev, slave, orig_mtt->com.res_id, RES_MTT);
+
+	return err;
+
+}
+
+int mlx4_MODIFY_CQ_wrapper(struct mlx4_dev *dev, int slave,
+			   struct mlx4_vhcr *vhcr,
+			   struct mlx4_cmd_mailbox *inbox,
+			   struct mlx4_cmd_mailbox *outbox,
+			   struct mlx4_cmd_info *cmd)
+{
+	int cqn = vhcr->in_modifier;
+	struct res_cq *cq;
+	int err;
+
+	err = get_res(dev, slave, cqn, RES_CQ, &cq);
+	if (err)
+		return err;
+
+	if (cq->com.from_state != RES_CQ_HW)
+		goto ex_put;
+
+	if (vhcr->op_modifier == 0) {
+		err = handle_resize(dev, slave, vhcr, inbox, outbox, cmd, cq);
+		goto ex_put;
+	}
+
+	err = mlx4_DMA_wrapper(dev, slave, vhcr, inbox, outbox, cmd);
+ex_put:
+	put_res(dev, slave, cqn, RES_CQ);
+
+	return err;
+}
+
+static int srq_get_mtt_size(struct mlx4_srq_context *srqc)
+{
+	int log_srq_size = (be32_to_cpu(srqc->state_logsize_srqn) >> 24) & 0xf;
+	int log_rq_stride = srqc->logstride & 7;
+	int page_shift = (srqc->log_page_size & 0x3f) + 12;
+
+	if (log_srq_size + log_rq_stride + 4 < page_shift)
+		return 1;
+
+	return 1 << (log_srq_size + log_rq_stride + 4 - page_shift);
+}
+
+int mlx4_SW2HW_SRQ_wrapper(struct mlx4_dev *dev, int slave,
+			   struct mlx4_vhcr *vhcr,
+			   struct mlx4_cmd_mailbox *inbox,
+			   struct mlx4_cmd_mailbox *outbox,
+			   struct mlx4_cmd_info *cmd)
+{
+	int err;
+	int srqn = vhcr->in_modifier;
+	struct res_mtt *mtt;
+	struct res_srq *srq = NULL;
+	struct mlx4_srq_context *srqc = inbox->buf;
+	int mtt_base = srq_get_mtt_addr(srqc) / dev->caps.mtt_entry_sz;
+
+	if (srqn != (be32_to_cpu(srqc->state_logsize_srqn) & 0xffffff))
+		return -EINVAL;
+
+	err = srq_res_start_move_to(dev, slave, srqn, RES_SRQ_HW, &srq);
+	if (err)
+		return err;
+	err = get_res(dev, slave, mtt_base, RES_MTT, &mtt);
+	if (err)
+		goto ex_abort;
+	err = check_mtt_range(dev, slave, mtt_base, srq_get_mtt_size(srqc),
+			      mtt);
+	if (err)
+		goto ex_put_mtt;
+
+	err = mlx4_DMA_wrapper(dev, slave, vhcr, inbox, outbox, cmd);
+	if (err)
+		goto ex_put_mtt;
+
+	atomic_inc(&mtt->ref_count);
+	srq->mtt = mtt;
+	put_res(dev, slave, mtt->com.res_id, RES_MTT);
+	res_end_move(dev, slave, RES_SRQ, srqn);
+	return 0;
+
+ex_put_mtt:
+	put_res(dev, slave, mtt->com.res_id, RES_MTT);
+ex_abort:
+	res_abort_move(dev, slave, RES_SRQ, srqn);
+
+	return err;
+}
+
+int mlx4_HW2SW_SRQ_wrapper(struct mlx4_dev *dev, int slave,
+			   struct mlx4_vhcr *vhcr,
+			   struct mlx4_cmd_mailbox *inbox,
+			   struct mlx4_cmd_mailbox *outbox,
+			   struct mlx4_cmd_info *cmd)
+{
+	int err;
+	int srqn = vhcr->in_modifier;
+	struct res_srq *srq = NULL;
+
+	err = srq_res_start_move_to(dev, slave, srqn, RES_SRQ_ALLOCATED, &srq);
+	if (err)
+		return err;
+	err = mlx4_DMA_wrapper(dev, slave, vhcr, inbox, outbox, cmd);
+	if (err)
+		goto ex_abort;
+	atomic_dec(&srq->mtt->ref_count);
+	if (srq->cq)
+		atomic_dec(&srq->cq->ref_count);
+	res_end_move(dev, slave, RES_SRQ, srqn);
+
+	return 0;
+
+ex_abort:
+	res_abort_move(dev, slave, RES_SRQ, srqn);
+
+	return err;
+}
+
+int mlx4_QUERY_SRQ_wrapper(struct mlx4_dev *dev, int slave,
+			   struct mlx4_vhcr *vhcr,
+			   struct mlx4_cmd_mailbox *inbox,
+			   struct mlx4_cmd_mailbox *outbox,
+			   struct mlx4_cmd_info *cmd)
+{
+	int err;
+	int srqn = vhcr->in_modifier;
+	struct res_srq *srq;
+
+	err = get_res(dev, slave, srqn, RES_SRQ, &srq);
+	if (err)
+		return err;
+	if (srq->com.from_state != RES_SRQ_HW) {
+		err = -EBUSY;
+		goto out;
+	}
+	err = mlx4_DMA_wrapper(dev, slave, vhcr, inbox, outbox, cmd);
+out:
+	put_res(dev, slave, srqn, RES_SRQ);
+	return err;
+}
+
+int mlx4_ARM_SRQ_wrapper(struct mlx4_dev *dev, int slave,
+			 struct mlx4_vhcr *vhcr,
+			 struct mlx4_cmd_mailbox *inbox,
+			 struct mlx4_cmd_mailbox *outbox,
+			 struct mlx4_cmd_info *cmd)
+{
+	int err;
+	int srqn = vhcr->in_modifier;
+	struct res_srq *srq;
+
+	err = get_res(dev, slave, srqn, RES_SRQ, &srq);
+	if (err)
+		return err;
+
+	if (srq->com.from_state != RES_SRQ_HW) {
+		err = -EBUSY;
+		goto out;
+	}
+
+	err = mlx4_DMA_wrapper(dev, slave, vhcr, inbox, outbox, cmd);
+out:
+	put_res(dev, slave, srqn, RES_SRQ);
+	return err;
+}
+
+int mlx4_GEN_QP_wrapper(struct mlx4_dev *dev, int slave,
+			struct mlx4_vhcr *vhcr,
+			struct mlx4_cmd_mailbox *inbox,
+			struct mlx4_cmd_mailbox *outbox,
+			struct mlx4_cmd_info *cmd)
+{
+	int err;
+	int qpn = vhcr->in_modifier & 0x7fffff;
+	struct res_qp *qp;
+
+	err = get_res(dev, slave, qpn, RES_QP, &qp);
+	if (err)
+		return err;
+	if (qp->com.from_state != RES_QP_HW) {
+		err = -EBUSY;
+		goto out;
+	}
+
+	err = mlx4_DMA_wrapper(dev, slave, vhcr, inbox, outbox, cmd);
+out:
+	put_res(dev, slave, qpn, RES_QP);
+	return err;
+}
+
+int mlx4_INIT2INIT_QP_wrapper(struct mlx4_dev *dev, int slave,
+			      struct mlx4_vhcr *vhcr,
+			      struct mlx4_cmd_mailbox *inbox,
+			      struct mlx4_cmd_mailbox *outbox,
+			      struct mlx4_cmd_info *cmd)
+{
+	struct mlx4_qp_context *context = inbox->buf + 8;
+	adjust_proxy_tun_qkey(dev, vhcr, context);
+	update_pkey_index(dev, slave, inbox);
+	return mlx4_GEN_QP_wrapper(dev, slave, vhcr, inbox, outbox, cmd);
+}
+
+static int adjust_qp_sched_queue(struct mlx4_dev *dev, int slave,
+				  struct mlx4_qp_context *qpc,
+				  struct mlx4_cmd_mailbox *inbox)
+{
+	enum mlx4_qp_optpar optpar = be32_to_cpu(*(__be32 *)inbox->buf);
+	u8 pri_sched_queue;
+	int port = mlx4_slave_convert_port(
+		   dev, slave, (qpc->pri_path.sched_queue >> 6 & 1) + 1) - 1;
+
+	if (port < 0)
+		return -EINVAL;
+
+	pri_sched_queue = (qpc->pri_path.sched_queue & ~(1 << 6)) |
+			  ((port & 1) << 6);
+
+	if (optpar & (MLX4_QP_OPTPAR_PRIMARY_ADDR_PATH | MLX4_QP_OPTPAR_SCHED_QUEUE) ||
+	    qpc->pri_path.sched_queue || mlx4_is_eth(dev, port + 1)) {
+		qpc->pri_path.sched_queue = pri_sched_queue;
+	}
+
+	if (optpar & MLX4_QP_OPTPAR_ALT_ADDR_PATH) {
+		port = mlx4_slave_convert_port(
+				dev, slave, (qpc->alt_path.sched_queue >> 6 & 1)
+				+ 1) - 1;
+		if (port < 0)
+			return -EINVAL;
+		qpc->alt_path.sched_queue =
+			(qpc->alt_path.sched_queue & ~(1 << 6)) |
+			(port & 1) << 6;
+	}
+	return 0;
+}
+
+static int roce_verify_mac(struct mlx4_dev *dev, int slave,
+				struct mlx4_qp_context *qpc,
+				struct mlx4_cmd_mailbox *inbox)
+{
+	u64 mac;
+	int port;
+	u32 ts = (be32_to_cpu(qpc->flags) >> 16) & 0xff;
+	u8 sched = *(u8 *)(inbox->buf + 64);
+	u8 smac_ix;
+
+	port = (sched >> 6 & 1) + 1;
+	if (mlx4_is_eth(dev, port) && (ts != MLX4_QP_ST_MLX)) {
+		smac_ix = qpc->pri_path.grh_mylmc & 0x7f;
+		if (mac_find_smac_ix_in_slave(dev, slave, port, smac_ix, &mac))
+			return -ENOENT;
+	}
+	return 0;
+}
+
+int mlx4_INIT2RTR_QP_wrapper(struct mlx4_dev *dev, int slave,
+			     struct mlx4_vhcr *vhcr,
+			     struct mlx4_cmd_mailbox *inbox,
+			     struct mlx4_cmd_mailbox *outbox,
+			     struct mlx4_cmd_info *cmd)
+{
+	int err;
+	struct mlx4_qp_context *qpc = inbox->buf + 8;
+	int qpn = vhcr->in_modifier & 0x7fffff;
+	struct res_qp *qp;
+	u8 orig_sched_queue;
+	u8 orig_vlan_control = qpc->pri_path.vlan_control;
+	u8 orig_fvl_rx = qpc->pri_path.fvl_rx;
+	u8 orig_pri_path_fl = qpc->pri_path.fl;
+	u8 orig_vlan_index = qpc->pri_path.vlan_index;
+	u8 orig_feup = qpc->pri_path.feup;
+
+	err = adjust_qp_sched_queue(dev, slave, qpc, inbox);
+	if (err)
+		return err;
+	err = verify_qp_parameters(dev, vhcr, inbox, QP_TRANS_INIT2RTR, slave);
+	if (err)
+		return err;
+
+	if (roce_verify_mac(dev, slave, qpc, inbox))
+		return -EINVAL;
+
+	update_pkey_index(dev, slave, inbox);
+	update_gid(dev, inbox, (u8)slave);
+	adjust_proxy_tun_qkey(dev, vhcr, qpc);
+	orig_sched_queue = qpc->pri_path.sched_queue;
+
+	err = get_res(dev, slave, qpn, RES_QP, &qp);
+	if (err)
+		return err;
+	if (qp->com.from_state != RES_QP_HW) {
+		err = -EBUSY;
+		goto out;
+	}
+
+	err = update_vport_qp_param(dev, inbox, slave, qpn);
+	if (err)
+		goto out;
+
+	err = mlx4_DMA_wrapper(dev, slave, vhcr, inbox, outbox, cmd);
+out:
+	/* if no error, save sched queue value passed in by VF. This is
+	 * essentially the QOS value provided by the VF. This will be useful
+	 * if we allow dynamic changes from VST back to VGT
+	 */
+	if (!err) {
+		qp->sched_queue = orig_sched_queue;
+		qp->vlan_control = orig_vlan_control;
+		qp->fvl_rx	=  orig_fvl_rx;
+		qp->pri_path_fl = orig_pri_path_fl;
+		qp->vlan_index  = orig_vlan_index;
+		qp->feup	= orig_feup;
+	}
+	put_res(dev, slave, qpn, RES_QP);
+	return err;
+}
+
+int mlx4_RTR2RTS_QP_wrapper(struct mlx4_dev *dev, int slave,
+			    struct mlx4_vhcr *vhcr,
+			    struct mlx4_cmd_mailbox *inbox,
+			    struct mlx4_cmd_mailbox *outbox,
+			    struct mlx4_cmd_info *cmd)
+{
+	int err;
+	struct mlx4_qp_context *context = inbox->buf + 8;
+
+	err = adjust_qp_sched_queue(dev, slave, context, inbox);
+	if (err)
+		return err;
+	err = verify_qp_parameters(dev, vhcr, inbox, QP_TRANS_RTR2RTS, slave);
+	if (err)
+		return err;
+
+	update_pkey_index(dev, slave, inbox);
+	update_gid(dev, inbox, (u8)slave);
+	adjust_proxy_tun_qkey(dev, vhcr, context);
+	return mlx4_GEN_QP_wrapper(dev, slave, vhcr, inbox, outbox, cmd);
+}
+
+int mlx4_RTS2RTS_QP_wrapper(struct mlx4_dev *dev, int slave,
+			    struct mlx4_vhcr *vhcr,
+			    struct mlx4_cmd_mailbox *inbox,
+			    struct mlx4_cmd_mailbox *outbox,
+			    struct mlx4_cmd_info *cmd)
+{
+	int err;
+	struct mlx4_qp_context *context = inbox->buf + 8;
+
+	err = adjust_qp_sched_queue(dev, slave, context, inbox);
+	if (err)
+		return err;
+	err = verify_qp_parameters(dev, vhcr, inbox, QP_TRANS_RTS2RTS, slave);
+	if (err)
+		return err;
+
+	update_pkey_index(dev, slave, inbox);
+	update_gid(dev, inbox, (u8)slave);
+	adjust_proxy_tun_qkey(dev, vhcr, context);
+	return mlx4_GEN_QP_wrapper(dev, slave, vhcr, inbox, outbox, cmd);
+}
+
+
+int mlx4_SQERR2RTS_QP_wrapper(struct mlx4_dev *dev, int slave,
+			      struct mlx4_vhcr *vhcr,
+			      struct mlx4_cmd_mailbox *inbox,
+			      struct mlx4_cmd_mailbox *outbox,
+			      struct mlx4_cmd_info *cmd)
+{
+	struct mlx4_qp_context *context = inbox->buf + 8;
+	int err = adjust_qp_sched_queue(dev, slave, context, inbox);
+	if (err)
+		return err;
+	adjust_proxy_tun_qkey(dev, vhcr, context);
+	return mlx4_GEN_QP_wrapper(dev, slave, vhcr, inbox, outbox, cmd);
+}
+
+int mlx4_SQD2SQD_QP_wrapper(struct mlx4_dev *dev, int slave,
+			    struct mlx4_vhcr *vhcr,
+			    struct mlx4_cmd_mailbox *inbox,
+			    struct mlx4_cmd_mailbox *outbox,
+			    struct mlx4_cmd_info *cmd)
+{
+	int err;
+	struct mlx4_qp_context *context = inbox->buf + 8;
+
+	err = adjust_qp_sched_queue(dev, slave, context, inbox);
+	if (err)
+		return err;
+	err = verify_qp_parameters(dev, vhcr, inbox, QP_TRANS_SQD2SQD, slave);
+	if (err)
+		return err;
+
+	adjust_proxy_tun_qkey(dev, vhcr, context);
+	update_gid(dev, inbox, (u8)slave);
+	update_pkey_index(dev, slave, inbox);
+	return mlx4_GEN_QP_wrapper(dev, slave, vhcr, inbox, outbox, cmd);
+}
+
+int mlx4_SQD2RTS_QP_wrapper(struct mlx4_dev *dev, int slave,
+			    struct mlx4_vhcr *vhcr,
+			    struct mlx4_cmd_mailbox *inbox,
+			    struct mlx4_cmd_mailbox *outbox,
+			    struct mlx4_cmd_info *cmd)
+{
+	int err;
+	struct mlx4_qp_context *context = inbox->buf + 8;
+
+	err = adjust_qp_sched_queue(dev, slave, context, inbox);
+	if (err)
+		return err;
+	err = verify_qp_parameters(dev, vhcr, inbox, QP_TRANS_SQD2RTS, slave);
+	if (err)
+		return err;
+
+	adjust_proxy_tun_qkey(dev, vhcr, context);
+	update_gid(dev, inbox, (u8)slave);
+	update_pkey_index(dev, slave, inbox);
+	return mlx4_GEN_QP_wrapper(dev, slave, vhcr, inbox, outbox, cmd);
+}
+
+int mlx4_2RST_QP_wrapper(struct mlx4_dev *dev, int slave,
+			 struct mlx4_vhcr *vhcr,
+			 struct mlx4_cmd_mailbox *inbox,
+			 struct mlx4_cmd_mailbox *outbox,
+			 struct mlx4_cmd_info *cmd)
+{
+	int err;
+	int qpn = vhcr->in_modifier & 0x7fffff;
+	struct res_qp *qp;
+
+	err = qp_res_start_move_to(dev, slave, qpn, RES_QP_MAPPED, &qp, 0);
+	if (err)
+		return err;
+	err = mlx4_DMA_wrapper(dev, slave, vhcr, inbox, outbox, cmd);
+	if (err)
+		goto ex_abort;
+
+	atomic_dec(&qp->mtt->ref_count);
+	atomic_dec(&qp->rcq->ref_count);
+	atomic_dec(&qp->scq->ref_count);
+	if (qp->srq)
+		atomic_dec(&qp->srq->ref_count);
+	res_end_move(dev, slave, RES_QP, qpn);
+	return 0;
+
+ex_abort:
+	res_abort_move(dev, slave, RES_QP, qpn);
+
+	return err;
+}
+
+static struct res_gid *find_gid(struct mlx4_dev *dev, int slave,
+				struct res_qp *rqp, u8 *gid)
+{
+	struct res_gid *res;
+
+	list_for_each_entry(res, &rqp->mcg_list, list) {
+		if (!memcmp(res->gid, gid, 16))
+			return res;
+	}
+	return NULL;
+}
+
+static int add_mcg_res(struct mlx4_dev *dev, int slave, struct res_qp *rqp,
+		       u8 *gid, enum mlx4_protocol prot,
+		       enum mlx4_steer_type steer, u64 reg_id)
+{
+	struct res_gid *res;
+	int err;
+
+	res = kzalloc(sizeof(*res), GFP_KERNEL);
+	if (!res)
+		return -ENOMEM;
+
+	spin_lock_irq(&rqp->mcg_spl);
+	if (find_gid(dev, slave, rqp, gid)) {
+		kfree(res);
+		err = -EEXIST;
+	} else {
+		memcpy(res->gid, gid, 16);
+		res->prot = prot;
+		res->steer = steer;
+		res->reg_id = reg_id;
+		list_add_tail(&res->list, &rqp->mcg_list);
+		err = 0;
+	}
+	spin_unlock_irq(&rqp->mcg_spl);
+
+	return err;
+}
+
+static int rem_mcg_res(struct mlx4_dev *dev, int slave, struct res_qp *rqp,
+		       u8 *gid, enum mlx4_protocol prot,
+		       enum mlx4_steer_type steer, u64 *reg_id)
+{
+	struct res_gid *res;
+	int err;
+
+	spin_lock_irq(&rqp->mcg_spl);
+	res = find_gid(dev, slave, rqp, gid);
+	if (!res || res->prot != prot || res->steer != steer)
+		err = -EINVAL;
+	else {
+		*reg_id = res->reg_id;
+		list_del(&res->list);
+		kfree(res);
+		err = 0;
+	}
+	spin_unlock_irq(&rqp->mcg_spl);
+
+	return err;
+}
+
+static int qp_attach(struct mlx4_dev *dev, int slave, struct mlx4_qp *qp,
+		     u8 gid[16], int block_loopback, enum mlx4_protocol prot,
+		     enum mlx4_steer_type type, u64 *reg_id)
+{
+	switch (dev->caps.steering_mode) {
+	case MLX4_STEERING_MODE_DEVICE_MANAGED: {
+		int port = mlx4_slave_convert_port(dev, slave, gid[5]);
+		if (port < 0)
+			return port;
+		return mlx4_trans_to_dmfs_attach(dev, qp, gid, port,
+						block_loopback, prot,
+						reg_id);
+	}
+	case MLX4_STEERING_MODE_B0:
+		if (prot == MLX4_PROT_ETH) {
+			int port = mlx4_slave_convert_port(dev, slave, gid[5]);
+			if (port < 0)
+				return port;
+			gid[5] = port;
+		}
+		return mlx4_qp_attach_common(dev, qp, gid,
+					    block_loopback, prot, type);
+	default:
+		return -EINVAL;
+	}
+}
+
+static int qp_detach(struct mlx4_dev *dev, struct mlx4_qp *qp,
+		     u8 gid[16], enum mlx4_protocol prot,
+		     enum mlx4_steer_type type, u64 reg_id)
+{
+	switch (dev->caps.steering_mode) {
+	case MLX4_STEERING_MODE_DEVICE_MANAGED:
+		return mlx4_flow_detach(dev, reg_id);
+	case MLX4_STEERING_MODE_B0:
+		return mlx4_qp_detach_common(dev, qp, gid, prot, type);
+	default:
+		return -EINVAL;
+	}
+}
+
+static int mlx4_adjust_port(struct mlx4_dev *dev, int slave,
+			    u8 *gid, enum mlx4_protocol prot)
+{
+	int real_port;
+
+	if (prot != MLX4_PROT_ETH)
+		return 0;
+
+	if (dev->caps.steering_mode == MLX4_STEERING_MODE_B0 ||
+	    dev->caps.steering_mode == MLX4_STEERING_MODE_DEVICE_MANAGED) {
+		real_port = mlx4_slave_convert_port(dev, slave, gid[5]);
+		if (real_port < 0)
+			return -EINVAL;
+		gid[5] = real_port;
+	}
+
+	return 0;
+}
+
+int mlx4_QP_ATTACH_wrapper(struct mlx4_dev *dev, int slave,
+			       struct mlx4_vhcr *vhcr,
+			       struct mlx4_cmd_mailbox *inbox,
+			       struct mlx4_cmd_mailbox *outbox,
+			       struct mlx4_cmd_info *cmd)
+{
+	struct mlx4_qp qp; /* dummy for calling attach/detach */
+	u8 *gid = inbox->buf;
+	enum mlx4_protocol prot = (vhcr->in_modifier >> 28) & 0x7;
+	int err;
+	int qpn;
+	struct res_qp *rqp;
+	u64 reg_id = 0;
+	int attach = vhcr->op_modifier;
+	int block_loopback = vhcr->in_modifier >> 31;
+	u8 steer_type_mask = 2;
+	enum mlx4_steer_type type = (gid[7] & steer_type_mask) >> 1;
+
+	qpn = vhcr->in_modifier & 0xffffff;
+	err = get_res(dev, slave, qpn, RES_QP, &rqp);
+	if (err)
+		return err;
+
+	qp.qpn = qpn;
+	if (attach) {
+		err = qp_attach(dev, slave, &qp, gid, block_loopback, prot,
+				type, &reg_id);
+		if (err) {
+			pr_err("Fail to attach rule to qp 0x%x\n", qpn);
+			goto ex_put;
+		}
+		err = add_mcg_res(dev, slave, rqp, gid, prot, type, reg_id);
+		if (err)
+			goto ex_detach;
+	} else {
+		err = mlx4_adjust_port(dev, slave, gid, prot);
+		if (err)
+			goto ex_put;
+
+		err = rem_mcg_res(dev, slave, rqp, gid, prot, type, &reg_id);
+		if (err)
+			goto ex_put;
+
+		err = qp_detach(dev, &qp, gid, prot, type, reg_id);
+		if (err)
+			pr_err("Fail to detach rule from qp 0x%x reg_id = 0x%llx\n",
+			       qpn, reg_id);
+	}
+	put_res(dev, slave, qpn, RES_QP);
+	return err;
+
+ex_detach:
+	qp_detach(dev, &qp, gid, prot, type, reg_id);
+ex_put:
+	put_res(dev, slave, qpn, RES_QP);
+	return err;
+}
+
+/*
+ * MAC validation for Flow Steering rules.
+ * VF can attach rules only with a mac address which is assigned to it.
+ */
+static int validate_eth_header_mac(int slave, struct _rule_hw *eth_header,
+				   struct list_head *rlist)
+{
+	struct mac_res *res, *tmp;
+	__be64 be_mac;
+
+	/* make sure it isn't multicast or broadcast mac*/
+	if (!is_multicast_ether_addr(eth_header->eth.dst_mac) &&
+	    !is_broadcast_ether_addr(eth_header->eth.dst_mac)) {
+		list_for_each_entry_safe(res, tmp, rlist, list) {
+			be_mac = cpu_to_be64(res->mac << 16);
+			if (ether_addr_equal((u8 *)&be_mac, eth_header->eth.dst_mac))
+				return 0;
+		}
+		pr_err("MAC %pM doesn't belong to VF %d, Steering rule rejected\n",
+		       eth_header->eth.dst_mac, slave);
+		return -EINVAL;
+	}
+	return 0;
+}
+
+/*
+ * In case of missing eth header, append eth header with a MAC address
+ * assigned to the VF.
+ */
+static int add_eth_header(struct mlx4_dev *dev, int slave,
+			  struct mlx4_cmd_mailbox *inbox,
+			  struct list_head *rlist, int header_id)
+{
+	struct mac_res *res, *tmp;
+	u8 port;
+	struct mlx4_net_trans_rule_hw_ctrl *ctrl;
+	struct mlx4_net_trans_rule_hw_eth *eth_header;
+	struct mlx4_net_trans_rule_hw_ipv4 *ip_header;
+	struct mlx4_net_trans_rule_hw_tcp_udp *l4_header;
+	__be64 be_mac = 0;
+	__be64 mac_msk = cpu_to_be64(MLX4_MAC_MASK << 16);
+
+	ctrl = (struct mlx4_net_trans_rule_hw_ctrl *)inbox->buf;
+	port = ctrl->port;
+	eth_header = (struct mlx4_net_trans_rule_hw_eth *)(ctrl + 1);
+
+	/* Clear a space in the inbox for eth header */
+	switch (header_id) {
+	case MLX4_NET_TRANS_RULE_ID_IPV4:
+		ip_header =
+			(struct mlx4_net_trans_rule_hw_ipv4 *)(eth_header + 1);
+		memmove(ip_header, eth_header,
+			sizeof(*ip_header) + sizeof(*l4_header));
+		break;
+	case MLX4_NET_TRANS_RULE_ID_TCP:
+	case MLX4_NET_TRANS_RULE_ID_UDP:
+		l4_header = (struct mlx4_net_trans_rule_hw_tcp_udp *)
+			    (eth_header + 1);
+		memmove(l4_header, eth_header, sizeof(*l4_header));
+		break;
+	default:
+		return -EINVAL;
+	}
+	list_for_each_entry_safe(res, tmp, rlist, list) {
+		if (port == res->port) {
+			be_mac = cpu_to_be64(res->mac << 16);
+			break;
+		}
+	}
+	if (!be_mac) {
+		pr_err("Failed adding eth header to FS rule, Can't find matching MAC for port %d\n",
+		       port);
+		return -EINVAL;
+	}
+
+	memset(eth_header, 0, sizeof(*eth_header));
+	eth_header->size = sizeof(*eth_header) >> 2;
+	eth_header->id = cpu_to_be16(__sw_id_hw[MLX4_NET_TRANS_RULE_ID_ETH]);
+	memcpy(eth_header->dst_mac, &be_mac, ETH_ALEN);
+	memcpy(eth_header->dst_mac_msk, &mac_msk, ETH_ALEN);
+
+	return 0;
+
+}
+
+#define MLX4_UPD_QP_PATH_MASK_SUPPORTED      (                                \
+	1ULL << MLX4_UPD_QP_PATH_MASK_MAC_INDEX                     |\
+	1ULL << MLX4_UPD_QP_PATH_MASK_ETH_SRC_CHECK_MC_LB)
+int mlx4_UPDATE_QP_wrapper(struct mlx4_dev *dev, int slave,
+			   struct mlx4_vhcr *vhcr,
+			   struct mlx4_cmd_mailbox *inbox,
+			   struct mlx4_cmd_mailbox *outbox,
+			   struct mlx4_cmd_info *cmd_info)
+{
+	int err;
+	u32 qpn = vhcr->in_modifier & 0xffffff;
+	struct res_qp *rqp;
+	u64 mac;
+	unsigned port;
+	u64 pri_addr_path_mask;
+	struct mlx4_update_qp_context *cmd;
+	int smac_index;
+
+	cmd = (struct mlx4_update_qp_context *)inbox->buf;
+
+	pri_addr_path_mask = be64_to_cpu(cmd->primary_addr_path_mask);
+	if (cmd->qp_mask || cmd->secondary_addr_path_mask ||
+	    (pri_addr_path_mask & ~MLX4_UPD_QP_PATH_MASK_SUPPORTED))
+		return -EPERM;
+
+	if ((pri_addr_path_mask &
+	     (1ULL << MLX4_UPD_QP_PATH_MASK_ETH_SRC_CHECK_MC_LB)) &&
+		!(dev->caps.flags2 &
+		  MLX4_DEV_CAP_FLAG2_UPDATE_QP_SRC_CHECK_LB)) {
+		mlx4_warn(dev, "Src check LB for slave %d isn't supported\n",
+			  slave);
+		return -EOPNOTSUPP;
+	}
+
+	/* Just change the smac for the QP */
+	err = get_res(dev, slave, qpn, RES_QP, &rqp);
+	if (err) {
+		mlx4_err(dev, "Updating qpn 0x%x for slave %d rejected\n", qpn, slave);
+		return err;
+	}
+
+	port = (rqp->sched_queue >> 6 & 1) + 1;
+
+	if (pri_addr_path_mask & (1ULL << MLX4_UPD_QP_PATH_MASK_MAC_INDEX)) {
+		smac_index = cmd->qp_context.pri_path.grh_mylmc;
+		err = mac_find_smac_ix_in_slave(dev, slave, port,
+						smac_index, &mac);
+
+		if (err) {
+			mlx4_err(dev, "Failed to update qpn 0x%x, MAC is invalid. smac_ix: %d\n",
+				 qpn, smac_index);
+			goto err_mac;
+		}
+	}
+
+	err = mlx4_cmd(dev, inbox->dma,
+		       vhcr->in_modifier, 0,
+		       MLX4_CMD_UPDATE_QP, MLX4_CMD_TIME_CLASS_A,
+		       MLX4_CMD_NATIVE);
+	if (err) {
+		mlx4_err(dev, "Failed to update qpn on qpn 0x%x, command failed\n", qpn);
+		goto err_mac;
+	}
+
+err_mac:
+	put_res(dev, slave, qpn, RES_QP);
+	return err;
+}
+
+static u32 qp_attach_mbox_size(void *mbox)
+{
+	u32 size = sizeof(struct mlx4_net_trans_rule_hw_ctrl);
+	struct _rule_hw  *rule_header;
+
+	rule_header = (struct _rule_hw *)(mbox + size);
+
+	while (rule_header->size) {
+		size += rule_header->size * sizeof(u32);
+		rule_header += 1;
+	}
+	return size;
+}
+
+static int mlx4_do_mirror_rule(struct mlx4_dev *dev, struct res_fs_rule *fs_rule);
+
+int mlx4_QP_FLOW_STEERING_ATTACH_wrapper(struct mlx4_dev *dev, int slave,
+					 struct mlx4_vhcr *vhcr,
+					 struct mlx4_cmd_mailbox *inbox,
+					 struct mlx4_cmd_mailbox *outbox,
+					 struct mlx4_cmd_info *cmd)
+{
+
+	struct mlx4_priv *priv = mlx4_priv(dev);
+	struct mlx4_resource_tracker *tracker = &priv->mfunc.master.res_tracker;
+	struct list_head *rlist = &tracker->slave_list[slave].res_list[RES_MAC];
+	int err;
+	int qpn;
+	struct res_qp *rqp;
+	struct mlx4_net_trans_rule_hw_ctrl *ctrl;
+	struct _rule_hw  *rule_header;
+	int header_id;
+	struct res_fs_rule *rrule;
+	u32 mbox_size;
+
+	if (dev->caps.steering_mode !=
+	    MLX4_STEERING_MODE_DEVICE_MANAGED)
+		return -EOPNOTSUPP;
+
+	ctrl = (struct mlx4_net_trans_rule_hw_ctrl *)inbox->buf;
+	err = mlx4_slave_convert_port(dev, slave, ctrl->port);
+	if (err <= 0)
+		return -EINVAL;
+	ctrl->port = err;
+	qpn = be32_to_cpu(ctrl->qpn) & 0xffffff;
+	err = get_res(dev, slave, qpn, RES_QP, &rqp);
+	if (err) {
+		pr_err("Steering rule with qpn 0x%x rejected\n", qpn);
+		return err;
+	}
+	rule_header = (struct _rule_hw *)(ctrl + 1);
+	header_id = map_hw_to_sw_id(be16_to_cpu(rule_header->id));
+
+	if (header_id == MLX4_NET_TRANS_RULE_ID_ETH)
+		mlx4_handle_eth_header_mcast_prio(ctrl, rule_header);
+
+	switch (header_id) {
+	case MLX4_NET_TRANS_RULE_ID_ETH:
+		if (validate_eth_header_mac(slave, rule_header, rlist)) {
+			err = -EINVAL;
+			goto err_put_qp;
+		}
+		break;
+	case MLX4_NET_TRANS_RULE_ID_IB:
+		break;
+	case MLX4_NET_TRANS_RULE_ID_IPV4:
+	case MLX4_NET_TRANS_RULE_ID_TCP:
+	case MLX4_NET_TRANS_RULE_ID_UDP:
+		pr_warn("Can't attach FS rule without L2 headers, adding L2 header\n");
+		if (add_eth_header(dev, slave, inbox, rlist, header_id)) {
+			err = -EINVAL;
+			goto err_put_qp;
+		}
+		vhcr->in_modifier +=
+			sizeof(struct mlx4_net_trans_rule_hw_eth) >> 2;
+		break;
+	default:
+		pr_err("Corrupted mailbox\n");
+		err = -EINVAL;
+		goto err_put_qp;
+	}
+
+	err = mlx4_cmd_imm(dev, inbox->dma, &vhcr->out_param,
+			   vhcr->in_modifier, 0,
+			   MLX4_QP_FLOW_STEERING_ATTACH, MLX4_CMD_TIME_CLASS_A,
+			   MLX4_CMD_NATIVE);
+	if (err)
+		goto err_put_qp;
+
+
+	err = add_res_range(dev, slave, vhcr->out_param, 1, RES_FS_RULE, qpn);
+	if (err) {
+		mlx4_err(dev, "Fail to add flow steering resources\n");
+		goto err_detach;
+	}
+
+	err = get_res(dev, slave, vhcr->out_param, RES_FS_RULE, &rrule);
+	if (err)
+		goto err_detach;
+
+	mbox_size = qp_attach_mbox_size(inbox->buf);
+	rrule->mirr_mbox = kmalloc(mbox_size, GFP_KERNEL);
+	if (!rrule->mirr_mbox) {
+		err = -ENOMEM;
+		goto err_put_rule;
+	}
+	rrule->mirr_mbox_size = mbox_size;
+	rrule->mirr_rule_id = 0;
+	memcpy(rrule->mirr_mbox, inbox->buf, mbox_size);
+
+	/* set different port */
+	ctrl = (struct mlx4_net_trans_rule_hw_ctrl *)rrule->mirr_mbox;
+	if (ctrl->port == 1)
+		ctrl->port = 2;
+	else
+		ctrl->port = 1;
+
+	if (mlx4_is_bonded(dev))
+		mlx4_do_mirror_rule(dev, rrule);
+
+	atomic_inc(&rqp->ref_count);
+
+err_put_rule:
+	put_res(dev, slave, vhcr->out_param, RES_FS_RULE);
+err_detach:
+	/* detach rule on error */
+	if (err)
+		mlx4_cmd(dev, vhcr->out_param, 0, 0,
+			 MLX4_QP_FLOW_STEERING_DETACH, MLX4_CMD_TIME_CLASS_A,
+			 MLX4_CMD_NATIVE);
+err_put_qp:
+	put_res(dev, slave, qpn, RES_QP);
+	return err;
+}
+
+static int mlx4_undo_mirror_rule(struct mlx4_dev *dev, struct res_fs_rule *fs_rule)
+{
+	int err;
+
+	err = rem_res_range(dev, fs_rule->com.owner, fs_rule->com.res_id, 1, RES_FS_RULE, 0);
+	if (err) {
+		mlx4_err(dev, "Fail to remove flow steering resources\n");
+		return err;
+	}
+
+	mlx4_cmd(dev, fs_rule->com.res_id, 0, 0, MLX4_QP_FLOW_STEERING_DETACH,
+		 MLX4_CMD_TIME_CLASS_A, MLX4_CMD_NATIVE);
+	return 0;
+}
+
+int mlx4_QP_FLOW_STEERING_DETACH_wrapper(struct mlx4_dev *dev, int slave,
+					 struct mlx4_vhcr *vhcr,
+					 struct mlx4_cmd_mailbox *inbox,
+					 struct mlx4_cmd_mailbox *outbox,
+					 struct mlx4_cmd_info *cmd)
+{
+	int err;
+	struct res_qp *rqp;
+	struct res_fs_rule *rrule;
+	u64 mirr_reg_id;
+	int qpn;
+
+	if (dev->caps.steering_mode !=
+	    MLX4_STEERING_MODE_DEVICE_MANAGED)
+		return -EOPNOTSUPP;
+
+	err = get_res(dev, slave, vhcr->in_param, RES_FS_RULE, &rrule);
+	if (err)
+		return err;
+
+	if (!rrule->mirr_mbox) {
+		mlx4_err(dev, "Mirror rules cannot be removed explicitly\n");
+		put_res(dev, slave, vhcr->in_param, RES_FS_RULE);
+		return -EINVAL;
+	}
+	mirr_reg_id = rrule->mirr_rule_id;
+	kfree(rrule->mirr_mbox);
+	qpn = rrule->qpn;
+
+	/* Release the rule form busy state before removal */
+	put_res(dev, slave, vhcr->in_param, RES_FS_RULE);
+	err = get_res(dev, slave, qpn, RES_QP, &rqp);
+	if (err)
+		return err;
+
+	if (mirr_reg_id && mlx4_is_bonded(dev)) {
+		err = get_res(dev, slave, mirr_reg_id, RES_FS_RULE, &rrule);
+		if (err) {
+			mlx4_err(dev, "Fail to get resource of mirror rule\n");
+		} else {
+			put_res(dev, slave, mirr_reg_id, RES_FS_RULE);
+			mlx4_undo_mirror_rule(dev, rrule);
+		}
+	}
+	err = rem_res_range(dev, slave, vhcr->in_param, 1, RES_FS_RULE, 0);
+	if (err) {
+		mlx4_err(dev, "Fail to remove flow steering resources\n");
+		goto out;
+	}
+
+	err = mlx4_cmd(dev, vhcr->in_param, 0, 0,
+		       MLX4_QP_FLOW_STEERING_DETACH, MLX4_CMD_TIME_CLASS_A,
+		       MLX4_CMD_NATIVE);
+	if (!err)
+		atomic_dec(&rqp->ref_count);
+out:
+	put_res(dev, slave, qpn, RES_QP);
+	return err;
+}
+
+enum {
+	BUSY_MAX_RETRIES = 10
+};
+
+int mlx4_QUERY_IF_STAT_wrapper(struct mlx4_dev *dev, int slave,
+			       struct mlx4_vhcr *vhcr,
+			       struct mlx4_cmd_mailbox *inbox,
+			       struct mlx4_cmd_mailbox *outbox,
+			       struct mlx4_cmd_info *cmd)
+{
+	int err;
+	int index = vhcr->in_modifier & 0xffff;
+
+	err = get_res(dev, slave, index, RES_COUNTER, NULL);
+	if (err)
+		return err;
+
+	err = mlx4_DMA_wrapper(dev, slave, vhcr, inbox, outbox, cmd);
+	put_res(dev, slave, index, RES_COUNTER);
+	return err;
+}
+
+static void detach_qp(struct mlx4_dev *dev, int slave, struct res_qp *rqp)
+{
+	struct res_gid *rgid;
+	struct res_gid *tmp;
+	struct mlx4_qp qp; /* dummy for calling attach/detach */
+
+	list_for_each_entry_safe(rgid, tmp, &rqp->mcg_list, list) {
+		switch (dev->caps.steering_mode) {
+		case MLX4_STEERING_MODE_DEVICE_MANAGED:
+			mlx4_flow_detach(dev, rgid->reg_id);
+			break;
+		case MLX4_STEERING_MODE_B0:
+			qp.qpn = rqp->local_qpn;
+			(void) mlx4_qp_detach_common(dev, &qp, rgid->gid,
+						     rgid->prot, rgid->steer);
+			break;
+		}
+		list_del(&rgid->list);
+		kfree(rgid);
+	}
+}
+
+static int _move_all_busy(struct mlx4_dev *dev, int slave,
+			  enum mlx4_resource type, int print)
+{
+	struct mlx4_priv *priv = mlx4_priv(dev);
+	struct mlx4_resource_tracker *tracker =
+		&priv->mfunc.master.res_tracker;
+	struct list_head *rlist = &tracker->slave_list[slave].res_list[type];
+	struct res_common *r;
+	struct res_common *tmp;
+	int busy;
+
+	busy = 0;
+	spin_lock_irq(mlx4_tlock(dev));
+	list_for_each_entry_safe(r, tmp, rlist, list) {
+		if (r->owner == slave) {
+			if (!r->removing) {
+				if (r->state == RES_ANY_BUSY) {
+					if (print)
+						mlx4_dbg(dev,
+							 "%s id 0x%llx is busy\n",
+							  resource_str(type),
+							  r->res_id);
+					++busy;
+				} else {
+					r->from_state = r->state;
+					r->state = RES_ANY_BUSY;
+					r->removing = 1;
+				}
+			}
+		}
+	}
+	spin_unlock_irq(mlx4_tlock(dev));
+
+	return busy;
+}
+
+static int move_all_busy(struct mlx4_dev *dev, int slave,
+			 enum mlx4_resource type)
+{
+	unsigned long begin;
+	int busy;
+
+	begin = jiffies;
+	do {
+		busy = _move_all_busy(dev, slave, type, 0);
+		if (time_after(jiffies, begin + 5 * HZ))
+			break;
+		if (busy)
+			cond_resched();
+	} while (busy);
+
+	if (busy)
+		busy = _move_all_busy(dev, slave, type, 1);
+
+	return busy;
+}
+static void rem_slave_qps(struct mlx4_dev *dev, int slave)
+{
+	struct mlx4_priv *priv = mlx4_priv(dev);
+	struct mlx4_resource_tracker *tracker = &priv->mfunc.master.res_tracker;
+	struct list_head *qp_list =
+		&tracker->slave_list[slave].res_list[RES_QP];
+	struct res_qp *qp;
+	struct res_qp *tmp;
+	int state;
+	u64 in_param;
+	int qpn;
+	int err;
+
+	err = move_all_busy(dev, slave, RES_QP);
+	if (err)
+		mlx4_warn(dev, "rem_slave_qps: Could not move all qps to busy for slave %d\n",
+			  slave);
+
+	spin_lock_irq(mlx4_tlock(dev));
+	list_for_each_entry_safe(qp, tmp, qp_list, com.list) {
+		spin_unlock_irq(mlx4_tlock(dev));
+		if (qp->com.owner == slave) {
+			qpn = qp->com.res_id;
+			detach_qp(dev, slave, qp);
+			state = qp->com.from_state;
+			while (state != 0) {
+				switch (state) {
+				case RES_QP_RESERVED:
+					spin_lock_irq(mlx4_tlock(dev));
+					rb_erase(&qp->com.node,
+						 &tracker->res_tree[RES_QP]);
+					list_del(&qp->com.list);
+					spin_unlock_irq(mlx4_tlock(dev));
+					if (!valid_reserved(dev, slave, qpn)) {
+						__mlx4_qp_release_range(dev, qpn, 1);
+						mlx4_release_resource(dev, slave,
+								      RES_QP, 1, 0);
+					}
+					kfree(qp);
+					state = 0;
+					break;
+				case RES_QP_MAPPED:
+					if (!valid_reserved(dev, slave, qpn))
+						__mlx4_qp_free_icm(dev, qpn);
+					state = RES_QP_RESERVED;
+					break;
+				case RES_QP_HW:
+					in_param = slave;
+					err = mlx4_cmd(dev, in_param,
+						       qp->local_qpn, 2,
+						       MLX4_CMD_2RST_QP,
+						       MLX4_CMD_TIME_CLASS_A,
+						       MLX4_CMD_NATIVE);
+					if (err)
+						mlx4_dbg(dev, "rem_slave_qps: failed to move slave %d qpn %d to reset\n",
+							 slave, qp->local_qpn);
+					atomic_dec(&qp->rcq->ref_count);
+					atomic_dec(&qp->scq->ref_count);
+					atomic_dec(&qp->mtt->ref_count);
+					if (qp->srq)
+						atomic_dec(&qp->srq->ref_count);
+					state = RES_QP_MAPPED;
+					break;
+				default:
+					state = 0;
+				}
+			}
+		}
+		spin_lock_irq(mlx4_tlock(dev));
+	}
+	spin_unlock_irq(mlx4_tlock(dev));
+}
+
+static void rem_slave_srqs(struct mlx4_dev *dev, int slave)
+{
+	struct mlx4_priv *priv = mlx4_priv(dev);
+	struct mlx4_resource_tracker *tracker = &priv->mfunc.master.res_tracker;
+	struct list_head *srq_list =
+		&tracker->slave_list[slave].res_list[RES_SRQ];
+	struct res_srq *srq;
+	struct res_srq *tmp;
+	int state;
+	u64 in_param;
+	LIST_HEAD(tlist);
+	int srqn;
+	int err;
+
+	err = move_all_busy(dev, slave, RES_SRQ);
+	if (err)
+		mlx4_warn(dev, "rem_slave_srqs: Could not move all srqs - too busy for slave %d\n",
+			  slave);
+
+	spin_lock_irq(mlx4_tlock(dev));
+	list_for_each_entry_safe(srq, tmp, srq_list, com.list) {
+		spin_unlock_irq(mlx4_tlock(dev));
+		if (srq->com.owner == slave) {
+			srqn = srq->com.res_id;
+			state = srq->com.from_state;
+			while (state != 0) {
+				switch (state) {
+				case RES_SRQ_ALLOCATED:
+					__mlx4_srq_free_icm(dev, srqn);
+					spin_lock_irq(mlx4_tlock(dev));
+					rb_erase(&srq->com.node,
+						 &tracker->res_tree[RES_SRQ]);
+					list_del(&srq->com.list);
+					spin_unlock_irq(mlx4_tlock(dev));
+					mlx4_release_resource(dev, slave,
+							      RES_SRQ, 1, 0);
+					kfree(srq);
+					state = 0;
+					break;
+
+				case RES_SRQ_HW:
+					in_param = slave;
+					err = mlx4_cmd(dev, in_param, srqn, 1,
+						       MLX4_CMD_HW2SW_SRQ,
+						       MLX4_CMD_TIME_CLASS_A,
+						       MLX4_CMD_NATIVE);
+					if (err)
+						mlx4_dbg(dev, "rem_slave_srqs: failed to move slave %d srq %d to SW ownership\n",
+							 slave, srqn);
+
+					atomic_dec(&srq->mtt->ref_count);
+					if (srq->cq)
+						atomic_dec(&srq->cq->ref_count);
+					state = RES_SRQ_ALLOCATED;
+					break;
+
+				default:
+					state = 0;
+				}
+			}
+		}
+		spin_lock_irq(mlx4_tlock(dev));
+	}
+	spin_unlock_irq(mlx4_tlock(dev));
+}
+
+static void rem_slave_cqs(struct mlx4_dev *dev, int slave)
+{
+	struct mlx4_priv *priv = mlx4_priv(dev);
+	struct mlx4_resource_tracker *tracker = &priv->mfunc.master.res_tracker;
+	struct list_head *cq_list =
+		&tracker->slave_list[slave].res_list[RES_CQ];
+	struct res_cq *cq;
+	struct res_cq *tmp;
+	int state;
+	u64 in_param;
+	LIST_HEAD(tlist);
+	int cqn;
+	int err;
+
+	err = move_all_busy(dev, slave, RES_CQ);
+	if (err)
+		mlx4_warn(dev, "rem_slave_cqs: Could not move all cqs - too busy for slave %d\n",
+			  slave);
+
+	spin_lock_irq(mlx4_tlock(dev));
+	list_for_each_entry_safe(cq, tmp, cq_list, com.list) {
+		spin_unlock_irq(mlx4_tlock(dev));
+		if (cq->com.owner == slave && !atomic_read(&cq->ref_count)) {
+			cqn = cq->com.res_id;
+			state = cq->com.from_state;
+			while (state != 0) {
+				switch (state) {
+				case RES_CQ_ALLOCATED:
+					__mlx4_cq_free_icm(dev, cqn);
+					spin_lock_irq(mlx4_tlock(dev));
+					rb_erase(&cq->com.node,
+						 &tracker->res_tree[RES_CQ]);
+					list_del(&cq->com.list);
+					spin_unlock_irq(mlx4_tlock(dev));
+					mlx4_release_resource(dev, slave,
+							      RES_CQ, 1, 0);
+					kfree(cq);
+					state = 0;
+					break;
+
+				case RES_CQ_HW:
+					in_param = slave;
+					err = mlx4_cmd(dev, in_param, cqn, 1,
+						       MLX4_CMD_HW2SW_CQ,
+						       MLX4_CMD_TIME_CLASS_A,
+						       MLX4_CMD_NATIVE);
+					if (err)
+						mlx4_dbg(dev, "rem_slave_cqs: failed to move slave %d cq %d to SW ownership\n",
+							 slave, cqn);
+					atomic_dec(&cq->mtt->ref_count);
+					state = RES_CQ_ALLOCATED;
+					break;
+
+				default:
+					state = 0;
+				}
+			}
+		}
+		spin_lock_irq(mlx4_tlock(dev));
+	}
+	spin_unlock_irq(mlx4_tlock(dev));
+}
+
+static void rem_slave_mrs(struct mlx4_dev *dev, int slave)
+{
+	struct mlx4_priv *priv = mlx4_priv(dev);
+	struct mlx4_resource_tracker *tracker = &priv->mfunc.master.res_tracker;
+	struct list_head *mpt_list =
+		&tracker->slave_list[slave].res_list[RES_MPT];
+	struct res_mpt *mpt;
+	struct res_mpt *tmp;
+	int state;
+	u64 in_param;
+	LIST_HEAD(tlist);
+	int mptn;
+	int err;
+
+	err = move_all_busy(dev, slave, RES_MPT);
+	if (err)
+		mlx4_warn(dev, "rem_slave_mrs: Could not move all mpts - too busy for slave %d\n",
+			  slave);
+
+	spin_lock_irq(mlx4_tlock(dev));
+	list_for_each_entry_safe(mpt, tmp, mpt_list, com.list) {
+		spin_unlock_irq(mlx4_tlock(dev));
+		if (mpt->com.owner == slave) {
+			mptn = mpt->com.res_id;
+			state = mpt->com.from_state;
+			while (state != 0) {
+				switch (state) {
+				case RES_MPT_RESERVED:
+					__mlx4_mpt_release(dev, mpt->key);
+					spin_lock_irq(mlx4_tlock(dev));
+					rb_erase(&mpt->com.node,
+						 &tracker->res_tree[RES_MPT]);
+					list_del(&mpt->com.list);
+					spin_unlock_irq(mlx4_tlock(dev));
+					mlx4_release_resource(dev, slave,
+							      RES_MPT, 1, 0);
+					kfree(mpt);
+					state = 0;
+					break;
+
+				case RES_MPT_MAPPED:
+					__mlx4_mpt_free_icm(dev, mpt->key);
+					state = RES_MPT_RESERVED;
+					break;
+
+				case RES_MPT_HW:
+					in_param = slave;
+					err = mlx4_cmd(dev, in_param, mptn, 0,
+						     MLX4_CMD_HW2SW_MPT,
+						     MLX4_CMD_TIME_CLASS_A,
+						     MLX4_CMD_NATIVE);
+					if (err)
+						mlx4_dbg(dev, "rem_slave_mrs: failed to move slave %d mpt %d to SW ownership\n",
+							 slave, mptn);
+					if (mpt->mtt)
+						atomic_dec(&mpt->mtt->ref_count);
+					state = RES_MPT_MAPPED;
+					break;
+				default:
+					state = 0;
+				}
+			}
+		}
+		spin_lock_irq(mlx4_tlock(dev));
+	}
+	spin_unlock_irq(mlx4_tlock(dev));
+}
+
+static void rem_slave_mtts(struct mlx4_dev *dev, int slave)
+{
+	struct mlx4_priv *priv = mlx4_priv(dev);
+	struct mlx4_resource_tracker *tracker =
+		&priv->mfunc.master.res_tracker;
+	struct list_head *mtt_list =
+		&tracker->slave_list[slave].res_list[RES_MTT];
+	struct res_mtt *mtt;
+	struct res_mtt *tmp;
+	int state;
+	LIST_HEAD(tlist);
+	int base;
+	int err;
+
+	err = move_all_busy(dev, slave, RES_MTT);
+	if (err)
+		mlx4_warn(dev, "rem_slave_mtts: Could not move all mtts  - too busy for slave %d\n",
+			  slave);
+
+	spin_lock_irq(mlx4_tlock(dev));
+	list_for_each_entry_safe(mtt, tmp, mtt_list, com.list) {
+		spin_unlock_irq(mlx4_tlock(dev));
+		if (mtt->com.owner == slave) {
+			base = mtt->com.res_id;
+			state = mtt->com.from_state;
+			while (state != 0) {
+				switch (state) {
+				case RES_MTT_ALLOCATED:
+					__mlx4_free_mtt_range(dev, base,
+							      mtt->order);
+					spin_lock_irq(mlx4_tlock(dev));
+					rb_erase(&mtt->com.node,
+						 &tracker->res_tree[RES_MTT]);
+					list_del(&mtt->com.list);
+					spin_unlock_irq(mlx4_tlock(dev));
+					mlx4_release_resource(dev, slave, RES_MTT,
+							      1 << mtt->order, 0);
+					kfree(mtt);
+					state = 0;
+					break;
+
+				default:
+					state = 0;
+				}
+			}
+		}
+		spin_lock_irq(mlx4_tlock(dev));
+	}
+	spin_unlock_irq(mlx4_tlock(dev));
+}
+
+static int mlx4_do_mirror_rule(struct mlx4_dev *dev, struct res_fs_rule *fs_rule)
+{
+	struct mlx4_cmd_mailbox *mailbox;
+	int err;
+	struct res_fs_rule *mirr_rule;
+	u64 reg_id;
+
+	mailbox = mlx4_alloc_cmd_mailbox(dev);
+	if (IS_ERR(mailbox))
+		return PTR_ERR(mailbox);
+
+	if (!fs_rule->mirr_mbox) {
+		mlx4_err(dev, "rule mirroring mailbox is null\n");
+		mlx4_free_cmd_mailbox(dev, mailbox);
+		return -EINVAL;
+	}
+	memcpy(mailbox->buf, fs_rule->mirr_mbox, fs_rule->mirr_mbox_size);
+	err = mlx4_cmd_imm(dev, mailbox->dma, &reg_id, fs_rule->mirr_mbox_size >> 2, 0,
+			   MLX4_QP_FLOW_STEERING_ATTACH, MLX4_CMD_TIME_CLASS_A,
+			   MLX4_CMD_NATIVE);
+	mlx4_free_cmd_mailbox(dev, mailbox);
+
+	if (err)
+		goto err;
+
+	err = add_res_range(dev, fs_rule->com.owner, reg_id, 1, RES_FS_RULE, fs_rule->qpn);
+	if (err)
+		goto err_detach;
+
+	err = get_res(dev, fs_rule->com.owner, reg_id, RES_FS_RULE, &mirr_rule);
+	if (err)
+		goto err_rem;
+
+	fs_rule->mirr_rule_id = reg_id;
+	mirr_rule->mirr_rule_id = 0;
+	mirr_rule->mirr_mbox_size = 0;
+	mirr_rule->mirr_mbox = NULL;
+	put_res(dev, fs_rule->com.owner, reg_id, RES_FS_RULE);
+
+	return 0;
+err_rem:
+	rem_res_range(dev, fs_rule->com.owner, reg_id, 1, RES_FS_RULE, 0);
+err_detach:
+	mlx4_cmd(dev, reg_id, 0, 0, MLX4_QP_FLOW_STEERING_DETACH,
+		 MLX4_CMD_TIME_CLASS_A, MLX4_CMD_NATIVE);
+err:
+	return err;
+}
+
+static int mlx4_mirror_fs_rules(struct mlx4_dev *dev, bool bond)
+{
+	struct mlx4_priv *priv = mlx4_priv(dev);
+	struct mlx4_resource_tracker *tracker =
+		&priv->mfunc.master.res_tracker;
+	struct rb_root *root = &tracker->res_tree[RES_FS_RULE];
+	struct rb_node *p;
+	struct res_fs_rule *fs_rule;
+	int err = 0;
+	LIST_HEAD(mirr_list);
+
+	for (p = rb_first(root); p; p = rb_next(p)) {
+		fs_rule = rb_entry(p, struct res_fs_rule, com.node);
+		if ((bond && fs_rule->mirr_mbox_size) ||
+		    (!bond && !fs_rule->mirr_mbox_size))
+			list_add_tail(&fs_rule->mirr_list, &mirr_list);
+	}
+
+	list_for_each_entry(fs_rule, &mirr_list, mirr_list) {
+		if (bond)
+			err += mlx4_do_mirror_rule(dev, fs_rule);
+		else
+			err += mlx4_undo_mirror_rule(dev, fs_rule);
+	}
+	return err;
+}
+
+int mlx4_bond_fs_rules(struct mlx4_dev *dev)
+{
+	return mlx4_mirror_fs_rules(dev, true);
+}
+
+int mlx4_unbond_fs_rules(struct mlx4_dev *dev)
+{
+	return mlx4_mirror_fs_rules(dev, false);
+}
+
+static void rem_slave_fs_rule(struct mlx4_dev *dev, int slave)
+{
+	struct mlx4_priv *priv = mlx4_priv(dev);
+	struct mlx4_resource_tracker *tracker =
+		&priv->mfunc.master.res_tracker;
+	struct list_head *fs_rule_list =
+		&tracker->slave_list[slave].res_list[RES_FS_RULE];
+	struct res_fs_rule *fs_rule;
+	struct res_fs_rule *tmp;
+	int state;
+	u64 base;
+	int err;
+
+	err = move_all_busy(dev, slave, RES_FS_RULE);
+	if (err)
+		mlx4_warn(dev, "rem_slave_fs_rule: Could not move all mtts to busy for slave %d\n",
+			  slave);
+
+	spin_lock_irq(mlx4_tlock(dev));
+	list_for_each_entry_safe(fs_rule, tmp, fs_rule_list, com.list) {
+		spin_unlock_irq(mlx4_tlock(dev));
+		if (fs_rule->com.owner == slave) {
+			base = fs_rule->com.res_id;
+			state = fs_rule->com.from_state;
+			while (state != 0) {
+				switch (state) {
+				case RES_FS_RULE_ALLOCATED:
+					/* detach rule */
+					err = mlx4_cmd(dev, base, 0, 0,
+						       MLX4_QP_FLOW_STEERING_DETACH,
+						       MLX4_CMD_TIME_CLASS_A,
+						       MLX4_CMD_NATIVE);
+
+					spin_lock_irq(mlx4_tlock(dev));
+					rb_erase(&fs_rule->com.node,
+						 &tracker->res_tree[RES_FS_RULE]);
+					list_del(&fs_rule->com.list);
+					spin_unlock_irq(mlx4_tlock(dev));
+					kfree(fs_rule->mirr_mbox);
+					kfree(fs_rule);
+					state = 0;
+					break;
+
+				default:
+					state = 0;
+				}
+			}
+		}
+		spin_lock_irq(mlx4_tlock(dev));
+	}
+	spin_unlock_irq(mlx4_tlock(dev));
+}
+
+static void rem_slave_eqs(struct mlx4_dev *dev, int slave)
+{
+	struct mlx4_priv *priv = mlx4_priv(dev);
+	struct mlx4_resource_tracker *tracker = &priv->mfunc.master.res_tracker;
+	struct list_head *eq_list =
+		&tracker->slave_list[slave].res_list[RES_EQ];
+	struct res_eq *eq;
+	struct res_eq *tmp;
+	int err;
+	int state;
+	LIST_HEAD(tlist);
+	int eqn;
+
+	err = move_all_busy(dev, slave, RES_EQ);
+	if (err)
+		mlx4_warn(dev, "rem_slave_eqs: Could not move all eqs - too busy for slave %d\n",
+			  slave);
+
+	spin_lock_irq(mlx4_tlock(dev));
+	list_for_each_entry_safe(eq, tmp, eq_list, com.list) {
+		spin_unlock_irq(mlx4_tlock(dev));
+		if (eq->com.owner == slave) {
+			eqn = eq->com.res_id;
+			state = eq->com.from_state;
+			while (state != 0) {
+				switch (state) {
+				case RES_EQ_RESERVED:
+					spin_lock_irq(mlx4_tlock(dev));
+					rb_erase(&eq->com.node,
+						 &tracker->res_tree[RES_EQ]);
+					list_del(&eq->com.list);
+					spin_unlock_irq(mlx4_tlock(dev));
+					kfree(eq);
+					state = 0;
+					break;
+
+				case RES_EQ_HW:
+					err = mlx4_cmd(dev, slave, eqn & 0x3ff,
+						       1, MLX4_CMD_HW2SW_EQ,
+						       MLX4_CMD_TIME_CLASS_A,
+						       MLX4_CMD_NATIVE);
+					if (err)
+						mlx4_dbg(dev, "rem_slave_eqs: failed to move slave %d eqs %d to SW ownership\n",
+							 slave, eqn & 0x3ff);
+					atomic_dec(&eq->mtt->ref_count);
+					state = RES_EQ_RESERVED;
+					break;
+
+				default:
+					state = 0;
+				}
+			}
+		}
+		spin_lock_irq(mlx4_tlock(dev));
+	}
+	spin_unlock_irq(mlx4_tlock(dev));
+}
+
+static void rem_slave_counters(struct mlx4_dev *dev, int slave)
+{
+	struct mlx4_priv *priv = mlx4_priv(dev);
+	struct mlx4_resource_tracker *tracker = &priv->mfunc.master.res_tracker;
+	struct list_head *counter_list =
+		&tracker->slave_list[slave].res_list[RES_COUNTER];
+	struct res_counter *counter;
+	struct res_counter *tmp;
+	int err;
+	int *counters_arr = NULL;
+	int i, j;
+
+	err = move_all_busy(dev, slave, RES_COUNTER);
+	if (err)
+		mlx4_warn(dev, "rem_slave_counters: Could not move all counters - too busy for slave %d\n",
+			  slave);
+
+	counters_arr = kmalloc_array(dev->caps.max_counters,
+				     sizeof(*counters_arr), GFP_KERNEL);
+	if (!counters_arr)
+		return;
+
+	do {
+		i = 0;
+		j = 0;
+		spin_lock_irq(mlx4_tlock(dev));
+		list_for_each_entry_safe(counter, tmp, counter_list, com.list) {
+			if (counter->com.owner == slave) {
+				counters_arr[i++] = counter->com.res_id;
+				rb_erase(&counter->com.node,
+					 &tracker->res_tree[RES_COUNTER]);
+				list_del(&counter->com.list);
+				kfree(counter);
+			}
+		}
+		spin_unlock_irq(mlx4_tlock(dev));
+
+		while (j < i) {
+			__mlx4_counter_free(dev, counters_arr[j++]);
+			mlx4_release_resource(dev, slave, RES_COUNTER, 1, 0);
+		}
+	} while (i);
+
+	kfree(counters_arr);
+}
+
+static void rem_slave_xrcdns(struct mlx4_dev *dev, int slave)
+{
+	struct mlx4_priv *priv = mlx4_priv(dev);
+	struct mlx4_resource_tracker *tracker = &priv->mfunc.master.res_tracker;
+	struct list_head *xrcdn_list =
+		&tracker->slave_list[slave].res_list[RES_XRCD];
+	struct res_xrcdn *xrcd;
+	struct res_xrcdn *tmp;
+	int err;
+	int xrcdn;
+
+	err = move_all_busy(dev, slave, RES_XRCD);
+	if (err)
+		mlx4_warn(dev, "rem_slave_xrcdns: Could not move all xrcdns - too busy for slave %d\n",
+			  slave);
+
+	spin_lock_irq(mlx4_tlock(dev));
+	list_for_each_entry_safe(xrcd, tmp, xrcdn_list, com.list) {
+		if (xrcd->com.owner == slave) {
+			xrcdn = xrcd->com.res_id;
+			rb_erase(&xrcd->com.node, &tracker->res_tree[RES_XRCD]);
+			list_del(&xrcd->com.list);
+			kfree(xrcd);
+			__mlx4_xrcd_free(dev, xrcdn);
+		}
+	}
+	spin_unlock_irq(mlx4_tlock(dev));
+}
+
+void mlx4_delete_all_resources_for_slave(struct mlx4_dev *dev, int slave)
+{
+	struct mlx4_priv *priv = mlx4_priv(dev);
+	mlx4_reset_roce_gids(dev, slave);
+	mutex_lock(&priv->mfunc.master.res_tracker.slave_list[slave].mutex);
+	rem_slave_vlans(dev, slave);
+	rem_slave_macs(dev, slave);
+	rem_slave_fs_rule(dev, slave);
+	rem_slave_qps(dev, slave);
+	rem_slave_srqs(dev, slave);
+	rem_slave_cqs(dev, slave);
+	rem_slave_mrs(dev, slave);
+	rem_slave_eqs(dev, slave);
+	rem_slave_mtts(dev, slave);
+	rem_slave_counters(dev, slave);
+	rem_slave_xrcdns(dev, slave);
+	mutex_unlock(&priv->mfunc.master.res_tracker.slave_list[slave].mutex);
+}
+
+static void update_qos_vpp(struct mlx4_update_qp_context *ctx,
+			   struct mlx4_vf_immed_vlan_work *work)
+{
+	ctx->qp_mask |= cpu_to_be64(1ULL << MLX4_UPD_QP_MASK_QOS_VPP);
+	ctx->qp_context.qos_vport = work->qos_vport;
+}
+
+void mlx4_vf_immed_vlan_work_handler(struct work_struct *_work)
+{
+	struct mlx4_vf_immed_vlan_work *work =
+		container_of(_work, struct mlx4_vf_immed_vlan_work, work);
+	struct mlx4_cmd_mailbox *mailbox;
+	struct mlx4_update_qp_context *upd_context;
+	struct mlx4_dev *dev = &work->priv->dev;
+	struct mlx4_resource_tracker *tracker =
+		&work->priv->mfunc.master.res_tracker;
+	struct list_head *qp_list =
+		&tracker->slave_list[work->slave].res_list[RES_QP];
+	struct res_qp *qp;
+	struct res_qp *tmp;
+	u64 qp_path_mask_vlan_ctrl =
+		       ((1ULL << MLX4_UPD_QP_PATH_MASK_ETH_TX_BLOCK_UNTAGGED) |
+		       (1ULL << MLX4_UPD_QP_PATH_MASK_ETH_TX_BLOCK_1P) |
+		       (1ULL << MLX4_UPD_QP_PATH_MASK_ETH_TX_BLOCK_TAGGED) |
+		       (1ULL << MLX4_UPD_QP_PATH_MASK_ETH_RX_BLOCK_UNTAGGED) |
+		       (1ULL << MLX4_UPD_QP_PATH_MASK_ETH_RX_BLOCK_1P) |
+		       (1ULL << MLX4_UPD_QP_PATH_MASK_ETH_RX_BLOCK_TAGGED));
+
+	u64 qp_path_mask = ((1ULL << MLX4_UPD_QP_PATH_MASK_VLAN_INDEX) |
+		       (1ULL << MLX4_UPD_QP_PATH_MASK_FVL) |
+		       (1ULL << MLX4_UPD_QP_PATH_MASK_CV) |
+		       (1ULL << MLX4_UPD_QP_PATH_MASK_SV) |
+		       (1ULL << MLX4_UPD_QP_PATH_MASK_ETH_HIDE_CQE_VLAN) |
+		       (1ULL << MLX4_UPD_QP_PATH_MASK_FEUP) |
+		       (1ULL << MLX4_UPD_QP_PATH_MASK_FVL_RX) |
+		       (1ULL << MLX4_UPD_QP_PATH_MASK_SCHED_QUEUE));
+
+	int err;
+	int port, errors = 0;
+	u8 vlan_control;
+
+	if (mlx4_is_slave(dev)) {
+		mlx4_warn(dev, "Trying to update-qp in slave %d\n",
+			  work->slave);
+		goto out;
+	}
+
+	mailbox = mlx4_alloc_cmd_mailbox(dev);
+	if (IS_ERR(mailbox))
+		goto out;
+	if (work->flags & MLX4_VF_IMMED_VLAN_FLAG_LINK_DISABLE) /* block all */
+		vlan_control = MLX4_VLAN_CTRL_ETH_TX_BLOCK_TAGGED |
+			MLX4_VLAN_CTRL_ETH_TX_BLOCK_PRIO_TAGGED |
+			MLX4_VLAN_CTRL_ETH_TX_BLOCK_UNTAGGED |
+			MLX4_VLAN_CTRL_ETH_RX_BLOCK_PRIO_TAGGED |
+			MLX4_VLAN_CTRL_ETH_RX_BLOCK_UNTAGGED |
+			MLX4_VLAN_CTRL_ETH_RX_BLOCK_TAGGED;
+	else if (!work->vlan_id)
+		vlan_control = MLX4_VLAN_CTRL_ETH_TX_BLOCK_TAGGED |
+			MLX4_VLAN_CTRL_ETH_RX_BLOCK_TAGGED;
+	else if (work->vlan_proto == htons(ETH_P_8021AD))
+		vlan_control = MLX4_VLAN_CTRL_ETH_TX_BLOCK_PRIO_TAGGED |
+			MLX4_VLAN_CTRL_ETH_TX_BLOCK_TAGGED |
+			MLX4_VLAN_CTRL_ETH_RX_BLOCK_PRIO_TAGGED |
+			MLX4_VLAN_CTRL_ETH_RX_BLOCK_UNTAGGED;
+	else  /* vst 802.1Q */
+		vlan_control = MLX4_VLAN_CTRL_ETH_TX_BLOCK_TAGGED |
+			MLX4_VLAN_CTRL_ETH_RX_BLOCK_PRIO_TAGGED |
+			MLX4_VLAN_CTRL_ETH_RX_BLOCK_UNTAGGED;
+
+	upd_context = mailbox->buf;
+	upd_context->qp_mask = cpu_to_be64(1ULL << MLX4_UPD_QP_MASK_VSD);
+
+	spin_lock_irq(mlx4_tlock(dev));
+	list_for_each_entry_safe(qp, tmp, qp_list, com.list) {
+		spin_unlock_irq(mlx4_tlock(dev));
+		if (qp->com.owner == work->slave) {
+			if (qp->com.from_state != RES_QP_HW ||
+			    !qp->sched_queue ||  /* no INIT2RTR trans yet */
+			    mlx4_is_qp_reserved(dev, qp->local_qpn) ||
+			    qp->qpc_flags & (1 << MLX4_RSS_QPC_FLAG_OFFSET)) {
+				spin_lock_irq(mlx4_tlock(dev));
+				continue;
+			}
+			port = (qp->sched_queue >> 6 & 1) + 1;
+			if (port != work->port) {
+				spin_lock_irq(mlx4_tlock(dev));
+				continue;
+			}
+			if (MLX4_QP_ST_RC == ((qp->qpc_flags >> 16) & 0xff))
+				upd_context->primary_addr_path_mask = cpu_to_be64(qp_path_mask);
+			else
+				upd_context->primary_addr_path_mask =
+					cpu_to_be64(qp_path_mask | qp_path_mask_vlan_ctrl);
+			if (work->vlan_id == MLX4_VGT) {
+				upd_context->qp_context.param3 = qp->param3;
+				upd_context->qp_context.pri_path.vlan_control = qp->vlan_control;
+				upd_context->qp_context.pri_path.fvl_rx = qp->fvl_rx;
+				upd_context->qp_context.pri_path.vlan_index = qp->vlan_index;
+				upd_context->qp_context.pri_path.fl = qp->pri_path_fl;
+				upd_context->qp_context.pri_path.feup = qp->feup;
+				upd_context->qp_context.pri_path.sched_queue =
+					qp->sched_queue;
+			} else {
+				upd_context->qp_context.param3 = qp->param3 & ~cpu_to_be32(MLX4_STRIP_VLAN);
+				upd_context->qp_context.pri_path.vlan_control = vlan_control;
+				upd_context->qp_context.pri_path.vlan_index = work->vlan_ix;
+				upd_context->qp_context.pri_path.fvl_rx =
+					qp->fvl_rx | MLX4_FVL_RX_FORCE_ETH_VLAN;
+				upd_context->qp_context.pri_path.fl =
+					qp->pri_path_fl | MLX4_FL_ETH_HIDE_CQE_VLAN;
+				if (work->vlan_proto == htons(ETH_P_8021AD))
+					upd_context->qp_context.pri_path.fl |= MLX4_FL_SV;
+				else
+					upd_context->qp_context.pri_path.fl |= MLX4_FL_CV;
+				upd_context->qp_context.pri_path.feup =
+					qp->feup | MLX4_FEUP_FORCE_ETH_UP | MLX4_FVL_FORCE_ETH_VLAN;
+				upd_context->qp_context.pri_path.sched_queue =
+					qp->sched_queue & 0xC7;
+				upd_context->qp_context.pri_path.sched_queue |=
+					((work->qos & 0x7) << 3);
+
+				if (dev->caps.flags2 &
+				    MLX4_DEV_CAP_FLAG2_QOS_VPP)
+					update_qos_vpp(upd_context, work);
+			}
+
+			err = mlx4_cmd(dev, mailbox->dma,
+				       qp->local_qpn & 0xffffff,
+				       0, MLX4_CMD_UPDATE_QP,
+				       MLX4_CMD_TIME_CLASS_C, MLX4_CMD_NATIVE);
+			if (err) {
+				mlx4_info(dev, "UPDATE_QP failed for slave %d, port %d, qpn %d (%d)\n",
+					  work->slave, port, qp->local_qpn, err);
+				errors++;
+			}
+		}
+		spin_lock_irq(mlx4_tlock(dev));
+	}
+	spin_unlock_irq(mlx4_tlock(dev));
+	mlx4_free_cmd_mailbox(dev, mailbox);
+
+	if (errors)
+		mlx4_err(dev, "%d UPDATE_QP failures for slave %d, port %d\n",
+			 errors, work->slave, work->port);
+
+	/* unregister previous vlan_id if needed and we had no errors
+	 * while updating the QPs
+	 */
+	if (work->flags & MLX4_VF_IMMED_VLAN_FLAG_VLAN && !errors &&
+	    NO_INDX != work->orig_vlan_ix)
+		__mlx4_unregister_vlan(&work->priv->dev, work->port,
+				       work->orig_vlan_id);
+out:
+	kfree(work);
+	return;
+}
diff --git a/drivers/net/ethernet/mellanox/mlx4/sense.c b/drivers/net/ethernet/mellanox/mlx4/sense.c
new file mode 100644
index 000000000..094773d88
--- /dev/null
+++ b/drivers/net/ethernet/mellanox/mlx4/sense.c
@@ -0,0 +1,143 @@
+/*
+ * Copyright (c) 2007 Mellanox Technologies. All rights reserved.
+ *
+ * This software is available to you under a choice of one of two
+ * licenses.  You may choose to be licensed under the terms of the GNU
+ * General Public License (GPL) Version 2, available from the file
+ * COPYING in the main directory of this source tree, or the
+ * OpenIB.org BSD license below:
+ *
+ *     Redistribution and use in source and binary forms, with or
+ *     without modification, are permitted provided that the following
+ *     conditions are met:
+ *
+ *      - Redistributions of source code must retain the above
+ *        copyright notice, this list of conditions and the following
+ *        disclaimer.
+ *
+ *      - Redistributions in binary form must reproduce the above
+ *        copyright notice, this list of conditions and the following
+ *        disclaimer in the documentation and/or other materials
+ *        provided with the distribution.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
+ * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
+ * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS
+ * BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN
+ * ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
+ * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+ * SOFTWARE.
+ *
+ */
+
+#include <linux/errno.h>
+#include <linux/if_ether.h>
+
+#include <linux/mlx4/cmd.h>
+
+#include "mlx4.h"
+
+int mlx4_SENSE_PORT(struct mlx4_dev *dev, int port,
+		    enum mlx4_port_type *type)
+{
+	u64 out_param;
+	int err = 0;
+
+	err = mlx4_cmd_imm(dev, 0, &out_param, port, 0,
+			   MLX4_CMD_SENSE_PORT, MLX4_CMD_TIME_CLASS_B,
+			   MLX4_CMD_WRAPPED);
+	if (err) {
+		mlx4_err(dev, "Sense command failed for port: %d\n", port);
+		return err;
+	}
+
+	if (out_param > 2) {
+		mlx4_err(dev, "Sense returned illegal value: 0x%llx\n", out_param);
+		return -EINVAL;
+	}
+
+	*type = out_param;
+	return 0;
+}
+
+void mlx4_do_sense_ports(struct mlx4_dev *dev,
+			 enum mlx4_port_type *stype,
+			 enum mlx4_port_type *defaults)
+{
+	struct mlx4_sense *sense = &mlx4_priv(dev)->sense;
+	int err;
+	int i;
+
+	for (i = 1; i <= dev->caps.num_ports; i++) {
+		stype[i - 1] = 0;
+		if (sense->do_sense_port[i] && sense->sense_allowed[i] &&
+		    dev->caps.possible_type[i] == MLX4_PORT_TYPE_AUTO) {
+			err = mlx4_SENSE_PORT(dev, i, &stype[i - 1]);
+			if (err)
+				stype[i - 1] = defaults[i - 1];
+		} else
+			stype[i - 1] = defaults[i - 1];
+	}
+
+	/*
+	 * If sensed nothing, remain in current configuration.
+	 */
+	for (i = 0; i < dev->caps.num_ports; i++)
+		stype[i] = stype[i] ? stype[i] : defaults[i];
+
+}
+
+static void mlx4_sense_port(struct work_struct *work)
+{
+	struct delayed_work *delay = to_delayed_work(work);
+	struct mlx4_sense *sense = container_of(delay, struct mlx4_sense,
+						sense_poll);
+	struct mlx4_dev *dev = sense->dev;
+	struct mlx4_priv *priv = mlx4_priv(dev);
+	enum mlx4_port_type stype[MLX4_MAX_PORTS];
+
+	mutex_lock(&priv->port_mutex);
+	mlx4_do_sense_ports(dev, stype, &dev->caps.port_type[1]);
+
+	if (mlx4_check_port_params(dev, stype))
+		goto sense_again;
+
+	if (mlx4_change_port_types(dev, stype))
+		mlx4_err(dev, "Failed to change port_types\n");
+
+sense_again:
+	mutex_unlock(&priv->port_mutex);
+	queue_delayed_work(mlx4_wq , &sense->sense_poll,
+			   round_jiffies_relative(MLX4_SENSE_RANGE));
+}
+
+void mlx4_start_sense(struct mlx4_dev *dev)
+{
+	struct mlx4_priv *priv = mlx4_priv(dev);
+	struct mlx4_sense *sense = &priv->sense;
+
+	if (!(dev->caps.flags & MLX4_DEV_CAP_FLAG_DPDP))
+		return;
+
+	queue_delayed_work(mlx4_wq , &sense->sense_poll,
+			   round_jiffies_relative(MLX4_SENSE_RANGE));
+}
+
+void mlx4_stop_sense(struct mlx4_dev *dev)
+{
+	cancel_delayed_work_sync(&mlx4_priv(dev)->sense.sense_poll);
+}
+
+void  mlx4_sense_init(struct mlx4_dev *dev)
+{
+	struct mlx4_priv *priv = mlx4_priv(dev);
+	struct mlx4_sense *sense = &priv->sense;
+	int port;
+
+	sense->dev = dev;
+	for (port = 1; port <= dev->caps.num_ports; port++)
+		sense->do_sense_port[port] = 1;
+
+	INIT_DEFERRABLE_WORK(&sense->sense_poll, mlx4_sense_port);
+}
diff --git a/drivers/net/ethernet/mellanox/mlx4/srq.c b/drivers/net/ethernet/mellanox/mlx4/srq.c
new file mode 100644
index 000000000..cbe4d9746
--- /dev/null
+++ b/drivers/net/ethernet/mellanox/mlx4/srq.c
@@ -0,0 +1,309 @@
+/*
+ * Copyright (c) 2006, 2007 Cisco Systems, Inc. All rights reserved.
+ * Copyright (c) 2007, 2008 Mellanox Technologies. All rights reserved.
+ *
+ * This software is available to you under a choice of one of two
+ * licenses.  You may choose to be licensed under the terms of the GNU
+ * General Public License (GPL) Version 2, available from the file
+ * COPYING in the main directory of this source tree, or the
+ * OpenIB.org BSD license below:
+ *
+ *     Redistribution and use in source and binary forms, with or
+ *     without modification, are permitted provided that the following
+ *     conditions are met:
+ *
+ *      - Redistributions of source code must retain the above
+ *        copyright notice, this list of conditions and the following
+ *        disclaimer.
+ *
+ *      - Redistributions in binary form must reproduce the above
+ *        copyright notice, this list of conditions and the following
+ *        disclaimer in the documentation and/or other materials
+ *        provided with the distribution.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
+ * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
+ * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS
+ * BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN
+ * ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
+ * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+ * SOFTWARE.
+ */
+
+
+#include <linux/mlx4/cmd.h>
+#include <linux/mlx4/srq.h>
+#include <linux/export.h>
+#include <linux/gfp.h>
+
+#include "mlx4.h"
+#include "icm.h"
+
+void mlx4_srq_event(struct mlx4_dev *dev, u32 srqn, int event_type)
+{
+	struct mlx4_srq_table *srq_table = &mlx4_priv(dev)->srq_table;
+	struct mlx4_srq *srq;
+
+	rcu_read_lock();
+	srq = radix_tree_lookup(&srq_table->tree, srqn & (dev->caps.num_srqs - 1));
+	rcu_read_unlock();
+	if (srq)
+		refcount_inc(&srq->refcount);
+	else {
+		mlx4_warn(dev, "Async event for bogus SRQ %08x\n", srqn);
+		return;
+	}
+
+	srq->event(srq, event_type);
+
+	if (refcount_dec_and_test(&srq->refcount))
+		complete(&srq->free);
+}
+
+static int mlx4_SW2HW_SRQ(struct mlx4_dev *dev, struct mlx4_cmd_mailbox *mailbox,
+			  int srq_num)
+{
+	return mlx4_cmd(dev, mailbox->dma, srq_num, 0,
+			MLX4_CMD_SW2HW_SRQ, MLX4_CMD_TIME_CLASS_A,
+			MLX4_CMD_WRAPPED);
+}
+
+static int mlx4_HW2SW_SRQ(struct mlx4_dev *dev, struct mlx4_cmd_mailbox *mailbox,
+			  int srq_num)
+{
+	return mlx4_cmd_box(dev, 0, mailbox ? mailbox->dma : 0, srq_num,
+			    mailbox ? 0 : 1, MLX4_CMD_HW2SW_SRQ,
+			    MLX4_CMD_TIME_CLASS_A, MLX4_CMD_WRAPPED);
+}
+
+static int mlx4_ARM_SRQ(struct mlx4_dev *dev, int srq_num, int limit_watermark)
+{
+	return mlx4_cmd(dev, limit_watermark, srq_num, 0, MLX4_CMD_ARM_SRQ,
+			MLX4_CMD_TIME_CLASS_B, MLX4_CMD_WRAPPED);
+}
+
+static int mlx4_QUERY_SRQ(struct mlx4_dev *dev, struct mlx4_cmd_mailbox *mailbox,
+			  int srq_num)
+{
+	return mlx4_cmd_box(dev, 0, mailbox->dma, srq_num, 0, MLX4_CMD_QUERY_SRQ,
+			    MLX4_CMD_TIME_CLASS_A, MLX4_CMD_WRAPPED);
+}
+
+int __mlx4_srq_alloc_icm(struct mlx4_dev *dev, int *srqn)
+{
+	struct mlx4_srq_table *srq_table = &mlx4_priv(dev)->srq_table;
+	int err;
+
+
+	*srqn = mlx4_bitmap_alloc(&srq_table->bitmap);
+	if (*srqn == -1)
+		return -ENOMEM;
+
+	err = mlx4_table_get(dev, &srq_table->table, *srqn);
+	if (err)
+		goto err_out;
+
+	err = mlx4_table_get(dev, &srq_table->cmpt_table, *srqn);
+	if (err)
+		goto err_put;
+	return 0;
+
+err_put:
+	mlx4_table_put(dev, &srq_table->table, *srqn);
+
+err_out:
+	mlx4_bitmap_free(&srq_table->bitmap, *srqn, MLX4_NO_RR);
+	return err;
+}
+
+static int mlx4_srq_alloc_icm(struct mlx4_dev *dev, int *srqn)
+{
+	u64 out_param;
+	int err;
+
+	if (mlx4_is_mfunc(dev)) {
+		err = mlx4_cmd_imm(dev, 0, &out_param, RES_SRQ,
+				   RES_OP_RESERVE_AND_MAP,
+				   MLX4_CMD_ALLOC_RES,
+				   MLX4_CMD_TIME_CLASS_A, MLX4_CMD_WRAPPED);
+		if (!err)
+			*srqn = get_param_l(&out_param);
+
+		return err;
+	}
+	return __mlx4_srq_alloc_icm(dev, srqn);
+}
+
+void __mlx4_srq_free_icm(struct mlx4_dev *dev, int srqn)
+{
+	struct mlx4_srq_table *srq_table = &mlx4_priv(dev)->srq_table;
+
+	mlx4_table_put(dev, &srq_table->cmpt_table, srqn);
+	mlx4_table_put(dev, &srq_table->table, srqn);
+	mlx4_bitmap_free(&srq_table->bitmap, srqn, MLX4_NO_RR);
+}
+
+static void mlx4_srq_free_icm(struct mlx4_dev *dev, int srqn)
+{
+	u64 in_param = 0;
+
+	if (mlx4_is_mfunc(dev)) {
+		set_param_l(&in_param, srqn);
+		if (mlx4_cmd(dev, in_param, RES_SRQ, RES_OP_RESERVE_AND_MAP,
+			     MLX4_CMD_FREE_RES,
+			     MLX4_CMD_TIME_CLASS_A, MLX4_CMD_WRAPPED))
+			mlx4_warn(dev, "Failed freeing cq:%d\n", srqn);
+		return;
+	}
+	__mlx4_srq_free_icm(dev, srqn);
+}
+
+int mlx4_srq_alloc(struct mlx4_dev *dev, u32 pdn, u32 cqn, u16 xrcd,
+		   struct mlx4_mtt *mtt, u64 db_rec, struct mlx4_srq *srq)
+{
+	struct mlx4_srq_table *srq_table = &mlx4_priv(dev)->srq_table;
+	struct mlx4_cmd_mailbox *mailbox;
+	struct mlx4_srq_context *srq_context;
+	u64 mtt_addr;
+	int err;
+
+	err = mlx4_srq_alloc_icm(dev, &srq->srqn);
+	if (err)
+		return err;
+
+	spin_lock_irq(&srq_table->lock);
+	err = radix_tree_insert(&srq_table->tree, srq->srqn, srq);
+	spin_unlock_irq(&srq_table->lock);
+	if (err)
+		goto err_icm;
+
+	mailbox = mlx4_alloc_cmd_mailbox(dev);
+	if (IS_ERR(mailbox)) {
+		err = PTR_ERR(mailbox);
+		goto err_radix;
+	}
+
+	srq_context = mailbox->buf;
+	srq_context->state_logsize_srqn = cpu_to_be32((ilog2(srq->max) << 24) |
+						      srq->srqn);
+	srq_context->logstride          = srq->wqe_shift - 4;
+	srq_context->xrcd		= cpu_to_be16(xrcd);
+	srq_context->pg_offset_cqn	= cpu_to_be32(cqn & 0xffffff);
+	srq_context->log_page_size      = mtt->page_shift - MLX4_ICM_PAGE_SHIFT;
+
+	mtt_addr = mlx4_mtt_addr(dev, mtt);
+	srq_context->mtt_base_addr_h    = mtt_addr >> 32;
+	srq_context->mtt_base_addr_l    = cpu_to_be32(mtt_addr & 0xffffffff);
+	srq_context->pd			= cpu_to_be32(pdn);
+	srq_context->db_rec_addr        = cpu_to_be64(db_rec);
+
+	err = mlx4_SW2HW_SRQ(dev, mailbox, srq->srqn);
+	mlx4_free_cmd_mailbox(dev, mailbox);
+	if (err)
+		goto err_radix;
+
+	refcount_set(&srq->refcount, 1);
+	init_completion(&srq->free);
+
+	return 0;
+
+err_radix:
+	spin_lock_irq(&srq_table->lock);
+	radix_tree_delete(&srq_table->tree, srq->srqn);
+	spin_unlock_irq(&srq_table->lock);
+
+err_icm:
+	mlx4_srq_free_icm(dev, srq->srqn);
+	return err;
+}
+EXPORT_SYMBOL_GPL(mlx4_srq_alloc);
+
+void mlx4_srq_free(struct mlx4_dev *dev, struct mlx4_srq *srq)
+{
+	struct mlx4_srq_table *srq_table = &mlx4_priv(dev)->srq_table;
+	int err;
+
+	err = mlx4_HW2SW_SRQ(dev, NULL, srq->srqn);
+	if (err)
+		mlx4_warn(dev, "HW2SW_SRQ failed (%d) for SRQN %06x\n", err, srq->srqn);
+
+	spin_lock_irq(&srq_table->lock);
+	radix_tree_delete(&srq_table->tree, srq->srqn);
+	spin_unlock_irq(&srq_table->lock);
+
+	if (refcount_dec_and_test(&srq->refcount))
+		complete(&srq->free);
+	wait_for_completion(&srq->free);
+
+	mlx4_srq_free_icm(dev, srq->srqn);
+}
+EXPORT_SYMBOL_GPL(mlx4_srq_free);
+
+int mlx4_srq_arm(struct mlx4_dev *dev, struct mlx4_srq *srq, int limit_watermark)
+{
+	return mlx4_ARM_SRQ(dev, srq->srqn, limit_watermark);
+}
+EXPORT_SYMBOL_GPL(mlx4_srq_arm);
+
+int mlx4_srq_query(struct mlx4_dev *dev, struct mlx4_srq *srq, int *limit_watermark)
+{
+	struct mlx4_cmd_mailbox *mailbox;
+	struct mlx4_srq_context *srq_context;
+	int err;
+
+	mailbox = mlx4_alloc_cmd_mailbox(dev);
+	if (IS_ERR(mailbox))
+		return PTR_ERR(mailbox);
+
+	srq_context = mailbox->buf;
+
+	err = mlx4_QUERY_SRQ(dev, mailbox, srq->srqn);
+	if (err)
+		goto err_out;
+	*limit_watermark = be16_to_cpu(srq_context->limit_watermark);
+
+err_out:
+	mlx4_free_cmd_mailbox(dev, mailbox);
+	return err;
+}
+EXPORT_SYMBOL_GPL(mlx4_srq_query);
+
+int mlx4_init_srq_table(struct mlx4_dev *dev)
+{
+	struct mlx4_srq_table *srq_table = &mlx4_priv(dev)->srq_table;
+	int err;
+
+	spin_lock_init(&srq_table->lock);
+	INIT_RADIX_TREE(&srq_table->tree, GFP_ATOMIC);
+	if (mlx4_is_slave(dev))
+		return 0;
+
+	err = mlx4_bitmap_init(&srq_table->bitmap, dev->caps.num_srqs,
+			       dev->caps.num_srqs - 1, dev->caps.reserved_srqs, 0);
+	if (err)
+		return err;
+
+	return 0;
+}
+
+void mlx4_cleanup_srq_table(struct mlx4_dev *dev)
+{
+	if (mlx4_is_slave(dev))
+		return;
+	mlx4_bitmap_cleanup(&mlx4_priv(dev)->srq_table.bitmap);
+}
+
+struct mlx4_srq *mlx4_srq_lookup(struct mlx4_dev *dev, u32 srqn)
+{
+	struct mlx4_srq_table *srq_table = &mlx4_priv(dev)->srq_table;
+	struct mlx4_srq *srq;
+
+	rcu_read_lock();
+	srq = radix_tree_lookup(&srq_table->tree,
+				srqn & (dev->caps.num_srqs - 1));
+	rcu_read_unlock();
+
+	return srq;
+}
+EXPORT_SYMBOL_GPL(mlx4_srq_lookup);