Adding upstream version 4.19.249.upstream/4.19.249

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

114 files changed, 57917 insertions, 0 deletions

diff --git a/drivers/net/ethernet/mellanox/mlx5/core/Kconfig b/drivers/net/ethernet/mellanox/mlx5/core/Kconfig
new file mode 100644
index 000000000..b7e3b8902
--- /dev/null
+++ b/drivers/net/ethernet/mellanox/mlx5/core/Kconfig
@@ -0,0 +1,120 @@
+#
+# Mellanox driver configuration
+#
+
+config MLX5_CORE
+	tristate "Mellanox 5th generation network adapters (ConnectX series) core driver"
+	depends on MAY_USE_DEVLINK
+	depends on PCI
+	imply PTP_1588_CLOCK
+	imply VXLAN
+	imply MLXFW
+	default n
+	---help---
+	  Core driver for low level functionality of the ConnectX-4 and
+	  Connect-IB cards by Mellanox Technologies.
+
+config MLX5_ACCEL
+	bool
+
+config MLX5_FPGA
+        bool "Mellanox Technologies Innova support"
+        depends on MLX5_CORE
+	select MLX5_ACCEL
+        ---help---
+          Build support for the Innova family of network cards by Mellanox
+          Technologies. Innova network cards are comprised of a ConnectX chip
+          and an FPGA chip on one board. If you select this option, the
+          mlx5_core driver will include the Innova FPGA core and allow building
+          sandbox-specific client drivers.
+
+config MLX5_CORE_EN
+	bool "Mellanox 5th generation network adapters (ConnectX series) Ethernet support"
+	depends on NETDEVICES && ETHERNET && INET && PCI && MLX5_CORE
+	depends on IPV6=y || IPV6=n || MLX5_CORE=m
+	select PAGE_POOL
+	default n
+	---help---
+	  Ethernet support in Mellanox Technologies ConnectX-4 NIC.
+
+config MLX5_EN_ARFS
+	bool "Mellanox MLX5 ethernet accelerated receive flow steering (ARFS) support"
+	depends on MLX5_CORE_EN && RFS_ACCEL
+	default y
+	---help---
+	  Mellanox MLX5 ethernet hardware-accelerated receive flow steering support,
+	  Enables ethernet netdevice arfs support and ntuple filtering.
+
+config MLX5_EN_RXNFC
+	bool "Mellanox MLX5 ethernet rx nfc flow steering support"
+	depends on MLX5_CORE_EN
+	default y
+	---help---
+	  Mellanox MLX5 ethernet rx nfc flow steering support
+	  Enables ethtool receive network flow classification, which allows user defined
+	  flow rules to direct traffic into arbitrary rx queue via ethtool set/get_rxnfc
+	  API.
+
+config MLX5_MPFS
+        bool "Mellanox Technologies MLX5 MPFS support"
+        depends on MLX5_CORE_EN
+	default y
+        ---help---
+	  Mellanox Technologies Ethernet Multi-Physical Function Switch (MPFS)
+          support in ConnectX NIC. MPFs is required for when multi-PF configuration
+          is enabled to allow passing user configured unicast MAC addresses to the
+          requesting PF.
+
+config MLX5_ESWITCH
+	bool "Mellanox Technologies MLX5 SRIOV E-Switch support"
+	depends on MLX5_CORE_EN && NET_SWITCHDEV
+	default y
+	---help---
+	  Mellanox Technologies Ethernet SRIOV E-Switch support in ConnectX NIC.
+          E-Switch provides internal SRIOV packet steering and switching for the
+          enabled VFs and PF in two available modes:
+                Legacy SRIOV mode (L2 mac vlan steering based).
+                Switchdev mode (eswitch offloads).
+
+config MLX5_CORE_EN_DCB
+	bool "Data Center Bridging (DCB) Support"
+	default y
+	depends on MLX5_CORE_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 MLX5_CORE_IPOIB
+	bool "Mellanox 5th generation network adapters (connectX series) IPoIB offloads support"
+	depends on MLX5_CORE_EN
+	default n
+	---help---
+	  MLX5 IPoIB offloads & acceleration support.
+
+config MLX5_EN_IPSEC
+	bool "IPSec XFRM cryptography-offload accelaration"
+	depends on MLX5_ACCEL
+	depends on MLX5_CORE_EN
+	depends on XFRM_OFFLOAD
+	depends on INET_ESP_OFFLOAD || INET6_ESP_OFFLOAD
+	default n
+	---help---
+	  Build support for IPsec cryptography-offload accelaration in the NIC.
+	  Note: Support for hardware with this capability needs to be selected
+	  for this option to become available.
+
+config MLX5_EN_TLS
+	bool "TLS cryptography-offload accelaration"
+	depends on MLX5_CORE_EN
+	depends on TLS_DEVICE
+	depends on TLS=y || MLX5_CORE=m
+	depends on MLX5_ACCEL
+	default n
+	---help---
+	  Build support for TLS cryptography-offload accelaration in the NIC.
+	  Note: Support for hardware with this capability needs to be selected
+	  for this option to become available.
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/Makefile b/drivers/net/ethernet/mellanox/mlx5/core/Makefile
new file mode 100644
index 000000000..d324a3884
--- /dev/null
+++ b/drivers/net/ethernet/mellanox/mlx5/core/Makefile
@@ -0,0 +1,61 @@
+# SPDX-License-Identifier: GPL-2.0
+#
+# Makefile for Mellanox 5th generation network adapters
+# (ConnectX series) core & netdev driver
+#
+
+subdir-ccflags-y += -I$(src)
+
+obj-$(CONFIG_MLX5_CORE) += mlx5_core.o
+
+#
+# mlx5 core basic
+#
+mlx5_core-y :=	main.o cmd.o debugfs.o fw.o eq.o uar.o pagealloc.o \
+		health.o mcg.o cq.o srq.o alloc.o qp.o port.o mr.o pd.o \
+		mad.o transobj.o vport.o sriov.o fs_cmd.o fs_core.o \
+		fs_counters.o rl.o lag.o dev.o wq.o lib/gid.o  \
+		diag/fs_tracepoint.o diag/fw_tracer.o
+
+#
+# Netdev basic
+#
+mlx5_core-$(CONFIG_MLX5_CORE_EN) += en_main.o en_common.o en_fs.o en_ethtool.o \
+		en_tx.o en_rx.o en_dim.o en_txrx.o en/xdp.o en_stats.o \
+		en_selftest.o en/port.o
+
+#
+# Netdev extra
+#
+mlx5_core-$(CONFIG_MLX5_EN_ARFS)     += en_arfs.o
+mlx5_core-$(CONFIG_MLX5_EN_RXNFC)    += en_fs_ethtool.o
+mlx5_core-$(CONFIG_MLX5_CORE_EN_DCB) += en_dcbnl.o en/port_buffer.o
+mlx5_core-$(CONFIG_MLX5_ESWITCH)     += en_rep.o en_tc.o
+
+#
+# Core extra
+#
+mlx5_core-$(CONFIG_MLX5_ESWITCH)   += eswitch.o eswitch_offloads.o
+mlx5_core-$(CONFIG_MLX5_MPFS)      += lib/mpfs.o
+mlx5_core-$(CONFIG_VXLAN)          += lib/vxlan.o
+mlx5_core-$(CONFIG_PTP_1588_CLOCK) += lib/clock.o
+
+#
+# Ipoib netdev
+#
+mlx5_core-$(CONFIG_MLX5_CORE_IPOIB) += ipoib/ipoib.o ipoib/ethtool.o ipoib/ipoib_vlan.o
+
+#
+# Accelerations & FPGA
+#
+mlx5_core-$(CONFIG_MLX5_ACCEL) += accel/ipsec.o accel/tls.o
+
+mlx5_core-$(CONFIG_MLX5_FPGA) += fpga/cmd.o fpga/core.o fpga/conn.o fpga/sdk.o \
+				 fpga/ipsec.o fpga/tls.o
+
+mlx5_core-$(CONFIG_MLX5_EN_IPSEC) += en_accel/ipsec.o en_accel/ipsec_rxtx.o \
+				     en_accel/ipsec_stats.o
+
+mlx5_core-$(CONFIG_MLX5_EN_TLS) += en_accel/tls.o en_accel/tls_rxtx.o en_accel/tls_stats.o
+
+CFLAGS_tracepoint.o := -I$(src)
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/accel/Makefile b/drivers/net/ethernet/mellanox/mlx5/core/accel/Makefile
new file mode 100644
index 000000000..d8e17110f
--- /dev/null
+++ b/drivers/net/ethernet/mellanox/mlx5/core/accel/Makefile
@@ -0,0 +1 @@
+subdir-ccflags-y += -I$(src)/..
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/accel/accel.h b/drivers/net/ethernet/mellanox/mlx5/core/accel/accel.h
new file mode 100644
index 000000000..c13260467
--- /dev/null
+++ b/drivers/net/ethernet/mellanox/mlx5/core/accel/accel.h
@@ -0,0 +1,37 @@
+#ifndef __MLX5E_ACCEL_H__
+#define __MLX5E_ACCEL_H__
+
+#ifdef CONFIG_MLX5_ACCEL
+
+#include <linux/skbuff.h>
+#include <linux/netdevice.h>
+#include "en.h"
+
+static inline bool is_metadata_hdr_valid(struct sk_buff *skb)
+{
+	__be16 *ethtype;
+
+	if (unlikely(skb->len < ETH_HLEN + MLX5E_METADATA_ETHER_LEN))
+		return false;
+	ethtype = (__be16 *)(skb->data + ETH_ALEN * 2);
+	if (*ethtype != cpu_to_be16(MLX5E_METADATA_ETHER_TYPE))
+		return false;
+	return true;
+}
+
+static inline void remove_metadata_hdr(struct sk_buff *skb)
+{
+	struct ethhdr *old_eth;
+	struct ethhdr *new_eth;
+
+	/* Remove the metadata from the buffer */
+	old_eth = (struct ethhdr *)skb->data;
+	new_eth = (struct ethhdr *)(skb->data + MLX5E_METADATA_ETHER_LEN);
+	memmove(new_eth, old_eth, 2 * ETH_ALEN);
+	/* Ethertype is already in its new place */
+	skb_pull_inline(skb, MLX5E_METADATA_ETHER_LEN);
+}
+
+#endif /* CONFIG_MLX5_ACCEL */
+
+#endif /* __MLX5E_EN_ACCEL_H__ */
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/accel/ipsec.c b/drivers/net/ethernet/mellanox/mlx5/core/accel/ipsec.c
new file mode 100644
index 000000000..9f1b19397
--- /dev/null
+++ b/drivers/net/ethernet/mellanox/mlx5/core/accel/ipsec.c
@@ -0,0 +1,109 @@
+/*
+ * Copyright (c) 2017 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/mlx5/device.h>
+
+#include "accel/ipsec.h"
+#include "mlx5_core.h"
+#include "fpga/ipsec.h"
+
+u32 mlx5_accel_ipsec_device_caps(struct mlx5_core_dev *mdev)
+{
+	return mlx5_fpga_ipsec_device_caps(mdev);
+}
+EXPORT_SYMBOL_GPL(mlx5_accel_ipsec_device_caps);
+
+unsigned int mlx5_accel_ipsec_counters_count(struct mlx5_core_dev *mdev)
+{
+	return mlx5_fpga_ipsec_counters_count(mdev);
+}
+
+int mlx5_accel_ipsec_counters_read(struct mlx5_core_dev *mdev, u64 *counters,
+				   unsigned int count)
+{
+	return mlx5_fpga_ipsec_counters_read(mdev, counters, count);
+}
+
+void *mlx5_accel_esp_create_hw_context(struct mlx5_core_dev *mdev,
+				       struct mlx5_accel_esp_xfrm *xfrm,
+				       const __be32 saddr[4],
+				       const __be32 daddr[4],
+				       const __be32 spi, bool is_ipv6)
+{
+	return mlx5_fpga_ipsec_create_sa_ctx(mdev, xfrm, saddr, daddr,
+					     spi, is_ipv6);
+}
+
+void mlx5_accel_esp_free_hw_context(void *context)
+{
+	mlx5_fpga_ipsec_delete_sa_ctx(context);
+}
+
+int mlx5_accel_ipsec_init(struct mlx5_core_dev *mdev)
+{
+	return mlx5_fpga_ipsec_init(mdev);
+}
+
+void mlx5_accel_ipsec_cleanup(struct mlx5_core_dev *mdev)
+{
+	mlx5_fpga_ipsec_cleanup(mdev);
+}
+
+struct mlx5_accel_esp_xfrm *
+mlx5_accel_esp_create_xfrm(struct mlx5_core_dev *mdev,
+			   const struct mlx5_accel_esp_xfrm_attrs *attrs,
+			   u32 flags)
+{
+	struct mlx5_accel_esp_xfrm *xfrm;
+
+	xfrm = mlx5_fpga_esp_create_xfrm(mdev, attrs, flags);
+	if (IS_ERR(xfrm))
+		return xfrm;
+
+	xfrm->mdev = mdev;
+	return xfrm;
+}
+EXPORT_SYMBOL_GPL(mlx5_accel_esp_create_xfrm);
+
+void mlx5_accel_esp_destroy_xfrm(struct mlx5_accel_esp_xfrm *xfrm)
+{
+	mlx5_fpga_esp_destroy_xfrm(xfrm);
+}
+EXPORT_SYMBOL_GPL(mlx5_accel_esp_destroy_xfrm);
+
+int mlx5_accel_esp_modify_xfrm(struct mlx5_accel_esp_xfrm *xfrm,
+			       const struct mlx5_accel_esp_xfrm_attrs *attrs)
+{
+	return mlx5_fpga_esp_modify_xfrm(xfrm, attrs);
+}
+EXPORT_SYMBOL_GPL(mlx5_accel_esp_modify_xfrm);
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/accel/ipsec.h b/drivers/net/ethernet/mellanox/mlx5/core/accel/ipsec.h
new file mode 100644
index 000000000..024dbd22a
--- /dev/null
+++ b/drivers/net/ethernet/mellanox/mlx5/core/accel/ipsec.h
@@ -0,0 +1,88 @@
+/*
+ * Copyright (c) 2017 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 __MLX5_ACCEL_IPSEC_H__
+#define __MLX5_ACCEL_IPSEC_H__
+
+#include <linux/mlx5/driver.h>
+#include <linux/mlx5/accel.h>
+
+#ifdef CONFIG_MLX5_ACCEL
+
+#define MLX5_IPSEC_DEV(mdev) (mlx5_accel_ipsec_device_caps(mdev) & \
+			      MLX5_ACCEL_IPSEC_CAP_DEVICE)
+
+unsigned int mlx5_accel_ipsec_counters_count(struct mlx5_core_dev *mdev);
+int mlx5_accel_ipsec_counters_read(struct mlx5_core_dev *mdev, u64 *counters,
+				   unsigned int count);
+
+void *mlx5_accel_esp_create_hw_context(struct mlx5_core_dev *mdev,
+				       struct mlx5_accel_esp_xfrm *xfrm,
+				       const __be32 saddr[4],
+				       const __be32 daddr[4],
+				       const __be32 spi, bool is_ipv6);
+void mlx5_accel_esp_free_hw_context(void *context);
+
+int mlx5_accel_ipsec_init(struct mlx5_core_dev *mdev);
+void mlx5_accel_ipsec_cleanup(struct mlx5_core_dev *mdev);
+
+#else
+
+#define MLX5_IPSEC_DEV(mdev) false
+
+static inline void *
+mlx5_accel_esp_create_hw_context(struct mlx5_core_dev *mdev,
+				 struct mlx5_accel_esp_xfrm *xfrm,
+				 const __be32 saddr[4],
+				 const __be32 daddr[4],
+				 const __be32 spi, bool is_ipv6)
+{
+	return NULL;
+}
+
+static inline void mlx5_accel_esp_free_hw_context(void *context)
+{
+}
+
+static inline int mlx5_accel_ipsec_init(struct mlx5_core_dev *mdev)
+{
+	return 0;
+}
+
+static inline void mlx5_accel_ipsec_cleanup(struct mlx5_core_dev *mdev)
+{
+}
+
+#endif
+
+#endif	/* __MLX5_ACCEL_IPSEC_H__ */
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/accel/tls.c b/drivers/net/ethernet/mellanox/mlx5/core/accel/tls.c
new file mode 100644
index 000000000..da7bd2636
--- /dev/null
+++ b/drivers/net/ethernet/mellanox/mlx5/core/accel/tls.c
@@ -0,0 +1,80 @@
+/*
+ * 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 <linux/mlx5/device.h>
+
+#include "accel/tls.h"
+#include "mlx5_core.h"
+#include "fpga/tls.h"
+
+int mlx5_accel_tls_add_flow(struct mlx5_core_dev *mdev, void *flow,
+			    struct tls_crypto_info *crypto_info,
+			    u32 start_offload_tcp_sn, u32 *p_swid,
+			    bool direction_sx)
+{
+	return mlx5_fpga_tls_add_flow(mdev, flow, crypto_info,
+				      start_offload_tcp_sn, p_swid,
+				      direction_sx);
+}
+
+void mlx5_accel_tls_del_flow(struct mlx5_core_dev *mdev, u32 swid,
+			     bool direction_sx)
+{
+	mlx5_fpga_tls_del_flow(mdev, swid, GFP_KERNEL, direction_sx);
+}
+
+int mlx5_accel_tls_resync_rx(struct mlx5_core_dev *mdev, u32 handle, u32 seq,
+			     u64 rcd_sn)
+{
+	return mlx5_fpga_tls_resync_rx(mdev, handle, seq, rcd_sn);
+}
+
+bool mlx5_accel_is_tls_device(struct mlx5_core_dev *mdev)
+{
+	return mlx5_fpga_is_tls_device(mdev);
+}
+
+u32 mlx5_accel_tls_device_caps(struct mlx5_core_dev *mdev)
+{
+	return mlx5_fpga_tls_device_caps(mdev);
+}
+
+int mlx5_accel_tls_init(struct mlx5_core_dev *mdev)
+{
+	return mlx5_fpga_tls_init(mdev);
+}
+
+void mlx5_accel_tls_cleanup(struct mlx5_core_dev *mdev)
+{
+	mlx5_fpga_tls_cleanup(mdev);
+}
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/accel/tls.h b/drivers/net/ethernet/mellanox/mlx5/core/accel/tls.h
new file mode 100644
index 000000000..def4093eb
--- /dev/null
+++ b/drivers/net/ethernet/mellanox/mlx5/core/accel/tls.h
@@ -0,0 +1,94 @@
+/*
+ * 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.
+ *
+ */
+
+#ifndef __MLX5_ACCEL_TLS_H__
+#define __MLX5_ACCEL_TLS_H__
+
+#include <linux/mlx5/driver.h>
+#include <linux/tls.h>
+
+#ifdef CONFIG_MLX5_ACCEL
+
+enum {
+	MLX5_ACCEL_TLS_TX = BIT(0),
+	MLX5_ACCEL_TLS_RX = BIT(1),
+	MLX5_ACCEL_TLS_V12 = BIT(2),
+	MLX5_ACCEL_TLS_V13 = BIT(3),
+	MLX5_ACCEL_TLS_LRO = BIT(4),
+	MLX5_ACCEL_TLS_IPV6 = BIT(5),
+	MLX5_ACCEL_TLS_AES_GCM128 = BIT(30),
+	MLX5_ACCEL_TLS_AES_GCM256 = BIT(31),
+};
+
+struct mlx5_ifc_tls_flow_bits {
+	u8         src_port[0x10];
+	u8         dst_port[0x10];
+	union mlx5_ifc_ipv6_layout_ipv4_layout_auto_bits src_ipv4_src_ipv6;
+	union mlx5_ifc_ipv6_layout_ipv4_layout_auto_bits dst_ipv4_dst_ipv6;
+	u8         ipv6[0x1];
+	u8         direction_sx[0x1];
+	u8         reserved_at_2[0x1e];
+};
+
+int mlx5_accel_tls_add_flow(struct mlx5_core_dev *mdev, void *flow,
+			    struct tls_crypto_info *crypto_info,
+			    u32 start_offload_tcp_sn, u32 *p_swid,
+			    bool direction_sx);
+void mlx5_accel_tls_del_flow(struct mlx5_core_dev *mdev, u32 swid,
+			     bool direction_sx);
+int mlx5_accel_tls_resync_rx(struct mlx5_core_dev *mdev, u32 handle, u32 seq,
+			     u64 rcd_sn);
+bool mlx5_accel_is_tls_device(struct mlx5_core_dev *mdev);
+u32 mlx5_accel_tls_device_caps(struct mlx5_core_dev *mdev);
+int mlx5_accel_tls_init(struct mlx5_core_dev *mdev);
+void mlx5_accel_tls_cleanup(struct mlx5_core_dev *mdev);
+
+#else
+
+static inline int
+mlx5_accel_tls_add_flow(struct mlx5_core_dev *mdev, void *flow,
+			struct tls_crypto_info *crypto_info,
+			u32 start_offload_tcp_sn, u32 *p_swid,
+			bool direction_sx) { return -ENOTSUPP; }
+static inline void mlx5_accel_tls_del_flow(struct mlx5_core_dev *mdev, u32 swid,
+					   bool direction_sx) { }
+static inline int mlx5_accel_tls_resync_rx(struct mlx5_core_dev *mdev, u32 handle,
+					   u32 seq, u64 rcd_sn) { return 0; }
+static inline bool mlx5_accel_is_tls_device(struct mlx5_core_dev *mdev) { return false; }
+static inline u32 mlx5_accel_tls_device_caps(struct mlx5_core_dev *mdev) { return 0; }
+static inline int mlx5_accel_tls_init(struct mlx5_core_dev *mdev) { return 0; }
+static inline void mlx5_accel_tls_cleanup(struct mlx5_core_dev *mdev) { }
+
+#endif
+
+#endif	/* __MLX5_ACCEL_TLS_H__ */
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/alloc.c b/drivers/net/ethernet/mellanox/mlx5/core/alloc.c
new file mode 100644
index 000000000..456f30007
--- /dev/null
+++ b/drivers/net/ethernet/mellanox/mlx5/core/alloc.c
@@ -0,0 +1,311 @@
+/*
+ * Copyright (c) 2013-2015, 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 <linux/mlx5/driver.h>
+
+#include "mlx5_core.h"
+
+struct mlx5_db_pgdir {
+	struct list_head	list;
+	unsigned long	       *bitmap;
+	__be32		       *db_page;
+	dma_addr_t		db_dma;
+};
+
+/* Handling for queue buffers -- we allocate a bunch of memory and
+ * register it in a memory region at HCA virtual address 0.
+ */
+
+static void *mlx5_dma_zalloc_coherent_node(struct mlx5_core_dev *dev,
+					   size_t size, dma_addr_t *dma_handle,
+					   int node)
+{
+	struct mlx5_priv *priv = &dev->priv;
+	int original_node;
+	void *cpu_handle;
+
+	mutex_lock(&priv->alloc_mutex);
+	original_node = dev_to_node(&dev->pdev->dev);
+	set_dev_node(&dev->pdev->dev, node);
+	cpu_handle = dma_zalloc_coherent(&dev->pdev->dev, size,
+					 dma_handle, GFP_KERNEL);
+	set_dev_node(&dev->pdev->dev, original_node);
+	mutex_unlock(&priv->alloc_mutex);
+	return cpu_handle;
+}
+
+int mlx5_buf_alloc_node(struct mlx5_core_dev *dev, int size,
+			struct mlx5_frag_buf *buf, int node)
+{
+	dma_addr_t t;
+
+	buf->size = size;
+	buf->npages       = 1;
+	buf->page_shift   = (u8)get_order(size) + PAGE_SHIFT;
+
+	buf->frags = kzalloc(sizeof(*buf->frags), GFP_KERNEL);
+	if (!buf->frags)
+		return -ENOMEM;
+
+	buf->frags->buf   = mlx5_dma_zalloc_coherent_node(dev, size,
+							  &t, node);
+	if (!buf->frags->buf)
+		goto err_out;
+
+	buf->frags->map = t;
+
+	while (t & ((1 << buf->page_shift) - 1)) {
+		--buf->page_shift;
+		buf->npages *= 2;
+	}
+
+	return 0;
+err_out:
+	kfree(buf->frags);
+	return -ENOMEM;
+}
+
+int mlx5_buf_alloc(struct mlx5_core_dev *dev,
+		   int size, struct mlx5_frag_buf *buf)
+{
+	return mlx5_buf_alloc_node(dev, size, buf, dev->priv.numa_node);
+}
+EXPORT_SYMBOL(mlx5_buf_alloc);
+
+void mlx5_buf_free(struct mlx5_core_dev *dev, struct mlx5_frag_buf *buf)
+{
+	dma_free_coherent(&dev->pdev->dev, buf->size, buf->frags->buf,
+			  buf->frags->map);
+
+	kfree(buf->frags);
+}
+EXPORT_SYMBOL_GPL(mlx5_buf_free);
+
+int mlx5_frag_buf_alloc_node(struct mlx5_core_dev *dev, int size,
+			     struct mlx5_frag_buf *buf, int node)
+{
+	int i;
+
+	buf->size = size;
+	buf->npages = DIV_ROUND_UP(size, PAGE_SIZE);
+	buf->page_shift = PAGE_SHIFT;
+	buf->frags = kcalloc(buf->npages, sizeof(struct mlx5_buf_list),
+			     GFP_KERNEL);
+	if (!buf->frags)
+		goto err_out;
+
+	for (i = 0; i < buf->npages; i++) {
+		struct mlx5_buf_list *frag = &buf->frags[i];
+		int frag_sz = min_t(int, size, PAGE_SIZE);
+
+		frag->buf = mlx5_dma_zalloc_coherent_node(dev, frag_sz,
+							  &frag->map, node);
+		if (!frag->buf)
+			goto err_free_buf;
+		if (frag->map & ((1 << buf->page_shift) - 1)) {
+			dma_free_coherent(&dev->pdev->dev, frag_sz,
+					  buf->frags[i].buf, buf->frags[i].map);
+			mlx5_core_warn(dev, "unexpected map alignment: %pad, page_shift=%d\n",
+				       &frag->map, buf->page_shift);
+			goto err_free_buf;
+		}
+		size -= frag_sz;
+	}
+
+	return 0;
+
+err_free_buf:
+	while (i--)
+		dma_free_coherent(&dev->pdev->dev, PAGE_SIZE, buf->frags[i].buf,
+				  buf->frags[i].map);
+	kfree(buf->frags);
+err_out:
+	return -ENOMEM;
+}
+EXPORT_SYMBOL_GPL(mlx5_frag_buf_alloc_node);
+
+void mlx5_frag_buf_free(struct mlx5_core_dev *dev, struct mlx5_frag_buf *buf)
+{
+	int size = buf->size;
+	int i;
+
+	for (i = 0; i < buf->npages; i++) {
+		int frag_sz = min_t(int, size, PAGE_SIZE);
+
+		dma_free_coherent(&dev->pdev->dev, frag_sz, buf->frags[i].buf,
+				  buf->frags[i].map);
+		size -= frag_sz;
+	}
+	kfree(buf->frags);
+}
+EXPORT_SYMBOL_GPL(mlx5_frag_buf_free);
+
+static struct mlx5_db_pgdir *mlx5_alloc_db_pgdir(struct mlx5_core_dev *dev,
+						 int node)
+{
+	u32 db_per_page = PAGE_SIZE / cache_line_size();
+	struct mlx5_db_pgdir *pgdir;
+
+	pgdir = kzalloc(sizeof(*pgdir), GFP_KERNEL);
+	if (!pgdir)
+		return NULL;
+
+	pgdir->bitmap = kcalloc(BITS_TO_LONGS(db_per_page),
+				sizeof(unsigned long),
+				GFP_KERNEL);
+
+	if (!pgdir->bitmap) {
+		kfree(pgdir);
+		return NULL;
+	}
+
+	bitmap_fill(pgdir->bitmap, db_per_page);
+
+	pgdir->db_page = mlx5_dma_zalloc_coherent_node(dev, PAGE_SIZE,
+						       &pgdir->db_dma, node);
+	if (!pgdir->db_page) {
+		kfree(pgdir->bitmap);
+		kfree(pgdir);
+		return NULL;
+	}
+
+	return pgdir;
+}
+
+static int mlx5_alloc_db_from_pgdir(struct mlx5_db_pgdir *pgdir,
+				    struct mlx5_db *db)
+{
+	u32 db_per_page = PAGE_SIZE / cache_line_size();
+	int offset;
+	int i;
+
+	i = find_first_bit(pgdir->bitmap, db_per_page);
+	if (i >= db_per_page)
+		return -ENOMEM;
+
+	__clear_bit(i, pgdir->bitmap);
+
+	db->u.pgdir = pgdir;
+	db->index   = i;
+	offset = db->index * cache_line_size();
+	db->db      = pgdir->db_page + offset / sizeof(*pgdir->db_page);
+	db->dma     = pgdir->db_dma  + offset;
+
+	db->db[0] = 0;
+	db->db[1] = 0;
+
+	return 0;
+}
+
+int mlx5_db_alloc_node(struct mlx5_core_dev *dev, struct mlx5_db *db, int node)
+{
+	struct mlx5_db_pgdir *pgdir;
+	int ret = 0;
+
+	mutex_lock(&dev->priv.pgdir_mutex);
+
+	list_for_each_entry(pgdir, &dev->priv.pgdir_list, list)
+		if (!mlx5_alloc_db_from_pgdir(pgdir, db))
+			goto out;
+
+	pgdir = mlx5_alloc_db_pgdir(dev, node);
+	if (!pgdir) {
+		ret = -ENOMEM;
+		goto out;
+	}
+
+	list_add(&pgdir->list, &dev->priv.pgdir_list);
+
+	/* This should never fail -- we just allocated an empty page: */
+	WARN_ON(mlx5_alloc_db_from_pgdir(pgdir, db));
+
+out:
+	mutex_unlock(&dev->priv.pgdir_mutex);
+
+	return ret;
+}
+EXPORT_SYMBOL_GPL(mlx5_db_alloc_node);
+
+int mlx5_db_alloc(struct mlx5_core_dev *dev, struct mlx5_db *db)
+{
+	return mlx5_db_alloc_node(dev, db, dev->priv.numa_node);
+}
+EXPORT_SYMBOL_GPL(mlx5_db_alloc);
+
+void mlx5_db_free(struct mlx5_core_dev *dev, struct mlx5_db *db)
+{
+	u32 db_per_page = PAGE_SIZE / cache_line_size();
+
+	mutex_lock(&dev->priv.pgdir_mutex);
+
+	__set_bit(db->index, db->u.pgdir->bitmap);
+
+	if (bitmap_full(db->u.pgdir->bitmap, db_per_page)) {
+		dma_free_coherent(&(dev->pdev->dev), PAGE_SIZE,
+				  db->u.pgdir->db_page, db->u.pgdir->db_dma);
+		list_del(&db->u.pgdir->list);
+		kfree(db->u.pgdir->bitmap);
+		kfree(db->u.pgdir);
+	}
+
+	mutex_unlock(&dev->priv.pgdir_mutex);
+}
+EXPORT_SYMBOL_GPL(mlx5_db_free);
+
+void mlx5_fill_page_array(struct mlx5_frag_buf *buf, __be64 *pas)
+{
+	u64 addr;
+	int i;
+
+	for (i = 0; i < buf->npages; i++) {
+		addr = buf->frags->map + (i << buf->page_shift);
+
+		pas[i] = cpu_to_be64(addr);
+	}
+}
+EXPORT_SYMBOL_GPL(mlx5_fill_page_array);
+
+void mlx5_fill_page_frag_array(struct mlx5_frag_buf *buf, __be64 *pas)
+{
+	int i;
+
+	for (i = 0; i < buf->npages; i++)
+		pas[i] = cpu_to_be64(buf->frags[i].map);
+}
+EXPORT_SYMBOL_GPL(mlx5_fill_page_frag_array);
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/cmd.c b/drivers/net/ethernet/mellanox/mlx5/core/cmd.c
new file mode 100644
index 000000000..a68608276
--- /dev/null
+++ b/drivers/net/ethernet/mellanox/mlx5/core/cmd.c
@@ -0,0 +1,1919 @@
+/*
+ * Copyright (c) 2013-2016, 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/highmem.h>
+#include <linux/module.h>
+#include <linux/errno.h>
+#include <linux/pci.h>
+#include <linux/dma-mapping.h>
+#include <linux/slab.h>
+#include <linux/delay.h>
+#include <linux/random.h>
+#include <linux/io-mapping.h>
+#include <linux/mlx5/driver.h>
+#include <linux/debugfs.h>
+
+#include "mlx5_core.h"
+
+enum {
+	CMD_IF_REV = 5,
+};
+
+enum {
+	CMD_MODE_POLLING,
+	CMD_MODE_EVENTS
+};
+
+enum {
+	MLX5_CMD_DELIVERY_STAT_OK			= 0x0,
+	MLX5_CMD_DELIVERY_STAT_SIGNAT_ERR		= 0x1,
+	MLX5_CMD_DELIVERY_STAT_TOK_ERR			= 0x2,
+	MLX5_CMD_DELIVERY_STAT_BAD_BLK_NUM_ERR		= 0x3,
+	MLX5_CMD_DELIVERY_STAT_OUT_PTR_ALIGN_ERR	= 0x4,
+	MLX5_CMD_DELIVERY_STAT_IN_PTR_ALIGN_ERR		= 0x5,
+	MLX5_CMD_DELIVERY_STAT_FW_ERR			= 0x6,
+	MLX5_CMD_DELIVERY_STAT_IN_LENGTH_ERR		= 0x7,
+	MLX5_CMD_DELIVERY_STAT_OUT_LENGTH_ERR		= 0x8,
+	MLX5_CMD_DELIVERY_STAT_RES_FLD_NOT_CLR_ERR	= 0x9,
+	MLX5_CMD_DELIVERY_STAT_CMD_DESCR_ERR		= 0x10,
+};
+
+static struct mlx5_cmd_work_ent *alloc_cmd(struct mlx5_cmd *cmd,
+					   struct mlx5_cmd_msg *in,
+					   struct mlx5_cmd_msg *out,
+					   void *uout, int uout_size,
+					   mlx5_cmd_cbk_t cbk,
+					   void *context, int page_queue)
+{
+	gfp_t alloc_flags = cbk ? GFP_ATOMIC : GFP_KERNEL;
+	struct mlx5_cmd_work_ent *ent;
+
+	ent = kzalloc(sizeof(*ent), alloc_flags);
+	if (!ent)
+		return ERR_PTR(-ENOMEM);
+
+	ent->in		= in;
+	ent->out	= out;
+	ent->uout	= uout;
+	ent->uout_size	= uout_size;
+	ent->callback	= cbk;
+	ent->context	= context;
+	ent->cmd	= cmd;
+	ent->page_queue = page_queue;
+
+	return ent;
+}
+
+static u8 alloc_token(struct mlx5_cmd *cmd)
+{
+	u8 token;
+
+	spin_lock(&cmd->token_lock);
+	cmd->token++;
+	if (cmd->token == 0)
+		cmd->token++;
+	token = cmd->token;
+	spin_unlock(&cmd->token_lock);
+
+	return token;
+}
+
+static int alloc_ent(struct mlx5_cmd *cmd)
+{
+	unsigned long flags;
+	int ret;
+
+	spin_lock_irqsave(&cmd->alloc_lock, flags);
+	ret = find_first_bit(&cmd->bitmask, cmd->max_reg_cmds);
+	if (ret < cmd->max_reg_cmds)
+		clear_bit(ret, &cmd->bitmask);
+	spin_unlock_irqrestore(&cmd->alloc_lock, flags);
+
+	return ret < cmd->max_reg_cmds ? ret : -ENOMEM;
+}
+
+static void free_ent(struct mlx5_cmd *cmd, int idx)
+{
+	unsigned long flags;
+
+	spin_lock_irqsave(&cmd->alloc_lock, flags);
+	set_bit(idx, &cmd->bitmask);
+	spin_unlock_irqrestore(&cmd->alloc_lock, flags);
+}
+
+static struct mlx5_cmd_layout *get_inst(struct mlx5_cmd *cmd, int idx)
+{
+	return cmd->cmd_buf + (idx << cmd->log_stride);
+}
+
+static int mlx5_calc_cmd_blocks(struct mlx5_cmd_msg *msg)
+{
+	int size = msg->len;
+	int blen = size - min_t(int, sizeof(msg->first.data), size);
+
+	return DIV_ROUND_UP(blen, MLX5_CMD_DATA_BLOCK_SIZE);
+}
+
+static u8 xor8_buf(void *buf, size_t offset, int len)
+{
+	u8 *ptr = buf;
+	u8 sum = 0;
+	int i;
+	int end = len + offset;
+
+	for (i = offset; i < end; i++)
+		sum ^= ptr[i];
+
+	return sum;
+}
+
+static int verify_block_sig(struct mlx5_cmd_prot_block *block)
+{
+	size_t rsvd0_off = offsetof(struct mlx5_cmd_prot_block, rsvd0);
+	int xor_len = sizeof(*block) - sizeof(block->data) - 1;
+
+	if (xor8_buf(block, rsvd0_off, xor_len) != 0xff)
+		return -EINVAL;
+
+	if (xor8_buf(block, 0, sizeof(*block)) != 0xff)
+		return -EINVAL;
+
+	return 0;
+}
+
+static void calc_block_sig(struct mlx5_cmd_prot_block *block)
+{
+	int ctrl_xor_len = sizeof(*block) - sizeof(block->data) - 2;
+	size_t rsvd0_off = offsetof(struct mlx5_cmd_prot_block, rsvd0);
+
+	block->ctrl_sig = ~xor8_buf(block, rsvd0_off, ctrl_xor_len);
+	block->sig = ~xor8_buf(block, 0, sizeof(*block) - 1);
+}
+
+static void calc_chain_sig(struct mlx5_cmd_msg *msg)
+{
+	struct mlx5_cmd_mailbox *next = msg->next;
+	int n = mlx5_calc_cmd_blocks(msg);
+	int i = 0;
+
+	for (i = 0; i < n && next; i++)  {
+		calc_block_sig(next->buf);
+		next = next->next;
+	}
+}
+
+static void set_signature(struct mlx5_cmd_work_ent *ent, int csum)
+{
+	ent->lay->sig = ~xor8_buf(ent->lay, 0,  sizeof(*ent->lay));
+	if (csum) {
+		calc_chain_sig(ent->in);
+		calc_chain_sig(ent->out);
+	}
+}
+
+static void poll_timeout(struct mlx5_cmd_work_ent *ent)
+{
+	unsigned long poll_end = jiffies + msecs_to_jiffies(MLX5_CMD_TIMEOUT_MSEC + 1000);
+	u8 own;
+
+	do {
+		own = READ_ONCE(ent->lay->status_own);
+		if (!(own & CMD_OWNER_HW)) {
+			ent->ret = 0;
+			return;
+		}
+		cond_resched();
+	} while (time_before(jiffies, poll_end));
+
+	ent->ret = -ETIMEDOUT;
+}
+
+static void free_cmd(struct mlx5_cmd_work_ent *ent)
+{
+	kfree(ent);
+}
+
+static int verify_signature(struct mlx5_cmd_work_ent *ent)
+{
+	struct mlx5_cmd_mailbox *next = ent->out->next;
+	int n = mlx5_calc_cmd_blocks(ent->out);
+	int err;
+	u8 sig;
+	int i = 0;
+
+	sig = xor8_buf(ent->lay, 0, sizeof(*ent->lay));
+	if (sig != 0xff)
+		return -EINVAL;
+
+	for (i = 0; i < n && next; i++) {
+		err = verify_block_sig(next->buf);
+		if (err)
+			return err;
+
+		next = next->next;
+	}
+
+	return 0;
+}
+
+static void dump_buf(void *buf, int size, int data_only, int offset)
+{
+	__be32 *p = buf;
+	int i;
+
+	for (i = 0; i < size; i += 16) {
+		pr_debug("%03x: %08x %08x %08x %08x\n", offset, be32_to_cpu(p[0]),
+			 be32_to_cpu(p[1]), be32_to_cpu(p[2]),
+			 be32_to_cpu(p[3]));
+		p += 4;
+		offset += 16;
+	}
+	if (!data_only)
+		pr_debug("\n");
+}
+
+static int mlx5_internal_err_ret_value(struct mlx5_core_dev *dev, u16 op,
+				       u32 *synd, u8 *status)
+{
+	*synd = 0;
+	*status = 0;
+
+	switch (op) {
+	case MLX5_CMD_OP_TEARDOWN_HCA:
+	case MLX5_CMD_OP_DISABLE_HCA:
+	case MLX5_CMD_OP_MANAGE_PAGES:
+	case MLX5_CMD_OP_DESTROY_MKEY:
+	case MLX5_CMD_OP_DESTROY_EQ:
+	case MLX5_CMD_OP_DESTROY_CQ:
+	case MLX5_CMD_OP_DESTROY_QP:
+	case MLX5_CMD_OP_DESTROY_PSV:
+	case MLX5_CMD_OP_DESTROY_SRQ:
+	case MLX5_CMD_OP_DESTROY_XRC_SRQ:
+	case MLX5_CMD_OP_DESTROY_XRQ:
+	case MLX5_CMD_OP_DESTROY_DCT:
+	case MLX5_CMD_OP_DEALLOC_Q_COUNTER:
+	case MLX5_CMD_OP_DESTROY_SCHEDULING_ELEMENT:
+	case MLX5_CMD_OP_DESTROY_QOS_PARA_VPORT:
+	case MLX5_CMD_OP_DEALLOC_PD:
+	case MLX5_CMD_OP_DEALLOC_UAR:
+	case MLX5_CMD_OP_DETACH_FROM_MCG:
+	case MLX5_CMD_OP_DEALLOC_XRCD:
+	case MLX5_CMD_OP_DEALLOC_TRANSPORT_DOMAIN:
+	case MLX5_CMD_OP_DELETE_VXLAN_UDP_DPORT:
+	case MLX5_CMD_OP_DELETE_L2_TABLE_ENTRY:
+	case MLX5_CMD_OP_DESTROY_LAG:
+	case MLX5_CMD_OP_DESTROY_VPORT_LAG:
+	case MLX5_CMD_OP_DESTROY_TIR:
+	case MLX5_CMD_OP_DESTROY_SQ:
+	case MLX5_CMD_OP_DESTROY_RQ:
+	case MLX5_CMD_OP_DESTROY_RMP:
+	case MLX5_CMD_OP_DESTROY_TIS:
+	case MLX5_CMD_OP_DESTROY_RQT:
+	case MLX5_CMD_OP_DESTROY_FLOW_TABLE:
+	case MLX5_CMD_OP_DESTROY_FLOW_GROUP:
+	case MLX5_CMD_OP_DELETE_FLOW_TABLE_ENTRY:
+	case MLX5_CMD_OP_DEALLOC_FLOW_COUNTER:
+	case MLX5_CMD_OP_2ERR_QP:
+	case MLX5_CMD_OP_2RST_QP:
+	case MLX5_CMD_OP_MODIFY_NIC_VPORT_CONTEXT:
+	case MLX5_CMD_OP_MODIFY_FLOW_TABLE:
+	case MLX5_CMD_OP_SET_FLOW_TABLE_ENTRY:
+	case MLX5_CMD_OP_SET_FLOW_TABLE_ROOT:
+	case MLX5_CMD_OP_DEALLOC_ENCAP_HEADER:
+	case MLX5_CMD_OP_DEALLOC_MODIFY_HEADER_CONTEXT:
+	case MLX5_CMD_OP_FPGA_DESTROY_QP:
+	case MLX5_CMD_OP_DESTROY_GENERAL_OBJECT:
+		return MLX5_CMD_STAT_OK;
+
+	case MLX5_CMD_OP_QUERY_HCA_CAP:
+	case MLX5_CMD_OP_QUERY_ADAPTER:
+	case MLX5_CMD_OP_INIT_HCA:
+	case MLX5_CMD_OP_ENABLE_HCA:
+	case MLX5_CMD_OP_QUERY_PAGES:
+	case MLX5_CMD_OP_SET_HCA_CAP:
+	case MLX5_CMD_OP_QUERY_ISSI:
+	case MLX5_CMD_OP_SET_ISSI:
+	case MLX5_CMD_OP_CREATE_MKEY:
+	case MLX5_CMD_OP_QUERY_MKEY:
+	case MLX5_CMD_OP_QUERY_SPECIAL_CONTEXTS:
+	case MLX5_CMD_OP_PAGE_FAULT_RESUME:
+	case MLX5_CMD_OP_CREATE_EQ:
+	case MLX5_CMD_OP_QUERY_EQ:
+	case MLX5_CMD_OP_GEN_EQE:
+	case MLX5_CMD_OP_CREATE_CQ:
+	case MLX5_CMD_OP_QUERY_CQ:
+	case MLX5_CMD_OP_MODIFY_CQ:
+	case MLX5_CMD_OP_CREATE_QP:
+	case MLX5_CMD_OP_RST2INIT_QP:
+	case MLX5_CMD_OP_INIT2RTR_QP:
+	case MLX5_CMD_OP_RTR2RTS_QP:
+	case MLX5_CMD_OP_RTS2RTS_QP:
+	case MLX5_CMD_OP_SQERR2RTS_QP:
+	case MLX5_CMD_OP_QUERY_QP:
+	case MLX5_CMD_OP_SQD_RTS_QP:
+	case MLX5_CMD_OP_INIT2INIT_QP:
+	case MLX5_CMD_OP_CREATE_PSV:
+	case MLX5_CMD_OP_CREATE_SRQ:
+	case MLX5_CMD_OP_QUERY_SRQ:
+	case MLX5_CMD_OP_ARM_RQ:
+	case MLX5_CMD_OP_CREATE_XRC_SRQ:
+	case MLX5_CMD_OP_QUERY_XRC_SRQ:
+	case MLX5_CMD_OP_ARM_XRC_SRQ:
+	case MLX5_CMD_OP_CREATE_XRQ:
+	case MLX5_CMD_OP_QUERY_XRQ:
+	case MLX5_CMD_OP_ARM_XRQ:
+	case MLX5_CMD_OP_CREATE_DCT:
+	case MLX5_CMD_OP_DRAIN_DCT:
+	case MLX5_CMD_OP_QUERY_DCT:
+	case MLX5_CMD_OP_ARM_DCT_FOR_KEY_VIOLATION:
+	case MLX5_CMD_OP_QUERY_VPORT_STATE:
+	case MLX5_CMD_OP_MODIFY_VPORT_STATE:
+	case MLX5_CMD_OP_QUERY_ESW_VPORT_CONTEXT:
+	case MLX5_CMD_OP_MODIFY_ESW_VPORT_CONTEXT:
+	case MLX5_CMD_OP_QUERY_NIC_VPORT_CONTEXT:
+	case MLX5_CMD_OP_QUERY_ROCE_ADDRESS:
+	case MLX5_CMD_OP_SET_ROCE_ADDRESS:
+	case MLX5_CMD_OP_QUERY_HCA_VPORT_CONTEXT:
+	case MLX5_CMD_OP_MODIFY_HCA_VPORT_CONTEXT:
+	case MLX5_CMD_OP_QUERY_HCA_VPORT_GID:
+	case MLX5_CMD_OP_QUERY_HCA_VPORT_PKEY:
+	case MLX5_CMD_OP_QUERY_VNIC_ENV:
+	case MLX5_CMD_OP_QUERY_VPORT_COUNTER:
+	case MLX5_CMD_OP_ALLOC_Q_COUNTER:
+	case MLX5_CMD_OP_QUERY_Q_COUNTER:
+	case MLX5_CMD_OP_SET_PP_RATE_LIMIT:
+	case MLX5_CMD_OP_QUERY_RATE_LIMIT:
+	case MLX5_CMD_OP_CREATE_SCHEDULING_ELEMENT:
+	case MLX5_CMD_OP_QUERY_SCHEDULING_ELEMENT:
+	case MLX5_CMD_OP_MODIFY_SCHEDULING_ELEMENT:
+	case MLX5_CMD_OP_CREATE_QOS_PARA_VPORT:
+	case MLX5_CMD_OP_ALLOC_PD:
+	case MLX5_CMD_OP_ALLOC_UAR:
+	case MLX5_CMD_OP_CONFIG_INT_MODERATION:
+	case MLX5_CMD_OP_ACCESS_REG:
+	case MLX5_CMD_OP_ATTACH_TO_MCG:
+	case MLX5_CMD_OP_GET_DROPPED_PACKET_LOG:
+	case MLX5_CMD_OP_MAD_IFC:
+	case MLX5_CMD_OP_QUERY_MAD_DEMUX:
+	case MLX5_CMD_OP_SET_MAD_DEMUX:
+	case MLX5_CMD_OP_NOP:
+	case MLX5_CMD_OP_ALLOC_XRCD:
+	case MLX5_CMD_OP_ALLOC_TRANSPORT_DOMAIN:
+	case MLX5_CMD_OP_QUERY_CONG_STATUS:
+	case MLX5_CMD_OP_MODIFY_CONG_STATUS:
+	case MLX5_CMD_OP_QUERY_CONG_PARAMS:
+	case MLX5_CMD_OP_MODIFY_CONG_PARAMS:
+	case MLX5_CMD_OP_QUERY_CONG_STATISTICS:
+	case MLX5_CMD_OP_ADD_VXLAN_UDP_DPORT:
+	case MLX5_CMD_OP_SET_L2_TABLE_ENTRY:
+	case MLX5_CMD_OP_QUERY_L2_TABLE_ENTRY:
+	case MLX5_CMD_OP_CREATE_LAG:
+	case MLX5_CMD_OP_MODIFY_LAG:
+	case MLX5_CMD_OP_QUERY_LAG:
+	case MLX5_CMD_OP_CREATE_VPORT_LAG:
+	case MLX5_CMD_OP_CREATE_TIR:
+	case MLX5_CMD_OP_MODIFY_TIR:
+	case MLX5_CMD_OP_QUERY_TIR:
+	case MLX5_CMD_OP_CREATE_SQ:
+	case MLX5_CMD_OP_MODIFY_SQ:
+	case MLX5_CMD_OP_QUERY_SQ:
+	case MLX5_CMD_OP_CREATE_RQ:
+	case MLX5_CMD_OP_MODIFY_RQ:
+	case MLX5_CMD_OP_QUERY_RQ:
+	case MLX5_CMD_OP_CREATE_RMP:
+	case MLX5_CMD_OP_MODIFY_RMP:
+	case MLX5_CMD_OP_QUERY_RMP:
+	case MLX5_CMD_OP_CREATE_TIS:
+	case MLX5_CMD_OP_MODIFY_TIS:
+	case MLX5_CMD_OP_QUERY_TIS:
+	case MLX5_CMD_OP_CREATE_RQT:
+	case MLX5_CMD_OP_MODIFY_RQT:
+	case MLX5_CMD_OP_QUERY_RQT:
+
+	case MLX5_CMD_OP_CREATE_FLOW_TABLE:
+	case MLX5_CMD_OP_QUERY_FLOW_TABLE:
+	case MLX5_CMD_OP_CREATE_FLOW_GROUP:
+	case MLX5_CMD_OP_QUERY_FLOW_GROUP:
+	case MLX5_CMD_OP_QUERY_FLOW_TABLE_ENTRY:
+	case MLX5_CMD_OP_ALLOC_FLOW_COUNTER:
+	case MLX5_CMD_OP_QUERY_FLOW_COUNTER:
+	case MLX5_CMD_OP_ALLOC_ENCAP_HEADER:
+	case MLX5_CMD_OP_ALLOC_MODIFY_HEADER_CONTEXT:
+	case MLX5_CMD_OP_FPGA_CREATE_QP:
+	case MLX5_CMD_OP_FPGA_MODIFY_QP:
+	case MLX5_CMD_OP_FPGA_QUERY_QP:
+	case MLX5_CMD_OP_FPGA_QUERY_QP_COUNTERS:
+	case MLX5_CMD_OP_CREATE_GENERAL_OBJECT:
+	case MLX5_CMD_OP_MODIFY_GENERAL_OBJECT:
+	case MLX5_CMD_OP_QUERY_GENERAL_OBJECT:
+		*status = MLX5_DRIVER_STATUS_ABORTED;
+		*synd = MLX5_DRIVER_SYND;
+		return -EIO;
+	default:
+		mlx5_core_err(dev, "Unknown FW command (%d)\n", op);
+		return -EINVAL;
+	}
+}
+
+const char *mlx5_command_str(int command)
+{
+#define MLX5_COMMAND_STR_CASE(__cmd) case MLX5_CMD_OP_ ## __cmd: return #__cmd
+
+	switch (command) {
+	MLX5_COMMAND_STR_CASE(QUERY_HCA_CAP);
+	MLX5_COMMAND_STR_CASE(QUERY_ADAPTER);
+	MLX5_COMMAND_STR_CASE(INIT_HCA);
+	MLX5_COMMAND_STR_CASE(TEARDOWN_HCA);
+	MLX5_COMMAND_STR_CASE(ENABLE_HCA);
+	MLX5_COMMAND_STR_CASE(DISABLE_HCA);
+	MLX5_COMMAND_STR_CASE(QUERY_PAGES);
+	MLX5_COMMAND_STR_CASE(MANAGE_PAGES);
+	MLX5_COMMAND_STR_CASE(SET_HCA_CAP);
+	MLX5_COMMAND_STR_CASE(QUERY_ISSI);
+	MLX5_COMMAND_STR_CASE(SET_ISSI);
+	MLX5_COMMAND_STR_CASE(SET_DRIVER_VERSION);
+	MLX5_COMMAND_STR_CASE(CREATE_MKEY);
+	MLX5_COMMAND_STR_CASE(QUERY_MKEY);
+	MLX5_COMMAND_STR_CASE(DESTROY_MKEY);
+	MLX5_COMMAND_STR_CASE(QUERY_SPECIAL_CONTEXTS);
+	MLX5_COMMAND_STR_CASE(PAGE_FAULT_RESUME);
+	MLX5_COMMAND_STR_CASE(CREATE_EQ);
+	MLX5_COMMAND_STR_CASE(DESTROY_EQ);
+	MLX5_COMMAND_STR_CASE(QUERY_EQ);
+	MLX5_COMMAND_STR_CASE(GEN_EQE);
+	MLX5_COMMAND_STR_CASE(CREATE_CQ);
+	MLX5_COMMAND_STR_CASE(DESTROY_CQ);
+	MLX5_COMMAND_STR_CASE(QUERY_CQ);
+	MLX5_COMMAND_STR_CASE(MODIFY_CQ);
+	MLX5_COMMAND_STR_CASE(CREATE_QP);
+	MLX5_COMMAND_STR_CASE(DESTROY_QP);
+	MLX5_COMMAND_STR_CASE(RST2INIT_QP);
+	MLX5_COMMAND_STR_CASE(INIT2RTR_QP);
+	MLX5_COMMAND_STR_CASE(RTR2RTS_QP);
+	MLX5_COMMAND_STR_CASE(RTS2RTS_QP);
+	MLX5_COMMAND_STR_CASE(SQERR2RTS_QP);
+	MLX5_COMMAND_STR_CASE(2ERR_QP);
+	MLX5_COMMAND_STR_CASE(2RST_QP);
+	MLX5_COMMAND_STR_CASE(QUERY_QP);
+	MLX5_COMMAND_STR_CASE(SQD_RTS_QP);
+	MLX5_COMMAND_STR_CASE(INIT2INIT_QP);
+	MLX5_COMMAND_STR_CASE(CREATE_PSV);
+	MLX5_COMMAND_STR_CASE(DESTROY_PSV);
+	MLX5_COMMAND_STR_CASE(CREATE_SRQ);
+	MLX5_COMMAND_STR_CASE(DESTROY_SRQ);
+	MLX5_COMMAND_STR_CASE(QUERY_SRQ);
+	MLX5_COMMAND_STR_CASE(ARM_RQ);
+	MLX5_COMMAND_STR_CASE(CREATE_XRC_SRQ);
+	MLX5_COMMAND_STR_CASE(DESTROY_XRC_SRQ);
+	MLX5_COMMAND_STR_CASE(QUERY_XRC_SRQ);
+	MLX5_COMMAND_STR_CASE(ARM_XRC_SRQ);
+	MLX5_COMMAND_STR_CASE(CREATE_DCT);
+	MLX5_COMMAND_STR_CASE(DESTROY_DCT);
+	MLX5_COMMAND_STR_CASE(DRAIN_DCT);
+	MLX5_COMMAND_STR_CASE(QUERY_DCT);
+	MLX5_COMMAND_STR_CASE(ARM_DCT_FOR_KEY_VIOLATION);
+	MLX5_COMMAND_STR_CASE(QUERY_VPORT_STATE);
+	MLX5_COMMAND_STR_CASE(MODIFY_VPORT_STATE);
+	MLX5_COMMAND_STR_CASE(QUERY_ESW_VPORT_CONTEXT);
+	MLX5_COMMAND_STR_CASE(MODIFY_ESW_VPORT_CONTEXT);
+	MLX5_COMMAND_STR_CASE(QUERY_NIC_VPORT_CONTEXT);
+	MLX5_COMMAND_STR_CASE(MODIFY_NIC_VPORT_CONTEXT);
+	MLX5_COMMAND_STR_CASE(QUERY_ROCE_ADDRESS);
+	MLX5_COMMAND_STR_CASE(SET_ROCE_ADDRESS);
+	MLX5_COMMAND_STR_CASE(QUERY_HCA_VPORT_CONTEXT);
+	MLX5_COMMAND_STR_CASE(MODIFY_HCA_VPORT_CONTEXT);
+	MLX5_COMMAND_STR_CASE(QUERY_HCA_VPORT_GID);
+	MLX5_COMMAND_STR_CASE(QUERY_HCA_VPORT_PKEY);
+	MLX5_COMMAND_STR_CASE(QUERY_VNIC_ENV);
+	MLX5_COMMAND_STR_CASE(QUERY_VPORT_COUNTER);
+	MLX5_COMMAND_STR_CASE(ALLOC_Q_COUNTER);
+	MLX5_COMMAND_STR_CASE(DEALLOC_Q_COUNTER);
+	MLX5_COMMAND_STR_CASE(QUERY_Q_COUNTER);
+	MLX5_COMMAND_STR_CASE(SET_PP_RATE_LIMIT);
+	MLX5_COMMAND_STR_CASE(QUERY_RATE_LIMIT);
+	MLX5_COMMAND_STR_CASE(CREATE_SCHEDULING_ELEMENT);
+	MLX5_COMMAND_STR_CASE(DESTROY_SCHEDULING_ELEMENT);
+	MLX5_COMMAND_STR_CASE(QUERY_SCHEDULING_ELEMENT);
+	MLX5_COMMAND_STR_CASE(MODIFY_SCHEDULING_ELEMENT);
+	MLX5_COMMAND_STR_CASE(CREATE_QOS_PARA_VPORT);
+	MLX5_COMMAND_STR_CASE(DESTROY_QOS_PARA_VPORT);
+	MLX5_COMMAND_STR_CASE(ALLOC_PD);
+	MLX5_COMMAND_STR_CASE(DEALLOC_PD);
+	MLX5_COMMAND_STR_CASE(ALLOC_UAR);
+	MLX5_COMMAND_STR_CASE(DEALLOC_UAR);
+	MLX5_COMMAND_STR_CASE(CONFIG_INT_MODERATION);
+	MLX5_COMMAND_STR_CASE(ACCESS_REG);
+	MLX5_COMMAND_STR_CASE(ATTACH_TO_MCG);
+	MLX5_COMMAND_STR_CASE(DETACH_FROM_MCG);
+	MLX5_COMMAND_STR_CASE(GET_DROPPED_PACKET_LOG);
+	MLX5_COMMAND_STR_CASE(MAD_IFC);
+	MLX5_COMMAND_STR_CASE(QUERY_MAD_DEMUX);
+	MLX5_COMMAND_STR_CASE(SET_MAD_DEMUX);
+	MLX5_COMMAND_STR_CASE(NOP);
+	MLX5_COMMAND_STR_CASE(ALLOC_XRCD);
+	MLX5_COMMAND_STR_CASE(DEALLOC_XRCD);
+	MLX5_COMMAND_STR_CASE(ALLOC_TRANSPORT_DOMAIN);
+	MLX5_COMMAND_STR_CASE(DEALLOC_TRANSPORT_DOMAIN);
+	MLX5_COMMAND_STR_CASE(QUERY_CONG_STATUS);
+	MLX5_COMMAND_STR_CASE(MODIFY_CONG_STATUS);
+	MLX5_COMMAND_STR_CASE(QUERY_CONG_PARAMS);
+	MLX5_COMMAND_STR_CASE(MODIFY_CONG_PARAMS);
+	MLX5_COMMAND_STR_CASE(QUERY_CONG_STATISTICS);
+	MLX5_COMMAND_STR_CASE(ADD_VXLAN_UDP_DPORT);
+	MLX5_COMMAND_STR_CASE(DELETE_VXLAN_UDP_DPORT);
+	MLX5_COMMAND_STR_CASE(SET_L2_TABLE_ENTRY);
+	MLX5_COMMAND_STR_CASE(QUERY_L2_TABLE_ENTRY);
+	MLX5_COMMAND_STR_CASE(DELETE_L2_TABLE_ENTRY);
+	MLX5_COMMAND_STR_CASE(SET_WOL_ROL);
+	MLX5_COMMAND_STR_CASE(QUERY_WOL_ROL);
+	MLX5_COMMAND_STR_CASE(CREATE_LAG);
+	MLX5_COMMAND_STR_CASE(MODIFY_LAG);
+	MLX5_COMMAND_STR_CASE(QUERY_LAG);
+	MLX5_COMMAND_STR_CASE(DESTROY_LAG);
+	MLX5_COMMAND_STR_CASE(CREATE_VPORT_LAG);
+	MLX5_COMMAND_STR_CASE(DESTROY_VPORT_LAG);
+	MLX5_COMMAND_STR_CASE(CREATE_TIR);
+	MLX5_COMMAND_STR_CASE(MODIFY_TIR);
+	MLX5_COMMAND_STR_CASE(DESTROY_TIR);
+	MLX5_COMMAND_STR_CASE(QUERY_TIR);
+	MLX5_COMMAND_STR_CASE(CREATE_SQ);
+	MLX5_COMMAND_STR_CASE(MODIFY_SQ);
+	MLX5_COMMAND_STR_CASE(DESTROY_SQ);
+	MLX5_COMMAND_STR_CASE(QUERY_SQ);
+	MLX5_COMMAND_STR_CASE(CREATE_RQ);
+	MLX5_COMMAND_STR_CASE(MODIFY_RQ);
+	MLX5_COMMAND_STR_CASE(DESTROY_RQ);
+	MLX5_COMMAND_STR_CASE(QUERY_RQ);
+	MLX5_COMMAND_STR_CASE(CREATE_RMP);
+	MLX5_COMMAND_STR_CASE(MODIFY_RMP);
+	MLX5_COMMAND_STR_CASE(DESTROY_RMP);
+	MLX5_COMMAND_STR_CASE(QUERY_RMP);
+	MLX5_COMMAND_STR_CASE(CREATE_TIS);
+	MLX5_COMMAND_STR_CASE(MODIFY_TIS);
+	MLX5_COMMAND_STR_CASE(DESTROY_TIS);
+	MLX5_COMMAND_STR_CASE(QUERY_TIS);
+	MLX5_COMMAND_STR_CASE(CREATE_RQT);
+	MLX5_COMMAND_STR_CASE(MODIFY_RQT);
+	MLX5_COMMAND_STR_CASE(DESTROY_RQT);
+	MLX5_COMMAND_STR_CASE(QUERY_RQT);
+	MLX5_COMMAND_STR_CASE(SET_FLOW_TABLE_ROOT);
+	MLX5_COMMAND_STR_CASE(CREATE_FLOW_TABLE);
+	MLX5_COMMAND_STR_CASE(DESTROY_FLOW_TABLE);
+	MLX5_COMMAND_STR_CASE(QUERY_FLOW_TABLE);
+	MLX5_COMMAND_STR_CASE(CREATE_FLOW_GROUP);
+	MLX5_COMMAND_STR_CASE(DESTROY_FLOW_GROUP);
+	MLX5_COMMAND_STR_CASE(QUERY_FLOW_GROUP);
+	MLX5_COMMAND_STR_CASE(SET_FLOW_TABLE_ENTRY);
+	MLX5_COMMAND_STR_CASE(QUERY_FLOW_TABLE_ENTRY);
+	MLX5_COMMAND_STR_CASE(DELETE_FLOW_TABLE_ENTRY);
+	MLX5_COMMAND_STR_CASE(ALLOC_FLOW_COUNTER);
+	MLX5_COMMAND_STR_CASE(DEALLOC_FLOW_COUNTER);
+	MLX5_COMMAND_STR_CASE(QUERY_FLOW_COUNTER);
+	MLX5_COMMAND_STR_CASE(MODIFY_FLOW_TABLE);
+	MLX5_COMMAND_STR_CASE(ALLOC_ENCAP_HEADER);
+	MLX5_COMMAND_STR_CASE(DEALLOC_ENCAP_HEADER);
+	MLX5_COMMAND_STR_CASE(ALLOC_MODIFY_HEADER_CONTEXT);
+	MLX5_COMMAND_STR_CASE(DEALLOC_MODIFY_HEADER_CONTEXT);
+	MLX5_COMMAND_STR_CASE(FPGA_CREATE_QP);
+	MLX5_COMMAND_STR_CASE(FPGA_MODIFY_QP);
+	MLX5_COMMAND_STR_CASE(FPGA_QUERY_QP);
+	MLX5_COMMAND_STR_CASE(FPGA_QUERY_QP_COUNTERS);
+	MLX5_COMMAND_STR_CASE(FPGA_DESTROY_QP);
+	MLX5_COMMAND_STR_CASE(CREATE_XRQ);
+	MLX5_COMMAND_STR_CASE(DESTROY_XRQ);
+	MLX5_COMMAND_STR_CASE(QUERY_XRQ);
+	MLX5_COMMAND_STR_CASE(ARM_XRQ);
+	MLX5_COMMAND_STR_CASE(CREATE_GENERAL_OBJECT);
+	MLX5_COMMAND_STR_CASE(DESTROY_GENERAL_OBJECT);
+	MLX5_COMMAND_STR_CASE(MODIFY_GENERAL_OBJECT);
+	MLX5_COMMAND_STR_CASE(QUERY_GENERAL_OBJECT);
+	MLX5_COMMAND_STR_CASE(QUERY_MODIFY_HEADER_CONTEXT);
+	default: return "unknown command opcode";
+	}
+}
+
+static const char *cmd_status_str(u8 status)
+{
+	switch (status) {
+	case MLX5_CMD_STAT_OK:
+		return "OK";
+	case MLX5_CMD_STAT_INT_ERR:
+		return "internal error";
+	case MLX5_CMD_STAT_BAD_OP_ERR:
+		return "bad operation";
+	case MLX5_CMD_STAT_BAD_PARAM_ERR:
+		return "bad parameter";
+	case MLX5_CMD_STAT_BAD_SYS_STATE_ERR:
+		return "bad system state";
+	case MLX5_CMD_STAT_BAD_RES_ERR:
+		return "bad resource";
+	case MLX5_CMD_STAT_RES_BUSY:
+		return "resource busy";
+	case MLX5_CMD_STAT_LIM_ERR:
+		return "limits exceeded";
+	case MLX5_CMD_STAT_BAD_RES_STATE_ERR:
+		return "bad resource state";
+	case MLX5_CMD_STAT_IX_ERR:
+		return "bad index";
+	case MLX5_CMD_STAT_NO_RES_ERR:
+		return "no resources";
+	case MLX5_CMD_STAT_BAD_INP_LEN_ERR:
+		return "bad input length";
+	case MLX5_CMD_STAT_BAD_OUTP_LEN_ERR:
+		return "bad output length";
+	case MLX5_CMD_STAT_BAD_QP_STATE_ERR:
+		return "bad QP state";
+	case MLX5_CMD_STAT_BAD_PKT_ERR:
+		return "bad packet (discarded)";
+	case MLX5_CMD_STAT_BAD_SIZE_OUTS_CQES_ERR:
+		return "bad size too many outstanding CQEs";
+	default:
+		return "unknown status";
+	}
+}
+
+static int cmd_status_to_err(u8 status)
+{
+	switch (status) {
+	case MLX5_CMD_STAT_OK:				return 0;
+	case MLX5_CMD_STAT_INT_ERR:			return -EIO;
+	case MLX5_CMD_STAT_BAD_OP_ERR:			return -EINVAL;
+	case MLX5_CMD_STAT_BAD_PARAM_ERR:		return -EINVAL;
+	case MLX5_CMD_STAT_BAD_SYS_STATE_ERR:		return -EIO;
+	case MLX5_CMD_STAT_BAD_RES_ERR:			return -EINVAL;
+	case MLX5_CMD_STAT_RES_BUSY:			return -EBUSY;
+	case MLX5_CMD_STAT_LIM_ERR:			return -ENOMEM;
+	case MLX5_CMD_STAT_BAD_RES_STATE_ERR:		return -EINVAL;
+	case MLX5_CMD_STAT_IX_ERR:			return -EINVAL;
+	case MLX5_CMD_STAT_NO_RES_ERR:			return -EAGAIN;
+	case MLX5_CMD_STAT_BAD_INP_LEN_ERR:		return -EIO;
+	case MLX5_CMD_STAT_BAD_OUTP_LEN_ERR:		return -EIO;
+	case MLX5_CMD_STAT_BAD_QP_STATE_ERR:		return -EINVAL;
+	case MLX5_CMD_STAT_BAD_PKT_ERR:			return -EINVAL;
+	case MLX5_CMD_STAT_BAD_SIZE_OUTS_CQES_ERR:	return -EINVAL;
+	default:					return -EIO;
+	}
+}
+
+struct mlx5_ifc_mbox_out_bits {
+	u8         status[0x8];
+	u8         reserved_at_8[0x18];
+
+	u8         syndrome[0x20];
+
+	u8         reserved_at_40[0x40];
+};
+
+struct mlx5_ifc_mbox_in_bits {
+	u8         opcode[0x10];
+	u8         uid[0x10];
+
+	u8         reserved_at_20[0x10];
+	u8         op_mod[0x10];
+
+	u8         reserved_at_40[0x40];
+};
+
+void mlx5_cmd_mbox_status(void *out, u8 *status, u32 *syndrome)
+{
+	*status = MLX5_GET(mbox_out, out, status);
+	*syndrome = MLX5_GET(mbox_out, out, syndrome);
+}
+
+static int mlx5_cmd_check(struct mlx5_core_dev *dev, void *in, void *out)
+{
+	u32 syndrome;
+	u8  status;
+	u16 opcode;
+	u16 op_mod;
+	u16 uid;
+
+	mlx5_cmd_mbox_status(out, &status, &syndrome);
+	if (!status)
+		return 0;
+
+	opcode = MLX5_GET(mbox_in, in, opcode);
+	op_mod = MLX5_GET(mbox_in, in, op_mod);
+	uid    = MLX5_GET(mbox_in, in, uid);
+
+	if (!uid && opcode != MLX5_CMD_OP_DESTROY_MKEY)
+		mlx5_core_err_rl(dev,
+			"%s(0x%x) op_mod(0x%x) failed, status %s(0x%x), syndrome (0x%x)\n",
+			mlx5_command_str(opcode), opcode, op_mod,
+			cmd_status_str(status), status, syndrome);
+	else
+		mlx5_core_dbg(dev,
+		      "%s(0x%x) op_mod(0x%x) failed, status %s(0x%x), syndrome (0x%x)\n",
+		      mlx5_command_str(opcode),
+		      opcode, op_mod,
+		      cmd_status_str(status),
+		      status,
+		      syndrome);
+
+	return cmd_status_to_err(status);
+}
+
+static void dump_command(struct mlx5_core_dev *dev,
+			 struct mlx5_cmd_work_ent *ent, int input)
+{
+	struct mlx5_cmd_msg *msg = input ? ent->in : ent->out;
+	u16 op = MLX5_GET(mbox_in, ent->lay->in, opcode);
+	struct mlx5_cmd_mailbox *next = msg->next;
+	int n = mlx5_calc_cmd_blocks(msg);
+	int data_only;
+	u32 offset = 0;
+	int dump_len;
+	int i;
+
+	data_only = !!(mlx5_core_debug_mask & (1 << MLX5_CMD_DATA));
+
+	if (data_only)
+		mlx5_core_dbg_mask(dev, 1 << MLX5_CMD_DATA,
+				   "dump command data %s(0x%x) %s\n",
+				   mlx5_command_str(op), op,
+				   input ? "INPUT" : "OUTPUT");
+	else
+		mlx5_core_dbg(dev, "dump command %s(0x%x) %s\n",
+			      mlx5_command_str(op), op,
+			      input ? "INPUT" : "OUTPUT");
+
+	if (data_only) {
+		if (input) {
+			dump_buf(ent->lay->in, sizeof(ent->lay->in), 1, offset);
+			offset += sizeof(ent->lay->in);
+		} else {
+			dump_buf(ent->lay->out, sizeof(ent->lay->out), 1, offset);
+			offset += sizeof(ent->lay->out);
+		}
+	} else {
+		dump_buf(ent->lay, sizeof(*ent->lay), 0, offset);
+		offset += sizeof(*ent->lay);
+	}
+
+	for (i = 0; i < n && next; i++)  {
+		if (data_only) {
+			dump_len = min_t(int, MLX5_CMD_DATA_BLOCK_SIZE, msg->len - offset);
+			dump_buf(next->buf, dump_len, 1, offset);
+			offset += MLX5_CMD_DATA_BLOCK_SIZE;
+		} else {
+			mlx5_core_dbg(dev, "command block:\n");
+			dump_buf(next->buf, sizeof(struct mlx5_cmd_prot_block), 0, offset);
+			offset += sizeof(struct mlx5_cmd_prot_block);
+		}
+		next = next->next;
+	}
+
+	if (data_only)
+		pr_debug("\n");
+}
+
+static u16 msg_to_opcode(struct mlx5_cmd_msg *in)
+{
+	return MLX5_GET(mbox_in, in->first.data, opcode);
+}
+
+static void cb_timeout_handler(struct work_struct *work)
+{
+	struct delayed_work *dwork = container_of(work, struct delayed_work,
+						  work);
+	struct mlx5_cmd_work_ent *ent = container_of(dwork,
+						     struct mlx5_cmd_work_ent,
+						     cb_timeout_work);
+	struct mlx5_core_dev *dev = container_of(ent->cmd, struct mlx5_core_dev,
+						 cmd);
+
+	ent->ret = -ETIMEDOUT;
+	mlx5_core_warn(dev, "%s(0x%x) timeout. Will cause a leak of a command resource\n",
+		       mlx5_command_str(msg_to_opcode(ent->in)),
+		       msg_to_opcode(ent->in));
+	mlx5_cmd_comp_handler(dev, 1UL << ent->idx, true);
+}
+
+static void free_msg(struct mlx5_core_dev *dev, struct mlx5_cmd_msg *msg);
+static void mlx5_free_cmd_msg(struct mlx5_core_dev *dev,
+			      struct mlx5_cmd_msg *msg);
+
+static void cmd_work_handler(struct work_struct *work)
+{
+	struct mlx5_cmd_work_ent *ent = container_of(work, struct mlx5_cmd_work_ent, work);
+	struct mlx5_cmd *cmd = ent->cmd;
+	struct mlx5_core_dev *dev = container_of(cmd, struct mlx5_core_dev, cmd);
+	unsigned long cb_timeout = msecs_to_jiffies(MLX5_CMD_TIMEOUT_MSEC);
+	struct mlx5_cmd_layout *lay;
+	struct semaphore *sem;
+	unsigned long flags;
+	bool poll_cmd = ent->polling;
+	int alloc_ret;
+	int cmd_mode;
+
+	complete(&ent->handling);
+	sem = ent->page_queue ? &cmd->pages_sem : &cmd->sem;
+	down(sem);
+	if (!ent->page_queue) {
+		alloc_ret = alloc_ent(cmd);
+		if (alloc_ret < 0) {
+			mlx5_core_err(dev, "failed to allocate command entry\n");
+			if (ent->callback) {
+				ent->callback(-EAGAIN, ent->context);
+				mlx5_free_cmd_msg(dev, ent->out);
+				free_msg(dev, ent->in);
+				free_cmd(ent);
+			} else {
+				ent->ret = -EAGAIN;
+				complete(&ent->done);
+			}
+			up(sem);
+			return;
+		}
+		ent->idx = alloc_ret;
+	} else {
+		ent->idx = cmd->max_reg_cmds;
+		spin_lock_irqsave(&cmd->alloc_lock, flags);
+		clear_bit(ent->idx, &cmd->bitmask);
+		spin_unlock_irqrestore(&cmd->alloc_lock, flags);
+	}
+
+	cmd->ent_arr[ent->idx] = ent;
+	lay = get_inst(cmd, ent->idx);
+	ent->lay = lay;
+	memset(lay, 0, sizeof(*lay));
+	memcpy(lay->in, ent->in->first.data, sizeof(lay->in));
+	ent->op = be32_to_cpu(lay->in[0]) >> 16;
+	if (ent->in->next)
+		lay->in_ptr = cpu_to_be64(ent->in->next->dma);
+	lay->inlen = cpu_to_be32(ent->in->len);
+	if (ent->out->next)
+		lay->out_ptr = cpu_to_be64(ent->out->next->dma);
+	lay->outlen = cpu_to_be32(ent->out->len);
+	lay->type = MLX5_PCI_CMD_XPORT;
+	lay->token = ent->token;
+	lay->status_own = CMD_OWNER_HW;
+	set_signature(ent, !cmd->checksum_disabled);
+	dump_command(dev, ent, 1);
+	ent->ts1 = ktime_get_ns();
+	cmd_mode = cmd->mode;
+
+	if (ent->callback)
+		schedule_delayed_work(&ent->cb_timeout_work, cb_timeout);
+	set_bit(MLX5_CMD_ENT_STATE_PENDING_COMP, &ent->state);
+
+	/* Skip sending command to fw if internal error */
+	if (pci_channel_offline(dev->pdev) ||
+	    dev->state == MLX5_DEVICE_STATE_INTERNAL_ERROR) {
+		u8 status = 0;
+		u32 drv_synd;
+
+		ent->ret = mlx5_internal_err_ret_value(dev, msg_to_opcode(ent->in), &drv_synd, &status);
+		MLX5_SET(mbox_out, ent->out, status, status);
+		MLX5_SET(mbox_out, ent->out, syndrome, drv_synd);
+
+		mlx5_cmd_comp_handler(dev, 1UL << ent->idx, true);
+		/* no doorbell, no need to keep the entry */
+		free_ent(cmd, ent->idx);
+		if (ent->callback)
+			free_cmd(ent);
+		return;
+	}
+
+	/* ring doorbell after the descriptor is valid */
+	mlx5_core_dbg(dev, "writing 0x%x to command doorbell\n", 1 << ent->idx);
+	wmb();
+	iowrite32be(1 << ent->idx, &dev->iseg->cmd_dbell);
+	mmiowb();
+	/* if not in polling don't use ent after this point */
+	if (cmd_mode == CMD_MODE_POLLING || poll_cmd) {
+		poll_timeout(ent);
+		/* make sure we read the descriptor after ownership is SW */
+		rmb();
+		mlx5_cmd_comp_handler(dev, 1UL << ent->idx, (ent->ret == -ETIMEDOUT));
+	}
+}
+
+static const char *deliv_status_to_str(u8 status)
+{
+	switch (status) {
+	case MLX5_CMD_DELIVERY_STAT_OK:
+		return "no errors";
+	case MLX5_CMD_DELIVERY_STAT_SIGNAT_ERR:
+		return "signature error";
+	case MLX5_CMD_DELIVERY_STAT_TOK_ERR:
+		return "token error";
+	case MLX5_CMD_DELIVERY_STAT_BAD_BLK_NUM_ERR:
+		return "bad block number";
+	case MLX5_CMD_DELIVERY_STAT_OUT_PTR_ALIGN_ERR:
+		return "output pointer not aligned to block size";
+	case MLX5_CMD_DELIVERY_STAT_IN_PTR_ALIGN_ERR:
+		return "input pointer not aligned to block size";
+	case MLX5_CMD_DELIVERY_STAT_FW_ERR:
+		return "firmware internal error";
+	case MLX5_CMD_DELIVERY_STAT_IN_LENGTH_ERR:
+		return "command input length error";
+	case MLX5_CMD_DELIVERY_STAT_OUT_LENGTH_ERR:
+		return "command output length error";
+	case MLX5_CMD_DELIVERY_STAT_RES_FLD_NOT_CLR_ERR:
+		return "reserved fields not cleared";
+	case MLX5_CMD_DELIVERY_STAT_CMD_DESCR_ERR:
+		return "bad command descriptor type";
+	default:
+		return "unknown status code";
+	}
+}
+
+static int wait_func(struct mlx5_core_dev *dev, struct mlx5_cmd_work_ent *ent)
+{
+	unsigned long timeout = msecs_to_jiffies(MLX5_CMD_TIMEOUT_MSEC);
+	struct mlx5_cmd *cmd = &dev->cmd;
+	int err;
+
+	if (!wait_for_completion_timeout(&ent->handling, timeout) &&
+	    cancel_work_sync(&ent->work)) {
+		ent->ret = -ECANCELED;
+		goto out_err;
+	}
+	if (cmd->mode == CMD_MODE_POLLING || ent->polling) {
+		wait_for_completion(&ent->done);
+	} else if (!wait_for_completion_timeout(&ent->done, timeout)) {
+		ent->ret = -ETIMEDOUT;
+		mlx5_cmd_comp_handler(dev, 1UL << ent->idx, true);
+	}
+
+out_err:
+	err = ent->ret;
+
+	if (err == -ETIMEDOUT) {
+		mlx5_core_warn(dev, "%s(0x%x) timeout. Will cause a leak of a command resource\n",
+			       mlx5_command_str(msg_to_opcode(ent->in)),
+			       msg_to_opcode(ent->in));
+	} else if (err == -ECANCELED) {
+		mlx5_core_warn(dev, "%s(0x%x) canceled on out of queue timeout.\n",
+			       mlx5_command_str(msg_to_opcode(ent->in)),
+			       msg_to_opcode(ent->in));
+	}
+	mlx5_core_dbg(dev, "err %d, delivery status %s(%d)\n",
+		      err, deliv_status_to_str(ent->status), ent->status);
+
+	return err;
+}
+
+/*  Notes:
+ *    1. Callback functions may not sleep
+ *    2. page queue commands do not support asynchrous completion
+ */
+static int mlx5_cmd_invoke(struct mlx5_core_dev *dev, struct mlx5_cmd_msg *in,
+			   struct mlx5_cmd_msg *out, void *uout, int uout_size,
+			   mlx5_cmd_cbk_t callback,
+			   void *context, int page_queue, u8 *status,
+			   u8 token, bool force_polling)
+{
+	struct mlx5_cmd *cmd = &dev->cmd;
+	struct mlx5_cmd_work_ent *ent;
+	struct mlx5_cmd_stats *stats;
+	int err = 0;
+	s64 ds;
+	u16 op;
+
+	if (callback && page_queue)
+		return -EINVAL;
+
+	ent = alloc_cmd(cmd, in, out, uout, uout_size, callback, context,
+			page_queue);
+	if (IS_ERR(ent))
+		return PTR_ERR(ent);
+
+	ent->token = token;
+	ent->polling = force_polling;
+
+	init_completion(&ent->handling);
+	if (!callback)
+		init_completion(&ent->done);
+
+	INIT_DELAYED_WORK(&ent->cb_timeout_work, cb_timeout_handler);
+	INIT_WORK(&ent->work, cmd_work_handler);
+	if (page_queue) {
+		cmd_work_handler(&ent->work);
+	} else if (!queue_work(cmd->wq, &ent->work)) {
+		mlx5_core_warn(dev, "failed to queue work\n");
+		err = -ENOMEM;
+		goto out_free;
+	}
+
+	if (callback)
+		goto out;
+
+	err = wait_func(dev, ent);
+	if (err == -ETIMEDOUT)
+		goto out;
+	if (err == -ECANCELED)
+		goto out_free;
+
+	ds = ent->ts2 - ent->ts1;
+	op = MLX5_GET(mbox_in, in->first.data, opcode);
+	if (op < ARRAY_SIZE(cmd->stats)) {
+		stats = &cmd->stats[op];
+		spin_lock_irq(&stats->lock);
+		stats->sum += ds;
+		++stats->n;
+		spin_unlock_irq(&stats->lock);
+	}
+	mlx5_core_dbg_mask(dev, 1 << MLX5_CMD_TIME,
+			   "fw exec time for %s is %lld nsec\n",
+			   mlx5_command_str(op), ds);
+	*status = ent->status;
+
+out_free:
+	free_cmd(ent);
+out:
+	return err;
+}
+
+static ssize_t dbg_write(struct file *filp, const char __user *buf,
+			 size_t count, loff_t *pos)
+{
+	struct mlx5_core_dev *dev = filp->private_data;
+	struct mlx5_cmd_debug *dbg = &dev->cmd.dbg;
+	char lbuf[3];
+	int err;
+
+	if (!dbg->in_msg || !dbg->out_msg)
+		return -ENOMEM;
+
+	if (count < sizeof(lbuf) - 1)
+		return -EINVAL;
+
+	if (copy_from_user(lbuf, buf, sizeof(lbuf) - 1))
+		return -EFAULT;
+
+	lbuf[sizeof(lbuf) - 1] = 0;
+
+	if (strcmp(lbuf, "go"))
+		return -EINVAL;
+
+	err = mlx5_cmd_exec(dev, dbg->in_msg, dbg->inlen, dbg->out_msg, dbg->outlen);
+
+	return err ? err : count;
+}
+
+static const struct file_operations fops = {
+	.owner	= THIS_MODULE,
+	.open	= simple_open,
+	.write	= dbg_write,
+};
+
+static int mlx5_copy_to_msg(struct mlx5_cmd_msg *to, void *from, int size,
+			    u8 token)
+{
+	struct mlx5_cmd_prot_block *block;
+	struct mlx5_cmd_mailbox *next;
+	int copy;
+
+	if (!to || !from)
+		return -ENOMEM;
+
+	copy = min_t(int, size, sizeof(to->first.data));
+	memcpy(to->first.data, from, copy);
+	size -= copy;
+	from += copy;
+
+	next = to->next;
+	while (size) {
+		if (!next) {
+			/* this is a BUG */
+			return -ENOMEM;
+		}
+
+		copy = min_t(int, size, MLX5_CMD_DATA_BLOCK_SIZE);
+		block = next->buf;
+		memcpy(block->data, from, copy);
+		from += copy;
+		size -= copy;
+		block->token = token;
+		next = next->next;
+	}
+
+	return 0;
+}
+
+static int mlx5_copy_from_msg(void *to, struct mlx5_cmd_msg *from, int size)
+{
+	struct mlx5_cmd_prot_block *block;
+	struct mlx5_cmd_mailbox *next;
+	int copy;
+
+	if (!to || !from)
+		return -ENOMEM;
+
+	copy = min_t(int, size, sizeof(from->first.data));
+	memcpy(to, from->first.data, copy);
+	size -= copy;
+	to += copy;
+
+	next = from->next;
+	while (size) {
+		if (!next) {
+			/* this is a BUG */
+			return -ENOMEM;
+		}
+
+		copy = min_t(int, size, MLX5_CMD_DATA_BLOCK_SIZE);
+		block = next->buf;
+
+		memcpy(to, block->data, copy);
+		to += copy;
+		size -= copy;
+		next = next->next;
+	}
+
+	return 0;
+}
+
+static struct mlx5_cmd_mailbox *alloc_cmd_box(struct mlx5_core_dev *dev,
+					      gfp_t flags)
+{
+	struct mlx5_cmd_mailbox *mailbox;
+
+	mailbox = kmalloc(sizeof(*mailbox), flags);
+	if (!mailbox)
+		return ERR_PTR(-ENOMEM);
+
+	mailbox->buf = dma_pool_zalloc(dev->cmd.pool, flags,
+				       &mailbox->dma);
+	if (!mailbox->buf) {
+		mlx5_core_dbg(dev, "failed allocation\n");
+		kfree(mailbox);
+		return ERR_PTR(-ENOMEM);
+	}
+	mailbox->next = NULL;
+
+	return mailbox;
+}
+
+static void free_cmd_box(struct mlx5_core_dev *dev,
+			 struct mlx5_cmd_mailbox *mailbox)
+{
+	dma_pool_free(dev->cmd.pool, mailbox->buf, mailbox->dma);
+	kfree(mailbox);
+}
+
+static struct mlx5_cmd_msg *mlx5_alloc_cmd_msg(struct mlx5_core_dev *dev,
+					       gfp_t flags, int size,
+					       u8 token)
+{
+	struct mlx5_cmd_mailbox *tmp, *head = NULL;
+	struct mlx5_cmd_prot_block *block;
+	struct mlx5_cmd_msg *msg;
+	int err;
+	int n;
+	int i;
+
+	msg = kzalloc(sizeof(*msg), flags);
+	if (!msg)
+		return ERR_PTR(-ENOMEM);
+
+	msg->len = size;
+	n = mlx5_calc_cmd_blocks(msg);
+
+	for (i = 0; i < n; i++) {
+		tmp = alloc_cmd_box(dev, flags);
+		if (IS_ERR(tmp)) {
+			mlx5_core_warn(dev, "failed allocating block\n");
+			err = PTR_ERR(tmp);
+			goto err_alloc;
+		}
+
+		block = tmp->buf;
+		tmp->next = head;
+		block->next = cpu_to_be64(tmp->next ? tmp->next->dma : 0);
+		block->block_num = cpu_to_be32(n - i - 1);
+		block->token = token;
+		head = tmp;
+	}
+	msg->next = head;
+	return msg;
+
+err_alloc:
+	while (head) {
+		tmp = head->next;
+		free_cmd_box(dev, head);
+		head = tmp;
+	}
+	kfree(msg);
+
+	return ERR_PTR(err);
+}
+
+static void mlx5_free_cmd_msg(struct mlx5_core_dev *dev,
+			      struct mlx5_cmd_msg *msg)
+{
+	struct mlx5_cmd_mailbox *head = msg->next;
+	struct mlx5_cmd_mailbox *next;
+
+	while (head) {
+		next = head->next;
+		free_cmd_box(dev, head);
+		head = next;
+	}
+	kfree(msg);
+}
+
+static ssize_t data_write(struct file *filp, const char __user *buf,
+			  size_t count, loff_t *pos)
+{
+	struct mlx5_core_dev *dev = filp->private_data;
+	struct mlx5_cmd_debug *dbg = &dev->cmd.dbg;
+	void *ptr;
+
+	if (*pos != 0)
+		return -EINVAL;
+
+	kfree(dbg->in_msg);
+	dbg->in_msg = NULL;
+	dbg->inlen = 0;
+	ptr = memdup_user(buf, count);
+	if (IS_ERR(ptr))
+		return PTR_ERR(ptr);
+	dbg->in_msg = ptr;
+	dbg->inlen = count;
+
+	*pos = count;
+
+	return count;
+}
+
+static ssize_t data_read(struct file *filp, char __user *buf, size_t count,
+			 loff_t *pos)
+{
+	struct mlx5_core_dev *dev = filp->private_data;
+	struct mlx5_cmd_debug *dbg = &dev->cmd.dbg;
+
+	if (!dbg->out_msg)
+		return -ENOMEM;
+
+	return simple_read_from_buffer(buf, count, pos, dbg->out_msg,
+				       dbg->outlen);
+}
+
+static const struct file_operations dfops = {
+	.owner	= THIS_MODULE,
+	.open	= simple_open,
+	.write	= data_write,
+	.read	= data_read,
+};
+
+static ssize_t outlen_read(struct file *filp, char __user *buf, size_t count,
+			   loff_t *pos)
+{
+	struct mlx5_core_dev *dev = filp->private_data;
+	struct mlx5_cmd_debug *dbg = &dev->cmd.dbg;
+	char outlen[8];
+	int err;
+
+	err = snprintf(outlen, sizeof(outlen), "%d", dbg->outlen);
+	if (err < 0)
+		return err;
+
+	return simple_read_from_buffer(buf, count, pos, outlen, err);
+}
+
+static ssize_t outlen_write(struct file *filp, const char __user *buf,
+			    size_t count, loff_t *pos)
+{
+	struct mlx5_core_dev *dev = filp->private_data;
+	struct mlx5_cmd_debug *dbg = &dev->cmd.dbg;
+	char outlen_str[8] = {0};
+	int outlen;
+	void *ptr;
+	int err;
+
+	if (*pos != 0 || count > 6)
+		return -EINVAL;
+
+	kfree(dbg->out_msg);
+	dbg->out_msg = NULL;
+	dbg->outlen = 0;
+
+	if (copy_from_user(outlen_str, buf, count))
+		return -EFAULT;
+
+	err = sscanf(outlen_str, "%d", &outlen);
+	if (err < 0)
+		return err;
+
+	ptr = kzalloc(outlen, GFP_KERNEL);
+	if (!ptr)
+		return -ENOMEM;
+
+	dbg->out_msg = ptr;
+	dbg->outlen = outlen;
+
+	*pos = count;
+
+	return count;
+}
+
+static const struct file_operations olfops = {
+	.owner	= THIS_MODULE,
+	.open	= simple_open,
+	.write	= outlen_write,
+	.read	= outlen_read,
+};
+
+static void set_wqname(struct mlx5_core_dev *dev)
+{
+	struct mlx5_cmd *cmd = &dev->cmd;
+
+	snprintf(cmd->wq_name, sizeof(cmd->wq_name), "mlx5_cmd_%s",
+		 dev_name(&dev->pdev->dev));
+}
+
+static void clean_debug_files(struct mlx5_core_dev *dev)
+{
+	struct mlx5_cmd_debug *dbg = &dev->cmd.dbg;
+
+	if (!mlx5_debugfs_root)
+		return;
+
+	mlx5_cmdif_debugfs_cleanup(dev);
+	debugfs_remove_recursive(dbg->dbg_root);
+}
+
+static int create_debugfs_files(struct mlx5_core_dev *dev)
+{
+	struct mlx5_cmd_debug *dbg = &dev->cmd.dbg;
+	int err = -ENOMEM;
+
+	if (!mlx5_debugfs_root)
+		return 0;
+
+	dbg->dbg_root = debugfs_create_dir("cmd", dev->priv.dbg_root);
+	if (!dbg->dbg_root)
+		return err;
+
+	dbg->dbg_in = debugfs_create_file("in", 0400, dbg->dbg_root,
+					  dev, &dfops);
+	if (!dbg->dbg_in)
+		goto err_dbg;
+
+	dbg->dbg_out = debugfs_create_file("out", 0200, dbg->dbg_root,
+					   dev, &dfops);
+	if (!dbg->dbg_out)
+		goto err_dbg;
+
+	dbg->dbg_outlen = debugfs_create_file("out_len", 0600, dbg->dbg_root,
+					      dev, &olfops);
+	if (!dbg->dbg_outlen)
+		goto err_dbg;
+
+	dbg->dbg_status = debugfs_create_u8("status", 0600, dbg->dbg_root,
+					    &dbg->status);
+	if (!dbg->dbg_status)
+		goto err_dbg;
+
+	dbg->dbg_run = debugfs_create_file("run", 0200, dbg->dbg_root, dev, &fops);
+	if (!dbg->dbg_run)
+		goto err_dbg;
+
+	mlx5_cmdif_debugfs_init(dev);
+
+	return 0;
+
+err_dbg:
+	clean_debug_files(dev);
+	return err;
+}
+
+static void mlx5_cmd_change_mod(struct mlx5_core_dev *dev, int mode)
+{
+	struct mlx5_cmd *cmd = &dev->cmd;
+	int i;
+
+	for (i = 0; i < cmd->max_reg_cmds; i++)
+		down(&cmd->sem);
+	down(&cmd->pages_sem);
+
+	cmd->mode = mode;
+
+	up(&cmd->pages_sem);
+	for (i = 0; i < cmd->max_reg_cmds; i++)
+		up(&cmd->sem);
+}
+
+void mlx5_cmd_use_events(struct mlx5_core_dev *dev)
+{
+	mlx5_cmd_change_mod(dev, CMD_MODE_EVENTS);
+}
+
+void mlx5_cmd_use_polling(struct mlx5_core_dev *dev)
+{
+	mlx5_cmd_change_mod(dev, CMD_MODE_POLLING);
+}
+
+static void free_msg(struct mlx5_core_dev *dev, struct mlx5_cmd_msg *msg)
+{
+	unsigned long flags;
+
+	if (msg->parent) {
+		spin_lock_irqsave(&msg->parent->lock, flags);
+		list_add_tail(&msg->list, &msg->parent->head);
+		spin_unlock_irqrestore(&msg->parent->lock, flags);
+	} else {
+		mlx5_free_cmd_msg(dev, msg);
+	}
+}
+
+void mlx5_cmd_comp_handler(struct mlx5_core_dev *dev, u64 vec, bool forced)
+{
+	struct mlx5_cmd *cmd = &dev->cmd;
+	struct mlx5_cmd_work_ent *ent;
+	mlx5_cmd_cbk_t callback;
+	void *context;
+	int err;
+	int i;
+	s64 ds;
+	struct mlx5_cmd_stats *stats;
+	unsigned long flags;
+	unsigned long vector;
+
+	/* there can be at most 32 command queues */
+	vector = vec & 0xffffffff;
+	for (i = 0; i < (1 << cmd->log_sz); i++) {
+		if (test_bit(i, &vector)) {
+			struct semaphore *sem;
+
+			ent = cmd->ent_arr[i];
+
+			/* if we already completed the command, ignore it */
+			if (!test_and_clear_bit(MLX5_CMD_ENT_STATE_PENDING_COMP,
+						&ent->state)) {
+				/* only real completion can free the cmd slot */
+				if (!forced) {
+					mlx5_core_err(dev, "Command completion arrived after timeout (entry idx = %d).\n",
+						      ent->idx);
+					free_ent(cmd, ent->idx);
+					free_cmd(ent);
+				}
+				continue;
+			}
+
+			if (ent->callback)
+				cancel_delayed_work(&ent->cb_timeout_work);
+			if (ent->page_queue)
+				sem = &cmd->pages_sem;
+			else
+				sem = &cmd->sem;
+			ent->ts2 = ktime_get_ns();
+			memcpy(ent->out->first.data, ent->lay->out, sizeof(ent->lay->out));
+			dump_command(dev, ent, 0);
+			if (!ent->ret) {
+				if (!cmd->checksum_disabled)
+					ent->ret = verify_signature(ent);
+				else
+					ent->ret = 0;
+				if (vec & MLX5_TRIGGERED_CMD_COMP)
+					ent->status = MLX5_DRIVER_STATUS_ABORTED;
+				else
+					ent->status = ent->lay->status_own >> 1;
+
+				mlx5_core_dbg(dev, "command completed. ret 0x%x, delivery status %s(0x%x)\n",
+					      ent->ret, deliv_status_to_str(ent->status), ent->status);
+			}
+
+			/* only real completion will free the entry slot */
+			if (!forced)
+				free_ent(cmd, ent->idx);
+
+			if (ent->callback) {
+				ds = ent->ts2 - ent->ts1;
+				if (ent->op < ARRAY_SIZE(cmd->stats)) {
+					stats = &cmd->stats[ent->op];
+					spin_lock_irqsave(&stats->lock, flags);
+					stats->sum += ds;
+					++stats->n;
+					spin_unlock_irqrestore(&stats->lock, flags);
+				}
+
+				callback = ent->callback;
+				context = ent->context;
+				err = ent->ret;
+				if (!err) {
+					err = mlx5_copy_from_msg(ent->uout,
+								 ent->out,
+								 ent->uout_size);
+
+					err = err ? err : mlx5_cmd_check(dev,
+									ent->in->first.data,
+									ent->uout);
+				}
+
+				mlx5_free_cmd_msg(dev, ent->out);
+				free_msg(dev, ent->in);
+
+				err = err ? err : ent->status;
+				if (!forced)
+					free_cmd(ent);
+				callback(err, context);
+			} else {
+				complete(&ent->done);
+			}
+			up(sem);
+		}
+	}
+}
+EXPORT_SYMBOL(mlx5_cmd_comp_handler);
+
+static int status_to_err(u8 status)
+{
+	return status ? -1 : 0; /* TBD more meaningful codes */
+}
+
+static struct mlx5_cmd_msg *alloc_msg(struct mlx5_core_dev *dev, int in_size,
+				      gfp_t gfp)
+{
+	struct mlx5_cmd_msg *msg = ERR_PTR(-ENOMEM);
+	struct cmd_msg_cache *ch = NULL;
+	struct mlx5_cmd *cmd = &dev->cmd;
+	int i;
+
+	if (in_size <= 16)
+		goto cache_miss;
+
+	for (i = 0; i < MLX5_NUM_COMMAND_CACHES; i++) {
+		ch = &cmd->cache[i];
+		if (in_size > ch->max_inbox_size)
+			continue;
+		spin_lock_irq(&ch->lock);
+		if (list_empty(&ch->head)) {
+			spin_unlock_irq(&ch->lock);
+			continue;
+		}
+		msg = list_entry(ch->head.next, typeof(*msg), list);
+		/* For cached lists, we must explicitly state what is
+		 * the real size
+		 */
+		msg->len = in_size;
+		list_del(&msg->list);
+		spin_unlock_irq(&ch->lock);
+		break;
+	}
+
+	if (!IS_ERR(msg))
+		return msg;
+
+cache_miss:
+	msg = mlx5_alloc_cmd_msg(dev, gfp, in_size, 0);
+	return msg;
+}
+
+static int is_manage_pages(void *in)
+{
+	return MLX5_GET(mbox_in, in, opcode) == MLX5_CMD_OP_MANAGE_PAGES;
+}
+
+static int cmd_exec(struct mlx5_core_dev *dev, void *in, int in_size, void *out,
+		    int out_size, mlx5_cmd_cbk_t callback, void *context,
+		    bool force_polling)
+{
+	struct mlx5_cmd_msg *inb;
+	struct mlx5_cmd_msg *outb;
+	int pages_queue;
+	gfp_t gfp;
+	int err;
+	u8 status = 0;
+	u32 drv_synd;
+	u8 token;
+
+	if (pci_channel_offline(dev->pdev) ||
+	    dev->state == MLX5_DEVICE_STATE_INTERNAL_ERROR) {
+		u16 opcode = MLX5_GET(mbox_in, in, opcode);
+
+		err = mlx5_internal_err_ret_value(dev, opcode, &drv_synd, &status);
+		MLX5_SET(mbox_out, out, status, status);
+		MLX5_SET(mbox_out, out, syndrome, drv_synd);
+		return err;
+	}
+
+	pages_queue = is_manage_pages(in);
+	gfp = callback ? GFP_ATOMIC : GFP_KERNEL;
+
+	inb = alloc_msg(dev, in_size, gfp);
+	if (IS_ERR(inb)) {
+		err = PTR_ERR(inb);
+		return err;
+	}
+
+	token = alloc_token(&dev->cmd);
+
+	err = mlx5_copy_to_msg(inb, in, in_size, token);
+	if (err) {
+		mlx5_core_warn(dev, "err %d\n", err);
+		goto out_in;
+	}
+
+	outb = mlx5_alloc_cmd_msg(dev, gfp, out_size, token);
+	if (IS_ERR(outb)) {
+		err = PTR_ERR(outb);
+		goto out_in;
+	}
+
+	err = mlx5_cmd_invoke(dev, inb, outb, out, out_size, callback, context,
+			      pages_queue, &status, token, force_polling);
+	if (err)
+		goto out_out;
+
+	mlx5_core_dbg(dev, "err %d, status %d\n", err, status);
+	if (status) {
+		err = status_to_err(status);
+		goto out_out;
+	}
+
+	if (!callback)
+		err = mlx5_copy_from_msg(out, outb, out_size);
+
+out_out:
+	if (!callback)
+		mlx5_free_cmd_msg(dev, outb);
+
+out_in:
+	if (!callback)
+		free_msg(dev, inb);
+	return err;
+}
+
+int mlx5_cmd_exec(struct mlx5_core_dev *dev, void *in, int in_size, void *out,
+		  int out_size)
+{
+	int err;
+
+	err = cmd_exec(dev, in, in_size, out, out_size, NULL, NULL, false);
+	return err ? : mlx5_cmd_check(dev, in, out);
+}
+EXPORT_SYMBOL(mlx5_cmd_exec);
+
+int mlx5_cmd_exec_cb(struct mlx5_core_dev *dev, void *in, int in_size,
+		     void *out, int out_size, mlx5_cmd_cbk_t callback,
+		     void *context)
+{
+	return cmd_exec(dev, in, in_size, out, out_size, callback, context,
+			false);
+}
+EXPORT_SYMBOL(mlx5_cmd_exec_cb);
+
+int mlx5_cmd_exec_polling(struct mlx5_core_dev *dev, void *in, int in_size,
+			  void *out, int out_size)
+{
+	int err;
+
+	err = cmd_exec(dev, in, in_size, out, out_size, NULL, NULL, true);
+
+	return err ? : mlx5_cmd_check(dev, in, out);
+}
+EXPORT_SYMBOL(mlx5_cmd_exec_polling);
+
+static void destroy_msg_cache(struct mlx5_core_dev *dev)
+{
+	struct cmd_msg_cache *ch;
+	struct mlx5_cmd_msg *msg;
+	struct mlx5_cmd_msg *n;
+	int i;
+
+	for (i = 0; i < MLX5_NUM_COMMAND_CACHES; i++) {
+		ch = &dev->cmd.cache[i];
+		list_for_each_entry_safe(msg, n, &ch->head, list) {
+			list_del(&msg->list);
+			mlx5_free_cmd_msg(dev, msg);
+		}
+	}
+}
+
+static unsigned cmd_cache_num_ent[MLX5_NUM_COMMAND_CACHES] = {
+	512, 32, 16, 8, 2
+};
+
+static unsigned cmd_cache_ent_size[MLX5_NUM_COMMAND_CACHES] = {
+	16 + MLX5_CMD_DATA_BLOCK_SIZE,
+	16 + MLX5_CMD_DATA_BLOCK_SIZE * 2,
+	16 + MLX5_CMD_DATA_BLOCK_SIZE * 16,
+	16 + MLX5_CMD_DATA_BLOCK_SIZE * 256,
+	16 + MLX5_CMD_DATA_BLOCK_SIZE * 512,
+};
+
+static void create_msg_cache(struct mlx5_core_dev *dev)
+{
+	struct mlx5_cmd *cmd = &dev->cmd;
+	struct cmd_msg_cache *ch;
+	struct mlx5_cmd_msg *msg;
+	int i;
+	int k;
+
+	/* Initialize and fill the caches with initial entries */
+	for (k = 0; k < MLX5_NUM_COMMAND_CACHES; k++) {
+		ch = &cmd->cache[k];
+		spin_lock_init(&ch->lock);
+		INIT_LIST_HEAD(&ch->head);
+		ch->num_ent = cmd_cache_num_ent[k];
+		ch->max_inbox_size = cmd_cache_ent_size[k];
+		for (i = 0; i < ch->num_ent; i++) {
+			msg = mlx5_alloc_cmd_msg(dev, GFP_KERNEL | __GFP_NOWARN,
+						 ch->max_inbox_size, 0);
+			if (IS_ERR(msg))
+				break;
+			msg->parent = ch;
+			list_add_tail(&msg->list, &ch->head);
+		}
+	}
+}
+
+static int alloc_cmd_page(struct mlx5_core_dev *dev, struct mlx5_cmd *cmd)
+{
+	struct device *ddev = &dev->pdev->dev;
+
+	cmd->cmd_alloc_buf = dma_zalloc_coherent(ddev, MLX5_ADAPTER_PAGE_SIZE,
+						 &cmd->alloc_dma, GFP_KERNEL);
+	if (!cmd->cmd_alloc_buf)
+		return -ENOMEM;
+
+	/* make sure it is aligned to 4K */
+	if (!((uintptr_t)cmd->cmd_alloc_buf & (MLX5_ADAPTER_PAGE_SIZE - 1))) {
+		cmd->cmd_buf = cmd->cmd_alloc_buf;
+		cmd->dma = cmd->alloc_dma;
+		cmd->alloc_size = MLX5_ADAPTER_PAGE_SIZE;
+		return 0;
+	}
+
+	dma_free_coherent(ddev, MLX5_ADAPTER_PAGE_SIZE, cmd->cmd_alloc_buf,
+			  cmd->alloc_dma);
+	cmd->cmd_alloc_buf = dma_zalloc_coherent(ddev,
+						 2 * MLX5_ADAPTER_PAGE_SIZE - 1,
+						 &cmd->alloc_dma, GFP_KERNEL);
+	if (!cmd->cmd_alloc_buf)
+		return -ENOMEM;
+
+	cmd->cmd_buf = PTR_ALIGN(cmd->cmd_alloc_buf, MLX5_ADAPTER_PAGE_SIZE);
+	cmd->dma = ALIGN(cmd->alloc_dma, MLX5_ADAPTER_PAGE_SIZE);
+	cmd->alloc_size = 2 * MLX5_ADAPTER_PAGE_SIZE - 1;
+	return 0;
+}
+
+static void free_cmd_page(struct mlx5_core_dev *dev, struct mlx5_cmd *cmd)
+{
+	struct device *ddev = &dev->pdev->dev;
+
+	dma_free_coherent(ddev, cmd->alloc_size, cmd->cmd_alloc_buf,
+			  cmd->alloc_dma);
+}
+
+int mlx5_cmd_init(struct mlx5_core_dev *dev)
+{
+	int size = sizeof(struct mlx5_cmd_prot_block);
+	int align = roundup_pow_of_two(size);
+	struct mlx5_cmd *cmd = &dev->cmd;
+	u32 cmd_h, cmd_l;
+	u16 cmd_if_rev;
+	int err;
+	int i;
+
+	memset(cmd, 0, sizeof(*cmd));
+	cmd_if_rev = cmdif_rev(dev);
+	if (cmd_if_rev != CMD_IF_REV) {
+		dev_err(&dev->pdev->dev,
+			"Driver cmdif rev(%d) differs from firmware's(%d)\n",
+			CMD_IF_REV, cmd_if_rev);
+		return -EINVAL;
+	}
+
+	cmd->pool = dma_pool_create("mlx5_cmd", &dev->pdev->dev, size, align,
+				    0);
+	if (!cmd->pool)
+		return -ENOMEM;
+
+	err = alloc_cmd_page(dev, cmd);
+	if (err)
+		goto err_free_pool;
+
+	cmd_l = ioread32be(&dev->iseg->cmdq_addr_l_sz) & 0xff;
+	cmd->log_sz = cmd_l >> 4 & 0xf;
+	cmd->log_stride = cmd_l & 0xf;
+	if (1 << cmd->log_sz > MLX5_MAX_COMMANDS) {
+		dev_err(&dev->pdev->dev, "firmware reports too many outstanding commands %d\n",
+			1 << cmd->log_sz);
+		err = -EINVAL;
+		goto err_free_page;
+	}
+
+	if (cmd->log_sz + cmd->log_stride > MLX5_ADAPTER_PAGE_SHIFT) {
+		dev_err(&dev->pdev->dev, "command queue size overflow\n");
+		err = -EINVAL;
+		goto err_free_page;
+	}
+
+	cmd->checksum_disabled = 1;
+	cmd->max_reg_cmds = (1 << cmd->log_sz) - 1;
+	cmd->bitmask = (1UL << cmd->max_reg_cmds) - 1;
+
+	cmd->cmdif_rev = ioread32be(&dev->iseg->cmdif_rev_fw_sub) >> 16;
+	if (cmd->cmdif_rev > CMD_IF_REV) {
+		dev_err(&dev->pdev->dev, "driver does not support command interface version. driver %d, firmware %d\n",
+			CMD_IF_REV, cmd->cmdif_rev);
+		err = -EOPNOTSUPP;
+		goto err_free_page;
+	}
+
+	spin_lock_init(&cmd->alloc_lock);
+	spin_lock_init(&cmd->token_lock);
+	for (i = 0; i < ARRAY_SIZE(cmd->stats); i++)
+		spin_lock_init(&cmd->stats[i].lock);
+
+	sema_init(&cmd->sem, cmd->max_reg_cmds);
+	sema_init(&cmd->pages_sem, 1);
+
+	cmd_h = (u32)((u64)(cmd->dma) >> 32);
+	cmd_l = (u32)(cmd->dma);
+	if (cmd_l & 0xfff) {
+		dev_err(&dev->pdev->dev, "invalid command queue address\n");
+		err = -ENOMEM;
+		goto err_free_page;
+	}
+
+	iowrite32be(cmd_h, &dev->iseg->cmdq_addr_h);
+	iowrite32be(cmd_l, &dev->iseg->cmdq_addr_l_sz);
+
+	/* Make sure firmware sees the complete address before we proceed */
+	wmb();
+
+	mlx5_core_dbg(dev, "descriptor at dma 0x%llx\n", (unsigned long long)(cmd->dma));
+
+	cmd->mode = CMD_MODE_POLLING;
+
+	create_msg_cache(dev);
+
+	set_wqname(dev);
+	cmd->wq = create_singlethread_workqueue(cmd->wq_name);
+	if (!cmd->wq) {
+		dev_err(&dev->pdev->dev, "failed to create command workqueue\n");
+		err = -ENOMEM;
+		goto err_cache;
+	}
+
+	err = create_debugfs_files(dev);
+	if (err) {
+		err = -ENOMEM;
+		goto err_wq;
+	}
+
+	return 0;
+
+err_wq:
+	destroy_workqueue(cmd->wq);
+
+err_cache:
+	destroy_msg_cache(dev);
+
+err_free_page:
+	free_cmd_page(dev, cmd);
+
+err_free_pool:
+	dma_pool_destroy(cmd->pool);
+
+	return err;
+}
+EXPORT_SYMBOL(mlx5_cmd_init);
+
+void mlx5_cmd_cleanup(struct mlx5_core_dev *dev)
+{
+	struct mlx5_cmd *cmd = &dev->cmd;
+
+	clean_debug_files(dev);
+	destroy_workqueue(cmd->wq);
+	destroy_msg_cache(dev);
+	free_cmd_page(dev, cmd);
+	dma_pool_destroy(cmd->pool);
+}
+EXPORT_SYMBOL(mlx5_cmd_cleanup);
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/cq.c b/drivers/net/ethernet/mellanox/mlx5/core/cq.c
new file mode 100644
index 000000000..a4179122a
--- /dev/null
+++ b/drivers/net/ethernet/mellanox/mlx5/core/cq.c
@@ -0,0 +1,221 @@
+/*
+ * Copyright (c) 2013-2015, 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/module.h>
+#include <linux/hardirq.h>
+#include <linux/mlx5/driver.h>
+#include <linux/mlx5/cmd.h>
+#include <rdma/ib_verbs.h>
+#include <linux/mlx5/cq.h>
+#include "mlx5_core.h"
+
+#define TASKLET_MAX_TIME 2
+#define TASKLET_MAX_TIME_JIFFIES msecs_to_jiffies(TASKLET_MAX_TIME)
+
+void mlx5_cq_tasklet_cb(unsigned long data)
+{
+	unsigned long flags;
+	unsigned long end = jiffies + TASKLET_MAX_TIME_JIFFIES;
+	struct mlx5_eq_tasklet *ctx = (struct mlx5_eq_tasklet *)data;
+	struct mlx5_core_cq *mcq;
+	struct mlx5_core_cq *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);
+		mlx5_cq_put(mcq);
+		if (time_after(jiffies, end))
+			break;
+	}
+
+	if (!list_empty(&ctx->process_list))
+		tasklet_schedule(&ctx->task);
+}
+
+static void mlx5_add_cq_to_tasklet(struct mlx5_core_cq *cq)
+{
+	unsigned long flags;
+	struct mlx5_eq_tasklet *tasklet_ctx = cq->tasklet_ctx.priv;
+
+	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)) {
+		mlx5_cq_hold(cq);
+		list_add_tail(&cq->tasklet_ctx.list, &tasklet_ctx->list);
+	}
+	spin_unlock_irqrestore(&tasklet_ctx->lock, flags);
+}
+
+int mlx5_core_create_cq(struct mlx5_core_dev *dev, struct mlx5_core_cq *cq,
+			u32 *in, int inlen)
+{
+	int eqn = MLX5_GET(cqc, MLX5_ADDR_OF(create_cq_in, in, cq_context), c_eqn);
+	u32 dout[MLX5_ST_SZ_DW(destroy_cq_out)];
+	u32 out[MLX5_ST_SZ_DW(create_cq_out)];
+	u32 din[MLX5_ST_SZ_DW(destroy_cq_in)];
+	struct mlx5_eq *eq;
+	int err;
+
+	eq = mlx5_eqn2eq(dev, eqn);
+	if (IS_ERR(eq))
+		return PTR_ERR(eq);
+
+	memset(out, 0, sizeof(out));
+	MLX5_SET(create_cq_in, in, opcode, MLX5_CMD_OP_CREATE_CQ);
+	err = mlx5_cmd_exec(dev, in, inlen, out, sizeof(out));
+	if (err)
+		return err;
+
+	cq->cqn = MLX5_GET(create_cq_out, out, cqn);
+	cq->cons_index = 0;
+	cq->arm_sn     = 0;
+	cq->eq         = eq;
+	refcount_set(&cq->refcount, 1);
+	init_completion(&cq->free);
+	if (!cq->comp)
+		cq->comp = mlx5_add_cq_to_tasklet;
+	/* assuming CQ will be deleted before the EQ */
+	cq->tasklet_ctx.priv = &eq->tasklet_ctx;
+	INIT_LIST_HEAD(&cq->tasklet_ctx.list);
+
+	/* Add to comp EQ CQ tree to recv comp events */
+	err = mlx5_eq_add_cq(eq, cq);
+	if (err)
+		goto err_cmd;
+
+	/* Add to async EQ CQ tree to recv async events */
+	err = mlx5_eq_add_cq(&dev->priv.eq_table.async_eq, cq);
+	if (err)
+		goto err_cq_add;
+
+	cq->pid = current->pid;
+	err = mlx5_debug_cq_add(dev, cq);
+	if (err)
+		mlx5_core_dbg(dev, "failed adding CP 0x%x to debug file system\n",
+			      cq->cqn);
+
+	cq->uar = dev->priv.uar;
+
+	return 0;
+
+err_cq_add:
+	mlx5_eq_del_cq(eq, cq);
+err_cmd:
+	memset(din, 0, sizeof(din));
+	memset(dout, 0, sizeof(dout));
+	MLX5_SET(destroy_cq_in, din, opcode, MLX5_CMD_OP_DESTROY_CQ);
+	MLX5_SET(destroy_cq_in, din, cqn, cq->cqn);
+	mlx5_cmd_exec(dev, din, sizeof(din), dout, sizeof(dout));
+	return err;
+}
+EXPORT_SYMBOL(mlx5_core_create_cq);
+
+int mlx5_core_destroy_cq(struct mlx5_core_dev *dev, struct mlx5_core_cq *cq)
+{
+	u32 out[MLX5_ST_SZ_DW(destroy_cq_out)] = {0};
+	u32 in[MLX5_ST_SZ_DW(destroy_cq_in)] = {0};
+	int err;
+
+	err = mlx5_eq_del_cq(&dev->priv.eq_table.async_eq, cq);
+	if (err)
+		return err;
+
+	err = mlx5_eq_del_cq(cq->eq, cq);
+	if (err)
+		return err;
+
+	MLX5_SET(destroy_cq_in, in, opcode, MLX5_CMD_OP_DESTROY_CQ);
+	MLX5_SET(destroy_cq_in, in, cqn, cq->cqn);
+	err = mlx5_cmd_exec(dev, in, sizeof(in), out, sizeof(out));
+	if (err)
+		return err;
+
+	synchronize_irq(cq->irqn);
+
+	mlx5_debug_cq_remove(dev, cq);
+	mlx5_cq_put(cq);
+	wait_for_completion(&cq->free);
+
+	return 0;
+}
+EXPORT_SYMBOL(mlx5_core_destroy_cq);
+
+int mlx5_core_query_cq(struct mlx5_core_dev *dev, struct mlx5_core_cq *cq,
+		       u32 *out, int outlen)
+{
+	u32 in[MLX5_ST_SZ_DW(query_cq_in)] = {0};
+
+	MLX5_SET(query_cq_in, in, opcode, MLX5_CMD_OP_QUERY_CQ);
+	MLX5_SET(query_cq_in, in, cqn, cq->cqn);
+	return mlx5_cmd_exec(dev, in, sizeof(in), out, outlen);
+}
+EXPORT_SYMBOL(mlx5_core_query_cq);
+
+int mlx5_core_modify_cq(struct mlx5_core_dev *dev, struct mlx5_core_cq *cq,
+			u32 *in, int inlen)
+{
+	u32 out[MLX5_ST_SZ_DW(modify_cq_out)] = {0};
+
+	MLX5_SET(modify_cq_in, in, opcode, MLX5_CMD_OP_MODIFY_CQ);
+	return mlx5_cmd_exec(dev, in, inlen, out, sizeof(out));
+}
+EXPORT_SYMBOL(mlx5_core_modify_cq);
+
+int mlx5_core_modify_cq_moderation(struct mlx5_core_dev *dev,
+				   struct mlx5_core_cq *cq,
+				   u16 cq_period,
+				   u16 cq_max_count)
+{
+	u32 in[MLX5_ST_SZ_DW(modify_cq_in)] = {0};
+	void *cqc;
+
+	MLX5_SET(modify_cq_in, in, cqn, cq->cqn);
+	cqc = MLX5_ADDR_OF(modify_cq_in, in, cq_context);
+	MLX5_SET(cqc, cqc, cq_period, cq_period);
+	MLX5_SET(cqc, cqc, cq_max_count, cq_max_count);
+	MLX5_SET(modify_cq_in, in,
+		 modify_field_select_resize_field_select.modify_field_select.modify_field_select,
+		 MLX5_CQ_MODIFY_PERIOD | MLX5_CQ_MODIFY_COUNT);
+
+	return mlx5_core_modify_cq(dev, cq, in, sizeof(in));
+}
+EXPORT_SYMBOL(mlx5_core_modify_cq_moderation);
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/debugfs.c b/drivers/net/ethernet/mellanox/mlx5/core/debugfs.c
new file mode 100644
index 000000000..90fabd612
--- /dev/null
+++ b/drivers/net/ethernet/mellanox/mlx5/core/debugfs.c
@@ -0,0 +1,593 @@
+/*
+ * Copyright (c) 2013-2015, 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/module.h>
+#include <linux/debugfs.h>
+#include <linux/mlx5/qp.h>
+#include <linux/mlx5/cq.h>
+#include <linux/mlx5/driver.h>
+#include "mlx5_core.h"
+
+enum {
+	QP_PID,
+	QP_STATE,
+	QP_XPORT,
+	QP_MTU,
+	QP_N_RECV,
+	QP_RECV_SZ,
+	QP_N_SEND,
+	QP_LOG_PG_SZ,
+	QP_RQPN,
+};
+
+static char *qp_fields[] = {
+	[QP_PID]	= "pid",
+	[QP_STATE]	= "state",
+	[QP_XPORT]	= "transport",
+	[QP_MTU]	= "mtu",
+	[QP_N_RECV]	= "num_recv",
+	[QP_RECV_SZ]	= "rcv_wqe_sz",
+	[QP_N_SEND]	= "num_send",
+	[QP_LOG_PG_SZ]	= "log2_page_sz",
+	[QP_RQPN]	= "remote_qpn",
+};
+
+enum {
+	EQ_NUM_EQES,
+	EQ_INTR,
+	EQ_LOG_PG_SZ,
+};
+
+static char *eq_fields[] = {
+	[EQ_NUM_EQES]	= "num_eqes",
+	[EQ_INTR]	= "intr",
+	[EQ_LOG_PG_SZ]	= "log_page_size",
+};
+
+enum {
+	CQ_PID,
+	CQ_NUM_CQES,
+	CQ_LOG_PG_SZ,
+};
+
+static char *cq_fields[] = {
+	[CQ_PID]	= "pid",
+	[CQ_NUM_CQES]	= "num_cqes",
+	[CQ_LOG_PG_SZ]	= "log_page_size",
+};
+
+struct dentry *mlx5_debugfs_root;
+EXPORT_SYMBOL(mlx5_debugfs_root);
+
+void mlx5_register_debugfs(void)
+{
+	mlx5_debugfs_root = debugfs_create_dir("mlx5", NULL);
+	if (IS_ERR_OR_NULL(mlx5_debugfs_root))
+		mlx5_debugfs_root = NULL;
+}
+
+void mlx5_unregister_debugfs(void)
+{
+	debugfs_remove(mlx5_debugfs_root);
+}
+
+int mlx5_qp_debugfs_init(struct mlx5_core_dev *dev)
+{
+	if (!mlx5_debugfs_root)
+		return 0;
+
+	atomic_set(&dev->num_qps, 0);
+
+	dev->priv.qp_debugfs = debugfs_create_dir("QPs",  dev->priv.dbg_root);
+	if (!dev->priv.qp_debugfs)
+		return -ENOMEM;
+
+	return 0;
+}
+
+void mlx5_qp_debugfs_cleanup(struct mlx5_core_dev *dev)
+{
+	if (!mlx5_debugfs_root)
+		return;
+
+	debugfs_remove_recursive(dev->priv.qp_debugfs);
+}
+
+int mlx5_eq_debugfs_init(struct mlx5_core_dev *dev)
+{
+	if (!mlx5_debugfs_root)
+		return 0;
+
+	dev->priv.eq_debugfs = debugfs_create_dir("EQs",  dev->priv.dbg_root);
+	if (!dev->priv.eq_debugfs)
+		return -ENOMEM;
+
+	return 0;
+}
+
+void mlx5_eq_debugfs_cleanup(struct mlx5_core_dev *dev)
+{
+	if (!mlx5_debugfs_root)
+		return;
+
+	debugfs_remove_recursive(dev->priv.eq_debugfs);
+}
+
+static ssize_t average_read(struct file *filp, char __user *buf, size_t count,
+			    loff_t *pos)
+{
+	struct mlx5_cmd_stats *stats;
+	u64 field = 0;
+	int ret;
+	char tbuf[22];
+
+	stats = filp->private_data;
+	spin_lock_irq(&stats->lock);
+	if (stats->n)
+		field = div64_u64(stats->sum, stats->n);
+	spin_unlock_irq(&stats->lock);
+	ret = snprintf(tbuf, sizeof(tbuf), "%llu\n", field);
+	return simple_read_from_buffer(buf, count, pos, tbuf, ret);
+}
+
+static ssize_t average_write(struct file *filp, const char __user *buf,
+			     size_t count, loff_t *pos)
+{
+	struct mlx5_cmd_stats *stats;
+
+	stats = filp->private_data;
+	spin_lock_irq(&stats->lock);
+	stats->sum = 0;
+	stats->n = 0;
+	spin_unlock_irq(&stats->lock);
+
+	*pos += count;
+
+	return count;
+}
+
+static const struct file_operations stats_fops = {
+	.owner	= THIS_MODULE,
+	.open	= simple_open,
+	.read	= average_read,
+	.write	= average_write,
+};
+
+int mlx5_cmdif_debugfs_init(struct mlx5_core_dev *dev)
+{
+	struct mlx5_cmd_stats *stats;
+	struct dentry **cmd;
+	const char *namep;
+	int err;
+	int i;
+
+	if (!mlx5_debugfs_root)
+		return 0;
+
+	cmd = &dev->priv.cmdif_debugfs;
+	*cmd = debugfs_create_dir("commands", dev->priv.dbg_root);
+	if (!*cmd)
+		return -ENOMEM;
+
+	for (i = 0; i < ARRAY_SIZE(dev->cmd.stats); i++) {
+		stats = &dev->cmd.stats[i];
+		namep = mlx5_command_str(i);
+		if (strcmp(namep, "unknown command opcode")) {
+			stats->root = debugfs_create_dir(namep, *cmd);
+			if (!stats->root) {
+				mlx5_core_warn(dev, "failed adding command %d\n",
+					       i);
+				err = -ENOMEM;
+				goto out;
+			}
+
+			stats->avg = debugfs_create_file("average", 0400,
+							 stats->root, stats,
+							 &stats_fops);
+			if (!stats->avg) {
+				mlx5_core_warn(dev, "failed creating debugfs file\n");
+				err = -ENOMEM;
+				goto out;
+			}
+
+			stats->count = debugfs_create_u64("n", 0400,
+							  stats->root,
+							  &stats->n);
+			if (!stats->count) {
+				mlx5_core_warn(dev, "failed creating debugfs file\n");
+				err = -ENOMEM;
+				goto out;
+			}
+		}
+	}
+
+	return 0;
+out:
+	debugfs_remove_recursive(dev->priv.cmdif_debugfs);
+	return err;
+}
+
+void mlx5_cmdif_debugfs_cleanup(struct mlx5_core_dev *dev)
+{
+	if (!mlx5_debugfs_root)
+		return;
+
+	debugfs_remove_recursive(dev->priv.cmdif_debugfs);
+}
+
+int mlx5_cq_debugfs_init(struct mlx5_core_dev *dev)
+{
+	if (!mlx5_debugfs_root)
+		return 0;
+
+	dev->priv.cq_debugfs = debugfs_create_dir("CQs",  dev->priv.dbg_root);
+	if (!dev->priv.cq_debugfs)
+		return -ENOMEM;
+
+	return 0;
+}
+
+void mlx5_cq_debugfs_cleanup(struct mlx5_core_dev *dev)
+{
+	if (!mlx5_debugfs_root)
+		return;
+
+	debugfs_remove_recursive(dev->priv.cq_debugfs);
+}
+
+static u64 qp_read_field(struct mlx5_core_dev *dev, struct mlx5_core_qp *qp,
+			 int index, int *is_str)
+{
+	int outlen = MLX5_ST_SZ_BYTES(query_qp_out);
+	struct mlx5_qp_context *ctx;
+	u64 param = 0;
+	u32 *out;
+	int err;
+	int no_sq;
+
+	out = kzalloc(outlen, GFP_KERNEL);
+	if (!out)
+		return param;
+
+	err = mlx5_core_qp_query(dev, qp, out, outlen);
+	if (err) {
+		mlx5_core_warn(dev, "failed to query qp err=%d\n", err);
+		goto out;
+	}
+
+	*is_str = 0;
+
+	/* FIXME: use MLX5_GET rather than mlx5_qp_context manual struct */
+	ctx = (struct mlx5_qp_context *)MLX5_ADDR_OF(query_qp_out, out, qpc);
+
+	switch (index) {
+	case QP_PID:
+		param = qp->pid;
+		break;
+	case QP_STATE:
+		param = (unsigned long)mlx5_qp_state_str(be32_to_cpu(ctx->flags) >> 28);
+		*is_str = 1;
+		break;
+	case QP_XPORT:
+		param = (unsigned long)mlx5_qp_type_str((be32_to_cpu(ctx->flags) >> 16) & 0xff);
+		*is_str = 1;
+		break;
+	case QP_MTU:
+		switch (ctx->mtu_msgmax >> 5) {
+		case IB_MTU_256:
+			param = 256;
+			break;
+		case IB_MTU_512:
+			param = 512;
+			break;
+		case IB_MTU_1024:
+			param = 1024;
+			break;
+		case IB_MTU_2048:
+			param = 2048;
+			break;
+		case IB_MTU_4096:
+			param = 4096;
+			break;
+		default:
+			param = 0;
+		}
+		break;
+	case QP_N_RECV:
+		param = 1 << ((ctx->rq_size_stride >> 3) & 0xf);
+		break;
+	case QP_RECV_SZ:
+		param = 1 << ((ctx->rq_size_stride & 7) + 4);
+		break;
+	case QP_N_SEND:
+		no_sq = be16_to_cpu(ctx->sq_crq_size) >> 15;
+		if (!no_sq)
+			param = 1 << (be16_to_cpu(ctx->sq_crq_size) >> 11);
+		else
+			param = 0;
+		break;
+	case QP_LOG_PG_SZ:
+		param = (be32_to_cpu(ctx->log_pg_sz_remote_qpn) >> 24) & 0x1f;
+		param += 12;
+		break;
+	case QP_RQPN:
+		param = be32_to_cpu(ctx->log_pg_sz_remote_qpn) & 0xffffff;
+		break;
+	}
+
+out:
+	kfree(out);
+	return param;
+}
+
+static u64 eq_read_field(struct mlx5_core_dev *dev, struct mlx5_eq *eq,
+			 int index)
+{
+	int outlen = MLX5_ST_SZ_BYTES(query_eq_out);
+	u64 param = 0;
+	void *ctx;
+	u32 *out;
+	int err;
+
+	out = kzalloc(outlen, GFP_KERNEL);
+	if (!out)
+		return param;
+
+	err = mlx5_core_eq_query(dev, eq, out, outlen);
+	if (err) {
+		mlx5_core_warn(dev, "failed to query eq\n");
+		goto out;
+	}
+	ctx = MLX5_ADDR_OF(query_eq_out, out, eq_context_entry);
+
+	switch (index) {
+	case EQ_NUM_EQES:
+		param = 1 << MLX5_GET(eqc, ctx, log_eq_size);
+		break;
+	case EQ_INTR:
+		param = MLX5_GET(eqc, ctx, intr);
+		break;
+	case EQ_LOG_PG_SZ:
+		param = MLX5_GET(eqc, ctx, log_page_size) + 12;
+		break;
+	}
+
+out:
+	kfree(out);
+	return param;
+}
+
+static u64 cq_read_field(struct mlx5_core_dev *dev, struct mlx5_core_cq *cq,
+			 int index)
+{
+	int outlen = MLX5_ST_SZ_BYTES(query_cq_out);
+	u64 param = 0;
+	void *ctx;
+	u32 *out;
+	int err;
+
+	out = kvzalloc(outlen, GFP_KERNEL);
+	if (!out)
+		return param;
+
+	err = mlx5_core_query_cq(dev, cq, out, outlen);
+	if (err) {
+		mlx5_core_warn(dev, "failed to query cq\n");
+		goto out;
+	}
+	ctx = MLX5_ADDR_OF(query_cq_out, out, cq_context);
+
+	switch (index) {
+	case CQ_PID:
+		param = cq->pid;
+		break;
+	case CQ_NUM_CQES:
+		param = 1 << MLX5_GET(cqc, ctx, log_cq_size);
+		break;
+	case CQ_LOG_PG_SZ:
+		param = MLX5_GET(cqc, ctx, log_page_size);
+		break;
+	}
+
+out:
+	kvfree(out);
+	return param;
+}
+
+static ssize_t dbg_read(struct file *filp, char __user *buf, size_t count,
+			loff_t *pos)
+{
+	struct mlx5_field_desc *desc;
+	struct mlx5_rsc_debug *d;
+	char tbuf[18];
+	int is_str = 0;
+	u64 field;
+	int ret;
+
+	desc = filp->private_data;
+	d = (void *)(desc - desc->i) - sizeof(*d);
+	switch (d->type) {
+	case MLX5_DBG_RSC_QP:
+		field = qp_read_field(d->dev, d->object, desc->i, &is_str);
+		break;
+
+	case MLX5_DBG_RSC_EQ:
+		field = eq_read_field(d->dev, d->object, desc->i);
+		break;
+
+	case MLX5_DBG_RSC_CQ:
+		field = cq_read_field(d->dev, d->object, desc->i);
+		break;
+
+	default:
+		mlx5_core_warn(d->dev, "invalid resource type %d\n", d->type);
+		return -EINVAL;
+	}
+
+	if (is_str)
+		ret = snprintf(tbuf, sizeof(tbuf), "%s\n", (const char *)(unsigned long)field);
+	else
+		ret = snprintf(tbuf, sizeof(tbuf), "0x%llx\n", field);
+
+	return simple_read_from_buffer(buf, count, pos, tbuf, ret);
+}
+
+static const struct file_operations fops = {
+	.owner	= THIS_MODULE,
+	.open	= simple_open,
+	.read	= dbg_read,
+};
+
+static int add_res_tree(struct mlx5_core_dev *dev, enum dbg_rsc_type type,
+			struct dentry *root, struct mlx5_rsc_debug **dbg,
+			int rsn, char **field, int nfile, void *data)
+{
+	struct mlx5_rsc_debug *d;
+	char resn[32];
+	int err;
+	int i;
+
+	d = kzalloc(struct_size(d, fields, nfile), GFP_KERNEL);
+	if (!d)
+		return -ENOMEM;
+
+	d->dev = dev;
+	d->object = data;
+	d->type = type;
+	sprintf(resn, "0x%x", rsn);
+	d->root = debugfs_create_dir(resn,  root);
+	if (!d->root) {
+		err = -ENOMEM;
+		goto out_free;
+	}
+
+	for (i = 0; i < nfile; i++) {
+		d->fields[i].i = i;
+		d->fields[i].dent = debugfs_create_file(field[i], 0400,
+							d->root, &d->fields[i],
+							&fops);
+		if (!d->fields[i].dent) {
+			err = -ENOMEM;
+			goto out_rem;
+		}
+	}
+	*dbg = d;
+
+	return 0;
+out_rem:
+	debugfs_remove_recursive(d->root);
+
+out_free:
+	kfree(d);
+	return err;
+}
+
+static void rem_res_tree(struct mlx5_rsc_debug *d)
+{
+	debugfs_remove_recursive(d->root);
+	kfree(d);
+}
+
+int mlx5_debug_qp_add(struct mlx5_core_dev *dev, struct mlx5_core_qp *qp)
+{
+	int err;
+
+	if (!mlx5_debugfs_root)
+		return 0;
+
+	err = add_res_tree(dev, MLX5_DBG_RSC_QP, dev->priv.qp_debugfs,
+			   &qp->dbg, qp->qpn, qp_fields,
+			   ARRAY_SIZE(qp_fields), qp);
+	if (err)
+		qp->dbg = NULL;
+
+	return err;
+}
+
+void mlx5_debug_qp_remove(struct mlx5_core_dev *dev, struct mlx5_core_qp *qp)
+{
+	if (!mlx5_debugfs_root)
+		return;
+
+	if (qp->dbg)
+		rem_res_tree(qp->dbg);
+}
+
+int mlx5_debug_eq_add(struct mlx5_core_dev *dev, struct mlx5_eq *eq)
+{
+	int err;
+
+	if (!mlx5_debugfs_root)
+		return 0;
+
+	err = add_res_tree(dev, MLX5_DBG_RSC_EQ, dev->priv.eq_debugfs,
+			   &eq->dbg, eq->eqn, eq_fields,
+			   ARRAY_SIZE(eq_fields), eq);
+	if (err)
+		eq->dbg = NULL;
+
+	return err;
+}
+
+void mlx5_debug_eq_remove(struct mlx5_core_dev *dev, struct mlx5_eq *eq)
+{
+	if (!mlx5_debugfs_root)
+		return;
+
+	if (eq->dbg)
+		rem_res_tree(eq->dbg);
+}
+
+int mlx5_debug_cq_add(struct mlx5_core_dev *dev, struct mlx5_core_cq *cq)
+{
+	int err;
+
+	if (!mlx5_debugfs_root)
+		return 0;
+
+	err = add_res_tree(dev, MLX5_DBG_RSC_CQ, dev->priv.cq_debugfs,
+			   &cq->dbg, cq->cqn, cq_fields,
+			   ARRAY_SIZE(cq_fields), cq);
+	if (err)
+		cq->dbg = NULL;
+
+	return err;
+}
+
+void mlx5_debug_cq_remove(struct mlx5_core_dev *dev, struct mlx5_core_cq *cq)
+{
+	if (!mlx5_debugfs_root)
+		return;
+
+	if (cq->dbg)
+		rem_res_tree(cq->dbg);
+}
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/dev.c b/drivers/net/ethernet/mellanox/mlx5/core/dev.c
new file mode 100644
index 000000000..3692d6a1c
--- /dev/null
+++ b/drivers/net/ethernet/mellanox/mlx5/core/dev.c
@@ -0,0 +1,498 @@
+/*
+ * Copyright (c) 2013-2015, 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/mlx5/driver.h>
+#include "mlx5_core.h"
+
+static LIST_HEAD(intf_list);
+static LIST_HEAD(mlx5_dev_list);
+/* intf dev list mutex */
+static DEFINE_MUTEX(mlx5_intf_mutex);
+
+struct mlx5_device_context {
+	struct list_head	list;
+	struct mlx5_interface  *intf;
+	void		       *context;
+	unsigned long		state;
+};
+
+struct mlx5_delayed_event {
+	struct list_head	list;
+	struct mlx5_core_dev	*dev;
+	enum mlx5_dev_event	event;
+	unsigned long		param;
+};
+
+enum {
+	MLX5_INTERFACE_ADDED,
+	MLX5_INTERFACE_ATTACHED,
+};
+
+static void add_delayed_event(struct mlx5_priv *priv,
+			      struct mlx5_core_dev *dev,
+			      enum mlx5_dev_event event,
+			      unsigned long param)
+{
+	struct mlx5_delayed_event *delayed_event;
+
+	delayed_event = kzalloc(sizeof(*delayed_event), GFP_ATOMIC);
+	if (!delayed_event) {
+		mlx5_core_err(dev, "event %d is missed\n", event);
+		return;
+	}
+
+	mlx5_core_dbg(dev, "Accumulating event %d\n", event);
+	delayed_event->dev = dev;
+	delayed_event->event = event;
+	delayed_event->param = param;
+	list_add_tail(&delayed_event->list, &priv->waiting_events_list);
+}
+
+static void delayed_event_release(struct mlx5_device_context *dev_ctx,
+				  struct mlx5_priv *priv)
+{
+	struct mlx5_core_dev *dev = container_of(priv, struct mlx5_core_dev, priv);
+	struct mlx5_delayed_event *de;
+	struct mlx5_delayed_event *n;
+	struct list_head temp;
+
+	INIT_LIST_HEAD(&temp);
+
+	spin_lock_irq(&priv->ctx_lock);
+
+	priv->is_accum_events = false;
+	list_splice_init(&priv->waiting_events_list, &temp);
+	if (!dev_ctx->context)
+		goto out;
+	list_for_each_entry_safe(de, n, &temp, list)
+		dev_ctx->intf->event(dev, dev_ctx->context, de->event, de->param);
+
+out:
+	spin_unlock_irq(&priv->ctx_lock);
+
+	list_for_each_entry_safe(de, n, &temp, list) {
+		list_del(&de->list);
+		kfree(de);
+	}
+}
+
+/* accumulating events that can come after mlx5_ib calls to
+ * ib_register_device, till adding that interface to the events list.
+ */
+static void delayed_event_start(struct mlx5_priv *priv)
+{
+	spin_lock_irq(&priv->ctx_lock);
+	priv->is_accum_events = true;
+	spin_unlock_irq(&priv->ctx_lock);
+}
+
+void mlx5_add_device(struct mlx5_interface *intf, struct mlx5_priv *priv)
+{
+	struct mlx5_device_context *dev_ctx;
+	struct mlx5_core_dev *dev = container_of(priv, struct mlx5_core_dev, priv);
+
+	if (!mlx5_lag_intf_add(intf, priv))
+		return;
+
+	dev_ctx = kzalloc(sizeof(*dev_ctx), GFP_KERNEL);
+	if (!dev_ctx)
+		return;
+
+	dev_ctx->intf = intf;
+
+	delayed_event_start(priv);
+
+	dev_ctx->context = intf->add(dev);
+	if (dev_ctx->context) {
+		set_bit(MLX5_INTERFACE_ADDED, &dev_ctx->state);
+		if (intf->attach)
+			set_bit(MLX5_INTERFACE_ATTACHED, &dev_ctx->state);
+
+		spin_lock_irq(&priv->ctx_lock);
+		list_add_tail(&dev_ctx->list, &priv->ctx_list);
+
+#ifdef CONFIG_INFINIBAND_ON_DEMAND_PAGING
+		if (dev_ctx->intf->pfault) {
+			if (priv->pfault) {
+				mlx5_core_err(dev, "multiple page fault handlers not supported");
+			} else {
+				priv->pfault_ctx = dev_ctx->context;
+				priv->pfault = dev_ctx->intf->pfault;
+			}
+		}
+#endif
+		spin_unlock_irq(&priv->ctx_lock);
+	}
+
+	delayed_event_release(dev_ctx, priv);
+
+	if (!dev_ctx->context)
+		kfree(dev_ctx);
+}
+
+static struct mlx5_device_context *mlx5_get_device(struct mlx5_interface *intf,
+						   struct mlx5_priv *priv)
+{
+	struct mlx5_device_context *dev_ctx;
+
+	list_for_each_entry(dev_ctx, &priv->ctx_list, list)
+		if (dev_ctx->intf == intf)
+			return dev_ctx;
+	return NULL;
+}
+
+void mlx5_remove_device(struct mlx5_interface *intf, struct mlx5_priv *priv)
+{
+	struct mlx5_device_context *dev_ctx;
+	struct mlx5_core_dev *dev = container_of(priv, struct mlx5_core_dev, priv);
+
+	dev_ctx = mlx5_get_device(intf, priv);
+	if (!dev_ctx)
+		return;
+
+#ifdef CONFIG_INFINIBAND_ON_DEMAND_PAGING
+	spin_lock_irq(&priv->ctx_lock);
+	if (priv->pfault == dev_ctx->intf->pfault)
+		priv->pfault = NULL;
+	spin_unlock_irq(&priv->ctx_lock);
+
+	synchronize_srcu(&priv->pfault_srcu);
+#endif
+
+	spin_lock_irq(&priv->ctx_lock);
+	list_del(&dev_ctx->list);
+	spin_unlock_irq(&priv->ctx_lock);
+
+	if (test_bit(MLX5_INTERFACE_ADDED, &dev_ctx->state))
+		intf->remove(dev, dev_ctx->context);
+
+	kfree(dev_ctx);
+}
+
+static void mlx5_attach_interface(struct mlx5_interface *intf, struct mlx5_priv *priv)
+{
+	struct mlx5_device_context *dev_ctx;
+	struct mlx5_core_dev *dev = container_of(priv, struct mlx5_core_dev, priv);
+
+	dev_ctx = mlx5_get_device(intf, priv);
+	if (!dev_ctx)
+		return;
+
+	delayed_event_start(priv);
+	if (intf->attach) {
+		if (test_bit(MLX5_INTERFACE_ATTACHED, &dev_ctx->state))
+			goto out;
+		if (intf->attach(dev, dev_ctx->context))
+			goto out;
+
+		set_bit(MLX5_INTERFACE_ATTACHED, &dev_ctx->state);
+	} else {
+		if (test_bit(MLX5_INTERFACE_ADDED, &dev_ctx->state))
+			goto out;
+		dev_ctx->context = intf->add(dev);
+		if (!dev_ctx->context)
+			goto out;
+
+		set_bit(MLX5_INTERFACE_ADDED, &dev_ctx->state);
+	}
+
+out:
+	delayed_event_release(dev_ctx, priv);
+}
+
+void mlx5_attach_device(struct mlx5_core_dev *dev)
+{
+	struct mlx5_priv *priv = &dev->priv;
+	struct mlx5_interface *intf;
+
+	mutex_lock(&mlx5_intf_mutex);
+	list_for_each_entry(intf, &intf_list, list)
+		mlx5_attach_interface(intf, priv);
+	mutex_unlock(&mlx5_intf_mutex);
+}
+
+static void mlx5_detach_interface(struct mlx5_interface *intf, struct mlx5_priv *priv)
+{
+	struct mlx5_device_context *dev_ctx;
+	struct mlx5_core_dev *dev = container_of(priv, struct mlx5_core_dev, priv);
+
+	dev_ctx = mlx5_get_device(intf, priv);
+	if (!dev_ctx)
+		return;
+
+	if (intf->detach) {
+		if (!test_bit(MLX5_INTERFACE_ATTACHED, &dev_ctx->state))
+			return;
+		intf->detach(dev, dev_ctx->context);
+		clear_bit(MLX5_INTERFACE_ATTACHED, &dev_ctx->state);
+	} else {
+		if (!test_bit(MLX5_INTERFACE_ADDED, &dev_ctx->state))
+			return;
+		intf->remove(dev, dev_ctx->context);
+		clear_bit(MLX5_INTERFACE_ADDED, &dev_ctx->state);
+	}
+}
+
+void mlx5_detach_device(struct mlx5_core_dev *dev)
+{
+	struct mlx5_priv *priv = &dev->priv;
+	struct mlx5_interface *intf;
+
+	mutex_lock(&mlx5_intf_mutex);
+	list_for_each_entry(intf, &intf_list, list)
+		mlx5_detach_interface(intf, priv);
+	mutex_unlock(&mlx5_intf_mutex);
+}
+
+bool mlx5_device_registered(struct mlx5_core_dev *dev)
+{
+	struct mlx5_priv *priv;
+	bool found = false;
+
+	mutex_lock(&mlx5_intf_mutex);
+	list_for_each_entry(priv, &mlx5_dev_list, dev_list)
+		if (priv == &dev->priv)
+			found = true;
+	mutex_unlock(&mlx5_intf_mutex);
+
+	return found;
+}
+
+int mlx5_register_device(struct mlx5_core_dev *dev)
+{
+	struct mlx5_priv *priv = &dev->priv;
+	struct mlx5_interface *intf;
+
+	mutex_lock(&mlx5_intf_mutex);
+	list_add_tail(&priv->dev_list, &mlx5_dev_list);
+	list_for_each_entry(intf, &intf_list, list)
+		mlx5_add_device(intf, priv);
+	mutex_unlock(&mlx5_intf_mutex);
+
+	return 0;
+}
+
+void mlx5_unregister_device(struct mlx5_core_dev *dev)
+{
+	struct mlx5_priv *priv = &dev->priv;
+	struct mlx5_interface *intf;
+
+	mutex_lock(&mlx5_intf_mutex);
+	list_for_each_entry_reverse(intf, &intf_list, list)
+		mlx5_remove_device(intf, priv);
+	list_del(&priv->dev_list);
+	mutex_unlock(&mlx5_intf_mutex);
+}
+
+int mlx5_register_interface(struct mlx5_interface *intf)
+{
+	struct mlx5_priv *priv;
+
+	if (!intf->add || !intf->remove)
+		return -EINVAL;
+
+	mutex_lock(&mlx5_intf_mutex);
+	list_add_tail(&intf->list, &intf_list);
+	list_for_each_entry(priv, &mlx5_dev_list, dev_list)
+		mlx5_add_device(intf, priv);
+	mutex_unlock(&mlx5_intf_mutex);
+
+	return 0;
+}
+EXPORT_SYMBOL(mlx5_register_interface);
+
+void mlx5_unregister_interface(struct mlx5_interface *intf)
+{
+	struct mlx5_priv *priv;
+
+	mutex_lock(&mlx5_intf_mutex);
+	list_for_each_entry(priv, &mlx5_dev_list, dev_list)
+		mlx5_remove_device(intf, priv);
+	list_del(&intf->list);
+	mutex_unlock(&mlx5_intf_mutex);
+}
+EXPORT_SYMBOL(mlx5_unregister_interface);
+
+/* Must be called with intf_mutex held */
+static bool mlx5_has_added_dev_by_protocol(struct mlx5_core_dev *mdev, int protocol)
+{
+	struct mlx5_device_context *dev_ctx;
+	struct mlx5_interface *intf;
+	bool found = false;
+
+	list_for_each_entry(intf, &intf_list, list) {
+		if (intf->protocol == protocol) {
+			dev_ctx = mlx5_get_device(intf, &mdev->priv);
+			if (dev_ctx && test_bit(MLX5_INTERFACE_ADDED, &dev_ctx->state))
+				found = true;
+			break;
+		}
+	}
+
+	return found;
+}
+
+void mlx5_reload_interface(struct mlx5_core_dev *mdev, int protocol)
+{
+	mutex_lock(&mlx5_intf_mutex);
+	if (mlx5_has_added_dev_by_protocol(mdev, protocol)) {
+		mlx5_remove_dev_by_protocol(mdev, protocol);
+		mlx5_add_dev_by_protocol(mdev, protocol);
+	}
+	mutex_unlock(&mlx5_intf_mutex);
+}
+
+void *mlx5_get_protocol_dev(struct mlx5_core_dev *mdev, int protocol)
+{
+	struct mlx5_priv *priv = &mdev->priv;
+	struct mlx5_device_context *dev_ctx;
+	unsigned long flags;
+	void *result = NULL;
+
+	spin_lock_irqsave(&priv->ctx_lock, flags);
+
+	list_for_each_entry(dev_ctx, &mdev->priv.ctx_list, list)
+		if ((dev_ctx->intf->protocol == protocol) &&
+		    dev_ctx->intf->get_dev) {
+			result = dev_ctx->intf->get_dev(dev_ctx->context);
+			break;
+		}
+
+	spin_unlock_irqrestore(&priv->ctx_lock, flags);
+
+	return result;
+}
+EXPORT_SYMBOL(mlx5_get_protocol_dev);
+
+/* Must be called with intf_mutex held */
+void mlx5_add_dev_by_protocol(struct mlx5_core_dev *dev, int protocol)
+{
+	struct mlx5_interface *intf;
+
+	list_for_each_entry(intf, &intf_list, list)
+		if (intf->protocol == protocol) {
+			mlx5_add_device(intf, &dev->priv);
+			break;
+		}
+}
+
+/* Must be called with intf_mutex held */
+void mlx5_remove_dev_by_protocol(struct mlx5_core_dev *dev, int protocol)
+{
+	struct mlx5_interface *intf;
+
+	list_for_each_entry(intf, &intf_list, list)
+		if (intf->protocol == protocol) {
+			mlx5_remove_device(intf, &dev->priv);
+			break;
+		}
+}
+
+static u32 mlx5_gen_pci_id(struct mlx5_core_dev *dev)
+{
+	return (u32)((pci_domain_nr(dev->pdev->bus) << 16) |
+		     (dev->pdev->bus->number << 8) |
+		     PCI_SLOT(dev->pdev->devfn));
+}
+
+/* Must be called with intf_mutex held */
+struct mlx5_core_dev *mlx5_get_next_phys_dev(struct mlx5_core_dev *dev)
+{
+	u32 pci_id = mlx5_gen_pci_id(dev);
+	struct mlx5_core_dev *res = NULL;
+	struct mlx5_core_dev *tmp_dev;
+	struct mlx5_priv *priv;
+
+	list_for_each_entry(priv, &mlx5_dev_list, dev_list) {
+		tmp_dev = container_of(priv, struct mlx5_core_dev, priv);
+		if ((dev != tmp_dev) && (mlx5_gen_pci_id(tmp_dev) == pci_id)) {
+			res = tmp_dev;
+			break;
+		}
+	}
+
+	return res;
+}
+
+void mlx5_core_event(struct mlx5_core_dev *dev, enum mlx5_dev_event event,
+		     unsigned long param)
+{
+	struct mlx5_priv *priv = &dev->priv;
+	struct mlx5_device_context *dev_ctx;
+	unsigned long flags;
+
+	spin_lock_irqsave(&priv->ctx_lock, flags);
+
+	if (priv->is_accum_events)
+		add_delayed_event(priv, dev, event, param);
+
+	/* After mlx5_detach_device, the dev_ctx->intf is still set and dev_ctx is
+	 * still in priv->ctx_list. In this case, only notify the dev_ctx if its
+	 * ADDED or ATTACHED bit are set.
+	 */
+	list_for_each_entry(dev_ctx, &priv->ctx_list, list)
+		if (dev_ctx->intf->event &&
+		    (test_bit(MLX5_INTERFACE_ADDED, &dev_ctx->state) ||
+		     test_bit(MLX5_INTERFACE_ATTACHED, &dev_ctx->state)))
+			dev_ctx->intf->event(dev, dev_ctx->context, event, param);
+
+	spin_unlock_irqrestore(&priv->ctx_lock, flags);
+}
+
+#ifdef CONFIG_INFINIBAND_ON_DEMAND_PAGING
+void mlx5_core_page_fault(struct mlx5_core_dev *dev,
+			  struct mlx5_pagefault *pfault)
+{
+	struct mlx5_priv *priv = &dev->priv;
+	int srcu_idx;
+
+	srcu_idx = srcu_read_lock(&priv->pfault_srcu);
+	if (priv->pfault)
+		priv->pfault(dev, priv->pfault_ctx, pfault);
+	srcu_read_unlock(&priv->pfault_srcu, srcu_idx);
+}
+#endif
+
+void mlx5_dev_list_lock(void)
+{
+	mutex_lock(&mlx5_intf_mutex);
+}
+
+void mlx5_dev_list_unlock(void)
+{
+	mutex_unlock(&mlx5_intf_mutex);
+}
+
+int mlx5_dev_list_trylock(void)
+{
+	return mutex_trylock(&mlx5_intf_mutex);
+}
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/diag/Makefile b/drivers/net/ethernet/mellanox/mlx5/core/diag/Makefile
new file mode 100644
index 000000000..d8e17110f
--- /dev/null
+++ b/drivers/net/ethernet/mellanox/mlx5/core/diag/Makefile
@@ -0,0 +1 @@
+subdir-ccflags-y += -I$(src)/..
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/diag/fs_tracepoint.c b/drivers/net/ethernet/mellanox/mlx5/core/diag/fs_tracepoint.c
new file mode 100644
index 000000000..0f11fff32
--- /dev/null
+++ b/drivers/net/ethernet/mellanox/mlx5/core/diag/fs_tracepoint.c
@@ -0,0 +1,267 @@
+/*
+ * Copyright (c) 2017, 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.
+ */
+
+#define CREATE_TRACE_POINTS
+
+#include "fs_tracepoint.h"
+#include <linux/stringify.h>
+
+#define DECLARE_MASK_VAL(type, name) struct {type m; type v; } name
+#define MASK_VAL(type, spec, name, mask, val, fld)	\
+		DECLARE_MASK_VAL(type, name) =		\
+			{.m = MLX5_GET(spec, mask, fld),\
+			 .v = MLX5_GET(spec, val, fld)}
+#define MASK_VAL_BE(type, spec, name, mask, val, fld)	\
+		    DECLARE_MASK_VAL(type, name) =	\
+			{.m = MLX5_GET_BE(type, spec, mask, fld),\
+			 .v = MLX5_GET_BE(type, spec, val, fld)}
+#define GET_MASKED_VAL(name) (name.m & name.v)
+
+#define GET_MASK_VAL(name, type, mask, val, fld)	\
+		(name.m = MLX5_GET(type, mask, fld),	\
+		 name.v = MLX5_GET(type, val, fld),	\
+		 name.m & name.v)
+#define PRINT_MASKED_VAL(name, p, format) {		\
+	if (name.m)			\
+		trace_seq_printf(p, __stringify(name) "=" format " ", name.v); \
+	}
+#define PRINT_MASKED_VALP(name, cast, p, format) {	\
+	if (name.m)			\
+		trace_seq_printf(p, __stringify(name) "=" format " ",	       \
+				 (cast)&name.v);\
+	}
+
+static void print_lyr_2_4_hdrs(struct trace_seq *p,
+			       const u32 *mask, const u32 *value)
+{
+#define MASK_VAL_L2(type, name, fld) \
+	MASK_VAL(type, fte_match_set_lyr_2_4, name, mask, value, fld)
+	DECLARE_MASK_VAL(u64, smac) = {
+		.m = MLX5_GET(fte_match_set_lyr_2_4, mask, smac_47_16) << 16 |
+		     MLX5_GET(fte_match_set_lyr_2_4, mask, smac_15_0),
+		.v = MLX5_GET(fte_match_set_lyr_2_4, value, smac_47_16) << 16 |
+		     MLX5_GET(fte_match_set_lyr_2_4, value, smac_15_0)};
+	DECLARE_MASK_VAL(u64, dmac) = {
+		.m = MLX5_GET(fte_match_set_lyr_2_4, mask, dmac_47_16) << 16 |
+		     MLX5_GET(fte_match_set_lyr_2_4, mask, dmac_15_0),
+		.v = MLX5_GET(fte_match_set_lyr_2_4, value, dmac_47_16) << 16 |
+		     MLX5_GET(fte_match_set_lyr_2_4, value, dmac_15_0)};
+	MASK_VAL_L2(u16, ethertype, ethertype);
+
+	PRINT_MASKED_VALP(smac, u8 *, p, "%pM");
+	PRINT_MASKED_VALP(dmac, u8 *, p, "%pM");
+	PRINT_MASKED_VAL(ethertype, p, "%04x");
+
+	if (ethertype.m == 0xffff) {
+		if (ethertype.v == ETH_P_IP) {
+#define MASK_VAL_L2_BE(type, name, fld) \
+	MASK_VAL_BE(type, fte_match_set_lyr_2_4, name, mask, value, fld)
+			MASK_VAL_L2_BE(u32, src_ipv4,
+				       src_ipv4_src_ipv6.ipv4_layout.ipv4);
+			MASK_VAL_L2_BE(u32, dst_ipv4,
+				       dst_ipv4_dst_ipv6.ipv4_layout.ipv4);
+
+			PRINT_MASKED_VALP(src_ipv4, typeof(&src_ipv4.v), p,
+					  "%pI4");
+			PRINT_MASKED_VALP(dst_ipv4, typeof(&dst_ipv4.v), p,
+					  "%pI4");
+		} else if (ethertype.v == ETH_P_IPV6) {
+			static const struct in6_addr full_ones = {
+				.in6_u.u6_addr32 = {__constant_htonl(0xffffffff),
+						    __constant_htonl(0xffffffff),
+						    __constant_htonl(0xffffffff),
+						    __constant_htonl(0xffffffff)},
+			};
+			DECLARE_MASK_VAL(struct in6_addr, src_ipv6);
+			DECLARE_MASK_VAL(struct in6_addr, dst_ipv6);
+
+			memcpy(src_ipv6.m.in6_u.u6_addr8,
+			       MLX5_ADDR_OF(fte_match_set_lyr_2_4, mask,
+					    src_ipv4_src_ipv6.ipv6_layout.ipv6),
+			       sizeof(src_ipv6.m));
+			memcpy(dst_ipv6.m.in6_u.u6_addr8,
+			       MLX5_ADDR_OF(fte_match_set_lyr_2_4, mask,
+					    dst_ipv4_dst_ipv6.ipv6_layout.ipv6),
+			       sizeof(dst_ipv6.m));
+			memcpy(src_ipv6.v.in6_u.u6_addr8,
+			       MLX5_ADDR_OF(fte_match_set_lyr_2_4, value,
+					    src_ipv4_src_ipv6.ipv6_layout.ipv6),
+			       sizeof(src_ipv6.v));
+			memcpy(dst_ipv6.v.in6_u.u6_addr8,
+			       MLX5_ADDR_OF(fte_match_set_lyr_2_4, value,
+					    dst_ipv4_dst_ipv6.ipv6_layout.ipv6),
+			       sizeof(dst_ipv6.v));
+
+			if (!memcmp(&src_ipv6.m, &full_ones, sizeof(full_ones)))
+				trace_seq_printf(p, "src_ipv6=%pI6 ",
+						 &src_ipv6.v);
+			if (!memcmp(&dst_ipv6.m, &full_ones, sizeof(full_ones)))
+				trace_seq_printf(p, "dst_ipv6=%pI6 ",
+						 &dst_ipv6.v);
+		}
+	}
+
+#define PRINT_MASKED_VAL_L2(type, name, fld, p, format) {\
+	MASK_VAL_L2(type, name, fld);		         \
+	PRINT_MASKED_VAL(name, p, format);		 \
+}
+
+	PRINT_MASKED_VAL_L2(u8, ip_protocol, ip_protocol, p, "%02x");
+	PRINT_MASKED_VAL_L2(u16, tcp_flags, tcp_flags, p, "%x");
+	PRINT_MASKED_VAL_L2(u16, tcp_sport, tcp_sport, p, "%u");
+	PRINT_MASKED_VAL_L2(u16, tcp_dport, tcp_dport, p, "%u");
+	PRINT_MASKED_VAL_L2(u16, udp_sport, udp_sport, p, "%u");
+	PRINT_MASKED_VAL_L2(u16, udp_dport, udp_dport, p, "%u");
+	PRINT_MASKED_VAL_L2(u16, first_vid, first_vid, p, "%04x");
+	PRINT_MASKED_VAL_L2(u8, first_prio, first_prio, p, "%x");
+	PRINT_MASKED_VAL_L2(u8, first_cfi, first_cfi, p, "%d");
+	PRINT_MASKED_VAL_L2(u8, ip_dscp, ip_dscp, p, "%02x");
+	PRINT_MASKED_VAL_L2(u8, ip_ecn, ip_ecn, p, "%x");
+	PRINT_MASKED_VAL_L2(u8, cvlan_tag, cvlan_tag, p, "%d");
+	PRINT_MASKED_VAL_L2(u8, svlan_tag, svlan_tag, p, "%d");
+	PRINT_MASKED_VAL_L2(u8, frag, frag, p, "%d");
+}
+
+static void print_misc_parameters_hdrs(struct trace_seq *p,
+				       const u32 *mask, const u32 *value)
+{
+#define MASK_VAL_MISC(type, name, fld) \
+	MASK_VAL(type, fte_match_set_misc, name, mask, value, fld)
+#define PRINT_MASKED_VAL_MISC(type, name, fld, p, format) {\
+	MASK_VAL_MISC(type, name, fld);			   \
+	PRINT_MASKED_VAL(name, p, format);		   \
+}
+	DECLARE_MASK_VAL(u64, gre_key) = {
+		.m = MLX5_GET(fte_match_set_misc, mask, gre_key_h) << 8 |
+		     MLX5_GET(fte_match_set_misc, mask, gre_key_l),
+		.v = MLX5_GET(fte_match_set_misc, value, gre_key_h) << 8 |
+		     MLX5_GET(fte_match_set_misc, value, gre_key_l)};
+
+	PRINT_MASKED_VAL(gre_key, p, "%llu");
+	PRINT_MASKED_VAL_MISC(u32, source_sqn, source_sqn, p, "%u");
+	PRINT_MASKED_VAL_MISC(u16, source_port, source_port, p, "%u");
+	PRINT_MASKED_VAL_MISC(u8, outer_second_prio, outer_second_prio,
+			      p, "%u");
+	PRINT_MASKED_VAL_MISC(u8, outer_second_cfi, outer_second_cfi, p, "%u");
+	PRINT_MASKED_VAL_MISC(u16, outer_second_vid, outer_second_vid, p, "%u");
+	PRINT_MASKED_VAL_MISC(u8, inner_second_prio, inner_second_prio,
+			      p, "%u");
+	PRINT_MASKED_VAL_MISC(u8, inner_second_cfi, inner_second_cfi, p, "%u");
+	PRINT_MASKED_VAL_MISC(u16, inner_second_vid, inner_second_vid, p, "%u");
+
+	PRINT_MASKED_VAL_MISC(u8, outer_second_cvlan_tag,
+			      outer_second_cvlan_tag, p, "%u");
+	PRINT_MASKED_VAL_MISC(u8, inner_second_cvlan_tag,
+			      inner_second_cvlan_tag, p, "%u");
+	PRINT_MASKED_VAL_MISC(u8, outer_second_svlan_tag,
+			      outer_second_svlan_tag, p, "%u");
+	PRINT_MASKED_VAL_MISC(u8, inner_second_svlan_tag,
+			      inner_second_svlan_tag, p, "%u");
+
+	PRINT_MASKED_VAL_MISC(u8, gre_protocol, gre_protocol, p, "%u");
+
+	PRINT_MASKED_VAL_MISC(u32, vxlan_vni, vxlan_vni, p, "%u");
+	PRINT_MASKED_VAL_MISC(u32, outer_ipv6_flow_label, outer_ipv6_flow_label,
+			      p, "%x");
+	PRINT_MASKED_VAL_MISC(u32, inner_ipv6_flow_label, inner_ipv6_flow_label,
+			      p, "%x");
+}
+
+const char *parse_fs_hdrs(struct trace_seq *p,
+			  u8 match_criteria_enable,
+			  const u32 *mask_outer,
+			  const u32 *mask_misc,
+			  const u32 *mask_inner,
+			  const u32 *value_outer,
+			  const u32 *value_misc,
+			  const u32 *value_inner)
+{
+	const char *ret = trace_seq_buffer_ptr(p);
+
+	if (match_criteria_enable &
+	    1 << MLX5_CREATE_FLOW_GROUP_IN_MATCH_CRITERIA_ENABLE_OUTER_HEADERS) {
+		trace_seq_printf(p, "[outer] ");
+		print_lyr_2_4_hdrs(p, mask_outer, value_outer);
+	}
+
+	if (match_criteria_enable &
+	    1 << MLX5_CREATE_FLOW_GROUP_IN_MATCH_CRITERIA_ENABLE_MISC_PARAMETERS) {
+		trace_seq_printf(p, "[misc] ");
+		print_misc_parameters_hdrs(p, mask_misc, value_misc);
+	}
+	if (match_criteria_enable &
+	    1 << MLX5_CREATE_FLOW_GROUP_IN_MATCH_CRITERIA_ENABLE_INNER_HEADERS) {
+		trace_seq_printf(p, "[inner] ");
+		print_lyr_2_4_hdrs(p, mask_inner, value_inner);
+	}
+	trace_seq_putc(p, 0);
+	return ret;
+}
+
+const char *parse_fs_dst(struct trace_seq *p,
+			 const struct mlx5_flow_destination *dst,
+			 u32 counter_id)
+{
+	const char *ret = trace_seq_buffer_ptr(p);
+
+	switch (dst->type) {
+	case MLX5_FLOW_DESTINATION_TYPE_VPORT:
+		trace_seq_printf(p, "vport=%u\n", dst->vport.num);
+		break;
+	case MLX5_FLOW_DESTINATION_TYPE_FLOW_TABLE:
+		trace_seq_printf(p, "ft=%p\n", dst->ft);
+		break;
+	case MLX5_FLOW_DESTINATION_TYPE_FLOW_TABLE_NUM:
+		trace_seq_printf(p, "ft_num=%u\n", dst->ft_num);
+		break;
+	case MLX5_FLOW_DESTINATION_TYPE_TIR:
+		trace_seq_printf(p, "tir=%u\n", dst->tir_num);
+		break;
+	case MLX5_FLOW_DESTINATION_TYPE_COUNTER:
+		trace_seq_printf(p, "counter_id=%u\n", counter_id);
+		break;
+	case MLX5_FLOW_DESTINATION_TYPE_PORT:
+		trace_seq_printf(p, "port\n");
+		break;
+	}
+
+	trace_seq_putc(p, 0);
+	return ret;
+}
+
+EXPORT_TRACEPOINT_SYMBOL(mlx5_fs_add_fg);
+EXPORT_TRACEPOINT_SYMBOL(mlx5_fs_del_fg);
+EXPORT_TRACEPOINT_SYMBOL(mlx5_fs_set_fte);
+EXPORT_TRACEPOINT_SYMBOL(mlx5_fs_del_fte);
+EXPORT_TRACEPOINT_SYMBOL(mlx5_fs_add_rule);
+EXPORT_TRACEPOINT_SYMBOL(mlx5_fs_del_rule);
+
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/diag/fs_tracepoint.h b/drivers/net/ethernet/mellanox/mlx5/core/diag/fs_tracepoint.h
new file mode 100644
index 000000000..0240aee91
--- /dev/null
+++ b/drivers/net/ethernet/mellanox/mlx5/core/diag/fs_tracepoint.h
@@ -0,0 +1,286 @@
+/*
+ * Copyright (c) 2017, 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.
+ */
+
+#if !defined(_MLX5_FS_TP_) || defined(TRACE_HEADER_MULTI_READ)
+#define _MLX5_FS_TP_
+
+#include <linux/tracepoint.h>
+#include <linux/trace_seq.h>
+#include "../fs_core.h"
+
+#undef TRACE_SYSTEM
+#define TRACE_SYSTEM mlx5
+
+#define __parse_fs_hdrs(match_criteria_enable, mouter, mmisc, minner, vouter, \
+			vinner, vmisc)					      \
+	parse_fs_hdrs(p, match_criteria_enable, mouter, mmisc, minner, vouter,\
+		      vinner, vmisc)
+
+const char *parse_fs_hdrs(struct trace_seq *p,
+			  u8 match_criteria_enable,
+			  const u32 *mask_outer,
+			  const u32 *mask_misc,
+			  const u32 *mask_inner,
+			  const u32 *value_outer,
+			  const u32 *value_misc,
+			  const u32 *value_inner);
+
+#define __parse_fs_dst(dst, counter_id) \
+	parse_fs_dst(p, (const struct mlx5_flow_destination *)dst, counter_id)
+
+const char *parse_fs_dst(struct trace_seq *p,
+			 const struct mlx5_flow_destination *dst,
+			 u32 counter_id);
+
+TRACE_EVENT(mlx5_fs_add_fg,
+	    TP_PROTO(const struct mlx5_flow_group *fg),
+	    TP_ARGS(fg),
+	    TP_STRUCT__entry(
+		__field(const struct mlx5_flow_group *, fg)
+		__field(const struct mlx5_flow_table *, ft)
+		__field(u32, start_index)
+		__field(u32, end_index)
+		__field(u32, id)
+		__field(u8, mask_enable)
+		__array(u32, mask_outer, MLX5_ST_SZ_DW(fte_match_set_lyr_2_4))
+		__array(u32, mask_inner, MLX5_ST_SZ_DW(fte_match_set_lyr_2_4))
+		__array(u32, mask_misc, MLX5_ST_SZ_DW(fte_match_set_misc))
+	    ),
+	    TP_fast_assign(
+			   __entry->fg = fg;
+			   fs_get_obj(__entry->ft, fg->node.parent);
+			   __entry->start_index = fg->start_index;
+			   __entry->end_index = fg->start_index + fg->max_ftes;
+			   __entry->id = fg->id;
+			   __entry->mask_enable = fg->mask.match_criteria_enable;
+			   memcpy(__entry->mask_outer,
+				  MLX5_ADDR_OF(fte_match_param,
+					       &fg->mask.match_criteria,
+					       outer_headers),
+				  sizeof(__entry->mask_outer));
+			   memcpy(__entry->mask_inner,
+				  MLX5_ADDR_OF(fte_match_param,
+					       &fg->mask.match_criteria,
+					       inner_headers),
+				  sizeof(__entry->mask_inner));
+			   memcpy(__entry->mask_misc,
+				  MLX5_ADDR_OF(fte_match_param,
+					       &fg->mask.match_criteria,
+					       misc_parameters),
+				  sizeof(__entry->mask_misc));
+
+	    ),
+	    TP_printk("fg=%p ft=%p id=%u start=%u end=%u bit_mask=%02x %s\n",
+		      __entry->fg, __entry->ft, __entry->id,
+		      __entry->start_index, __entry->end_index,
+		      __entry->mask_enable,
+		      __parse_fs_hdrs(__entry->mask_enable,
+				      __entry->mask_outer,
+				      __entry->mask_misc,
+				      __entry->mask_inner,
+				      __entry->mask_outer,
+				      __entry->mask_misc,
+				      __entry->mask_inner))
+	    );
+
+TRACE_EVENT(mlx5_fs_del_fg,
+	    TP_PROTO(const struct mlx5_flow_group *fg),
+	    TP_ARGS(fg),
+	    TP_STRUCT__entry(
+		__field(const struct mlx5_flow_group *, fg)
+		__field(u32, id)
+	    ),
+	    TP_fast_assign(
+			   __entry->fg = fg;
+			   __entry->id = fg->id;
+
+	    ),
+	    TP_printk("fg=%p id=%u\n",
+		      __entry->fg, __entry->id)
+	    );
+
+#define ACTION_FLAGS \
+	{MLX5_FLOW_CONTEXT_ACTION_ALLOW,	 "ALLOW"},\
+	{MLX5_FLOW_CONTEXT_ACTION_DROP,		 "DROP"},\
+	{MLX5_FLOW_CONTEXT_ACTION_FWD_DEST,	 "FWD"},\
+	{MLX5_FLOW_CONTEXT_ACTION_COUNT,	 "CNT"},\
+	{MLX5_FLOW_CONTEXT_ACTION_ENCAP,	 "ENCAP"},\
+	{MLX5_FLOW_CONTEXT_ACTION_DECAP,	 "DECAP"},\
+	{MLX5_FLOW_CONTEXT_ACTION_MOD_HDR,	 "MOD_HDR"},\
+	{MLX5_FLOW_CONTEXT_ACTION_VLAN_PUSH,	 "VLAN_PUSH"},\
+	{MLX5_FLOW_CONTEXT_ACTION_VLAN_POP,	 "VLAN_POP"},\
+	{MLX5_FLOW_CONTEXT_ACTION_VLAN_PUSH_2,	 "VLAN_PUSH_2"},\
+	{MLX5_FLOW_CONTEXT_ACTION_VLAN_POP_2,	 "VLAN_POP_2"},\
+	{MLX5_FLOW_CONTEXT_ACTION_FWD_NEXT_PRIO, "NEXT_PRIO"}
+
+TRACE_EVENT(mlx5_fs_set_fte,
+	    TP_PROTO(const struct fs_fte *fte, int new_fte),
+	    TP_ARGS(fte, new_fte),
+	    TP_STRUCT__entry(
+		__field(const struct fs_fte *, fte)
+		__field(const struct mlx5_flow_group *, fg)
+		__field(u32, group_index)
+		__field(u32, index)
+		__field(u32, action)
+		__field(u32, flow_tag)
+		__field(u8,  mask_enable)
+		__field(int, new_fte)
+		__array(u32, mask_outer, MLX5_ST_SZ_DW(fte_match_set_lyr_2_4))
+		__array(u32, mask_inner, MLX5_ST_SZ_DW(fte_match_set_lyr_2_4))
+		__array(u32, mask_misc, MLX5_ST_SZ_DW(fte_match_set_misc))
+		__array(u32, value_outer, MLX5_ST_SZ_DW(fte_match_set_lyr_2_4))
+		__array(u32, value_inner, MLX5_ST_SZ_DW(fte_match_set_lyr_2_4))
+		__array(u32, value_misc, MLX5_ST_SZ_DW(fte_match_set_misc))
+	    ),
+	    TP_fast_assign(
+			   __entry->fte = fte;
+			   __entry->new_fte = new_fte;
+			   fs_get_obj(__entry->fg, fte->node.parent);
+			   __entry->group_index = __entry->fg->id;
+			   __entry->index = fte->index;
+			   __entry->action = fte->action.action;
+			   __entry->mask_enable = __entry->fg->mask.match_criteria_enable;
+			   __entry->flow_tag = fte->action.flow_tag;
+			   memcpy(__entry->mask_outer,
+				  MLX5_ADDR_OF(fte_match_param,
+					       &__entry->fg->mask.match_criteria,
+					       outer_headers),
+				  sizeof(__entry->mask_outer));
+			   memcpy(__entry->mask_inner,
+				  MLX5_ADDR_OF(fte_match_param,
+					       &__entry->fg->mask.match_criteria,
+					       inner_headers),
+				  sizeof(__entry->mask_inner));
+			   memcpy(__entry->mask_misc,
+				  MLX5_ADDR_OF(fte_match_param,
+					       &__entry->fg->mask.match_criteria,
+					       misc_parameters),
+				  sizeof(__entry->mask_misc));
+			   memcpy(__entry->value_outer,
+				  MLX5_ADDR_OF(fte_match_param,
+					       &fte->val,
+					       outer_headers),
+				  sizeof(__entry->value_outer));
+			   memcpy(__entry->value_inner,
+				  MLX5_ADDR_OF(fte_match_param,
+					       &fte->val,
+					       inner_headers),
+				  sizeof(__entry->value_inner));
+			   memcpy(__entry->value_misc,
+				  MLX5_ADDR_OF(fte_match_param,
+					       &fte->val,
+					       misc_parameters),
+				  sizeof(__entry->value_misc));
+
+	    ),
+	    TP_printk("op=%s fte=%p fg=%p index=%u group_index=%u action=<%s> flow_tag=%x %s\n",
+		      __entry->new_fte ? "add" : "set",
+		      __entry->fte, __entry->fg, __entry->index,
+		      __entry->group_index, __print_flags(__entry->action, "|",
+							  ACTION_FLAGS),
+		      __entry->flow_tag,
+		      __parse_fs_hdrs(__entry->mask_enable,
+				      __entry->mask_outer,
+				      __entry->mask_misc,
+				      __entry->mask_inner,
+				      __entry->value_outer,
+				      __entry->value_misc,
+				      __entry->value_inner))
+	    );
+
+TRACE_EVENT(mlx5_fs_del_fte,
+	    TP_PROTO(const struct fs_fte *fte),
+	    TP_ARGS(fte),
+	    TP_STRUCT__entry(
+		__field(const struct fs_fte *, fte)
+		__field(u32, index)
+	    ),
+	    TP_fast_assign(
+			   __entry->fte = fte;
+			   __entry->index = fte->index;
+
+	    ),
+	    TP_printk("fte=%p index=%u\n",
+		      __entry->fte, __entry->index)
+	    );
+
+TRACE_EVENT(mlx5_fs_add_rule,
+	    TP_PROTO(const struct mlx5_flow_rule *rule),
+	    TP_ARGS(rule),
+	    TP_STRUCT__entry(
+		__field(const struct mlx5_flow_rule *, rule)
+		__field(const struct fs_fte *, fte)
+		__field(u32, sw_action)
+		__field(u32, index)
+		__field(u32, counter_id)
+		__array(u8, destination, sizeof(struct mlx5_flow_destination))
+	    ),
+	    TP_fast_assign(
+			   __entry->rule = rule;
+			   fs_get_obj(__entry->fte, rule->node.parent);
+			   __entry->index = __entry->fte->dests_size - 1;
+			   __entry->sw_action = rule->sw_action;
+			   memcpy(__entry->destination,
+				  &rule->dest_attr,
+				  sizeof(__entry->destination));
+			   if (rule->dest_attr.type & MLX5_FLOW_DESTINATION_TYPE_COUNTER &&
+			       rule->dest_attr.counter)
+				__entry->counter_id =
+				rule->dest_attr.counter->id;
+	    ),
+	    TP_printk("rule=%p fte=%p index=%u sw_action=<%s> [dst] %s\n",
+		      __entry->rule, __entry->fte, __entry->index,
+		      __print_flags(__entry->sw_action, "|", ACTION_FLAGS),
+		      __parse_fs_dst(__entry->destination, __entry->counter_id))
+	    );
+
+TRACE_EVENT(mlx5_fs_del_rule,
+	    TP_PROTO(const struct mlx5_flow_rule *rule),
+	    TP_ARGS(rule),
+	    TP_STRUCT__entry(
+		__field(const struct mlx5_flow_rule *, rule)
+		__field(const struct fs_fte *, fte)
+	    ),
+	    TP_fast_assign(
+			   __entry->rule = rule;
+			   fs_get_obj(__entry->fte, rule->node.parent);
+	    ),
+	    TP_printk("rule=%p fte=%p\n",
+		      __entry->rule, __entry->fte)
+	    );
+#endif
+
+#undef TRACE_INCLUDE_PATH
+#define TRACE_INCLUDE_PATH ./diag
+#undef TRACE_INCLUDE_FILE
+#define TRACE_INCLUDE_FILE fs_tracepoint
+#include <trace/define_trace.h>
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/diag/fw_tracer.c b/drivers/net/ethernet/mellanox/mlx5/core/diag/fw_tracer.c
new file mode 100644
index 000000000..a22e932a0
--- /dev/null
+++ b/drivers/net/ethernet/mellanox/mlx5/core/diag/fw_tracer.c
@@ -0,0 +1,950 @@
+/*
+ * 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.
+ */
+#define CREATE_TRACE_POINTS
+#include "fw_tracer.h"
+#include "fw_tracer_tracepoint.h"
+
+static int mlx5_query_mtrc_caps(struct mlx5_fw_tracer *tracer)
+{
+	u32 *string_db_base_address_out = tracer->str_db.base_address_out;
+	u32 *string_db_size_out = tracer->str_db.size_out;
+	struct mlx5_core_dev *dev = tracer->dev;
+	u32 out[MLX5_ST_SZ_DW(mtrc_cap)] = {0};
+	u32 in[MLX5_ST_SZ_DW(mtrc_cap)] = {0};
+	void *mtrc_cap_sp;
+	int err, i;
+
+	err = mlx5_core_access_reg(dev, in, sizeof(in), out, sizeof(out),
+				   MLX5_REG_MTRC_CAP, 0, 0);
+	if (err) {
+		mlx5_core_warn(dev, "FWTracer: Error reading tracer caps %d\n",
+			       err);
+		return err;
+	}
+
+	if (!MLX5_GET(mtrc_cap, out, trace_to_memory)) {
+		mlx5_core_dbg(dev, "FWTracer: Device does not support logging traces to memory\n");
+		return -ENOTSUPP;
+	}
+
+	tracer->trc_ver = MLX5_GET(mtrc_cap, out, trc_ver);
+	tracer->str_db.first_string_trace =
+			MLX5_GET(mtrc_cap, out, first_string_trace);
+	tracer->str_db.num_string_trace =
+			MLX5_GET(mtrc_cap, out, num_string_trace);
+	tracer->str_db.num_string_db = MLX5_GET(mtrc_cap, out, num_string_db);
+	tracer->owner = !!MLX5_GET(mtrc_cap, out, trace_owner);
+
+	for (i = 0; i < tracer->str_db.num_string_db; i++) {
+		mtrc_cap_sp = MLX5_ADDR_OF(mtrc_cap, out, string_db_param[i]);
+		string_db_base_address_out[i] = MLX5_GET(mtrc_string_db_param,
+							 mtrc_cap_sp,
+							 string_db_base_address);
+		string_db_size_out[i] = MLX5_GET(mtrc_string_db_param,
+						 mtrc_cap_sp, string_db_size);
+	}
+
+	return err;
+}
+
+static int mlx5_set_mtrc_caps_trace_owner(struct mlx5_fw_tracer *tracer,
+					  u32 *out, u32 out_size,
+					  u8 trace_owner)
+{
+	struct mlx5_core_dev *dev = tracer->dev;
+	u32 in[MLX5_ST_SZ_DW(mtrc_cap)] = {0};
+
+	MLX5_SET(mtrc_cap, in, trace_owner, trace_owner);
+
+	return mlx5_core_access_reg(dev, in, sizeof(in), out, out_size,
+				    MLX5_REG_MTRC_CAP, 0, 1);
+}
+
+static int mlx5_fw_tracer_ownership_acquire(struct mlx5_fw_tracer *tracer)
+{
+	struct mlx5_core_dev *dev = tracer->dev;
+	u32 out[MLX5_ST_SZ_DW(mtrc_cap)] = {0};
+	int err;
+
+	err = mlx5_set_mtrc_caps_trace_owner(tracer, out, sizeof(out),
+					     MLX5_FW_TRACER_ACQUIRE_OWNERSHIP);
+	if (err) {
+		mlx5_core_warn(dev, "FWTracer: Acquire tracer ownership failed %d\n",
+			       err);
+		return err;
+	}
+
+	tracer->owner = !!MLX5_GET(mtrc_cap, out, trace_owner);
+
+	if (!tracer->owner)
+		return -EBUSY;
+
+	return 0;
+}
+
+static void mlx5_fw_tracer_ownership_release(struct mlx5_fw_tracer *tracer)
+{
+	u32 out[MLX5_ST_SZ_DW(mtrc_cap)] = {0};
+
+	mlx5_set_mtrc_caps_trace_owner(tracer, out, sizeof(out),
+				       MLX5_FW_TRACER_RELEASE_OWNERSHIP);
+	tracer->owner = false;
+}
+
+static int mlx5_fw_tracer_create_log_buf(struct mlx5_fw_tracer *tracer)
+{
+	struct mlx5_core_dev *dev = tracer->dev;
+	struct device *ddev = &dev->pdev->dev;
+	dma_addr_t dma;
+	void *buff;
+	gfp_t gfp;
+	int err;
+
+	tracer->buff.size = TRACE_BUFFER_SIZE_BYTE;
+
+	gfp = GFP_KERNEL | __GFP_ZERO;
+	buff = (void *)__get_free_pages(gfp,
+					get_order(tracer->buff.size));
+	if (!buff) {
+		err = -ENOMEM;
+		mlx5_core_warn(dev, "FWTracer: Failed to allocate pages, %d\n", err);
+		return err;
+	}
+	tracer->buff.log_buf = buff;
+
+	dma = dma_map_single(ddev, buff, tracer->buff.size, DMA_FROM_DEVICE);
+	if (dma_mapping_error(ddev, dma)) {
+		mlx5_core_warn(dev, "FWTracer: Unable to map DMA: %d\n",
+			       dma_mapping_error(ddev, dma));
+		err = -ENOMEM;
+		goto free_pages;
+	}
+	tracer->buff.dma = dma;
+
+	return 0;
+
+free_pages:
+	free_pages((unsigned long)tracer->buff.log_buf, get_order(tracer->buff.size));
+
+	return err;
+}
+
+static void mlx5_fw_tracer_destroy_log_buf(struct mlx5_fw_tracer *tracer)
+{
+	struct mlx5_core_dev *dev = tracer->dev;
+	struct device *ddev = &dev->pdev->dev;
+
+	if (!tracer->buff.log_buf)
+		return;
+
+	dma_unmap_single(ddev, tracer->buff.dma, tracer->buff.size, DMA_FROM_DEVICE);
+	free_pages((unsigned long)tracer->buff.log_buf, get_order(tracer->buff.size));
+}
+
+static int mlx5_fw_tracer_create_mkey(struct mlx5_fw_tracer *tracer)
+{
+	struct mlx5_core_dev *dev = tracer->dev;
+	int err, inlen, i;
+	__be64 *mtt;
+	void *mkc;
+	u32 *in;
+
+	inlen = MLX5_ST_SZ_BYTES(create_mkey_in) +
+			sizeof(*mtt) * round_up(TRACER_BUFFER_PAGE_NUM, 2);
+
+	in = kvzalloc(inlen, GFP_KERNEL);
+	if (!in)
+		return -ENOMEM;
+
+	MLX5_SET(create_mkey_in, in, translations_octword_actual_size,
+		 DIV_ROUND_UP(TRACER_BUFFER_PAGE_NUM, 2));
+	mtt = (u64 *)MLX5_ADDR_OF(create_mkey_in, in, klm_pas_mtt);
+	for (i = 0 ; i < TRACER_BUFFER_PAGE_NUM ; i++)
+		mtt[i] = cpu_to_be64(tracer->buff.dma + i * PAGE_SIZE);
+
+	mkc = MLX5_ADDR_OF(create_mkey_in, in, memory_key_mkey_entry);
+	MLX5_SET(mkc, mkc, access_mode_1_0, MLX5_MKC_ACCESS_MODE_MTT);
+	MLX5_SET(mkc, mkc, lr, 1);
+	MLX5_SET(mkc, mkc, lw, 1);
+	MLX5_SET(mkc, mkc, pd, tracer->buff.pdn);
+	MLX5_SET(mkc, mkc, bsf_octword_size, 0);
+	MLX5_SET(mkc, mkc, qpn, 0xffffff);
+	MLX5_SET(mkc, mkc, log_page_size, PAGE_SHIFT);
+	MLX5_SET(mkc, mkc, translations_octword_size,
+		 DIV_ROUND_UP(TRACER_BUFFER_PAGE_NUM, 2));
+	MLX5_SET64(mkc, mkc, start_addr, tracer->buff.dma);
+	MLX5_SET64(mkc, mkc, len, tracer->buff.size);
+	err = mlx5_core_create_mkey(dev, &tracer->buff.mkey, in, inlen);
+	if (err)
+		mlx5_core_warn(dev, "FWTracer: Failed to create mkey, %d\n", err);
+
+	kvfree(in);
+
+	return err;
+}
+
+static void mlx5_fw_tracer_free_strings_db(struct mlx5_fw_tracer *tracer)
+{
+	u32 num_string_db = tracer->str_db.num_string_db;
+	int i;
+
+	for (i = 0; i < num_string_db; i++) {
+		kfree(tracer->str_db.buffer[i]);
+		tracer->str_db.buffer[i] = NULL;
+	}
+}
+
+static int mlx5_fw_tracer_allocate_strings_db(struct mlx5_fw_tracer *tracer)
+{
+	u32 *string_db_size_out = tracer->str_db.size_out;
+	u32 num_string_db = tracer->str_db.num_string_db;
+	int i;
+
+	for (i = 0; i < num_string_db; i++) {
+		tracer->str_db.buffer[i] = kzalloc(string_db_size_out[i], GFP_KERNEL);
+		if (!tracer->str_db.buffer[i])
+			goto free_strings_db;
+	}
+
+	return 0;
+
+free_strings_db:
+	mlx5_fw_tracer_free_strings_db(tracer);
+	return -ENOMEM;
+}
+
+static void mlx5_tracer_read_strings_db(struct work_struct *work)
+{
+	struct mlx5_fw_tracer *tracer = container_of(work, struct mlx5_fw_tracer,
+						     read_fw_strings_work);
+	u32 num_of_reads, num_string_db = tracer->str_db.num_string_db;
+	struct mlx5_core_dev *dev = tracer->dev;
+	u32 in[MLX5_ST_SZ_DW(mtrc_cap)] = {0};
+	u32 leftovers, offset;
+	int err = 0, i, j;
+	u32 *out, outlen;
+	void *out_value;
+
+	outlen = MLX5_ST_SZ_BYTES(mtrc_stdb) + STRINGS_DB_READ_SIZE_BYTES;
+	out = kzalloc(outlen, GFP_KERNEL);
+	if (!out) {
+		err = -ENOMEM;
+		goto out;
+	}
+
+	for (i = 0; i < num_string_db; i++) {
+		offset = 0;
+		MLX5_SET(mtrc_stdb, in, string_db_index, i);
+		num_of_reads = tracer->str_db.size_out[i] /
+				STRINGS_DB_READ_SIZE_BYTES;
+		leftovers = (tracer->str_db.size_out[i] %
+				STRINGS_DB_READ_SIZE_BYTES) /
+					STRINGS_DB_LEFTOVER_SIZE_BYTES;
+
+		MLX5_SET(mtrc_stdb, in, read_size, STRINGS_DB_READ_SIZE_BYTES);
+		for (j = 0; j < num_of_reads; j++) {
+			MLX5_SET(mtrc_stdb, in, start_offset, offset);
+
+			err = mlx5_core_access_reg(dev, in, sizeof(in), out,
+						   outlen, MLX5_REG_MTRC_STDB,
+						   0, 1);
+			if (err) {
+				mlx5_core_dbg(dev, "FWTracer: Failed to read strings DB %d\n",
+					      err);
+				goto out_free;
+			}
+
+			out_value = MLX5_ADDR_OF(mtrc_stdb, out, string_db_data);
+			memcpy(tracer->str_db.buffer[i] + offset, out_value,
+			       STRINGS_DB_READ_SIZE_BYTES);
+			offset += STRINGS_DB_READ_SIZE_BYTES;
+		}
+
+		/* Strings database is aligned to 64, need to read leftovers*/
+		MLX5_SET(mtrc_stdb, in, read_size,
+			 STRINGS_DB_LEFTOVER_SIZE_BYTES);
+		for (j = 0; j < leftovers; j++) {
+			MLX5_SET(mtrc_stdb, in, start_offset, offset);
+
+			err = mlx5_core_access_reg(dev, in, sizeof(in), out,
+						   outlen, MLX5_REG_MTRC_STDB,
+						   0, 1);
+			if (err) {
+				mlx5_core_dbg(dev, "FWTracer: Failed to read strings DB %d\n",
+					      err);
+				goto out_free;
+			}
+
+			out_value = MLX5_ADDR_OF(mtrc_stdb, out, string_db_data);
+			memcpy(tracer->str_db.buffer[i] + offset, out_value,
+			       STRINGS_DB_LEFTOVER_SIZE_BYTES);
+			offset += STRINGS_DB_LEFTOVER_SIZE_BYTES;
+		}
+	}
+
+	tracer->str_db.loaded = true;
+
+out_free:
+	kfree(out);
+out:
+	return;
+}
+
+static void mlx5_fw_tracer_arm(struct mlx5_core_dev *dev)
+{
+	u32 out[MLX5_ST_SZ_DW(mtrc_ctrl)] = {0};
+	u32 in[MLX5_ST_SZ_DW(mtrc_ctrl)] = {0};
+	int err;
+
+	MLX5_SET(mtrc_ctrl, in, arm_event, 1);
+
+	err = mlx5_core_access_reg(dev, in, sizeof(in), out, sizeof(out),
+				   MLX5_REG_MTRC_CTRL, 0, 1);
+	if (err)
+		mlx5_core_warn(dev, "FWTracer: Failed to arm tracer event %d\n", err);
+}
+
+static const char *VAL_PARM		= "%llx";
+static const char *REPLACE_64_VAL_PARM	= "%x%x";
+static const char *PARAM_CHAR		= "%";
+
+static int mlx5_tracer_message_hash(u32 message_id)
+{
+	return jhash_1word(message_id, 0) & (MESSAGE_HASH_SIZE - 1);
+}
+
+static struct tracer_string_format *mlx5_tracer_message_insert(struct mlx5_fw_tracer *tracer,
+							       struct tracer_event *tracer_event)
+{
+	struct hlist_head *head =
+		&tracer->hash[mlx5_tracer_message_hash(tracer_event->string_event.tmsn)];
+	struct tracer_string_format *cur_string;
+
+	cur_string = kzalloc(sizeof(*cur_string), GFP_KERNEL);
+	if (!cur_string)
+		return NULL;
+
+	hlist_add_head(&cur_string->hlist, head);
+
+	return cur_string;
+}
+
+static struct tracer_string_format *mlx5_tracer_get_string(struct mlx5_fw_tracer *tracer,
+							   struct tracer_event *tracer_event)
+{
+	struct tracer_string_format *cur_string;
+	u32 str_ptr, offset;
+	int i;
+
+	str_ptr = tracer_event->string_event.string_param;
+
+	for (i = 0; i < tracer->str_db.num_string_db; i++) {
+		if (str_ptr > tracer->str_db.base_address_out[i] &&
+		    str_ptr < tracer->str_db.base_address_out[i] +
+		    tracer->str_db.size_out[i]) {
+			offset = str_ptr - tracer->str_db.base_address_out[i];
+			/* add it to the hash */
+			cur_string = mlx5_tracer_message_insert(tracer, tracer_event);
+			if (!cur_string)
+				return NULL;
+			cur_string->string = (char *)(tracer->str_db.buffer[i] +
+							offset);
+			return cur_string;
+		}
+	}
+
+	return NULL;
+}
+
+static void mlx5_tracer_clean_message(struct tracer_string_format *str_frmt)
+{
+	hlist_del(&str_frmt->hlist);
+	kfree(str_frmt);
+}
+
+static int mlx5_tracer_get_num_of_params(char *str)
+{
+	char *substr, *pstr = str;
+	int num_of_params = 0;
+
+	/* replace %llx with %x%x */
+	substr = strstr(pstr, VAL_PARM);
+	while (substr) {
+		memcpy(substr, REPLACE_64_VAL_PARM, 4);
+		pstr = substr;
+		substr = strstr(pstr, VAL_PARM);
+	}
+
+	/* count all the % characters */
+	substr = strstr(str, PARAM_CHAR);
+	while (substr) {
+		num_of_params += 1;
+		str = substr + 1;
+		substr = strstr(str, PARAM_CHAR);
+	}
+
+	return num_of_params;
+}
+
+static struct tracer_string_format *mlx5_tracer_message_find(struct hlist_head *head,
+							     u8 event_id, u32 tmsn)
+{
+	struct tracer_string_format *message;
+
+	hlist_for_each_entry(message, head, hlist)
+		if (message->event_id == event_id && message->tmsn == tmsn)
+			return message;
+
+	return NULL;
+}
+
+static struct tracer_string_format *mlx5_tracer_message_get(struct mlx5_fw_tracer *tracer,
+							    struct tracer_event *tracer_event)
+{
+	struct hlist_head *head =
+		&tracer->hash[mlx5_tracer_message_hash(tracer_event->string_event.tmsn)];
+
+	return mlx5_tracer_message_find(head, tracer_event->event_id, tracer_event->string_event.tmsn);
+}
+
+static void poll_trace(struct mlx5_fw_tracer *tracer,
+		       struct tracer_event *tracer_event, u64 *trace)
+{
+	u32 timestamp_low, timestamp_mid, timestamp_high, urts;
+
+	tracer_event->event_id = MLX5_GET(tracer_event, trace, event_id);
+	tracer_event->lost_event = MLX5_GET(tracer_event, trace, lost);
+
+	switch (tracer_event->event_id) {
+	case TRACER_EVENT_TYPE_TIMESTAMP:
+		tracer_event->type = TRACER_EVENT_TYPE_TIMESTAMP;
+		urts = MLX5_GET(tracer_timestamp_event, trace, urts);
+		if (tracer->trc_ver == 0)
+			tracer_event->timestamp_event.unreliable = !!(urts >> 2);
+		else
+			tracer_event->timestamp_event.unreliable = !!(urts & 1);
+
+		timestamp_low = MLX5_GET(tracer_timestamp_event,
+					 trace, timestamp7_0);
+		timestamp_mid = MLX5_GET(tracer_timestamp_event,
+					 trace, timestamp39_8);
+		timestamp_high = MLX5_GET(tracer_timestamp_event,
+					  trace, timestamp52_40);
+
+		tracer_event->timestamp_event.timestamp =
+				((u64)timestamp_high << 40) |
+				((u64)timestamp_mid << 8) |
+				(u64)timestamp_low;
+		break;
+	default:
+		if (tracer_event->event_id >= tracer->str_db.first_string_trace ||
+		    tracer_event->event_id <= tracer->str_db.first_string_trace +
+					      tracer->str_db.num_string_trace) {
+			tracer_event->type = TRACER_EVENT_TYPE_STRING;
+			tracer_event->string_event.timestamp =
+				MLX5_GET(tracer_string_event, trace, timestamp);
+			tracer_event->string_event.string_param =
+				MLX5_GET(tracer_string_event, trace, string_param);
+			tracer_event->string_event.tmsn =
+				MLX5_GET(tracer_string_event, trace, tmsn);
+			tracer_event->string_event.tdsn =
+				MLX5_GET(tracer_string_event, trace, tdsn);
+		} else {
+			tracer_event->type = TRACER_EVENT_TYPE_UNRECOGNIZED;
+		}
+		break;
+	}
+}
+
+static u64 get_block_timestamp(struct mlx5_fw_tracer *tracer, u64 *ts_event)
+{
+	struct tracer_event tracer_event;
+	u8 event_id;
+
+	event_id = MLX5_GET(tracer_event, ts_event, event_id);
+
+	if (event_id == TRACER_EVENT_TYPE_TIMESTAMP)
+		poll_trace(tracer, &tracer_event, ts_event);
+	else
+		tracer_event.timestamp_event.timestamp = 0;
+
+	return tracer_event.timestamp_event.timestamp;
+}
+
+static void mlx5_fw_tracer_clean_print_hash(struct mlx5_fw_tracer *tracer)
+{
+	struct tracer_string_format *str_frmt;
+	struct hlist_node *n;
+	int i;
+
+	for (i = 0; i < MESSAGE_HASH_SIZE; i++) {
+		hlist_for_each_entry_safe(str_frmt, n, &tracer->hash[i], hlist)
+			mlx5_tracer_clean_message(str_frmt);
+	}
+}
+
+static void mlx5_fw_tracer_clean_ready_list(struct mlx5_fw_tracer *tracer)
+{
+	struct tracer_string_format *str_frmt, *tmp_str;
+
+	list_for_each_entry_safe(str_frmt, tmp_str, &tracer->ready_strings_list,
+				 list)
+		list_del(&str_frmt->list);
+}
+
+static void mlx5_tracer_print_trace(struct tracer_string_format *str_frmt,
+				    struct mlx5_core_dev *dev,
+				    u64 trace_timestamp)
+{
+	char	tmp[512];
+
+	snprintf(tmp, sizeof(tmp), str_frmt->string,
+		 str_frmt->params[0],
+		 str_frmt->params[1],
+		 str_frmt->params[2],
+		 str_frmt->params[3],
+		 str_frmt->params[4],
+		 str_frmt->params[5],
+		 str_frmt->params[6]);
+
+	trace_mlx5_fw(dev->tracer, trace_timestamp, str_frmt->lost,
+		      str_frmt->event_id, tmp);
+
+	/* remove it from hash */
+	mlx5_tracer_clean_message(str_frmt);
+}
+
+static int mlx5_tracer_handle_string_trace(struct mlx5_fw_tracer *tracer,
+					   struct tracer_event *tracer_event)
+{
+	struct tracer_string_format *cur_string;
+
+	if (tracer_event->string_event.tdsn == 0) {
+		cur_string = mlx5_tracer_get_string(tracer, tracer_event);
+		if (!cur_string)
+			return -1;
+
+		cur_string->num_of_params = mlx5_tracer_get_num_of_params(cur_string->string);
+		cur_string->last_param_num = 0;
+		cur_string->event_id = tracer_event->event_id;
+		cur_string->tmsn = tracer_event->string_event.tmsn;
+		cur_string->timestamp = tracer_event->string_event.timestamp;
+		cur_string->lost = tracer_event->lost_event;
+		if (cur_string->num_of_params == 0) /* trace with no params */
+			list_add_tail(&cur_string->list, &tracer->ready_strings_list);
+	} else {
+		cur_string = mlx5_tracer_message_get(tracer, tracer_event);
+		if (!cur_string) {
+			pr_debug("%s Got string event for unknown string tdsm: %d\n",
+				 __func__, tracer_event->string_event.tmsn);
+			return -1;
+		}
+		cur_string->last_param_num += 1;
+		if (cur_string->last_param_num > TRACER_MAX_PARAMS) {
+			pr_debug("%s Number of params exceeds the max (%d)\n",
+				 __func__, TRACER_MAX_PARAMS);
+			list_add_tail(&cur_string->list, &tracer->ready_strings_list);
+			return 0;
+		}
+		/* keep the new parameter */
+		cur_string->params[cur_string->last_param_num - 1] =
+			tracer_event->string_event.string_param;
+		if (cur_string->last_param_num == cur_string->num_of_params)
+			list_add_tail(&cur_string->list, &tracer->ready_strings_list);
+	}
+
+	return 0;
+}
+
+static void mlx5_tracer_handle_timestamp_trace(struct mlx5_fw_tracer *tracer,
+					       struct tracer_event *tracer_event)
+{
+	struct tracer_timestamp_event timestamp_event =
+						tracer_event->timestamp_event;
+	struct tracer_string_format *str_frmt, *tmp_str;
+	struct mlx5_core_dev *dev = tracer->dev;
+	u64 trace_timestamp;
+
+	list_for_each_entry_safe(str_frmt, tmp_str, &tracer->ready_strings_list, list) {
+		list_del(&str_frmt->list);
+		if (str_frmt->timestamp < (timestamp_event.timestamp & MASK_6_0))
+			trace_timestamp = (timestamp_event.timestamp & MASK_52_7) |
+					  (str_frmt->timestamp & MASK_6_0);
+		else
+			trace_timestamp = ((timestamp_event.timestamp & MASK_52_7) - 1) |
+					  (str_frmt->timestamp & MASK_6_0);
+
+		mlx5_tracer_print_trace(str_frmt, dev, trace_timestamp);
+	}
+}
+
+static int mlx5_tracer_handle_trace(struct mlx5_fw_tracer *tracer,
+				    struct tracer_event *tracer_event)
+{
+	if (tracer_event->type == TRACER_EVENT_TYPE_STRING) {
+		mlx5_tracer_handle_string_trace(tracer, tracer_event);
+	} else if (tracer_event->type == TRACER_EVENT_TYPE_TIMESTAMP) {
+		if (!tracer_event->timestamp_event.unreliable)
+			mlx5_tracer_handle_timestamp_trace(tracer, tracer_event);
+	} else {
+		pr_debug("%s Got unrecognised type %d for parsing, exiting..\n",
+			 __func__, tracer_event->type);
+	}
+	return 0;
+}
+
+static void mlx5_fw_tracer_handle_traces(struct work_struct *work)
+{
+	struct mlx5_fw_tracer *tracer =
+			container_of(work, struct mlx5_fw_tracer, handle_traces_work);
+	u64 block_timestamp, last_block_timestamp, tmp_trace_block[TRACES_PER_BLOCK];
+	u32 block_count, start_offset, prev_start_offset, prev_consumer_index;
+	u32 trace_event_size = MLX5_ST_SZ_BYTES(tracer_event);
+	struct mlx5_core_dev *dev = tracer->dev;
+	struct tracer_event tracer_event;
+	int i;
+
+	mlx5_core_dbg(dev, "FWTracer: Handle Trace event, owner=(%d)\n", tracer->owner);
+	if (!tracer->owner)
+		return;
+
+	if (unlikely(!tracer->str_db.loaded))
+		goto arm;
+
+	block_count = tracer->buff.size / TRACER_BLOCK_SIZE_BYTE;
+	start_offset = tracer->buff.consumer_index * TRACER_BLOCK_SIZE_BYTE;
+
+	/* Copy the block to local buffer to avoid HW override while being processed*/
+	memcpy(tmp_trace_block, tracer->buff.log_buf + start_offset,
+	       TRACER_BLOCK_SIZE_BYTE);
+
+	block_timestamp =
+		get_block_timestamp(tracer, &tmp_trace_block[TRACES_PER_BLOCK - 1]);
+
+	while (block_timestamp > tracer->last_timestamp) {
+		/* Check block override if its not the first block */
+		if (!tracer->last_timestamp) {
+			u64 *ts_event;
+			/* To avoid block override be the HW in case of buffer
+			 * wraparound, the time stamp of the previous block
+			 * should be compared to the last timestamp handled
+			 * by the driver.
+			 */
+			prev_consumer_index =
+				(tracer->buff.consumer_index - 1) & (block_count - 1);
+			prev_start_offset = prev_consumer_index * TRACER_BLOCK_SIZE_BYTE;
+
+			ts_event = tracer->buff.log_buf + prev_start_offset +
+				   (TRACES_PER_BLOCK - 1) * trace_event_size;
+			last_block_timestamp = get_block_timestamp(tracer, ts_event);
+			/* If previous timestamp different from last stored
+			 * timestamp then there is a good chance that the
+			 * current buffer is overwritten and therefore should
+			 * not be parsed.
+			 */
+			if (tracer->last_timestamp != last_block_timestamp) {
+				mlx5_core_warn(dev, "FWTracer: Events were lost\n");
+				tracer->last_timestamp = block_timestamp;
+				tracer->buff.consumer_index =
+					(tracer->buff.consumer_index + 1) & (block_count - 1);
+				break;
+			}
+		}
+
+		/* Parse events */
+		for (i = 0; i < TRACES_PER_BLOCK ; i++) {
+			poll_trace(tracer, &tracer_event, &tmp_trace_block[i]);
+			mlx5_tracer_handle_trace(tracer, &tracer_event);
+		}
+
+		tracer->buff.consumer_index =
+			(tracer->buff.consumer_index + 1) & (block_count - 1);
+
+		tracer->last_timestamp = block_timestamp;
+		start_offset = tracer->buff.consumer_index * TRACER_BLOCK_SIZE_BYTE;
+		memcpy(tmp_trace_block, tracer->buff.log_buf + start_offset,
+		       TRACER_BLOCK_SIZE_BYTE);
+		block_timestamp = get_block_timestamp(tracer,
+						      &tmp_trace_block[TRACES_PER_BLOCK - 1]);
+	}
+
+arm:
+	mlx5_fw_tracer_arm(dev);
+}
+
+static int mlx5_fw_tracer_set_mtrc_conf(struct mlx5_fw_tracer *tracer)
+{
+	struct mlx5_core_dev *dev = tracer->dev;
+	u32 out[MLX5_ST_SZ_DW(mtrc_conf)] = {0};
+	u32 in[MLX5_ST_SZ_DW(mtrc_conf)] = {0};
+	int err;
+
+	MLX5_SET(mtrc_conf, in, trace_mode, TRACE_TO_MEMORY);
+	MLX5_SET(mtrc_conf, in, log_trace_buffer_size,
+		 ilog2(TRACER_BUFFER_PAGE_NUM));
+	MLX5_SET(mtrc_conf, in, trace_mkey, tracer->buff.mkey.key);
+
+	err = mlx5_core_access_reg(dev, in, sizeof(in), out, sizeof(out),
+				   MLX5_REG_MTRC_CONF, 0, 1);
+	if (err)
+		mlx5_core_warn(dev, "FWTracer: Failed to set tracer configurations %d\n", err);
+
+	return err;
+}
+
+static int mlx5_fw_tracer_set_mtrc_ctrl(struct mlx5_fw_tracer *tracer, u8 status, u8 arm)
+{
+	struct mlx5_core_dev *dev = tracer->dev;
+	u32 out[MLX5_ST_SZ_DW(mtrc_ctrl)] = {0};
+	u32 in[MLX5_ST_SZ_DW(mtrc_ctrl)] = {0};
+	int err;
+
+	MLX5_SET(mtrc_ctrl, in, modify_field_select, TRACE_STATUS);
+	MLX5_SET(mtrc_ctrl, in, trace_status, status);
+	MLX5_SET(mtrc_ctrl, in, arm_event, arm);
+
+	err = mlx5_core_access_reg(dev, in, sizeof(in), out, sizeof(out),
+				   MLX5_REG_MTRC_CTRL, 0, 1);
+
+	if (!err && status)
+		tracer->last_timestamp = 0;
+
+	return err;
+}
+
+static int mlx5_fw_tracer_start(struct mlx5_fw_tracer *tracer)
+{
+	struct mlx5_core_dev *dev = tracer->dev;
+	int err;
+
+	err = mlx5_fw_tracer_ownership_acquire(tracer);
+	if (err) {
+		mlx5_core_dbg(dev, "FWTracer: Ownership was not granted %d\n", err);
+		/* Don't fail since ownership can be acquired on a later FW event */
+		return 0;
+	}
+
+	err = mlx5_fw_tracer_set_mtrc_conf(tracer);
+	if (err) {
+		mlx5_core_warn(dev, "FWTracer: Failed to set tracer configuration %d\n", err);
+		goto release_ownership;
+	}
+
+	/* enable tracer & trace events */
+	err = mlx5_fw_tracer_set_mtrc_ctrl(tracer, 1, 1);
+	if (err) {
+		mlx5_core_warn(dev, "FWTracer: Failed to enable tracer %d\n", err);
+		goto release_ownership;
+	}
+
+	mlx5_core_dbg(dev, "FWTracer: Ownership granted and active\n");
+	return 0;
+
+release_ownership:
+	mlx5_fw_tracer_ownership_release(tracer);
+	return err;
+}
+
+static void mlx5_fw_tracer_ownership_change(struct work_struct *work)
+{
+	struct mlx5_fw_tracer *tracer =
+		container_of(work, struct mlx5_fw_tracer, ownership_change_work);
+
+	mlx5_core_dbg(tracer->dev, "FWTracer: ownership changed, current=(%d)\n", tracer->owner);
+	if (tracer->owner) {
+		tracer->owner = false;
+		tracer->buff.consumer_index = 0;
+		return;
+	}
+
+	mlx5_fw_tracer_start(tracer);
+}
+
+/* Create software resources (Buffers, etc ..) */
+struct mlx5_fw_tracer *mlx5_fw_tracer_create(struct mlx5_core_dev *dev)
+{
+	struct mlx5_fw_tracer *tracer = NULL;
+	int err;
+
+	if (!MLX5_CAP_MCAM_REG(dev, tracer_registers)) {
+		mlx5_core_dbg(dev, "FWTracer: Tracer capability not present\n");
+		return NULL;
+	}
+
+	tracer = kvzalloc(sizeof(*tracer), GFP_KERNEL);
+	if (!tracer)
+		return ERR_PTR(-ENOMEM);
+
+	tracer->work_queue = create_singlethread_workqueue("mlx5_fw_tracer");
+	if (!tracer->work_queue) {
+		err = -ENOMEM;
+		goto free_tracer;
+	}
+
+	tracer->dev = dev;
+
+	INIT_LIST_HEAD(&tracer->ready_strings_list);
+	INIT_WORK(&tracer->ownership_change_work, mlx5_fw_tracer_ownership_change);
+	INIT_WORK(&tracer->read_fw_strings_work, mlx5_tracer_read_strings_db);
+	INIT_WORK(&tracer->handle_traces_work, mlx5_fw_tracer_handle_traces);
+
+
+	err = mlx5_query_mtrc_caps(tracer);
+	if (err) {
+		mlx5_core_dbg(dev, "FWTracer: Failed to query capabilities %d\n", err);
+		goto destroy_workqueue;
+	}
+
+	err = mlx5_fw_tracer_create_log_buf(tracer);
+	if (err) {
+		mlx5_core_warn(dev, "FWTracer: Create log buffer failed %d\n", err);
+		goto destroy_workqueue;
+	}
+
+	err = mlx5_fw_tracer_allocate_strings_db(tracer);
+	if (err) {
+		mlx5_core_warn(dev, "FWTracer: Allocate strings database failed %d\n", err);
+		goto free_log_buf;
+	}
+
+	mlx5_core_dbg(dev, "FWTracer: Tracer created\n");
+
+	return tracer;
+
+free_log_buf:
+	mlx5_fw_tracer_destroy_log_buf(tracer);
+destroy_workqueue:
+	tracer->dev = NULL;
+	destroy_workqueue(tracer->work_queue);
+free_tracer:
+	kvfree(tracer);
+	return ERR_PTR(err);
+}
+
+/* Create HW resources + start tracer
+ * must be called before Async EQ is created
+ */
+int mlx5_fw_tracer_init(struct mlx5_fw_tracer *tracer)
+{
+	struct mlx5_core_dev *dev;
+	int err;
+
+	if (IS_ERR_OR_NULL(tracer))
+		return 0;
+
+	dev = tracer->dev;
+
+	if (!tracer->str_db.loaded)
+		queue_work(tracer->work_queue, &tracer->read_fw_strings_work);
+
+	err = mlx5_core_alloc_pd(dev, &tracer->buff.pdn);
+	if (err) {
+		mlx5_core_warn(dev, "FWTracer: Failed to allocate PD %d\n", err);
+		return err;
+	}
+
+	err = mlx5_fw_tracer_create_mkey(tracer);
+	if (err) {
+		mlx5_core_warn(dev, "FWTracer: Failed to create mkey %d\n", err);
+		goto err_dealloc_pd;
+	}
+
+	mlx5_fw_tracer_start(tracer);
+
+	return 0;
+
+err_dealloc_pd:
+	mlx5_core_dealloc_pd(dev, tracer->buff.pdn);
+	return err;
+}
+
+/* Stop tracer + Cleanup HW resources
+ * must be called after Async EQ is destroyed
+ */
+void mlx5_fw_tracer_cleanup(struct mlx5_fw_tracer *tracer)
+{
+	if (IS_ERR_OR_NULL(tracer))
+		return;
+
+	mlx5_core_dbg(tracer->dev, "FWTracer: Cleanup, is owner ? (%d)\n",
+		      tracer->owner);
+
+	cancel_work_sync(&tracer->ownership_change_work);
+	cancel_work_sync(&tracer->handle_traces_work);
+
+	if (tracer->owner)
+		mlx5_fw_tracer_ownership_release(tracer);
+
+	mlx5_core_destroy_mkey(tracer->dev, &tracer->buff.mkey);
+	mlx5_core_dealloc_pd(tracer->dev, tracer->buff.pdn);
+}
+
+/* Free software resources (Buffers, etc ..) */
+void mlx5_fw_tracer_destroy(struct mlx5_fw_tracer *tracer)
+{
+	if (IS_ERR_OR_NULL(tracer))
+		return;
+
+	mlx5_core_dbg(tracer->dev, "FWTracer: Destroy\n");
+
+	cancel_work_sync(&tracer->read_fw_strings_work);
+	mlx5_fw_tracer_clean_ready_list(tracer);
+	mlx5_fw_tracer_clean_print_hash(tracer);
+	mlx5_fw_tracer_free_strings_db(tracer);
+	mlx5_fw_tracer_destroy_log_buf(tracer);
+	flush_workqueue(tracer->work_queue);
+	destroy_workqueue(tracer->work_queue);
+	kvfree(tracer);
+}
+
+void mlx5_fw_tracer_event(struct mlx5_core_dev *dev, struct mlx5_eqe *eqe)
+{
+	struct mlx5_fw_tracer *tracer = dev->tracer;
+
+	if (!tracer)
+		return;
+
+	switch (eqe->sub_type) {
+	case MLX5_TRACER_SUBTYPE_OWNERSHIP_CHANGE:
+		if (test_bit(MLX5_INTERFACE_STATE_UP, &dev->intf_state))
+			queue_work(tracer->work_queue, &tracer->ownership_change_work);
+		break;
+	case MLX5_TRACER_SUBTYPE_TRACES_AVAILABLE:
+		queue_work(tracer->work_queue, &tracer->handle_traces_work);
+		break;
+	default:
+		mlx5_core_dbg(dev, "FWTracer: Event with unrecognized subtype: sub_type %d\n",
+			      eqe->sub_type);
+	}
+}
+
+EXPORT_TRACEPOINT_SYMBOL(mlx5_fw);
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/diag/fw_tracer.h b/drivers/net/ethernet/mellanox/mlx5/core/diag/fw_tracer.h
new file mode 100644
index 000000000..0347f2dd5
--- /dev/null
+++ b/drivers/net/ethernet/mellanox/mlx5/core/diag/fw_tracer.h
@@ -0,0 +1,175 @@
+/*
+ * 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.
+ */
+
+#ifndef __LIB_TRACER_H__
+#define __LIB_TRACER_H__
+
+#include <linux/mlx5/driver.h>
+#include "mlx5_core.h"
+
+#define STRINGS_DB_SECTIONS_NUM 8
+#define STRINGS_DB_READ_SIZE_BYTES 256
+#define STRINGS_DB_LEFTOVER_SIZE_BYTES 64
+#define TRACER_BUFFER_PAGE_NUM 64
+#define TRACER_BUFFER_CHUNK 4096
+#define TRACE_BUFFER_SIZE_BYTE (TRACER_BUFFER_PAGE_NUM * TRACER_BUFFER_CHUNK)
+
+#define TRACER_BLOCK_SIZE_BYTE 256
+#define TRACES_PER_BLOCK 32
+
+#define TRACER_MAX_PARAMS 7
+#define MESSAGE_HASH_BITS 6
+#define MESSAGE_HASH_SIZE BIT(MESSAGE_HASH_BITS)
+
+#define MASK_52_7 (0x1FFFFFFFFFFF80)
+#define MASK_6_0  (0x7F)
+
+struct mlx5_fw_tracer {
+	struct mlx5_core_dev *dev;
+	bool owner;
+	u8   trc_ver;
+	struct workqueue_struct *work_queue;
+	struct work_struct ownership_change_work;
+	struct work_struct read_fw_strings_work;
+
+	/* Strings DB */
+	struct {
+		u8 first_string_trace;
+		u8 num_string_trace;
+		u32 num_string_db;
+		u32 base_address_out[STRINGS_DB_SECTIONS_NUM];
+		u32 size_out[STRINGS_DB_SECTIONS_NUM];
+		void *buffer[STRINGS_DB_SECTIONS_NUM];
+		bool loaded;
+	} str_db;
+
+	/* Log Buffer */
+	struct {
+		u32 pdn;
+		void *log_buf;
+		dma_addr_t dma;
+		u32 size;
+		struct mlx5_core_mkey mkey;
+		u32 consumer_index;
+	} buff;
+
+	u64 last_timestamp;
+	struct work_struct handle_traces_work;
+	struct hlist_head hash[MESSAGE_HASH_SIZE];
+	struct list_head ready_strings_list;
+};
+
+struct tracer_string_format {
+	char *string;
+	int params[TRACER_MAX_PARAMS];
+	int num_of_params;
+	int last_param_num;
+	u8 event_id;
+	u32 tmsn;
+	struct hlist_node hlist;
+	struct list_head list;
+	u32 timestamp;
+	bool lost;
+};
+
+enum mlx5_fw_tracer_ownership_state {
+	MLX5_FW_TRACER_RELEASE_OWNERSHIP,
+	MLX5_FW_TRACER_ACQUIRE_OWNERSHIP,
+};
+
+enum tracer_ctrl_fields_select {
+	TRACE_STATUS = 1 << 0,
+};
+
+enum tracer_event_type {
+	TRACER_EVENT_TYPE_STRING,
+	TRACER_EVENT_TYPE_TIMESTAMP = 0xFF,
+	TRACER_EVENT_TYPE_UNRECOGNIZED,
+};
+
+enum tracing_mode {
+	TRACE_TO_MEMORY = 1 << 0,
+};
+
+struct tracer_timestamp_event {
+	u64        timestamp;
+	u8         unreliable;
+};
+
+struct tracer_string_event {
+	u32        timestamp;
+	u32        tmsn;
+	u32        tdsn;
+	u32        string_param;
+};
+
+struct tracer_event {
+	bool      lost_event;
+	u32       type;
+	u8        event_id;
+	union {
+		struct tracer_string_event string_event;
+		struct tracer_timestamp_event timestamp_event;
+	};
+};
+
+struct mlx5_ifc_tracer_event_bits {
+	u8         lost[0x1];
+	u8         timestamp[0x7];
+	u8         event_id[0x8];
+	u8         event_data[0x30];
+};
+
+struct mlx5_ifc_tracer_string_event_bits {
+	u8         lost[0x1];
+	u8         timestamp[0x7];
+	u8         event_id[0x8];
+	u8         tmsn[0xd];
+	u8         tdsn[0x3];
+	u8         string_param[0x20];
+};
+
+struct mlx5_ifc_tracer_timestamp_event_bits {
+	u8         timestamp7_0[0x8];
+	u8         event_id[0x8];
+	u8         urts[0x3];
+	u8         timestamp52_40[0xd];
+	u8         timestamp39_8[0x20];
+};
+
+struct mlx5_fw_tracer *mlx5_fw_tracer_create(struct mlx5_core_dev *dev);
+int mlx5_fw_tracer_init(struct mlx5_fw_tracer *tracer);
+void mlx5_fw_tracer_cleanup(struct mlx5_fw_tracer *tracer);
+void mlx5_fw_tracer_destroy(struct mlx5_fw_tracer *tracer);
+void mlx5_fw_tracer_event(struct mlx5_core_dev *dev, struct mlx5_eqe *eqe);
+
+#endif
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/diag/fw_tracer_tracepoint.h b/drivers/net/ethernet/mellanox/mlx5/core/diag/fw_tracer_tracepoint.h
new file mode 100644
index 000000000..83f90e9af
--- /dev/null
+++ b/drivers/net/ethernet/mellanox/mlx5/core/diag/fw_tracer_tracepoint.h
@@ -0,0 +1,78 @@
+/*
+ * 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.
+ */
+
+#if !defined(__LIB_TRACER_TRACEPOINT_H__) || defined(TRACE_HEADER_MULTI_READ)
+#define __LIB_TRACER_TRACEPOINT_H__
+
+#include <linux/tracepoint.h>
+#include "fw_tracer.h"
+
+#undef TRACE_SYSTEM
+#define TRACE_SYSTEM mlx5
+
+/* Tracepoint for FWTracer messages: */
+TRACE_EVENT(mlx5_fw,
+	TP_PROTO(const struct mlx5_fw_tracer *tracer, u64 trace_timestamp,
+		 bool lost, u8 event_id, const char *msg),
+
+	TP_ARGS(tracer, trace_timestamp, lost, event_id, msg),
+
+	TP_STRUCT__entry(
+		__string(dev_name, dev_name(&tracer->dev->pdev->dev))
+		__field(u64, trace_timestamp)
+		__field(bool, lost)
+		__field(u8, event_id)
+		__string(msg, msg)
+	),
+
+	TP_fast_assign(
+		__assign_str(dev_name, dev_name(&tracer->dev->pdev->dev));
+		__entry->trace_timestamp = trace_timestamp;
+		__entry->lost = lost;
+		__entry->event_id = event_id;
+		__assign_str(msg, msg);
+	),
+
+	TP_printk("%s [0x%llx] %d [0x%x] %s",
+		  __get_str(dev_name),
+		  __entry->trace_timestamp,
+		  __entry->lost, __entry->event_id,
+		  __get_str(msg))
+);
+
+#endif
+
+#undef TRACE_INCLUDE_PATH
+#undef TRACE_INCLUDE_FILE
+#define TRACE_INCLUDE_PATH ./diag
+#define TRACE_INCLUDE_FILE fw_tracer_tracepoint
+#include <trace/define_trace.h>
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en.h b/drivers/net/ethernet/mellanox/mlx5/core/en.h
new file mode 100644
index 000000000..d79e177f8
--- /dev/null
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en.h
@@ -0,0 +1,979 @@
+/*
+ * Copyright (c) 2015-2016, 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 __MLX5_EN_H__
+#define __MLX5_EN_H__
+
+#include <linux/if_vlan.h>
+#include <linux/etherdevice.h>
+#include <linux/timecounter.h>
+#include <linux/net_tstamp.h>
+#include <linux/ptp_clock_kernel.h>
+#include <linux/crash_dump.h>
+#include <linux/mlx5/driver.h>
+#include <linux/mlx5/qp.h>
+#include <linux/mlx5/cq.h>
+#include <linux/mlx5/port.h>
+#include <linux/mlx5/vport.h>
+#include <linux/mlx5/transobj.h>
+#include <linux/mlx5/fs.h>
+#include <linux/rhashtable.h>
+#include <net/switchdev.h>
+#include <net/xdp.h>
+#include <linux/net_dim.h>
+#include "wq.h"
+#include "mlx5_core.h"
+#include "en_stats.h"
+#include "en/fs.h"
+
+extern const struct net_device_ops mlx5e_netdev_ops;
+struct page_pool;
+
+#define MLX5E_METADATA_ETHER_TYPE (0x8CE4)
+#define MLX5E_METADATA_ETHER_LEN 8
+
+#define MLX5_SET_CFG(p, f, v) MLX5_SET(create_flow_group_in, p, f, v)
+
+#define MLX5E_ETH_HARD_MTU (ETH_HLEN + VLAN_HLEN + ETH_FCS_LEN)
+
+#define MLX5E_HW2SW_MTU(params, hwmtu) ((hwmtu) - ((params)->hard_mtu))
+#define MLX5E_SW2HW_MTU(params, swmtu) ((swmtu) + ((params)->hard_mtu))
+
+#define MLX5E_MAX_PRIORITY      8
+#define MLX5E_MAX_DSCP          64
+#define MLX5E_MAX_NUM_TC	8
+
+#define MLX5_RX_HEADROOM NET_SKB_PAD
+#define MLX5_SKB_FRAG_SZ(len)	(SKB_DATA_ALIGN(len) +	\
+				 SKB_DATA_ALIGN(sizeof(struct skb_shared_info)))
+
+#define MLX5_MPWRQ_MIN_LOG_STRIDE_SZ(mdev) \
+	(6 + MLX5_CAP_GEN(mdev, cache_line_128byte)) /* HW restriction */
+#define MLX5_MPWRQ_LOG_STRIDE_SZ(mdev, req) \
+	max_t(u32, MLX5_MPWRQ_MIN_LOG_STRIDE_SZ(mdev), req)
+#define MLX5_MPWRQ_DEF_LOG_STRIDE_SZ(mdev)       MLX5_MPWRQ_LOG_STRIDE_SZ(mdev, 6)
+#define MLX5_MPWRQ_CQE_CMPRS_LOG_STRIDE_SZ(mdev) MLX5_MPWRQ_LOG_STRIDE_SZ(mdev, 8)
+#define MLX5E_MPWQE_STRIDE_SZ(mdev, cqe_cmprs) \
+	(cqe_cmprs ? MLX5_MPWRQ_CQE_CMPRS_LOG_STRIDE_SZ(mdev) : \
+	MLX5_MPWRQ_DEF_LOG_STRIDE_SZ(mdev))
+
+#define MLX5_MPWRQ_LOG_WQE_SZ			18
+#define MLX5_MPWRQ_WQE_PAGE_ORDER  (MLX5_MPWRQ_LOG_WQE_SZ - PAGE_SHIFT > 0 ? \
+				    MLX5_MPWRQ_LOG_WQE_SZ - PAGE_SHIFT : 0)
+#define MLX5_MPWRQ_PAGES_PER_WQE		BIT(MLX5_MPWRQ_WQE_PAGE_ORDER)
+
+#define MLX5_MTT_OCTW(npages) (ALIGN(npages, 8) / 2)
+#define MLX5E_REQUIRED_WQE_MTTS		(ALIGN(MLX5_MPWRQ_PAGES_PER_WQE, 8))
+#define MLX5E_LOG_ALIGNED_MPWQE_PPW	(ilog2(MLX5E_REQUIRED_WQE_MTTS))
+#define MLX5E_REQUIRED_MTTS(wqes)	(wqes * MLX5E_REQUIRED_WQE_MTTS)
+#define MLX5E_MAX_RQ_NUM_MTTS	\
+	((1 << 16) * 2) /* So that MLX5_MTT_OCTW(num_mtts) fits into u16 */
+#define MLX5E_ORDER2_MAX_PACKET_MTU (order_base_2(10 * 1024))
+#define MLX5E_PARAMS_MAXIMUM_LOG_RQ_SIZE_MPW	\
+		(ilog2(MLX5E_MAX_RQ_NUM_MTTS / MLX5E_REQUIRED_WQE_MTTS))
+#define MLX5E_LOG_MAX_RQ_NUM_PACKETS_MPW \
+	(MLX5E_PARAMS_MAXIMUM_LOG_RQ_SIZE_MPW + \
+	 (MLX5_MPWRQ_LOG_WQE_SZ - MLX5E_ORDER2_MAX_PACKET_MTU))
+
+#define MLX5E_MIN_SKB_FRAG_SZ		(MLX5_SKB_FRAG_SZ(MLX5_RX_HEADROOM))
+#define MLX5E_LOG_MAX_RX_WQE_BULK	\
+	(ilog2(PAGE_SIZE / roundup_pow_of_two(MLX5E_MIN_SKB_FRAG_SZ)))
+
+#define MLX5E_PARAMS_MINIMUM_LOG_SQ_SIZE                0x6
+#define MLX5E_PARAMS_DEFAULT_LOG_SQ_SIZE                0xa
+#define MLX5E_PARAMS_MAXIMUM_LOG_SQ_SIZE                0xd
+
+#define MLX5E_PARAMS_MINIMUM_LOG_RQ_SIZE (1 + MLX5E_LOG_MAX_RX_WQE_BULK)
+#define MLX5E_PARAMS_DEFAULT_LOG_RQ_SIZE                0xa
+#define MLX5E_PARAMS_MAXIMUM_LOG_RQ_SIZE min_t(u8, 0xd,	\
+					       MLX5E_LOG_MAX_RQ_NUM_PACKETS_MPW)
+
+#define MLX5E_PARAMS_MINIMUM_LOG_RQ_SIZE_MPW            0x2
+
+#define MLX5E_RX_MAX_HEAD (256)
+
+#define MLX5E_PARAMS_DEFAULT_LRO_WQE_SZ                 (64 * 1024)
+#define MLX5E_DEFAULT_LRO_TIMEOUT                       32
+#define MLX5E_LRO_TIMEOUT_ARR_SIZE                      4
+
+#define MLX5E_PARAMS_DEFAULT_RX_CQ_MODERATION_USEC      0x10
+#define MLX5E_PARAMS_DEFAULT_RX_CQ_MODERATION_USEC_FROM_CQE 0x3
+#define MLX5E_PARAMS_DEFAULT_RX_CQ_MODERATION_PKTS      0x20
+#define MLX5E_PARAMS_DEFAULT_TX_CQ_MODERATION_USEC      0x10
+#define MLX5E_PARAMS_DEFAULT_TX_CQ_MODERATION_USEC_FROM_CQE 0x10
+#define MLX5E_PARAMS_DEFAULT_TX_CQ_MODERATION_PKTS      0x20
+#define MLX5E_PARAMS_DEFAULT_MIN_RX_WQES                0x80
+#define MLX5E_PARAMS_DEFAULT_MIN_RX_WQES_MPW            0x2
+
+#define MLX5E_LOG_INDIR_RQT_SIZE       0x7
+#define MLX5E_INDIR_RQT_SIZE           BIT(MLX5E_LOG_INDIR_RQT_SIZE)
+#define MLX5E_MIN_NUM_CHANNELS         0x1
+#define MLX5E_MAX_NUM_CHANNELS         (MLX5E_INDIR_RQT_SIZE >> 1)
+#define MLX5E_MAX_NUM_SQS              (MLX5E_MAX_NUM_CHANNELS * MLX5E_MAX_NUM_TC)
+#define MLX5E_TX_CQ_POLL_BUDGET        128
+#define MLX5E_SQ_RECOVER_MIN_INTERVAL  500 /* msecs */
+
+#define MLX5E_UMR_WQE_INLINE_SZ \
+	(sizeof(struct mlx5e_umr_wqe) + \
+	 ALIGN(MLX5_MPWRQ_PAGES_PER_WQE * sizeof(struct mlx5_mtt), \
+	       MLX5_UMR_MTT_ALIGNMENT))
+#define MLX5E_UMR_WQEBBS \
+	(DIV_ROUND_UP(MLX5E_UMR_WQE_INLINE_SZ, MLX5_SEND_WQE_BB))
+#define MLX5E_ICOSQ_MAX_WQEBBS MLX5E_UMR_WQEBBS
+
+#define MLX5E_NUM_MAIN_GROUPS 9
+
+#define MLX5E_MSG_LEVEL			NETIF_MSG_LINK
+
+#define mlx5e_dbg(mlevel, priv, format, ...)                    \
+do {                                                            \
+	if (NETIF_MSG_##mlevel & (priv)->msglevel)              \
+		netdev_warn(priv->netdev, format,               \
+			    ##__VA_ARGS__);                     \
+} while (0)
+
+
+static inline u16 mlx5_min_rx_wqes(int wq_type, u32 wq_size)
+{
+	switch (wq_type) {
+	case MLX5_WQ_TYPE_LINKED_LIST_STRIDING_RQ:
+		return min_t(u16, MLX5E_PARAMS_DEFAULT_MIN_RX_WQES_MPW,
+			     wq_size / 2);
+	default:
+		return min_t(u16, MLX5E_PARAMS_DEFAULT_MIN_RX_WQES,
+			     wq_size / 2);
+	}
+}
+
+static inline int mlx5e_get_max_num_channels(struct mlx5_core_dev *mdev)
+{
+	return is_kdump_kernel() ?
+		MLX5E_MIN_NUM_CHANNELS :
+		min_t(int, mdev->priv.eq_table.num_comp_vectors,
+		      MLX5E_MAX_NUM_CHANNELS);
+}
+
+struct mlx5e_tx_wqe {
+	struct mlx5_wqe_ctrl_seg ctrl;
+	struct mlx5_wqe_eth_seg  eth;
+	struct mlx5_wqe_data_seg data[0];
+};
+
+struct mlx5e_rx_wqe_ll {
+	struct mlx5_wqe_srq_next_seg  next;
+	struct mlx5_wqe_data_seg      data[0];
+};
+
+struct mlx5e_rx_wqe_cyc {
+	struct mlx5_wqe_data_seg      data[0];
+};
+
+struct mlx5e_umr_wqe {
+	struct mlx5_wqe_ctrl_seg       ctrl;
+	struct mlx5_wqe_umr_ctrl_seg   uctrl;
+	struct mlx5_mkey_seg           mkc;
+	struct mlx5_mtt                inline_mtts[0];
+};
+
+extern const char mlx5e_self_tests[][ETH_GSTRING_LEN];
+
+static const char mlx5e_priv_flags[][ETH_GSTRING_LEN] = {
+	"rx_cqe_moder",
+	"tx_cqe_moder",
+	"rx_cqe_compress",
+	"rx_striding_rq",
+	"rx_no_csum_complete",
+};
+
+enum mlx5e_priv_flag {
+	MLX5E_PFLAG_RX_CQE_BASED_MODER = (1 << 0),
+	MLX5E_PFLAG_TX_CQE_BASED_MODER = (1 << 1),
+	MLX5E_PFLAG_RX_CQE_COMPRESS = (1 << 2),
+	MLX5E_PFLAG_RX_STRIDING_RQ = (1 << 3),
+	MLX5E_PFLAG_RX_NO_CSUM_COMPLETE = (1 << 4),
+};
+
+#define MLX5E_SET_PFLAG(params, pflag, enable)			\
+	do {							\
+		if (enable)					\
+			(params)->pflags |= (pflag);		\
+		else						\
+			(params)->pflags &= ~(pflag);		\
+	} while (0)
+
+#define MLX5E_GET_PFLAG(params, pflag) (!!((params)->pflags & (pflag)))
+
+#ifdef CONFIG_MLX5_CORE_EN_DCB
+#define MLX5E_MAX_BW_ALLOC 100 /* Max percentage of BW allocation */
+#endif
+
+struct mlx5e_params {
+	u8  log_sq_size;
+	u8  rq_wq_type;
+	u8  log_rq_mtu_frames;
+	u16 num_channels;
+	u8  num_tc;
+	bool rx_cqe_compress_def;
+	struct net_dim_cq_moder rx_cq_moderation;
+	struct net_dim_cq_moder tx_cq_moderation;
+	bool lro_en;
+	u32 lro_wqe_sz;
+	u8  tx_min_inline_mode;
+	u8  rss_hfunc;
+	u8  toeplitz_hash_key[40];
+	u32 indirection_rqt[MLX5E_INDIR_RQT_SIZE];
+	bool vlan_strip_disable;
+	bool scatter_fcs_en;
+	bool rx_dim_enabled;
+	bool tx_dim_enabled;
+	u32 lro_timeout;
+	u32 pflags;
+	struct bpf_prog *xdp_prog;
+	unsigned int sw_mtu;
+	int hard_mtu;
+};
+
+#ifdef CONFIG_MLX5_CORE_EN_DCB
+struct mlx5e_cee_config {
+	/* bw pct for priority group */
+	u8                         pg_bw_pct[CEE_DCBX_MAX_PGS];
+	u8                         prio_to_pg_map[CEE_DCBX_MAX_PRIO];
+	bool                       pfc_setting[CEE_DCBX_MAX_PRIO];
+	bool                       pfc_enable;
+};
+
+enum {
+	MLX5_DCB_CHG_RESET,
+	MLX5_DCB_NO_CHG,
+	MLX5_DCB_CHG_NO_RESET,
+};
+
+struct mlx5e_dcbx {
+	enum mlx5_dcbx_oper_mode   mode;
+	struct mlx5e_cee_config    cee_cfg; /* pending configuration */
+	u8                         dscp_app_cnt;
+
+	/* The only setting that cannot be read from FW */
+	u8                         tc_tsa[IEEE_8021QAZ_MAX_TCS];
+	u8                         cap;
+
+	/* Buffer configuration */
+	bool                       manual_buffer;
+	u32                        cable_len;
+	u32                        xoff;
+};
+
+struct mlx5e_dcbx_dp {
+	u8                         dscp2prio[MLX5E_MAX_DSCP];
+	u8                         trust_state;
+};
+#endif
+
+enum {
+	MLX5E_RQ_STATE_ENABLED,
+	MLX5E_RQ_STATE_AM,
+	MLX5E_RQ_STATE_NO_CSUM_COMPLETE,
+};
+
+struct mlx5e_cq {
+	/* data path - accessed per cqe */
+	struct mlx5_cqwq           wq;
+
+	/* data path - accessed per napi poll */
+	u16                        event_ctr;
+	struct napi_struct        *napi;
+	struct mlx5_core_cq        mcq;
+	struct mlx5e_channel      *channel;
+
+	/* cqe decompression */
+	struct mlx5_cqe64          title;
+	struct mlx5_mini_cqe8      mini_arr[MLX5_MINI_CQE_ARRAY_SIZE];
+	u8                         mini_arr_idx;
+	u16                        decmprs_left;
+	u16                        decmprs_wqe_counter;
+
+	/* control */
+	struct mlx5_core_dev      *mdev;
+	struct mlx5_wq_ctrl        wq_ctrl;
+} ____cacheline_aligned_in_smp;
+
+struct mlx5e_tx_wqe_info {
+	struct sk_buff *skb;
+	u32 num_bytes;
+	u8  num_wqebbs;
+	u8  num_dma;
+};
+
+enum mlx5e_dma_map_type {
+	MLX5E_DMA_MAP_SINGLE,
+	MLX5E_DMA_MAP_PAGE
+};
+
+struct mlx5e_sq_dma {
+	dma_addr_t              addr;
+	u32                     size;
+	enum mlx5e_dma_map_type type;
+};
+
+enum {
+	MLX5E_SQ_STATE_ENABLED,
+	MLX5E_SQ_STATE_RECOVERING,
+	MLX5E_SQ_STATE_IPSEC,
+	MLX5E_SQ_STATE_AM,
+	MLX5E_SQ_STATE_TLS,
+	MLX5E_SQ_STATE_REDIRECT,
+};
+
+struct mlx5e_sq_wqe_info {
+	u8  opcode;
+};
+
+struct mlx5e_txqsq {
+	/* data path */
+
+	/* dirtied @completion */
+	u16                        cc;
+	u32                        dma_fifo_cc;
+	struct net_dim             dim; /* Adaptive Moderation */
+
+	/* dirtied @xmit */
+	u16                        pc ____cacheline_aligned_in_smp;
+	u32                        dma_fifo_pc;
+
+	struct mlx5e_cq            cq;
+
+	/* read only */
+	struct mlx5_wq_cyc         wq;
+	u32                        dma_fifo_mask;
+	struct mlx5e_sq_stats     *stats;
+	struct {
+		struct mlx5e_sq_dma       *dma_fifo;
+		struct mlx5e_tx_wqe_info  *wqe_info;
+	} db;
+	void __iomem              *uar_map;
+	struct netdev_queue       *txq;
+	u32                        sqn;
+	u8                         min_inline_mode;
+	struct device             *pdev;
+	__be32                     mkey_be;
+	unsigned long              state;
+	struct hwtstamp_config    *tstamp;
+	struct mlx5_clock         *clock;
+
+	/* control path */
+	struct mlx5_wq_ctrl        wq_ctrl;
+	struct mlx5e_channel      *channel;
+	int                        txq_ix;
+	u32                        rate_limit;
+	struct mlx5e_txqsq_recover {
+		struct work_struct         recover_work;
+		u64                        last_recover;
+	} recover;
+} ____cacheline_aligned_in_smp;
+
+struct mlx5e_dma_info {
+	struct page     *page;
+	dma_addr_t      addr;
+};
+
+struct mlx5e_xdp_info {
+	struct xdp_frame      *xdpf;
+	dma_addr_t            dma_addr;
+	struct mlx5e_dma_info di;
+};
+
+struct mlx5e_xdpsq {
+	/* data path */
+
+	/* dirtied @completion */
+	u16                        cc;
+	bool                       redirect_flush;
+
+	/* dirtied @xmit */
+	u16                        pc ____cacheline_aligned_in_smp;
+	bool                       doorbell;
+
+	struct mlx5e_cq            cq;
+
+	/* read only */
+	struct mlx5_wq_cyc         wq;
+	struct mlx5e_xdpsq_stats  *stats;
+	struct {
+		struct mlx5e_xdp_info     *xdpi;
+	} db;
+	void __iomem              *uar_map;
+	u32                        sqn;
+	struct device             *pdev;
+	__be32                     mkey_be;
+	u8                         min_inline_mode;
+	unsigned long              state;
+	unsigned int               hw_mtu;
+
+	/* control path */
+	struct mlx5_wq_ctrl        wq_ctrl;
+	struct mlx5e_channel      *channel;
+} ____cacheline_aligned_in_smp;
+
+struct mlx5e_icosq {
+	/* data path */
+
+	/* dirtied @xmit */
+	u16                        pc ____cacheline_aligned_in_smp;
+
+	struct mlx5e_cq            cq;
+
+	/* write@xmit, read@completion */
+	struct {
+		struct mlx5e_sq_wqe_info *ico_wqe;
+	} db;
+
+	/* read only */
+	struct mlx5_wq_cyc         wq;
+	void __iomem              *uar_map;
+	u32                        sqn;
+	unsigned long              state;
+
+	/* control path */
+	struct mlx5_wq_ctrl        wq_ctrl;
+	struct mlx5e_channel      *channel;
+} ____cacheline_aligned_in_smp;
+
+static inline bool
+mlx5e_wqc_has_room_for(struct mlx5_wq_cyc *wq, u16 cc, u16 pc, u16 n)
+{
+	return (mlx5_wq_cyc_ctr2ix(wq, cc - pc) >= n) || (cc == pc);
+}
+
+struct mlx5e_wqe_frag_info {
+	struct mlx5e_dma_info *di;
+	u32 offset;
+	bool last_in_page;
+};
+
+struct mlx5e_umr_dma_info {
+	struct mlx5e_dma_info  dma_info[MLX5_MPWRQ_PAGES_PER_WQE];
+};
+
+struct mlx5e_mpw_info {
+	struct mlx5e_umr_dma_info umr;
+	u16 consumed_strides;
+	DECLARE_BITMAP(xdp_xmit_bitmap, MLX5_MPWRQ_PAGES_PER_WQE);
+};
+
+#define MLX5E_MAX_RX_FRAGS 4
+
+/* a single cache unit is capable to serve one napi call (for non-striding rq)
+ * or a MPWQE (for striding rq).
+ */
+#define MLX5E_CACHE_UNIT	(MLX5_MPWRQ_PAGES_PER_WQE > NAPI_POLL_WEIGHT ? \
+				 MLX5_MPWRQ_PAGES_PER_WQE : NAPI_POLL_WEIGHT)
+#define MLX5E_CACHE_SIZE	(4 * roundup_pow_of_two(MLX5E_CACHE_UNIT))
+struct mlx5e_page_cache {
+	u32 head;
+	u32 tail;
+	struct mlx5e_dma_info page_cache[MLX5E_CACHE_SIZE];
+};
+
+struct mlx5e_rq;
+typedef void (*mlx5e_fp_handle_rx_cqe)(struct mlx5e_rq*, struct mlx5_cqe64*);
+typedef struct sk_buff *
+(*mlx5e_fp_skb_from_cqe_mpwrq)(struct mlx5e_rq *rq, struct mlx5e_mpw_info *wi,
+			       u16 cqe_bcnt, u32 head_offset, u32 page_idx);
+typedef struct sk_buff *
+(*mlx5e_fp_skb_from_cqe)(struct mlx5e_rq *rq, struct mlx5_cqe64 *cqe,
+			 struct mlx5e_wqe_frag_info *wi, u32 cqe_bcnt);
+typedef bool (*mlx5e_fp_post_rx_wqes)(struct mlx5e_rq *rq);
+typedef void (*mlx5e_fp_dealloc_wqe)(struct mlx5e_rq*, u16);
+
+enum mlx5e_rq_flag {
+	MLX5E_RQ_FLAG_XDP_XMIT = BIT(0),
+};
+
+struct mlx5e_rq_frag_info {
+	int frag_size;
+	int frag_stride;
+};
+
+struct mlx5e_rq_frags_info {
+	struct mlx5e_rq_frag_info arr[MLX5E_MAX_RX_FRAGS];
+	u8 num_frags;
+	u8 log_num_frags;
+	u8 wqe_bulk;
+};
+
+struct mlx5e_rq {
+	/* data path */
+	union {
+		struct {
+			struct mlx5_wq_cyc          wq;
+			struct mlx5e_wqe_frag_info *frags;
+			struct mlx5e_dma_info      *di;
+			struct mlx5e_rq_frags_info  info;
+			mlx5e_fp_skb_from_cqe       skb_from_cqe;
+		} wqe;
+		struct {
+			struct mlx5_wq_ll      wq;
+			struct mlx5e_umr_wqe   umr_wqe;
+			struct mlx5e_mpw_info *info;
+			mlx5e_fp_skb_from_cqe_mpwrq skb_from_cqe_mpwrq;
+			u16                    num_strides;
+			u8                     log_stride_sz;
+			bool                   umr_in_progress;
+		} mpwqe;
+	};
+	struct {
+		u16            headroom;
+		u8             map_dir;   /* dma map direction */
+	} buff;
+
+	struct mlx5e_channel  *channel;
+	struct device         *pdev;
+	struct net_device     *netdev;
+	struct mlx5e_rq_stats *stats;
+	struct mlx5e_cq        cq;
+	struct mlx5e_page_cache page_cache;
+	struct hwtstamp_config *tstamp;
+	struct mlx5_clock      *clock;
+
+	mlx5e_fp_handle_rx_cqe handle_rx_cqe;
+	mlx5e_fp_post_rx_wqes  post_wqes;
+	mlx5e_fp_dealloc_wqe   dealloc_wqe;
+
+	unsigned long          state;
+	int                    ix;
+	unsigned int           hw_mtu;
+
+	struct net_dim         dim; /* Dynamic Interrupt Moderation */
+
+	/* XDP */
+	struct bpf_prog       *xdp_prog;
+	struct mlx5e_xdpsq     xdpsq;
+	DECLARE_BITMAP(flags, 8);
+	struct page_pool      *page_pool;
+
+	/* control */
+	struct mlx5_wq_ctrl    wq_ctrl;
+	__be32                 mkey_be;
+	u8                     wq_type;
+	u32                    rqn;
+	struct mlx5_core_dev  *mdev;
+	struct mlx5_core_mkey  umr_mkey;
+
+	/* XDP read-mostly */
+	struct xdp_rxq_info    xdp_rxq;
+} ____cacheline_aligned_in_smp;
+
+struct mlx5e_channel {
+	/* data path */
+	struct mlx5e_rq            rq;
+	struct mlx5e_txqsq         sq[MLX5E_MAX_NUM_TC];
+	struct mlx5e_icosq         icosq;   /* internal control operations */
+	bool                       xdp;
+	struct napi_struct         napi;
+	struct device             *pdev;
+	struct net_device         *netdev;
+	__be32                     mkey_be;
+	u8                         num_tc;
+
+	/* XDP_REDIRECT */
+	struct mlx5e_xdpsq         xdpsq;
+
+	/* data path - accessed per napi poll */
+	struct irq_desc *irq_desc;
+	struct mlx5e_ch_stats     *stats;
+
+	/* control */
+	struct mlx5e_priv         *priv;
+	struct mlx5_core_dev      *mdev;
+	struct hwtstamp_config    *tstamp;
+	int                        ix;
+	int                        cpu;
+};
+
+struct mlx5e_channels {
+	struct mlx5e_channel **c;
+	unsigned int           num;
+	struct mlx5e_params    params;
+};
+
+struct mlx5e_channel_stats {
+	struct mlx5e_ch_stats ch;
+	struct mlx5e_sq_stats sq[MLX5E_MAX_NUM_TC];
+	struct mlx5e_rq_stats rq;
+	struct mlx5e_xdpsq_stats rq_xdpsq;
+	struct mlx5e_xdpsq_stats xdpsq;
+} ____cacheline_aligned_in_smp;
+
+enum {
+	MLX5E_STATE_ASYNC_EVENTS_ENABLED,
+	MLX5E_STATE_OPENED,
+	MLX5E_STATE_DESTROYING,
+	MLX5E_STATE_XDP_TX_ENABLED,
+};
+
+struct mlx5e_rqt {
+	u32              rqtn;
+	bool		 enabled;
+};
+
+struct mlx5e_tir {
+	u32		  tirn;
+	struct mlx5e_rqt  rqt;
+	struct list_head  list;
+};
+
+enum {
+	MLX5E_TC_PRIO = 0,
+	MLX5E_NIC_PRIO
+};
+
+struct mlx5e_priv {
+	/* priv data path fields - start */
+	struct mlx5e_txqsq *txq2sq[MLX5E_MAX_NUM_CHANNELS * MLX5E_MAX_NUM_TC];
+	int channel_tc2txq[MLX5E_MAX_NUM_CHANNELS][MLX5E_MAX_NUM_TC];
+#ifdef CONFIG_MLX5_CORE_EN_DCB
+	struct mlx5e_dcbx_dp       dcbx_dp;
+#endif
+	/* priv data path fields - end */
+
+	u32                        msglevel;
+	unsigned long              state;
+	struct mutex               state_lock; /* Protects Interface state */
+	struct mlx5e_rq            drop_rq;
+
+	struct mlx5e_channels      channels;
+	u32                        tisn[MLX5E_MAX_NUM_TC];
+	struct mlx5e_rqt           indir_rqt;
+	struct mlx5e_tir           indir_tir[MLX5E_NUM_INDIR_TIRS];
+	struct mlx5e_tir           inner_indir_tir[MLX5E_NUM_INDIR_TIRS];
+	struct mlx5e_tir           direct_tir[MLX5E_MAX_NUM_CHANNELS];
+	u32                        tx_rates[MLX5E_MAX_NUM_SQS];
+
+	struct mlx5e_flow_steering fs;
+
+	struct workqueue_struct    *wq;
+	struct work_struct         update_carrier_work;
+	struct work_struct         set_rx_mode_work;
+	struct work_struct         tx_timeout_work;
+	struct delayed_work        update_stats_work;
+
+	struct mlx5_core_dev      *mdev;
+	struct net_device         *netdev;
+	struct mlx5e_stats         stats;
+	struct mlx5e_channel_stats channel_stats[MLX5E_MAX_NUM_CHANNELS];
+	u8                         max_opened_tc;
+	struct hwtstamp_config     tstamp;
+	u16                        q_counter;
+	u16                        drop_rq_q_counter;
+#ifdef CONFIG_MLX5_CORE_EN_DCB
+	struct mlx5e_dcbx          dcbx;
+#endif
+
+	const struct mlx5e_profile *profile;
+	void                      *ppriv;
+#ifdef CONFIG_MLX5_EN_IPSEC
+	struct mlx5e_ipsec        *ipsec;
+#endif
+#ifdef CONFIG_MLX5_EN_TLS
+	struct mlx5e_tls          *tls;
+#endif
+};
+
+struct mlx5e_profile {
+	void	(*init)(struct mlx5_core_dev *mdev,
+			struct net_device *netdev,
+			const struct mlx5e_profile *profile, void *ppriv);
+	void	(*cleanup)(struct mlx5e_priv *priv);
+	int	(*init_rx)(struct mlx5e_priv *priv);
+	void	(*cleanup_rx)(struct mlx5e_priv *priv);
+	int	(*init_tx)(struct mlx5e_priv *priv);
+	void	(*cleanup_tx)(struct mlx5e_priv *priv);
+	void	(*enable)(struct mlx5e_priv *priv);
+	void	(*disable)(struct mlx5e_priv *priv);
+	void	(*update_stats)(struct mlx5e_priv *priv);
+	void	(*update_carrier)(struct mlx5e_priv *priv);
+	int	(*max_nch)(struct mlx5_core_dev *mdev);
+	struct {
+		mlx5e_fp_handle_rx_cqe handle_rx_cqe;
+		mlx5e_fp_handle_rx_cqe handle_rx_cqe_mpwqe;
+	} rx_handlers;
+	int	max_tc;
+};
+
+void mlx5e_build_ptys2ethtool_map(void);
+
+u16 mlx5e_select_queue(struct net_device *dev, struct sk_buff *skb,
+		       struct net_device *sb_dev,
+		       select_queue_fallback_t fallback);
+netdev_tx_t mlx5e_xmit(struct sk_buff *skb, struct net_device *dev);
+netdev_tx_t mlx5e_sq_xmit(struct mlx5e_txqsq *sq, struct sk_buff *skb,
+			  struct mlx5e_tx_wqe *wqe, u16 pi);
+
+void mlx5e_completion_event(struct mlx5_core_cq *mcq);
+void mlx5e_cq_error_event(struct mlx5_core_cq *mcq, enum mlx5_event event);
+int mlx5e_napi_poll(struct napi_struct *napi, int budget);
+bool mlx5e_poll_tx_cq(struct mlx5e_cq *cq, int napi_budget);
+int mlx5e_poll_rx_cq(struct mlx5e_cq *cq, int budget);
+void mlx5e_free_txqsq_descs(struct mlx5e_txqsq *sq);
+
+bool mlx5e_check_fragmented_striding_rq_cap(struct mlx5_core_dev *mdev);
+bool mlx5e_striding_rq_possible(struct mlx5_core_dev *mdev,
+				struct mlx5e_params *params);
+
+void mlx5e_page_dma_unmap(struct mlx5e_rq *rq, struct mlx5e_dma_info *dma_info);
+void mlx5e_page_release(struct mlx5e_rq *rq, struct mlx5e_dma_info *dma_info,
+			bool recycle);
+void mlx5e_handle_rx_cqe(struct mlx5e_rq *rq, struct mlx5_cqe64 *cqe);
+void mlx5e_handle_rx_cqe_mpwrq(struct mlx5e_rq *rq, struct mlx5_cqe64 *cqe);
+bool mlx5e_post_rx_wqes(struct mlx5e_rq *rq);
+bool mlx5e_post_rx_mpwqes(struct mlx5e_rq *rq);
+void mlx5e_dealloc_rx_wqe(struct mlx5e_rq *rq, u16 ix);
+void mlx5e_dealloc_rx_mpwqe(struct mlx5e_rq *rq, u16 ix);
+struct sk_buff *
+mlx5e_skb_from_cqe_mpwrq_linear(struct mlx5e_rq *rq, struct mlx5e_mpw_info *wi,
+				u16 cqe_bcnt, u32 head_offset, u32 page_idx);
+struct sk_buff *
+mlx5e_skb_from_cqe_mpwrq_nonlinear(struct mlx5e_rq *rq, struct mlx5e_mpw_info *wi,
+				   u16 cqe_bcnt, u32 head_offset, u32 page_idx);
+struct sk_buff *
+mlx5e_skb_from_cqe_linear(struct mlx5e_rq *rq, struct mlx5_cqe64 *cqe,
+			  struct mlx5e_wqe_frag_info *wi, u32 cqe_bcnt);
+struct sk_buff *
+mlx5e_skb_from_cqe_nonlinear(struct mlx5e_rq *rq, struct mlx5_cqe64 *cqe,
+			     struct mlx5e_wqe_frag_info *wi, u32 cqe_bcnt);
+
+void mlx5e_update_stats(struct mlx5e_priv *priv);
+
+void mlx5e_init_l2_addr(struct mlx5e_priv *priv);
+int mlx5e_self_test_num(struct mlx5e_priv *priv);
+void mlx5e_self_test(struct net_device *ndev, struct ethtool_test *etest,
+		     u64 *buf);
+void mlx5e_set_rx_mode_work(struct work_struct *work);
+
+int mlx5e_hwstamp_set(struct mlx5e_priv *priv, struct ifreq *ifr);
+int mlx5e_hwstamp_get(struct mlx5e_priv *priv, struct ifreq *ifr);
+int mlx5e_modify_rx_cqe_compression_locked(struct mlx5e_priv *priv, bool val);
+
+int mlx5e_vlan_rx_add_vid(struct net_device *dev, __always_unused __be16 proto,
+			  u16 vid);
+int mlx5e_vlan_rx_kill_vid(struct net_device *dev, __always_unused __be16 proto,
+			   u16 vid);
+void mlx5e_timestamp_init(struct mlx5e_priv *priv);
+
+struct mlx5e_redirect_rqt_param {
+	bool is_rss;
+	union {
+		u32 rqn; /* Direct RQN (Non-RSS) */
+		struct {
+			u8 hfunc;
+			struct mlx5e_channels *channels;
+		} rss; /* RSS data */
+	};
+};
+
+int mlx5e_redirect_rqt(struct mlx5e_priv *priv, u32 rqtn, int sz,
+		       struct mlx5e_redirect_rqt_param rrp);
+void mlx5e_build_indir_tir_ctx_hash(struct mlx5e_params *params,
+				    enum mlx5e_traffic_types tt,
+				    void *tirc, bool inner);
+
+int mlx5e_open_locked(struct net_device *netdev);
+int mlx5e_close_locked(struct net_device *netdev);
+
+int mlx5e_open_channels(struct mlx5e_priv *priv,
+			struct mlx5e_channels *chs);
+void mlx5e_close_channels(struct mlx5e_channels *chs);
+
+/* Function pointer to be used to modify WH settings while
+ * switching channels
+ */
+typedef int (*mlx5e_fp_hw_modify)(struct mlx5e_priv *priv);
+void mlx5e_switch_priv_channels(struct mlx5e_priv *priv,
+				struct mlx5e_channels *new_chs,
+				mlx5e_fp_hw_modify hw_modify);
+void mlx5e_activate_priv_channels(struct mlx5e_priv *priv);
+void mlx5e_deactivate_priv_channels(struct mlx5e_priv *priv);
+
+void mlx5e_build_default_indir_rqt(u32 *indirection_rqt, int len,
+				   int num_channels);
+void mlx5e_set_tx_cq_mode_params(struct mlx5e_params *params,
+				 u8 cq_period_mode);
+void mlx5e_set_rx_cq_mode_params(struct mlx5e_params *params,
+				 u8 cq_period_mode);
+void mlx5e_set_rq_type(struct mlx5_core_dev *mdev, struct mlx5e_params *params);
+void mlx5e_init_rq_type_params(struct mlx5_core_dev *mdev,
+			       struct mlx5e_params *params);
+
+static inline bool mlx5e_tunnel_inner_ft_supported(struct mlx5_core_dev *mdev)
+{
+	return (MLX5_CAP_ETH(mdev, tunnel_stateless_gre) &&
+		MLX5_CAP_FLOWTABLE_NIC_RX(mdev, ft_field_support.inner_ip_version));
+}
+
+static inline void mlx5e_sq_fetch_wqe(struct mlx5e_txqsq *sq,
+				      struct mlx5e_tx_wqe **wqe,
+				      u16 *pi)
+{
+	struct mlx5_wq_cyc *wq = &sq->wq;
+
+	*pi  = mlx5_wq_cyc_ctr2ix(wq, sq->pc);
+	*wqe = mlx5_wq_cyc_get_wqe(wq, *pi);
+	memset(*wqe, 0, sizeof(**wqe));
+}
+
+static inline
+struct mlx5e_tx_wqe *mlx5e_post_nop(struct mlx5_wq_cyc *wq, u32 sqn, u16 *pc)
+{
+	u16                         pi   = mlx5_wq_cyc_ctr2ix(wq, *pc);
+	struct mlx5e_tx_wqe        *wqe  = mlx5_wq_cyc_get_wqe(wq, pi);
+	struct mlx5_wqe_ctrl_seg   *cseg = &wqe->ctrl;
+
+	memset(cseg, 0, sizeof(*cseg));
+
+	cseg->opmod_idx_opcode = cpu_to_be32((*pc << 8) | MLX5_OPCODE_NOP);
+	cseg->qpn_ds           = cpu_to_be32((sqn << 8) | 0x01);
+
+	(*pc)++;
+
+	return wqe;
+}
+
+static inline
+void mlx5e_notify_hw(struct mlx5_wq_cyc *wq, u16 pc,
+		     void __iomem *uar_map,
+		     struct mlx5_wqe_ctrl_seg *ctrl)
+{
+	ctrl->fm_ce_se = MLX5_WQE_CTRL_CQ_UPDATE;
+	/* ensure wqe is visible to device before updating doorbell record */
+	dma_wmb();
+
+	*wq->db = cpu_to_be32(pc);
+
+	/* ensure doorbell record is visible to device before ringing the
+	 * doorbell
+	 */
+	wmb();
+
+	mlx5_write64((__be32 *)ctrl, uar_map, NULL);
+}
+
+static inline void mlx5e_cq_arm(struct mlx5e_cq *cq)
+{
+	struct mlx5_core_cq *mcq;
+
+	mcq = &cq->mcq;
+	mlx5_cq_arm(mcq, MLX5_CQ_DB_REQ_NOT, mcq->uar->map, cq->wq.cc);
+}
+
+extern const struct ethtool_ops mlx5e_ethtool_ops;
+#ifdef CONFIG_MLX5_CORE_EN_DCB
+extern const struct dcbnl_rtnl_ops mlx5e_dcbnl_ops;
+int mlx5e_dcbnl_ieee_setets_core(struct mlx5e_priv *priv, struct ieee_ets *ets);
+void mlx5e_dcbnl_initialize(struct mlx5e_priv *priv);
+void mlx5e_dcbnl_init_app(struct mlx5e_priv *priv);
+void mlx5e_dcbnl_delete_app(struct mlx5e_priv *priv);
+#endif
+
+int mlx5e_create_tir(struct mlx5_core_dev *mdev,
+		     struct mlx5e_tir *tir, u32 *in, int inlen);
+void mlx5e_destroy_tir(struct mlx5_core_dev *mdev,
+		       struct mlx5e_tir *tir);
+int mlx5e_create_mdev_resources(struct mlx5_core_dev *mdev);
+void mlx5e_destroy_mdev_resources(struct mlx5_core_dev *mdev);
+int mlx5e_refresh_tirs(struct mlx5e_priv *priv, bool enable_uc_lb);
+
+/* common netdev helpers */
+int mlx5e_create_indirect_rqt(struct mlx5e_priv *priv);
+
+int mlx5e_create_indirect_tirs(struct mlx5e_priv *priv);
+void mlx5e_destroy_indirect_tirs(struct mlx5e_priv *priv);
+
+int mlx5e_create_direct_rqts(struct mlx5e_priv *priv);
+void mlx5e_destroy_direct_rqts(struct mlx5e_priv *priv);
+int mlx5e_create_direct_tirs(struct mlx5e_priv *priv);
+void mlx5e_destroy_direct_tirs(struct mlx5e_priv *priv);
+void mlx5e_destroy_rqt(struct mlx5e_priv *priv, struct mlx5e_rqt *rqt);
+
+int mlx5e_create_tis(struct mlx5_core_dev *mdev, int tc,
+		     u32 underlay_qpn, u32 *tisn);
+void mlx5e_destroy_tis(struct mlx5_core_dev *mdev, u32 tisn);
+
+int mlx5e_create_tises(struct mlx5e_priv *priv);
+void mlx5e_cleanup_nic_tx(struct mlx5e_priv *priv);
+int mlx5e_close(struct net_device *netdev);
+int mlx5e_open(struct net_device *netdev);
+void mlx5e_update_stats_work(struct work_struct *work);
+
+int mlx5e_bits_invert(unsigned long a, int size);
+
+typedef int (*change_hw_mtu_cb)(struct mlx5e_priv *priv);
+int mlx5e_change_mtu(struct net_device *netdev, int new_mtu,
+		     change_hw_mtu_cb set_mtu_cb);
+
+/* ethtool helpers */
+void mlx5e_ethtool_get_drvinfo(struct mlx5e_priv *priv,
+			       struct ethtool_drvinfo *drvinfo);
+void mlx5e_ethtool_get_strings(struct mlx5e_priv *priv,
+			       uint32_t stringset, uint8_t *data);
+int mlx5e_ethtool_get_sset_count(struct mlx5e_priv *priv, int sset);
+void mlx5e_ethtool_get_ethtool_stats(struct mlx5e_priv *priv,
+				     struct ethtool_stats *stats, u64 *data);
+void mlx5e_ethtool_get_ringparam(struct mlx5e_priv *priv,
+				 struct ethtool_ringparam *param);
+int mlx5e_ethtool_set_ringparam(struct mlx5e_priv *priv,
+				struct ethtool_ringparam *param);
+void mlx5e_ethtool_get_channels(struct mlx5e_priv *priv,
+				struct ethtool_channels *ch);
+int mlx5e_ethtool_set_channels(struct mlx5e_priv *priv,
+			       struct ethtool_channels *ch);
+int mlx5e_ethtool_get_coalesce(struct mlx5e_priv *priv,
+			       struct ethtool_coalesce *coal);
+int mlx5e_ethtool_set_coalesce(struct mlx5e_priv *priv,
+			       struct ethtool_coalesce *coal);
+int mlx5e_ethtool_get_ts_info(struct mlx5e_priv *priv,
+			      struct ethtool_ts_info *info);
+int mlx5e_ethtool_flash_device(struct mlx5e_priv *priv,
+			       struct ethtool_flash *flash);
+
+/* mlx5e generic netdev management API */
+struct net_device*
+mlx5e_create_netdev(struct mlx5_core_dev *mdev, const struct mlx5e_profile *profile,
+		    void *ppriv);
+int mlx5e_attach_netdev(struct mlx5e_priv *priv);
+void mlx5e_detach_netdev(struct mlx5e_priv *priv);
+void mlx5e_destroy_netdev(struct mlx5e_priv *priv);
+void mlx5e_build_nic_params(struct mlx5_core_dev *mdev,
+			    struct mlx5e_params *params,
+			    u16 max_channels, u16 mtu);
+u8 mlx5e_params_calculate_tx_min_inline(struct mlx5_core_dev *mdev);
+void mlx5e_rx_dim_work(struct work_struct *work);
+void mlx5e_tx_dim_work(struct work_struct *work);
+#endif /* __MLX5_EN_H__ */
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en/Makefile b/drivers/net/ethernet/mellanox/mlx5/core/en/Makefile
new file mode 100644
index 000000000..d8e17110f
--- /dev/null
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en/Makefile
@@ -0,0 +1 @@
+subdir-ccflags-y += -I$(src)/..
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en/fs.h b/drivers/net/ethernet/mellanox/mlx5/core/en/fs.h
new file mode 100644
index 000000000..1431232c9
--- /dev/null
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en/fs.h
@@ -0,0 +1,212 @@
+/* SPDX-License-Identifier: GPL-2.0 OR Linux-OpenIB */
+/* Copyright (c) 2018 Mellanox Technologies. */
+
+#ifndef __MLX5E_FLOW_STEER_H__
+#define __MLX5E_FLOW_STEER_H__
+
+enum {
+	MLX5E_TC_FT_LEVEL = 0,
+	MLX5E_TC_TTC_FT_LEVEL,
+};
+
+struct mlx5e_tc_table {
+	struct mlx5_flow_table		*t;
+
+	struct rhashtable               ht;
+
+	DECLARE_HASHTABLE(mod_hdr_tbl, 8);
+	DECLARE_HASHTABLE(hairpin_tbl, 8);
+
+	struct notifier_block     netdevice_nb;
+};
+
+struct mlx5e_flow_table {
+	int num_groups;
+	struct mlx5_flow_table *t;
+	struct mlx5_flow_group **g;
+};
+
+struct mlx5e_l2_rule {
+	u8  addr[ETH_ALEN + 2];
+	struct mlx5_flow_handle *rule;
+};
+
+#define MLX5E_L2_ADDR_HASH_SIZE BIT(BITS_PER_BYTE)
+
+struct mlx5e_vlan_table {
+	struct mlx5e_flow_table		ft;
+	DECLARE_BITMAP(active_cvlans, VLAN_N_VID);
+	DECLARE_BITMAP(active_svlans, VLAN_N_VID);
+	struct mlx5_flow_handle	*active_cvlans_rule[VLAN_N_VID];
+	struct mlx5_flow_handle	*active_svlans_rule[VLAN_N_VID];
+	struct mlx5_flow_handle	*untagged_rule;
+	struct mlx5_flow_handle	*any_cvlan_rule;
+	struct mlx5_flow_handle	*any_svlan_rule;
+	bool			cvlan_filter_disabled;
+};
+
+struct mlx5e_l2_table {
+	struct mlx5e_flow_table    ft;
+	struct hlist_head          netdev_uc[MLX5E_L2_ADDR_HASH_SIZE];
+	struct hlist_head          netdev_mc[MLX5E_L2_ADDR_HASH_SIZE];
+	struct mlx5e_l2_rule	   broadcast;
+	struct mlx5e_l2_rule	   allmulti;
+	struct mlx5e_l2_rule	   promisc;
+	bool                       broadcast_enabled;
+	bool                       allmulti_enabled;
+	bool                       promisc_enabled;
+};
+
+enum mlx5e_traffic_types {
+	MLX5E_TT_IPV4_TCP,
+	MLX5E_TT_IPV6_TCP,
+	MLX5E_TT_IPV4_UDP,
+	MLX5E_TT_IPV6_UDP,
+	MLX5E_TT_IPV4_IPSEC_AH,
+	MLX5E_TT_IPV6_IPSEC_AH,
+	MLX5E_TT_IPV4_IPSEC_ESP,
+	MLX5E_TT_IPV6_IPSEC_ESP,
+	MLX5E_TT_IPV4,
+	MLX5E_TT_IPV6,
+	MLX5E_TT_ANY,
+	MLX5E_NUM_TT,
+	MLX5E_NUM_INDIR_TIRS = MLX5E_TT_ANY,
+};
+
+enum mlx5e_tunnel_types {
+	MLX5E_TT_IPV4_GRE,
+	MLX5E_TT_IPV6_GRE,
+	MLX5E_NUM_TUNNEL_TT,
+};
+
+/* L3/L4 traffic type classifier */
+struct mlx5e_ttc_table {
+	struct mlx5e_flow_table  ft;
+	struct mlx5_flow_handle	 *rules[MLX5E_NUM_TT];
+	struct mlx5_flow_handle  *tunnel_rules[MLX5E_NUM_TUNNEL_TT];
+};
+
+/* NIC prio FTS */
+enum {
+	MLX5E_VLAN_FT_LEVEL = 0,
+	MLX5E_L2_FT_LEVEL,
+	MLX5E_TTC_FT_LEVEL,
+	MLX5E_INNER_TTC_FT_LEVEL,
+#ifdef CONFIG_MLX5_EN_ARFS
+	MLX5E_ARFS_FT_LEVEL
+#endif
+};
+
+#ifdef CONFIG_MLX5_EN_RXNFC
+
+struct mlx5e_ethtool_table {
+	struct mlx5_flow_table *ft;
+	int                    num_rules;
+};
+
+#define ETHTOOL_NUM_L3_L4_FTS 7
+#define ETHTOOL_NUM_L2_FTS 4
+
+struct mlx5e_ethtool_steering {
+	struct mlx5e_ethtool_table      l3_l4_ft[ETHTOOL_NUM_L3_L4_FTS];
+	struct mlx5e_ethtool_table      l2_ft[ETHTOOL_NUM_L2_FTS];
+	struct list_head                rules;
+	int                             tot_num_rules;
+};
+
+void mlx5e_ethtool_init_steering(struct mlx5e_priv *priv);
+void mlx5e_ethtool_cleanup_steering(struct mlx5e_priv *priv);
+int mlx5e_set_rxnfc(struct net_device *dev, struct ethtool_rxnfc *cmd);
+int mlx5e_get_rxnfc(struct net_device *dev,
+		    struct ethtool_rxnfc *info, u32 *rule_locs);
+#else
+static inline void mlx5e_ethtool_init_steering(struct mlx5e_priv *priv)    { }
+static inline void mlx5e_ethtool_cleanup_steering(struct mlx5e_priv *priv) { }
+#endif /* CONFIG_MLX5_EN_RXNFC */
+
+#ifdef CONFIG_MLX5_EN_ARFS
+#define ARFS_HASH_SHIFT BITS_PER_BYTE
+#define ARFS_HASH_SIZE BIT(BITS_PER_BYTE)
+
+struct arfs_table {
+	struct mlx5e_flow_table  ft;
+	struct mlx5_flow_handle	 *default_rule;
+	struct hlist_head	 rules_hash[ARFS_HASH_SIZE];
+};
+
+enum  arfs_type {
+	ARFS_IPV4_TCP,
+	ARFS_IPV6_TCP,
+	ARFS_IPV4_UDP,
+	ARFS_IPV6_UDP,
+	ARFS_NUM_TYPES,
+};
+
+struct mlx5e_arfs_tables {
+	struct arfs_table arfs_tables[ARFS_NUM_TYPES];
+	/* Protect aRFS rules list */
+	spinlock_t                     arfs_lock;
+	struct list_head               rules;
+	int                            last_filter_id;
+	struct workqueue_struct        *wq;
+};
+
+int mlx5e_arfs_create_tables(struct mlx5e_priv *priv);
+void mlx5e_arfs_destroy_tables(struct mlx5e_priv *priv);
+int mlx5e_arfs_enable(struct mlx5e_priv *priv);
+int mlx5e_arfs_disable(struct mlx5e_priv *priv);
+int mlx5e_rx_flow_steer(struct net_device *dev, const struct sk_buff *skb,
+			u16 rxq_index, u32 flow_id);
+#else
+static inline int mlx5e_arfs_create_tables(struct mlx5e_priv *priv) { return 0; }
+static inline void mlx5e_arfs_destroy_tables(struct mlx5e_priv *priv) {}
+static inline int mlx5e_arfs_enable(struct mlx5e_priv *priv) { return -EOPNOTSUPP; }
+static inline int mlx5e_arfs_disable(struct mlx5e_priv *priv) {	return -EOPNOTSUPP; }
+#endif
+
+struct mlx5e_flow_steering {
+	struct mlx5_flow_namespace      *ns;
+#ifdef CONFIG_MLX5_EN_RXNFC
+	struct mlx5e_ethtool_steering   ethtool;
+#endif
+	struct mlx5e_tc_table           tc;
+	struct mlx5e_vlan_table         vlan;
+	struct mlx5e_l2_table           l2;
+	struct mlx5e_ttc_table          ttc;
+	struct mlx5e_ttc_table          inner_ttc;
+#ifdef CONFIG_MLX5_EN_ARFS
+	struct mlx5e_arfs_tables        arfs;
+#endif
+};
+
+struct ttc_params {
+	struct mlx5_flow_table_attr ft_attr;
+	u32 any_tt_tirn;
+	u32 indir_tirn[MLX5E_NUM_INDIR_TIRS];
+	struct mlx5e_ttc_table *inner_ttc;
+};
+
+void mlx5e_set_ttc_basic_params(struct mlx5e_priv *priv, struct ttc_params *ttc_params);
+void mlx5e_set_ttc_ft_params(struct ttc_params *ttc_params);
+void mlx5e_set_inner_ttc_ft_params(struct ttc_params *ttc_params);
+
+int mlx5e_create_ttc_table(struct mlx5e_priv *priv, struct ttc_params *params,
+			   struct mlx5e_ttc_table *ttc);
+void mlx5e_destroy_ttc_table(struct mlx5e_priv *priv,
+			     struct mlx5e_ttc_table *ttc);
+
+int mlx5e_create_inner_ttc_table(struct mlx5e_priv *priv, struct ttc_params *params,
+				 struct mlx5e_ttc_table *ttc);
+void mlx5e_destroy_inner_ttc_table(struct mlx5e_priv *priv,
+				   struct mlx5e_ttc_table *ttc);
+
+void mlx5e_destroy_flow_table(struct mlx5e_flow_table *ft);
+
+void mlx5e_enable_cvlan_filter(struct mlx5e_priv *priv);
+void mlx5e_disable_cvlan_filter(struct mlx5e_priv *priv);
+
+int mlx5e_create_flow_steering(struct mlx5e_priv *priv);
+void mlx5e_destroy_flow_steering(struct mlx5e_priv *priv);
+
+#endif /* __MLX5E_FLOW_STEER_H__ */
+
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en/port.c b/drivers/net/ethernet/mellanox/mlx5/core/en/port.c
new file mode 100644
index 000000000..12e1682f9
--- /dev/null
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en/port.c
@@ -0,0 +1,235 @@
+/*
+ * 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 "port.h"
+
+/* speed in units of 1Mb */
+static const u32 mlx5e_link_speed[MLX5E_LINK_MODES_NUMBER] = {
+	[MLX5E_1000BASE_CX_SGMII] = 1000,
+	[MLX5E_1000BASE_KX]       = 1000,
+	[MLX5E_10GBASE_CX4]       = 10000,
+	[MLX5E_10GBASE_KX4]       = 10000,
+	[MLX5E_10GBASE_KR]        = 10000,
+	[MLX5E_20GBASE_KR2]       = 20000,
+	[MLX5E_40GBASE_CR4]       = 40000,
+	[MLX5E_40GBASE_KR4]       = 40000,
+	[MLX5E_56GBASE_R4]        = 56000,
+	[MLX5E_10GBASE_CR]        = 10000,
+	[MLX5E_10GBASE_SR]        = 10000,
+	[MLX5E_10GBASE_ER]        = 10000,
+	[MLX5E_40GBASE_SR4]       = 40000,
+	[MLX5E_40GBASE_LR4]       = 40000,
+	[MLX5E_50GBASE_SR2]       = 50000,
+	[MLX5E_100GBASE_CR4]      = 100000,
+	[MLX5E_100GBASE_SR4]      = 100000,
+	[MLX5E_100GBASE_KR4]      = 100000,
+	[MLX5E_100GBASE_LR4]      = 100000,
+	[MLX5E_100BASE_TX]        = 100,
+	[MLX5E_1000BASE_T]        = 1000,
+	[MLX5E_10GBASE_T]         = 10000,
+	[MLX5E_25GBASE_CR]        = 25000,
+	[MLX5E_25GBASE_KR]        = 25000,
+	[MLX5E_25GBASE_SR]        = 25000,
+	[MLX5E_50GBASE_CR2]       = 50000,
+	[MLX5E_50GBASE_KR2]       = 50000,
+};
+
+u32 mlx5e_port_ptys2speed(u32 eth_proto_oper)
+{
+	unsigned long temp = eth_proto_oper;
+	u32 speed = 0;
+	int i;
+
+	i = find_first_bit(&temp, MLX5E_LINK_MODES_NUMBER);
+	if (i < MLX5E_LINK_MODES_NUMBER)
+		speed = mlx5e_link_speed[i];
+
+	return speed;
+}
+
+int mlx5e_port_linkspeed(struct mlx5_core_dev *mdev, u32 *speed)
+{
+	u32 out[MLX5_ST_SZ_DW(ptys_reg)] = {};
+	u32 eth_proto_oper;
+	int err;
+
+	err = mlx5_query_port_ptys(mdev, out, sizeof(out), MLX5_PTYS_EN, 1);
+	if (err)
+		return err;
+
+	eth_proto_oper = MLX5_GET(ptys_reg, out, eth_proto_oper);
+	*speed = mlx5e_port_ptys2speed(eth_proto_oper);
+	if (!(*speed))
+		err = -EINVAL;
+
+	return err;
+}
+
+int mlx5e_port_max_linkspeed(struct mlx5_core_dev *mdev, u32 *speed)
+{
+	u32 max_speed = 0;
+	u32 proto_cap;
+	int err;
+	int i;
+
+	err = mlx5_query_port_proto_cap(mdev, &proto_cap, MLX5_PTYS_EN);
+	if (err)
+		return err;
+
+	for (i = 0; i < MLX5E_LINK_MODES_NUMBER; ++i)
+		if (proto_cap & MLX5E_PROT_MASK(i))
+			max_speed = max(max_speed, mlx5e_link_speed[i]);
+
+	*speed = max_speed;
+	return 0;
+}
+
+u32 mlx5e_port_speed2linkmodes(u32 speed)
+{
+	u32 link_modes = 0;
+	int i;
+
+	for (i = 0; i < MLX5E_LINK_MODES_NUMBER; ++i) {
+		if (mlx5e_link_speed[i] == speed)
+			link_modes |= MLX5E_PROT_MASK(i);
+	}
+
+	return link_modes;
+}
+
+int mlx5e_port_query_pbmc(struct mlx5_core_dev *mdev, void *out)
+{
+	int sz = MLX5_ST_SZ_BYTES(pbmc_reg);
+	void *in;
+	int err;
+
+	in = kzalloc(sz, GFP_KERNEL);
+	if (!in)
+		return -ENOMEM;
+
+	MLX5_SET(pbmc_reg, in, local_port, 1);
+	err = mlx5_core_access_reg(mdev, in, sz, out, sz, MLX5_REG_PBMC, 0, 0);
+
+	kfree(in);
+	return err;
+}
+
+int mlx5e_port_set_pbmc(struct mlx5_core_dev *mdev, void *in)
+{
+	int sz = MLX5_ST_SZ_BYTES(pbmc_reg);
+	void *out;
+	int err;
+
+	out = kzalloc(sz, GFP_KERNEL);
+	if (!out)
+		return -ENOMEM;
+
+	MLX5_SET(pbmc_reg, in, local_port, 1);
+	err = mlx5_core_access_reg(mdev, in, sz, out, sz, MLX5_REG_PBMC, 0, 1);
+
+	kfree(out);
+	return err;
+}
+
+/* buffer[i]: buffer that priority i mapped to */
+int mlx5e_port_query_priority2buffer(struct mlx5_core_dev *mdev, u8 *buffer)
+{
+	int sz = MLX5_ST_SZ_BYTES(pptb_reg);
+	u32 prio_x_buff;
+	void *out;
+	void *in;
+	int prio;
+	int err;
+
+	in = kzalloc(sz, GFP_KERNEL);
+	out = kzalloc(sz, GFP_KERNEL);
+	if (!in || !out) {
+		err = -ENOMEM;
+		goto out;
+	}
+
+	MLX5_SET(pptb_reg, in, local_port, 1);
+	err = mlx5_core_access_reg(mdev, in, sz, out, sz, MLX5_REG_PPTB, 0, 0);
+	if (err)
+		goto out;
+
+	prio_x_buff = MLX5_GET(pptb_reg, out, prio_x_buff);
+	for (prio = 0; prio < 8; prio++) {
+		buffer[prio] = (u8)(prio_x_buff >> (4 * prio)) & 0xF;
+		mlx5_core_dbg(mdev, "prio %d, buffer %d\n", prio, buffer[prio]);
+	}
+out:
+	kfree(in);
+	kfree(out);
+	return err;
+}
+
+int mlx5e_port_set_priority2buffer(struct mlx5_core_dev *mdev, u8 *buffer)
+{
+	int sz = MLX5_ST_SZ_BYTES(pptb_reg);
+	u32 prio_x_buff;
+	void *out;
+	void *in;
+	int prio;
+	int err;
+
+	in = kzalloc(sz, GFP_KERNEL);
+	out = kzalloc(sz, GFP_KERNEL);
+	if (!in || !out) {
+		err = -ENOMEM;
+		goto out;
+	}
+
+	/* First query the pptb register */
+	MLX5_SET(pptb_reg, in, local_port, 1);
+	err = mlx5_core_access_reg(mdev, in, sz, out, sz, MLX5_REG_PPTB, 0, 0);
+	if (err)
+		goto out;
+
+	memcpy(in, out, sz);
+	MLX5_SET(pptb_reg, in, local_port, 1);
+
+	/* Update the pm and prio_x_buff */
+	MLX5_SET(pptb_reg, in, pm, 0xFF);
+
+	prio_x_buff = 0;
+	for (prio = 0; prio < 8; prio++)
+		prio_x_buff |= (buffer[prio] << (4 * prio));
+	MLX5_SET(pptb_reg, in, prio_x_buff, prio_x_buff);
+
+	err = mlx5_core_access_reg(mdev, in, sz, out, sz, MLX5_REG_PPTB, 0, 1);
+
+out:
+	kfree(in);
+	kfree(out);
+	return err;
+}
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en/port.h b/drivers/net/ethernet/mellanox/mlx5/core/en/port.h
new file mode 100644
index 000000000..f8cbd8194
--- /dev/null
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en/port.h
@@ -0,0 +1,48 @@
+/*
+ * 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.
+ */
+
+#ifndef __MLX5E_EN_PORT_H
+#define __MLX5E_EN_PORT_H
+
+#include <linux/mlx5/driver.h>
+#include "en.h"
+
+u32 mlx5e_port_ptys2speed(u32 eth_proto_oper);
+int mlx5e_port_linkspeed(struct mlx5_core_dev *mdev, u32 *speed);
+int mlx5e_port_max_linkspeed(struct mlx5_core_dev *mdev, u32 *speed);
+u32 mlx5e_port_speed2linkmodes(u32 speed);
+
+int mlx5e_port_query_pbmc(struct mlx5_core_dev *mdev, void *out);
+int mlx5e_port_set_pbmc(struct mlx5_core_dev *mdev, void *in);
+int mlx5e_port_query_priority2buffer(struct mlx5_core_dev *mdev, u8 *buffer);
+int mlx5e_port_set_priority2buffer(struct mlx5_core_dev *mdev, u8 *buffer);
+#endif
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en/port_buffer.c b/drivers/net/ethernet/mellanox/mlx5/core/en/port_buffer.c
new file mode 100644
index 000000000..28d56e44e
--- /dev/null
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en/port_buffer.c
@@ -0,0 +1,357 @@
+/*
+ * 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 "port_buffer.h"
+
+int mlx5e_port_query_buffer(struct mlx5e_priv *priv,
+			    struct mlx5e_port_buffer *port_buffer)
+{
+	struct mlx5_core_dev *mdev = priv->mdev;
+	int sz = MLX5_ST_SZ_BYTES(pbmc_reg);
+	u32 total_used = 0;
+	void *buffer;
+	void *out;
+	int err;
+	int i;
+
+	out = kzalloc(sz, GFP_KERNEL);
+	if (!out)
+		return -ENOMEM;
+
+	err = mlx5e_port_query_pbmc(mdev, out);
+	if (err)
+		goto out;
+
+	for (i = 0; i < MLX5E_MAX_BUFFER; i++) {
+		buffer = MLX5_ADDR_OF(pbmc_reg, out, buffer[i]);
+		port_buffer->buffer[i].lossy =
+			MLX5_GET(bufferx_reg, buffer, lossy);
+		port_buffer->buffer[i].epsb =
+			MLX5_GET(bufferx_reg, buffer, epsb);
+		port_buffer->buffer[i].size =
+			MLX5_GET(bufferx_reg, buffer, size) << MLX5E_BUFFER_CELL_SHIFT;
+		port_buffer->buffer[i].xon =
+			MLX5_GET(bufferx_reg, buffer, xon_threshold) << MLX5E_BUFFER_CELL_SHIFT;
+		port_buffer->buffer[i].xoff =
+			MLX5_GET(bufferx_reg, buffer, xoff_threshold) << MLX5E_BUFFER_CELL_SHIFT;
+		total_used += port_buffer->buffer[i].size;
+
+		mlx5e_dbg(HW, priv, "buffer %d: size=%d, xon=%d, xoff=%d, epsb=%d, lossy=%d\n", i,
+			  port_buffer->buffer[i].size,
+			  port_buffer->buffer[i].xon,
+			  port_buffer->buffer[i].xoff,
+			  port_buffer->buffer[i].epsb,
+			  port_buffer->buffer[i].lossy);
+	}
+
+	port_buffer->port_buffer_size =
+		MLX5_GET(pbmc_reg, out, port_buffer_size) << MLX5E_BUFFER_CELL_SHIFT;
+	port_buffer->spare_buffer_size =
+		port_buffer->port_buffer_size - total_used;
+
+	mlx5e_dbg(HW, priv, "total buffer size=%d, spare buffer size=%d\n",
+		  port_buffer->port_buffer_size,
+		  port_buffer->spare_buffer_size);
+out:
+	kfree(out);
+	return err;
+}
+
+static int port_set_buffer(struct mlx5e_priv *priv,
+			   struct mlx5e_port_buffer *port_buffer)
+{
+	struct mlx5_core_dev *mdev = priv->mdev;
+	int sz = MLX5_ST_SZ_BYTES(pbmc_reg);
+	void *buffer;
+	void *in;
+	int err;
+	int i;
+
+	in = kzalloc(sz, GFP_KERNEL);
+	if (!in)
+		return -ENOMEM;
+
+	err = mlx5e_port_query_pbmc(mdev, in);
+	if (err)
+		goto out;
+
+	for (i = 0; i < MLX5E_MAX_BUFFER; i++) {
+		buffer = MLX5_ADDR_OF(pbmc_reg, in, buffer[i]);
+
+		MLX5_SET(bufferx_reg, buffer, size,
+			 port_buffer->buffer[i].size >> MLX5E_BUFFER_CELL_SHIFT);
+		MLX5_SET(bufferx_reg, buffer, lossy,
+			 port_buffer->buffer[i].lossy);
+		MLX5_SET(bufferx_reg, buffer, xoff_threshold,
+			 port_buffer->buffer[i].xoff >> MLX5E_BUFFER_CELL_SHIFT);
+		MLX5_SET(bufferx_reg, buffer, xon_threshold,
+			 port_buffer->buffer[i].xon >> MLX5E_BUFFER_CELL_SHIFT);
+	}
+
+	err = mlx5e_port_set_pbmc(mdev, in);
+out:
+	kfree(in);
+	return err;
+}
+
+/* xoff = ((301+2.16 * len [m]) * speed [Gbps] + 2.72 MTU [B])
+ * minimum speed value is 40Gbps
+ */
+static u32 calculate_xoff(struct mlx5e_priv *priv, unsigned int mtu)
+{
+	u32 speed;
+	u32 xoff;
+	int err;
+
+	err = mlx5e_port_linkspeed(priv->mdev, &speed);
+	if (err)
+		speed = SPEED_40000;
+	speed = max_t(u32, speed, SPEED_40000);
+
+	xoff = (301 + 216 * priv->dcbx.cable_len / 100) * speed / 1000 + 272 * mtu / 100;
+
+	mlx5e_dbg(HW, priv, "%s: xoff=%d\n", __func__, xoff);
+	return xoff;
+}
+
+static int update_xoff_threshold(struct mlx5e_port_buffer *port_buffer,
+				 u32 xoff, unsigned int max_mtu)
+{
+	int i;
+
+	for (i = 0; i < MLX5E_MAX_BUFFER; i++) {
+		if (port_buffer->buffer[i].lossy) {
+			port_buffer->buffer[i].xoff = 0;
+			port_buffer->buffer[i].xon  = 0;
+			continue;
+		}
+
+		if (port_buffer->buffer[i].size <
+		    (xoff + max_mtu + (1 << MLX5E_BUFFER_CELL_SHIFT))) {
+			pr_err("buffer_size[%d]=%d is not enough for lossless buffer\n",
+			       i, port_buffer->buffer[i].size);
+			return -ENOMEM;
+		}
+
+		port_buffer->buffer[i].xoff = port_buffer->buffer[i].size - xoff;
+		port_buffer->buffer[i].xon  =
+			port_buffer->buffer[i].xoff - max_mtu;
+	}
+
+	return 0;
+}
+
+/**
+ * update_buffer_lossy()
+ *   max_mtu: netdev's max_mtu
+ *   pfc_en: <input> current pfc configuration
+ *   buffer: <input> current prio to buffer mapping
+ *   xoff:   <input> xoff value
+ *   port_buffer: <output> port receive buffer configuration
+ *   change: <output>
+ *
+ *   Update buffer configuration based on pfc configuraiton and priority
+ *   to buffer mapping.
+ *   Buffer's lossy bit is changed to:
+ *     lossless if there is at least one PFC enabled priority mapped to this buffer
+ *     lossy if all priorities mapped to this buffer are PFC disabled
+ *
+ *   Return:
+ *     Return 0 if no error.
+ *     Set change to true if buffer configuration is modified.
+ */
+static int update_buffer_lossy(unsigned int max_mtu,
+			       u8 pfc_en, u8 *buffer, u32 xoff,
+			       struct mlx5e_port_buffer *port_buffer,
+			       bool *change)
+{
+	bool changed = false;
+	u8 lossy_count;
+	u8 prio_count;
+	u8 lossy;
+	int prio;
+	int err;
+	int i;
+
+	for (i = 0; i < MLX5E_MAX_BUFFER; i++) {
+		prio_count = 0;
+		lossy_count = 0;
+
+		for (prio = 0; prio < MLX5E_MAX_PRIORITY; prio++) {
+			if (buffer[prio] != i)
+				continue;
+
+			prio_count++;
+			lossy_count += !(pfc_en & (1 << prio));
+		}
+
+		if (lossy_count == prio_count)
+			lossy = 1;
+		else /* lossy_count < prio_count */
+			lossy = 0;
+
+		if (lossy != port_buffer->buffer[i].lossy) {
+			port_buffer->buffer[i].lossy = lossy;
+			changed = true;
+		}
+	}
+
+	if (changed) {
+		err = update_xoff_threshold(port_buffer, xoff, max_mtu);
+		if (err)
+			return err;
+
+		*change = true;
+	}
+
+	return 0;
+}
+
+static int fill_pfc_en(struct mlx5_core_dev *mdev, u8 *pfc_en)
+{
+	u32 g_rx_pause, g_tx_pause;
+	int err;
+
+	err = mlx5_query_port_pause(mdev, &g_rx_pause, &g_tx_pause);
+	if (err)
+		return err;
+
+	/* If global pause enabled, set all active buffers to lossless.
+	 * Otherwise, check PFC setting.
+	 */
+	if (g_rx_pause || g_tx_pause)
+		*pfc_en = 0xff;
+	else
+		err = mlx5_query_port_pfc(mdev, pfc_en, NULL);
+
+	return err;
+}
+
+#define MINIMUM_MAX_MTU 9216
+int mlx5e_port_manual_buffer_config(struct mlx5e_priv *priv,
+				    u32 change, unsigned int mtu,
+				    struct ieee_pfc *pfc,
+				    u32 *buffer_size,
+				    u8 *prio2buffer)
+{
+	struct mlx5e_port_buffer port_buffer;
+	u32 xoff = calculate_xoff(priv, mtu);
+	bool update_prio2buffer = false;
+	u8 buffer[MLX5E_MAX_PRIORITY];
+	bool update_buffer = false;
+	unsigned int max_mtu;
+	u32 total_used = 0;
+	u8 curr_pfc_en;
+	int err;
+	int i;
+
+	mlx5e_dbg(HW, priv, "%s: change=%x\n", __func__, change);
+	max_mtu = max_t(unsigned int, priv->netdev->max_mtu, MINIMUM_MAX_MTU);
+
+	err = mlx5e_port_query_buffer(priv, &port_buffer);
+	if (err)
+		return err;
+
+	if (change & MLX5E_PORT_BUFFER_CABLE_LEN) {
+		update_buffer = true;
+		err = update_xoff_threshold(&port_buffer, xoff, max_mtu);
+		if (err)
+			return err;
+	}
+
+	if (change & MLX5E_PORT_BUFFER_PFC) {
+		err = mlx5e_port_query_priority2buffer(priv->mdev, buffer);
+		if (err)
+			return err;
+
+		err = update_buffer_lossy(max_mtu, pfc->pfc_en, buffer, xoff,
+					  &port_buffer, &update_buffer);
+		if (err)
+			return err;
+	}
+
+	if (change & MLX5E_PORT_BUFFER_PRIO2BUFFER) {
+		update_prio2buffer = true;
+		err = fill_pfc_en(priv->mdev, &curr_pfc_en);
+		if (err)
+			return err;
+
+		err = update_buffer_lossy(max_mtu, curr_pfc_en, prio2buffer,
+					  xoff, &port_buffer, &update_buffer);
+		if (err)
+			return err;
+	}
+
+	if (change & MLX5E_PORT_BUFFER_SIZE) {
+		for (i = 0; i < MLX5E_MAX_BUFFER; i++) {
+			mlx5e_dbg(HW, priv, "%s: buffer[%d]=%d\n", __func__, i, buffer_size[i]);
+			if (!port_buffer.buffer[i].lossy && !buffer_size[i]) {
+				mlx5e_dbg(HW, priv, "%s: lossless buffer[%d] size cannot be zero\n",
+					  __func__, i);
+				return -EINVAL;
+			}
+
+			port_buffer.buffer[i].size = buffer_size[i];
+			total_used += buffer_size[i];
+		}
+
+		mlx5e_dbg(HW, priv, "%s: total buffer requested=%d\n", __func__, total_used);
+
+		if (total_used > port_buffer.port_buffer_size)
+			return -EINVAL;
+
+		update_buffer = true;
+		err = update_xoff_threshold(&port_buffer, xoff, max_mtu);
+		if (err)
+			return err;
+	}
+
+	/* Need to update buffer configuration if xoff value is changed */
+	if (!update_buffer && xoff != priv->dcbx.xoff) {
+		update_buffer = true;
+		err = update_xoff_threshold(&port_buffer, xoff, max_mtu);
+		if (err)
+			return err;
+	}
+	priv->dcbx.xoff = xoff;
+
+	/* Apply the settings */
+	if (update_buffer) {
+		err = port_set_buffer(priv, &port_buffer);
+		if (err)
+			return err;
+	}
+
+	if (update_prio2buffer)
+		err = mlx5e_port_set_priority2buffer(priv->mdev, prio2buffer);
+
+	return err;
+}
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en/port_buffer.h b/drivers/net/ethernet/mellanox/mlx5/core/en/port_buffer.h
new file mode 100644
index 000000000..34f55b81a
--- /dev/null
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en/port_buffer.h
@@ -0,0 +1,75 @@
+/*
+ * 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.
+ */
+#ifndef __MLX5_EN_PORT_BUFFER_H__
+#define __MLX5_EN_PORT_BUFFER_H__
+
+#include "en.h"
+#include "port.h"
+
+#define MLX5E_MAX_BUFFER 8
+#define MLX5E_BUFFER_CELL_SHIFT 7
+#define MLX5E_DEFAULT_CABLE_LEN 7 /* 7 meters */
+
+#define MLX5_BUFFER_SUPPORTED(mdev) (MLX5_CAP_GEN(mdev, pcam_reg) && \
+				     MLX5_CAP_PCAM_REG(mdev, pbmc) && \
+				     MLX5_CAP_PCAM_REG(mdev, pptb))
+
+enum {
+	MLX5E_PORT_BUFFER_CABLE_LEN   = BIT(0),
+	MLX5E_PORT_BUFFER_PFC         = BIT(1),
+	MLX5E_PORT_BUFFER_PRIO2BUFFER = BIT(2),
+	MLX5E_PORT_BUFFER_SIZE        = BIT(3),
+};
+
+struct mlx5e_bufferx_reg {
+	u8   lossy;
+	u8   epsb;
+	u32  size;
+	u32  xoff;
+	u32  xon;
+};
+
+struct mlx5e_port_buffer {
+	u32                       port_buffer_size;
+	u32                       spare_buffer_size;
+	struct mlx5e_bufferx_reg  buffer[MLX5E_MAX_BUFFER];
+};
+
+int mlx5e_port_manual_buffer_config(struct mlx5e_priv *priv,
+				    u32 change, unsigned int mtu,
+				    struct ieee_pfc *pfc,
+				    u32 *buffer_size,
+				    u8 *prio2buffer);
+
+int mlx5e_port_query_buffer(struct mlx5e_priv *priv,
+			    struct mlx5e_port_buffer *port_buffer);
+#endif
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en/xdp.c b/drivers/net/ethernet/mellanox/mlx5/core/en/xdp.c
new file mode 100644
index 000000000..599114ab7
--- /dev/null
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en/xdp.c
@@ -0,0 +1,325 @@
+/*
+ * 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 <linux/bpf_trace.h>
+#include "en/xdp.h"
+
+int mlx5e_xdp_max_mtu(struct mlx5e_params *params)
+{
+	int hr = NET_IP_ALIGN + XDP_PACKET_HEADROOM;
+
+	/* Let S := SKB_DATA_ALIGN(sizeof(struct skb_shared_info)).
+	 * The condition checked in mlx5e_rx_is_linear_skb is:
+	 *   SKB_DATA_ALIGN(sw_mtu + hard_mtu + hr) + S <= PAGE_SIZE         (1)
+	 *   (Note that hw_mtu == sw_mtu + hard_mtu.)
+	 * What is returned from this function is:
+	 *   max_mtu = PAGE_SIZE - S - hr - hard_mtu                         (2)
+	 * After assigning sw_mtu := max_mtu, the left side of (1) turns to
+	 * SKB_DATA_ALIGN(PAGE_SIZE - S) + S, which is equal to PAGE_SIZE,
+	 * because both PAGE_SIZE and S are already aligned. Any number greater
+	 * than max_mtu would make the left side of (1) greater than PAGE_SIZE,
+	 * so max_mtu is the maximum MTU allowed.
+	 */
+
+	return MLX5E_HW2SW_MTU(params, SKB_MAX_HEAD(hr));
+}
+
+static inline bool
+mlx5e_xmit_xdp_buff(struct mlx5e_xdpsq *sq, struct mlx5e_dma_info *di,
+		    struct xdp_buff *xdp)
+{
+	struct mlx5e_xdp_info xdpi;
+
+	xdpi.xdpf = convert_to_xdp_frame(xdp);
+	if (unlikely(!xdpi.xdpf))
+		return false;
+	xdpi.dma_addr = di->addr + (xdpi.xdpf->data - (void *)xdpi.xdpf);
+	dma_sync_single_for_device(sq->pdev, xdpi.dma_addr,
+				   xdpi.xdpf->len, PCI_DMA_TODEVICE);
+	xdpi.di = *di;
+
+	return mlx5e_xmit_xdp_frame(sq, &xdpi);
+}
+
+/* returns true if packet was consumed by xdp */
+bool mlx5e_xdp_handle(struct mlx5e_rq *rq, struct mlx5e_dma_info *di,
+		      void *va, u16 *rx_headroom, u32 *len)
+{
+	struct bpf_prog *prog = READ_ONCE(rq->xdp_prog);
+	struct xdp_buff xdp;
+	u32 act;
+	int err;
+
+	if (!prog)
+		return false;
+
+	xdp.data = va + *rx_headroom;
+	xdp_set_data_meta_invalid(&xdp);
+	xdp.data_end = xdp.data + *len;
+	xdp.data_hard_start = va;
+	xdp.rxq = &rq->xdp_rxq;
+
+	act = bpf_prog_run_xdp(prog, &xdp);
+	switch (act) {
+	case XDP_PASS:
+		*rx_headroom = xdp.data - xdp.data_hard_start;
+		*len = xdp.data_end - xdp.data;
+		return false;
+	case XDP_TX:
+		if (unlikely(!mlx5e_xmit_xdp_buff(&rq->xdpsq, di, &xdp)))
+			goto xdp_abort;
+		__set_bit(MLX5E_RQ_FLAG_XDP_XMIT, rq->flags); /* non-atomic */
+		return true;
+	case XDP_REDIRECT:
+		/* When XDP enabled then page-refcnt==1 here */
+		err = xdp_do_redirect(rq->netdev, &xdp, prog);
+		if (unlikely(err))
+			goto xdp_abort;
+		__set_bit(MLX5E_RQ_FLAG_XDP_XMIT, rq->flags);
+		rq->xdpsq.redirect_flush = true;
+		mlx5e_page_dma_unmap(rq, di);
+		rq->stats->xdp_redirect++;
+		return true;
+	default:
+		bpf_warn_invalid_xdp_action(act);
+		/* fall through */
+	case XDP_ABORTED:
+xdp_abort:
+		trace_xdp_exception(rq->netdev, prog, act);
+		/* fall through */
+	case XDP_DROP:
+		rq->stats->xdp_drop++;
+		return true;
+	}
+}
+
+bool mlx5e_xmit_xdp_frame(struct mlx5e_xdpsq *sq, struct mlx5e_xdp_info *xdpi)
+{
+	struct mlx5_wq_cyc       *wq   = &sq->wq;
+	u16                       pi   = mlx5_wq_cyc_ctr2ix(wq, sq->pc);
+	struct mlx5e_tx_wqe      *wqe  = mlx5_wq_cyc_get_wqe(wq, pi);
+
+	struct mlx5_wqe_ctrl_seg *cseg = &wqe->ctrl;
+	struct mlx5_wqe_eth_seg  *eseg = &wqe->eth;
+	struct mlx5_wqe_data_seg *dseg = wqe->data;
+
+	struct xdp_frame *xdpf = xdpi->xdpf;
+	dma_addr_t dma_addr  = xdpi->dma_addr;
+	unsigned int dma_len = xdpf->len;
+
+	struct mlx5e_xdpsq_stats *stats = sq->stats;
+
+	prefetchw(wqe);
+
+	if (unlikely(dma_len < MLX5E_XDP_MIN_INLINE || sq->hw_mtu < dma_len)) {
+		stats->err++;
+		return false;
+	}
+
+	if (unlikely(!mlx5e_wqc_has_room_for(wq, sq->cc, sq->pc, 1))) {
+		if (sq->doorbell) {
+			/* SQ is full, ring doorbell */
+			mlx5e_xmit_xdp_doorbell(sq);
+			sq->doorbell = false;
+		}
+		stats->full++;
+		return false;
+	}
+
+	cseg->fm_ce_se = 0;
+
+	/* copy the inline part if required */
+	if (sq->min_inline_mode != MLX5_INLINE_MODE_NONE) {
+		memcpy(eseg->inline_hdr.start, xdpf->data, MLX5E_XDP_MIN_INLINE);
+		eseg->inline_hdr.sz = cpu_to_be16(MLX5E_XDP_MIN_INLINE);
+		dma_len  -= MLX5E_XDP_MIN_INLINE;
+		dma_addr += MLX5E_XDP_MIN_INLINE;
+		dseg++;
+	}
+
+	/* write the dma part */
+	dseg->addr       = cpu_to_be64(dma_addr);
+	dseg->byte_count = cpu_to_be32(dma_len);
+
+	cseg->opmod_idx_opcode = cpu_to_be32((sq->pc << 8) | MLX5_OPCODE_SEND);
+
+	/* move page to reference to sq responsibility,
+	 * and mark so it's not put back in page-cache.
+	 */
+	sq->db.xdpi[pi] = *xdpi;
+	sq->pc++;
+
+	sq->doorbell = true;
+
+	stats->xmit++;
+	return true;
+}
+
+bool mlx5e_poll_xdpsq_cq(struct mlx5e_cq *cq)
+{
+	struct mlx5e_xdpsq *sq;
+	struct mlx5_cqe64 *cqe;
+	struct mlx5e_rq *rq;
+	bool is_redirect;
+	u16 sqcc;
+	int i;
+
+	sq = container_of(cq, struct mlx5e_xdpsq, cq);
+
+	if (unlikely(!test_bit(MLX5E_SQ_STATE_ENABLED, &sq->state)))
+		return false;
+
+	cqe = mlx5_cqwq_get_cqe(&cq->wq);
+	if (!cqe)
+		return false;
+
+	is_redirect = test_bit(MLX5E_SQ_STATE_REDIRECT, &sq->state);
+	rq = container_of(sq, struct mlx5e_rq, xdpsq);
+
+	/* sq->cc must be updated only after mlx5_cqwq_update_db_record(),
+	 * otherwise a cq overrun may occur
+	 */
+	sqcc = sq->cc;
+
+	i = 0;
+	do {
+		u16 wqe_counter;
+		bool last_wqe;
+
+		mlx5_cqwq_pop(&cq->wq);
+
+		wqe_counter = be16_to_cpu(cqe->wqe_counter);
+
+		do {
+			u16 ci = mlx5_wq_cyc_ctr2ix(&sq->wq, sqcc);
+			struct mlx5e_xdp_info *xdpi = &sq->db.xdpi[ci];
+
+			last_wqe = (sqcc == wqe_counter);
+			sqcc++;
+
+			if (is_redirect) {
+				dma_unmap_single(sq->pdev, xdpi->dma_addr,
+						 xdpi->xdpf->len, DMA_TO_DEVICE);
+				xdp_return_frame(xdpi->xdpf);
+			} else {
+				/* Recycle RX page */
+				mlx5e_page_release(rq, &xdpi->di, true);
+			}
+		} while (!last_wqe);
+	} while ((++i < MLX5E_TX_CQ_POLL_BUDGET) && (cqe = mlx5_cqwq_get_cqe(&cq->wq)));
+
+	sq->stats->cqes += i;
+
+	mlx5_cqwq_update_db_record(&cq->wq);
+
+	/* ensure cq space is freed before enabling more cqes */
+	wmb();
+
+	sq->cc = sqcc;
+	return (i == MLX5E_TX_CQ_POLL_BUDGET);
+}
+
+void mlx5e_free_xdpsq_descs(struct mlx5e_xdpsq *sq)
+{
+	struct mlx5e_rq *rq;
+	bool is_redirect;
+
+	is_redirect = test_bit(MLX5E_SQ_STATE_REDIRECT, &sq->state);
+	rq = is_redirect ? NULL : container_of(sq, struct mlx5e_rq, xdpsq);
+
+	while (sq->cc != sq->pc) {
+		u16 ci = mlx5_wq_cyc_ctr2ix(&sq->wq, sq->cc);
+		struct mlx5e_xdp_info *xdpi = &sq->db.xdpi[ci];
+
+		sq->cc++;
+
+		if (is_redirect) {
+			dma_unmap_single(sq->pdev, xdpi->dma_addr,
+					 xdpi->xdpf->len, DMA_TO_DEVICE);
+			xdp_return_frame(xdpi->xdpf);
+		} else {
+			/* Recycle RX page */
+			mlx5e_page_release(rq, &xdpi->di, false);
+		}
+	}
+}
+
+int mlx5e_xdp_xmit(struct net_device *dev, int n, struct xdp_frame **frames,
+		   u32 flags)
+{
+	struct mlx5e_priv *priv = netdev_priv(dev);
+	struct mlx5e_xdpsq *sq;
+	int drops = 0;
+	int sq_num;
+	int i;
+
+	/* this flag is sufficient, no need to test internal sq state */
+	if (unlikely(!mlx5e_xdp_tx_is_enabled(priv)))
+		return -ENETDOWN;
+
+	if (unlikely(flags & ~XDP_XMIT_FLAGS_MASK))
+		return -EINVAL;
+
+	sq_num = smp_processor_id();
+
+	if (unlikely(sq_num >= priv->channels.num))
+		return -ENXIO;
+
+	sq = &priv->channels.c[sq_num]->xdpsq;
+
+	for (i = 0; i < n; i++) {
+		struct xdp_frame *xdpf = frames[i];
+		struct mlx5e_xdp_info xdpi;
+
+		xdpi.dma_addr = dma_map_single(sq->pdev, xdpf->data, xdpf->len,
+					       DMA_TO_DEVICE);
+		if (unlikely(dma_mapping_error(sq->pdev, xdpi.dma_addr))) {
+			xdp_return_frame_rx_napi(xdpf);
+			drops++;
+			continue;
+		}
+
+		xdpi.xdpf = xdpf;
+
+		if (unlikely(!mlx5e_xmit_xdp_frame(sq, &xdpi))) {
+			dma_unmap_single(sq->pdev, xdpi.dma_addr,
+					 xdpf->len, DMA_TO_DEVICE);
+			xdp_return_frame_rx_napi(xdpf);
+			drops++;
+		}
+	}
+
+	if (flags & XDP_XMIT_FLUSH)
+		mlx5e_xmit_xdp_doorbell(sq);
+
+	return n - drops;
+}
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en/xdp.h b/drivers/net/ethernet/mellanox/mlx5/core/en/xdp.h
new file mode 100644
index 000000000..827ceef5f
--- /dev/null
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en/xdp.h
@@ -0,0 +1,79 @@
+/*
+ * 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.
+ */
+#ifndef __MLX5_EN_XDP_H__
+#define __MLX5_EN_XDP_H__
+
+#include "en.h"
+
+#define MLX5E_XDP_MIN_INLINE (ETH_HLEN + VLAN_HLEN)
+#define MLX5E_XDP_TX_DS_COUNT \
+	((sizeof(struct mlx5e_tx_wqe) / MLX5_SEND_WQE_DS) + 1 /* SG DS */)
+
+int mlx5e_xdp_max_mtu(struct mlx5e_params *params);
+bool mlx5e_xdp_handle(struct mlx5e_rq *rq, struct mlx5e_dma_info *di,
+		      void *va, u16 *rx_headroom, u32 *len);
+bool mlx5e_poll_xdpsq_cq(struct mlx5e_cq *cq);
+void mlx5e_free_xdpsq_descs(struct mlx5e_xdpsq *sq);
+
+bool mlx5e_xmit_xdp_frame(struct mlx5e_xdpsq *sq, struct mlx5e_xdp_info *xdpi);
+int mlx5e_xdp_xmit(struct net_device *dev, int n, struct xdp_frame **frames,
+		   u32 flags);
+
+static inline void mlx5e_xdp_tx_enable(struct mlx5e_priv *priv)
+{
+	set_bit(MLX5E_STATE_XDP_TX_ENABLED, &priv->state);
+}
+
+static inline void mlx5e_xdp_tx_disable(struct mlx5e_priv *priv)
+{
+	clear_bit(MLX5E_STATE_XDP_TX_ENABLED, &priv->state);
+	/* let other device's napi(s) see our new state */
+	synchronize_rcu();
+}
+
+static inline bool mlx5e_xdp_tx_is_enabled(struct mlx5e_priv *priv)
+{
+	return test_bit(MLX5E_STATE_XDP_TX_ENABLED, &priv->state);
+}
+
+static inline void mlx5e_xmit_xdp_doorbell(struct mlx5e_xdpsq *sq)
+{
+	struct mlx5_wq_cyc *wq = &sq->wq;
+	struct mlx5e_tx_wqe *wqe;
+	u16 pi = mlx5_wq_cyc_ctr2ix(wq, sq->pc - 1); /* last pi */
+
+	wqe  = mlx5_wq_cyc_get_wqe(wq, pi);
+
+	mlx5e_notify_hw(wq, sq->pc, sq->uar_map, &wqe->ctrl);
+}
+
+#endif
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_accel/Makefile b/drivers/net/ethernet/mellanox/mlx5/core/en_accel/Makefile
new file mode 100644
index 000000000..d8e17110f
--- /dev/null
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_accel/Makefile
@@ -0,0 +1 @@
+subdir-ccflags-y += -I$(src)/..
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_accel/en_accel.h b/drivers/net/ethernet/mellanox/mlx5/core/en_accel/en_accel.h
new file mode 100644
index 000000000..1dd225380
--- /dev/null
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_accel/en_accel.h
@@ -0,0 +1,80 @@
+/*
+ * 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.
+ *
+ */
+
+#ifndef __MLX5E_EN_ACCEL_H__
+#define __MLX5E_EN_ACCEL_H__
+
+#include <linux/skbuff.h>
+#include <linux/netdevice.h>
+#include "en_accel/ipsec_rxtx.h"
+#include "en_accel/tls_rxtx.h"
+#include "en.h"
+
+static inline void
+mlx5e_udp_gso_handle_tx_skb(struct sk_buff *skb)
+{
+	int payload_len = skb_shinfo(skb)->gso_size + sizeof(struct udphdr);
+
+	udp_hdr(skb)->len = htons(payload_len);
+}
+
+static inline struct sk_buff *
+mlx5e_accel_handle_tx(struct sk_buff *skb,
+		      struct mlx5e_txqsq *sq,
+		      struct net_device *dev,
+		      struct mlx5e_tx_wqe **wqe,
+		      u16 *pi)
+{
+#ifdef CONFIG_MLX5_EN_TLS
+	if (test_bit(MLX5E_SQ_STATE_TLS, &sq->state)) {
+		skb = mlx5e_tls_handle_tx_skb(dev, sq, skb, wqe, pi);
+		if (unlikely(!skb))
+			return NULL;
+	}
+#endif
+
+#ifdef CONFIG_MLX5_EN_IPSEC
+	if (test_bit(MLX5E_SQ_STATE_IPSEC, &sq->state)) {
+		skb = mlx5e_ipsec_handle_tx_skb(dev, *wqe, skb);
+		if (unlikely(!skb))
+			return NULL;
+	}
+#endif
+
+	if (skb_is_gso(skb) && skb_shinfo(skb)->gso_type & SKB_GSO_UDP_L4)
+		mlx5e_udp_gso_handle_tx_skb(skb);
+
+	return skb;
+}
+
+#endif /* __MLX5E_EN_ACCEL_H__ */
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_accel/ipsec.c b/drivers/net/ethernet/mellanox/mlx5/core/en_accel/ipsec.c
new file mode 100644
index 000000000..c467f5e98
--- /dev/null
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_accel/ipsec.c
@@ -0,0 +1,547 @@
+/*
+ * Copyright (c) 2017 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 <crypto/internal/geniv.h>
+#include <crypto/aead.h>
+#include <linux/inetdevice.h>
+#include <linux/netdevice.h>
+#include <linux/module.h>
+
+#include "en.h"
+#include "en_accel/ipsec.h"
+#include "en_accel/ipsec_rxtx.h"
+
+
+static struct mlx5e_ipsec_sa_entry *to_ipsec_sa_entry(struct xfrm_state *x)
+{
+	struct mlx5e_ipsec_sa_entry *sa;
+
+	if (!x)
+		return NULL;
+
+	sa = (struct mlx5e_ipsec_sa_entry *)x->xso.offload_handle;
+	if (!sa)
+		return NULL;
+
+	WARN_ON(sa->x != x);
+	return sa;
+}
+
+struct xfrm_state *mlx5e_ipsec_sadb_rx_lookup(struct mlx5e_ipsec *ipsec,
+					      unsigned int handle)
+{
+	struct mlx5e_ipsec_sa_entry *sa_entry;
+	struct xfrm_state *ret = NULL;
+
+	rcu_read_lock();
+	hash_for_each_possible_rcu(ipsec->sadb_rx, sa_entry, hlist, handle)
+		if (sa_entry->handle == handle) {
+			ret = sa_entry->x;
+			xfrm_state_hold(ret);
+			break;
+		}
+	rcu_read_unlock();
+
+	return ret;
+}
+
+static int mlx5e_ipsec_sadb_rx_add(struct mlx5e_ipsec_sa_entry *sa_entry)
+{
+	struct mlx5e_ipsec *ipsec = sa_entry->ipsec;
+	unsigned long flags;
+	int ret;
+
+	ret = ida_simple_get(&ipsec->halloc, 1, 0, GFP_KERNEL);
+	if (ret < 0)
+		return ret;
+
+	spin_lock_irqsave(&ipsec->sadb_rx_lock, flags);
+	sa_entry->handle = ret;
+	hash_add_rcu(ipsec->sadb_rx, &sa_entry->hlist, sa_entry->handle);
+	spin_unlock_irqrestore(&ipsec->sadb_rx_lock, flags);
+
+	return 0;
+}
+
+static void mlx5e_ipsec_sadb_rx_del(struct mlx5e_ipsec_sa_entry *sa_entry)
+{
+	struct mlx5e_ipsec *ipsec = sa_entry->ipsec;
+	unsigned long flags;
+
+	spin_lock_irqsave(&ipsec->sadb_rx_lock, flags);
+	hash_del_rcu(&sa_entry->hlist);
+	spin_unlock_irqrestore(&ipsec->sadb_rx_lock, flags);
+}
+
+static void mlx5e_ipsec_sadb_rx_free(struct mlx5e_ipsec_sa_entry *sa_entry)
+{
+	struct mlx5e_ipsec *ipsec = sa_entry->ipsec;
+
+	/* xfrm already doing sync rcu between del and free callbacks */
+
+	ida_simple_remove(&ipsec->halloc, sa_entry->handle);
+}
+
+static bool mlx5e_ipsec_update_esn_state(struct mlx5e_ipsec_sa_entry *sa_entry)
+{
+	struct xfrm_replay_state_esn *replay_esn;
+	u32 seq_bottom;
+	u8 overlap;
+	u32 *esn;
+
+	if (!(sa_entry->x->props.flags & XFRM_STATE_ESN)) {
+		sa_entry->esn_state.trigger = 0;
+		return false;
+	}
+
+	replay_esn = sa_entry->x->replay_esn;
+	seq_bottom = replay_esn->seq - replay_esn->replay_window + 1;
+	overlap = sa_entry->esn_state.overlap;
+
+	sa_entry->esn_state.esn = xfrm_replay_seqhi(sa_entry->x,
+						    htonl(seq_bottom));
+	esn = &sa_entry->esn_state.esn;
+
+	sa_entry->esn_state.trigger = 1;
+	if (unlikely(overlap && seq_bottom < MLX5E_IPSEC_ESN_SCOPE_MID)) {
+		++(*esn);
+		sa_entry->esn_state.overlap = 0;
+		return true;
+	} else if (unlikely(!overlap &&
+			    (seq_bottom >= MLX5E_IPSEC_ESN_SCOPE_MID))) {
+		sa_entry->esn_state.overlap = 1;
+		return true;
+	}
+
+	return false;
+}
+
+static void
+mlx5e_ipsec_build_accel_xfrm_attrs(struct mlx5e_ipsec_sa_entry *sa_entry,
+				   struct mlx5_accel_esp_xfrm_attrs *attrs)
+{
+	struct xfrm_state *x = sa_entry->x;
+	struct aes_gcm_keymat *aes_gcm = &attrs->keymat.aes_gcm;
+	struct aead_geniv_ctx *geniv_ctx;
+	struct crypto_aead *aead;
+	unsigned int crypto_data_len, key_len;
+	int ivsize;
+
+	memset(attrs, 0, sizeof(*attrs));
+
+	/* key */
+	crypto_data_len = (x->aead->alg_key_len + 7) / 8;
+	key_len = crypto_data_len - 4; /* 4 bytes salt at end */
+
+	memcpy(aes_gcm->aes_key, x->aead->alg_key, key_len);
+	aes_gcm->key_len = key_len * 8;
+
+	/* salt and seq_iv */
+	aead = x->data;
+	geniv_ctx = crypto_aead_ctx(aead);
+	ivsize = crypto_aead_ivsize(aead);
+	memcpy(&aes_gcm->seq_iv, &geniv_ctx->salt, ivsize);
+	memcpy(&aes_gcm->salt, x->aead->alg_key + key_len,
+	       sizeof(aes_gcm->salt));
+
+	/* iv len */
+	aes_gcm->icv_len = x->aead->alg_icv_len;
+
+	/* esn */
+	if (sa_entry->esn_state.trigger) {
+		attrs->flags |= MLX5_ACCEL_ESP_FLAGS_ESN_TRIGGERED;
+		attrs->esn = sa_entry->esn_state.esn;
+		if (sa_entry->esn_state.overlap)
+			attrs->flags |= MLX5_ACCEL_ESP_FLAGS_ESN_STATE_OVERLAP;
+	}
+
+	/* rx handle */
+	attrs->sa_handle = sa_entry->handle;
+
+	/* algo type */
+	attrs->keymat_type = MLX5_ACCEL_ESP_KEYMAT_AES_GCM;
+
+	/* action */
+	attrs->action = (!(x->xso.flags & XFRM_OFFLOAD_INBOUND)) ?
+			MLX5_ACCEL_ESP_ACTION_ENCRYPT :
+			MLX5_ACCEL_ESP_ACTION_DECRYPT;
+	/* flags */
+	attrs->flags |= (x->props.mode == XFRM_MODE_TRANSPORT) ?
+			MLX5_ACCEL_ESP_FLAGS_TRANSPORT :
+			MLX5_ACCEL_ESP_FLAGS_TUNNEL;
+}
+
+static inline int mlx5e_xfrm_validate_state(struct xfrm_state *x)
+{
+	struct net_device *netdev = x->xso.dev;
+	struct mlx5e_priv *priv;
+
+	priv = netdev_priv(netdev);
+
+	if (x->props.aalgo != SADB_AALG_NONE) {
+		netdev_info(netdev, "Cannot offload authenticated xfrm states\n");
+		return -EINVAL;
+	}
+	if (x->props.ealgo != SADB_X_EALG_AES_GCM_ICV16) {
+		netdev_info(netdev, "Only AES-GCM-ICV16 xfrm state may be offloaded\n");
+		return -EINVAL;
+	}
+	if (x->props.calgo != SADB_X_CALG_NONE) {
+		netdev_info(netdev, "Cannot offload compressed xfrm states\n");
+		return -EINVAL;
+	}
+	if (x->props.flags & XFRM_STATE_ESN &&
+	    !(mlx5_accel_ipsec_device_caps(priv->mdev) &
+	    MLX5_ACCEL_IPSEC_CAP_ESN)) {
+		netdev_info(netdev, "Cannot offload ESN xfrm states\n");
+		return -EINVAL;
+	}
+	if (x->props.family != AF_INET &&
+	    x->props.family != AF_INET6) {
+		netdev_info(netdev, "Only IPv4/6 xfrm states may be offloaded\n");
+		return -EINVAL;
+	}
+	if (x->props.mode != XFRM_MODE_TRANSPORT &&
+	    x->props.mode != XFRM_MODE_TUNNEL) {
+		dev_info(&netdev->dev, "Only transport and tunnel xfrm states may be offloaded\n");
+		return -EINVAL;
+	}
+	if (x->id.proto != IPPROTO_ESP) {
+		netdev_info(netdev, "Only ESP xfrm state may be offloaded\n");
+		return -EINVAL;
+	}
+	if (x->encap) {
+		netdev_info(netdev, "Encapsulated xfrm state may not be offloaded\n");
+		return -EINVAL;
+	}
+	if (!x->aead) {
+		netdev_info(netdev, "Cannot offload xfrm states without aead\n");
+		return -EINVAL;
+	}
+	if (x->aead->alg_icv_len != 128) {
+		netdev_info(netdev, "Cannot offload xfrm states with AEAD ICV length other than 128bit\n");
+		return -EINVAL;
+	}
+	if ((x->aead->alg_key_len != 128 + 32) &&
+	    (x->aead->alg_key_len != 256 + 32)) {
+		netdev_info(netdev, "Cannot offload xfrm states with AEAD key length other than 128/256 bit\n");
+		return -EINVAL;
+	}
+	if (x->tfcpad) {
+		netdev_info(netdev, "Cannot offload xfrm states with tfc padding\n");
+		return -EINVAL;
+	}
+	if (!x->geniv) {
+		netdev_info(netdev, "Cannot offload xfrm states without geniv\n");
+		return -EINVAL;
+	}
+	if (strcmp(x->geniv, "seqiv")) {
+		netdev_info(netdev, "Cannot offload xfrm states with geniv other than seqiv\n");
+		return -EINVAL;
+	}
+	if (x->props.family == AF_INET6 &&
+	    !(mlx5_accel_ipsec_device_caps(priv->mdev) &
+	     MLX5_ACCEL_IPSEC_CAP_IPV6)) {
+		netdev_info(netdev, "IPv6 xfrm state offload is not supported by this device\n");
+		return -EINVAL;
+	}
+	return 0;
+}
+
+static int mlx5e_xfrm_add_state(struct xfrm_state *x)
+{
+	struct mlx5e_ipsec_sa_entry *sa_entry = NULL;
+	struct net_device *netdev = x->xso.dev;
+	struct mlx5_accel_esp_xfrm_attrs attrs;
+	struct mlx5e_priv *priv;
+	__be32 saddr[4] = {0}, daddr[4] = {0}, spi;
+	bool is_ipv6 = false;
+	int err;
+
+	priv = netdev_priv(netdev);
+
+	err = mlx5e_xfrm_validate_state(x);
+	if (err)
+		return err;
+
+	sa_entry = kzalloc(sizeof(*sa_entry), GFP_KERNEL);
+	if (!sa_entry) {
+		err = -ENOMEM;
+		goto out;
+	}
+
+	sa_entry->x = x;
+	sa_entry->ipsec = priv->ipsec;
+
+	/* Add the SA to handle processed incoming packets before the add SA
+	 * completion was received
+	 */
+	if (x->xso.flags & XFRM_OFFLOAD_INBOUND) {
+		err = mlx5e_ipsec_sadb_rx_add(sa_entry);
+		if (err) {
+			netdev_info(netdev, "Failed adding to SADB_RX: %d\n", err);
+			goto err_entry;
+		}
+	} else {
+		sa_entry->set_iv_op = (x->props.flags & XFRM_STATE_ESN) ?
+				mlx5e_ipsec_set_iv_esn : mlx5e_ipsec_set_iv;
+	}
+
+	/* check esn */
+	mlx5e_ipsec_update_esn_state(sa_entry);
+
+	/* create xfrm */
+	mlx5e_ipsec_build_accel_xfrm_attrs(sa_entry, &attrs);
+	sa_entry->xfrm =
+		mlx5_accel_esp_create_xfrm(priv->mdev, &attrs,
+					   MLX5_ACCEL_XFRM_FLAG_REQUIRE_METADATA);
+	if (IS_ERR(sa_entry->xfrm)) {
+		err = PTR_ERR(sa_entry->xfrm);
+		goto err_sadb_rx;
+	}
+
+	/* create hw context */
+	if (x->props.family == AF_INET) {
+		saddr[3] = x->props.saddr.a4;
+		daddr[3] = x->id.daddr.a4;
+	} else {
+		memcpy(saddr, x->props.saddr.a6, sizeof(saddr));
+		memcpy(daddr, x->id.daddr.a6, sizeof(daddr));
+		is_ipv6 = true;
+	}
+	spi = x->id.spi;
+	sa_entry->hw_context =
+			mlx5_accel_esp_create_hw_context(priv->mdev,
+							 sa_entry->xfrm,
+							 saddr, daddr, spi,
+							 is_ipv6);
+	if (IS_ERR(sa_entry->hw_context)) {
+		err = PTR_ERR(sa_entry->hw_context);
+		goto err_xfrm;
+	}
+
+	x->xso.offload_handle = (unsigned long)sa_entry;
+	goto out;
+
+err_xfrm:
+	mlx5_accel_esp_destroy_xfrm(sa_entry->xfrm);
+err_sadb_rx:
+	if (x->xso.flags & XFRM_OFFLOAD_INBOUND) {
+		mlx5e_ipsec_sadb_rx_del(sa_entry);
+		mlx5e_ipsec_sadb_rx_free(sa_entry);
+	}
+err_entry:
+	kfree(sa_entry);
+out:
+	return err;
+}
+
+static void mlx5e_xfrm_del_state(struct xfrm_state *x)
+{
+	struct mlx5e_ipsec_sa_entry *sa_entry = to_ipsec_sa_entry(x);
+
+	if (!sa_entry)
+		return;
+
+	if (x->xso.flags & XFRM_OFFLOAD_INBOUND)
+		mlx5e_ipsec_sadb_rx_del(sa_entry);
+}
+
+static void mlx5e_xfrm_free_state(struct xfrm_state *x)
+{
+	struct mlx5e_ipsec_sa_entry *sa_entry = to_ipsec_sa_entry(x);
+
+	if (!sa_entry)
+		return;
+
+	if (sa_entry->hw_context) {
+		flush_workqueue(sa_entry->ipsec->wq);
+		mlx5_accel_esp_free_hw_context(sa_entry->hw_context);
+		mlx5_accel_esp_destroy_xfrm(sa_entry->xfrm);
+	}
+
+	if (x->xso.flags & XFRM_OFFLOAD_INBOUND)
+		mlx5e_ipsec_sadb_rx_free(sa_entry);
+
+	kfree(sa_entry);
+}
+
+int mlx5e_ipsec_init(struct mlx5e_priv *priv)
+{
+	struct mlx5e_ipsec *ipsec = NULL;
+
+	if (!MLX5_IPSEC_DEV(priv->mdev)) {
+		netdev_dbg(priv->netdev, "Not an IPSec offload device\n");
+		return 0;
+	}
+
+	ipsec = kzalloc(sizeof(*ipsec), GFP_KERNEL);
+	if (!ipsec)
+		return -ENOMEM;
+
+	hash_init(ipsec->sadb_rx);
+	spin_lock_init(&ipsec->sadb_rx_lock);
+	ida_init(&ipsec->halloc);
+	ipsec->en_priv = priv;
+	ipsec->en_priv->ipsec = ipsec;
+	ipsec->no_trailer = !!(mlx5_accel_ipsec_device_caps(priv->mdev) &
+			       MLX5_ACCEL_IPSEC_CAP_RX_NO_TRAILER);
+	ipsec->wq = alloc_ordered_workqueue("mlx5e_ipsec: %s", 0,
+					    priv->netdev->name);
+	if (!ipsec->wq) {
+		kfree(ipsec);
+		return -ENOMEM;
+	}
+	netdev_dbg(priv->netdev, "IPSec attached to netdevice\n");
+	return 0;
+}
+
+void mlx5e_ipsec_cleanup(struct mlx5e_priv *priv)
+{
+	struct mlx5e_ipsec *ipsec = priv->ipsec;
+
+	if (!ipsec)
+		return;
+
+	drain_workqueue(ipsec->wq);
+	destroy_workqueue(ipsec->wq);
+
+	ida_destroy(&ipsec->halloc);
+	kfree(ipsec);
+	priv->ipsec = NULL;
+}
+
+static bool mlx5e_ipsec_offload_ok(struct sk_buff *skb, struct xfrm_state *x)
+{
+	if (x->props.family == AF_INET) {
+		/* Offload with IPv4 options is not supported yet */
+		if (ip_hdr(skb)->ihl > 5)
+			return false;
+	} else {
+		/* Offload with IPv6 extension headers is not support yet */
+		if (ipv6_ext_hdr(ipv6_hdr(skb)->nexthdr))
+			return false;
+	}
+
+	return true;
+}
+
+struct mlx5e_ipsec_modify_state_work {
+	struct work_struct		work;
+	struct mlx5_accel_esp_xfrm_attrs attrs;
+	struct mlx5e_ipsec_sa_entry	*sa_entry;
+};
+
+static void _update_xfrm_state(struct work_struct *work)
+{
+	int ret;
+	struct mlx5e_ipsec_modify_state_work *modify_work =
+		container_of(work, struct mlx5e_ipsec_modify_state_work, work);
+	struct mlx5e_ipsec_sa_entry *sa_entry = modify_work->sa_entry;
+
+	ret = mlx5_accel_esp_modify_xfrm(sa_entry->xfrm,
+					 &modify_work->attrs);
+	if (ret)
+		netdev_warn(sa_entry->ipsec->en_priv->netdev,
+			    "Not an IPSec offload device\n");
+
+	kfree(modify_work);
+}
+
+static void mlx5e_xfrm_advance_esn_state(struct xfrm_state *x)
+{
+	struct mlx5e_ipsec_sa_entry *sa_entry = to_ipsec_sa_entry(x);
+	struct mlx5e_ipsec_modify_state_work *modify_work;
+	bool need_update;
+
+	if (!sa_entry)
+		return;
+
+	need_update = mlx5e_ipsec_update_esn_state(sa_entry);
+	if (!need_update)
+		return;
+
+	modify_work = kzalloc(sizeof(*modify_work), GFP_ATOMIC);
+	if (!modify_work)
+		return;
+
+	mlx5e_ipsec_build_accel_xfrm_attrs(sa_entry, &modify_work->attrs);
+	modify_work->sa_entry = sa_entry;
+
+	INIT_WORK(&modify_work->work, _update_xfrm_state);
+	WARN_ON(!queue_work(sa_entry->ipsec->wq, &modify_work->work));
+}
+
+static const struct xfrmdev_ops mlx5e_ipsec_xfrmdev_ops = {
+	.xdo_dev_state_add	= mlx5e_xfrm_add_state,
+	.xdo_dev_state_delete	= mlx5e_xfrm_del_state,
+	.xdo_dev_state_free	= mlx5e_xfrm_free_state,
+	.xdo_dev_offload_ok	= mlx5e_ipsec_offload_ok,
+	.xdo_dev_state_advance_esn = mlx5e_xfrm_advance_esn_state,
+};
+
+void mlx5e_ipsec_build_netdev(struct mlx5e_priv *priv)
+{
+	struct mlx5_core_dev *mdev = priv->mdev;
+	struct net_device *netdev = priv->netdev;
+
+	if (!(mlx5_accel_ipsec_device_caps(mdev) & MLX5_ACCEL_IPSEC_CAP_ESP) ||
+	    !MLX5_CAP_ETH(mdev, swp)) {
+		mlx5_core_dbg(mdev, "mlx5e: ESP and SWP offload not supported\n");
+		return;
+	}
+
+	mlx5_core_info(mdev, "mlx5e: IPSec ESP acceleration enabled\n");
+	netdev->xfrmdev_ops = &mlx5e_ipsec_xfrmdev_ops;
+	netdev->features |= NETIF_F_HW_ESP;
+	netdev->hw_enc_features |= NETIF_F_HW_ESP;
+
+	if (!MLX5_CAP_ETH(mdev, swp_csum)) {
+		mlx5_core_dbg(mdev, "mlx5e: SWP checksum not supported\n");
+		return;
+	}
+
+	netdev->features |= NETIF_F_HW_ESP_TX_CSUM;
+	netdev->hw_enc_features |= NETIF_F_HW_ESP_TX_CSUM;
+
+	if (!(mlx5_accel_ipsec_device_caps(mdev) & MLX5_ACCEL_IPSEC_CAP_LSO) ||
+	    !MLX5_CAP_ETH(mdev, swp_lso)) {
+		mlx5_core_dbg(mdev, "mlx5e: ESP LSO not supported\n");
+		return;
+	}
+
+	mlx5_core_dbg(mdev, "mlx5e: ESP GSO capability turned on\n");
+	netdev->features |= NETIF_F_GSO_ESP;
+	netdev->hw_features |= NETIF_F_GSO_ESP;
+	netdev->hw_enc_features |= NETIF_F_GSO_ESP;
+}
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_accel/ipsec.h b/drivers/net/ethernet/mellanox/mlx5/core/en_accel/ipsec.h
new file mode 100644
index 000000000..93bf10e65
--- /dev/null
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_accel/ipsec.h
@@ -0,0 +1,161 @@
+/*
+ * Copyright (c) 2017 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 __MLX5E_IPSEC_H__
+#define __MLX5E_IPSEC_H__
+
+#ifdef CONFIG_MLX5_EN_IPSEC
+
+#include <linux/mlx5/device.h>
+#include <net/xfrm.h>
+#include <linux/idr.h>
+
+#include "accel/ipsec.h"
+
+#define MLX5E_IPSEC_SADB_RX_BITS 10
+#define MLX5E_IPSEC_ESN_SCOPE_MID 0x80000000L
+
+struct mlx5e_priv;
+
+struct mlx5e_ipsec_sw_stats {
+	atomic64_t ipsec_rx_drop_sp_alloc;
+	atomic64_t ipsec_rx_drop_sadb_miss;
+	atomic64_t ipsec_rx_drop_syndrome;
+	atomic64_t ipsec_tx_drop_bundle;
+	atomic64_t ipsec_tx_drop_no_state;
+	atomic64_t ipsec_tx_drop_not_ip;
+	atomic64_t ipsec_tx_drop_trailer;
+	atomic64_t ipsec_tx_drop_metadata;
+};
+
+struct mlx5e_ipsec_stats {
+	u64 ipsec_dec_in_packets;
+	u64 ipsec_dec_out_packets;
+	u64 ipsec_dec_bypass_packets;
+	u64 ipsec_enc_in_packets;
+	u64 ipsec_enc_out_packets;
+	u64 ipsec_enc_bypass_packets;
+	u64 ipsec_dec_drop_packets;
+	u64 ipsec_dec_auth_fail_packets;
+	u64 ipsec_enc_drop_packets;
+	u64 ipsec_add_sa_success;
+	u64 ipsec_add_sa_fail;
+	u64 ipsec_del_sa_success;
+	u64 ipsec_del_sa_fail;
+	u64 ipsec_cmd_drop;
+};
+
+struct mlx5e_ipsec {
+	struct mlx5e_priv *en_priv;
+	DECLARE_HASHTABLE(sadb_rx, MLX5E_IPSEC_SADB_RX_BITS);
+	bool no_trailer;
+	spinlock_t sadb_rx_lock; /* Protects sadb_rx and halloc */
+	struct ida halloc;
+	struct mlx5e_ipsec_sw_stats sw_stats;
+	struct mlx5e_ipsec_stats stats;
+	struct workqueue_struct *wq;
+};
+
+struct mlx5e_ipsec_esn_state {
+	u32 esn;
+	u8 trigger: 1;
+	u8 overlap: 1;
+};
+
+struct mlx5e_ipsec_sa_entry {
+	struct hlist_node hlist; /* Item in SADB_RX hashtable */
+	struct mlx5e_ipsec_esn_state esn_state;
+	unsigned int handle; /* Handle in SADB_RX */
+	struct xfrm_state *x;
+	struct mlx5e_ipsec *ipsec;
+	struct mlx5_accel_esp_xfrm *xfrm;
+	void *hw_context;
+	void (*set_iv_op)(struct sk_buff *skb, struct xfrm_state *x,
+			  struct xfrm_offload *xo);
+};
+
+void mlx5e_ipsec_build_inverse_table(void);
+int mlx5e_ipsec_init(struct mlx5e_priv *priv);
+void mlx5e_ipsec_cleanup(struct mlx5e_priv *priv);
+void mlx5e_ipsec_build_netdev(struct mlx5e_priv *priv);
+
+int mlx5e_ipsec_get_count(struct mlx5e_priv *priv);
+int mlx5e_ipsec_get_strings(struct mlx5e_priv *priv, uint8_t *data);
+void mlx5e_ipsec_update_stats(struct mlx5e_priv *priv);
+int mlx5e_ipsec_get_stats(struct mlx5e_priv *priv, u64 *data);
+
+struct xfrm_state *mlx5e_ipsec_sadb_rx_lookup(struct mlx5e_ipsec *dev,
+					      unsigned int handle);
+
+#else
+
+static inline void mlx5e_ipsec_build_inverse_table(void)
+{
+}
+
+static inline int mlx5e_ipsec_init(struct mlx5e_priv *priv)
+{
+	return 0;
+}
+
+static inline void mlx5e_ipsec_cleanup(struct mlx5e_priv *priv)
+{
+}
+
+static inline void mlx5e_ipsec_build_netdev(struct mlx5e_priv *priv)
+{
+}
+
+static inline int mlx5e_ipsec_get_count(struct mlx5e_priv *priv)
+{
+	return 0;
+}
+
+static inline int mlx5e_ipsec_get_strings(struct mlx5e_priv *priv,
+					  uint8_t *data)
+{
+	return 0;
+}
+
+static inline void mlx5e_ipsec_update_stats(struct mlx5e_priv *priv)
+{
+}
+
+static inline int mlx5e_ipsec_get_stats(struct mlx5e_priv *priv, u64 *data)
+{
+	return 0;
+}
+
+#endif
+
+#endif	/* __MLX5E_IPSEC_H__ */
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_accel/ipsec_rxtx.c b/drivers/net/ethernet/mellanox/mlx5/core/en_accel/ipsec_rxtx.c
new file mode 100644
index 000000000..128a82b1d
--- /dev/null
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_accel/ipsec_rxtx.c
@@ -0,0 +1,401 @@
+/*
+ * Copyright (c) 2017 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 <crypto/aead.h>
+#include <net/xfrm.h>
+#include <net/esp.h>
+
+#include "en_accel/ipsec_rxtx.h"
+#include "en_accel/ipsec.h"
+#include "accel/accel.h"
+#include "en.h"
+
+enum {
+	MLX5E_IPSEC_RX_SYNDROME_DECRYPTED = 0x11,
+	MLX5E_IPSEC_RX_SYNDROME_AUTH_FAILED = 0x12,
+	MLX5E_IPSEC_RX_SYNDROME_BAD_PROTO = 0x17,
+};
+
+struct mlx5e_ipsec_rx_metadata {
+	unsigned char   nexthdr;
+	__be32		sa_handle;
+} __packed;
+
+enum {
+	MLX5E_IPSEC_TX_SYNDROME_OFFLOAD = 0x8,
+	MLX5E_IPSEC_TX_SYNDROME_OFFLOAD_WITH_LSO_TCP = 0x9,
+};
+
+struct mlx5e_ipsec_tx_metadata {
+	__be16 mss_inv;         /* 1/MSS in 16bit fixed point, only for LSO */
+	__be16 seq;             /* LSBs of the first TCP seq, only for LSO */
+	u8     esp_next_proto;  /* Next protocol of ESP */
+} __packed;
+
+struct mlx5e_ipsec_metadata {
+	unsigned char syndrome;
+	union {
+		unsigned char raw[5];
+		/* from FPGA to host, on successful decrypt */
+		struct mlx5e_ipsec_rx_metadata rx;
+		/* from host to FPGA */
+		struct mlx5e_ipsec_tx_metadata tx;
+	} __packed content;
+	/* packet type ID field	*/
+	__be16 ethertype;
+} __packed;
+
+#define MAX_LSO_MSS 2048
+
+/* Pre-calculated (Q0.16) fixed-point inverse 1/x function */
+static __be16 mlx5e_ipsec_inverse_table[MAX_LSO_MSS];
+
+static inline __be16 mlx5e_ipsec_mss_inv(struct sk_buff *skb)
+{
+	return mlx5e_ipsec_inverse_table[skb_shinfo(skb)->gso_size];
+}
+
+static struct mlx5e_ipsec_metadata *mlx5e_ipsec_add_metadata(struct sk_buff *skb)
+{
+	struct mlx5e_ipsec_metadata *mdata;
+	struct ethhdr *eth;
+
+	if (unlikely(skb_cow_head(skb, sizeof(*mdata))))
+		return ERR_PTR(-ENOMEM);
+
+	eth = (struct ethhdr *)skb_push(skb, sizeof(*mdata));
+	skb->mac_header -= sizeof(*mdata);
+	mdata = (struct mlx5e_ipsec_metadata *)(eth + 1);
+
+	memmove(skb->data, skb->data + sizeof(*mdata),
+		2 * ETH_ALEN);
+
+	eth->h_proto = cpu_to_be16(MLX5E_METADATA_ETHER_TYPE);
+
+	memset(mdata->content.raw, 0, sizeof(mdata->content.raw));
+	return mdata;
+}
+
+static int mlx5e_ipsec_remove_trailer(struct sk_buff *skb, struct xfrm_state *x)
+{
+	unsigned int alen = crypto_aead_authsize(x->data);
+	struct ipv6hdr *ipv6hdr = ipv6_hdr(skb);
+	struct iphdr *ipv4hdr = ip_hdr(skb);
+	unsigned int trailer_len;
+	u8 plen;
+	int ret;
+
+	ret = skb_copy_bits(skb, skb->len - alen - 2, &plen, 1);
+	if (unlikely(ret))
+		return ret;
+
+	trailer_len = alen + plen + 2;
+
+	pskb_trim(skb, skb->len - trailer_len);
+	if (skb->protocol == htons(ETH_P_IP)) {
+		ipv4hdr->tot_len = htons(ntohs(ipv4hdr->tot_len) - trailer_len);
+		ip_send_check(ipv4hdr);
+	} else {
+		ipv6hdr->payload_len = htons(ntohs(ipv6hdr->payload_len) -
+					     trailer_len);
+	}
+	return 0;
+}
+
+static void mlx5e_ipsec_set_swp(struct sk_buff *skb,
+				struct mlx5_wqe_eth_seg *eseg, u8 mode,
+				struct xfrm_offload *xo)
+{
+	u8 proto;
+
+	/* Tunnel Mode:
+	 * SWP:      OutL3       InL3  InL4
+	 * Pkt: MAC  IP     ESP  IP    L4
+	 *
+	 * Transport Mode:
+	 * SWP:      OutL3       InL4
+	 *           InL3
+	 * Pkt: MAC  IP     ESP  L4
+	 *
+	 * Offsets are in 2-byte words, counting from start of frame
+	 */
+	eseg->swp_outer_l3_offset = skb_network_offset(skb) / 2;
+	if (skb->protocol == htons(ETH_P_IPV6))
+		eseg->swp_flags |= MLX5_ETH_WQE_SWP_OUTER_L3_IPV6;
+
+	if (mode == XFRM_MODE_TUNNEL) {
+		eseg->swp_inner_l3_offset = skb_inner_network_offset(skb) / 2;
+		if (xo->proto == IPPROTO_IPV6) {
+			eseg->swp_flags |= MLX5_ETH_WQE_SWP_INNER_L3_IPV6;
+			proto = inner_ipv6_hdr(skb)->nexthdr;
+		} else {
+			proto = inner_ip_hdr(skb)->protocol;
+		}
+	} else {
+		eseg->swp_inner_l3_offset = skb_network_offset(skb) / 2;
+		if (skb->protocol == htons(ETH_P_IPV6))
+			eseg->swp_flags |= MLX5_ETH_WQE_SWP_INNER_L3_IPV6;
+		proto = xo->proto;
+	}
+	switch (proto) {
+	case IPPROTO_UDP:
+		eseg->swp_flags |= MLX5_ETH_WQE_SWP_INNER_L4_UDP;
+		/* Fall through */
+	case IPPROTO_TCP:
+		eseg->swp_inner_l4_offset = skb_inner_transport_offset(skb) / 2;
+		break;
+	}
+}
+
+void mlx5e_ipsec_set_iv_esn(struct sk_buff *skb, struct xfrm_state *x,
+			    struct xfrm_offload *xo)
+{
+	struct xfrm_replay_state_esn *replay_esn = x->replay_esn;
+	__u32 oseq = replay_esn->oseq;
+	int iv_offset;
+	__be64 seqno;
+	u32 seq_hi;
+
+	if (unlikely(skb_is_gso(skb) && oseq < MLX5E_IPSEC_ESN_SCOPE_MID &&
+		     MLX5E_IPSEC_ESN_SCOPE_MID < (oseq - skb_shinfo(skb)->gso_segs))) {
+		seq_hi = xo->seq.hi - 1;
+	} else {
+		seq_hi = xo->seq.hi;
+	}
+
+	/* Place the SN in the IV field */
+	seqno = cpu_to_be64(xo->seq.low + ((u64)seq_hi << 32));
+	iv_offset = skb_transport_offset(skb) + sizeof(struct ip_esp_hdr);
+	skb_store_bits(skb, iv_offset, &seqno, 8);
+}
+
+void mlx5e_ipsec_set_iv(struct sk_buff *skb, struct xfrm_state *x,
+			struct xfrm_offload *xo)
+{
+	int iv_offset;
+	__be64 seqno;
+
+	/* Place the SN in the IV field */
+	seqno = cpu_to_be64(xo->seq.low + ((u64)xo->seq.hi << 32));
+	iv_offset = skb_transport_offset(skb) + sizeof(struct ip_esp_hdr);
+	skb_store_bits(skb, iv_offset, &seqno, 8);
+}
+
+static void mlx5e_ipsec_set_metadata(struct sk_buff *skb,
+				     struct mlx5e_ipsec_metadata *mdata,
+				     struct xfrm_offload *xo)
+{
+	struct ip_esp_hdr *esph;
+	struct tcphdr *tcph;
+
+	if (skb_is_gso(skb)) {
+		/* Add LSO metadata indication */
+		esph = ip_esp_hdr(skb);
+		tcph = inner_tcp_hdr(skb);
+		netdev_dbg(skb->dev, "   Offloading GSO packet outer L3 %u; L4 %u; Inner L3 %u; L4 %u\n",
+			   skb->network_header,
+			   skb->transport_header,
+			   skb->inner_network_header,
+			   skb->inner_transport_header);
+		netdev_dbg(skb->dev, "   Offloading GSO packet of len %u; mss %u; TCP sp %u dp %u seq 0x%x ESP seq 0x%x\n",
+			   skb->len, skb_shinfo(skb)->gso_size,
+			   ntohs(tcph->source), ntohs(tcph->dest),
+			   ntohl(tcph->seq), ntohl(esph->seq_no));
+		mdata->syndrome = MLX5E_IPSEC_TX_SYNDROME_OFFLOAD_WITH_LSO_TCP;
+		mdata->content.tx.mss_inv = mlx5e_ipsec_mss_inv(skb);
+		mdata->content.tx.seq = htons(ntohl(tcph->seq) & 0xFFFF);
+	} else {
+		mdata->syndrome = MLX5E_IPSEC_TX_SYNDROME_OFFLOAD;
+	}
+	mdata->content.tx.esp_next_proto = xo->proto;
+
+	netdev_dbg(skb->dev, "   TX metadata syndrome %u proto %u mss_inv %04x seq %04x\n",
+		   mdata->syndrome, mdata->content.tx.esp_next_proto,
+		   ntohs(mdata->content.tx.mss_inv),
+		   ntohs(mdata->content.tx.seq));
+}
+
+struct sk_buff *mlx5e_ipsec_handle_tx_skb(struct net_device *netdev,
+					  struct mlx5e_tx_wqe *wqe,
+					  struct sk_buff *skb)
+{
+	struct mlx5e_priv *priv = netdev_priv(netdev);
+	struct xfrm_offload *xo = xfrm_offload(skb);
+	struct mlx5e_ipsec_metadata *mdata;
+	struct mlx5e_ipsec_sa_entry *sa_entry;
+	struct xfrm_state *x;
+
+	if (!xo)
+		return skb;
+
+	if (unlikely(skb->sp->len != 1)) {
+		atomic64_inc(&priv->ipsec->sw_stats.ipsec_tx_drop_bundle);
+		goto drop;
+	}
+
+	x = xfrm_input_state(skb);
+	if (unlikely(!x)) {
+		atomic64_inc(&priv->ipsec->sw_stats.ipsec_tx_drop_no_state);
+		goto drop;
+	}
+
+	if (unlikely(!x->xso.offload_handle ||
+		     (skb->protocol != htons(ETH_P_IP) &&
+		      skb->protocol != htons(ETH_P_IPV6)))) {
+		atomic64_inc(&priv->ipsec->sw_stats.ipsec_tx_drop_not_ip);
+		goto drop;
+	}
+
+	if (!skb_is_gso(skb))
+		if (unlikely(mlx5e_ipsec_remove_trailer(skb, x))) {
+			atomic64_inc(&priv->ipsec->sw_stats.ipsec_tx_drop_trailer);
+			goto drop;
+		}
+	mdata = mlx5e_ipsec_add_metadata(skb);
+	if (IS_ERR(mdata)) {
+		atomic64_inc(&priv->ipsec->sw_stats.ipsec_tx_drop_metadata);
+		goto drop;
+	}
+	mlx5e_ipsec_set_swp(skb, &wqe->eth, x->props.mode, xo);
+	sa_entry = (struct mlx5e_ipsec_sa_entry *)x->xso.offload_handle;
+	sa_entry->set_iv_op(skb, x, xo);
+	mlx5e_ipsec_set_metadata(skb, mdata, xo);
+
+	return skb;
+
+drop:
+	kfree_skb(skb);
+	return NULL;
+}
+
+static inline struct xfrm_state *
+mlx5e_ipsec_build_sp(struct net_device *netdev, struct sk_buff *skb,
+		     struct mlx5e_ipsec_metadata *mdata)
+{
+	struct mlx5e_priv *priv = netdev_priv(netdev);
+	struct xfrm_offload *xo;
+	struct xfrm_state *xs;
+	u32 sa_handle;
+
+	skb->sp = secpath_dup(skb->sp);
+	if (unlikely(!skb->sp)) {
+		atomic64_inc(&priv->ipsec->sw_stats.ipsec_rx_drop_sp_alloc);
+		return NULL;
+	}
+
+	sa_handle = be32_to_cpu(mdata->content.rx.sa_handle);
+	xs = mlx5e_ipsec_sadb_rx_lookup(priv->ipsec, sa_handle);
+	if (unlikely(!xs)) {
+		atomic64_inc(&priv->ipsec->sw_stats.ipsec_rx_drop_sadb_miss);
+		return NULL;
+	}
+
+	skb->sp->xvec[skb->sp->len++] = xs;
+	skb->sp->olen++;
+
+	xo = xfrm_offload(skb);
+	xo->flags = CRYPTO_DONE;
+	switch (mdata->syndrome) {
+	case MLX5E_IPSEC_RX_SYNDROME_DECRYPTED:
+		xo->status = CRYPTO_SUCCESS;
+		if (likely(priv->ipsec->no_trailer)) {
+			xo->flags |= XFRM_ESP_NO_TRAILER;
+			xo->proto = mdata->content.rx.nexthdr;
+		}
+		break;
+	case MLX5E_IPSEC_RX_SYNDROME_AUTH_FAILED:
+		xo->status = CRYPTO_TUNNEL_ESP_AUTH_FAILED;
+		break;
+	case MLX5E_IPSEC_RX_SYNDROME_BAD_PROTO:
+		xo->status = CRYPTO_INVALID_PROTOCOL;
+		break;
+	default:
+		atomic64_inc(&priv->ipsec->sw_stats.ipsec_rx_drop_syndrome);
+		return NULL;
+	}
+	return xs;
+}
+
+struct sk_buff *mlx5e_ipsec_handle_rx_skb(struct net_device *netdev,
+					  struct sk_buff *skb, u32 *cqe_bcnt)
+{
+	struct mlx5e_ipsec_metadata *mdata;
+	struct xfrm_state *xs;
+
+	if (!is_metadata_hdr_valid(skb))
+		return skb;
+
+	/* Use the metadata */
+	mdata = (struct mlx5e_ipsec_metadata *)(skb->data + ETH_HLEN);
+	xs = mlx5e_ipsec_build_sp(netdev, skb, mdata);
+	if (unlikely(!xs)) {
+		kfree_skb(skb);
+		return NULL;
+	}
+
+	remove_metadata_hdr(skb);
+	*cqe_bcnt -= MLX5E_METADATA_ETHER_LEN;
+
+	return skb;
+}
+
+bool mlx5e_ipsec_feature_check(struct sk_buff *skb, struct net_device *netdev,
+			       netdev_features_t features)
+{
+	struct xfrm_state *x;
+
+	if (skb->sp && skb->sp->len) {
+		x = skb->sp->xvec[0];
+		if (x && x->xso.offload_handle)
+			return true;
+	}
+	return false;
+}
+
+void mlx5e_ipsec_build_inverse_table(void)
+{
+	u16 mss_inv;
+	u32 mss;
+
+	/* Calculate 1/x inverse table for use in GSO data path.
+	 * Using this table, we provide the IPSec accelerator with the value of
+	 * 1/gso_size so that it can infer the position of each segment inside
+	 * the GSO, and increment the ESP sequence number, and generate the IV.
+	 * The HW needs this value in Q0.16 fixed-point number format
+	 */
+	mlx5e_ipsec_inverse_table[1] = htons(0xFFFF);
+	for (mss = 2; mss < MAX_LSO_MSS; mss++) {
+		mss_inv = div_u64(1ULL << 32, mss) >> 16;
+		mlx5e_ipsec_inverse_table[mss] = htons(mss_inv);
+	}
+}
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_accel/ipsec_rxtx.h b/drivers/net/ethernet/mellanox/mlx5/core/en_accel/ipsec_rxtx.h
new file mode 100644
index 000000000..ca47c0540
--- /dev/null
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_accel/ipsec_rxtx.h
@@ -0,0 +1,60 @@
+/*
+ * Copyright (c) 2017 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 __MLX5E_IPSEC_RXTX_H__
+#define __MLX5E_IPSEC_RXTX_H__
+
+#ifdef CONFIG_MLX5_EN_IPSEC
+
+#include <linux/skbuff.h>
+#include <net/xfrm.h>
+#include "en.h"
+
+struct sk_buff *mlx5e_ipsec_handle_rx_skb(struct net_device *netdev,
+					  struct sk_buff *skb, u32 *cqe_bcnt);
+void mlx5e_ipsec_handle_rx_cqe(struct mlx5e_rq *rq, struct mlx5_cqe64 *cqe);
+
+void mlx5e_ipsec_inverse_table_init(void);
+bool mlx5e_ipsec_feature_check(struct sk_buff *skb, struct net_device *netdev,
+			       netdev_features_t features);
+void mlx5e_ipsec_set_iv_esn(struct sk_buff *skb, struct xfrm_state *x,
+			    struct xfrm_offload *xo);
+void mlx5e_ipsec_set_iv(struct sk_buff *skb, struct xfrm_state *x,
+			struct xfrm_offload *xo);
+struct sk_buff *mlx5e_ipsec_handle_tx_skb(struct net_device *netdev,
+					  struct mlx5e_tx_wqe *wqe,
+					  struct sk_buff *skb);
+
+#endif /* CONFIG_MLX5_EN_IPSEC */
+
+#endif /* __MLX5E_IPSEC_RXTX_H__ */
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_accel/ipsec_stats.c b/drivers/net/ethernet/mellanox/mlx5/core/en_accel/ipsec_stats.c
new file mode 100644
index 000000000..6fea59223
--- /dev/null
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_accel/ipsec_stats.c
@@ -0,0 +1,133 @@
+/*
+ * Copyright (c) 2017 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/ethtool.h>
+#include <net/sock.h>
+
+#include "en.h"
+#include "accel/ipsec.h"
+#include "fpga/sdk.h"
+#include "en_accel/ipsec.h"
+
+static const struct counter_desc mlx5e_ipsec_hw_stats_desc[] = {
+	{ MLX5E_DECLARE_STAT(struct mlx5e_ipsec_stats, ipsec_dec_in_packets) },
+	{ MLX5E_DECLARE_STAT(struct mlx5e_ipsec_stats, ipsec_dec_out_packets) },
+	{ MLX5E_DECLARE_STAT(struct mlx5e_ipsec_stats, ipsec_dec_bypass_packets) },
+	{ MLX5E_DECLARE_STAT(struct mlx5e_ipsec_stats, ipsec_enc_in_packets) },
+	{ MLX5E_DECLARE_STAT(struct mlx5e_ipsec_stats, ipsec_enc_out_packets) },
+	{ MLX5E_DECLARE_STAT(struct mlx5e_ipsec_stats, ipsec_enc_bypass_packets) },
+	{ MLX5E_DECLARE_STAT(struct mlx5e_ipsec_stats, ipsec_dec_drop_packets) },
+	{ MLX5E_DECLARE_STAT(struct mlx5e_ipsec_stats, ipsec_dec_auth_fail_packets) },
+	{ MLX5E_DECLARE_STAT(struct mlx5e_ipsec_stats, ipsec_enc_drop_packets) },
+	{ MLX5E_DECLARE_STAT(struct mlx5e_ipsec_stats, ipsec_add_sa_success) },
+	{ MLX5E_DECLARE_STAT(struct mlx5e_ipsec_stats, ipsec_add_sa_fail) },
+	{ MLX5E_DECLARE_STAT(struct mlx5e_ipsec_stats, ipsec_del_sa_success) },
+	{ MLX5E_DECLARE_STAT(struct mlx5e_ipsec_stats, ipsec_del_sa_fail) },
+	{ MLX5E_DECLARE_STAT(struct mlx5e_ipsec_stats, ipsec_cmd_drop) },
+};
+
+static const struct counter_desc mlx5e_ipsec_sw_stats_desc[] = {
+	{ MLX5E_DECLARE_STAT(struct mlx5e_ipsec_sw_stats, ipsec_rx_drop_sp_alloc) },
+	{ MLX5E_DECLARE_STAT(struct mlx5e_ipsec_sw_stats, ipsec_rx_drop_sadb_miss) },
+	{ MLX5E_DECLARE_STAT(struct mlx5e_ipsec_sw_stats, ipsec_rx_drop_syndrome) },
+	{ MLX5E_DECLARE_STAT(struct mlx5e_ipsec_sw_stats, ipsec_tx_drop_bundle) },
+	{ MLX5E_DECLARE_STAT(struct mlx5e_ipsec_sw_stats, ipsec_tx_drop_no_state) },
+	{ MLX5E_DECLARE_STAT(struct mlx5e_ipsec_sw_stats, ipsec_tx_drop_not_ip) },
+	{ MLX5E_DECLARE_STAT(struct mlx5e_ipsec_sw_stats, ipsec_tx_drop_trailer) },
+	{ MLX5E_DECLARE_STAT(struct mlx5e_ipsec_sw_stats, ipsec_tx_drop_metadata) },
+};
+
+#define MLX5E_READ_CTR_ATOMIC64(ptr, dsc, i) \
+	atomic64_read((atomic64_t *)((char *)(ptr) + (dsc)[i].offset))
+
+#define NUM_IPSEC_HW_COUNTERS ARRAY_SIZE(mlx5e_ipsec_hw_stats_desc)
+#define NUM_IPSEC_SW_COUNTERS ARRAY_SIZE(mlx5e_ipsec_sw_stats_desc)
+
+#define NUM_IPSEC_COUNTERS (NUM_IPSEC_HW_COUNTERS + NUM_IPSEC_SW_COUNTERS)
+
+int mlx5e_ipsec_get_count(struct mlx5e_priv *priv)
+{
+	if (!priv->ipsec)
+		return 0;
+
+	return NUM_IPSEC_COUNTERS;
+}
+
+int mlx5e_ipsec_get_strings(struct mlx5e_priv *priv, uint8_t *data)
+{
+	unsigned int i, idx = 0;
+
+	if (!priv->ipsec)
+		return 0;
+
+	for (i = 0; i < NUM_IPSEC_HW_COUNTERS; i++)
+		strcpy(data + (idx++) * ETH_GSTRING_LEN,
+		       mlx5e_ipsec_hw_stats_desc[i].format);
+
+	for (i = 0; i < NUM_IPSEC_SW_COUNTERS; i++)
+		strcpy(data + (idx++) * ETH_GSTRING_LEN,
+		       mlx5e_ipsec_sw_stats_desc[i].format);
+
+	return NUM_IPSEC_COUNTERS;
+}
+
+void mlx5e_ipsec_update_stats(struct mlx5e_priv *priv)
+{
+	int ret;
+
+	if (!priv->ipsec)
+		return;
+
+	ret = mlx5_accel_ipsec_counters_read(priv->mdev, (u64 *)&priv->ipsec->stats,
+					     NUM_IPSEC_HW_COUNTERS);
+	if (ret)
+		memset(&priv->ipsec->stats, 0, sizeof(priv->ipsec->stats));
+}
+
+int mlx5e_ipsec_get_stats(struct mlx5e_priv *priv, u64 *data)
+{
+	int i, idx = 0;
+
+	if (!priv->ipsec)
+		return 0;
+
+	for (i = 0; i < NUM_IPSEC_HW_COUNTERS; i++)
+		data[idx++] = MLX5E_READ_CTR64_CPU(&priv->ipsec->stats,
+						   mlx5e_ipsec_hw_stats_desc, i);
+
+	for (i = 0; i < NUM_IPSEC_SW_COUNTERS; i++)
+		data[idx++] = MLX5E_READ_CTR_ATOMIC64(&priv->ipsec->sw_stats,
+						      mlx5e_ipsec_sw_stats_desc, i);
+
+	return NUM_IPSEC_COUNTERS;
+}
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_accel/tls.c b/drivers/net/ethernet/mellanox/mlx5/core/en_accel/tls.c
new file mode 100644
index 000000000..e88340e19
--- /dev/null
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_accel/tls.c
@@ -0,0 +1,231 @@
+/*
+ * 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 <linux/netdevice.h>
+#include <net/ipv6.h>
+#include "en_accel/tls.h"
+#include "accel/tls.h"
+
+static void mlx5e_tls_set_ipv4_flow(void *flow, struct sock *sk)
+{
+	struct inet_sock *inet = inet_sk(sk);
+
+	MLX5_SET(tls_flow, flow, ipv6, 0);
+	memcpy(MLX5_ADDR_OF(tls_flow, flow, dst_ipv4_dst_ipv6.ipv4_layout.ipv4),
+	       &inet->inet_daddr, MLX5_FLD_SZ_BYTES(ipv4_layout, ipv4));
+	memcpy(MLX5_ADDR_OF(tls_flow, flow, src_ipv4_src_ipv6.ipv4_layout.ipv4),
+	       &inet->inet_rcv_saddr, MLX5_FLD_SZ_BYTES(ipv4_layout, ipv4));
+}
+
+#if IS_ENABLED(CONFIG_IPV6)
+static void mlx5e_tls_set_ipv6_flow(void *flow, struct sock *sk)
+{
+	struct ipv6_pinfo *np = inet6_sk(sk);
+
+	MLX5_SET(tls_flow, flow, ipv6, 1);
+	memcpy(MLX5_ADDR_OF(tls_flow, flow, dst_ipv4_dst_ipv6.ipv6_layout.ipv6),
+	       &sk->sk_v6_daddr, MLX5_FLD_SZ_BYTES(ipv6_layout, ipv6));
+	memcpy(MLX5_ADDR_OF(tls_flow, flow, src_ipv4_src_ipv6.ipv6_layout.ipv6),
+	       &np->saddr, MLX5_FLD_SZ_BYTES(ipv6_layout, ipv6));
+}
+#endif
+
+static void mlx5e_tls_set_flow_tcp_ports(void *flow, struct sock *sk)
+{
+	struct inet_sock *inet = inet_sk(sk);
+
+	memcpy(MLX5_ADDR_OF(tls_flow, flow, src_port), &inet->inet_sport,
+	       MLX5_FLD_SZ_BYTES(tls_flow, src_port));
+	memcpy(MLX5_ADDR_OF(tls_flow, flow, dst_port), &inet->inet_dport,
+	       MLX5_FLD_SZ_BYTES(tls_flow, dst_port));
+}
+
+static int mlx5e_tls_set_flow(void *flow, struct sock *sk, u32 caps)
+{
+	switch (sk->sk_family) {
+	case AF_INET:
+		mlx5e_tls_set_ipv4_flow(flow, sk);
+		break;
+#if IS_ENABLED(CONFIG_IPV6)
+	case AF_INET6:
+		if (!sk->sk_ipv6only &&
+		    ipv6_addr_type(&sk->sk_v6_daddr) == IPV6_ADDR_MAPPED) {
+			mlx5e_tls_set_ipv4_flow(flow, sk);
+			break;
+		}
+		if (!(caps & MLX5_ACCEL_TLS_IPV6))
+			goto error_out;
+
+		mlx5e_tls_set_ipv6_flow(flow, sk);
+		break;
+#endif
+	default:
+		goto error_out;
+	}
+
+	mlx5e_tls_set_flow_tcp_ports(flow, sk);
+	return 0;
+error_out:
+	return -EINVAL;
+}
+
+static int mlx5e_tls_add(struct net_device *netdev, struct sock *sk,
+			 enum tls_offload_ctx_dir direction,
+			 struct tls_crypto_info *crypto_info,
+			 u32 start_offload_tcp_sn)
+{
+	struct mlx5e_priv *priv = netdev_priv(netdev);
+	struct tls_context *tls_ctx = tls_get_ctx(sk);
+	struct mlx5_core_dev *mdev = priv->mdev;
+	u32 caps = mlx5_accel_tls_device_caps(mdev);
+	int ret = -ENOMEM;
+	void *flow;
+	u32 swid;
+
+	flow = kzalloc(MLX5_ST_SZ_BYTES(tls_flow), GFP_KERNEL);
+	if (!flow)
+		return ret;
+
+	ret = mlx5e_tls_set_flow(flow, sk, caps);
+	if (ret)
+		goto free_flow;
+
+	ret = mlx5_accel_tls_add_flow(mdev, flow, crypto_info,
+				      start_offload_tcp_sn, &swid,
+				      direction == TLS_OFFLOAD_CTX_DIR_TX);
+	if (ret < 0)
+		goto free_flow;
+
+	if (direction == TLS_OFFLOAD_CTX_DIR_TX) {
+		struct mlx5e_tls_offload_context_tx *tx_ctx =
+		    mlx5e_get_tls_tx_context(tls_ctx);
+
+		tx_ctx->swid = htonl(swid);
+		tx_ctx->expected_seq = start_offload_tcp_sn;
+	} else {
+		struct mlx5e_tls_offload_context_rx *rx_ctx =
+		    mlx5e_get_tls_rx_context(tls_ctx);
+
+		rx_ctx->handle = htonl(swid);
+	}
+
+	return 0;
+free_flow:
+	kfree(flow);
+	return ret;
+}
+
+static void mlx5e_tls_del(struct net_device *netdev,
+			  struct tls_context *tls_ctx,
+			  enum tls_offload_ctx_dir direction)
+{
+	struct mlx5e_priv *priv = netdev_priv(netdev);
+	unsigned int handle;
+
+	handle = ntohl((direction == TLS_OFFLOAD_CTX_DIR_TX) ?
+		       mlx5e_get_tls_tx_context(tls_ctx)->swid :
+		       mlx5e_get_tls_rx_context(tls_ctx)->handle);
+
+	mlx5_accel_tls_del_flow(priv->mdev, handle,
+				direction == TLS_OFFLOAD_CTX_DIR_TX);
+}
+
+static void mlx5e_tls_resync_rx(struct net_device *netdev, struct sock *sk,
+				u32 seq, u64 rcd_sn)
+{
+	struct tls_context *tls_ctx = tls_get_ctx(sk);
+	struct mlx5e_priv *priv = netdev_priv(netdev);
+	struct mlx5e_tls_offload_context_rx *rx_ctx;
+
+	rx_ctx = mlx5e_get_tls_rx_context(tls_ctx);
+
+	netdev_info(netdev, "resyncing seq %d rcd %lld\n", seq,
+		    be64_to_cpu(rcd_sn));
+	mlx5_accel_tls_resync_rx(priv->mdev, rx_ctx->handle, seq, rcd_sn);
+	atomic64_inc(&priv->tls->sw_stats.rx_tls_resync_reply);
+}
+
+static const struct tlsdev_ops mlx5e_tls_ops = {
+	.tls_dev_add = mlx5e_tls_add,
+	.tls_dev_del = mlx5e_tls_del,
+	.tls_dev_resync_rx = mlx5e_tls_resync_rx,
+};
+
+void mlx5e_tls_build_netdev(struct mlx5e_priv *priv)
+{
+	struct net_device *netdev = priv->netdev;
+	u32 caps;
+
+	if (!mlx5_accel_is_tls_device(priv->mdev))
+		return;
+
+	caps = mlx5_accel_tls_device_caps(priv->mdev);
+	if (caps & MLX5_ACCEL_TLS_TX) {
+		netdev->features          |= NETIF_F_HW_TLS_TX;
+		netdev->hw_features       |= NETIF_F_HW_TLS_TX;
+	}
+
+	if (caps & MLX5_ACCEL_TLS_RX) {
+		netdev->features          |= NETIF_F_HW_TLS_RX;
+		netdev->hw_features       |= NETIF_F_HW_TLS_RX;
+	}
+
+	if (!(caps & MLX5_ACCEL_TLS_LRO)) {
+		netdev->features          &= ~NETIF_F_LRO;
+		netdev->hw_features       &= ~NETIF_F_LRO;
+	}
+
+	netdev->tlsdev_ops = &mlx5e_tls_ops;
+}
+
+int mlx5e_tls_init(struct mlx5e_priv *priv)
+{
+	struct mlx5e_tls *tls = kzalloc(sizeof(*tls), GFP_KERNEL);
+
+	if (!tls)
+		return -ENOMEM;
+
+	priv->tls = tls;
+	return 0;
+}
+
+void mlx5e_tls_cleanup(struct mlx5e_priv *priv)
+{
+	struct mlx5e_tls *tls = priv->tls;
+
+	if (!tls)
+		return;
+
+	kfree(tls);
+	priv->tls = NULL;
+}
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_accel/tls.h b/drivers/net/ethernet/mellanox/mlx5/core/en_accel/tls.h
new file mode 100644
index 000000000..3f5d72163
--- /dev/null
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_accel/tls.h
@@ -0,0 +1,106 @@
+/*
+ * 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.
+ *
+ */
+#ifndef __MLX5E_TLS_H__
+#define __MLX5E_TLS_H__
+
+#ifdef CONFIG_MLX5_EN_TLS
+
+#include <net/tls.h>
+#include "en.h"
+
+struct mlx5e_tls_sw_stats {
+	atomic64_t tx_tls_drop_metadata;
+	atomic64_t tx_tls_drop_resync_alloc;
+	atomic64_t tx_tls_drop_no_sync_data;
+	atomic64_t tx_tls_drop_bypass_required;
+	atomic64_t rx_tls_drop_resync_request;
+	atomic64_t rx_tls_resync_request;
+	atomic64_t rx_tls_resync_reply;
+	atomic64_t rx_tls_auth_fail;
+};
+
+struct mlx5e_tls {
+	struct mlx5e_tls_sw_stats sw_stats;
+};
+
+struct mlx5e_tls_offload_context_tx {
+	struct tls_offload_context_tx base;
+	u32 expected_seq;
+	__be32 swid;
+};
+
+static inline struct mlx5e_tls_offload_context_tx *
+mlx5e_get_tls_tx_context(struct tls_context *tls_ctx)
+{
+	BUILD_BUG_ON(sizeof(struct mlx5e_tls_offload_context_tx) >
+		     TLS_OFFLOAD_CONTEXT_SIZE_TX);
+	return container_of(tls_offload_ctx_tx(tls_ctx),
+			    struct mlx5e_tls_offload_context_tx,
+			    base);
+}
+
+struct mlx5e_tls_offload_context_rx {
+	struct tls_offload_context_rx base;
+	__be32 handle;
+};
+
+static inline struct mlx5e_tls_offload_context_rx *
+mlx5e_get_tls_rx_context(struct tls_context *tls_ctx)
+{
+	BUILD_BUG_ON(sizeof(struct mlx5e_tls_offload_context_rx) >
+		     TLS_OFFLOAD_CONTEXT_SIZE_RX);
+	return container_of(tls_offload_ctx_rx(tls_ctx),
+			    struct mlx5e_tls_offload_context_rx,
+			    base);
+}
+
+void mlx5e_tls_build_netdev(struct mlx5e_priv *priv);
+int mlx5e_tls_init(struct mlx5e_priv *priv);
+void mlx5e_tls_cleanup(struct mlx5e_priv *priv);
+
+int mlx5e_tls_get_count(struct mlx5e_priv *priv);
+int mlx5e_tls_get_strings(struct mlx5e_priv *priv, uint8_t *data);
+int mlx5e_tls_get_stats(struct mlx5e_priv *priv, u64 *data);
+
+#else
+
+static inline void mlx5e_tls_build_netdev(struct mlx5e_priv *priv) { }
+static inline int mlx5e_tls_init(struct mlx5e_priv *priv) { return 0; }
+static inline void mlx5e_tls_cleanup(struct mlx5e_priv *priv) { }
+static inline int mlx5e_tls_get_count(struct mlx5e_priv *priv) { return 0; }
+static inline int mlx5e_tls_get_strings(struct mlx5e_priv *priv, uint8_t *data) { return 0; }
+static inline int mlx5e_tls_get_stats(struct mlx5e_priv *priv, u64 *data) { return 0; }
+
+#endif
+
+#endif /* __MLX5E_TLS_H__ */
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_accel/tls_rxtx.c b/drivers/net/ethernet/mellanox/mlx5/core/en_accel/tls_rxtx.c
new file mode 100644
index 000000000..be137d4a9
--- /dev/null
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_accel/tls_rxtx.c
@@ -0,0 +1,383 @@
+/*
+ * 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 "en_accel/tls.h"
+#include "en_accel/tls_rxtx.h"
+#include "accel/accel.h"
+
+#include <net/inet6_hashtables.h>
+#include <linux/ipv6.h>
+
+#define SYNDROM_DECRYPTED  0x30
+#define SYNDROM_RESYNC_REQUEST 0x31
+#define SYNDROM_AUTH_FAILED 0x32
+
+#define SYNDROME_OFFLOAD_REQUIRED 32
+#define SYNDROME_SYNC 33
+
+struct sync_info {
+	u64 rcd_sn;
+	s32 sync_len;
+	int nr_frags;
+	skb_frag_t frags[MAX_SKB_FRAGS];
+};
+
+struct recv_metadata_content {
+	u8 syndrome;
+	u8 reserved;
+	__be32 sync_seq;
+} __packed;
+
+struct send_metadata_content {
+	/* One byte of syndrome followed by 3 bytes of swid */
+	__be32 syndrome_swid;
+	__be16 first_seq;
+} __packed;
+
+struct mlx5e_tls_metadata {
+	union {
+		/* from fpga to host */
+		struct recv_metadata_content recv;
+		/* from host to fpga */
+		struct send_metadata_content send;
+		unsigned char raw[6];
+	} __packed content;
+	/* packet type ID field	*/
+	__be16 ethertype;
+} __packed;
+
+static int mlx5e_tls_add_metadata(struct sk_buff *skb, __be32 swid)
+{
+	struct mlx5e_tls_metadata *pet;
+	struct ethhdr *eth;
+
+	if (skb_cow_head(skb, sizeof(struct mlx5e_tls_metadata)))
+		return -ENOMEM;
+
+	eth = (struct ethhdr *)skb_push(skb, sizeof(struct mlx5e_tls_metadata));
+	skb->mac_header -= sizeof(struct mlx5e_tls_metadata);
+	pet = (struct mlx5e_tls_metadata *)(eth + 1);
+
+	memmove(skb->data, skb->data + sizeof(struct mlx5e_tls_metadata),
+		2 * ETH_ALEN);
+
+	eth->h_proto = cpu_to_be16(MLX5E_METADATA_ETHER_TYPE);
+	pet->content.send.syndrome_swid =
+		htonl(SYNDROME_OFFLOAD_REQUIRED << 24) | swid;
+
+	return 0;
+}
+
+static int mlx5e_tls_get_sync_data(struct mlx5e_tls_offload_context_tx *context,
+				   u32 tcp_seq, struct sync_info *info)
+{
+	int remaining, i = 0, ret = -EINVAL;
+	struct tls_record_info *record;
+	unsigned long flags;
+	s32 sync_size;
+
+	spin_lock_irqsave(&context->base.lock, flags);
+	record = tls_get_record(&context->base, tcp_seq, &info->rcd_sn);
+
+	if (unlikely(!record))
+		goto out;
+
+	sync_size = tcp_seq - tls_record_start_seq(record);
+	info->sync_len = sync_size;
+	if (unlikely(sync_size < 0)) {
+		if (tls_record_is_start_marker(record))
+			goto done;
+
+		goto out;
+	}
+
+	remaining = sync_size;
+	while (remaining > 0) {
+		info->frags[i] = record->frags[i];
+		__skb_frag_ref(&info->frags[i]);
+		remaining -= skb_frag_size(&info->frags[i]);
+
+		if (remaining < 0)
+			skb_frag_size_add(&info->frags[i], remaining);
+
+		i++;
+	}
+	info->nr_frags = i;
+done:
+	ret = 0;
+out:
+	spin_unlock_irqrestore(&context->base.lock, flags);
+	return ret;
+}
+
+static void mlx5e_tls_complete_sync_skb(struct sk_buff *skb,
+					struct sk_buff *nskb, u32 tcp_seq,
+					int headln, __be64 rcd_sn)
+{
+	struct mlx5e_tls_metadata *pet;
+	u8 syndrome = SYNDROME_SYNC;
+	struct iphdr *iph;
+	struct tcphdr *th;
+	int data_len, mss;
+
+	nskb->dev = skb->dev;
+	skb_reset_mac_header(nskb);
+	skb_set_network_header(nskb, skb_network_offset(skb));
+	skb_set_transport_header(nskb, skb_transport_offset(skb));
+	memcpy(nskb->data, skb->data, headln);
+	memcpy(nskb->data + headln, &rcd_sn, sizeof(rcd_sn));
+
+	iph = ip_hdr(nskb);
+	iph->tot_len = htons(nskb->len - skb_network_offset(nskb));
+	th = tcp_hdr(nskb);
+	data_len = nskb->len - headln;
+	tcp_seq -= data_len;
+	th->seq = htonl(tcp_seq);
+
+	mss = nskb->dev->mtu - (headln - skb_network_offset(nskb));
+	skb_shinfo(nskb)->gso_size = 0;
+	if (data_len > mss) {
+		skb_shinfo(nskb)->gso_size = mss;
+		skb_shinfo(nskb)->gso_segs = DIV_ROUND_UP(data_len, mss);
+	}
+	skb_shinfo(nskb)->gso_type = skb_shinfo(skb)->gso_type;
+
+	pet = (struct mlx5e_tls_metadata *)(nskb->data + sizeof(struct ethhdr));
+	memcpy(pet, &syndrome, sizeof(syndrome));
+	pet->content.send.first_seq = htons(tcp_seq);
+
+	/* MLX5 devices don't care about the checksum partial start, offset
+	 * and pseudo header
+	 */
+	nskb->ip_summed = CHECKSUM_PARTIAL;
+
+	nskb->xmit_more = 1;
+	nskb->queue_mapping = skb->queue_mapping;
+}
+
+static struct sk_buff *
+mlx5e_tls_handle_ooo(struct mlx5e_tls_offload_context_tx *context,
+		     struct mlx5e_txqsq *sq, struct sk_buff *skb,
+		     struct mlx5e_tx_wqe **wqe,
+		     u16 *pi,
+		     struct mlx5e_tls *tls)
+{
+	u32 tcp_seq = ntohl(tcp_hdr(skb)->seq);
+	struct sync_info info;
+	struct sk_buff *nskb;
+	int linear_len = 0;
+	int headln;
+	int i;
+
+	sq->stats->tls_ooo++;
+
+	if (mlx5e_tls_get_sync_data(context, tcp_seq, &info)) {
+		/* We might get here if a retransmission reaches the driver
+		 * after the relevant record is acked.
+		 * It should be safe to drop the packet in this case
+		 */
+		atomic64_inc(&tls->sw_stats.tx_tls_drop_no_sync_data);
+		goto err_out;
+	}
+
+	if (unlikely(info.sync_len < 0)) {
+		u32 payload;
+
+		headln = skb_transport_offset(skb) + tcp_hdrlen(skb);
+		payload = skb->len - headln;
+		if (likely(payload <= -info.sync_len))
+			/* SKB payload doesn't require offload
+			 */
+			return skb;
+
+		atomic64_inc(&tls->sw_stats.tx_tls_drop_bypass_required);
+		goto err_out;
+	}
+
+	if (unlikely(mlx5e_tls_add_metadata(skb, context->swid))) {
+		atomic64_inc(&tls->sw_stats.tx_tls_drop_metadata);
+		goto err_out;
+	}
+
+	headln = skb_transport_offset(skb) + tcp_hdrlen(skb);
+	linear_len += headln + sizeof(info.rcd_sn);
+	nskb = alloc_skb(linear_len, GFP_ATOMIC);
+	if (unlikely(!nskb)) {
+		atomic64_inc(&tls->sw_stats.tx_tls_drop_resync_alloc);
+		goto err_out;
+	}
+
+	context->expected_seq = tcp_seq + skb->len - headln;
+	skb_put(nskb, linear_len);
+	for (i = 0; i < info.nr_frags; i++)
+		skb_shinfo(nskb)->frags[i] = info.frags[i];
+
+	skb_shinfo(nskb)->nr_frags = info.nr_frags;
+	nskb->data_len = info.sync_len;
+	nskb->len += info.sync_len;
+	sq->stats->tls_resync_bytes += nskb->len;
+	mlx5e_tls_complete_sync_skb(skb, nskb, tcp_seq, headln,
+				    cpu_to_be64(info.rcd_sn));
+	mlx5e_sq_xmit(sq, nskb, *wqe, *pi);
+	mlx5e_sq_fetch_wqe(sq, wqe, pi);
+	return skb;
+
+err_out:
+	dev_kfree_skb_any(skb);
+	return NULL;
+}
+
+struct sk_buff *mlx5e_tls_handle_tx_skb(struct net_device *netdev,
+					struct mlx5e_txqsq *sq,
+					struct sk_buff *skb,
+					struct mlx5e_tx_wqe **wqe,
+					u16 *pi)
+{
+	struct mlx5e_priv *priv = netdev_priv(netdev);
+	struct mlx5e_tls_offload_context_tx *context;
+	struct tls_context *tls_ctx;
+	u32 expected_seq;
+	int datalen;
+	u32 skb_seq;
+
+	if (!skb->sk || !tls_is_sk_tx_device_offloaded(skb->sk))
+		goto out;
+
+	datalen = skb->len - (skb_transport_offset(skb) + tcp_hdrlen(skb));
+	if (!datalen)
+		goto out;
+
+	tls_ctx = tls_get_ctx(skb->sk);
+	if (unlikely(tls_ctx->netdev != netdev))
+		goto out;
+
+	skb_seq = ntohl(tcp_hdr(skb)->seq);
+	context = mlx5e_get_tls_tx_context(tls_ctx);
+	expected_seq = context->expected_seq;
+
+	if (unlikely(expected_seq != skb_seq)) {
+		skb = mlx5e_tls_handle_ooo(context, sq, skb, wqe, pi, priv->tls);
+		goto out;
+	}
+
+	if (unlikely(mlx5e_tls_add_metadata(skb, context->swid))) {
+		atomic64_inc(&priv->tls->sw_stats.tx_tls_drop_metadata);
+		dev_kfree_skb_any(skb);
+		skb = NULL;
+		goto out;
+	}
+
+	context->expected_seq = skb_seq + datalen;
+out:
+	return skb;
+}
+
+static int tls_update_resync_sn(struct net_device *netdev,
+				struct sk_buff *skb,
+				struct mlx5e_tls_metadata *mdata)
+{
+	struct sock *sk = NULL;
+	struct iphdr *iph;
+	struct tcphdr *th;
+	__be32 seq;
+
+	if (mdata->ethertype != htons(ETH_P_IP))
+		return -EINVAL;
+
+	iph = (struct iphdr *)(mdata + 1);
+
+	th = ((void *)iph) + iph->ihl * 4;
+
+	if (iph->version == 4) {
+		sk = inet_lookup_established(dev_net(netdev), &tcp_hashinfo,
+					     iph->saddr, th->source, iph->daddr,
+					     th->dest, netdev->ifindex);
+#if IS_ENABLED(CONFIG_IPV6)
+	} else {
+		struct ipv6hdr *ipv6h = (struct ipv6hdr *)iph;
+
+		sk = __inet6_lookup_established(dev_net(netdev), &tcp_hashinfo,
+						&ipv6h->saddr, th->source,
+						&ipv6h->daddr, ntohs(th->dest),
+						netdev->ifindex, 0);
+#endif
+	}
+	if (!sk || sk->sk_state == TCP_TIME_WAIT) {
+		struct mlx5e_priv *priv = netdev_priv(netdev);
+
+		atomic64_inc(&priv->tls->sw_stats.rx_tls_drop_resync_request);
+		goto out;
+	}
+
+	skb->sk = sk;
+	skb->destructor = sock_edemux;
+
+	memcpy(&seq, &mdata->content.recv.sync_seq, sizeof(seq));
+	tls_offload_rx_resync_request(sk, seq);
+out:
+	return 0;
+}
+
+void mlx5e_tls_handle_rx_skb(struct net_device *netdev, struct sk_buff *skb,
+			     u32 *cqe_bcnt)
+{
+	struct mlx5e_tls_metadata *mdata;
+	struct mlx5e_priv *priv;
+
+	if (!is_metadata_hdr_valid(skb))
+		return;
+
+	/* Use the metadata */
+	mdata = (struct mlx5e_tls_metadata *)(skb->data + ETH_HLEN);
+	switch (mdata->content.recv.syndrome) {
+	case SYNDROM_DECRYPTED:
+		skb->decrypted = 1;
+		break;
+	case SYNDROM_RESYNC_REQUEST:
+		tls_update_resync_sn(netdev, skb, mdata);
+		priv = netdev_priv(netdev);
+		atomic64_inc(&priv->tls->sw_stats.rx_tls_resync_request);
+		break;
+	case SYNDROM_AUTH_FAILED:
+		/* Authentication failure will be observed and verified by kTLS */
+		priv = netdev_priv(netdev);
+		atomic64_inc(&priv->tls->sw_stats.rx_tls_auth_fail);
+		break;
+	default:
+		/* Bypass the metadata header to others */
+		return;
+	}
+
+	remove_metadata_hdr(skb);
+	*cqe_bcnt -= MLX5E_METADATA_ETHER_LEN;
+}
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_accel/tls_rxtx.h b/drivers/net/ethernet/mellanox/mlx5/core/en_accel/tls_rxtx.h
new file mode 100644
index 000000000..311667ec7
--- /dev/null
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_accel/tls_rxtx.h
@@ -0,0 +1,53 @@
+/*
+ * 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.
+ *
+ */
+
+#ifndef __MLX5E_TLS_RXTX_H__
+#define __MLX5E_TLS_RXTX_H__
+
+#ifdef CONFIG_MLX5_EN_TLS
+
+#include <linux/skbuff.h>
+#include "en.h"
+
+struct sk_buff *mlx5e_tls_handle_tx_skb(struct net_device *netdev,
+					struct mlx5e_txqsq *sq,
+					struct sk_buff *skb,
+					struct mlx5e_tx_wqe **wqe,
+					u16 *pi);
+
+void mlx5e_tls_handle_rx_skb(struct net_device *netdev, struct sk_buff *skb,
+			     u32 *cqe_bcnt);
+
+#endif /* CONFIG_MLX5_EN_TLS */
+
+#endif /* __MLX5E_TLS_RXTX_H__ */
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_accel/tls_stats.c b/drivers/net/ethernet/mellanox/mlx5/core/en_accel/tls_stats.c
new file mode 100644
index 000000000..01468ec27
--- /dev/null
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_accel/tls_stats.c
@@ -0,0 +1,89 @@
+/*
+ * 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 <linux/ethtool.h>
+#include <net/sock.h>
+
+#include "en.h"
+#include "accel/tls.h"
+#include "fpga/sdk.h"
+#include "en_accel/tls.h"
+
+static const struct counter_desc mlx5e_tls_sw_stats_desc[] = {
+	{ MLX5E_DECLARE_STAT(struct mlx5e_tls_sw_stats, tx_tls_drop_metadata) },
+	{ MLX5E_DECLARE_STAT(struct mlx5e_tls_sw_stats, tx_tls_drop_resync_alloc) },
+	{ MLX5E_DECLARE_STAT(struct mlx5e_tls_sw_stats, tx_tls_drop_no_sync_data) },
+	{ MLX5E_DECLARE_STAT(struct mlx5e_tls_sw_stats, tx_tls_drop_bypass_required) },
+};
+
+#define MLX5E_READ_CTR_ATOMIC64(ptr, dsc, i) \
+	atomic64_read((atomic64_t *)((char *)(ptr) + (dsc)[i].offset))
+
+#define NUM_TLS_SW_COUNTERS ARRAY_SIZE(mlx5e_tls_sw_stats_desc)
+
+int mlx5e_tls_get_count(struct mlx5e_priv *priv)
+{
+	if (!priv->tls)
+		return 0;
+
+	return NUM_TLS_SW_COUNTERS;
+}
+
+int mlx5e_tls_get_strings(struct mlx5e_priv *priv, uint8_t *data)
+{
+	unsigned int i, idx = 0;
+
+	if (!priv->tls)
+		return 0;
+
+	for (i = 0; i < NUM_TLS_SW_COUNTERS; i++)
+		strcpy(data + (idx++) * ETH_GSTRING_LEN,
+		       mlx5e_tls_sw_stats_desc[i].format);
+
+	return NUM_TLS_SW_COUNTERS;
+}
+
+int mlx5e_tls_get_stats(struct mlx5e_priv *priv, u64 *data)
+{
+	int i, idx = 0;
+
+	if (!priv->tls)
+		return 0;
+
+	for (i = 0; i < NUM_TLS_SW_COUNTERS; i++)
+		data[idx++] =
+		    MLX5E_READ_CTR_ATOMIC64(&priv->tls->sw_stats,
+					    mlx5e_tls_sw_stats_desc, i);
+
+	return NUM_TLS_SW_COUNTERS;
+}
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_arfs.c b/drivers/net/ethernet/mellanox/mlx5/core/en_arfs.c
new file mode 100644
index 000000000..a4be04deb
--- /dev/null
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_arfs.c
@@ -0,0 +1,710 @@
+/*
+ * Copyright (c) 2016, 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/hash.h>
+#include <linux/mlx5/fs.h>
+#include <linux/ip.h>
+#include <linux/ipv6.h>
+#include "en.h"
+
+struct arfs_tuple {
+	__be16 etype;
+	u8     ip_proto;
+	union {
+		__be32 src_ipv4;
+		struct in6_addr src_ipv6;
+	};
+	union {
+		__be32 dst_ipv4;
+		struct in6_addr dst_ipv6;
+	};
+	__be16 src_port;
+	__be16 dst_port;
+};
+
+struct arfs_rule {
+	struct mlx5e_priv	*priv;
+	struct work_struct      arfs_work;
+	struct mlx5_flow_handle *rule;
+	struct hlist_node	hlist;
+	int			rxq;
+	/* Flow ID passed to ndo_rx_flow_steer */
+	int			flow_id;
+	/* Filter ID returned by ndo_rx_flow_steer */
+	int			filter_id;
+	struct arfs_tuple	tuple;
+};
+
+#define mlx5e_for_each_arfs_rule(hn, tmp, arfs_tables, i, j) \
+	for (i = 0; i < ARFS_NUM_TYPES; i++) \
+		mlx5e_for_each_hash_arfs_rule(hn, tmp, arfs_tables[i].rules_hash, j)
+
+#define mlx5e_for_each_hash_arfs_rule(hn, tmp, hash, j) \
+	for (j = 0; j < ARFS_HASH_SIZE; j++) \
+		hlist_for_each_entry_safe(hn, tmp, &hash[j], hlist)
+
+static enum mlx5e_traffic_types arfs_get_tt(enum arfs_type type)
+{
+	switch (type) {
+	case ARFS_IPV4_TCP:
+		return MLX5E_TT_IPV4_TCP;
+	case ARFS_IPV4_UDP:
+		return MLX5E_TT_IPV4_UDP;
+	case ARFS_IPV6_TCP:
+		return MLX5E_TT_IPV6_TCP;
+	case ARFS_IPV6_UDP:
+		return MLX5E_TT_IPV6_UDP;
+	default:
+		return -EINVAL;
+	}
+}
+
+static int arfs_disable(struct mlx5e_priv *priv)
+{
+	struct mlx5_flow_destination dest = {};
+	struct mlx5e_tir *tir = priv->indir_tir;
+	int err = 0;
+	int tt;
+	int i;
+
+	dest.type = MLX5_FLOW_DESTINATION_TYPE_TIR;
+	for (i = 0; i < ARFS_NUM_TYPES; i++) {
+		dest.tir_num = tir[i].tirn;
+		tt = arfs_get_tt(i);
+		/* Modify ttc rules destination to bypass the aRFS tables*/
+		err = mlx5_modify_rule_destination(priv->fs.ttc.rules[tt],
+						   &dest, NULL);
+		if (err) {
+			netdev_err(priv->netdev,
+				   "%s: modify ttc destination failed\n",
+				   __func__);
+			return err;
+		}
+	}
+	return 0;
+}
+
+static void arfs_del_rules(struct mlx5e_priv *priv);
+
+int mlx5e_arfs_disable(struct mlx5e_priv *priv)
+{
+	arfs_del_rules(priv);
+
+	return arfs_disable(priv);
+}
+
+int mlx5e_arfs_enable(struct mlx5e_priv *priv)
+{
+	struct mlx5_flow_destination dest = {};
+	int err = 0;
+	int tt;
+	int i;
+
+	dest.type = MLX5_FLOW_DESTINATION_TYPE_FLOW_TABLE;
+	for (i = 0; i < ARFS_NUM_TYPES; i++) {
+		dest.ft = priv->fs.arfs.arfs_tables[i].ft.t;
+		tt = arfs_get_tt(i);
+		/* Modify ttc rules destination to point on the aRFS FTs */
+		err = mlx5_modify_rule_destination(priv->fs.ttc.rules[tt],
+						   &dest, NULL);
+		if (err) {
+			netdev_err(priv->netdev,
+				   "%s: modify ttc destination failed err=%d\n",
+				   __func__, err);
+			arfs_disable(priv);
+			return err;
+		}
+	}
+	return 0;
+}
+
+static void arfs_destroy_table(struct arfs_table *arfs_t)
+{
+	mlx5_del_flow_rules(arfs_t->default_rule);
+	mlx5e_destroy_flow_table(&arfs_t->ft);
+}
+
+void mlx5e_arfs_destroy_tables(struct mlx5e_priv *priv)
+{
+	int i;
+
+	if (!(priv->netdev->hw_features & NETIF_F_NTUPLE))
+		return;
+
+	arfs_del_rules(priv);
+	destroy_workqueue(priv->fs.arfs.wq);
+	for (i = 0; i < ARFS_NUM_TYPES; i++) {
+		if (!IS_ERR_OR_NULL(priv->fs.arfs.arfs_tables[i].ft.t))
+			arfs_destroy_table(&priv->fs.arfs.arfs_tables[i]);
+	}
+}
+
+static int arfs_add_default_rule(struct mlx5e_priv *priv,
+				 enum arfs_type type)
+{
+	struct arfs_table *arfs_t = &priv->fs.arfs.arfs_tables[type];
+	struct mlx5e_tir *tir = priv->indir_tir;
+	struct mlx5_flow_destination dest = {};
+	MLX5_DECLARE_FLOW_ACT(flow_act);
+	struct mlx5_flow_spec *spec;
+	enum mlx5e_traffic_types tt;
+	int err = 0;
+
+	spec = kvzalloc(sizeof(*spec), GFP_KERNEL);
+	if (!spec) {
+		err = -ENOMEM;
+		goto out;
+	}
+
+	dest.type = MLX5_FLOW_DESTINATION_TYPE_TIR;
+	tt = arfs_get_tt(type);
+	if (tt == -EINVAL) {
+		netdev_err(priv->netdev, "%s: bad arfs_type: %d\n",
+			   __func__, type);
+		err = -EINVAL;
+		goto out;
+	}
+
+	dest.tir_num = tir[tt].tirn;
+
+	arfs_t->default_rule = mlx5_add_flow_rules(arfs_t->ft.t, spec,
+						   &flow_act,
+						   &dest, 1);
+	if (IS_ERR(arfs_t->default_rule)) {
+		err = PTR_ERR(arfs_t->default_rule);
+		arfs_t->default_rule = NULL;
+		netdev_err(priv->netdev, "%s: add rule failed, arfs type=%d\n",
+			   __func__, type);
+	}
+out:
+	kvfree(spec);
+	return err;
+}
+
+#define MLX5E_ARFS_NUM_GROUPS	2
+#define MLX5E_ARFS_GROUP1_SIZE	(BIT(16) - 1)
+#define MLX5E_ARFS_GROUP2_SIZE	BIT(0)
+#define MLX5E_ARFS_TABLE_SIZE	(MLX5E_ARFS_GROUP1_SIZE +\
+				 MLX5E_ARFS_GROUP2_SIZE)
+static int arfs_create_groups(struct mlx5e_flow_table *ft,
+			      enum  arfs_type type)
+{
+	int inlen = MLX5_ST_SZ_BYTES(create_flow_group_in);
+	void *outer_headers_c;
+	int ix = 0;
+	u32 *in;
+	int err;
+	u8 *mc;
+
+	ft->g = kcalloc(MLX5E_ARFS_NUM_GROUPS,
+			sizeof(*ft->g), GFP_KERNEL);
+	in = kvzalloc(inlen, GFP_KERNEL);
+	if  (!in || !ft->g) {
+		kvfree(ft->g);
+		kvfree(in);
+		return -ENOMEM;
+	}
+
+	mc = MLX5_ADDR_OF(create_flow_group_in, in, match_criteria);
+	outer_headers_c = MLX5_ADDR_OF(fte_match_param, mc,
+				       outer_headers);
+	MLX5_SET_TO_ONES(fte_match_set_lyr_2_4, outer_headers_c, ethertype);
+	switch (type) {
+	case ARFS_IPV4_TCP:
+	case ARFS_IPV6_TCP:
+		MLX5_SET_TO_ONES(fte_match_set_lyr_2_4, outer_headers_c, tcp_dport);
+		MLX5_SET_TO_ONES(fte_match_set_lyr_2_4, outer_headers_c, tcp_sport);
+		break;
+	case ARFS_IPV4_UDP:
+	case ARFS_IPV6_UDP:
+		MLX5_SET_TO_ONES(fte_match_set_lyr_2_4, outer_headers_c, udp_dport);
+		MLX5_SET_TO_ONES(fte_match_set_lyr_2_4, outer_headers_c, udp_sport);
+		break;
+	default:
+		err = -EINVAL;
+		goto out;
+	}
+
+	switch (type) {
+	case ARFS_IPV4_TCP:
+	case ARFS_IPV4_UDP:
+		MLX5_SET_TO_ONES(fte_match_set_lyr_2_4, outer_headers_c,
+				 src_ipv4_src_ipv6.ipv4_layout.ipv4);
+		MLX5_SET_TO_ONES(fte_match_set_lyr_2_4, outer_headers_c,
+				 dst_ipv4_dst_ipv6.ipv4_layout.ipv4);
+		break;
+	case ARFS_IPV6_TCP:
+	case ARFS_IPV6_UDP:
+		memset(MLX5_ADDR_OF(fte_match_set_lyr_2_4, outer_headers_c,
+				    src_ipv4_src_ipv6.ipv6_layout.ipv6),
+		       0xff, 16);
+		memset(MLX5_ADDR_OF(fte_match_set_lyr_2_4, outer_headers_c,
+				    dst_ipv4_dst_ipv6.ipv6_layout.ipv6),
+		       0xff, 16);
+		break;
+	default:
+		err = -EINVAL;
+		goto out;
+	}
+
+	MLX5_SET_CFG(in, match_criteria_enable, MLX5_MATCH_OUTER_HEADERS);
+	MLX5_SET_CFG(in, start_flow_index, ix);
+	ix += MLX5E_ARFS_GROUP1_SIZE;
+	MLX5_SET_CFG(in, end_flow_index, ix - 1);
+	ft->g[ft->num_groups] = mlx5_create_flow_group(ft->t, in);
+	if (IS_ERR(ft->g[ft->num_groups]))
+		goto err;
+	ft->num_groups++;
+
+	memset(in, 0, inlen);
+	MLX5_SET_CFG(in, start_flow_index, ix);
+	ix += MLX5E_ARFS_GROUP2_SIZE;
+	MLX5_SET_CFG(in, end_flow_index, ix - 1);
+	ft->g[ft->num_groups] = mlx5_create_flow_group(ft->t, in);
+	if (IS_ERR(ft->g[ft->num_groups]))
+		goto err;
+	ft->num_groups++;
+
+	kvfree(in);
+	return 0;
+
+err:
+	err = PTR_ERR(ft->g[ft->num_groups]);
+	ft->g[ft->num_groups] = NULL;
+out:
+	kvfree(in);
+
+	return err;
+}
+
+static int arfs_create_table(struct mlx5e_priv *priv,
+			     enum arfs_type type)
+{
+	struct mlx5e_arfs_tables *arfs = &priv->fs.arfs;
+	struct mlx5e_flow_table *ft = &arfs->arfs_tables[type].ft;
+	struct mlx5_flow_table_attr ft_attr = {};
+	int err;
+
+	ft->num_groups = 0;
+
+	ft_attr.max_fte = MLX5E_ARFS_TABLE_SIZE;
+	ft_attr.level = MLX5E_ARFS_FT_LEVEL;
+	ft_attr.prio = MLX5E_NIC_PRIO;
+
+	ft->t = mlx5_create_flow_table(priv->fs.ns, &ft_attr);
+	if (IS_ERR(ft->t)) {
+		err = PTR_ERR(ft->t);
+		ft->t = NULL;
+		return err;
+	}
+
+	err = arfs_create_groups(ft, type);
+	if (err)
+		goto err;
+
+	err = arfs_add_default_rule(priv, type);
+	if (err)
+		goto err;
+
+	return 0;
+err:
+	mlx5e_destroy_flow_table(ft);
+	return err;
+}
+
+int mlx5e_arfs_create_tables(struct mlx5e_priv *priv)
+{
+	int err = 0;
+	int i;
+
+	if (!(priv->netdev->hw_features & NETIF_F_NTUPLE))
+		return 0;
+
+	spin_lock_init(&priv->fs.arfs.arfs_lock);
+	INIT_LIST_HEAD(&priv->fs.arfs.rules);
+	priv->fs.arfs.wq = create_singlethread_workqueue("mlx5e_arfs");
+	if (!priv->fs.arfs.wq)
+		return -ENOMEM;
+
+	for (i = 0; i < ARFS_NUM_TYPES; i++) {
+		err = arfs_create_table(priv, i);
+		if (err)
+			goto err;
+	}
+	return 0;
+err:
+	mlx5e_arfs_destroy_tables(priv);
+	return err;
+}
+
+#define MLX5E_ARFS_EXPIRY_QUOTA 60
+
+static void arfs_may_expire_flow(struct mlx5e_priv *priv)
+{
+	struct arfs_rule *arfs_rule;
+	struct hlist_node *htmp;
+	int quota = 0;
+	int i;
+	int j;
+
+	HLIST_HEAD(del_list);
+	spin_lock_bh(&priv->fs.arfs.arfs_lock);
+	mlx5e_for_each_arfs_rule(arfs_rule, htmp, priv->fs.arfs.arfs_tables, i, j) {
+		if (!work_pending(&arfs_rule->arfs_work) &&
+		    rps_may_expire_flow(priv->netdev,
+					arfs_rule->rxq, arfs_rule->flow_id,
+					arfs_rule->filter_id)) {
+			hlist_del_init(&arfs_rule->hlist);
+			hlist_add_head(&arfs_rule->hlist, &del_list);
+			if (quota++ > MLX5E_ARFS_EXPIRY_QUOTA)
+				break;
+		}
+	}
+	spin_unlock_bh(&priv->fs.arfs.arfs_lock);
+	hlist_for_each_entry_safe(arfs_rule, htmp, &del_list, hlist) {
+		if (arfs_rule->rule)
+			mlx5_del_flow_rules(arfs_rule->rule);
+		hlist_del(&arfs_rule->hlist);
+		kfree(arfs_rule);
+	}
+}
+
+static void arfs_del_rules(struct mlx5e_priv *priv)
+{
+	struct hlist_node *htmp;
+	struct arfs_rule *rule;
+	int i;
+	int j;
+
+	HLIST_HEAD(del_list);
+	spin_lock_bh(&priv->fs.arfs.arfs_lock);
+	mlx5e_for_each_arfs_rule(rule, htmp, priv->fs.arfs.arfs_tables, i, j) {
+		hlist_del_init(&rule->hlist);
+		hlist_add_head(&rule->hlist, &del_list);
+	}
+	spin_unlock_bh(&priv->fs.arfs.arfs_lock);
+
+	hlist_for_each_entry_safe(rule, htmp, &del_list, hlist) {
+		cancel_work_sync(&rule->arfs_work);
+		if (rule->rule)
+			mlx5_del_flow_rules(rule->rule);
+		hlist_del(&rule->hlist);
+		kfree(rule);
+	}
+}
+
+static struct hlist_head *
+arfs_hash_bucket(struct arfs_table *arfs_t, __be16 src_port,
+		 __be16 dst_port)
+{
+	unsigned long l;
+	int bucket_idx;
+
+	l = (__force unsigned long)src_port |
+	    ((__force unsigned long)dst_port << 2);
+
+	bucket_idx = hash_long(l, ARFS_HASH_SHIFT);
+
+	return &arfs_t->rules_hash[bucket_idx];
+}
+
+static struct arfs_table *arfs_get_table(struct mlx5e_arfs_tables *arfs,
+					 u8 ip_proto, __be16 etype)
+{
+	if (etype == htons(ETH_P_IP) && ip_proto == IPPROTO_TCP)
+		return &arfs->arfs_tables[ARFS_IPV4_TCP];
+	if (etype == htons(ETH_P_IP) && ip_proto == IPPROTO_UDP)
+		return &arfs->arfs_tables[ARFS_IPV4_UDP];
+	if (etype == htons(ETH_P_IPV6) && ip_proto == IPPROTO_TCP)
+		return &arfs->arfs_tables[ARFS_IPV6_TCP];
+	if (etype == htons(ETH_P_IPV6) && ip_proto == IPPROTO_UDP)
+		return &arfs->arfs_tables[ARFS_IPV6_UDP];
+
+	return NULL;
+}
+
+static struct mlx5_flow_handle *arfs_add_rule(struct mlx5e_priv *priv,
+					      struct arfs_rule *arfs_rule)
+{
+	struct mlx5e_arfs_tables *arfs = &priv->fs.arfs;
+	struct arfs_tuple *tuple = &arfs_rule->tuple;
+	struct mlx5_flow_handle *rule = NULL;
+	struct mlx5_flow_destination dest = {};
+	MLX5_DECLARE_FLOW_ACT(flow_act);
+	struct arfs_table *arfs_table;
+	struct mlx5_flow_spec *spec;
+	struct mlx5_flow_table *ft;
+	int err = 0;
+
+	spec = kvzalloc(sizeof(*spec), GFP_KERNEL);
+	if (!spec) {
+		err = -ENOMEM;
+		goto out;
+	}
+	spec->match_criteria_enable = MLX5_MATCH_OUTER_HEADERS;
+	MLX5_SET_TO_ONES(fte_match_param, spec->match_criteria,
+			 outer_headers.ethertype);
+	MLX5_SET(fte_match_param, spec->match_value, outer_headers.ethertype,
+		 ntohs(tuple->etype));
+	arfs_table = arfs_get_table(arfs, tuple->ip_proto, tuple->etype);
+	if (!arfs_table) {
+		err = -EINVAL;
+		goto out;
+	}
+
+	ft = arfs_table->ft.t;
+	if (tuple->ip_proto == IPPROTO_TCP) {
+		MLX5_SET_TO_ONES(fte_match_param, spec->match_criteria,
+				 outer_headers.tcp_dport);
+		MLX5_SET_TO_ONES(fte_match_param, spec->match_criteria,
+				 outer_headers.tcp_sport);
+		MLX5_SET(fte_match_param, spec->match_value, outer_headers.tcp_dport,
+			 ntohs(tuple->dst_port));
+		MLX5_SET(fte_match_param, spec->match_value, outer_headers.tcp_sport,
+			 ntohs(tuple->src_port));
+	} else {
+		MLX5_SET_TO_ONES(fte_match_param, spec->match_criteria,
+				 outer_headers.udp_dport);
+		MLX5_SET_TO_ONES(fte_match_param, spec->match_criteria,
+				 outer_headers.udp_sport);
+		MLX5_SET(fte_match_param, spec->match_value, outer_headers.udp_dport,
+			 ntohs(tuple->dst_port));
+		MLX5_SET(fte_match_param, spec->match_value, outer_headers.udp_sport,
+			 ntohs(tuple->src_port));
+	}
+	if (tuple->etype == htons(ETH_P_IP)) {
+		memcpy(MLX5_ADDR_OF(fte_match_param, spec->match_value,
+				    outer_headers.src_ipv4_src_ipv6.ipv4_layout.ipv4),
+		       &tuple->src_ipv4,
+		       4);
+		memcpy(MLX5_ADDR_OF(fte_match_param, spec->match_value,
+				    outer_headers.dst_ipv4_dst_ipv6.ipv4_layout.ipv4),
+		       &tuple->dst_ipv4,
+		       4);
+		MLX5_SET_TO_ONES(fte_match_param, spec->match_criteria,
+				 outer_headers.src_ipv4_src_ipv6.ipv4_layout.ipv4);
+		MLX5_SET_TO_ONES(fte_match_param, spec->match_criteria,
+				 outer_headers.dst_ipv4_dst_ipv6.ipv4_layout.ipv4);
+	} else {
+		memcpy(MLX5_ADDR_OF(fte_match_param, spec->match_value,
+				    outer_headers.src_ipv4_src_ipv6.ipv6_layout.ipv6),
+		       &tuple->src_ipv6,
+		       16);
+		memcpy(MLX5_ADDR_OF(fte_match_param, spec->match_value,
+				    outer_headers.dst_ipv4_dst_ipv6.ipv6_layout.ipv6),
+		       &tuple->dst_ipv6,
+		       16);
+		memset(MLX5_ADDR_OF(fte_match_param, spec->match_criteria,
+				    outer_headers.src_ipv4_src_ipv6.ipv6_layout.ipv6),
+		       0xff,
+		       16);
+		memset(MLX5_ADDR_OF(fte_match_param, spec->match_criteria,
+				    outer_headers.dst_ipv4_dst_ipv6.ipv6_layout.ipv6),
+		       0xff,
+		       16);
+	}
+	dest.type = MLX5_FLOW_DESTINATION_TYPE_TIR;
+	dest.tir_num = priv->direct_tir[arfs_rule->rxq].tirn;
+	rule = mlx5_add_flow_rules(ft, spec, &flow_act, &dest, 1);
+	if (IS_ERR(rule)) {
+		err = PTR_ERR(rule);
+		netdev_err(priv->netdev, "%s: add rule(filter id=%d, rq idx=%d) failed, err=%d\n",
+			   __func__, arfs_rule->filter_id, arfs_rule->rxq, err);
+	}
+
+out:
+	kvfree(spec);
+	return err ? ERR_PTR(err) : rule;
+}
+
+static void arfs_modify_rule_rq(struct mlx5e_priv *priv,
+				struct mlx5_flow_handle *rule, u16 rxq)
+{
+	struct mlx5_flow_destination dst = {};
+	int err = 0;
+
+	dst.type = MLX5_FLOW_DESTINATION_TYPE_TIR;
+	dst.tir_num = priv->direct_tir[rxq].tirn;
+	err =  mlx5_modify_rule_destination(rule, &dst, NULL);
+	if (err)
+		netdev_warn(priv->netdev,
+			    "Failed to modify aRFS rule destination to rq=%d\n", rxq);
+}
+
+static void arfs_handle_work(struct work_struct *work)
+{
+	struct arfs_rule *arfs_rule = container_of(work,
+						   struct arfs_rule,
+						   arfs_work);
+	struct mlx5e_priv *priv = arfs_rule->priv;
+	struct mlx5_flow_handle *rule;
+
+	mutex_lock(&priv->state_lock);
+	if (!test_bit(MLX5E_STATE_OPENED, &priv->state)) {
+		spin_lock_bh(&priv->fs.arfs.arfs_lock);
+		hlist_del(&arfs_rule->hlist);
+		spin_unlock_bh(&priv->fs.arfs.arfs_lock);
+
+		mutex_unlock(&priv->state_lock);
+		kfree(arfs_rule);
+		goto out;
+	}
+	mutex_unlock(&priv->state_lock);
+
+	if (!arfs_rule->rule) {
+		rule = arfs_add_rule(priv, arfs_rule);
+		if (IS_ERR(rule))
+			goto out;
+		arfs_rule->rule = rule;
+	} else {
+		arfs_modify_rule_rq(priv, arfs_rule->rule,
+				    arfs_rule->rxq);
+	}
+out:
+	arfs_may_expire_flow(priv);
+}
+
+static struct arfs_rule *arfs_alloc_rule(struct mlx5e_priv *priv,
+					 struct arfs_table *arfs_t,
+					 const struct flow_keys *fk,
+					 u16 rxq, u32 flow_id)
+{
+	struct arfs_rule *rule;
+	struct arfs_tuple *tuple;
+
+	rule = kzalloc(sizeof(*rule), GFP_ATOMIC);
+	if (!rule)
+		return NULL;
+
+	rule->priv = priv;
+	rule->rxq = rxq;
+	INIT_WORK(&rule->arfs_work, arfs_handle_work);
+
+	tuple = &rule->tuple;
+	tuple->etype = fk->basic.n_proto;
+	tuple->ip_proto = fk->basic.ip_proto;
+	if (tuple->etype == htons(ETH_P_IP)) {
+		tuple->src_ipv4 = fk->addrs.v4addrs.src;
+		tuple->dst_ipv4 = fk->addrs.v4addrs.dst;
+	} else {
+		memcpy(&tuple->src_ipv6, &fk->addrs.v6addrs.src,
+		       sizeof(struct in6_addr));
+		memcpy(&tuple->dst_ipv6, &fk->addrs.v6addrs.dst,
+		       sizeof(struct in6_addr));
+	}
+	tuple->src_port = fk->ports.src;
+	tuple->dst_port = fk->ports.dst;
+
+	rule->flow_id = flow_id;
+	rule->filter_id = priv->fs.arfs.last_filter_id++ % RPS_NO_FILTER;
+
+	hlist_add_head(&rule->hlist,
+		       arfs_hash_bucket(arfs_t, tuple->src_port,
+					tuple->dst_port));
+	return rule;
+}
+
+static bool arfs_cmp(const struct arfs_tuple *tuple, const struct flow_keys *fk)
+{
+	if (tuple->src_port != fk->ports.src || tuple->dst_port != fk->ports.dst)
+		return false;
+	if (tuple->etype != fk->basic.n_proto)
+		return false;
+	if (tuple->etype == htons(ETH_P_IP))
+		return tuple->src_ipv4 == fk->addrs.v4addrs.src &&
+		       tuple->dst_ipv4 == fk->addrs.v4addrs.dst;
+	if (tuple->etype == htons(ETH_P_IPV6))
+		return !memcmp(&tuple->src_ipv6, &fk->addrs.v6addrs.src,
+			       sizeof(struct in6_addr)) &&
+		       !memcmp(&tuple->dst_ipv6, &fk->addrs.v6addrs.dst,
+			       sizeof(struct in6_addr));
+	return false;
+}
+
+static struct arfs_rule *arfs_find_rule(struct arfs_table *arfs_t,
+					const struct flow_keys *fk)
+{
+	struct arfs_rule *arfs_rule;
+	struct hlist_head *head;
+
+	head = arfs_hash_bucket(arfs_t, fk->ports.src, fk->ports.dst);
+	hlist_for_each_entry(arfs_rule, head, hlist) {
+		if (arfs_cmp(&arfs_rule->tuple, fk))
+			return arfs_rule;
+	}
+
+	return NULL;
+}
+
+int mlx5e_rx_flow_steer(struct net_device *dev, const struct sk_buff *skb,
+			u16 rxq_index, u32 flow_id)
+{
+	struct mlx5e_priv *priv = netdev_priv(dev);
+	struct mlx5e_arfs_tables *arfs = &priv->fs.arfs;
+	struct arfs_table *arfs_t;
+	struct arfs_rule *arfs_rule;
+	struct flow_keys fk;
+
+	if (!skb_flow_dissect_flow_keys(skb, &fk, 0))
+		return -EPROTONOSUPPORT;
+
+	if (fk.basic.n_proto != htons(ETH_P_IP) &&
+	    fk.basic.n_proto != htons(ETH_P_IPV6))
+		return -EPROTONOSUPPORT;
+
+	if (skb->encapsulation)
+		return -EPROTONOSUPPORT;
+
+	arfs_t = arfs_get_table(arfs, fk.basic.ip_proto, fk.basic.n_proto);
+	if (!arfs_t)
+		return -EPROTONOSUPPORT;
+
+	spin_lock_bh(&arfs->arfs_lock);
+	arfs_rule = arfs_find_rule(arfs_t, &fk);
+	if (arfs_rule) {
+		if (arfs_rule->rxq == rxq_index) {
+			spin_unlock_bh(&arfs->arfs_lock);
+			return arfs_rule->filter_id;
+		}
+		arfs_rule->rxq = rxq_index;
+	} else {
+		arfs_rule = arfs_alloc_rule(priv, arfs_t, &fk, rxq_index, flow_id);
+		if (!arfs_rule) {
+			spin_unlock_bh(&arfs->arfs_lock);
+			return -ENOMEM;
+		}
+	}
+	queue_work(priv->fs.arfs.wq, &arfs_rule->arfs_work);
+	spin_unlock_bh(&arfs->arfs_lock);
+	return arfs_rule->filter_id;
+}
+
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_common.c b/drivers/net/ethernet/mellanox/mlx5/core/en_common.c
new file mode 100644
index 000000000..124e4567a
--- /dev/null
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_common.c
@@ -0,0 +1,194 @@
+/*
+ * Copyright (c) 2016, 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 "en.h"
+
+/* mlx5e global resources should be placed in this file.
+ * Global resources are common to all the netdevices crated on the same nic.
+ */
+
+int mlx5e_create_tir(struct mlx5_core_dev *mdev,
+		     struct mlx5e_tir *tir, u32 *in, int inlen)
+{
+	int err;
+
+	err = mlx5_core_create_tir(mdev, in, inlen, &tir->tirn);
+	if (err)
+		return err;
+
+	mutex_lock(&mdev->mlx5e_res.td.list_lock);
+	list_add(&tir->list, &mdev->mlx5e_res.td.tirs_list);
+	mutex_unlock(&mdev->mlx5e_res.td.list_lock);
+
+	return 0;
+}
+
+void mlx5e_destroy_tir(struct mlx5_core_dev *mdev,
+		       struct mlx5e_tir *tir)
+{
+	mutex_lock(&mdev->mlx5e_res.td.list_lock);
+	mlx5_core_destroy_tir(mdev, tir->tirn);
+	list_del(&tir->list);
+	mutex_unlock(&mdev->mlx5e_res.td.list_lock);
+}
+
+static int mlx5e_create_mkey(struct mlx5_core_dev *mdev, u32 pdn,
+			     struct mlx5_core_mkey *mkey)
+{
+	int inlen = MLX5_ST_SZ_BYTES(create_mkey_in);
+	void *mkc;
+	u32 *in;
+	int err;
+
+	in = kvzalloc(inlen, GFP_KERNEL);
+	if (!in)
+		return -ENOMEM;
+
+	mkc = MLX5_ADDR_OF(create_mkey_in, in, memory_key_mkey_entry);
+	MLX5_SET(mkc, mkc, access_mode_1_0, MLX5_MKC_ACCESS_MODE_PA);
+	MLX5_SET(mkc, mkc, lw, 1);
+	MLX5_SET(mkc, mkc, lr, 1);
+
+	MLX5_SET(mkc, mkc, pd, pdn);
+	MLX5_SET(mkc, mkc, length64, 1);
+	MLX5_SET(mkc, mkc, qpn, 0xffffff);
+
+	err = mlx5_core_create_mkey(mdev, mkey, in, inlen);
+
+	kvfree(in);
+	return err;
+}
+
+int mlx5e_create_mdev_resources(struct mlx5_core_dev *mdev)
+{
+	struct mlx5e_resources *res = &mdev->mlx5e_res;
+	int err;
+
+	err = mlx5_core_alloc_pd(mdev, &res->pdn);
+	if (err) {
+		mlx5_core_err(mdev, "alloc pd failed, %d\n", err);
+		return err;
+	}
+
+	err = mlx5_core_alloc_transport_domain(mdev, &res->td.tdn);
+	if (err) {
+		mlx5_core_err(mdev, "alloc td failed, %d\n", err);
+		goto err_dealloc_pd;
+	}
+
+	err = mlx5e_create_mkey(mdev, res->pdn, &res->mkey);
+	if (err) {
+		mlx5_core_err(mdev, "create mkey failed, %d\n", err);
+		goto err_dealloc_transport_domain;
+	}
+
+	err = mlx5_alloc_bfreg(mdev, &res->bfreg, false, false);
+	if (err) {
+		mlx5_core_err(mdev, "alloc bfreg failed, %d\n", err);
+		goto err_destroy_mkey;
+	}
+
+	INIT_LIST_HEAD(&mdev->mlx5e_res.td.tirs_list);
+	mutex_init(&mdev->mlx5e_res.td.list_lock);
+
+	return 0;
+
+err_destroy_mkey:
+	mlx5_core_destroy_mkey(mdev, &res->mkey);
+err_dealloc_transport_domain:
+	mlx5_core_dealloc_transport_domain(mdev, res->td.tdn);
+err_dealloc_pd:
+	mlx5_core_dealloc_pd(mdev, res->pdn);
+	return err;
+}
+
+void mlx5e_destroy_mdev_resources(struct mlx5_core_dev *mdev)
+{
+	struct mlx5e_resources *res = &mdev->mlx5e_res;
+
+	mlx5_free_bfreg(mdev, &res->bfreg);
+	mlx5_core_destroy_mkey(mdev, &res->mkey);
+	mlx5_core_dealloc_transport_domain(mdev, res->td.tdn);
+	mlx5_core_dealloc_pd(mdev, res->pdn);
+	memset(res, 0, sizeof(*res));
+}
+
+int mlx5e_refresh_tirs(struct mlx5e_priv *priv, bool enable_uc_lb)
+{
+	struct mlx5_core_dev *mdev = priv->mdev;
+	struct mlx5e_tir *tir;
+	int err  = 0;
+	u32 tirn = 0;
+	int inlen;
+	void *in;
+
+	inlen = MLX5_ST_SZ_BYTES(modify_tir_in);
+	in = kvzalloc(inlen, GFP_KERNEL);
+	if (!in) {
+		err = -ENOMEM;
+		goto out;
+	}
+
+	if (enable_uc_lb)
+		MLX5_SET(modify_tir_in, in, ctx.self_lb_block,
+			 MLX5_TIRC_SELF_LB_BLOCK_BLOCK_UNICAST_);
+
+	MLX5_SET(modify_tir_in, in, bitmask.self_lb_en, 1);
+
+	mutex_lock(&mdev->mlx5e_res.td.list_lock);
+	list_for_each_entry(tir, &mdev->mlx5e_res.td.tirs_list, list) {
+		tirn = tir->tirn;
+		err = mlx5_core_modify_tir(mdev, tirn, in, inlen);
+		if (err)
+			goto out;
+	}
+
+out:
+	kvfree(in);
+	if (err)
+		netdev_err(priv->netdev, "refresh tir(0x%x) failed, %d\n", tirn, err);
+	mutex_unlock(&mdev->mlx5e_res.td.list_lock);
+
+	return err;
+}
+
+u8 mlx5e_params_calculate_tx_min_inline(struct mlx5_core_dev *mdev)
+{
+	u8 min_inline_mode;
+
+	mlx5_query_min_inline(mdev, &min_inline_mode);
+	if (min_inline_mode == MLX5_INLINE_MODE_NONE &&
+	    !MLX5_CAP_ETH(mdev, wqe_vlan_insert))
+		min_inline_mode = MLX5_INLINE_MODE_L2;
+
+	return min_inline_mode;
+}
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_dcbnl.c b/drivers/net/ethernet/mellanox/mlx5/core/en_dcbnl.c
new file mode 100644
index 000000000..722998d68
--- /dev/null
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_dcbnl.c
@@ -0,0 +1,1206 @@
+/*
+ * Copyright (c) 2016, 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/device.h>
+#include <linux/netdevice.h>
+#include "en.h"
+#include "en/port.h"
+#include "en/port_buffer.h"
+
+#define MLX5E_100MB (100000)
+#define MLX5E_1GB   (1000000)
+
+#define MLX5E_CEE_STATE_UP    1
+#define MLX5E_CEE_STATE_DOWN  0
+
+/* Max supported cable length is 1000 meters */
+#define MLX5E_MAX_CABLE_LENGTH 1000
+
+enum {
+	MLX5E_VENDOR_TC_GROUP_NUM = 7,
+	MLX5E_LOWEST_PRIO_GROUP   = 0,
+};
+
+#define MLX5_DSCP_SUPPORTED(mdev) (MLX5_CAP_GEN(mdev, qcam_reg)  && \
+				   MLX5_CAP_QCAM_REG(mdev, qpts) && \
+				   MLX5_CAP_QCAM_REG(mdev, qpdpm))
+
+static int mlx5e_set_trust_state(struct mlx5e_priv *priv, u8 trust_state);
+static int mlx5e_set_dscp2prio(struct mlx5e_priv *priv, u8 dscp, u8 prio);
+
+/* If dcbx mode is non-host set the dcbx mode to host.
+ */
+static int mlx5e_dcbnl_set_dcbx_mode(struct mlx5e_priv *priv,
+				     enum mlx5_dcbx_oper_mode mode)
+{
+	struct mlx5_core_dev *mdev = priv->mdev;
+	u32 param[MLX5_ST_SZ_DW(dcbx_param)];
+	int err;
+
+	err = mlx5_query_port_dcbx_param(mdev, param);
+	if (err)
+		return err;
+
+	MLX5_SET(dcbx_param, param, version_admin, mode);
+	if (mode != MLX5E_DCBX_PARAM_VER_OPER_HOST)
+		MLX5_SET(dcbx_param, param, willing_admin, 1);
+
+	return mlx5_set_port_dcbx_param(mdev, param);
+}
+
+static int mlx5e_dcbnl_switch_to_host_mode(struct mlx5e_priv *priv)
+{
+	struct mlx5e_dcbx *dcbx = &priv->dcbx;
+	int err;
+
+	if (!MLX5_CAP_GEN(priv->mdev, dcbx))
+		return 0;
+
+	if (dcbx->mode == MLX5E_DCBX_PARAM_VER_OPER_HOST)
+		return 0;
+
+	err = mlx5e_dcbnl_set_dcbx_mode(priv, MLX5E_DCBX_PARAM_VER_OPER_HOST);
+	if (err)
+		return err;
+
+	dcbx->mode = MLX5E_DCBX_PARAM_VER_OPER_HOST;
+	return 0;
+}
+
+static int mlx5e_dcbnl_ieee_getets(struct net_device *netdev,
+				   struct ieee_ets *ets)
+{
+	struct mlx5e_priv *priv = netdev_priv(netdev);
+	struct mlx5_core_dev *mdev = priv->mdev;
+	u8 tc_group[IEEE_8021QAZ_MAX_TCS];
+	bool is_tc_group_6_exist = false;
+	bool is_zero_bw_ets_tc = false;
+	int err = 0;
+	int i;
+
+	if (!MLX5_CAP_GEN(priv->mdev, ets))
+		return -EOPNOTSUPP;
+
+	ets->ets_cap = mlx5_max_tc(priv->mdev) + 1;
+	for (i = 0; i < ets->ets_cap; i++) {
+		err = mlx5_query_port_prio_tc(mdev, i, &ets->prio_tc[i]);
+		if (err)
+			return err;
+
+		err = mlx5_query_port_tc_group(mdev, i, &tc_group[i]);
+		if (err)
+			return err;
+
+		err = mlx5_query_port_tc_bw_alloc(mdev, i, &ets->tc_tx_bw[i]);
+		if (err)
+			return err;
+
+		if (ets->tc_tx_bw[i] < MLX5E_MAX_BW_ALLOC &&
+		    tc_group[i] == (MLX5E_LOWEST_PRIO_GROUP + 1))
+			is_zero_bw_ets_tc = true;
+
+		if (tc_group[i] == (MLX5E_VENDOR_TC_GROUP_NUM - 1))
+			is_tc_group_6_exist = true;
+	}
+
+	/* Report 0% ets tc if exits*/
+	if (is_zero_bw_ets_tc) {
+		for (i = 0; i < ets->ets_cap; i++)
+			if (tc_group[i] == MLX5E_LOWEST_PRIO_GROUP)
+				ets->tc_tx_bw[i] = 0;
+	}
+
+	/* Update tc_tsa based on fw setting*/
+	for (i = 0; i < ets->ets_cap; i++) {
+		if (ets->tc_tx_bw[i] < MLX5E_MAX_BW_ALLOC)
+			priv->dcbx.tc_tsa[i] = IEEE_8021QAZ_TSA_ETS;
+		else if (tc_group[i] == MLX5E_VENDOR_TC_GROUP_NUM &&
+			 !is_tc_group_6_exist)
+			priv->dcbx.tc_tsa[i] = IEEE_8021QAZ_TSA_VENDOR;
+	}
+	memcpy(ets->tc_tsa, priv->dcbx.tc_tsa, sizeof(ets->tc_tsa));
+
+	return err;
+}
+
+static void mlx5e_build_tc_group(struct ieee_ets *ets, u8 *tc_group, int max_tc)
+{
+	bool any_tc_mapped_to_ets = false;
+	bool ets_zero_bw = false;
+	int strict_group;
+	int i;
+
+	for (i = 0; i <= max_tc; i++) {
+		if (ets->tc_tsa[i] == IEEE_8021QAZ_TSA_ETS) {
+			any_tc_mapped_to_ets = true;
+			if (!ets->tc_tx_bw[i])
+				ets_zero_bw = true;
+		}
+	}
+
+	/* strict group has higher priority than ets group */
+	strict_group = MLX5E_LOWEST_PRIO_GROUP;
+	if (any_tc_mapped_to_ets)
+		strict_group++;
+	if (ets_zero_bw)
+		strict_group++;
+
+	for (i = 0; i <= max_tc; i++) {
+		switch (ets->tc_tsa[i]) {
+		case IEEE_8021QAZ_TSA_VENDOR:
+			tc_group[i] = MLX5E_VENDOR_TC_GROUP_NUM;
+			break;
+		case IEEE_8021QAZ_TSA_STRICT:
+			tc_group[i] = strict_group++;
+			break;
+		case IEEE_8021QAZ_TSA_ETS:
+			tc_group[i] = MLX5E_LOWEST_PRIO_GROUP;
+			if (ets->tc_tx_bw[i] && ets_zero_bw)
+				tc_group[i] = MLX5E_LOWEST_PRIO_GROUP + 1;
+			break;
+		}
+	}
+}
+
+static void mlx5e_build_tc_tx_bw(struct ieee_ets *ets, u8 *tc_tx_bw,
+				 u8 *tc_group, int max_tc)
+{
+	int bw_for_ets_zero_bw_tc = 0;
+	int last_ets_zero_bw_tc = -1;
+	int num_ets_zero_bw = 0;
+	int i;
+
+	for (i = 0; i <= max_tc; i++) {
+		if (ets->tc_tsa[i] == IEEE_8021QAZ_TSA_ETS &&
+		    !ets->tc_tx_bw[i]) {
+			num_ets_zero_bw++;
+			last_ets_zero_bw_tc = i;
+		}
+	}
+
+	if (num_ets_zero_bw)
+		bw_for_ets_zero_bw_tc = MLX5E_MAX_BW_ALLOC / num_ets_zero_bw;
+
+	for (i = 0; i <= max_tc; i++) {
+		switch (ets->tc_tsa[i]) {
+		case IEEE_8021QAZ_TSA_VENDOR:
+			tc_tx_bw[i] = MLX5E_MAX_BW_ALLOC;
+			break;
+		case IEEE_8021QAZ_TSA_STRICT:
+			tc_tx_bw[i] = MLX5E_MAX_BW_ALLOC;
+			break;
+		case IEEE_8021QAZ_TSA_ETS:
+			tc_tx_bw[i] = ets->tc_tx_bw[i] ?
+				      ets->tc_tx_bw[i] :
+				      bw_for_ets_zero_bw_tc;
+			break;
+		}
+	}
+
+	/* Make sure the total bw for ets zero bw group is 100% */
+	if (last_ets_zero_bw_tc != -1)
+		tc_tx_bw[last_ets_zero_bw_tc] +=
+			MLX5E_MAX_BW_ALLOC % num_ets_zero_bw;
+}
+
+/* If there are ETS BW 0,
+ *   Set ETS group # to 1 for all ETS non zero BW tcs. Their sum must be 100%.
+ *   Set group #0 to all the ETS BW 0 tcs and
+ *     equally splits the 100% BW between them
+ *   Report both group #0 and #1 as ETS type.
+ *     All the tcs in group #0 will be reported with 0% BW.
+ */
+int mlx5e_dcbnl_ieee_setets_core(struct mlx5e_priv *priv, struct ieee_ets *ets)
+{
+	struct mlx5_core_dev *mdev = priv->mdev;
+	u8 tc_tx_bw[IEEE_8021QAZ_MAX_TCS];
+	u8 tc_group[IEEE_8021QAZ_MAX_TCS];
+	int max_tc = mlx5_max_tc(mdev);
+	int err, i;
+
+	mlx5e_build_tc_group(ets, tc_group, max_tc);
+	mlx5e_build_tc_tx_bw(ets, tc_tx_bw, tc_group, max_tc);
+
+	err = mlx5_set_port_prio_tc(mdev, ets->prio_tc);
+	if (err)
+		return err;
+
+	err = mlx5_set_port_tc_group(mdev, tc_group);
+	if (err)
+		return err;
+
+	err = mlx5_set_port_tc_bw_alloc(mdev, tc_tx_bw);
+
+	if (err)
+		return err;
+
+	memcpy(priv->dcbx.tc_tsa, ets->tc_tsa, sizeof(ets->tc_tsa));
+
+	for (i = 0; i < IEEE_8021QAZ_MAX_TCS; i++) {
+		mlx5e_dbg(HW, priv, "%s: prio_%d <=> tc_%d\n",
+			  __func__, i, ets->prio_tc[i]);
+		mlx5e_dbg(HW, priv, "%s: tc_%d <=> tx_bw_%d%%, group_%d\n",
+			  __func__, i, tc_tx_bw[i], tc_group[i]);
+	}
+
+	return err;
+}
+
+static int mlx5e_dbcnl_validate_ets(struct net_device *netdev,
+				    struct ieee_ets *ets,
+				    bool zero_sum_allowed)
+{
+	bool have_ets_tc = false;
+	int bw_sum = 0;
+	int i;
+
+	/* Validate Priority */
+	for (i = 0; i < IEEE_8021QAZ_MAX_TCS; i++) {
+		if (ets->prio_tc[i] >= MLX5E_MAX_PRIORITY) {
+			netdev_err(netdev,
+				   "Failed to validate ETS: priority value greater than max(%d)\n",
+				    MLX5E_MAX_PRIORITY);
+			return -EINVAL;
+		}
+	}
+
+	/* Validate Bandwidth Sum */
+	for (i = 0; i < IEEE_8021QAZ_MAX_TCS; i++) {
+		if (ets->tc_tsa[i] == IEEE_8021QAZ_TSA_ETS) {
+			have_ets_tc = true;
+			bw_sum += ets->tc_tx_bw[i];
+		}
+	}
+
+	if (have_ets_tc && bw_sum != 100) {
+		if (bw_sum || (!bw_sum && !zero_sum_allowed))
+			netdev_err(netdev,
+				   "Failed to validate ETS: BW sum is illegal\n");
+		return -EINVAL;
+	}
+	return 0;
+}
+
+static int mlx5e_dcbnl_ieee_setets(struct net_device *netdev,
+				   struct ieee_ets *ets)
+{
+	struct mlx5e_priv *priv = netdev_priv(netdev);
+	int err;
+
+	if (!MLX5_CAP_GEN(priv->mdev, ets))
+		return -EOPNOTSUPP;
+
+	err = mlx5e_dbcnl_validate_ets(netdev, ets, false);
+	if (err)
+		return err;
+
+	err = mlx5e_dcbnl_ieee_setets_core(priv, ets);
+	if (err)
+		return err;
+
+	return 0;
+}
+
+static int mlx5e_dcbnl_ieee_getpfc(struct net_device *dev,
+				   struct ieee_pfc *pfc)
+{
+	struct mlx5e_priv *priv = netdev_priv(dev);
+	struct mlx5_core_dev *mdev = priv->mdev;
+	struct mlx5e_pport_stats *pstats = &priv->stats.pport;
+	int i;
+
+	pfc->pfc_cap = mlx5_max_tc(mdev) + 1;
+	for (i = 0; i < IEEE_8021QAZ_MAX_TCS; i++) {
+		pfc->requests[i]    = PPORT_PER_PRIO_GET(pstats, i, tx_pause);
+		pfc->indications[i] = PPORT_PER_PRIO_GET(pstats, i, rx_pause);
+	}
+
+	if (MLX5_BUFFER_SUPPORTED(mdev))
+		pfc->delay = priv->dcbx.cable_len;
+
+	return mlx5_query_port_pfc(mdev, &pfc->pfc_en, NULL);
+}
+
+static int mlx5e_dcbnl_ieee_setpfc(struct net_device *dev,
+				   struct ieee_pfc *pfc)
+{
+	struct mlx5e_priv *priv = netdev_priv(dev);
+	struct mlx5_core_dev *mdev = priv->mdev;
+	u32 old_cable_len = priv->dcbx.cable_len;
+	struct ieee_pfc pfc_new;
+	u32 changed = 0;
+	u8 curr_pfc_en;
+	int ret = 0;
+
+	/* pfc_en */
+	mlx5_query_port_pfc(mdev, &curr_pfc_en, NULL);
+	if (pfc->pfc_en != curr_pfc_en) {
+		ret = mlx5_set_port_pfc(mdev, pfc->pfc_en, pfc->pfc_en);
+		if (ret)
+			return ret;
+		mlx5_toggle_port_link(mdev);
+		changed |= MLX5E_PORT_BUFFER_PFC;
+	}
+
+	if (pfc->delay &&
+	    pfc->delay < MLX5E_MAX_CABLE_LENGTH &&
+	    pfc->delay != priv->dcbx.cable_len) {
+		priv->dcbx.cable_len = pfc->delay;
+		changed |= MLX5E_PORT_BUFFER_CABLE_LEN;
+	}
+
+	if (MLX5_BUFFER_SUPPORTED(mdev)) {
+		pfc_new.pfc_en = (changed & MLX5E_PORT_BUFFER_PFC) ? pfc->pfc_en : curr_pfc_en;
+		if (priv->dcbx.manual_buffer)
+			ret = mlx5e_port_manual_buffer_config(priv, changed,
+							      dev->mtu, &pfc_new,
+							      NULL, NULL);
+
+		if (ret && (changed & MLX5E_PORT_BUFFER_CABLE_LEN))
+			priv->dcbx.cable_len = old_cable_len;
+	}
+
+	if (!ret) {
+		mlx5e_dbg(HW, priv,
+			  "%s: PFC per priority bit mask: 0x%x\n",
+			  __func__, pfc->pfc_en);
+	}
+	return ret;
+}
+
+static u8 mlx5e_dcbnl_getdcbx(struct net_device *dev)
+{
+	struct mlx5e_priv *priv = netdev_priv(dev);
+
+	return priv->dcbx.cap;
+}
+
+static u8 mlx5e_dcbnl_setdcbx(struct net_device *dev, u8 mode)
+{
+	struct mlx5e_priv *priv = netdev_priv(dev);
+	struct mlx5e_dcbx *dcbx = &priv->dcbx;
+
+	if (mode & DCB_CAP_DCBX_LLD_MANAGED)
+		return 1;
+
+	if ((!mode) && MLX5_CAP_GEN(priv->mdev, dcbx)) {
+		if (dcbx->mode == MLX5E_DCBX_PARAM_VER_OPER_AUTO)
+			return 0;
+
+		/* set dcbx to fw controlled */
+		if (!mlx5e_dcbnl_set_dcbx_mode(priv, MLX5E_DCBX_PARAM_VER_OPER_AUTO)) {
+			dcbx->mode = MLX5E_DCBX_PARAM_VER_OPER_AUTO;
+			dcbx->cap &= ~DCB_CAP_DCBX_HOST;
+			return 0;
+		}
+
+		return 1;
+	}
+
+	if (!(mode & DCB_CAP_DCBX_HOST))
+		return 1;
+
+	if (mlx5e_dcbnl_switch_to_host_mode(netdev_priv(dev)))
+		return 1;
+
+	dcbx->cap = mode;
+
+	return 0;
+}
+
+static int mlx5e_dcbnl_ieee_setapp(struct net_device *dev, struct dcb_app *app)
+{
+	struct mlx5e_priv *priv = netdev_priv(dev);
+	struct dcb_app temp;
+	bool is_new;
+	int err;
+
+	if (!MLX5_CAP_GEN(priv->mdev, vport_group_manager) ||
+	    !MLX5_DSCP_SUPPORTED(priv->mdev))
+		return -EOPNOTSUPP;
+
+	if ((app->selector != IEEE_8021QAZ_APP_SEL_DSCP) ||
+	    (app->protocol >= MLX5E_MAX_DSCP))
+		return -EINVAL;
+
+	/* Save the old entry info */
+	temp.selector = IEEE_8021QAZ_APP_SEL_DSCP;
+	temp.protocol = app->protocol;
+	temp.priority = priv->dcbx_dp.dscp2prio[app->protocol];
+
+	/* Check if need to switch to dscp trust state */
+	if (!priv->dcbx.dscp_app_cnt) {
+		err =  mlx5e_set_trust_state(priv, MLX5_QPTS_TRUST_DSCP);
+		if (err)
+			return err;
+	}
+
+	/* Skip the fw command if new and old mapping are the same */
+	if (app->priority != priv->dcbx_dp.dscp2prio[app->protocol]) {
+		err = mlx5e_set_dscp2prio(priv, app->protocol, app->priority);
+		if (err)
+			goto fw_err;
+	}
+
+	/* Delete the old entry if exists */
+	is_new = false;
+	err = dcb_ieee_delapp(dev, &temp);
+	if (err)
+		is_new = true;
+
+	/* Add new entry and update counter */
+	err = dcb_ieee_setapp(dev, app);
+	if (err)
+		return err;
+
+	if (is_new)
+		priv->dcbx.dscp_app_cnt++;
+
+	return err;
+
+fw_err:
+	mlx5e_set_trust_state(priv, MLX5_QPTS_TRUST_PCP);
+	return err;
+}
+
+static int mlx5e_dcbnl_ieee_delapp(struct net_device *dev, struct dcb_app *app)
+{
+	struct mlx5e_priv *priv = netdev_priv(dev);
+	int err;
+
+	if  (!MLX5_CAP_GEN(priv->mdev, vport_group_manager) ||
+	     !MLX5_DSCP_SUPPORTED(priv->mdev))
+		return -EOPNOTSUPP;
+
+	if ((app->selector != IEEE_8021QAZ_APP_SEL_DSCP) ||
+	    (app->protocol >= MLX5E_MAX_DSCP))
+		return -EINVAL;
+
+	/* Skip if no dscp app entry */
+	if (!priv->dcbx.dscp_app_cnt)
+		return -ENOENT;
+
+	/* Check if the entry matches fw setting */
+	if (app->priority != priv->dcbx_dp.dscp2prio[app->protocol])
+		return -ENOENT;
+
+	/* Delete the app entry */
+	err = dcb_ieee_delapp(dev, app);
+	if (err)
+		return err;
+
+	/* Reset the priority mapping back to zero */
+	err = mlx5e_set_dscp2prio(priv, app->protocol, 0);
+	if (err)
+		goto fw_err;
+
+	priv->dcbx.dscp_app_cnt--;
+
+	/* Check if need to switch to pcp trust state */
+	if (!priv->dcbx.dscp_app_cnt)
+		err = mlx5e_set_trust_state(priv, MLX5_QPTS_TRUST_PCP);
+
+	return err;
+
+fw_err:
+	mlx5e_set_trust_state(priv, MLX5_QPTS_TRUST_PCP);
+	return err;
+}
+
+static int mlx5e_dcbnl_ieee_getmaxrate(struct net_device *netdev,
+				       struct ieee_maxrate *maxrate)
+{
+	struct mlx5e_priv *priv    = netdev_priv(netdev);
+	struct mlx5_core_dev *mdev = priv->mdev;
+	u8 max_bw_value[IEEE_8021QAZ_MAX_TCS];
+	u8 max_bw_unit[IEEE_8021QAZ_MAX_TCS];
+	int err;
+	int i;
+
+	err = mlx5_query_port_ets_rate_limit(mdev, max_bw_value, max_bw_unit);
+	if (err)
+		return err;
+
+	memset(maxrate->tc_maxrate, 0, sizeof(maxrate->tc_maxrate));
+
+	for (i = 0; i <= mlx5_max_tc(mdev); i++) {
+		switch (max_bw_unit[i]) {
+		case MLX5_100_MBPS_UNIT:
+			maxrate->tc_maxrate[i] = max_bw_value[i] * MLX5E_100MB;
+			break;
+		case MLX5_GBPS_UNIT:
+			maxrate->tc_maxrate[i] = max_bw_value[i] * MLX5E_1GB;
+			break;
+		case MLX5_BW_NO_LIMIT:
+			break;
+		default:
+			WARN(true, "non-supported BW unit");
+			break;
+		}
+	}
+
+	return 0;
+}
+
+static int mlx5e_dcbnl_ieee_setmaxrate(struct net_device *netdev,
+				       struct ieee_maxrate *maxrate)
+{
+	struct mlx5e_priv *priv    = netdev_priv(netdev);
+	struct mlx5_core_dev *mdev = priv->mdev;
+	u8 max_bw_value[IEEE_8021QAZ_MAX_TCS];
+	u8 max_bw_unit[IEEE_8021QAZ_MAX_TCS];
+	__u64 upper_limit_mbps = roundup(255 * MLX5E_100MB, MLX5E_1GB);
+	int i;
+
+	memset(max_bw_value, 0, sizeof(max_bw_value));
+	memset(max_bw_unit, 0, sizeof(max_bw_unit));
+
+	for (i = 0; i <= mlx5_max_tc(mdev); i++) {
+		if (!maxrate->tc_maxrate[i]) {
+			max_bw_unit[i]  = MLX5_BW_NO_LIMIT;
+			continue;
+		}
+		if (maxrate->tc_maxrate[i] < upper_limit_mbps) {
+			max_bw_value[i] = div_u64(maxrate->tc_maxrate[i],
+						  MLX5E_100MB);
+			max_bw_value[i] = max_bw_value[i] ? max_bw_value[i] : 1;
+			max_bw_unit[i]  = MLX5_100_MBPS_UNIT;
+		} else {
+			max_bw_value[i] = div_u64(maxrate->tc_maxrate[i],
+						  MLX5E_1GB);
+			max_bw_unit[i]  = MLX5_GBPS_UNIT;
+		}
+	}
+
+	for (i = 0; i < IEEE_8021QAZ_MAX_TCS; i++) {
+		mlx5e_dbg(HW, priv, "%s: tc_%d <=> max_bw %d Gbps\n",
+			  __func__, i, max_bw_value[i]);
+	}
+
+	return mlx5_modify_port_ets_rate_limit(mdev, max_bw_value, max_bw_unit);
+}
+
+static u8 mlx5e_dcbnl_setall(struct net_device *netdev)
+{
+	struct mlx5e_priv *priv = netdev_priv(netdev);
+	struct mlx5e_cee_config *cee_cfg = &priv->dcbx.cee_cfg;
+	struct mlx5_core_dev *mdev = priv->mdev;
+	struct ieee_ets ets;
+	struct ieee_pfc pfc;
+	int err = -EOPNOTSUPP;
+	int i;
+
+	if (!MLX5_CAP_GEN(mdev, ets))
+		goto out;
+
+	memset(&ets, 0, sizeof(ets));
+	memset(&pfc, 0, sizeof(pfc));
+
+	ets.ets_cap = IEEE_8021QAZ_MAX_TCS;
+	for (i = 0; i < CEE_DCBX_MAX_PGS; i++) {
+		ets.tc_tx_bw[i] = cee_cfg->pg_bw_pct[i];
+		ets.tc_rx_bw[i] = cee_cfg->pg_bw_pct[i];
+		ets.tc_tsa[i]   = IEEE_8021QAZ_TSA_ETS;
+		ets.prio_tc[i]  = cee_cfg->prio_to_pg_map[i];
+		mlx5e_dbg(HW, priv,
+			  "%s: Priority group %d: tx_bw %d, rx_bw %d, prio_tc %d\n",
+			  __func__, i, ets.tc_tx_bw[i], ets.tc_rx_bw[i],
+			  ets.prio_tc[i]);
+	}
+
+	err = mlx5e_dbcnl_validate_ets(netdev, &ets, true);
+	if (err)
+		goto out;
+
+	err = mlx5e_dcbnl_ieee_setets_core(priv, &ets);
+	if (err) {
+		netdev_err(netdev,
+			   "%s, Failed to set ETS: %d\n", __func__, err);
+		goto out;
+	}
+
+	/* Set PFC */
+	pfc.pfc_cap = mlx5_max_tc(mdev) + 1;
+	if (!cee_cfg->pfc_enable)
+		pfc.pfc_en = 0;
+	else
+		for (i = 0; i < CEE_DCBX_MAX_PRIO; i++)
+			pfc.pfc_en |= cee_cfg->pfc_setting[i] << i;
+
+	err = mlx5e_dcbnl_ieee_setpfc(netdev, &pfc);
+	if (err) {
+		netdev_err(netdev,
+			   "%s, Failed to set PFC: %d\n", __func__, err);
+		goto out;
+	}
+out:
+	return err ? MLX5_DCB_NO_CHG : MLX5_DCB_CHG_RESET;
+}
+
+static u8 mlx5e_dcbnl_getstate(struct net_device *netdev)
+{
+	return MLX5E_CEE_STATE_UP;
+}
+
+static void mlx5e_dcbnl_getpermhwaddr(struct net_device *netdev,
+				      u8 *perm_addr)
+{
+	struct mlx5e_priv *priv = netdev_priv(netdev);
+
+	if (!perm_addr)
+		return;
+
+	memset(perm_addr, 0xff, MAX_ADDR_LEN);
+
+	mlx5_query_nic_vport_mac_address(priv->mdev, 0, perm_addr);
+}
+
+static void mlx5e_dcbnl_setpgtccfgtx(struct net_device *netdev,
+				     int priority, u8 prio_type,
+				     u8 pgid, u8 bw_pct, u8 up_map)
+{
+	struct mlx5e_priv *priv = netdev_priv(netdev);
+	struct mlx5e_cee_config *cee_cfg = &priv->dcbx.cee_cfg;
+
+	if (priority >= CEE_DCBX_MAX_PRIO) {
+		netdev_err(netdev,
+			   "%s, priority is out of range\n", __func__);
+		return;
+	}
+
+	if (pgid >= CEE_DCBX_MAX_PGS) {
+		netdev_err(netdev,
+			   "%s, priority group is out of range\n", __func__);
+		return;
+	}
+
+	cee_cfg->prio_to_pg_map[priority] = pgid;
+}
+
+static void mlx5e_dcbnl_setpgbwgcfgtx(struct net_device *netdev,
+				      int pgid, u8 bw_pct)
+{
+	struct mlx5e_priv *priv = netdev_priv(netdev);
+	struct mlx5e_cee_config *cee_cfg = &priv->dcbx.cee_cfg;
+
+	if (pgid >= CEE_DCBX_MAX_PGS) {
+		netdev_err(netdev,
+			   "%s, priority group is out of range\n", __func__);
+		return;
+	}
+
+	cee_cfg->pg_bw_pct[pgid] = bw_pct;
+}
+
+static void mlx5e_dcbnl_getpgtccfgtx(struct net_device *netdev,
+				     int priority, u8 *prio_type,
+				     u8 *pgid, u8 *bw_pct, u8 *up_map)
+{
+	struct mlx5e_priv *priv = netdev_priv(netdev);
+	struct mlx5_core_dev *mdev = priv->mdev;
+
+	if (!MLX5_CAP_GEN(priv->mdev, ets)) {
+		netdev_err(netdev, "%s, ets is not supported\n", __func__);
+		return;
+	}
+
+	if (priority >= CEE_DCBX_MAX_PRIO) {
+		netdev_err(netdev,
+			   "%s, priority is out of range\n", __func__);
+		return;
+	}
+
+	*prio_type = 0;
+	*bw_pct = 0;
+	*up_map = 0;
+
+	if (mlx5_query_port_prio_tc(mdev, priority, pgid))
+		*pgid = 0;
+}
+
+static void mlx5e_dcbnl_getpgbwgcfgtx(struct net_device *netdev,
+				      int pgid, u8 *bw_pct)
+{
+	struct ieee_ets ets;
+
+	if (pgid >= CEE_DCBX_MAX_PGS) {
+		netdev_err(netdev,
+			   "%s, priority group is out of range\n", __func__);
+		return;
+	}
+
+	mlx5e_dcbnl_ieee_getets(netdev, &ets);
+	*bw_pct = ets.tc_tx_bw[pgid];
+}
+
+static void mlx5e_dcbnl_setpfccfg(struct net_device *netdev,
+				  int priority, u8 setting)
+{
+	struct mlx5e_priv *priv = netdev_priv(netdev);
+	struct mlx5e_cee_config *cee_cfg = &priv->dcbx.cee_cfg;
+
+	if (priority >= CEE_DCBX_MAX_PRIO) {
+		netdev_err(netdev,
+			   "%s, priority is out of range\n", __func__);
+		return;
+	}
+
+	if (setting > 1)
+		return;
+
+	cee_cfg->pfc_setting[priority] = setting;
+}
+
+static int
+mlx5e_dcbnl_get_priority_pfc(struct net_device *netdev,
+			     int priority, u8 *setting)
+{
+	struct ieee_pfc pfc;
+	int err;
+
+	err = mlx5e_dcbnl_ieee_getpfc(netdev, &pfc);
+
+	if (err)
+		*setting = 0;
+	else
+		*setting = (pfc.pfc_en >> priority) & 0x01;
+
+	return err;
+}
+
+static void mlx5e_dcbnl_getpfccfg(struct net_device *netdev,
+				  int priority, u8 *setting)
+{
+	if (priority >= CEE_DCBX_MAX_PRIO) {
+		netdev_err(netdev,
+			   "%s, priority is out of range\n", __func__);
+		return;
+	}
+
+	if (!setting)
+		return;
+
+	mlx5e_dcbnl_get_priority_pfc(netdev, priority, setting);
+}
+
+static u8 mlx5e_dcbnl_getcap(struct net_device *netdev,
+			     int capid, u8 *cap)
+{
+	struct mlx5e_priv *priv = netdev_priv(netdev);
+	struct mlx5_core_dev *mdev = priv->mdev;
+	u8 rval = 0;
+
+	switch (capid) {
+	case DCB_CAP_ATTR_PG:
+		*cap = true;
+		break;
+	case DCB_CAP_ATTR_PFC:
+		*cap = true;
+		break;
+	case DCB_CAP_ATTR_UP2TC:
+		*cap = false;
+		break;
+	case DCB_CAP_ATTR_PG_TCS:
+		*cap = 1 << mlx5_max_tc(mdev);
+		break;
+	case DCB_CAP_ATTR_PFC_TCS:
+		*cap = 1 << mlx5_max_tc(mdev);
+		break;
+	case DCB_CAP_ATTR_GSP:
+		*cap = false;
+		break;
+	case DCB_CAP_ATTR_BCN:
+		*cap = false;
+		break;
+	case DCB_CAP_ATTR_DCBX:
+		*cap = priv->dcbx.cap |
+		       DCB_CAP_DCBX_VER_CEE |
+		       DCB_CAP_DCBX_VER_IEEE;
+		break;
+	default:
+		*cap = 0;
+		rval = 1;
+		break;
+	}
+
+	return rval;
+}
+
+static int mlx5e_dcbnl_getnumtcs(struct net_device *netdev,
+				 int tcs_id, u8 *num)
+{
+	struct mlx5e_priv *priv = netdev_priv(netdev);
+	struct mlx5_core_dev *mdev = priv->mdev;
+
+	switch (tcs_id) {
+	case DCB_NUMTCS_ATTR_PG:
+	case DCB_NUMTCS_ATTR_PFC:
+		*num = mlx5_max_tc(mdev) + 1;
+		break;
+	default:
+		return -EINVAL;
+	}
+
+	return 0;
+}
+
+static u8 mlx5e_dcbnl_getpfcstate(struct net_device *netdev)
+{
+	struct ieee_pfc pfc;
+
+	if (mlx5e_dcbnl_ieee_getpfc(netdev, &pfc))
+		return MLX5E_CEE_STATE_DOWN;
+
+	return pfc.pfc_en ? MLX5E_CEE_STATE_UP : MLX5E_CEE_STATE_DOWN;
+}
+
+static void mlx5e_dcbnl_setpfcstate(struct net_device *netdev, u8 state)
+{
+	struct mlx5e_priv *priv = netdev_priv(netdev);
+	struct mlx5e_cee_config *cee_cfg = &priv->dcbx.cee_cfg;
+
+	if ((state != MLX5E_CEE_STATE_UP) && (state != MLX5E_CEE_STATE_DOWN))
+		return;
+
+	cee_cfg->pfc_enable = state;
+}
+
+static int mlx5e_dcbnl_getbuffer(struct net_device *dev,
+				 struct dcbnl_buffer *dcb_buffer)
+{
+	struct mlx5e_priv *priv = netdev_priv(dev);
+	struct mlx5_core_dev *mdev = priv->mdev;
+	struct mlx5e_port_buffer port_buffer;
+	u8 buffer[MLX5E_MAX_PRIORITY];
+	int i, err;
+
+	if (!MLX5_BUFFER_SUPPORTED(mdev))
+		return -EOPNOTSUPP;
+
+	err = mlx5e_port_query_priority2buffer(mdev, buffer);
+	if (err)
+		return err;
+
+	for (i = 0; i < MLX5E_MAX_PRIORITY; i++)
+		dcb_buffer->prio2buffer[i] = buffer[i];
+
+	err = mlx5e_port_query_buffer(priv, &port_buffer);
+	if (err)
+		return err;
+
+	for (i = 0; i < MLX5E_MAX_BUFFER; i++)
+		dcb_buffer->buffer_size[i] = port_buffer.buffer[i].size;
+	dcb_buffer->total_size = port_buffer.port_buffer_size;
+
+	return 0;
+}
+
+static int mlx5e_dcbnl_setbuffer(struct net_device *dev,
+				 struct dcbnl_buffer *dcb_buffer)
+{
+	struct mlx5e_priv *priv = netdev_priv(dev);
+	struct mlx5_core_dev *mdev = priv->mdev;
+	struct mlx5e_port_buffer port_buffer;
+	u8 old_prio2buffer[MLX5E_MAX_PRIORITY];
+	u32 *buffer_size = NULL;
+	u8 *prio2buffer = NULL;
+	u32 changed = 0;
+	int i, err;
+
+	if (!MLX5_BUFFER_SUPPORTED(mdev))
+		return -EOPNOTSUPP;
+
+	for (i = 0; i < DCBX_MAX_BUFFERS; i++)
+		mlx5_core_dbg(mdev, "buffer[%d]=%d\n", i, dcb_buffer->buffer_size[i]);
+
+	for (i = 0; i < MLX5E_MAX_PRIORITY; i++)
+		mlx5_core_dbg(mdev, "priority %d buffer%d\n", i, dcb_buffer->prio2buffer[i]);
+
+	err = mlx5e_port_query_priority2buffer(mdev, old_prio2buffer);
+	if (err)
+		return err;
+
+	for (i = 0; i < MLX5E_MAX_PRIORITY; i++) {
+		if (dcb_buffer->prio2buffer[i] != old_prio2buffer[i]) {
+			changed |= MLX5E_PORT_BUFFER_PRIO2BUFFER;
+			prio2buffer = dcb_buffer->prio2buffer;
+			break;
+		}
+	}
+
+	err = mlx5e_port_query_buffer(priv, &port_buffer);
+	if (err)
+		return err;
+
+	for (i = 0; i < MLX5E_MAX_BUFFER; i++) {
+		if (port_buffer.buffer[i].size != dcb_buffer->buffer_size[i]) {
+			changed |= MLX5E_PORT_BUFFER_SIZE;
+			buffer_size = dcb_buffer->buffer_size;
+			break;
+		}
+	}
+
+	if (!changed)
+		return 0;
+
+	priv->dcbx.manual_buffer = true;
+	err = mlx5e_port_manual_buffer_config(priv, changed, dev->mtu, NULL,
+					      buffer_size, prio2buffer);
+	return err;
+}
+
+const struct dcbnl_rtnl_ops mlx5e_dcbnl_ops = {
+	.ieee_getets	= mlx5e_dcbnl_ieee_getets,
+	.ieee_setets	= mlx5e_dcbnl_ieee_setets,
+	.ieee_getmaxrate = mlx5e_dcbnl_ieee_getmaxrate,
+	.ieee_setmaxrate = mlx5e_dcbnl_ieee_setmaxrate,
+	.ieee_getpfc	= mlx5e_dcbnl_ieee_getpfc,
+	.ieee_setpfc	= mlx5e_dcbnl_ieee_setpfc,
+	.ieee_setapp    = mlx5e_dcbnl_ieee_setapp,
+	.ieee_delapp    = mlx5e_dcbnl_ieee_delapp,
+	.getdcbx	= mlx5e_dcbnl_getdcbx,
+	.setdcbx	= mlx5e_dcbnl_setdcbx,
+	.dcbnl_getbuffer = mlx5e_dcbnl_getbuffer,
+	.dcbnl_setbuffer = mlx5e_dcbnl_setbuffer,
+
+/* CEE interfaces */
+	.setall         = mlx5e_dcbnl_setall,
+	.getstate       = mlx5e_dcbnl_getstate,
+	.getpermhwaddr  = mlx5e_dcbnl_getpermhwaddr,
+
+	.setpgtccfgtx   = mlx5e_dcbnl_setpgtccfgtx,
+	.setpgbwgcfgtx  = mlx5e_dcbnl_setpgbwgcfgtx,
+	.getpgtccfgtx   = mlx5e_dcbnl_getpgtccfgtx,
+	.getpgbwgcfgtx  = mlx5e_dcbnl_getpgbwgcfgtx,
+
+	.setpfccfg      = mlx5e_dcbnl_setpfccfg,
+	.getpfccfg      = mlx5e_dcbnl_getpfccfg,
+	.getcap         = mlx5e_dcbnl_getcap,
+	.getnumtcs      = mlx5e_dcbnl_getnumtcs,
+	.getpfcstate    = mlx5e_dcbnl_getpfcstate,
+	.setpfcstate    = mlx5e_dcbnl_setpfcstate,
+};
+
+static void mlx5e_dcbnl_query_dcbx_mode(struct mlx5e_priv *priv,
+					enum mlx5_dcbx_oper_mode *mode)
+{
+	u32 out[MLX5_ST_SZ_DW(dcbx_param)];
+
+	*mode = MLX5E_DCBX_PARAM_VER_OPER_HOST;
+
+	if (!mlx5_query_port_dcbx_param(priv->mdev, out))
+		*mode = MLX5_GET(dcbx_param, out, version_oper);
+
+	/* From driver's point of view, we only care if the mode
+	 * is host (HOST) or non-host (AUTO)
+	 */
+	if (*mode != MLX5E_DCBX_PARAM_VER_OPER_HOST)
+		*mode = MLX5E_DCBX_PARAM_VER_OPER_AUTO;
+}
+
+static void mlx5e_ets_init(struct mlx5e_priv *priv)
+{
+	struct ieee_ets ets;
+	int err;
+	int i;
+
+	if (!MLX5_CAP_GEN(priv->mdev, ets))
+		return;
+
+	memset(&ets, 0, sizeof(ets));
+	ets.ets_cap = mlx5_max_tc(priv->mdev) + 1;
+	for (i = 0; i < ets.ets_cap; i++) {
+		ets.tc_tx_bw[i] = MLX5E_MAX_BW_ALLOC;
+		ets.tc_tsa[i] = IEEE_8021QAZ_TSA_VENDOR;
+		ets.prio_tc[i] = i;
+	}
+
+	if (ets.ets_cap > 1) {
+		/* tclass[prio=0]=1, tclass[prio=1]=0, tclass[prio=i]=i (for i>1) */
+		ets.prio_tc[0] = 1;
+		ets.prio_tc[1] = 0;
+	}
+
+	err = mlx5e_dcbnl_ieee_setets_core(priv, &ets);
+	if (err)
+		netdev_err(priv->netdev,
+			   "%s, Failed to init ETS: %d\n", __func__, err);
+}
+
+enum {
+	INIT,
+	DELETE,
+};
+
+static void mlx5e_dcbnl_dscp_app(struct mlx5e_priv *priv, int action)
+{
+	struct dcb_app temp;
+	int i;
+
+	if (!MLX5_CAP_GEN(priv->mdev, vport_group_manager))
+		return;
+
+	if (!MLX5_DSCP_SUPPORTED(priv->mdev))
+		return;
+
+	/* No SEL_DSCP entry in non DSCP state */
+	if (priv->dcbx_dp.trust_state != MLX5_QPTS_TRUST_DSCP)
+		return;
+
+	temp.selector = IEEE_8021QAZ_APP_SEL_DSCP;
+	for (i = 0; i < MLX5E_MAX_DSCP; i++) {
+		temp.protocol = i;
+		temp.priority = priv->dcbx_dp.dscp2prio[i];
+		if (action == INIT)
+			dcb_ieee_setapp(priv->netdev, &temp);
+		else
+			dcb_ieee_delapp(priv->netdev, &temp);
+	}
+
+	priv->dcbx.dscp_app_cnt = (action == INIT) ? MLX5E_MAX_DSCP : 0;
+}
+
+void mlx5e_dcbnl_init_app(struct mlx5e_priv *priv)
+{
+	mlx5e_dcbnl_dscp_app(priv, INIT);
+}
+
+void mlx5e_dcbnl_delete_app(struct mlx5e_priv *priv)
+{
+	mlx5e_dcbnl_dscp_app(priv, DELETE);
+}
+
+static void mlx5e_trust_update_tx_min_inline_mode(struct mlx5e_priv *priv,
+						  struct mlx5e_params *params)
+{
+	params->tx_min_inline_mode = mlx5e_params_calculate_tx_min_inline(priv->mdev);
+	if (priv->dcbx_dp.trust_state == MLX5_QPTS_TRUST_DSCP &&
+	    params->tx_min_inline_mode == MLX5_INLINE_MODE_L2)
+		params->tx_min_inline_mode = MLX5_INLINE_MODE_IP;
+}
+
+static void mlx5e_trust_update_sq_inline_mode(struct mlx5e_priv *priv)
+{
+	struct mlx5e_channels new_channels = {};
+
+	mutex_lock(&priv->state_lock);
+
+	new_channels.params = priv->channels.params;
+	mlx5e_trust_update_tx_min_inline_mode(priv, &new_channels.params);
+
+	if (!test_bit(MLX5E_STATE_OPENED, &priv->state)) {
+		priv->channels.params = new_channels.params;
+		goto out;
+	}
+
+	/* Skip if tx_min_inline is the same */
+	if (new_channels.params.tx_min_inline_mode ==
+	    priv->channels.params.tx_min_inline_mode)
+		goto out;
+
+	if (mlx5e_open_channels(priv, &new_channels))
+		goto out;
+	mlx5e_switch_priv_channels(priv, &new_channels, NULL);
+
+out:
+	mutex_unlock(&priv->state_lock);
+}
+
+static int mlx5e_set_trust_state(struct mlx5e_priv *priv, u8 trust_state)
+{
+	int err;
+
+	err = mlx5_set_trust_state(priv->mdev, trust_state);
+	if (err)
+		return err;
+	priv->dcbx_dp.trust_state = trust_state;
+	mlx5e_trust_update_sq_inline_mode(priv);
+
+	return err;
+}
+
+static int mlx5e_set_dscp2prio(struct mlx5e_priv *priv, u8 dscp, u8 prio)
+{
+	int err;
+
+	err = mlx5_set_dscp2prio(priv->mdev, dscp, prio);
+	if (err)
+		return err;
+
+	priv->dcbx_dp.dscp2prio[dscp] = prio;
+	return err;
+}
+
+static int mlx5e_trust_initialize(struct mlx5e_priv *priv)
+{
+	struct mlx5_core_dev *mdev = priv->mdev;
+	int err;
+
+	priv->dcbx_dp.trust_state = MLX5_QPTS_TRUST_PCP;
+
+	if (!MLX5_DSCP_SUPPORTED(mdev))
+		return 0;
+
+	err = mlx5_query_trust_state(priv->mdev, &priv->dcbx_dp.trust_state);
+	if (err)
+		return err;
+
+	mlx5e_trust_update_tx_min_inline_mode(priv, &priv->channels.params);
+
+	err = mlx5_query_dscp2prio(priv->mdev, priv->dcbx_dp.dscp2prio);
+	if (err)
+		return err;
+
+	return 0;
+}
+
+void mlx5e_dcbnl_initialize(struct mlx5e_priv *priv)
+{
+	struct mlx5e_dcbx *dcbx = &priv->dcbx;
+
+	mlx5e_trust_initialize(priv);
+
+	if (!MLX5_CAP_GEN(priv->mdev, qos))
+		return;
+
+	if (MLX5_CAP_GEN(priv->mdev, dcbx))
+		mlx5e_dcbnl_query_dcbx_mode(priv, &dcbx->mode);
+
+	priv->dcbx.cap = DCB_CAP_DCBX_VER_CEE |
+			 DCB_CAP_DCBX_VER_IEEE;
+	if (priv->dcbx.mode == MLX5E_DCBX_PARAM_VER_OPER_HOST)
+		priv->dcbx.cap |= DCB_CAP_DCBX_HOST;
+
+	priv->dcbx.manual_buffer = false;
+	priv->dcbx.cable_len = MLX5E_DEFAULT_CABLE_LEN;
+
+	mlx5e_ets_init(priv);
+}
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_dim.c b/drivers/net/ethernet/mellanox/mlx5/core/en_dim.c
new file mode 100644
index 000000000..d67adf70a
--- /dev/null
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_dim.c
@@ -0,0 +1,62 @@
+/*
+ * Copyright (c) 2016, 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/net_dim.h>
+#include "en.h"
+
+static void
+mlx5e_complete_dim_work(struct net_dim *dim, struct net_dim_cq_moder moder,
+			struct mlx5_core_dev *mdev, struct mlx5_core_cq *mcq)
+{
+	mlx5_core_modify_cq_moderation(mdev, mcq, moder.usec, moder.pkts);
+	dim->state = NET_DIM_START_MEASURE;
+}
+
+void mlx5e_rx_dim_work(struct work_struct *work)
+{
+	struct net_dim *dim = container_of(work, struct net_dim, work);
+	struct mlx5e_rq *rq = container_of(dim, struct mlx5e_rq, dim);
+	struct net_dim_cq_moder cur_moder =
+		net_dim_get_rx_moderation(dim->mode, dim->profile_ix);
+
+	mlx5e_complete_dim_work(dim, cur_moder, rq->mdev, &rq->cq.mcq);
+}
+
+void mlx5e_tx_dim_work(struct work_struct *work)
+{
+	struct net_dim *dim = container_of(work, struct net_dim, work);
+	struct mlx5e_txqsq *sq = container_of(dim, struct mlx5e_txqsq, dim);
+	struct net_dim_cq_moder cur_moder =
+		net_dim_get_tx_moderation(dim->mode, dim->profile_ix);
+
+	mlx5e_complete_dim_work(dim, cur_moder, sq->cq.mdev, &sq->cq.mcq);
+}
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_ethtool.c b/drivers/net/ethernet/mellanox/mlx5/core/en_ethtool.c
new file mode 100644
index 000000000..a5c4e4f0f
--- /dev/null
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_ethtool.c
@@ -0,0 +1,1701 @@
+/*
+ * Copyright (c) 2015, 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 "en.h"
+#include "en/port.h"
+#include "lib/clock.h"
+
+void mlx5e_ethtool_get_drvinfo(struct mlx5e_priv *priv,
+			       struct ethtool_drvinfo *drvinfo)
+{
+	struct mlx5_core_dev *mdev = priv->mdev;
+
+	strlcpy(drvinfo->driver, DRIVER_NAME, sizeof(drvinfo->driver));
+	strlcpy(drvinfo->version, DRIVER_VERSION,
+		sizeof(drvinfo->version));
+	snprintf(drvinfo->fw_version, sizeof(drvinfo->fw_version),
+		 "%d.%d.%04d (%.16s)",
+		 fw_rev_maj(mdev), fw_rev_min(mdev), fw_rev_sub(mdev),
+		 mdev->board_id);
+	strlcpy(drvinfo->bus_info, pci_name(mdev->pdev),
+		sizeof(drvinfo->bus_info));
+}
+
+static void mlx5e_get_drvinfo(struct net_device *dev,
+			      struct ethtool_drvinfo *drvinfo)
+{
+	struct mlx5e_priv *priv = netdev_priv(dev);
+
+	mlx5e_ethtool_get_drvinfo(priv, drvinfo);
+}
+
+struct ptys2ethtool_config {
+	__ETHTOOL_DECLARE_LINK_MODE_MASK(supported);
+	__ETHTOOL_DECLARE_LINK_MODE_MASK(advertised);
+};
+
+static struct ptys2ethtool_config ptys2ethtool_table[MLX5E_LINK_MODES_NUMBER];
+
+#define MLX5_BUILD_PTYS2ETHTOOL_CONFIG(reg_, ...)                       \
+	({                                                              \
+		struct ptys2ethtool_config *cfg;                        \
+		const unsigned int modes[] = { __VA_ARGS__ };           \
+		unsigned int i;                                         \
+		cfg = &ptys2ethtool_table[reg_];                        \
+		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);           \
+		}                                                       \
+	})
+
+void mlx5e_build_ptys2ethtool_map(void)
+{
+	MLX5_BUILD_PTYS2ETHTOOL_CONFIG(MLX5E_1000BASE_CX_SGMII,
+				       ETHTOOL_LINK_MODE_1000baseKX_Full_BIT);
+	MLX5_BUILD_PTYS2ETHTOOL_CONFIG(MLX5E_1000BASE_KX,
+				       ETHTOOL_LINK_MODE_1000baseKX_Full_BIT);
+	MLX5_BUILD_PTYS2ETHTOOL_CONFIG(MLX5E_10GBASE_CX4,
+				       ETHTOOL_LINK_MODE_10000baseKX4_Full_BIT);
+	MLX5_BUILD_PTYS2ETHTOOL_CONFIG(MLX5E_10GBASE_KX4,
+				       ETHTOOL_LINK_MODE_10000baseKX4_Full_BIT);
+	MLX5_BUILD_PTYS2ETHTOOL_CONFIG(MLX5E_10GBASE_KR,
+				       ETHTOOL_LINK_MODE_10000baseKR_Full_BIT);
+	MLX5_BUILD_PTYS2ETHTOOL_CONFIG(MLX5E_20GBASE_KR2,
+				       ETHTOOL_LINK_MODE_20000baseKR2_Full_BIT);
+	MLX5_BUILD_PTYS2ETHTOOL_CONFIG(MLX5E_40GBASE_CR4,
+				       ETHTOOL_LINK_MODE_40000baseCR4_Full_BIT);
+	MLX5_BUILD_PTYS2ETHTOOL_CONFIG(MLX5E_40GBASE_KR4,
+				       ETHTOOL_LINK_MODE_40000baseKR4_Full_BIT);
+	MLX5_BUILD_PTYS2ETHTOOL_CONFIG(MLX5E_56GBASE_R4,
+				       ETHTOOL_LINK_MODE_56000baseKR4_Full_BIT);
+	MLX5_BUILD_PTYS2ETHTOOL_CONFIG(MLX5E_10GBASE_CR,
+				       ETHTOOL_LINK_MODE_10000baseKR_Full_BIT);
+	MLX5_BUILD_PTYS2ETHTOOL_CONFIG(MLX5E_10GBASE_SR,
+				       ETHTOOL_LINK_MODE_10000baseKR_Full_BIT);
+	MLX5_BUILD_PTYS2ETHTOOL_CONFIG(MLX5E_10GBASE_ER,
+				       ETHTOOL_LINK_MODE_10000baseKR_Full_BIT);
+	MLX5_BUILD_PTYS2ETHTOOL_CONFIG(MLX5E_40GBASE_SR4,
+				       ETHTOOL_LINK_MODE_40000baseSR4_Full_BIT);
+	MLX5_BUILD_PTYS2ETHTOOL_CONFIG(MLX5E_40GBASE_LR4,
+				       ETHTOOL_LINK_MODE_40000baseLR4_Full_BIT);
+	MLX5_BUILD_PTYS2ETHTOOL_CONFIG(MLX5E_50GBASE_SR2,
+				       ETHTOOL_LINK_MODE_50000baseSR2_Full_BIT);
+	MLX5_BUILD_PTYS2ETHTOOL_CONFIG(MLX5E_100GBASE_CR4,
+				       ETHTOOL_LINK_MODE_100000baseCR4_Full_BIT);
+	MLX5_BUILD_PTYS2ETHTOOL_CONFIG(MLX5E_100GBASE_SR4,
+				       ETHTOOL_LINK_MODE_100000baseSR4_Full_BIT);
+	MLX5_BUILD_PTYS2ETHTOOL_CONFIG(MLX5E_100GBASE_KR4,
+				       ETHTOOL_LINK_MODE_100000baseKR4_Full_BIT);
+	MLX5_BUILD_PTYS2ETHTOOL_CONFIG(MLX5E_100GBASE_LR4,
+				       ETHTOOL_LINK_MODE_100000baseLR4_ER4_Full_BIT);
+	MLX5_BUILD_PTYS2ETHTOOL_CONFIG(MLX5E_10GBASE_T,
+				       ETHTOOL_LINK_MODE_10000baseT_Full_BIT);
+	MLX5_BUILD_PTYS2ETHTOOL_CONFIG(MLX5E_25GBASE_CR,
+				       ETHTOOL_LINK_MODE_25000baseCR_Full_BIT);
+	MLX5_BUILD_PTYS2ETHTOOL_CONFIG(MLX5E_25GBASE_KR,
+				       ETHTOOL_LINK_MODE_25000baseKR_Full_BIT);
+	MLX5_BUILD_PTYS2ETHTOOL_CONFIG(MLX5E_25GBASE_SR,
+				       ETHTOOL_LINK_MODE_25000baseSR_Full_BIT);
+	MLX5_BUILD_PTYS2ETHTOOL_CONFIG(MLX5E_50GBASE_CR2,
+				       ETHTOOL_LINK_MODE_50000baseCR2_Full_BIT);
+	MLX5_BUILD_PTYS2ETHTOOL_CONFIG(MLX5E_50GBASE_KR2,
+				       ETHTOOL_LINK_MODE_50000baseKR2_Full_BIT);
+}
+
+int mlx5e_ethtool_get_sset_count(struct mlx5e_priv *priv, int sset)
+{
+	int i, num_stats = 0;
+
+	switch (sset) {
+	case ETH_SS_STATS:
+		for (i = 0; i < mlx5e_num_stats_grps; i++)
+			num_stats += mlx5e_stats_grps[i].get_num_stats(priv);
+		return num_stats;
+	case ETH_SS_PRIV_FLAGS:
+		return ARRAY_SIZE(mlx5e_priv_flags);
+	case ETH_SS_TEST:
+		return mlx5e_self_test_num(priv);
+	/* fallthrough */
+	default:
+		return -EOPNOTSUPP;
+	}
+}
+
+static int mlx5e_get_sset_count(struct net_device *dev, int sset)
+{
+	struct mlx5e_priv *priv = netdev_priv(dev);
+
+	return mlx5e_ethtool_get_sset_count(priv, sset);
+}
+
+static void mlx5e_fill_stats_strings(struct mlx5e_priv *priv, u8 *data)
+{
+	int i, idx = 0;
+
+	for (i = 0; i < mlx5e_num_stats_grps; i++)
+		idx = mlx5e_stats_grps[i].fill_strings(priv, data, idx);
+}
+
+void mlx5e_ethtool_get_strings(struct mlx5e_priv *priv, u32 stringset, u8 *data)
+{
+	int i;
+
+	switch (stringset) {
+	case ETH_SS_PRIV_FLAGS:
+		for (i = 0; i < ARRAY_SIZE(mlx5e_priv_flags); i++)
+			strcpy(data + i * ETH_GSTRING_LEN, mlx5e_priv_flags[i]);
+		break;
+
+	case ETH_SS_TEST:
+		for (i = 0; i < mlx5e_self_test_num(priv); i++)
+			strcpy(data + i * ETH_GSTRING_LEN,
+			       mlx5e_self_tests[i]);
+		break;
+
+	case ETH_SS_STATS:
+		mlx5e_fill_stats_strings(priv, data);
+		break;
+	}
+}
+
+static void mlx5e_get_strings(struct net_device *dev, u32 stringset, u8 *data)
+{
+	struct mlx5e_priv *priv = netdev_priv(dev);
+
+	mlx5e_ethtool_get_strings(priv, stringset, data);
+}
+
+void mlx5e_ethtool_get_ethtool_stats(struct mlx5e_priv *priv,
+				     struct ethtool_stats *stats, u64 *data)
+{
+	int i, idx = 0;
+
+	mutex_lock(&priv->state_lock);
+	mlx5e_update_stats(priv);
+	mutex_unlock(&priv->state_lock);
+
+	for (i = 0; i < mlx5e_num_stats_grps; i++)
+		idx = mlx5e_stats_grps[i].fill_stats(priv, data, idx);
+}
+
+static void mlx5e_get_ethtool_stats(struct net_device *dev,
+				    struct ethtool_stats *stats,
+				    u64 *data)
+{
+	struct mlx5e_priv *priv = netdev_priv(dev);
+
+	mlx5e_ethtool_get_ethtool_stats(priv, stats, data);
+}
+
+void mlx5e_ethtool_get_ringparam(struct mlx5e_priv *priv,
+				 struct ethtool_ringparam *param)
+{
+	param->rx_max_pending = 1 << MLX5E_PARAMS_MAXIMUM_LOG_RQ_SIZE;
+	param->tx_max_pending = 1 << MLX5E_PARAMS_MAXIMUM_LOG_SQ_SIZE;
+	param->rx_pending     = 1 << priv->channels.params.log_rq_mtu_frames;
+	param->tx_pending     = 1 << priv->channels.params.log_sq_size;
+}
+
+static void mlx5e_get_ringparam(struct net_device *dev,
+				struct ethtool_ringparam *param)
+{
+	struct mlx5e_priv *priv = netdev_priv(dev);
+
+	mlx5e_ethtool_get_ringparam(priv, param);
+}
+
+int mlx5e_ethtool_set_ringparam(struct mlx5e_priv *priv,
+				struct ethtool_ringparam *param)
+{
+	struct mlx5e_channels new_channels = {};
+	u8 log_rq_size;
+	u8 log_sq_size;
+	int err = 0;
+
+	if (param->rx_jumbo_pending) {
+		netdev_info(priv->netdev, "%s: rx_jumbo_pending not supported\n",
+			    __func__);
+		return -EINVAL;
+	}
+	if (param->rx_mini_pending) {
+		netdev_info(priv->netdev, "%s: rx_mini_pending not supported\n",
+			    __func__);
+		return -EINVAL;
+	}
+
+	if (param->rx_pending < (1 << MLX5E_PARAMS_MINIMUM_LOG_RQ_SIZE)) {
+		netdev_info(priv->netdev, "%s: rx_pending (%d) < min (%d)\n",
+			    __func__, param->rx_pending,
+			    1 << MLX5E_PARAMS_MINIMUM_LOG_RQ_SIZE);
+		return -EINVAL;
+	}
+
+	if (param->tx_pending < (1 << MLX5E_PARAMS_MINIMUM_LOG_SQ_SIZE)) {
+		netdev_info(priv->netdev, "%s: tx_pending (%d) < min (%d)\n",
+			    __func__, param->tx_pending,
+			    1 << MLX5E_PARAMS_MINIMUM_LOG_SQ_SIZE);
+		return -EINVAL;
+	}
+
+	log_rq_size = order_base_2(param->rx_pending);
+	log_sq_size = order_base_2(param->tx_pending);
+
+	if (log_rq_size == priv->channels.params.log_rq_mtu_frames &&
+	    log_sq_size == priv->channels.params.log_sq_size)
+		return 0;
+
+	mutex_lock(&priv->state_lock);
+
+	new_channels.params = priv->channels.params;
+	new_channels.params.log_rq_mtu_frames = log_rq_size;
+	new_channels.params.log_sq_size = log_sq_size;
+
+	if (!test_bit(MLX5E_STATE_OPENED, &priv->state)) {
+		priv->channels.params = new_channels.params;
+		goto unlock;
+	}
+
+	err = mlx5e_open_channels(priv, &new_channels);
+	if (err)
+		goto unlock;
+
+	mlx5e_switch_priv_channels(priv, &new_channels, NULL);
+
+unlock:
+	mutex_unlock(&priv->state_lock);
+
+	return err;
+}
+
+static int mlx5e_set_ringparam(struct net_device *dev,
+			       struct ethtool_ringparam *param)
+{
+	struct mlx5e_priv *priv = netdev_priv(dev);
+
+	return mlx5e_ethtool_set_ringparam(priv, param);
+}
+
+void mlx5e_ethtool_get_channels(struct mlx5e_priv *priv,
+				struct ethtool_channels *ch)
+{
+	ch->max_combined   = priv->profile->max_nch(priv->mdev);
+	ch->combined_count = priv->channels.params.num_channels;
+}
+
+static void mlx5e_get_channels(struct net_device *dev,
+			       struct ethtool_channels *ch)
+{
+	struct mlx5e_priv *priv = netdev_priv(dev);
+
+	mlx5e_ethtool_get_channels(priv, ch);
+}
+
+int mlx5e_ethtool_set_channels(struct mlx5e_priv *priv,
+			       struct ethtool_channels *ch)
+{
+	unsigned int count = ch->combined_count;
+	struct mlx5e_channels new_channels = {};
+	bool arfs_enabled;
+	int err = 0;
+
+	if (!count) {
+		netdev_info(priv->netdev, "%s: combined_count=0 not supported\n",
+			    __func__);
+		return -EINVAL;
+	}
+
+	if (priv->channels.params.num_channels == count)
+		return 0;
+
+	mutex_lock(&priv->state_lock);
+
+	new_channels.params = priv->channels.params;
+	new_channels.params.num_channels = count;
+	if (!netif_is_rxfh_configured(priv->netdev))
+		mlx5e_build_default_indir_rqt(new_channels.params.indirection_rqt,
+					      MLX5E_INDIR_RQT_SIZE, count);
+
+	if (!test_bit(MLX5E_STATE_OPENED, &priv->state)) {
+		priv->channels.params = new_channels.params;
+		goto out;
+	}
+
+	/* Create fresh channels with new parameters */
+	err = mlx5e_open_channels(priv, &new_channels);
+	if (err)
+		goto out;
+
+	arfs_enabled = priv->netdev->features & NETIF_F_NTUPLE;
+	if (arfs_enabled)
+		mlx5e_arfs_disable(priv);
+
+	/* Switch to new channels, set new parameters and close old ones */
+	mlx5e_switch_priv_channels(priv, &new_channels, NULL);
+
+	if (arfs_enabled) {
+		err = mlx5e_arfs_enable(priv);
+		if (err)
+			netdev_err(priv->netdev, "%s: mlx5e_arfs_enable failed: %d\n",
+				   __func__, err);
+	}
+
+out:
+	mutex_unlock(&priv->state_lock);
+
+	return err;
+}
+
+static int mlx5e_set_channels(struct net_device *dev,
+			      struct ethtool_channels *ch)
+{
+	struct mlx5e_priv *priv = netdev_priv(dev);
+
+	return mlx5e_ethtool_set_channels(priv, ch);
+}
+
+int mlx5e_ethtool_get_coalesce(struct mlx5e_priv *priv,
+			       struct ethtool_coalesce *coal)
+{
+	struct net_dim_cq_moder *rx_moder, *tx_moder;
+
+	if (!MLX5_CAP_GEN(priv->mdev, cq_moderation))
+		return -EOPNOTSUPP;
+
+	rx_moder = &priv->channels.params.rx_cq_moderation;
+	coal->rx_coalesce_usecs		= rx_moder->usec;
+	coal->rx_max_coalesced_frames	= rx_moder->pkts;
+	coal->use_adaptive_rx_coalesce	= priv->channels.params.rx_dim_enabled;
+
+	tx_moder = &priv->channels.params.tx_cq_moderation;
+	coal->tx_coalesce_usecs		= tx_moder->usec;
+	coal->tx_max_coalesced_frames	= tx_moder->pkts;
+	coal->use_adaptive_tx_coalesce	= priv->channels.params.tx_dim_enabled;
+
+	return 0;
+}
+
+static int mlx5e_get_coalesce(struct net_device *netdev,
+			      struct ethtool_coalesce *coal)
+{
+	struct mlx5e_priv *priv = netdev_priv(netdev);
+
+	return mlx5e_ethtool_get_coalesce(priv, coal);
+}
+
+#define MLX5E_MAX_COAL_TIME		MLX5_MAX_CQ_PERIOD
+#define MLX5E_MAX_COAL_FRAMES		MLX5_MAX_CQ_COUNT
+
+static void
+mlx5e_set_priv_channels_coalesce(struct mlx5e_priv *priv, struct ethtool_coalesce *coal)
+{
+	struct mlx5_core_dev *mdev = priv->mdev;
+	int tc;
+	int i;
+
+	for (i = 0; i < priv->channels.num; ++i) {
+		struct mlx5e_channel *c = priv->channels.c[i];
+
+		for (tc = 0; tc < c->num_tc; tc++) {
+			mlx5_core_modify_cq_moderation(mdev,
+						&c->sq[tc].cq.mcq,
+						coal->tx_coalesce_usecs,
+						coal->tx_max_coalesced_frames);
+		}
+
+		mlx5_core_modify_cq_moderation(mdev, &c->rq.cq.mcq,
+					       coal->rx_coalesce_usecs,
+					       coal->rx_max_coalesced_frames);
+	}
+}
+
+int mlx5e_ethtool_set_coalesce(struct mlx5e_priv *priv,
+			       struct ethtool_coalesce *coal)
+{
+	struct net_dim_cq_moder *rx_moder, *tx_moder;
+	struct mlx5_core_dev *mdev = priv->mdev;
+	struct mlx5e_channels new_channels = {};
+	int err = 0;
+	bool reset;
+
+	if (!MLX5_CAP_GEN(mdev, cq_moderation))
+		return -EOPNOTSUPP;
+
+	if (coal->tx_coalesce_usecs > MLX5E_MAX_COAL_TIME ||
+	    coal->rx_coalesce_usecs > MLX5E_MAX_COAL_TIME) {
+		netdev_info(priv->netdev, "%s: maximum coalesce time supported is %lu usecs\n",
+			    __func__, MLX5E_MAX_COAL_TIME);
+		return -ERANGE;
+	}
+
+	if (coal->tx_max_coalesced_frames > MLX5E_MAX_COAL_FRAMES ||
+	    coal->rx_max_coalesced_frames > MLX5E_MAX_COAL_FRAMES) {
+		netdev_info(priv->netdev, "%s: maximum coalesced frames supported is %lu\n",
+			    __func__, MLX5E_MAX_COAL_FRAMES);
+		return -ERANGE;
+	}
+
+	mutex_lock(&priv->state_lock);
+	new_channels.params = priv->channels.params;
+
+	rx_moder          = &new_channels.params.rx_cq_moderation;
+	rx_moder->usec    = coal->rx_coalesce_usecs;
+	rx_moder->pkts    = coal->rx_max_coalesced_frames;
+	new_channels.params.rx_dim_enabled = !!coal->use_adaptive_rx_coalesce;
+
+	tx_moder          = &new_channels.params.tx_cq_moderation;
+	tx_moder->usec    = coal->tx_coalesce_usecs;
+	tx_moder->pkts    = coal->tx_max_coalesced_frames;
+	new_channels.params.tx_dim_enabled = !!coal->use_adaptive_tx_coalesce;
+
+	if (!test_bit(MLX5E_STATE_OPENED, &priv->state)) {
+		priv->channels.params = new_channels.params;
+		goto out;
+	}
+	/* we are opened */
+
+	reset = (!!coal->use_adaptive_rx_coalesce != priv->channels.params.rx_dim_enabled) ||
+		(!!coal->use_adaptive_tx_coalesce != priv->channels.params.tx_dim_enabled);
+
+	if (!reset) {
+		mlx5e_set_priv_channels_coalesce(priv, coal);
+		priv->channels.params = new_channels.params;
+		goto out;
+	}
+
+	/* open fresh channels with new coal parameters */
+	err = mlx5e_open_channels(priv, &new_channels);
+	if (err)
+		goto out;
+
+	mlx5e_switch_priv_channels(priv, &new_channels, NULL);
+
+out:
+	mutex_unlock(&priv->state_lock);
+	return err;
+}
+
+static int mlx5e_set_coalesce(struct net_device *netdev,
+			      struct ethtool_coalesce *coal)
+{
+	struct mlx5e_priv *priv    = netdev_priv(netdev);
+
+	return mlx5e_ethtool_set_coalesce(priv, coal);
+}
+
+static void ptys2ethtool_supported_link(unsigned long *supported_modes,
+					u32 eth_proto_cap)
+{
+	unsigned long proto_cap = eth_proto_cap;
+	int proto;
+
+	for_each_set_bit(proto, &proto_cap, MLX5E_LINK_MODES_NUMBER)
+		bitmap_or(supported_modes, supported_modes,
+			  ptys2ethtool_table[proto].supported,
+			  __ETHTOOL_LINK_MODE_MASK_NBITS);
+}
+
+static void ptys2ethtool_adver_link(unsigned long *advertising_modes,
+				    u32 eth_proto_cap)
+{
+	unsigned long proto_cap = eth_proto_cap;
+	int proto;
+
+	for_each_set_bit(proto, &proto_cap, MLX5E_LINK_MODES_NUMBER)
+		bitmap_or(advertising_modes, advertising_modes,
+			  ptys2ethtool_table[proto].advertised,
+			  __ETHTOOL_LINK_MODE_MASK_NBITS);
+}
+
+static void ptys2ethtool_supported_advertised_port(struct ethtool_link_ksettings *link_ksettings,
+						   u32 eth_proto_cap,
+						   u8 connector_type)
+{
+	if (!connector_type || connector_type >= MLX5E_CONNECTOR_TYPE_NUMBER) {
+		if (eth_proto_cap & (MLX5E_PROT_MASK(MLX5E_10GBASE_CR)
+				   | MLX5E_PROT_MASK(MLX5E_10GBASE_SR)
+				   | MLX5E_PROT_MASK(MLX5E_40GBASE_CR4)
+				   | MLX5E_PROT_MASK(MLX5E_40GBASE_SR4)
+				   | MLX5E_PROT_MASK(MLX5E_100GBASE_SR4)
+				   | MLX5E_PROT_MASK(MLX5E_1000BASE_CX_SGMII))) {
+			ethtool_link_ksettings_add_link_mode(link_ksettings,
+							     supported,
+							     FIBRE);
+			ethtool_link_ksettings_add_link_mode(link_ksettings,
+							     advertising,
+							     FIBRE);
+		}
+
+		if (eth_proto_cap & (MLX5E_PROT_MASK(MLX5E_100GBASE_KR4)
+				   | MLX5E_PROT_MASK(MLX5E_40GBASE_KR4)
+				   | MLX5E_PROT_MASK(MLX5E_10GBASE_KR)
+				   | MLX5E_PROT_MASK(MLX5E_10GBASE_KX4)
+				   | MLX5E_PROT_MASK(MLX5E_1000BASE_KX))) {
+			ethtool_link_ksettings_add_link_mode(link_ksettings,
+							     supported,
+							     Backplane);
+			ethtool_link_ksettings_add_link_mode(link_ksettings,
+							     advertising,
+							     Backplane);
+		}
+		return;
+	}
+
+	switch (connector_type) {
+	case MLX5E_PORT_TP:
+		ethtool_link_ksettings_add_link_mode(link_ksettings,
+						     supported, TP);
+		ethtool_link_ksettings_add_link_mode(link_ksettings,
+						     advertising, TP);
+		break;
+	case MLX5E_PORT_AUI:
+		ethtool_link_ksettings_add_link_mode(link_ksettings,
+						     supported, AUI);
+		ethtool_link_ksettings_add_link_mode(link_ksettings,
+						     advertising, AUI);
+		break;
+	case MLX5E_PORT_BNC:
+		ethtool_link_ksettings_add_link_mode(link_ksettings,
+						     supported, BNC);
+		ethtool_link_ksettings_add_link_mode(link_ksettings,
+						     advertising, BNC);
+		break;
+	case MLX5E_PORT_MII:
+		ethtool_link_ksettings_add_link_mode(link_ksettings,
+						     supported, MII);
+		ethtool_link_ksettings_add_link_mode(link_ksettings,
+						     advertising, MII);
+		break;
+	case MLX5E_PORT_FIBRE:
+		ethtool_link_ksettings_add_link_mode(link_ksettings,
+						     supported, FIBRE);
+		ethtool_link_ksettings_add_link_mode(link_ksettings,
+						     advertising, FIBRE);
+		break;
+	case MLX5E_PORT_DA:
+		ethtool_link_ksettings_add_link_mode(link_ksettings,
+						     supported, Backplane);
+		ethtool_link_ksettings_add_link_mode(link_ksettings,
+						     advertising, Backplane);
+		break;
+	case MLX5E_PORT_NONE:
+	case MLX5E_PORT_OTHER:
+	default:
+		break;
+	}
+}
+
+static void get_speed_duplex(struct net_device *netdev,
+			     u32 eth_proto_oper,
+			     struct ethtool_link_ksettings *link_ksettings)
+{
+	u32 speed = SPEED_UNKNOWN;
+	u8 duplex = DUPLEX_UNKNOWN;
+
+	if (!netif_carrier_ok(netdev))
+		goto out;
+
+	speed = mlx5e_port_ptys2speed(eth_proto_oper);
+	if (!speed) {
+		speed = SPEED_UNKNOWN;
+		goto out;
+	}
+
+	duplex = DUPLEX_FULL;
+
+out:
+	link_ksettings->base.speed = speed;
+	link_ksettings->base.duplex = duplex;
+}
+
+static void get_supported(u32 eth_proto_cap,
+			  struct ethtool_link_ksettings *link_ksettings)
+{
+	unsigned long *supported = link_ksettings->link_modes.supported;
+
+	ptys2ethtool_supported_link(supported, eth_proto_cap);
+	ethtool_link_ksettings_add_link_mode(link_ksettings, supported, Pause);
+}
+
+static void get_advertising(u32 eth_proto_cap, u8 tx_pause,
+			    u8 rx_pause,
+			    struct ethtool_link_ksettings *link_ksettings)
+{
+	unsigned long *advertising = link_ksettings->link_modes.advertising;
+
+	ptys2ethtool_adver_link(advertising, eth_proto_cap);
+	if (rx_pause)
+		ethtool_link_ksettings_add_link_mode(link_ksettings, advertising, Pause);
+	if (tx_pause ^ rx_pause)
+		ethtool_link_ksettings_add_link_mode(link_ksettings, advertising, Asym_Pause);
+}
+
+static int ptys2connector_type[MLX5E_CONNECTOR_TYPE_NUMBER] = {
+		[MLX5E_PORT_UNKNOWN]            = PORT_OTHER,
+		[MLX5E_PORT_NONE]               = PORT_NONE,
+		[MLX5E_PORT_TP]                 = PORT_TP,
+		[MLX5E_PORT_AUI]                = PORT_AUI,
+		[MLX5E_PORT_BNC]                = PORT_BNC,
+		[MLX5E_PORT_MII]                = PORT_MII,
+		[MLX5E_PORT_FIBRE]              = PORT_FIBRE,
+		[MLX5E_PORT_DA]                 = PORT_DA,
+		[MLX5E_PORT_OTHER]              = PORT_OTHER,
+	};
+
+static u8 get_connector_port(u32 eth_proto, u8 connector_type)
+{
+	if (connector_type && connector_type < MLX5E_CONNECTOR_TYPE_NUMBER)
+		return ptys2connector_type[connector_type];
+
+	if (eth_proto &
+	    (MLX5E_PROT_MASK(MLX5E_10GBASE_SR)   |
+	     MLX5E_PROT_MASK(MLX5E_40GBASE_SR4)  |
+	     MLX5E_PROT_MASK(MLX5E_100GBASE_SR4) |
+	     MLX5E_PROT_MASK(MLX5E_1000BASE_CX_SGMII))) {
+		return PORT_FIBRE;
+	}
+
+	if (eth_proto &
+	    (MLX5E_PROT_MASK(MLX5E_40GBASE_CR4) |
+	     MLX5E_PROT_MASK(MLX5E_10GBASE_CR)  |
+	     MLX5E_PROT_MASK(MLX5E_100GBASE_CR4))) {
+		return PORT_DA;
+	}
+
+	if (eth_proto &
+	    (MLX5E_PROT_MASK(MLX5E_10GBASE_KX4) |
+	     MLX5E_PROT_MASK(MLX5E_10GBASE_KR)  |
+	     MLX5E_PROT_MASK(MLX5E_40GBASE_KR4) |
+	     MLX5E_PROT_MASK(MLX5E_100GBASE_KR4))) {
+		return PORT_NONE;
+	}
+
+	return PORT_OTHER;
+}
+
+static void get_lp_advertising(u32 eth_proto_lp,
+			       struct ethtool_link_ksettings *link_ksettings)
+{
+	unsigned long *lp_advertising = link_ksettings->link_modes.lp_advertising;
+
+	ptys2ethtool_adver_link(lp_advertising, eth_proto_lp);
+}
+
+static int mlx5e_get_link_ksettings(struct net_device *netdev,
+				    struct ethtool_link_ksettings *link_ksettings)
+{
+	struct mlx5e_priv *priv    = netdev_priv(netdev);
+	struct mlx5_core_dev *mdev = priv->mdev;
+	u32 out[MLX5_ST_SZ_DW(ptys_reg)] = {0};
+	u32 rx_pause = 0;
+	u32 tx_pause = 0;
+	u32 eth_proto_cap;
+	u32 eth_proto_admin;
+	u32 eth_proto_lp;
+	u32 eth_proto_oper;
+	u8 an_disable_admin;
+	u8 an_status;
+	u8 connector_type;
+	int err;
+
+	err = mlx5_query_port_ptys(mdev, out, sizeof(out), MLX5_PTYS_EN, 1);
+	if (err) {
+		netdev_err(netdev, "%s: query port ptys failed: %d\n",
+			   __func__, err);
+		goto err_query_ptys;
+	}
+
+	eth_proto_cap    = MLX5_GET(ptys_reg, out, eth_proto_capability);
+	eth_proto_admin  = MLX5_GET(ptys_reg, out, eth_proto_admin);
+	eth_proto_oper   = MLX5_GET(ptys_reg, out, eth_proto_oper);
+	eth_proto_lp     = MLX5_GET(ptys_reg, out, eth_proto_lp_advertise);
+	an_disable_admin = MLX5_GET(ptys_reg, out, an_disable_admin);
+	an_status        = MLX5_GET(ptys_reg, out, an_status);
+	connector_type   = MLX5_GET(ptys_reg, out, connector_type);
+
+	mlx5_query_port_pause(mdev, &rx_pause, &tx_pause);
+
+	ethtool_link_ksettings_zero_link_mode(link_ksettings, supported);
+	ethtool_link_ksettings_zero_link_mode(link_ksettings, advertising);
+
+	get_supported(eth_proto_cap, link_ksettings);
+	get_advertising(eth_proto_admin, tx_pause, rx_pause, link_ksettings);
+	get_speed_duplex(netdev, eth_proto_oper, link_ksettings);
+
+	eth_proto_oper = eth_proto_oper ? eth_proto_oper : eth_proto_cap;
+
+	link_ksettings->base.port = get_connector_port(eth_proto_oper,
+						       connector_type);
+	ptys2ethtool_supported_advertised_port(link_ksettings, eth_proto_admin,
+					       connector_type);
+	get_lp_advertising(eth_proto_lp, link_ksettings);
+
+	if (an_status == MLX5_AN_COMPLETE)
+		ethtool_link_ksettings_add_link_mode(link_ksettings,
+						     lp_advertising, Autoneg);
+
+	link_ksettings->base.autoneg = an_disable_admin ? AUTONEG_DISABLE :
+							  AUTONEG_ENABLE;
+	ethtool_link_ksettings_add_link_mode(link_ksettings, supported,
+					     Autoneg);
+	if (!an_disable_admin)
+		ethtool_link_ksettings_add_link_mode(link_ksettings,
+						     advertising, Autoneg);
+
+err_query_ptys:
+	return err;
+}
+
+static u32 mlx5e_ethtool2ptys_adver_link(const unsigned long *link_modes)
+{
+	u32 i, ptys_modes = 0;
+
+	for (i = 0; i < MLX5E_LINK_MODES_NUMBER; ++i) {
+		if (bitmap_intersects(ptys2ethtool_table[i].advertised,
+				      link_modes,
+				      __ETHTOOL_LINK_MODE_MASK_NBITS))
+			ptys_modes |= MLX5E_PROT_MASK(i);
+	}
+
+	return ptys_modes;
+}
+
+static int mlx5e_set_link_ksettings(struct net_device *netdev,
+				    const struct ethtool_link_ksettings *link_ksettings)
+{
+	struct mlx5e_priv *priv    = netdev_priv(netdev);
+	struct mlx5_core_dev *mdev = priv->mdev;
+	u32 eth_proto_cap, eth_proto_admin;
+	bool an_changes = false;
+	u8 an_disable_admin;
+	u8 an_disable_cap;
+	bool an_disable;
+	u32 link_modes;
+	u8 an_status;
+	u32 speed;
+	int err;
+
+	speed = link_ksettings->base.speed;
+
+	link_modes = link_ksettings->base.autoneg == AUTONEG_ENABLE ?
+		mlx5e_ethtool2ptys_adver_link(link_ksettings->link_modes.advertising) :
+		mlx5e_port_speed2linkmodes(speed);
+
+	err = mlx5_query_port_proto_cap(mdev, &eth_proto_cap, MLX5_PTYS_EN);
+	if (err) {
+		netdev_err(netdev, "%s: query port eth proto cap failed: %d\n",
+			   __func__, err);
+		goto out;
+	}
+
+	link_modes = link_modes & eth_proto_cap;
+	if (!link_modes) {
+		netdev_err(netdev, "%s: Not supported link mode(s) requested",
+			   __func__);
+		err = -EINVAL;
+		goto out;
+	}
+
+	err = mlx5_query_port_proto_admin(mdev, &eth_proto_admin, MLX5_PTYS_EN);
+	if (err) {
+		netdev_err(netdev, "%s: query port eth proto admin failed: %d\n",
+			   __func__, err);
+		goto out;
+	}
+
+	mlx5_query_port_autoneg(mdev, MLX5_PTYS_EN, &an_status,
+				&an_disable_cap, &an_disable_admin);
+
+	an_disable = link_ksettings->base.autoneg == AUTONEG_DISABLE;
+	an_changes = ((!an_disable && an_disable_admin) ||
+		      (an_disable && !an_disable_admin));
+
+	if (!an_changes && link_modes == eth_proto_admin)
+		goto out;
+
+	mlx5_set_port_ptys(mdev, an_disable, link_modes, MLX5_PTYS_EN);
+	mlx5_toggle_port_link(mdev);
+
+out:
+	return err;
+}
+
+static u32 mlx5e_get_rxfh_key_size(struct net_device *netdev)
+{
+	struct mlx5e_priv *priv = netdev_priv(netdev);
+
+	return sizeof(priv->channels.params.toeplitz_hash_key);
+}
+
+static u32 mlx5e_get_rxfh_indir_size(struct net_device *netdev)
+{
+	return MLX5E_INDIR_RQT_SIZE;
+}
+
+static int mlx5e_get_rxfh(struct net_device *netdev, u32 *indir, u8 *key,
+			  u8 *hfunc)
+{
+	struct mlx5e_priv *priv = netdev_priv(netdev);
+
+	if (indir)
+		memcpy(indir, priv->channels.params.indirection_rqt,
+		       sizeof(priv->channels.params.indirection_rqt));
+
+	if (key)
+		memcpy(key, priv->channels.params.toeplitz_hash_key,
+		       sizeof(priv->channels.params.toeplitz_hash_key));
+
+	if (hfunc)
+		*hfunc = priv->channels.params.rss_hfunc;
+
+	return 0;
+}
+
+static void mlx5e_modify_tirs_hash(struct mlx5e_priv *priv, void *in, int inlen)
+{
+	void *tirc = MLX5_ADDR_OF(modify_tir_in, in, ctx);
+	struct mlx5_core_dev *mdev = priv->mdev;
+	int ctxlen = MLX5_ST_SZ_BYTES(tirc);
+	int tt;
+
+	MLX5_SET(modify_tir_in, in, bitmask.hash, 1);
+
+	for (tt = 0; tt < MLX5E_NUM_INDIR_TIRS; tt++) {
+		memset(tirc, 0, ctxlen);
+		mlx5e_build_indir_tir_ctx_hash(&priv->channels.params, tt, tirc, false);
+		mlx5_core_modify_tir(mdev, priv->indir_tir[tt].tirn, in, inlen);
+	}
+
+	if (!mlx5e_tunnel_inner_ft_supported(priv->mdev))
+		return;
+
+	for (tt = 0; tt < MLX5E_NUM_INDIR_TIRS; tt++) {
+		memset(tirc, 0, ctxlen);
+		mlx5e_build_indir_tir_ctx_hash(&priv->channels.params, tt, tirc, true);
+		mlx5_core_modify_tir(mdev, priv->inner_indir_tir[tt].tirn, in, inlen);
+	}
+}
+
+static int mlx5e_set_rxfh(struct net_device *dev, const u32 *indir,
+			  const u8 *key, const u8 hfunc)
+{
+	struct mlx5e_priv *priv = netdev_priv(dev);
+	int inlen = MLX5_ST_SZ_BYTES(modify_tir_in);
+	bool hash_changed = false;
+	void *in;
+
+	if ((hfunc != ETH_RSS_HASH_NO_CHANGE) &&
+	    (hfunc != ETH_RSS_HASH_XOR) &&
+	    (hfunc != ETH_RSS_HASH_TOP))
+		return -EINVAL;
+
+	in = kvzalloc(inlen, GFP_KERNEL);
+	if (!in)
+		return -ENOMEM;
+
+	mutex_lock(&priv->state_lock);
+
+	if (hfunc != ETH_RSS_HASH_NO_CHANGE &&
+	    hfunc != priv->channels.params.rss_hfunc) {
+		priv->channels.params.rss_hfunc = hfunc;
+		hash_changed = true;
+	}
+
+	if (indir) {
+		memcpy(priv->channels.params.indirection_rqt, indir,
+		       sizeof(priv->channels.params.indirection_rqt));
+
+		if (test_bit(MLX5E_STATE_OPENED, &priv->state)) {
+			u32 rqtn = priv->indir_rqt.rqtn;
+			struct mlx5e_redirect_rqt_param rrp = {
+				.is_rss = true,
+				{
+					.rss = {
+						.hfunc = priv->channels.params.rss_hfunc,
+						.channels  = &priv->channels,
+					},
+				},
+			};
+
+			mlx5e_redirect_rqt(priv, rqtn, MLX5E_INDIR_RQT_SIZE, rrp);
+		}
+	}
+
+	if (key) {
+		memcpy(priv->channels.params.toeplitz_hash_key, key,
+		       sizeof(priv->channels.params.toeplitz_hash_key));
+		hash_changed = hash_changed ||
+			       priv->channels.params.rss_hfunc == ETH_RSS_HASH_TOP;
+	}
+
+	if (hash_changed)
+		mlx5e_modify_tirs_hash(priv, in, inlen);
+
+	mutex_unlock(&priv->state_lock);
+
+	kvfree(in);
+
+	return 0;
+}
+
+#define MLX5E_PFC_PREVEN_AUTO_TOUT_MSEC		100
+#define MLX5E_PFC_PREVEN_TOUT_MAX_MSEC		8000
+#define MLX5E_PFC_PREVEN_MINOR_PRECENT		85
+#define MLX5E_PFC_PREVEN_TOUT_MIN_MSEC		80
+#define MLX5E_DEVICE_STALL_MINOR_WATERMARK(critical_tout) \
+	max_t(u16, MLX5E_PFC_PREVEN_TOUT_MIN_MSEC, \
+	      (critical_tout * MLX5E_PFC_PREVEN_MINOR_PRECENT) / 100)
+
+static int mlx5e_get_pfc_prevention_tout(struct net_device *netdev,
+					 u16 *pfc_prevention_tout)
+{
+	struct mlx5e_priv *priv    = netdev_priv(netdev);
+	struct mlx5_core_dev *mdev = priv->mdev;
+
+	if (!MLX5_CAP_PCAM_FEATURE((priv)->mdev, pfcc_mask) ||
+	    !MLX5_CAP_DEBUG((priv)->mdev, stall_detect))
+		return -EOPNOTSUPP;
+
+	return mlx5_query_port_stall_watermark(mdev, pfc_prevention_tout, NULL);
+}
+
+static int mlx5e_set_pfc_prevention_tout(struct net_device *netdev,
+					 u16 pfc_preven)
+{
+	struct mlx5e_priv *priv = netdev_priv(netdev);
+	struct mlx5_core_dev *mdev = priv->mdev;
+	u16 critical_tout;
+	u16 minor;
+
+	if (!MLX5_CAP_PCAM_FEATURE((priv)->mdev, pfcc_mask) ||
+	    !MLX5_CAP_DEBUG((priv)->mdev, stall_detect))
+		return -EOPNOTSUPP;
+
+	critical_tout = (pfc_preven == PFC_STORM_PREVENTION_AUTO) ?
+			MLX5E_PFC_PREVEN_AUTO_TOUT_MSEC :
+			pfc_preven;
+
+	if (critical_tout != PFC_STORM_PREVENTION_DISABLE &&
+	    (critical_tout > MLX5E_PFC_PREVEN_TOUT_MAX_MSEC ||
+	     critical_tout < MLX5E_PFC_PREVEN_TOUT_MIN_MSEC)) {
+		netdev_info(netdev, "%s: pfc prevention tout not in range (%d-%d)\n",
+			    __func__, MLX5E_PFC_PREVEN_TOUT_MIN_MSEC,
+			    MLX5E_PFC_PREVEN_TOUT_MAX_MSEC);
+		return -EINVAL;
+	}
+
+	minor = MLX5E_DEVICE_STALL_MINOR_WATERMARK(critical_tout);
+	return mlx5_set_port_stall_watermark(mdev, critical_tout,
+					     minor);
+}
+
+static int mlx5e_get_tunable(struct net_device *dev,
+			     const struct ethtool_tunable *tuna,
+			     void *data)
+{
+	int err;
+
+	switch (tuna->id) {
+	case ETHTOOL_PFC_PREVENTION_TOUT:
+		err = mlx5e_get_pfc_prevention_tout(dev, data);
+		break;
+	default:
+		err = -EINVAL;
+		break;
+	}
+
+	return err;
+}
+
+static int mlx5e_set_tunable(struct net_device *dev,
+			     const struct ethtool_tunable *tuna,
+			     const void *data)
+{
+	struct mlx5e_priv *priv = netdev_priv(dev);
+	int err;
+
+	mutex_lock(&priv->state_lock);
+
+	switch (tuna->id) {
+	case ETHTOOL_PFC_PREVENTION_TOUT:
+		err = mlx5e_set_pfc_prevention_tout(dev, *(u16 *)data);
+		break;
+	default:
+		err = -EINVAL;
+		break;
+	}
+
+	mutex_unlock(&priv->state_lock);
+	return err;
+}
+
+static void mlx5e_get_pauseparam(struct net_device *netdev,
+				 struct ethtool_pauseparam *pauseparam)
+{
+	struct mlx5e_priv *priv    = netdev_priv(netdev);
+	struct mlx5_core_dev *mdev = priv->mdev;
+	int err;
+
+	err = mlx5_query_port_pause(mdev, &pauseparam->rx_pause,
+				    &pauseparam->tx_pause);
+	if (err) {
+		netdev_err(netdev, "%s: mlx5_query_port_pause failed:0x%x\n",
+			   __func__, err);
+	}
+}
+
+static int mlx5e_set_pauseparam(struct net_device *netdev,
+				struct ethtool_pauseparam *pauseparam)
+{
+	struct mlx5e_priv *priv    = netdev_priv(netdev);
+	struct mlx5_core_dev *mdev = priv->mdev;
+	int err;
+
+	if (!MLX5_CAP_GEN(mdev, vport_group_manager))
+		return -EOPNOTSUPP;
+
+	if (pauseparam->autoneg)
+		return -EINVAL;
+
+	err = mlx5_set_port_pause(mdev,
+				  pauseparam->rx_pause ? 1 : 0,
+				  pauseparam->tx_pause ? 1 : 0);
+	if (err) {
+		netdev_err(netdev, "%s: mlx5_set_port_pause failed:0x%x\n",
+			   __func__, err);
+	}
+
+	return err;
+}
+
+int mlx5e_ethtool_get_ts_info(struct mlx5e_priv *priv,
+			      struct ethtool_ts_info *info)
+{
+	struct mlx5_core_dev *mdev = priv->mdev;
+
+	info->phc_index = mlx5_clock_get_ptp_index(mdev);
+
+	if (!MLX5_CAP_GEN(priv->mdev, device_frequency_khz) ||
+	    info->phc_index == -1)
+		return 0;
+
+	info->so_timestamping = SOF_TIMESTAMPING_TX_HARDWARE |
+				SOF_TIMESTAMPING_RX_HARDWARE |
+				SOF_TIMESTAMPING_RAW_HARDWARE;
+
+	info->tx_types = BIT(HWTSTAMP_TX_OFF) |
+			 BIT(HWTSTAMP_TX_ON);
+
+	info->rx_filters = BIT(HWTSTAMP_FILTER_NONE) |
+			   BIT(HWTSTAMP_FILTER_ALL);
+
+	return 0;
+}
+
+static int mlx5e_get_ts_info(struct net_device *dev,
+			     struct ethtool_ts_info *info)
+{
+	struct mlx5e_priv *priv = netdev_priv(dev);
+
+	return mlx5e_ethtool_get_ts_info(priv, info);
+}
+
+static __u32 mlx5e_get_wol_supported(struct mlx5_core_dev *mdev)
+{
+	__u32 ret = 0;
+
+	if (MLX5_CAP_GEN(mdev, wol_g))
+		ret |= WAKE_MAGIC;
+
+	if (MLX5_CAP_GEN(mdev, wol_s))
+		ret |= WAKE_MAGICSECURE;
+
+	if (MLX5_CAP_GEN(mdev, wol_a))
+		ret |= WAKE_ARP;
+
+	if (MLX5_CAP_GEN(mdev, wol_b))
+		ret |= WAKE_BCAST;
+
+	if (MLX5_CAP_GEN(mdev, wol_m))
+		ret |= WAKE_MCAST;
+
+	if (MLX5_CAP_GEN(mdev, wol_u))
+		ret |= WAKE_UCAST;
+
+	if (MLX5_CAP_GEN(mdev, wol_p))
+		ret |= WAKE_PHY;
+
+	return ret;
+}
+
+static __u32 mlx5e_refomrat_wol_mode_mlx5_to_linux(u8 mode)
+{
+	__u32 ret = 0;
+
+	if (mode & MLX5_WOL_MAGIC)
+		ret |= WAKE_MAGIC;
+
+	if (mode & MLX5_WOL_SECURED_MAGIC)
+		ret |= WAKE_MAGICSECURE;
+
+	if (mode & MLX5_WOL_ARP)
+		ret |= WAKE_ARP;
+
+	if (mode & MLX5_WOL_BROADCAST)
+		ret |= WAKE_BCAST;
+
+	if (mode & MLX5_WOL_MULTICAST)
+		ret |= WAKE_MCAST;
+
+	if (mode & MLX5_WOL_UNICAST)
+		ret |= WAKE_UCAST;
+
+	if (mode & MLX5_WOL_PHY_ACTIVITY)
+		ret |= WAKE_PHY;
+
+	return ret;
+}
+
+static u8 mlx5e_refomrat_wol_mode_linux_to_mlx5(__u32 mode)
+{
+	u8 ret = 0;
+
+	if (mode & WAKE_MAGIC)
+		ret |= MLX5_WOL_MAGIC;
+
+	if (mode & WAKE_MAGICSECURE)
+		ret |= MLX5_WOL_SECURED_MAGIC;
+
+	if (mode & WAKE_ARP)
+		ret |= MLX5_WOL_ARP;
+
+	if (mode & WAKE_BCAST)
+		ret |= MLX5_WOL_BROADCAST;
+
+	if (mode & WAKE_MCAST)
+		ret |= MLX5_WOL_MULTICAST;
+
+	if (mode & WAKE_UCAST)
+		ret |= MLX5_WOL_UNICAST;
+
+	if (mode & WAKE_PHY)
+		ret |= MLX5_WOL_PHY_ACTIVITY;
+
+	return ret;
+}
+
+static void mlx5e_get_wol(struct net_device *netdev,
+			  struct ethtool_wolinfo *wol)
+{
+	struct mlx5e_priv *priv = netdev_priv(netdev);
+	struct mlx5_core_dev *mdev = priv->mdev;
+	u8 mlx5_wol_mode;
+	int err;
+
+	memset(wol, 0, sizeof(*wol));
+
+	wol->supported = mlx5e_get_wol_supported(mdev);
+	if (!wol->supported)
+		return;
+
+	err = mlx5_query_port_wol(mdev, &mlx5_wol_mode);
+	if (err)
+		return;
+
+	wol->wolopts = mlx5e_refomrat_wol_mode_mlx5_to_linux(mlx5_wol_mode);
+}
+
+static int mlx5e_set_wol(struct net_device *netdev, struct ethtool_wolinfo *wol)
+{
+	struct mlx5e_priv *priv = netdev_priv(netdev);
+	struct mlx5_core_dev *mdev = priv->mdev;
+	__u32 wol_supported = mlx5e_get_wol_supported(mdev);
+	u32 mlx5_wol_mode;
+
+	if (!wol_supported)
+		return -EOPNOTSUPP;
+
+	if (wol->wolopts & ~wol_supported)
+		return -EINVAL;
+
+	mlx5_wol_mode = mlx5e_refomrat_wol_mode_linux_to_mlx5(wol->wolopts);
+
+	return mlx5_set_port_wol(mdev, mlx5_wol_mode);
+}
+
+static u32 mlx5e_get_msglevel(struct net_device *dev)
+{
+	return ((struct mlx5e_priv *)netdev_priv(dev))->msglevel;
+}
+
+static void mlx5e_set_msglevel(struct net_device *dev, u32 val)
+{
+	((struct mlx5e_priv *)netdev_priv(dev))->msglevel = val;
+}
+
+static int mlx5e_set_phys_id(struct net_device *dev,
+			     enum ethtool_phys_id_state state)
+{
+	struct mlx5e_priv *priv = netdev_priv(dev);
+	struct mlx5_core_dev *mdev = priv->mdev;
+	u16 beacon_duration;
+
+	if (!MLX5_CAP_GEN(mdev, beacon_led))
+		return -EOPNOTSUPP;
+
+	switch (state) {
+	case ETHTOOL_ID_ACTIVE:
+		beacon_duration = MLX5_BEACON_DURATION_INF;
+		break;
+	case ETHTOOL_ID_INACTIVE:
+		beacon_duration = MLX5_BEACON_DURATION_OFF;
+		break;
+	default:
+		return -EOPNOTSUPP;
+	}
+
+	return mlx5_set_port_beacon(mdev, beacon_duration);
+}
+
+static int mlx5e_get_module_info(struct net_device *netdev,
+				 struct ethtool_modinfo *modinfo)
+{
+	struct mlx5e_priv *priv = netdev_priv(netdev);
+	struct mlx5_core_dev *dev = priv->mdev;
+	int size_read = 0;
+	u8 data[4];
+
+	size_read = mlx5_query_module_eeprom(dev, 0, 2, data);
+	if (size_read < 2)
+		return -EIO;
+
+	/* data[0] = identifier byte */
+	switch (data[0]) {
+	case MLX5_MODULE_ID_QSFP:
+		modinfo->type       = ETH_MODULE_SFF_8436;
+		modinfo->eeprom_len = ETH_MODULE_SFF_8436_LEN;
+		break;
+	case MLX5_MODULE_ID_QSFP_PLUS:
+	case MLX5_MODULE_ID_QSFP28:
+		/* data[1] = revision id */
+		if (data[0] == MLX5_MODULE_ID_QSFP28 || data[1] >= 0x3) {
+			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 MLX5_MODULE_ID_SFP:
+		modinfo->type       = ETH_MODULE_SFF_8472;
+		modinfo->eeprom_len = ETH_MODULE_SFF_8472_LEN;
+		break;
+	default:
+		netdev_err(priv->netdev, "%s: cable type not recognized:0x%x\n",
+			   __func__, data[0]);
+		return -EINVAL;
+	}
+
+	return 0;
+}
+
+static int mlx5e_get_module_eeprom(struct net_device *netdev,
+				   struct ethtool_eeprom *ee,
+				   u8 *data)
+{
+	struct mlx5e_priv *priv = netdev_priv(netdev);
+	struct mlx5_core_dev *mdev = priv->mdev;
+	int offset = ee->offset;
+	int size_read;
+	int i = 0;
+
+	if (!ee->len)
+		return -EINVAL;
+
+	memset(data, 0, ee->len);
+
+	while (i < ee->len) {
+		size_read = mlx5_query_module_eeprom(mdev, offset, ee->len - i,
+						     data + i);
+
+		if (!size_read)
+			/* Done reading */
+			return 0;
+
+		if (size_read < 0) {
+			netdev_err(priv->netdev, "%s: mlx5_query_eeprom failed:0x%x\n",
+				   __func__, size_read);
+			return size_read;
+		}
+
+		i += size_read;
+		offset += size_read;
+	}
+
+	return 0;
+}
+
+typedef int (*mlx5e_pflag_handler)(struct net_device *netdev, bool enable);
+
+static int set_pflag_cqe_based_moder(struct net_device *netdev, bool enable,
+				     bool is_rx_cq)
+{
+	struct mlx5e_priv *priv = netdev_priv(netdev);
+	struct mlx5_core_dev *mdev = priv->mdev;
+	struct mlx5e_channels new_channels = {};
+	bool mode_changed;
+	u8 cq_period_mode, current_cq_period_mode;
+	int err = 0;
+
+	cq_period_mode = enable ?
+		MLX5_CQ_PERIOD_MODE_START_FROM_CQE :
+		MLX5_CQ_PERIOD_MODE_START_FROM_EQE;
+	current_cq_period_mode = is_rx_cq ?
+		priv->channels.params.rx_cq_moderation.cq_period_mode :
+		priv->channels.params.tx_cq_moderation.cq_period_mode;
+	mode_changed = cq_period_mode != current_cq_period_mode;
+
+	if (cq_period_mode == MLX5_CQ_PERIOD_MODE_START_FROM_CQE &&
+	    !MLX5_CAP_GEN(mdev, cq_period_start_from_cqe))
+		return -EOPNOTSUPP;
+
+	if (!mode_changed)
+		return 0;
+
+	new_channels.params = priv->channels.params;
+	if (is_rx_cq)
+		mlx5e_set_rx_cq_mode_params(&new_channels.params, cq_period_mode);
+	else
+		mlx5e_set_tx_cq_mode_params(&new_channels.params, cq_period_mode);
+
+	if (!test_bit(MLX5E_STATE_OPENED, &priv->state)) {
+		priv->channels.params = new_channels.params;
+		return 0;
+	}
+
+	err = mlx5e_open_channels(priv, &new_channels);
+	if (err)
+		return err;
+
+	mlx5e_switch_priv_channels(priv, &new_channels, NULL);
+	return 0;
+}
+
+static int set_pflag_tx_cqe_based_moder(struct net_device *netdev, bool enable)
+{
+	return set_pflag_cqe_based_moder(netdev, enable, false);
+}
+
+static int set_pflag_rx_cqe_based_moder(struct net_device *netdev, bool enable)
+{
+	return set_pflag_cqe_based_moder(netdev, enable, true);
+}
+
+int mlx5e_modify_rx_cqe_compression_locked(struct mlx5e_priv *priv, bool new_val)
+{
+	bool curr_val = MLX5E_GET_PFLAG(&priv->channels.params, MLX5E_PFLAG_RX_CQE_COMPRESS);
+	struct mlx5e_channels new_channels = {};
+	int err = 0;
+
+	if (!MLX5_CAP_GEN(priv->mdev, cqe_compression))
+		return new_val ? -EOPNOTSUPP : 0;
+
+	if (curr_val == new_val)
+		return 0;
+
+	new_channels.params = priv->channels.params;
+	MLX5E_SET_PFLAG(&new_channels.params, MLX5E_PFLAG_RX_CQE_COMPRESS, new_val);
+
+	if (!test_bit(MLX5E_STATE_OPENED, &priv->state)) {
+		priv->channels.params = new_channels.params;
+		return 0;
+	}
+
+	err = mlx5e_open_channels(priv, &new_channels);
+	if (err)
+		return err;
+
+	mlx5e_switch_priv_channels(priv, &new_channels, NULL);
+	mlx5e_dbg(DRV, priv, "MLX5E: RxCqeCmprss was turned %s\n",
+		  MLX5E_GET_PFLAG(&priv->channels.params,
+				  MLX5E_PFLAG_RX_CQE_COMPRESS) ? "ON" : "OFF");
+
+	return 0;
+}
+
+static int set_pflag_rx_cqe_compress(struct net_device *netdev,
+				     bool enable)
+{
+	struct mlx5e_priv *priv = netdev_priv(netdev);
+	struct mlx5_core_dev *mdev = priv->mdev;
+	int err;
+
+	if (!MLX5_CAP_GEN(mdev, cqe_compression))
+		return -EOPNOTSUPP;
+
+	if (enable && priv->tstamp.rx_filter != HWTSTAMP_FILTER_NONE) {
+		netdev_err(netdev, "Can't enable cqe compression while timestamping is enabled.\n");
+		return -EINVAL;
+	}
+
+	err = mlx5e_modify_rx_cqe_compression_locked(priv, enable);
+	if (err)
+		return err;
+
+	priv->channels.params.rx_cqe_compress_def = enable;
+
+	return 0;
+}
+
+static int set_pflag_rx_striding_rq(struct net_device *netdev, bool enable)
+{
+	struct mlx5e_priv *priv = netdev_priv(netdev);
+	struct mlx5_core_dev *mdev = priv->mdev;
+	struct mlx5e_channels new_channels = {};
+	int err;
+
+	if (enable) {
+		if (!mlx5e_check_fragmented_striding_rq_cap(mdev))
+			return -EOPNOTSUPP;
+		if (!mlx5e_striding_rq_possible(mdev, &priv->channels.params))
+			return -EINVAL;
+	} else if (priv->channels.params.lro_en) {
+		netdev_warn(netdev, "Can't set legacy RQ with LRO, disable LRO first\n");
+		return -EINVAL;
+	}
+
+	new_channels.params = priv->channels.params;
+
+	MLX5E_SET_PFLAG(&new_channels.params, MLX5E_PFLAG_RX_STRIDING_RQ, enable);
+	mlx5e_set_rq_type(mdev, &new_channels.params);
+
+	if (!test_bit(MLX5E_STATE_OPENED, &priv->state)) {
+		priv->channels.params = new_channels.params;
+		return 0;
+	}
+
+	err = mlx5e_open_channels(priv, &new_channels);
+	if (err)
+		return err;
+
+	mlx5e_switch_priv_channels(priv, &new_channels, NULL);
+	return 0;
+}
+
+static int set_pflag_rx_no_csum_complete(struct net_device *netdev, bool enable)
+{
+	struct mlx5e_priv *priv = netdev_priv(netdev);
+	struct mlx5e_channels *channels = &priv->channels;
+	struct mlx5e_channel *c;
+	int i;
+
+	if (!test_bit(MLX5E_STATE_OPENED, &priv->state) ||
+	    priv->channels.params.xdp_prog)
+		return 0;
+
+	for (i = 0; i < channels->num; i++) {
+		c = channels->c[i];
+		if (enable)
+			__set_bit(MLX5E_RQ_STATE_NO_CSUM_COMPLETE, &c->rq.state);
+		else
+			__clear_bit(MLX5E_RQ_STATE_NO_CSUM_COMPLETE, &c->rq.state);
+	}
+
+	return 0;
+}
+
+static int mlx5e_handle_pflag(struct net_device *netdev,
+			      u32 wanted_flags,
+			      enum mlx5e_priv_flag flag,
+			      mlx5e_pflag_handler pflag_handler)
+{
+	struct mlx5e_priv *priv = netdev_priv(netdev);
+	bool enable = !!(wanted_flags & flag);
+	u32 changes = wanted_flags ^ priv->channels.params.pflags;
+	int err;
+
+	if (!(changes & flag))
+		return 0;
+
+	err = pflag_handler(netdev, enable);
+	if (err) {
+		netdev_err(netdev, "%s private flag 0x%x failed err %d\n",
+			   enable ? "Enable" : "Disable", flag, err);
+		return err;
+	}
+
+	MLX5E_SET_PFLAG(&priv->channels.params, flag, enable);
+	return 0;
+}
+
+static int mlx5e_set_priv_flags(struct net_device *netdev, u32 pflags)
+{
+	struct mlx5e_priv *priv = netdev_priv(netdev);
+	int err;
+
+	mutex_lock(&priv->state_lock);
+	err = mlx5e_handle_pflag(netdev, pflags,
+				 MLX5E_PFLAG_RX_CQE_BASED_MODER,
+				 set_pflag_rx_cqe_based_moder);
+	if (err)
+		goto out;
+
+	err = mlx5e_handle_pflag(netdev, pflags,
+				 MLX5E_PFLAG_TX_CQE_BASED_MODER,
+				 set_pflag_tx_cqe_based_moder);
+	if (err)
+		goto out;
+
+	err = mlx5e_handle_pflag(netdev, pflags,
+				 MLX5E_PFLAG_RX_CQE_COMPRESS,
+				 set_pflag_rx_cqe_compress);
+	if (err)
+		goto out;
+
+	err = mlx5e_handle_pflag(netdev, pflags,
+				 MLX5E_PFLAG_RX_STRIDING_RQ,
+				 set_pflag_rx_striding_rq);
+	if (err)
+		goto out;
+
+	err = mlx5e_handle_pflag(netdev, pflags,
+				 MLX5E_PFLAG_RX_NO_CSUM_COMPLETE,
+				 set_pflag_rx_no_csum_complete);
+
+out:
+	mutex_unlock(&priv->state_lock);
+
+	/* Need to fix some features.. */
+	netdev_update_features(netdev);
+
+	return err;
+}
+
+static u32 mlx5e_get_priv_flags(struct net_device *netdev)
+{
+	struct mlx5e_priv *priv = netdev_priv(netdev);
+
+	return priv->channels.params.pflags;
+}
+
+int mlx5e_ethtool_flash_device(struct mlx5e_priv *priv,
+			       struct ethtool_flash *flash)
+{
+	struct mlx5_core_dev *mdev = priv->mdev;
+	struct net_device *dev = priv->netdev;
+	const struct firmware *fw;
+	int err;
+
+	if (flash->region != ETHTOOL_FLASH_ALL_REGIONS)
+		return -EOPNOTSUPP;
+
+	err = request_firmware_direct(&fw, flash->data, &dev->dev);
+	if (err)
+		return err;
+
+	dev_hold(dev);
+	rtnl_unlock();
+
+	err = mlx5_firmware_flash(mdev, fw);
+	release_firmware(fw);
+
+	rtnl_lock();
+	dev_put(dev);
+	return err;
+}
+
+static int mlx5e_flash_device(struct net_device *dev,
+			      struct ethtool_flash *flash)
+{
+	struct mlx5e_priv *priv = netdev_priv(dev);
+
+	return mlx5e_ethtool_flash_device(priv, flash);
+}
+
+#ifndef CONFIG_MLX5_EN_RXNFC
+/* When CONFIG_MLX5_EN_RXNFC=n we only support ETHTOOL_GRXRINGS
+ * otherwise this function will be defined from en_fs_ethtool.c
+ */
+static int mlx5e_get_rxnfc(struct net_device *dev, struct ethtool_rxnfc *info, u32 *rule_locs)
+{
+	struct mlx5e_priv *priv = netdev_priv(dev);
+
+	if (info->cmd != ETHTOOL_GRXRINGS)
+		return -EOPNOTSUPP;
+	/* ring_count is needed by ethtool -x */
+	info->data = priv->channels.params.num_channels;
+	return 0;
+}
+#endif
+
+const struct ethtool_ops mlx5e_ethtool_ops = {
+	.get_drvinfo       = mlx5e_get_drvinfo,
+	.get_link          = ethtool_op_get_link,
+	.get_strings       = mlx5e_get_strings,
+	.get_sset_count    = mlx5e_get_sset_count,
+	.get_ethtool_stats = mlx5e_get_ethtool_stats,
+	.get_ringparam     = mlx5e_get_ringparam,
+	.set_ringparam     = mlx5e_set_ringparam,
+	.get_channels      = mlx5e_get_channels,
+	.set_channels      = mlx5e_set_channels,
+	.get_coalesce      = mlx5e_get_coalesce,
+	.set_coalesce      = mlx5e_set_coalesce,
+	.get_link_ksettings  = mlx5e_get_link_ksettings,
+	.set_link_ksettings  = mlx5e_set_link_ksettings,
+	.get_rxfh_key_size   = mlx5e_get_rxfh_key_size,
+	.get_rxfh_indir_size = mlx5e_get_rxfh_indir_size,
+	.get_rxfh          = mlx5e_get_rxfh,
+	.set_rxfh          = mlx5e_set_rxfh,
+	.get_rxnfc         = mlx5e_get_rxnfc,
+#ifdef CONFIG_MLX5_EN_RXNFC
+	.set_rxnfc         = mlx5e_set_rxnfc,
+#endif
+	.flash_device      = mlx5e_flash_device,
+	.get_tunable       = mlx5e_get_tunable,
+	.set_tunable       = mlx5e_set_tunable,
+	.get_pauseparam    = mlx5e_get_pauseparam,
+	.set_pauseparam    = mlx5e_set_pauseparam,
+	.get_ts_info       = mlx5e_get_ts_info,
+	.set_phys_id       = mlx5e_set_phys_id,
+	.get_wol	   = mlx5e_get_wol,
+	.set_wol	   = mlx5e_set_wol,
+	.get_module_info   = mlx5e_get_module_info,
+	.get_module_eeprom = mlx5e_get_module_eeprom,
+	.get_priv_flags    = mlx5e_get_priv_flags,
+	.set_priv_flags    = mlx5e_set_priv_flags,
+	.self_test         = mlx5e_self_test,
+	.get_msglevel      = mlx5e_get_msglevel,
+	.set_msglevel      = mlx5e_set_msglevel,
+};
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_fs.c b/drivers/net/ethernet/mellanox/mlx5/core/en_fs.c
new file mode 100644
index 000000000..c6eea6b6b
--- /dev/null
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_fs.c
@@ -0,0 +1,1576 @@
+/*
+ * Copyright (c) 2015, 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/list.h>
+#include <linux/ip.h>
+#include <linux/ipv6.h>
+#include <linux/tcp.h>
+#include <linux/mlx5/fs.h>
+#include "en.h"
+#include "lib/mpfs.h"
+
+static int mlx5e_add_l2_flow_rule(struct mlx5e_priv *priv,
+				  struct mlx5e_l2_rule *ai, int type);
+static void mlx5e_del_l2_flow_rule(struct mlx5e_priv *priv,
+				   struct mlx5e_l2_rule *ai);
+
+enum {
+	MLX5E_FULLMATCH = 0,
+	MLX5E_ALLMULTI  = 1,
+	MLX5E_PROMISC   = 2,
+};
+
+enum {
+	MLX5E_UC        = 0,
+	MLX5E_MC_IPV4   = 1,
+	MLX5E_MC_IPV6   = 2,
+	MLX5E_MC_OTHER  = 3,
+};
+
+enum {
+	MLX5E_ACTION_NONE = 0,
+	MLX5E_ACTION_ADD  = 1,
+	MLX5E_ACTION_DEL  = 2,
+};
+
+struct mlx5e_l2_hash_node {
+	struct hlist_node          hlist;
+	u8                         action;
+	struct mlx5e_l2_rule ai;
+	bool   mpfs;
+};
+
+static inline int mlx5e_hash_l2(u8 *addr)
+{
+	return addr[5];
+}
+
+static void mlx5e_add_l2_to_hash(struct hlist_head *hash, u8 *addr)
+{
+	struct mlx5e_l2_hash_node *hn;
+	int ix = mlx5e_hash_l2(addr);
+	int found = 0;
+
+	hlist_for_each_entry(hn, &hash[ix], hlist)
+		if (ether_addr_equal_64bits(hn->ai.addr, addr)) {
+			found = 1;
+			break;
+		}
+
+	if (found) {
+		hn->action = MLX5E_ACTION_NONE;
+		return;
+	}
+
+	hn = kzalloc(sizeof(*hn), GFP_ATOMIC);
+	if (!hn)
+		return;
+
+	ether_addr_copy(hn->ai.addr, addr);
+	hn->action = MLX5E_ACTION_ADD;
+
+	hlist_add_head(&hn->hlist, &hash[ix]);
+}
+
+static void mlx5e_del_l2_from_hash(struct mlx5e_l2_hash_node *hn)
+{
+	hlist_del(&hn->hlist);
+	kfree(hn);
+}
+
+static int mlx5e_vport_context_update_vlans(struct mlx5e_priv *priv)
+{
+	struct net_device *ndev = priv->netdev;
+	int max_list_size;
+	int list_size;
+	u16 *vlans;
+	int vlan;
+	int err;
+	int i;
+
+	list_size = 0;
+	for_each_set_bit(vlan, priv->fs.vlan.active_cvlans, VLAN_N_VID)
+		list_size++;
+
+	max_list_size = 1 << MLX5_CAP_GEN(priv->mdev, log_max_vlan_list);
+
+	if (list_size > max_list_size) {
+		netdev_warn(ndev,
+			    "netdev vlans list size (%d) > (%d) max vport list size, some vlans will be dropped\n",
+			    list_size, max_list_size);
+		list_size = max_list_size;
+	}
+
+	vlans = kcalloc(list_size, sizeof(*vlans), GFP_KERNEL);
+	if (!vlans)
+		return -ENOMEM;
+
+	i = 0;
+	for_each_set_bit(vlan, priv->fs.vlan.active_cvlans, VLAN_N_VID) {
+		if (i >= list_size)
+			break;
+		vlans[i++] = vlan;
+	}
+
+	err = mlx5_modify_nic_vport_vlans(priv->mdev, vlans, list_size);
+	if (err)
+		netdev_err(ndev, "Failed to modify vport vlans list err(%d)\n",
+			   err);
+
+	kfree(vlans);
+	return err;
+}
+
+enum mlx5e_vlan_rule_type {
+	MLX5E_VLAN_RULE_TYPE_UNTAGGED,
+	MLX5E_VLAN_RULE_TYPE_ANY_CTAG_VID,
+	MLX5E_VLAN_RULE_TYPE_ANY_STAG_VID,
+	MLX5E_VLAN_RULE_TYPE_MATCH_CTAG_VID,
+	MLX5E_VLAN_RULE_TYPE_MATCH_STAG_VID,
+};
+
+static int __mlx5e_add_vlan_rule(struct mlx5e_priv *priv,
+				 enum mlx5e_vlan_rule_type rule_type,
+				 u16 vid, struct mlx5_flow_spec *spec)
+{
+	struct mlx5_flow_table *ft = priv->fs.vlan.ft.t;
+	struct mlx5_flow_destination dest = {};
+	struct mlx5_flow_handle **rule_p;
+	MLX5_DECLARE_FLOW_ACT(flow_act);
+	int err = 0;
+
+	dest.type = MLX5_FLOW_DESTINATION_TYPE_FLOW_TABLE;
+	dest.ft = priv->fs.l2.ft.t;
+
+	spec->match_criteria_enable = MLX5_MATCH_OUTER_HEADERS;
+
+	switch (rule_type) {
+	case MLX5E_VLAN_RULE_TYPE_UNTAGGED:
+		/* cvlan_tag enabled in match criteria and
+		 * disabled in match value means both S & C tags
+		 * don't exist (untagged of both)
+		 */
+		rule_p = &priv->fs.vlan.untagged_rule;
+		MLX5_SET_TO_ONES(fte_match_param, spec->match_criteria,
+				 outer_headers.cvlan_tag);
+		break;
+	case MLX5E_VLAN_RULE_TYPE_ANY_CTAG_VID:
+		rule_p = &priv->fs.vlan.any_cvlan_rule;
+		MLX5_SET_TO_ONES(fte_match_param, spec->match_criteria,
+				 outer_headers.cvlan_tag);
+		MLX5_SET(fte_match_param, spec->match_value, outer_headers.cvlan_tag, 1);
+		break;
+	case MLX5E_VLAN_RULE_TYPE_ANY_STAG_VID:
+		rule_p = &priv->fs.vlan.any_svlan_rule;
+		MLX5_SET_TO_ONES(fte_match_param, spec->match_criteria,
+				 outer_headers.svlan_tag);
+		MLX5_SET(fte_match_param, spec->match_value, outer_headers.svlan_tag, 1);
+		break;
+	case MLX5E_VLAN_RULE_TYPE_MATCH_STAG_VID:
+		rule_p = &priv->fs.vlan.active_svlans_rule[vid];
+		MLX5_SET_TO_ONES(fte_match_param, spec->match_criteria,
+				 outer_headers.svlan_tag);
+		MLX5_SET(fte_match_param, spec->match_value, outer_headers.svlan_tag, 1);
+		MLX5_SET_TO_ONES(fte_match_param, spec->match_criteria,
+				 outer_headers.first_vid);
+		MLX5_SET(fte_match_param, spec->match_value, outer_headers.first_vid,
+			 vid);
+		break;
+	default: /* MLX5E_VLAN_RULE_TYPE_MATCH_CTAG_VID */
+		rule_p = &priv->fs.vlan.active_cvlans_rule[vid];
+		MLX5_SET_TO_ONES(fte_match_param, spec->match_criteria,
+				 outer_headers.cvlan_tag);
+		MLX5_SET(fte_match_param, spec->match_value, outer_headers.cvlan_tag, 1);
+		MLX5_SET_TO_ONES(fte_match_param, spec->match_criteria,
+				 outer_headers.first_vid);
+		MLX5_SET(fte_match_param, spec->match_value, outer_headers.first_vid,
+			 vid);
+		break;
+	}
+
+	if (WARN_ONCE(*rule_p, "VLAN rule already exists type %d", rule_type))
+		return 0;
+
+	*rule_p = mlx5_add_flow_rules(ft, spec, &flow_act, &dest, 1);
+
+	if (IS_ERR(*rule_p)) {
+		err = PTR_ERR(*rule_p);
+		*rule_p = NULL;
+		netdev_err(priv->netdev, "%s: add rule failed\n", __func__);
+	}
+
+	return err;
+}
+
+static int mlx5e_add_vlan_rule(struct mlx5e_priv *priv,
+			       enum mlx5e_vlan_rule_type rule_type, u16 vid)
+{
+	struct mlx5_flow_spec *spec;
+	int err = 0;
+
+	spec = kvzalloc(sizeof(*spec), GFP_KERNEL);
+	if (!spec)
+		return -ENOMEM;
+
+	if (rule_type == MLX5E_VLAN_RULE_TYPE_MATCH_CTAG_VID)
+		mlx5e_vport_context_update_vlans(priv);
+
+	err = __mlx5e_add_vlan_rule(priv, rule_type, vid, spec);
+
+	kvfree(spec);
+
+	return err;
+}
+
+static void mlx5e_del_vlan_rule(struct mlx5e_priv *priv,
+				enum mlx5e_vlan_rule_type rule_type, u16 vid)
+{
+	switch (rule_type) {
+	case MLX5E_VLAN_RULE_TYPE_UNTAGGED:
+		if (priv->fs.vlan.untagged_rule) {
+			mlx5_del_flow_rules(priv->fs.vlan.untagged_rule);
+			priv->fs.vlan.untagged_rule = NULL;
+		}
+		break;
+	case MLX5E_VLAN_RULE_TYPE_ANY_CTAG_VID:
+		if (priv->fs.vlan.any_cvlan_rule) {
+			mlx5_del_flow_rules(priv->fs.vlan.any_cvlan_rule);
+			priv->fs.vlan.any_cvlan_rule = NULL;
+		}
+		break;
+	case MLX5E_VLAN_RULE_TYPE_ANY_STAG_VID:
+		if (priv->fs.vlan.any_svlan_rule) {
+			mlx5_del_flow_rules(priv->fs.vlan.any_svlan_rule);
+			priv->fs.vlan.any_svlan_rule = NULL;
+		}
+		break;
+	case MLX5E_VLAN_RULE_TYPE_MATCH_STAG_VID:
+		if (priv->fs.vlan.active_svlans_rule[vid]) {
+			mlx5_del_flow_rules(priv->fs.vlan.active_svlans_rule[vid]);
+			priv->fs.vlan.active_svlans_rule[vid] = NULL;
+		}
+		break;
+	case MLX5E_VLAN_RULE_TYPE_MATCH_CTAG_VID:
+		if (priv->fs.vlan.active_cvlans_rule[vid]) {
+			mlx5_del_flow_rules(priv->fs.vlan.active_cvlans_rule[vid]);
+			priv->fs.vlan.active_cvlans_rule[vid] = NULL;
+		}
+		mlx5e_vport_context_update_vlans(priv);
+		break;
+	}
+}
+
+static void mlx5e_del_any_vid_rules(struct mlx5e_priv *priv)
+{
+	mlx5e_del_vlan_rule(priv, MLX5E_VLAN_RULE_TYPE_ANY_CTAG_VID, 0);
+	mlx5e_del_vlan_rule(priv, MLX5E_VLAN_RULE_TYPE_ANY_STAG_VID, 0);
+}
+
+static int mlx5e_add_any_vid_rules(struct mlx5e_priv *priv)
+{
+	int err;
+
+	err = mlx5e_add_vlan_rule(priv, MLX5E_VLAN_RULE_TYPE_ANY_CTAG_VID, 0);
+	if (err)
+		return err;
+
+	return mlx5e_add_vlan_rule(priv, MLX5E_VLAN_RULE_TYPE_ANY_STAG_VID, 0);
+}
+
+void mlx5e_enable_cvlan_filter(struct mlx5e_priv *priv)
+{
+	if (!priv->fs.vlan.cvlan_filter_disabled)
+		return;
+
+	priv->fs.vlan.cvlan_filter_disabled = false;
+	if (priv->netdev->flags & IFF_PROMISC)
+		return;
+	mlx5e_del_vlan_rule(priv, MLX5E_VLAN_RULE_TYPE_ANY_CTAG_VID, 0);
+}
+
+void mlx5e_disable_cvlan_filter(struct mlx5e_priv *priv)
+{
+	if (priv->fs.vlan.cvlan_filter_disabled)
+		return;
+
+	priv->fs.vlan.cvlan_filter_disabled = true;
+	if (priv->netdev->flags & IFF_PROMISC)
+		return;
+	mlx5e_add_vlan_rule(priv, MLX5E_VLAN_RULE_TYPE_ANY_CTAG_VID, 0);
+}
+
+static int mlx5e_vlan_rx_add_cvid(struct mlx5e_priv *priv, u16 vid)
+{
+	int err;
+
+	set_bit(vid, priv->fs.vlan.active_cvlans);
+
+	err = mlx5e_add_vlan_rule(priv, MLX5E_VLAN_RULE_TYPE_MATCH_CTAG_VID, vid);
+	if (err)
+		clear_bit(vid, priv->fs.vlan.active_cvlans);
+
+	return err;
+}
+
+static int mlx5e_vlan_rx_add_svid(struct mlx5e_priv *priv, u16 vid)
+{
+	struct net_device *netdev = priv->netdev;
+	int err;
+
+	set_bit(vid, priv->fs.vlan.active_svlans);
+
+	err = mlx5e_add_vlan_rule(priv, MLX5E_VLAN_RULE_TYPE_MATCH_STAG_VID, vid);
+	if (err) {
+		clear_bit(vid, priv->fs.vlan.active_svlans);
+		return err;
+	}
+
+	/* Need to fix some features.. */
+	netdev_update_features(netdev);
+	return err;
+}
+
+int mlx5e_vlan_rx_add_vid(struct net_device *dev, __be16 proto, u16 vid)
+{
+	struct mlx5e_priv *priv = netdev_priv(dev);
+
+	if (be16_to_cpu(proto) == ETH_P_8021Q)
+		return mlx5e_vlan_rx_add_cvid(priv, vid);
+	else if (be16_to_cpu(proto) == ETH_P_8021AD)
+		return mlx5e_vlan_rx_add_svid(priv, vid);
+
+	return -EOPNOTSUPP;
+}
+
+int mlx5e_vlan_rx_kill_vid(struct net_device *dev, __be16 proto, u16 vid)
+{
+	struct mlx5e_priv *priv = netdev_priv(dev);
+
+	if (be16_to_cpu(proto) == ETH_P_8021Q) {
+		clear_bit(vid, priv->fs.vlan.active_cvlans);
+		mlx5e_del_vlan_rule(priv, MLX5E_VLAN_RULE_TYPE_MATCH_CTAG_VID, vid);
+	} else if (be16_to_cpu(proto) == ETH_P_8021AD) {
+		clear_bit(vid, priv->fs.vlan.active_svlans);
+		mlx5e_del_vlan_rule(priv, MLX5E_VLAN_RULE_TYPE_MATCH_STAG_VID, vid);
+		netdev_update_features(dev);
+	}
+
+	return 0;
+}
+
+static void mlx5e_add_vlan_rules(struct mlx5e_priv *priv)
+{
+	int i;
+
+	mlx5e_add_vlan_rule(priv, MLX5E_VLAN_RULE_TYPE_UNTAGGED, 0);
+
+	for_each_set_bit(i, priv->fs.vlan.active_cvlans, VLAN_N_VID) {
+		mlx5e_add_vlan_rule(priv, MLX5E_VLAN_RULE_TYPE_MATCH_CTAG_VID, i);
+	}
+
+	for_each_set_bit(i, priv->fs.vlan.active_svlans, VLAN_N_VID)
+		mlx5e_add_vlan_rule(priv, MLX5E_VLAN_RULE_TYPE_MATCH_STAG_VID, i);
+
+	if (priv->fs.vlan.cvlan_filter_disabled)
+		mlx5e_add_any_vid_rules(priv);
+}
+
+static void mlx5e_del_vlan_rules(struct mlx5e_priv *priv)
+{
+	int i;
+
+	mlx5e_del_vlan_rule(priv, MLX5E_VLAN_RULE_TYPE_UNTAGGED, 0);
+
+	for_each_set_bit(i, priv->fs.vlan.active_cvlans, VLAN_N_VID) {
+		mlx5e_del_vlan_rule(priv, MLX5E_VLAN_RULE_TYPE_MATCH_CTAG_VID, i);
+	}
+
+	for_each_set_bit(i, priv->fs.vlan.active_svlans, VLAN_N_VID)
+		mlx5e_del_vlan_rule(priv, MLX5E_VLAN_RULE_TYPE_MATCH_STAG_VID, i);
+
+	WARN_ON_ONCE(!(test_bit(MLX5E_STATE_DESTROYING, &priv->state)));
+
+	/* must be called after DESTROY bit is set and
+	 * set_rx_mode is called and flushed
+	 */
+	if (priv->fs.vlan.cvlan_filter_disabled)
+		mlx5e_del_any_vid_rules(priv);
+}
+
+#define mlx5e_for_each_hash_node(hn, tmp, hash, i) \
+	for (i = 0; i < MLX5E_L2_ADDR_HASH_SIZE; i++) \
+		hlist_for_each_entry_safe(hn, tmp, &hash[i], hlist)
+
+static void mlx5e_execute_l2_action(struct mlx5e_priv *priv,
+				    struct mlx5e_l2_hash_node *hn)
+{
+	u8 action = hn->action;
+	u8 mac_addr[ETH_ALEN];
+	int l2_err = 0;
+
+	ether_addr_copy(mac_addr, hn->ai.addr);
+
+	switch (action) {
+	case MLX5E_ACTION_ADD:
+		mlx5e_add_l2_flow_rule(priv, &hn->ai, MLX5E_FULLMATCH);
+		if (!is_multicast_ether_addr(mac_addr)) {
+			l2_err = mlx5_mpfs_add_mac(priv->mdev, mac_addr);
+			hn->mpfs = !l2_err;
+		}
+		hn->action = MLX5E_ACTION_NONE;
+		break;
+
+	case MLX5E_ACTION_DEL:
+		if (!is_multicast_ether_addr(mac_addr) && hn->mpfs)
+			l2_err = mlx5_mpfs_del_mac(priv->mdev, mac_addr);
+		mlx5e_del_l2_flow_rule(priv, &hn->ai);
+		mlx5e_del_l2_from_hash(hn);
+		break;
+	}
+
+	if (l2_err)
+		netdev_warn(priv->netdev, "MPFS, failed to %s mac %pM, err(%d)\n",
+			    action == MLX5E_ACTION_ADD ? "add" : "del", mac_addr, l2_err);
+}
+
+static void mlx5e_sync_netdev_addr(struct mlx5e_priv *priv)
+{
+	struct net_device *netdev = priv->netdev;
+	struct netdev_hw_addr *ha;
+
+	netif_addr_lock_bh(netdev);
+
+	mlx5e_add_l2_to_hash(priv->fs.l2.netdev_uc,
+			     priv->netdev->dev_addr);
+
+	netdev_for_each_uc_addr(ha, netdev)
+		mlx5e_add_l2_to_hash(priv->fs.l2.netdev_uc, ha->addr);
+
+	netdev_for_each_mc_addr(ha, netdev)
+		mlx5e_add_l2_to_hash(priv->fs.l2.netdev_mc, ha->addr);
+
+	netif_addr_unlock_bh(netdev);
+}
+
+static void mlx5e_fill_addr_array(struct mlx5e_priv *priv, int list_type,
+				  u8 addr_array[][ETH_ALEN], int size)
+{
+	bool is_uc = (list_type == MLX5_NVPRT_LIST_TYPE_UC);
+	struct net_device *ndev = priv->netdev;
+	struct mlx5e_l2_hash_node *hn;
+	struct hlist_head *addr_list;
+	struct hlist_node *tmp;
+	int i = 0;
+	int hi;
+
+	addr_list = is_uc ? priv->fs.l2.netdev_uc : priv->fs.l2.netdev_mc;
+
+	if (is_uc) /* Make sure our own address is pushed first */
+		ether_addr_copy(addr_array[i++], ndev->dev_addr);
+	else if (priv->fs.l2.broadcast_enabled)
+		ether_addr_copy(addr_array[i++], ndev->broadcast);
+
+	mlx5e_for_each_hash_node(hn, tmp, addr_list, hi) {
+		if (ether_addr_equal(ndev->dev_addr, hn->ai.addr))
+			continue;
+		if (i >= size)
+			break;
+		ether_addr_copy(addr_array[i++], hn->ai.addr);
+	}
+}
+
+static void mlx5e_vport_context_update_addr_list(struct mlx5e_priv *priv,
+						 int list_type)
+{
+	bool is_uc = (list_type == MLX5_NVPRT_LIST_TYPE_UC);
+	struct mlx5e_l2_hash_node *hn;
+	u8 (*addr_array)[ETH_ALEN] = NULL;
+	struct hlist_head *addr_list;
+	struct hlist_node *tmp;
+	int max_size;
+	int size;
+	int err;
+	int hi;
+
+	size = is_uc ? 0 : (priv->fs.l2.broadcast_enabled ? 1 : 0);
+	max_size = is_uc ?
+		1 << MLX5_CAP_GEN(priv->mdev, log_max_current_uc_list) :
+		1 << MLX5_CAP_GEN(priv->mdev, log_max_current_mc_list);
+
+	addr_list = is_uc ? priv->fs.l2.netdev_uc : priv->fs.l2.netdev_mc;
+	mlx5e_for_each_hash_node(hn, tmp, addr_list, hi)
+		size++;
+
+	if (size > max_size) {
+		netdev_warn(priv->netdev,
+			    "netdev %s list size (%d) > (%d) max vport list size, some addresses will be dropped\n",
+			    is_uc ? "UC" : "MC", size, max_size);
+		size = max_size;
+	}
+
+	if (size) {
+		addr_array = kcalloc(size, ETH_ALEN, GFP_KERNEL);
+		if (!addr_array) {
+			err = -ENOMEM;
+			goto out;
+		}
+		mlx5e_fill_addr_array(priv, list_type, addr_array, size);
+	}
+
+	err = mlx5_modify_nic_vport_mac_list(priv->mdev, list_type, addr_array, size);
+out:
+	if (err)
+		netdev_err(priv->netdev,
+			   "Failed to modify vport %s list err(%d)\n",
+			   is_uc ? "UC" : "MC", err);
+	kfree(addr_array);
+}
+
+static void mlx5e_vport_context_update(struct mlx5e_priv *priv)
+{
+	struct mlx5e_l2_table *ea = &priv->fs.l2;
+
+	mlx5e_vport_context_update_addr_list(priv, MLX5_NVPRT_LIST_TYPE_UC);
+	mlx5e_vport_context_update_addr_list(priv, MLX5_NVPRT_LIST_TYPE_MC);
+	mlx5_modify_nic_vport_promisc(priv->mdev, 0,
+				      ea->allmulti_enabled,
+				      ea->promisc_enabled);
+}
+
+static void mlx5e_apply_netdev_addr(struct mlx5e_priv *priv)
+{
+	struct mlx5e_l2_hash_node *hn;
+	struct hlist_node *tmp;
+	int i;
+
+	mlx5e_for_each_hash_node(hn, tmp, priv->fs.l2.netdev_uc, i)
+		mlx5e_execute_l2_action(priv, hn);
+
+	mlx5e_for_each_hash_node(hn, tmp, priv->fs.l2.netdev_mc, i)
+		mlx5e_execute_l2_action(priv, hn);
+}
+
+static void mlx5e_handle_netdev_addr(struct mlx5e_priv *priv)
+{
+	struct mlx5e_l2_hash_node *hn;
+	struct hlist_node *tmp;
+	int i;
+
+	mlx5e_for_each_hash_node(hn, tmp, priv->fs.l2.netdev_uc, i)
+		hn->action = MLX5E_ACTION_DEL;
+	mlx5e_for_each_hash_node(hn, tmp, priv->fs.l2.netdev_mc, i)
+		hn->action = MLX5E_ACTION_DEL;
+
+	if (!test_bit(MLX5E_STATE_DESTROYING, &priv->state))
+		mlx5e_sync_netdev_addr(priv);
+
+	mlx5e_apply_netdev_addr(priv);
+}
+
+void mlx5e_set_rx_mode_work(struct work_struct *work)
+{
+	struct mlx5e_priv *priv = container_of(work, struct mlx5e_priv,
+					       set_rx_mode_work);
+
+	struct mlx5e_l2_table *ea = &priv->fs.l2;
+	struct net_device *ndev = priv->netdev;
+
+	bool rx_mode_enable   = !test_bit(MLX5E_STATE_DESTROYING, &priv->state);
+	bool promisc_enabled   = rx_mode_enable && (ndev->flags & IFF_PROMISC);
+	bool allmulti_enabled  = rx_mode_enable && (ndev->flags & IFF_ALLMULTI);
+	bool broadcast_enabled = rx_mode_enable;
+
+	bool enable_promisc    = !ea->promisc_enabled   &&  promisc_enabled;
+	bool disable_promisc   =  ea->promisc_enabled   && !promisc_enabled;
+	bool enable_allmulti   = !ea->allmulti_enabled  &&  allmulti_enabled;
+	bool disable_allmulti  =  ea->allmulti_enabled  && !allmulti_enabled;
+	bool enable_broadcast  = !ea->broadcast_enabled &&  broadcast_enabled;
+	bool disable_broadcast =  ea->broadcast_enabled && !broadcast_enabled;
+
+	if (enable_promisc) {
+		if (!priv->channels.params.vlan_strip_disable)
+			netdev_warn_once(ndev,
+					 "S-tagged traffic will be dropped while C-tag vlan stripping is enabled\n");
+		mlx5e_add_l2_flow_rule(priv, &ea->promisc, MLX5E_PROMISC);
+		if (!priv->fs.vlan.cvlan_filter_disabled)
+			mlx5e_add_any_vid_rules(priv);
+	}
+	if (enable_allmulti)
+		mlx5e_add_l2_flow_rule(priv, &ea->allmulti, MLX5E_ALLMULTI);
+	if (enable_broadcast)
+		mlx5e_add_l2_flow_rule(priv, &ea->broadcast, MLX5E_FULLMATCH);
+
+	mlx5e_handle_netdev_addr(priv);
+
+	if (disable_broadcast)
+		mlx5e_del_l2_flow_rule(priv, &ea->broadcast);
+	if (disable_allmulti)
+		mlx5e_del_l2_flow_rule(priv, &ea->allmulti);
+	if (disable_promisc) {
+		if (!priv->fs.vlan.cvlan_filter_disabled)
+			mlx5e_del_any_vid_rules(priv);
+		mlx5e_del_l2_flow_rule(priv, &ea->promisc);
+	}
+
+	ea->promisc_enabled   = promisc_enabled;
+	ea->allmulti_enabled  = allmulti_enabled;
+	ea->broadcast_enabled = broadcast_enabled;
+
+	mlx5e_vport_context_update(priv);
+}
+
+static void mlx5e_destroy_groups(struct mlx5e_flow_table *ft)
+{
+	int i;
+
+	for (i = ft->num_groups - 1; i >= 0; i--) {
+		if (!IS_ERR_OR_NULL(ft->g[i]))
+			mlx5_destroy_flow_group(ft->g[i]);
+		ft->g[i] = NULL;
+	}
+	ft->num_groups = 0;
+}
+
+void mlx5e_init_l2_addr(struct mlx5e_priv *priv)
+{
+	ether_addr_copy(priv->fs.l2.broadcast.addr, priv->netdev->broadcast);
+}
+
+void mlx5e_destroy_flow_table(struct mlx5e_flow_table *ft)
+{
+	mlx5e_destroy_groups(ft);
+	kfree(ft->g);
+	mlx5_destroy_flow_table(ft->t);
+	ft->t = NULL;
+}
+
+static void mlx5e_cleanup_ttc_rules(struct mlx5e_ttc_table *ttc)
+{
+	int i;
+
+	for (i = 0; i < MLX5E_NUM_TT; i++) {
+		if (!IS_ERR_OR_NULL(ttc->rules[i])) {
+			mlx5_del_flow_rules(ttc->rules[i]);
+			ttc->rules[i] = NULL;
+		}
+	}
+
+	for (i = 0; i < MLX5E_NUM_TUNNEL_TT; i++) {
+		if (!IS_ERR_OR_NULL(ttc->tunnel_rules[i])) {
+			mlx5_del_flow_rules(ttc->tunnel_rules[i]);
+			ttc->tunnel_rules[i] = NULL;
+		}
+	}
+}
+
+struct mlx5e_etype_proto {
+	u16 etype;
+	u8 proto;
+};
+
+static struct mlx5e_etype_proto ttc_rules[] = {
+	[MLX5E_TT_IPV4_TCP] = {
+		.etype = ETH_P_IP,
+		.proto = IPPROTO_TCP,
+	},
+	[MLX5E_TT_IPV6_TCP] = {
+		.etype = ETH_P_IPV6,
+		.proto = IPPROTO_TCP,
+	},
+	[MLX5E_TT_IPV4_UDP] = {
+		.etype = ETH_P_IP,
+		.proto = IPPROTO_UDP,
+	},
+	[MLX5E_TT_IPV6_UDP] = {
+		.etype = ETH_P_IPV6,
+		.proto = IPPROTO_UDP,
+	},
+	[MLX5E_TT_IPV4_IPSEC_AH] = {
+		.etype = ETH_P_IP,
+		.proto = IPPROTO_AH,
+	},
+	[MLX5E_TT_IPV6_IPSEC_AH] = {
+		.etype = ETH_P_IPV6,
+		.proto = IPPROTO_AH,
+	},
+	[MLX5E_TT_IPV4_IPSEC_ESP] = {
+		.etype = ETH_P_IP,
+		.proto = IPPROTO_ESP,
+	},
+	[MLX5E_TT_IPV6_IPSEC_ESP] = {
+		.etype = ETH_P_IPV6,
+		.proto = IPPROTO_ESP,
+	},
+	[MLX5E_TT_IPV4] = {
+		.etype = ETH_P_IP,
+		.proto = 0,
+	},
+	[MLX5E_TT_IPV6] = {
+		.etype = ETH_P_IPV6,
+		.proto = 0,
+	},
+	[MLX5E_TT_ANY] = {
+		.etype = 0,
+		.proto = 0,
+	},
+};
+
+static struct mlx5e_etype_proto ttc_tunnel_rules[] = {
+	[MLX5E_TT_IPV4_GRE] = {
+		.etype = ETH_P_IP,
+		.proto = IPPROTO_GRE,
+	},
+	[MLX5E_TT_IPV6_GRE] = {
+		.etype = ETH_P_IPV6,
+		.proto = IPPROTO_GRE,
+	},
+};
+
+static u8 mlx5e_etype_to_ipv(u16 ethertype)
+{
+	if (ethertype == ETH_P_IP)
+		return 4;
+
+	if (ethertype == ETH_P_IPV6)
+		return 6;
+
+	return 0;
+}
+
+static struct mlx5_flow_handle *
+mlx5e_generate_ttc_rule(struct mlx5e_priv *priv,
+			struct mlx5_flow_table *ft,
+			struct mlx5_flow_destination *dest,
+			u16 etype,
+			u8 proto)
+{
+	int match_ipv_outer = MLX5_CAP_FLOWTABLE_NIC_RX(priv->mdev, ft_field_support.outer_ip_version);
+	MLX5_DECLARE_FLOW_ACT(flow_act);
+	struct mlx5_flow_handle *rule;
+	struct mlx5_flow_spec *spec;
+	int err = 0;
+	u8 ipv;
+
+	spec = kvzalloc(sizeof(*spec), GFP_KERNEL);
+	if (!spec)
+		return ERR_PTR(-ENOMEM);
+
+	if (proto) {
+		spec->match_criteria_enable = MLX5_MATCH_OUTER_HEADERS;
+		MLX5_SET_TO_ONES(fte_match_param, spec->match_criteria, outer_headers.ip_protocol);
+		MLX5_SET(fte_match_param, spec->match_value, outer_headers.ip_protocol, proto);
+	}
+
+	ipv = mlx5e_etype_to_ipv(etype);
+	if (match_ipv_outer && ipv) {
+		spec->match_criteria_enable = MLX5_MATCH_OUTER_HEADERS;
+		MLX5_SET_TO_ONES(fte_match_param, spec->match_criteria, outer_headers.ip_version);
+		MLX5_SET(fte_match_param, spec->match_value, outer_headers.ip_version, ipv);
+	} else if (etype) {
+		spec->match_criteria_enable = MLX5_MATCH_OUTER_HEADERS;
+		MLX5_SET_TO_ONES(fte_match_param, spec->match_criteria, outer_headers.ethertype);
+		MLX5_SET(fte_match_param, spec->match_value, outer_headers.ethertype, etype);
+	}
+
+	rule = mlx5_add_flow_rules(ft, spec, &flow_act, dest, 1);
+	if (IS_ERR(rule)) {
+		err = PTR_ERR(rule);
+		netdev_err(priv->netdev, "%s: add rule failed\n", __func__);
+	}
+
+	kvfree(spec);
+	return err ? ERR_PTR(err) : rule;
+}
+
+static int mlx5e_generate_ttc_table_rules(struct mlx5e_priv *priv,
+					  struct ttc_params *params,
+					  struct mlx5e_ttc_table *ttc)
+{
+	struct mlx5_flow_destination dest = {};
+	struct mlx5_flow_handle **rules;
+	struct mlx5_flow_table *ft;
+	int tt;
+	int err;
+
+	ft = ttc->ft.t;
+	rules = ttc->rules;
+
+	dest.type = MLX5_FLOW_DESTINATION_TYPE_TIR;
+	for (tt = 0; tt < MLX5E_NUM_TT; tt++) {
+		if (tt == MLX5E_TT_ANY)
+			dest.tir_num = params->any_tt_tirn;
+		else
+			dest.tir_num = params->indir_tirn[tt];
+		rules[tt] = mlx5e_generate_ttc_rule(priv, ft, &dest,
+						    ttc_rules[tt].etype,
+						    ttc_rules[tt].proto);
+		if (IS_ERR(rules[tt]))
+			goto del_rules;
+	}
+
+	if (!params->inner_ttc || !mlx5e_tunnel_inner_ft_supported(priv->mdev))
+		return 0;
+
+	rules     = ttc->tunnel_rules;
+	dest.type = MLX5_FLOW_DESTINATION_TYPE_FLOW_TABLE;
+	dest.ft   = params->inner_ttc->ft.t;
+	for (tt = 0; tt < MLX5E_NUM_TUNNEL_TT; tt++) {
+		rules[tt] = mlx5e_generate_ttc_rule(priv, ft, &dest,
+						    ttc_tunnel_rules[tt].etype,
+						    ttc_tunnel_rules[tt].proto);
+		if (IS_ERR(rules[tt]))
+			goto del_rules;
+	}
+
+	return 0;
+
+del_rules:
+	err = PTR_ERR(rules[tt]);
+	rules[tt] = NULL;
+	mlx5e_cleanup_ttc_rules(ttc);
+	return err;
+}
+
+#define MLX5E_TTC_NUM_GROUPS	3
+#define MLX5E_TTC_GROUP1_SIZE	(BIT(3) + MLX5E_NUM_TUNNEL_TT)
+#define MLX5E_TTC_GROUP2_SIZE	 BIT(1)
+#define MLX5E_TTC_GROUP3_SIZE	 BIT(0)
+#define MLX5E_TTC_TABLE_SIZE	(MLX5E_TTC_GROUP1_SIZE +\
+				 MLX5E_TTC_GROUP2_SIZE +\
+				 MLX5E_TTC_GROUP3_SIZE)
+
+#define MLX5E_INNER_TTC_NUM_GROUPS	3
+#define MLX5E_INNER_TTC_GROUP1_SIZE	BIT(3)
+#define MLX5E_INNER_TTC_GROUP2_SIZE	BIT(1)
+#define MLX5E_INNER_TTC_GROUP3_SIZE	BIT(0)
+#define MLX5E_INNER_TTC_TABLE_SIZE	(MLX5E_INNER_TTC_GROUP1_SIZE +\
+					 MLX5E_INNER_TTC_GROUP2_SIZE +\
+					 MLX5E_INNER_TTC_GROUP3_SIZE)
+
+static int mlx5e_create_ttc_table_groups(struct mlx5e_ttc_table *ttc,
+					 bool use_ipv)
+{
+	int inlen = MLX5_ST_SZ_BYTES(create_flow_group_in);
+	struct mlx5e_flow_table *ft = &ttc->ft;
+	int ix = 0;
+	u32 *in;
+	int err;
+	u8 *mc;
+
+	ft->g = kcalloc(MLX5E_TTC_NUM_GROUPS,
+			sizeof(*ft->g), GFP_KERNEL);
+	if (!ft->g)
+		return -ENOMEM;
+	in = kvzalloc(inlen, GFP_KERNEL);
+	if (!in) {
+		kfree(ft->g);
+		ft->g = NULL;
+		return -ENOMEM;
+	}
+
+	/* L4 Group */
+	mc = MLX5_ADDR_OF(create_flow_group_in, in, match_criteria);
+	MLX5_SET_TO_ONES(fte_match_param, mc, outer_headers.ip_protocol);
+	if (use_ipv)
+		MLX5_SET_TO_ONES(fte_match_param, mc, outer_headers.ip_version);
+	else
+		MLX5_SET_TO_ONES(fte_match_param, mc, outer_headers.ethertype);
+	MLX5_SET_CFG(in, match_criteria_enable, MLX5_MATCH_OUTER_HEADERS);
+	MLX5_SET_CFG(in, start_flow_index, ix);
+	ix += MLX5E_TTC_GROUP1_SIZE;
+	MLX5_SET_CFG(in, end_flow_index, ix - 1);
+	ft->g[ft->num_groups] = mlx5_create_flow_group(ft->t, in);
+	if (IS_ERR(ft->g[ft->num_groups]))
+		goto err;
+	ft->num_groups++;
+
+	/* L3 Group */
+	MLX5_SET(fte_match_param, mc, outer_headers.ip_protocol, 0);
+	MLX5_SET_CFG(in, start_flow_index, ix);
+	ix += MLX5E_TTC_GROUP2_SIZE;
+	MLX5_SET_CFG(in, end_flow_index, ix - 1);
+	ft->g[ft->num_groups] = mlx5_create_flow_group(ft->t, in);
+	if (IS_ERR(ft->g[ft->num_groups]))
+		goto err;
+	ft->num_groups++;
+
+	/* Any Group */
+	memset(in, 0, inlen);
+	MLX5_SET_CFG(in, start_flow_index, ix);
+	ix += MLX5E_TTC_GROUP3_SIZE;
+	MLX5_SET_CFG(in, end_flow_index, ix - 1);
+	ft->g[ft->num_groups] = mlx5_create_flow_group(ft->t, in);
+	if (IS_ERR(ft->g[ft->num_groups]))
+		goto err;
+	ft->num_groups++;
+
+	kvfree(in);
+	return 0;
+
+err:
+	err = PTR_ERR(ft->g[ft->num_groups]);
+	ft->g[ft->num_groups] = NULL;
+	kvfree(in);
+
+	return err;
+}
+
+static struct mlx5_flow_handle *
+mlx5e_generate_inner_ttc_rule(struct mlx5e_priv *priv,
+			      struct mlx5_flow_table *ft,
+			      struct mlx5_flow_destination *dest,
+			      u16 etype, u8 proto)
+{
+	MLX5_DECLARE_FLOW_ACT(flow_act);
+	struct mlx5_flow_handle *rule;
+	struct mlx5_flow_spec *spec;
+	int err = 0;
+	u8 ipv;
+
+	spec = kvzalloc(sizeof(*spec), GFP_KERNEL);
+	if (!spec)
+		return ERR_PTR(-ENOMEM);
+
+	ipv = mlx5e_etype_to_ipv(etype);
+	if (etype && ipv) {
+		spec->match_criteria_enable = MLX5_MATCH_INNER_HEADERS;
+		MLX5_SET_TO_ONES(fte_match_param, spec->match_criteria, inner_headers.ip_version);
+		MLX5_SET(fte_match_param, spec->match_value, inner_headers.ip_version, ipv);
+	}
+
+	if (proto) {
+		spec->match_criteria_enable = MLX5_MATCH_INNER_HEADERS;
+		MLX5_SET_TO_ONES(fte_match_param, spec->match_criteria, inner_headers.ip_protocol);
+		MLX5_SET(fte_match_param, spec->match_value, inner_headers.ip_protocol, proto);
+	}
+
+	rule = mlx5_add_flow_rules(ft, spec, &flow_act, dest, 1);
+	if (IS_ERR(rule)) {
+		err = PTR_ERR(rule);
+		netdev_err(priv->netdev, "%s: add rule failed\n", __func__);
+	}
+
+	kvfree(spec);
+	return err ? ERR_PTR(err) : rule;
+}
+
+static int mlx5e_generate_inner_ttc_table_rules(struct mlx5e_priv *priv,
+						struct ttc_params *params,
+						struct mlx5e_ttc_table *ttc)
+{
+	struct mlx5_flow_destination dest = {};
+	struct mlx5_flow_handle **rules;
+	struct mlx5_flow_table *ft;
+	int err;
+	int tt;
+
+	ft = ttc->ft.t;
+	rules = ttc->rules;
+
+	dest.type = MLX5_FLOW_DESTINATION_TYPE_TIR;
+	for (tt = 0; tt < MLX5E_NUM_TT; tt++) {
+		if (tt == MLX5E_TT_ANY)
+			dest.tir_num = params->any_tt_tirn;
+		else
+			dest.tir_num = params->indir_tirn[tt];
+
+		rules[tt] = mlx5e_generate_inner_ttc_rule(priv, ft, &dest,
+							  ttc_rules[tt].etype,
+							  ttc_rules[tt].proto);
+		if (IS_ERR(rules[tt]))
+			goto del_rules;
+	}
+
+	return 0;
+
+del_rules:
+	err = PTR_ERR(rules[tt]);
+	rules[tt] = NULL;
+	mlx5e_cleanup_ttc_rules(ttc);
+	return err;
+}
+
+static int mlx5e_create_inner_ttc_table_groups(struct mlx5e_ttc_table *ttc)
+{
+	int inlen = MLX5_ST_SZ_BYTES(create_flow_group_in);
+	struct mlx5e_flow_table *ft = &ttc->ft;
+	int ix = 0;
+	u32 *in;
+	int err;
+	u8 *mc;
+
+	ft->g = kcalloc(MLX5E_INNER_TTC_NUM_GROUPS, sizeof(*ft->g), GFP_KERNEL);
+	if (!ft->g)
+		return -ENOMEM;
+	in = kvzalloc(inlen, GFP_KERNEL);
+	if (!in) {
+		kfree(ft->g);
+		ft->g = NULL;
+		return -ENOMEM;
+	}
+
+	/* L4 Group */
+	mc = MLX5_ADDR_OF(create_flow_group_in, in, match_criteria);
+	MLX5_SET_TO_ONES(fte_match_param, mc, inner_headers.ip_protocol);
+	MLX5_SET_TO_ONES(fte_match_param, mc, inner_headers.ip_version);
+	MLX5_SET_CFG(in, match_criteria_enable, MLX5_MATCH_INNER_HEADERS);
+	MLX5_SET_CFG(in, start_flow_index, ix);
+	ix += MLX5E_INNER_TTC_GROUP1_SIZE;
+	MLX5_SET_CFG(in, end_flow_index, ix - 1);
+	ft->g[ft->num_groups] = mlx5_create_flow_group(ft->t, in);
+	if (IS_ERR(ft->g[ft->num_groups]))
+		goto err;
+	ft->num_groups++;
+
+	/* L3 Group */
+	MLX5_SET(fte_match_param, mc, inner_headers.ip_protocol, 0);
+	MLX5_SET_CFG(in, start_flow_index, ix);
+	ix += MLX5E_INNER_TTC_GROUP2_SIZE;
+	MLX5_SET_CFG(in, end_flow_index, ix - 1);
+	ft->g[ft->num_groups] = mlx5_create_flow_group(ft->t, in);
+	if (IS_ERR(ft->g[ft->num_groups]))
+		goto err;
+	ft->num_groups++;
+
+	/* Any Group */
+	memset(in, 0, inlen);
+	MLX5_SET_CFG(in, start_flow_index, ix);
+	ix += MLX5E_INNER_TTC_GROUP3_SIZE;
+	MLX5_SET_CFG(in, end_flow_index, ix - 1);
+	ft->g[ft->num_groups] = mlx5_create_flow_group(ft->t, in);
+	if (IS_ERR(ft->g[ft->num_groups]))
+		goto err;
+	ft->num_groups++;
+
+	kvfree(in);
+	return 0;
+
+err:
+	err = PTR_ERR(ft->g[ft->num_groups]);
+	ft->g[ft->num_groups] = NULL;
+	kvfree(in);
+
+	return err;
+}
+
+void mlx5e_set_ttc_basic_params(struct mlx5e_priv *priv,
+				struct ttc_params *ttc_params)
+{
+	ttc_params->any_tt_tirn = priv->direct_tir[0].tirn;
+	ttc_params->inner_ttc = &priv->fs.inner_ttc;
+}
+
+void mlx5e_set_inner_ttc_ft_params(struct ttc_params *ttc_params)
+{
+	struct mlx5_flow_table_attr *ft_attr = &ttc_params->ft_attr;
+
+	ft_attr->max_fte = MLX5E_INNER_TTC_TABLE_SIZE;
+	ft_attr->level = MLX5E_INNER_TTC_FT_LEVEL;
+	ft_attr->prio = MLX5E_NIC_PRIO;
+}
+
+void mlx5e_set_ttc_ft_params(struct ttc_params *ttc_params)
+
+{
+	struct mlx5_flow_table_attr *ft_attr = &ttc_params->ft_attr;
+
+	ft_attr->max_fte = MLX5E_TTC_TABLE_SIZE;
+	ft_attr->level = MLX5E_TTC_FT_LEVEL;
+	ft_attr->prio = MLX5E_NIC_PRIO;
+}
+
+int mlx5e_create_inner_ttc_table(struct mlx5e_priv *priv, struct ttc_params *params,
+				 struct mlx5e_ttc_table *ttc)
+{
+	struct mlx5e_flow_table *ft = &ttc->ft;
+	int err;
+
+	if (!mlx5e_tunnel_inner_ft_supported(priv->mdev))
+		return 0;
+
+	ft->t = mlx5_create_flow_table(priv->fs.ns, &params->ft_attr);
+	if (IS_ERR(ft->t)) {
+		err = PTR_ERR(ft->t);
+		ft->t = NULL;
+		return err;
+	}
+
+	err = mlx5e_create_inner_ttc_table_groups(ttc);
+	if (err)
+		goto err;
+
+	err = mlx5e_generate_inner_ttc_table_rules(priv, params, ttc);
+	if (err)
+		goto err;
+
+	return 0;
+
+err:
+	mlx5e_destroy_flow_table(ft);
+	return err;
+}
+
+void mlx5e_destroy_inner_ttc_table(struct mlx5e_priv *priv,
+				   struct mlx5e_ttc_table *ttc)
+{
+	if (!mlx5e_tunnel_inner_ft_supported(priv->mdev))
+		return;
+
+	mlx5e_cleanup_ttc_rules(ttc);
+	mlx5e_destroy_flow_table(&ttc->ft);
+}
+
+void mlx5e_destroy_ttc_table(struct mlx5e_priv *priv,
+			     struct mlx5e_ttc_table *ttc)
+{
+	mlx5e_cleanup_ttc_rules(ttc);
+	mlx5e_destroy_flow_table(&ttc->ft);
+}
+
+int mlx5e_create_ttc_table(struct mlx5e_priv *priv, struct ttc_params *params,
+			   struct mlx5e_ttc_table *ttc)
+{
+	bool match_ipv_outer = MLX5_CAP_FLOWTABLE_NIC_RX(priv->mdev, ft_field_support.outer_ip_version);
+	struct mlx5e_flow_table *ft = &ttc->ft;
+	int err;
+
+	ft->t = mlx5_create_flow_table(priv->fs.ns, &params->ft_attr);
+	if (IS_ERR(ft->t)) {
+		err = PTR_ERR(ft->t);
+		ft->t = NULL;
+		return err;
+	}
+
+	err = mlx5e_create_ttc_table_groups(ttc, match_ipv_outer);
+	if (err)
+		goto err;
+
+	err = mlx5e_generate_ttc_table_rules(priv, params, ttc);
+	if (err)
+		goto err;
+
+	return 0;
+err:
+	mlx5e_destroy_flow_table(ft);
+	return err;
+}
+
+static void mlx5e_del_l2_flow_rule(struct mlx5e_priv *priv,
+				   struct mlx5e_l2_rule *ai)
+{
+	if (!IS_ERR_OR_NULL(ai->rule)) {
+		mlx5_del_flow_rules(ai->rule);
+		ai->rule = NULL;
+	}
+}
+
+static int mlx5e_add_l2_flow_rule(struct mlx5e_priv *priv,
+				  struct mlx5e_l2_rule *ai, int type)
+{
+	struct mlx5_flow_table *ft = priv->fs.l2.ft.t;
+	struct mlx5_flow_destination dest = {};
+	MLX5_DECLARE_FLOW_ACT(flow_act);
+	struct mlx5_flow_spec *spec;
+	int err = 0;
+	u8 *mc_dmac;
+	u8 *mv_dmac;
+
+	spec = kvzalloc(sizeof(*spec), GFP_KERNEL);
+	if (!spec)
+		return -ENOMEM;
+
+	mc_dmac = MLX5_ADDR_OF(fte_match_param, spec->match_criteria,
+			       outer_headers.dmac_47_16);
+	mv_dmac = MLX5_ADDR_OF(fte_match_param, spec->match_value,
+			       outer_headers.dmac_47_16);
+
+	dest.type = MLX5_FLOW_DESTINATION_TYPE_FLOW_TABLE;
+	dest.ft = priv->fs.ttc.ft.t;
+
+	switch (type) {
+	case MLX5E_FULLMATCH:
+		spec->match_criteria_enable = MLX5_MATCH_OUTER_HEADERS;
+		eth_broadcast_addr(mc_dmac);
+		ether_addr_copy(mv_dmac, ai->addr);
+		break;
+
+	case MLX5E_ALLMULTI:
+		spec->match_criteria_enable = MLX5_MATCH_OUTER_HEADERS;
+		mc_dmac[0] = 0x01;
+		mv_dmac[0] = 0x01;
+		break;
+
+	case MLX5E_PROMISC:
+		break;
+	}
+
+	ai->rule = mlx5_add_flow_rules(ft, spec, &flow_act, &dest, 1);
+	if (IS_ERR(ai->rule)) {
+		netdev_err(priv->netdev, "%s: add l2 rule(mac:%pM) failed\n",
+			   __func__, mv_dmac);
+		err = PTR_ERR(ai->rule);
+		ai->rule = NULL;
+	}
+
+	kvfree(spec);
+
+	return err;
+}
+
+#define MLX5E_NUM_L2_GROUPS	   3
+#define MLX5E_L2_GROUP1_SIZE	   BIT(0)
+#define MLX5E_L2_GROUP2_SIZE	   BIT(15)
+#define MLX5E_L2_GROUP3_SIZE	   BIT(0)
+#define MLX5E_L2_TABLE_SIZE	   (MLX5E_L2_GROUP1_SIZE +\
+				    MLX5E_L2_GROUP2_SIZE +\
+				    MLX5E_L2_GROUP3_SIZE)
+static int mlx5e_create_l2_table_groups(struct mlx5e_l2_table *l2_table)
+{
+	int inlen = MLX5_ST_SZ_BYTES(create_flow_group_in);
+	struct mlx5e_flow_table *ft = &l2_table->ft;
+	int ix = 0;
+	u8 *mc_dmac;
+	u32 *in;
+	int err;
+	u8 *mc;
+
+	ft->g = kcalloc(MLX5E_NUM_L2_GROUPS, sizeof(*ft->g), GFP_KERNEL);
+	if (!ft->g)
+		return -ENOMEM;
+	in = kvzalloc(inlen, GFP_KERNEL);
+	if (!in) {
+		kfree(ft->g);
+		return -ENOMEM;
+	}
+
+	mc = MLX5_ADDR_OF(create_flow_group_in, in, match_criteria);
+	mc_dmac = MLX5_ADDR_OF(fte_match_param, mc,
+			       outer_headers.dmac_47_16);
+	/* Flow Group for promiscuous */
+	MLX5_SET_CFG(in, start_flow_index, ix);
+	ix += MLX5E_L2_GROUP1_SIZE;
+	MLX5_SET_CFG(in, end_flow_index, ix - 1);
+	ft->g[ft->num_groups] = mlx5_create_flow_group(ft->t, in);
+	if (IS_ERR(ft->g[ft->num_groups]))
+		goto err_destroy_groups;
+	ft->num_groups++;
+
+	/* Flow Group for full match */
+	eth_broadcast_addr(mc_dmac);
+	MLX5_SET_CFG(in, match_criteria_enable, MLX5_MATCH_OUTER_HEADERS);
+	MLX5_SET_CFG(in, start_flow_index, ix);
+	ix += MLX5E_L2_GROUP2_SIZE;
+	MLX5_SET_CFG(in, end_flow_index, ix - 1);
+	ft->g[ft->num_groups] = mlx5_create_flow_group(ft->t, in);
+	if (IS_ERR(ft->g[ft->num_groups]))
+		goto err_destroy_groups;
+	ft->num_groups++;
+
+	/* Flow Group for allmulti */
+	eth_zero_addr(mc_dmac);
+	mc_dmac[0] = 0x01;
+	MLX5_SET_CFG(in, start_flow_index, ix);
+	ix += MLX5E_L2_GROUP3_SIZE;
+	MLX5_SET_CFG(in, end_flow_index, ix - 1);
+	ft->g[ft->num_groups] = mlx5_create_flow_group(ft->t, in);
+	if (IS_ERR(ft->g[ft->num_groups]))
+		goto err_destroy_groups;
+	ft->num_groups++;
+
+	kvfree(in);
+	return 0;
+
+err_destroy_groups:
+	err = PTR_ERR(ft->g[ft->num_groups]);
+	ft->g[ft->num_groups] = NULL;
+	mlx5e_destroy_groups(ft);
+	kvfree(in);
+	kfree(ft->g);
+
+	return err;
+}
+
+static void mlx5e_destroy_l2_table(struct mlx5e_priv *priv)
+{
+	mlx5e_destroy_flow_table(&priv->fs.l2.ft);
+}
+
+static int mlx5e_create_l2_table(struct mlx5e_priv *priv)
+{
+	struct mlx5e_l2_table *l2_table = &priv->fs.l2;
+	struct mlx5e_flow_table *ft = &l2_table->ft;
+	struct mlx5_flow_table_attr ft_attr = {};
+	int err;
+
+	ft->num_groups = 0;
+
+	ft_attr.max_fte = MLX5E_L2_TABLE_SIZE;
+	ft_attr.level = MLX5E_L2_FT_LEVEL;
+	ft_attr.prio = MLX5E_NIC_PRIO;
+
+	ft->t = mlx5_create_flow_table(priv->fs.ns, &ft_attr);
+	if (IS_ERR(ft->t)) {
+		err = PTR_ERR(ft->t);
+		ft->t = NULL;
+		return err;
+	}
+
+	err = mlx5e_create_l2_table_groups(l2_table);
+	if (err)
+		goto err_destroy_flow_table;
+
+	return 0;
+
+err_destroy_flow_table:
+	mlx5_destroy_flow_table(ft->t);
+	ft->t = NULL;
+
+	return err;
+}
+
+#define MLX5E_NUM_VLAN_GROUPS	4
+#define MLX5E_VLAN_GROUP0_SIZE	BIT(12)
+#define MLX5E_VLAN_GROUP1_SIZE	BIT(12)
+#define MLX5E_VLAN_GROUP2_SIZE	BIT(1)
+#define MLX5E_VLAN_GROUP3_SIZE	BIT(0)
+#define MLX5E_VLAN_TABLE_SIZE	(MLX5E_VLAN_GROUP0_SIZE +\
+				 MLX5E_VLAN_GROUP1_SIZE +\
+				 MLX5E_VLAN_GROUP2_SIZE +\
+				 MLX5E_VLAN_GROUP3_SIZE)
+
+static int __mlx5e_create_vlan_table_groups(struct mlx5e_flow_table *ft, u32 *in,
+					    int inlen)
+{
+	int err;
+	int ix = 0;
+	u8 *mc = MLX5_ADDR_OF(create_flow_group_in, in, match_criteria);
+
+	memset(in, 0, inlen);
+	MLX5_SET_CFG(in, match_criteria_enable, MLX5_MATCH_OUTER_HEADERS);
+	MLX5_SET_TO_ONES(fte_match_param, mc, outer_headers.cvlan_tag);
+	MLX5_SET_TO_ONES(fte_match_param, mc, outer_headers.first_vid);
+	MLX5_SET_CFG(in, start_flow_index, ix);
+	ix += MLX5E_VLAN_GROUP0_SIZE;
+	MLX5_SET_CFG(in, end_flow_index, ix - 1);
+	ft->g[ft->num_groups] = mlx5_create_flow_group(ft->t, in);
+	if (IS_ERR(ft->g[ft->num_groups]))
+		goto err_destroy_groups;
+	ft->num_groups++;
+
+	memset(in, 0, inlen);
+	MLX5_SET_CFG(in, match_criteria_enable, MLX5_MATCH_OUTER_HEADERS);
+	MLX5_SET_TO_ONES(fte_match_param, mc, outer_headers.svlan_tag);
+	MLX5_SET_TO_ONES(fte_match_param, mc, outer_headers.first_vid);
+	MLX5_SET_CFG(in, start_flow_index, ix);
+	ix += MLX5E_VLAN_GROUP1_SIZE;
+	MLX5_SET_CFG(in, end_flow_index, ix - 1);
+	ft->g[ft->num_groups] = mlx5_create_flow_group(ft->t, in);
+	if (IS_ERR(ft->g[ft->num_groups]))
+		goto err_destroy_groups;
+	ft->num_groups++;
+
+	memset(in, 0, inlen);
+	MLX5_SET_CFG(in, match_criteria_enable, MLX5_MATCH_OUTER_HEADERS);
+	MLX5_SET_TO_ONES(fte_match_param, mc, outer_headers.cvlan_tag);
+	MLX5_SET_CFG(in, start_flow_index, ix);
+	ix += MLX5E_VLAN_GROUP2_SIZE;
+	MLX5_SET_CFG(in, end_flow_index, ix - 1);
+	ft->g[ft->num_groups] = mlx5_create_flow_group(ft->t, in);
+	if (IS_ERR(ft->g[ft->num_groups]))
+		goto err_destroy_groups;
+	ft->num_groups++;
+
+	memset(in, 0, inlen);
+	MLX5_SET_CFG(in, match_criteria_enable, MLX5_MATCH_OUTER_HEADERS);
+	MLX5_SET_TO_ONES(fte_match_param, mc, outer_headers.svlan_tag);
+	MLX5_SET_CFG(in, start_flow_index, ix);
+	ix += MLX5E_VLAN_GROUP3_SIZE;
+	MLX5_SET_CFG(in, end_flow_index, ix - 1);
+	ft->g[ft->num_groups] = mlx5_create_flow_group(ft->t, in);
+	if (IS_ERR(ft->g[ft->num_groups]))
+		goto err_destroy_groups;
+	ft->num_groups++;
+
+	return 0;
+
+err_destroy_groups:
+	err = PTR_ERR(ft->g[ft->num_groups]);
+	ft->g[ft->num_groups] = NULL;
+	mlx5e_destroy_groups(ft);
+
+	return err;
+}
+
+static int mlx5e_create_vlan_table_groups(struct mlx5e_flow_table *ft)
+{
+	u32 *in;
+	int inlen = MLX5_ST_SZ_BYTES(create_flow_group_in);
+	int err;
+
+	in = kvzalloc(inlen, GFP_KERNEL);
+	if (!in)
+		return -ENOMEM;
+
+	err = __mlx5e_create_vlan_table_groups(ft, in, inlen);
+
+	kvfree(in);
+	return err;
+}
+
+static int mlx5e_create_vlan_table(struct mlx5e_priv *priv)
+{
+	struct mlx5e_flow_table *ft = &priv->fs.vlan.ft;
+	struct mlx5_flow_table_attr ft_attr = {};
+	int err;
+
+	ft->num_groups = 0;
+
+	ft_attr.max_fte = MLX5E_VLAN_TABLE_SIZE;
+	ft_attr.level = MLX5E_VLAN_FT_LEVEL;
+	ft_attr.prio = MLX5E_NIC_PRIO;
+
+	ft->t = mlx5_create_flow_table(priv->fs.ns, &ft_attr);
+
+	if (IS_ERR(ft->t)) {
+		err = PTR_ERR(ft->t);
+		ft->t = NULL;
+		return err;
+	}
+	ft->g = kcalloc(MLX5E_NUM_VLAN_GROUPS, sizeof(*ft->g), GFP_KERNEL);
+	if (!ft->g) {
+		err = -ENOMEM;
+		goto err_destroy_vlan_table;
+	}
+
+	err = mlx5e_create_vlan_table_groups(ft);
+	if (err)
+		goto err_free_g;
+
+	mlx5e_add_vlan_rules(priv);
+
+	return 0;
+
+err_free_g:
+	kfree(ft->g);
+err_destroy_vlan_table:
+	mlx5_destroy_flow_table(ft->t);
+	ft->t = NULL;
+
+	return err;
+}
+
+static void mlx5e_destroy_vlan_table(struct mlx5e_priv *priv)
+{
+	mlx5e_del_vlan_rules(priv);
+	mlx5e_destroy_flow_table(&priv->fs.vlan.ft);
+}
+
+int mlx5e_create_flow_steering(struct mlx5e_priv *priv)
+{
+	struct ttc_params ttc_params = {};
+	int tt, err;
+
+	priv->fs.ns = mlx5_get_flow_namespace(priv->mdev,
+					       MLX5_FLOW_NAMESPACE_KERNEL);
+
+	if (!priv->fs.ns)
+		return -EOPNOTSUPP;
+
+	err = mlx5e_arfs_create_tables(priv);
+	if (err) {
+		netdev_err(priv->netdev, "Failed to create arfs tables, err=%d\n",
+			   err);
+		priv->netdev->hw_features &= ~NETIF_F_NTUPLE;
+	}
+
+	mlx5e_set_ttc_basic_params(priv, &ttc_params);
+	mlx5e_set_inner_ttc_ft_params(&ttc_params);
+	for (tt = 0; tt < MLX5E_NUM_INDIR_TIRS; tt++)
+		ttc_params.indir_tirn[tt] = priv->inner_indir_tir[tt].tirn;
+
+	err = mlx5e_create_inner_ttc_table(priv, &ttc_params, &priv->fs.inner_ttc);
+	if (err) {
+		netdev_err(priv->netdev, "Failed to create inner ttc table, err=%d\n",
+			   err);
+		goto err_destroy_arfs_tables;
+	}
+
+	mlx5e_set_ttc_ft_params(&ttc_params);
+	for (tt = 0; tt < MLX5E_NUM_INDIR_TIRS; tt++)
+		ttc_params.indir_tirn[tt] = priv->indir_tir[tt].tirn;
+
+	err = mlx5e_create_ttc_table(priv, &ttc_params, &priv->fs.ttc);
+	if (err) {
+		netdev_err(priv->netdev, "Failed to create ttc table, err=%d\n",
+			   err);
+		goto err_destroy_inner_ttc_table;
+	}
+
+	err = mlx5e_create_l2_table(priv);
+	if (err) {
+		netdev_err(priv->netdev, "Failed to create l2 table, err=%d\n",
+			   err);
+		goto err_destroy_ttc_table;
+	}
+
+	err = mlx5e_create_vlan_table(priv);
+	if (err) {
+		netdev_err(priv->netdev, "Failed to create vlan table, err=%d\n",
+			   err);
+		goto err_destroy_l2_table;
+	}
+
+	mlx5e_ethtool_init_steering(priv);
+
+	return 0;
+
+err_destroy_l2_table:
+	mlx5e_destroy_l2_table(priv);
+err_destroy_ttc_table:
+	mlx5e_destroy_ttc_table(priv, &priv->fs.ttc);
+err_destroy_inner_ttc_table:
+	mlx5e_destroy_inner_ttc_table(priv, &priv->fs.inner_ttc);
+err_destroy_arfs_tables:
+	mlx5e_arfs_destroy_tables(priv);
+
+	return err;
+}
+
+void mlx5e_destroy_flow_steering(struct mlx5e_priv *priv)
+{
+	mlx5e_destroy_vlan_table(priv);
+	mlx5e_destroy_l2_table(priv);
+	mlx5e_destroy_ttc_table(priv, &priv->fs.ttc);
+	mlx5e_destroy_inner_ttc_table(priv, &priv->fs.inner_ttc);
+	mlx5e_arfs_destroy_tables(priv);
+	mlx5e_ethtool_cleanup_steering(priv);
+}
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_fs_ethtool.c b/drivers/net/ethernet/mellanox/mlx5/core/en_fs_ethtool.c
new file mode 100644
index 000000000..41cde926c
--- /dev/null
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_fs_ethtool.c
@@ -0,0 +1,844 @@
+/*
+ * Copyright (c) 2016, 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/mlx5/fs.h>
+#include "en.h"
+
+struct mlx5e_ethtool_rule {
+	struct list_head             list;
+	struct ethtool_rx_flow_spec  flow_spec;
+	struct mlx5_flow_handle	     *rule;
+	struct mlx5e_ethtool_table   *eth_ft;
+};
+
+static void put_flow_table(struct mlx5e_ethtool_table *eth_ft)
+{
+	if (!--eth_ft->num_rules) {
+		mlx5_destroy_flow_table(eth_ft->ft);
+		eth_ft->ft = NULL;
+	}
+}
+
+#define MLX5E_ETHTOOL_L3_L4_PRIO 0
+#define MLX5E_ETHTOOL_L2_PRIO (MLX5E_ETHTOOL_L3_L4_PRIO + ETHTOOL_NUM_L3_L4_FTS)
+#define MLX5E_ETHTOOL_NUM_ENTRIES 64000
+#define MLX5E_ETHTOOL_NUM_GROUPS  10
+static struct mlx5e_ethtool_table *get_flow_table(struct mlx5e_priv *priv,
+						  struct ethtool_rx_flow_spec *fs,
+						  int num_tuples)
+{
+	struct mlx5e_ethtool_table *eth_ft;
+	struct mlx5_flow_namespace *ns;
+	struct mlx5_flow_table *ft;
+	int max_tuples;
+	int table_size;
+	int prio;
+
+	switch (fs->flow_type & ~(FLOW_EXT | FLOW_MAC_EXT)) {
+	case TCP_V4_FLOW:
+	case UDP_V4_FLOW:
+	case TCP_V6_FLOW:
+	case UDP_V6_FLOW:
+		max_tuples = ETHTOOL_NUM_L3_L4_FTS;
+		prio = MLX5E_ETHTOOL_L3_L4_PRIO + (max_tuples - num_tuples);
+		eth_ft = &priv->fs.ethtool.l3_l4_ft[prio];
+		break;
+	case IP_USER_FLOW:
+	case IPV6_USER_FLOW:
+		max_tuples = ETHTOOL_NUM_L3_L4_FTS;
+		prio = MLX5E_ETHTOOL_L3_L4_PRIO + (max_tuples - num_tuples);
+		eth_ft = &priv->fs.ethtool.l3_l4_ft[prio];
+		break;
+	case ETHER_FLOW:
+		max_tuples = ETHTOOL_NUM_L2_FTS;
+		prio = max_tuples - num_tuples;
+		eth_ft = &priv->fs.ethtool.l2_ft[prio];
+		prio += MLX5E_ETHTOOL_L2_PRIO;
+		break;
+	default:
+		return ERR_PTR(-EINVAL);
+	}
+
+	eth_ft->num_rules++;
+	if (eth_ft->ft)
+		return eth_ft;
+
+	ns = mlx5_get_flow_namespace(priv->mdev,
+				     MLX5_FLOW_NAMESPACE_ETHTOOL);
+	if (!ns)
+		return ERR_PTR(-EOPNOTSUPP);
+
+	table_size = min_t(u32, BIT(MLX5_CAP_FLOWTABLE(priv->mdev,
+						       flow_table_properties_nic_receive.log_max_ft_size)),
+			   MLX5E_ETHTOOL_NUM_ENTRIES);
+	ft = mlx5_create_auto_grouped_flow_table(ns, prio,
+						 table_size,
+						 MLX5E_ETHTOOL_NUM_GROUPS, 0, 0);
+	if (IS_ERR(ft))
+		return (void *)ft;
+
+	eth_ft->ft = ft;
+	return eth_ft;
+}
+
+static void mask_spec(u8 *mask, u8 *val, size_t size)
+{
+	unsigned int i;
+
+	for (i = 0; i < size; i++, mask++, val++)
+		*((u8 *)val) = *((u8 *)mask) & *((u8 *)val);
+}
+
+#define MLX5E_FTE_SET(header_p, fld, v)  \
+	MLX5_SET(fte_match_set_lyr_2_4, header_p, fld, v)
+
+#define MLX5E_FTE_ADDR_OF(header_p, fld) \
+	MLX5_ADDR_OF(fte_match_set_lyr_2_4, header_p, fld)
+
+static void
+set_ip4(void *headers_c, void *headers_v, __be32 ip4src_m,
+	__be32 ip4src_v, __be32 ip4dst_m, __be32 ip4dst_v)
+{
+	if (ip4src_m) {
+		memcpy(MLX5E_FTE_ADDR_OF(headers_v, src_ipv4_src_ipv6.ipv4_layout.ipv4),
+		       &ip4src_v, sizeof(ip4src_v));
+		memset(MLX5E_FTE_ADDR_OF(headers_c, src_ipv4_src_ipv6.ipv4_layout.ipv4),
+		       0xff, sizeof(ip4src_m));
+	}
+	if (ip4dst_m) {
+		memcpy(MLX5E_FTE_ADDR_OF(headers_v, dst_ipv4_dst_ipv6.ipv4_layout.ipv4),
+		       &ip4dst_v, sizeof(ip4dst_v));
+		memset(MLX5E_FTE_ADDR_OF(headers_c, dst_ipv4_dst_ipv6.ipv4_layout.ipv4),
+		       0xff, sizeof(ip4dst_m));
+	}
+
+	MLX5E_FTE_SET(headers_c, ethertype, 0xffff);
+	MLX5E_FTE_SET(headers_v, ethertype, ETH_P_IP);
+}
+
+static void
+set_ip6(void *headers_c, void *headers_v, __be32 ip6src_m[4],
+	__be32 ip6src_v[4], __be32 ip6dst_m[4], __be32 ip6dst_v[4])
+{
+	u8 ip6_sz = MLX5_FLD_SZ_BYTES(ipv6_layout, ipv6);
+
+	if (!ipv6_addr_any((struct in6_addr *)ip6src_m)) {
+		memcpy(MLX5E_FTE_ADDR_OF(headers_v, src_ipv4_src_ipv6.ipv6_layout.ipv6),
+		       ip6src_v, ip6_sz);
+		memcpy(MLX5E_FTE_ADDR_OF(headers_c, src_ipv4_src_ipv6.ipv6_layout.ipv6),
+		       ip6src_m, ip6_sz);
+	}
+	if (!ipv6_addr_any((struct in6_addr *)ip6dst_m)) {
+		memcpy(MLX5E_FTE_ADDR_OF(headers_v, dst_ipv4_dst_ipv6.ipv6_layout.ipv6),
+		       ip6dst_v, ip6_sz);
+		memcpy(MLX5E_FTE_ADDR_OF(headers_c, dst_ipv4_dst_ipv6.ipv6_layout.ipv6),
+		       ip6dst_m, ip6_sz);
+	}
+
+	MLX5E_FTE_SET(headers_c, ethertype, 0xffff);
+	MLX5E_FTE_SET(headers_v, ethertype, ETH_P_IPV6);
+}
+
+static void
+set_tcp(void *headers_c, void *headers_v, __be16 psrc_m, __be16 psrc_v,
+	__be16 pdst_m, __be16 pdst_v)
+{
+	if (psrc_m) {
+		MLX5E_FTE_SET(headers_c, tcp_sport, 0xffff);
+		MLX5E_FTE_SET(headers_v, tcp_sport, ntohs(psrc_v));
+	}
+	if (pdst_m) {
+		MLX5E_FTE_SET(headers_c, tcp_dport, 0xffff);
+		MLX5E_FTE_SET(headers_v, tcp_dport, ntohs(pdst_v));
+	}
+
+	MLX5E_FTE_SET(headers_c, ip_protocol, 0xffff);
+	MLX5E_FTE_SET(headers_v, ip_protocol, IPPROTO_TCP);
+}
+
+static void
+set_udp(void *headers_c, void *headers_v, __be16 psrc_m, __be16 psrc_v,
+	__be16 pdst_m, __be16 pdst_v)
+{
+	if (psrc_m) {
+		MLX5E_FTE_SET(headers_c, udp_sport, 0xffff);
+		MLX5E_FTE_SET(headers_v, udp_sport, ntohs(psrc_v));
+	}
+
+	if (pdst_m) {
+		MLX5E_FTE_SET(headers_c, udp_dport, 0xffff);
+		MLX5E_FTE_SET(headers_v, udp_dport, ntohs(pdst_v));
+	}
+
+	MLX5E_FTE_SET(headers_c, ip_protocol, 0xffff);
+	MLX5E_FTE_SET(headers_v, ip_protocol, IPPROTO_UDP);
+}
+
+static void
+parse_tcp4(void *headers_c, void *headers_v, struct ethtool_rx_flow_spec *fs)
+{
+	struct ethtool_tcpip4_spec *l4_mask = &fs->m_u.tcp_ip4_spec;
+	struct ethtool_tcpip4_spec *l4_val  = &fs->h_u.tcp_ip4_spec;
+
+	set_ip4(headers_c, headers_v, l4_mask->ip4src, l4_val->ip4src,
+		l4_mask->ip4dst, l4_val->ip4dst);
+
+	set_tcp(headers_c, headers_v, l4_mask->psrc, l4_val->psrc,
+		l4_mask->pdst, l4_val->pdst);
+}
+
+static void
+parse_udp4(void *headers_c, void *headers_v, struct ethtool_rx_flow_spec *fs)
+{
+	struct ethtool_tcpip4_spec *l4_mask = &fs->m_u.udp_ip4_spec;
+	struct ethtool_tcpip4_spec *l4_val  = &fs->h_u.udp_ip4_spec;
+
+	set_ip4(headers_c, headers_v, l4_mask->ip4src, l4_val->ip4src,
+		l4_mask->ip4dst, l4_val->ip4dst);
+
+	set_udp(headers_c, headers_v, l4_mask->psrc, l4_val->psrc,
+		l4_mask->pdst, l4_val->pdst);
+}
+
+static void
+parse_ip4(void *headers_c, void *headers_v, struct ethtool_rx_flow_spec *fs)
+{
+	struct ethtool_usrip4_spec *l3_mask = &fs->m_u.usr_ip4_spec;
+	struct ethtool_usrip4_spec *l3_val  = &fs->h_u.usr_ip4_spec;
+
+	set_ip4(headers_c, headers_v, l3_mask->ip4src, l3_val->ip4src,
+		l3_mask->ip4dst, l3_val->ip4dst);
+
+	if (l3_mask->proto) {
+		MLX5E_FTE_SET(headers_c, ip_protocol, l3_mask->proto);
+		MLX5E_FTE_SET(headers_v, ip_protocol, l3_val->proto);
+	}
+}
+
+static void
+parse_ip6(void *headers_c, void *headers_v, struct ethtool_rx_flow_spec *fs)
+{
+	struct ethtool_usrip6_spec *l3_mask = &fs->m_u.usr_ip6_spec;
+	struct ethtool_usrip6_spec *l3_val  = &fs->h_u.usr_ip6_spec;
+
+	set_ip6(headers_c, headers_v, l3_mask->ip6src,
+		l3_val->ip6src, l3_mask->ip6dst, l3_val->ip6dst);
+
+	if (l3_mask->l4_proto) {
+		MLX5E_FTE_SET(headers_c, ip_protocol, l3_mask->l4_proto);
+		MLX5E_FTE_SET(headers_v, ip_protocol, l3_val->l4_proto);
+	}
+}
+
+static void
+parse_tcp6(void *headers_c, void *headers_v, struct ethtool_rx_flow_spec *fs)
+{
+	struct ethtool_tcpip6_spec *l4_mask = &fs->m_u.tcp_ip6_spec;
+	struct ethtool_tcpip6_spec *l4_val  = &fs->h_u.tcp_ip6_spec;
+
+	set_ip6(headers_c, headers_v, l4_mask->ip6src,
+		l4_val->ip6src, l4_mask->ip6dst, l4_val->ip6dst);
+
+	set_tcp(headers_c, headers_v, l4_mask->psrc, l4_val->psrc,
+		l4_mask->pdst, l4_val->pdst);
+}
+
+static void
+parse_udp6(void *headers_c, void *headers_v, struct ethtool_rx_flow_spec *fs)
+{
+	struct ethtool_tcpip6_spec *l4_mask = &fs->m_u.udp_ip6_spec;
+	struct ethtool_tcpip6_spec *l4_val  = &fs->h_u.udp_ip6_spec;
+
+	set_ip6(headers_c, headers_v, l4_mask->ip6src,
+		l4_val->ip6src, l4_mask->ip6dst, l4_val->ip6dst);
+
+	set_udp(headers_c, headers_v, l4_mask->psrc, l4_val->psrc,
+		l4_mask->pdst, l4_val->pdst);
+}
+
+static void
+parse_ether(void *headers_c, void *headers_v, struct ethtool_rx_flow_spec *fs)
+{
+	struct ethhdr *eth_mask = &fs->m_u.ether_spec;
+	struct ethhdr *eth_val = &fs->h_u.ether_spec;
+
+	mask_spec((u8 *)eth_mask, (u8 *)eth_val, sizeof(*eth_mask));
+	ether_addr_copy(MLX5E_FTE_ADDR_OF(headers_c, smac_47_16), eth_mask->h_source);
+	ether_addr_copy(MLX5E_FTE_ADDR_OF(headers_v, smac_47_16), eth_val->h_source);
+	ether_addr_copy(MLX5E_FTE_ADDR_OF(headers_c, dmac_47_16), eth_mask->h_dest);
+	ether_addr_copy(MLX5E_FTE_ADDR_OF(headers_v, dmac_47_16), eth_val->h_dest);
+	MLX5E_FTE_SET(headers_c, ethertype, ntohs(eth_mask->h_proto));
+	MLX5E_FTE_SET(headers_v, ethertype, ntohs(eth_val->h_proto));
+}
+
+static void
+set_cvlan(void *headers_c, void *headers_v, __be16 vlan_tci)
+{
+	MLX5E_FTE_SET(headers_c, cvlan_tag, 1);
+	MLX5E_FTE_SET(headers_v, cvlan_tag, 1);
+	MLX5E_FTE_SET(headers_c, first_vid, 0xfff);
+	MLX5E_FTE_SET(headers_v, first_vid, ntohs(vlan_tci));
+}
+
+static void
+set_dmac(void *headers_c, void *headers_v,
+	 unsigned char m_dest[ETH_ALEN], unsigned char v_dest[ETH_ALEN])
+{
+	ether_addr_copy(MLX5E_FTE_ADDR_OF(headers_c, dmac_47_16), m_dest);
+	ether_addr_copy(MLX5E_FTE_ADDR_OF(headers_v, dmac_47_16), v_dest);
+}
+
+static int set_flow_attrs(u32 *match_c, u32 *match_v,
+			  struct ethtool_rx_flow_spec *fs)
+{
+	void *outer_headers_c = MLX5_ADDR_OF(fte_match_param, match_c,
+					     outer_headers);
+	void *outer_headers_v = MLX5_ADDR_OF(fte_match_param, match_v,
+					     outer_headers);
+	u32 flow_type = fs->flow_type & ~(FLOW_EXT | FLOW_MAC_EXT);
+
+	switch (flow_type) {
+	case TCP_V4_FLOW:
+		parse_tcp4(outer_headers_c, outer_headers_v, fs);
+		break;
+	case UDP_V4_FLOW:
+		parse_udp4(outer_headers_c, outer_headers_v, fs);
+		break;
+	case IP_USER_FLOW:
+		parse_ip4(outer_headers_c, outer_headers_v, fs);
+		break;
+	case TCP_V6_FLOW:
+		parse_tcp6(outer_headers_c, outer_headers_v, fs);
+		break;
+	case UDP_V6_FLOW:
+		parse_udp6(outer_headers_c, outer_headers_v, fs);
+		break;
+	case IPV6_USER_FLOW:
+		parse_ip6(outer_headers_c, outer_headers_v, fs);
+		break;
+	case ETHER_FLOW:
+		parse_ether(outer_headers_c, outer_headers_v, fs);
+		break;
+	default:
+		return -EINVAL;
+	}
+
+	if ((fs->flow_type & FLOW_EXT) &&
+	    (fs->m_ext.vlan_tci & cpu_to_be16(VLAN_VID_MASK)))
+		set_cvlan(outer_headers_c, outer_headers_v, fs->h_ext.vlan_tci);
+
+	if (fs->flow_type & FLOW_MAC_EXT &&
+	    !is_zero_ether_addr(fs->m_ext.h_dest)) {
+		mask_spec(fs->m_ext.h_dest, fs->h_ext.h_dest, ETH_ALEN);
+		set_dmac(outer_headers_c, outer_headers_v, fs->m_ext.h_dest,
+			 fs->h_ext.h_dest);
+	}
+
+	return 0;
+}
+
+static void add_rule_to_list(struct mlx5e_priv *priv,
+			     struct mlx5e_ethtool_rule *rule)
+{
+	struct mlx5e_ethtool_rule *iter;
+	struct list_head *head = &priv->fs.ethtool.rules;
+
+	list_for_each_entry(iter, &priv->fs.ethtool.rules, list) {
+		if (iter->flow_spec.location > rule->flow_spec.location)
+			break;
+		head = &iter->list;
+	}
+	priv->fs.ethtool.tot_num_rules++;
+	list_add(&rule->list, head);
+}
+
+static bool outer_header_zero(u32 *match_criteria)
+{
+	int size = MLX5_FLD_SZ_BYTES(fte_match_param, outer_headers);
+	char *outer_headers_c = MLX5_ADDR_OF(fte_match_param, match_criteria,
+					     outer_headers);
+
+	return outer_headers_c[0] == 0 && !memcmp(outer_headers_c,
+						  outer_headers_c + 1,
+						  size - 1);
+}
+
+static struct mlx5_flow_handle *
+add_ethtool_flow_rule(struct mlx5e_priv *priv,
+		      struct mlx5_flow_table *ft,
+		      struct ethtool_rx_flow_spec *fs)
+{
+	struct mlx5_flow_destination *dst = NULL;
+	struct mlx5_flow_act flow_act = {0};
+	struct mlx5_flow_spec *spec;
+	struct mlx5_flow_handle *rule;
+	int err = 0;
+
+	spec = kvzalloc(sizeof(*spec), GFP_KERNEL);
+	if (!spec)
+		return ERR_PTR(-ENOMEM);
+	err = set_flow_attrs(spec->match_criteria, spec->match_value,
+			     fs);
+	if (err)
+		goto free;
+
+	if (fs->ring_cookie == RX_CLS_FLOW_DISC) {
+		flow_act.action = MLX5_FLOW_CONTEXT_ACTION_DROP;
+	} else {
+		dst = kzalloc(sizeof(*dst), GFP_KERNEL);
+		if (!dst) {
+			err = -ENOMEM;
+			goto free;
+		}
+
+		dst->type = MLX5_FLOW_DESTINATION_TYPE_TIR;
+		dst->tir_num = priv->direct_tir[fs->ring_cookie].tirn;
+		flow_act.action = MLX5_FLOW_CONTEXT_ACTION_FWD_DEST;
+	}
+
+	spec->match_criteria_enable = (!outer_header_zero(spec->match_criteria));
+	flow_act.flow_tag = MLX5_FS_DEFAULT_FLOW_TAG;
+	rule = mlx5_add_flow_rules(ft, spec, &flow_act, dst, dst ? 1 : 0);
+	if (IS_ERR(rule)) {
+		err = PTR_ERR(rule);
+		netdev_err(priv->netdev, "%s: failed to add ethtool steering rule: %d\n",
+			   __func__, err);
+		goto free;
+	}
+free:
+	kvfree(spec);
+	kfree(dst);
+	return err ? ERR_PTR(err) : rule;
+}
+
+static void del_ethtool_rule(struct mlx5e_priv *priv,
+			     struct mlx5e_ethtool_rule *eth_rule)
+{
+	if (eth_rule->rule)
+		mlx5_del_flow_rules(eth_rule->rule);
+	list_del(&eth_rule->list);
+	priv->fs.ethtool.tot_num_rules--;
+	put_flow_table(eth_rule->eth_ft);
+	kfree(eth_rule);
+}
+
+static struct mlx5e_ethtool_rule *find_ethtool_rule(struct mlx5e_priv *priv,
+						    int location)
+{
+	struct mlx5e_ethtool_rule *iter;
+
+	list_for_each_entry(iter, &priv->fs.ethtool.rules, list) {
+		if (iter->flow_spec.location == location)
+			return iter;
+	}
+	return NULL;
+}
+
+static struct mlx5e_ethtool_rule *get_ethtool_rule(struct mlx5e_priv *priv,
+						   int location)
+{
+	struct mlx5e_ethtool_rule *eth_rule;
+
+	eth_rule = find_ethtool_rule(priv, location);
+	if (eth_rule)
+		del_ethtool_rule(priv, eth_rule);
+
+	eth_rule = kzalloc(sizeof(*eth_rule), GFP_KERNEL);
+	if (!eth_rule)
+		return ERR_PTR(-ENOMEM);
+
+	add_rule_to_list(priv, eth_rule);
+	return eth_rule;
+}
+
+#define MAX_NUM_OF_ETHTOOL_RULES BIT(10)
+
+#define all_ones(field) (field == (__force typeof(field))-1)
+#define all_zeros_or_all_ones(field)		\
+	((field) == 0 || (field) == (__force typeof(field))-1)
+
+static int validate_ethter(struct ethtool_rx_flow_spec *fs)
+{
+	struct ethhdr *eth_mask = &fs->m_u.ether_spec;
+	int ntuples = 0;
+
+	if (!is_zero_ether_addr(eth_mask->h_dest))
+		ntuples++;
+	if (!is_zero_ether_addr(eth_mask->h_source))
+		ntuples++;
+	if (eth_mask->h_proto)
+		ntuples++;
+	return ntuples;
+}
+
+static int validate_tcpudp4(struct ethtool_rx_flow_spec *fs)
+{
+	struct ethtool_tcpip4_spec *l4_mask = &fs->m_u.tcp_ip4_spec;
+	int ntuples = 0;
+
+	if (l4_mask->tos)
+		return -EINVAL;
+
+	if (l4_mask->ip4src) {
+		if (!all_ones(l4_mask->ip4src))
+			return -EINVAL;
+		ntuples++;
+	}
+	if (l4_mask->ip4dst) {
+		if (!all_ones(l4_mask->ip4dst))
+			return -EINVAL;
+		ntuples++;
+	}
+	if (l4_mask->psrc) {
+		if (!all_ones(l4_mask->psrc))
+			return -EINVAL;
+		ntuples++;
+	}
+	if (l4_mask->pdst) {
+		if (!all_ones(l4_mask->pdst))
+			return -EINVAL;
+		ntuples++;
+	}
+	/* Flow is TCP/UDP */
+	return ++ntuples;
+}
+
+static int validate_ip4(struct ethtool_rx_flow_spec *fs)
+{
+	struct ethtool_usrip4_spec *l3_mask = &fs->m_u.usr_ip4_spec;
+	int ntuples = 0;
+
+	if (l3_mask->l4_4_bytes || l3_mask->tos ||
+	    fs->h_u.usr_ip4_spec.ip_ver != ETH_RX_NFC_IP4)
+		return -EINVAL;
+	if (l3_mask->ip4src) {
+		if (!all_ones(l3_mask->ip4src))
+			return -EINVAL;
+		ntuples++;
+	}
+	if (l3_mask->ip4dst) {
+		if (!all_ones(l3_mask->ip4dst))
+			return -EINVAL;
+		ntuples++;
+	}
+	if (l3_mask->proto)
+		ntuples++;
+	/* Flow is IPv4 */
+	return ++ntuples;
+}
+
+static int validate_ip6(struct ethtool_rx_flow_spec *fs)
+{
+	struct ethtool_usrip6_spec *l3_mask = &fs->m_u.usr_ip6_spec;
+	int ntuples = 0;
+
+	if (l3_mask->l4_4_bytes || l3_mask->tclass)
+		return -EINVAL;
+	if (!ipv6_addr_any((struct in6_addr *)l3_mask->ip6src))
+		ntuples++;
+
+	if (!ipv6_addr_any((struct in6_addr *)l3_mask->ip6dst))
+		ntuples++;
+	if (l3_mask->l4_proto)
+		ntuples++;
+	/* Flow is IPv6 */
+	return ++ntuples;
+}
+
+static int validate_tcpudp6(struct ethtool_rx_flow_spec *fs)
+{
+	struct ethtool_tcpip6_spec *l4_mask = &fs->m_u.tcp_ip6_spec;
+	int ntuples = 0;
+
+	if (l4_mask->tclass)
+		return -EINVAL;
+
+	if (!ipv6_addr_any((struct in6_addr *)l4_mask->ip6src))
+		ntuples++;
+
+	if (!ipv6_addr_any((struct in6_addr *)l4_mask->ip6dst))
+		ntuples++;
+
+	if (l4_mask->psrc) {
+		if (!all_ones(l4_mask->psrc))
+			return -EINVAL;
+		ntuples++;
+	}
+	if (l4_mask->pdst) {
+		if (!all_ones(l4_mask->pdst))
+			return -EINVAL;
+		ntuples++;
+	}
+	/* Flow is TCP/UDP */
+	return ++ntuples;
+}
+
+static int validate_vlan(struct ethtool_rx_flow_spec *fs)
+{
+	if (fs->m_ext.vlan_etype ||
+	    fs->m_ext.vlan_tci != cpu_to_be16(VLAN_VID_MASK))
+		return -EINVAL;
+
+	if (fs->m_ext.vlan_tci &&
+	    (be16_to_cpu(fs->h_ext.vlan_tci) >= VLAN_N_VID))
+		return -EINVAL;
+
+	return 1;
+}
+
+static int validate_flow(struct mlx5e_priv *priv,
+			 struct ethtool_rx_flow_spec *fs)
+{
+	int num_tuples = 0;
+	int ret = 0;
+
+	if (fs->location >= MAX_NUM_OF_ETHTOOL_RULES)
+		return -ENOSPC;
+
+	if (fs->ring_cookie >= priv->channels.params.num_channels &&
+	    fs->ring_cookie != RX_CLS_FLOW_DISC)
+		return -EINVAL;
+
+	switch (fs->flow_type & ~(FLOW_EXT | FLOW_MAC_EXT)) {
+	case ETHER_FLOW:
+		num_tuples += validate_ethter(fs);
+		break;
+	case TCP_V4_FLOW:
+	case UDP_V4_FLOW:
+		ret = validate_tcpudp4(fs);
+		if (ret < 0)
+			return ret;
+		num_tuples += ret;
+		break;
+	case IP_USER_FLOW:
+		ret = validate_ip4(fs);
+		if (ret < 0)
+			return ret;
+		num_tuples += ret;
+		break;
+	case TCP_V6_FLOW:
+	case UDP_V6_FLOW:
+		ret = validate_tcpudp6(fs);
+		if (ret < 0)
+			return ret;
+		num_tuples += ret;
+		break;
+	case IPV6_USER_FLOW:
+		ret = validate_ip6(fs);
+		if (ret < 0)
+			return ret;
+		num_tuples += ret;
+		break;
+	default:
+		return -ENOTSUPP;
+	}
+	if ((fs->flow_type & FLOW_EXT)) {
+		ret = validate_vlan(fs);
+		if (ret < 0)
+			return ret;
+		num_tuples += ret;
+	}
+
+	if (fs->flow_type & FLOW_MAC_EXT &&
+	    !is_zero_ether_addr(fs->m_ext.h_dest))
+		num_tuples++;
+
+	return num_tuples;
+}
+
+static int
+mlx5e_ethtool_flow_replace(struct mlx5e_priv *priv,
+			   struct ethtool_rx_flow_spec *fs)
+{
+	struct mlx5e_ethtool_table *eth_ft;
+	struct mlx5e_ethtool_rule *eth_rule;
+	struct mlx5_flow_handle *rule;
+	int num_tuples;
+	int err;
+
+	num_tuples = validate_flow(priv, fs);
+	if (num_tuples <= 0) {
+		netdev_warn(priv->netdev, "%s: flow is not valid %d\n",
+			    __func__, num_tuples);
+		return num_tuples;
+	}
+
+	eth_ft = get_flow_table(priv, fs, num_tuples);
+	if (IS_ERR(eth_ft))
+		return PTR_ERR(eth_ft);
+
+	eth_rule = get_ethtool_rule(priv, fs->location);
+	if (IS_ERR(eth_rule)) {
+		put_flow_table(eth_ft);
+		return PTR_ERR(eth_rule);
+	}
+
+	eth_rule->flow_spec = *fs;
+	eth_rule->eth_ft = eth_ft;
+	if (!eth_ft->ft) {
+		err = -EINVAL;
+		goto del_ethtool_rule;
+	}
+	rule = add_ethtool_flow_rule(priv, eth_ft->ft, fs);
+	if (IS_ERR(rule)) {
+		err = PTR_ERR(rule);
+		goto del_ethtool_rule;
+	}
+
+	eth_rule->rule = rule;
+
+	return 0;
+
+del_ethtool_rule:
+	del_ethtool_rule(priv, eth_rule);
+
+	return err;
+}
+
+static int
+mlx5e_ethtool_flow_remove(struct mlx5e_priv *priv, int location)
+{
+	struct mlx5e_ethtool_rule *eth_rule;
+	int err = 0;
+
+	if (location >= MAX_NUM_OF_ETHTOOL_RULES)
+		return -ENOSPC;
+
+	eth_rule = find_ethtool_rule(priv, location);
+	if (!eth_rule) {
+		err =  -ENOENT;
+		goto out;
+	}
+
+	del_ethtool_rule(priv, eth_rule);
+out:
+	return err;
+}
+
+static int
+mlx5e_ethtool_get_flow(struct mlx5e_priv *priv,
+		       struct ethtool_rxnfc *info, int location)
+{
+	struct mlx5e_ethtool_rule *eth_rule;
+
+	if (location < 0 || location >= MAX_NUM_OF_ETHTOOL_RULES)
+		return -EINVAL;
+
+	list_for_each_entry(eth_rule, &priv->fs.ethtool.rules, list) {
+		if (eth_rule->flow_spec.location == location) {
+			info->fs = eth_rule->flow_spec;
+			return 0;
+		}
+	}
+
+	return -ENOENT;
+}
+
+static int
+mlx5e_ethtool_get_all_flows(struct mlx5e_priv *priv,
+			    struct ethtool_rxnfc *info, u32 *rule_locs)
+{
+	int location = 0;
+	int idx = 0;
+	int err = 0;
+
+	info->data = MAX_NUM_OF_ETHTOOL_RULES;
+	while ((!err || err == -ENOENT) && idx < info->rule_cnt) {
+		err = mlx5e_ethtool_get_flow(priv, info, location);
+		if (!err)
+			rule_locs[idx++] = location;
+		location++;
+	}
+	return err;
+}
+
+void mlx5e_ethtool_cleanup_steering(struct mlx5e_priv *priv)
+{
+	struct mlx5e_ethtool_rule *iter;
+	struct mlx5e_ethtool_rule *temp;
+
+	list_for_each_entry_safe(iter, temp, &priv->fs.ethtool.rules, list)
+		del_ethtool_rule(priv, iter);
+}
+
+void mlx5e_ethtool_init_steering(struct mlx5e_priv *priv)
+{
+	INIT_LIST_HEAD(&priv->fs.ethtool.rules);
+}
+
+int mlx5e_set_rxnfc(struct net_device *dev, struct ethtool_rxnfc *cmd)
+{
+	int err = 0;
+	struct mlx5e_priv *priv = netdev_priv(dev);
+
+	switch (cmd->cmd) {
+	case ETHTOOL_SRXCLSRLINS:
+		err = mlx5e_ethtool_flow_replace(priv, &cmd->fs);
+		break;
+	case ETHTOOL_SRXCLSRLDEL:
+		err = mlx5e_ethtool_flow_remove(priv, cmd->fs.location);
+		break;
+	default:
+		err = -EOPNOTSUPP;
+		break;
+	}
+
+	return err;
+}
+
+int mlx5e_get_rxnfc(struct net_device *dev,
+		    struct ethtool_rxnfc *info, u32 *rule_locs)
+{
+	struct mlx5e_priv *priv = netdev_priv(dev);
+	int err = 0;
+
+	switch (info->cmd) {
+	case ETHTOOL_GRXRINGS:
+		info->data = priv->channels.params.num_channels;
+		break;
+	case ETHTOOL_GRXCLSRLCNT:
+		info->rule_cnt = priv->fs.ethtool.tot_num_rules;
+		break;
+	case ETHTOOL_GRXCLSRULE:
+		err = mlx5e_ethtool_get_flow(priv, info, info->fs.location);
+		break;
+	case ETHTOOL_GRXCLSRLALL:
+		err = mlx5e_ethtool_get_all_flows(priv, info, rule_locs);
+		break;
+	default:
+		err = -EOPNOTSUPP;
+		break;
+	}
+
+	return err;
+}
+
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_main.c b/drivers/net/ethernet/mellanox/mlx5/core/en_main.c
new file mode 100644
index 000000000..6ecb92f55
--- /dev/null
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_main.c
@@ -0,0 +1,5221 @@
+/*
+ * Copyright (c) 2015-2016, 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/tc_act/tc_gact.h>
+#include <net/pkt_cls.h>
+#include <linux/mlx5/fs.h>
+#include <net/vxlan.h>
+#include <linux/bpf.h>
+#include <net/page_pool.h>
+#include "eswitch.h"
+#include "en.h"
+#include "en_tc.h"
+#include "en_rep.h"
+#include "en_accel/ipsec.h"
+#include "en_accel/ipsec_rxtx.h"
+#include "en_accel/tls.h"
+#include "accel/ipsec.h"
+#include "accel/tls.h"
+#include "lib/vxlan.h"
+#include "lib/clock.h"
+#include "en/port.h"
+#include "en/xdp.h"
+
+struct mlx5e_rq_param {
+	u32			rqc[MLX5_ST_SZ_DW(rqc)];
+	struct mlx5_wq_param	wq;
+	struct mlx5e_rq_frags_info frags_info;
+};
+
+struct mlx5e_sq_param {
+	u32                        sqc[MLX5_ST_SZ_DW(sqc)];
+	struct mlx5_wq_param       wq;
+};
+
+struct mlx5e_cq_param {
+	u32                        cqc[MLX5_ST_SZ_DW(cqc)];
+	struct mlx5_wq_param       wq;
+	u16                        eq_ix;
+	u8                         cq_period_mode;
+};
+
+struct mlx5e_channel_param {
+	struct mlx5e_rq_param      rq;
+	struct mlx5e_sq_param      sq;
+	struct mlx5e_sq_param      xdp_sq;
+	struct mlx5e_sq_param      icosq;
+	struct mlx5e_cq_param      rx_cq;
+	struct mlx5e_cq_param      tx_cq;
+	struct mlx5e_cq_param      icosq_cq;
+};
+
+bool mlx5e_check_fragmented_striding_rq_cap(struct mlx5_core_dev *mdev)
+{
+	bool striding_rq_umr = MLX5_CAP_GEN(mdev, striding_rq) &&
+		MLX5_CAP_GEN(mdev, umr_ptr_rlky) &&
+		MLX5_CAP_ETH(mdev, reg_umr_sq);
+	u16 max_wqe_sz_cap = MLX5_CAP_GEN(mdev, max_wqe_sz_sq);
+	bool inline_umr = MLX5E_UMR_WQE_INLINE_SZ <= max_wqe_sz_cap;
+
+	if (!striding_rq_umr)
+		return false;
+	if (!inline_umr) {
+		mlx5_core_warn(mdev, "Cannot support Striding RQ: UMR WQE size (%d) exceeds maximum supported (%d).\n",
+			       (int)MLX5E_UMR_WQE_INLINE_SZ, max_wqe_sz_cap);
+		return false;
+	}
+	return true;
+}
+
+static u32 mlx5e_rx_get_linear_frag_sz(struct mlx5e_params *params)
+{
+	u16 hw_mtu = MLX5E_SW2HW_MTU(params, params->sw_mtu);
+	u16 linear_rq_headroom = params->xdp_prog ?
+		XDP_PACKET_HEADROOM : MLX5_RX_HEADROOM;
+	u32 frag_sz;
+
+	linear_rq_headroom += NET_IP_ALIGN;
+
+	frag_sz = MLX5_SKB_FRAG_SZ(linear_rq_headroom + hw_mtu);
+
+	if (params->xdp_prog && frag_sz < PAGE_SIZE)
+		frag_sz = PAGE_SIZE;
+
+	return frag_sz;
+}
+
+static u8 mlx5e_mpwqe_log_pkts_per_wqe(struct mlx5e_params *params)
+{
+	u32 linear_frag_sz = mlx5e_rx_get_linear_frag_sz(params);
+
+	return MLX5_MPWRQ_LOG_WQE_SZ - order_base_2(linear_frag_sz);
+}
+
+static bool mlx5e_rx_is_linear_skb(struct mlx5_core_dev *mdev,
+				   struct mlx5e_params *params)
+{
+	u32 frag_sz = mlx5e_rx_get_linear_frag_sz(params);
+
+	return !params->lro_en && frag_sz <= PAGE_SIZE;
+}
+
+#define MLX5_MAX_MPWQE_LOG_WQE_STRIDE_SZ ((BIT(__mlx5_bit_sz(wq, log_wqe_stride_size)) - 1) + \
+					  MLX5_MPWQE_LOG_STRIDE_SZ_BASE)
+static bool mlx5e_rx_mpwqe_is_linear_skb(struct mlx5_core_dev *mdev,
+					 struct mlx5e_params *params)
+{
+	u32 frag_sz = mlx5e_rx_get_linear_frag_sz(params);
+	s8 signed_log_num_strides_param;
+	u8 log_num_strides;
+
+	if (!mlx5e_rx_is_linear_skb(mdev, params))
+		return false;
+
+	if (order_base_2(frag_sz) > MLX5_MAX_MPWQE_LOG_WQE_STRIDE_SZ)
+		return false;
+
+	if (MLX5_CAP_GEN(mdev, ext_stride_num_range))
+		return true;
+
+	log_num_strides = MLX5_MPWRQ_LOG_WQE_SZ - order_base_2(frag_sz);
+	signed_log_num_strides_param =
+		(s8)log_num_strides - MLX5_MPWQE_LOG_NUM_STRIDES_BASE;
+
+	return signed_log_num_strides_param >= 0;
+}
+
+static u8 mlx5e_mpwqe_get_log_rq_size(struct mlx5e_params *params)
+{
+	if (params->log_rq_mtu_frames <
+	    mlx5e_mpwqe_log_pkts_per_wqe(params) + MLX5E_PARAMS_MINIMUM_LOG_RQ_SIZE_MPW)
+		return MLX5E_PARAMS_MINIMUM_LOG_RQ_SIZE_MPW;
+
+	return params->log_rq_mtu_frames - mlx5e_mpwqe_log_pkts_per_wqe(params);
+}
+
+static u8 mlx5e_mpwqe_get_log_stride_size(struct mlx5_core_dev *mdev,
+					  struct mlx5e_params *params)
+{
+	if (mlx5e_rx_mpwqe_is_linear_skb(mdev, params))
+		return order_base_2(mlx5e_rx_get_linear_frag_sz(params));
+
+	return MLX5E_MPWQE_STRIDE_SZ(mdev,
+		MLX5E_GET_PFLAG(params, MLX5E_PFLAG_RX_CQE_COMPRESS));
+}
+
+static u8 mlx5e_mpwqe_get_log_num_strides(struct mlx5_core_dev *mdev,
+					  struct mlx5e_params *params)
+{
+	return MLX5_MPWRQ_LOG_WQE_SZ -
+		mlx5e_mpwqe_get_log_stride_size(mdev, params);
+}
+
+static u16 mlx5e_get_rq_headroom(struct mlx5_core_dev *mdev,
+				 struct mlx5e_params *params)
+{
+	u16 linear_rq_headroom = params->xdp_prog ?
+		XDP_PACKET_HEADROOM : MLX5_RX_HEADROOM;
+	bool is_linear_skb;
+
+	linear_rq_headroom += NET_IP_ALIGN;
+
+	is_linear_skb = (params->rq_wq_type == MLX5_WQ_TYPE_CYCLIC) ?
+		mlx5e_rx_is_linear_skb(mdev, params) :
+		mlx5e_rx_mpwqe_is_linear_skb(mdev, params);
+
+	return is_linear_skb ? linear_rq_headroom : 0;
+}
+
+void mlx5e_init_rq_type_params(struct mlx5_core_dev *mdev,
+			       struct mlx5e_params *params)
+{
+	params->lro_wqe_sz = MLX5E_PARAMS_DEFAULT_LRO_WQE_SZ;
+	params->log_rq_mtu_frames = is_kdump_kernel() ?
+		MLX5E_PARAMS_MINIMUM_LOG_RQ_SIZE :
+		MLX5E_PARAMS_DEFAULT_LOG_RQ_SIZE;
+
+	mlx5_core_info(mdev, "MLX5E: StrdRq(%d) RqSz(%ld) StrdSz(%ld) RxCqeCmprss(%d)\n",
+		       params->rq_wq_type == MLX5_WQ_TYPE_LINKED_LIST_STRIDING_RQ,
+		       params->rq_wq_type == MLX5_WQ_TYPE_LINKED_LIST_STRIDING_RQ ?
+		       BIT(mlx5e_mpwqe_get_log_rq_size(params)) :
+		       BIT(params->log_rq_mtu_frames),
+		       BIT(mlx5e_mpwqe_get_log_stride_size(mdev, params)),
+		       MLX5E_GET_PFLAG(params, MLX5E_PFLAG_RX_CQE_COMPRESS));
+}
+
+bool mlx5e_striding_rq_possible(struct mlx5_core_dev *mdev,
+				struct mlx5e_params *params)
+{
+	return mlx5e_check_fragmented_striding_rq_cap(mdev) &&
+		!MLX5_IPSEC_DEV(mdev) &&
+		!(params->xdp_prog && !mlx5e_rx_mpwqe_is_linear_skb(mdev, params));
+}
+
+void mlx5e_set_rq_type(struct mlx5_core_dev *mdev, struct mlx5e_params *params)
+{
+	params->rq_wq_type = mlx5e_striding_rq_possible(mdev, params) &&
+		MLX5E_GET_PFLAG(params, MLX5E_PFLAG_RX_STRIDING_RQ) ?
+		MLX5_WQ_TYPE_LINKED_LIST_STRIDING_RQ :
+		MLX5_WQ_TYPE_CYCLIC;
+}
+
+static void mlx5e_update_carrier(struct mlx5e_priv *priv)
+{
+	struct mlx5_core_dev *mdev = priv->mdev;
+	u8 port_state;
+
+	port_state = mlx5_query_vport_state(mdev,
+					    MLX5_VPORT_STATE_OP_MOD_VNIC_VPORT,
+					    0);
+
+	if (port_state == VPORT_STATE_UP) {
+		netdev_info(priv->netdev, "Link up\n");
+		netif_carrier_on(priv->netdev);
+	} else {
+		netdev_info(priv->netdev, "Link down\n");
+		netif_carrier_off(priv->netdev);
+	}
+}
+
+static void mlx5e_update_carrier_work(struct work_struct *work)
+{
+	struct mlx5e_priv *priv = container_of(work, struct mlx5e_priv,
+					       update_carrier_work);
+
+	mutex_lock(&priv->state_lock);
+	if (test_bit(MLX5E_STATE_OPENED, &priv->state))
+		if (priv->profile->update_carrier)
+			priv->profile->update_carrier(priv);
+	mutex_unlock(&priv->state_lock);
+}
+
+void mlx5e_update_stats(struct mlx5e_priv *priv)
+{
+	int i;
+
+	for (i = mlx5e_num_stats_grps - 1; i >= 0; i--)
+		if (mlx5e_stats_grps[i].update_stats)
+			mlx5e_stats_grps[i].update_stats(priv);
+}
+
+static void mlx5e_update_ndo_stats(struct mlx5e_priv *priv)
+{
+	int i;
+
+	for (i = mlx5e_num_stats_grps - 1; i >= 0; i--)
+		if (mlx5e_stats_grps[i].update_stats_mask &
+		    MLX5E_NDO_UPDATE_STATS)
+			mlx5e_stats_grps[i].update_stats(priv);
+}
+
+void mlx5e_update_stats_work(struct work_struct *work)
+{
+	struct delayed_work *dwork = to_delayed_work(work);
+	struct mlx5e_priv *priv = container_of(dwork, struct mlx5e_priv,
+					       update_stats_work);
+
+	mutex_lock(&priv->state_lock);
+	priv->profile->update_stats(priv);
+	mutex_unlock(&priv->state_lock);
+}
+
+static void mlx5e_async_event(struct mlx5_core_dev *mdev, void *vpriv,
+			      enum mlx5_dev_event event, unsigned long param)
+{
+	struct mlx5e_priv *priv = vpriv;
+
+	if (!test_bit(MLX5E_STATE_ASYNC_EVENTS_ENABLED, &priv->state))
+		return;
+
+	switch (event) {
+	case MLX5_DEV_EVENT_PORT_UP:
+	case MLX5_DEV_EVENT_PORT_DOWN:
+		queue_work(priv->wq, &priv->update_carrier_work);
+		break;
+	default:
+		break;
+	}
+}
+
+static void mlx5e_enable_async_events(struct mlx5e_priv *priv)
+{
+	set_bit(MLX5E_STATE_ASYNC_EVENTS_ENABLED, &priv->state);
+}
+
+static void mlx5e_disable_async_events(struct mlx5e_priv *priv)
+{
+	clear_bit(MLX5E_STATE_ASYNC_EVENTS_ENABLED, &priv->state);
+	synchronize_irq(pci_irq_vector(priv->mdev->pdev, MLX5_EQ_VEC_ASYNC));
+}
+
+static inline void mlx5e_build_umr_wqe(struct mlx5e_rq *rq,
+				       struct mlx5e_icosq *sq,
+				       struct mlx5e_umr_wqe *wqe)
+{
+	struct mlx5_wqe_ctrl_seg      *cseg = &wqe->ctrl;
+	struct mlx5_wqe_umr_ctrl_seg *ucseg = &wqe->uctrl;
+	u8 ds_cnt = DIV_ROUND_UP(MLX5E_UMR_WQE_INLINE_SZ, MLX5_SEND_WQE_DS);
+
+	cseg->qpn_ds    = cpu_to_be32((sq->sqn << MLX5_WQE_CTRL_QPN_SHIFT) |
+				      ds_cnt);
+	cseg->fm_ce_se  = MLX5_WQE_CTRL_CQ_UPDATE;
+	cseg->imm       = rq->mkey_be;
+
+	ucseg->flags = MLX5_UMR_TRANSLATION_OFFSET_EN | MLX5_UMR_INLINE;
+	ucseg->xlt_octowords =
+		cpu_to_be16(MLX5_MTT_OCTW(MLX5_MPWRQ_PAGES_PER_WQE));
+	ucseg->mkey_mask     = cpu_to_be64(MLX5_MKEY_MASK_FREE);
+}
+
+static u32 mlx5e_rqwq_get_size(struct mlx5e_rq *rq)
+{
+	switch (rq->wq_type) {
+	case MLX5_WQ_TYPE_LINKED_LIST_STRIDING_RQ:
+		return mlx5_wq_ll_get_size(&rq->mpwqe.wq);
+	default:
+		return mlx5_wq_cyc_get_size(&rq->wqe.wq);
+	}
+}
+
+static u32 mlx5e_rqwq_get_cur_sz(struct mlx5e_rq *rq)
+{
+	switch (rq->wq_type) {
+	case MLX5_WQ_TYPE_LINKED_LIST_STRIDING_RQ:
+		return rq->mpwqe.wq.cur_sz;
+	default:
+		return rq->wqe.wq.cur_sz;
+	}
+}
+
+static int mlx5e_rq_alloc_mpwqe_info(struct mlx5e_rq *rq,
+				     struct mlx5e_channel *c)
+{
+	int wq_sz = mlx5_wq_ll_get_size(&rq->mpwqe.wq);
+
+	rq->mpwqe.info = kvzalloc_node(array_size(wq_sz,
+						  sizeof(*rq->mpwqe.info)),
+				       GFP_KERNEL, cpu_to_node(c->cpu));
+	if (!rq->mpwqe.info)
+		return -ENOMEM;
+
+	mlx5e_build_umr_wqe(rq, &c->icosq, &rq->mpwqe.umr_wqe);
+
+	return 0;
+}
+
+static int mlx5e_create_umr_mkey(struct mlx5_core_dev *mdev,
+				 u64 npages, u8 page_shift,
+				 struct mlx5_core_mkey *umr_mkey)
+{
+	int inlen = MLX5_ST_SZ_BYTES(create_mkey_in);
+	void *mkc;
+	u32 *in;
+	int err;
+
+	in = kvzalloc(inlen, GFP_KERNEL);
+	if (!in)
+		return -ENOMEM;
+
+	mkc = MLX5_ADDR_OF(create_mkey_in, in, memory_key_mkey_entry);
+
+	MLX5_SET(mkc, mkc, free, 1);
+	MLX5_SET(mkc, mkc, umr_en, 1);
+	MLX5_SET(mkc, mkc, lw, 1);
+	MLX5_SET(mkc, mkc, lr, 1);
+	MLX5_SET(mkc, mkc, access_mode_1_0, MLX5_MKC_ACCESS_MODE_MTT);
+
+	MLX5_SET(mkc, mkc, qpn, 0xffffff);
+	MLX5_SET(mkc, mkc, pd, mdev->mlx5e_res.pdn);
+	MLX5_SET64(mkc, mkc, len, npages << page_shift);
+	MLX5_SET(mkc, mkc, translations_octword_size,
+		 MLX5_MTT_OCTW(npages));
+	MLX5_SET(mkc, mkc, log_page_size, page_shift);
+
+	err = mlx5_core_create_mkey(mdev, umr_mkey, in, inlen);
+
+	kvfree(in);
+	return err;
+}
+
+static int mlx5e_create_rq_umr_mkey(struct mlx5_core_dev *mdev, struct mlx5e_rq *rq)
+{
+	u64 num_mtts = MLX5E_REQUIRED_MTTS(mlx5_wq_ll_get_size(&rq->mpwqe.wq));
+
+	return mlx5e_create_umr_mkey(mdev, num_mtts, PAGE_SHIFT, &rq->umr_mkey);
+}
+
+static inline u64 mlx5e_get_mpwqe_offset(struct mlx5e_rq *rq, u16 wqe_ix)
+{
+	return (wqe_ix << MLX5E_LOG_ALIGNED_MPWQE_PPW) << PAGE_SHIFT;
+}
+
+static void mlx5e_init_frags_partition(struct mlx5e_rq *rq)
+{
+	struct mlx5e_wqe_frag_info next_frag = {};
+	struct mlx5e_wqe_frag_info *prev = NULL;
+	int i;
+
+	next_frag.di = &rq->wqe.di[0];
+
+	for (i = 0; i < mlx5_wq_cyc_get_size(&rq->wqe.wq); i++) {
+		struct mlx5e_rq_frag_info *frag_info = &rq->wqe.info.arr[0];
+		struct mlx5e_wqe_frag_info *frag =
+			&rq->wqe.frags[i << rq->wqe.info.log_num_frags];
+		int f;
+
+		for (f = 0; f < rq->wqe.info.num_frags; f++, frag++) {
+			if (next_frag.offset + frag_info[f].frag_stride > PAGE_SIZE) {
+				next_frag.di++;
+				next_frag.offset = 0;
+				if (prev)
+					prev->last_in_page = true;
+			}
+			*frag = next_frag;
+
+			/* prepare next */
+			next_frag.offset += frag_info[f].frag_stride;
+			prev = frag;
+		}
+	}
+
+	if (prev)
+		prev->last_in_page = true;
+}
+
+static int mlx5e_init_di_list(struct mlx5e_rq *rq,
+			      struct mlx5e_params *params,
+			      int wq_sz, int cpu)
+{
+	int len = wq_sz << rq->wqe.info.log_num_frags;
+
+	rq->wqe.di = kvzalloc_node(array_size(len, sizeof(*rq->wqe.di)),
+				   GFP_KERNEL, cpu_to_node(cpu));
+	if (!rq->wqe.di)
+		return -ENOMEM;
+
+	mlx5e_init_frags_partition(rq);
+
+	return 0;
+}
+
+static void mlx5e_free_di_list(struct mlx5e_rq *rq)
+{
+	kvfree(rq->wqe.di);
+}
+
+static int mlx5e_alloc_rq(struct mlx5e_channel *c,
+			  struct mlx5e_params *params,
+			  struct mlx5e_rq_param *rqp,
+			  struct mlx5e_rq *rq)
+{
+	struct page_pool_params pp_params = { 0 };
+	struct mlx5_core_dev *mdev = c->mdev;
+	void *rqc = rqp->rqc;
+	void *rqc_wq = MLX5_ADDR_OF(rqc, rqc, wq);
+	u32 pool_size;
+	int wq_sz;
+	int err;
+	int i;
+
+	rqp->wq.db_numa_node = cpu_to_node(c->cpu);
+
+	rq->wq_type = params->rq_wq_type;
+	rq->pdev    = c->pdev;
+	rq->netdev  = c->netdev;
+	rq->tstamp  = c->tstamp;
+	rq->clock   = &mdev->clock;
+	rq->channel = c;
+	rq->ix      = c->ix;
+	rq->mdev    = mdev;
+	rq->hw_mtu  = MLX5E_SW2HW_MTU(params, params->sw_mtu);
+	rq->stats   = &c->priv->channel_stats[c->ix].rq;
+
+	rq->xdp_prog = params->xdp_prog ? bpf_prog_inc(params->xdp_prog) : NULL;
+	if (IS_ERR(rq->xdp_prog)) {
+		err = PTR_ERR(rq->xdp_prog);
+		rq->xdp_prog = NULL;
+		goto err_rq_wq_destroy;
+	}
+
+	err = xdp_rxq_info_reg(&rq->xdp_rxq, rq->netdev, rq->ix);
+	if (err < 0)
+		goto err_rq_wq_destroy;
+
+	rq->buff.map_dir = rq->xdp_prog ? DMA_BIDIRECTIONAL : DMA_FROM_DEVICE;
+	rq->buff.headroom = mlx5e_get_rq_headroom(mdev, params);
+	pool_size = 1 << params->log_rq_mtu_frames;
+
+	switch (rq->wq_type) {
+	case MLX5_WQ_TYPE_LINKED_LIST_STRIDING_RQ:
+		err = mlx5_wq_ll_create(mdev, &rqp->wq, rqc_wq, &rq->mpwqe.wq,
+					&rq->wq_ctrl);
+		if (err)
+			goto err_rq_wq_destroy;
+
+		rq->mpwqe.wq.db = &rq->mpwqe.wq.db[MLX5_RCV_DBR];
+
+		wq_sz = mlx5_wq_ll_get_size(&rq->mpwqe.wq);
+
+		pool_size = MLX5_MPWRQ_PAGES_PER_WQE << mlx5e_mpwqe_get_log_rq_size(params);
+
+		rq->post_wqes = mlx5e_post_rx_mpwqes;
+		rq->dealloc_wqe = mlx5e_dealloc_rx_mpwqe;
+
+		rq->handle_rx_cqe = c->priv->profile->rx_handlers.handle_rx_cqe_mpwqe;
+#ifdef CONFIG_MLX5_EN_IPSEC
+		if (MLX5_IPSEC_DEV(mdev)) {
+			err = -EINVAL;
+			netdev_err(c->netdev, "MPWQE RQ with IPSec offload not supported\n");
+			goto err_rq_wq_destroy;
+		}
+#endif
+		if (!rq->handle_rx_cqe) {
+			err = -EINVAL;
+			netdev_err(c->netdev, "RX handler of MPWQE RQ is not set, err %d\n", err);
+			goto err_rq_wq_destroy;
+		}
+
+		rq->mpwqe.skb_from_cqe_mpwrq =
+			mlx5e_rx_mpwqe_is_linear_skb(mdev, params) ?
+			mlx5e_skb_from_cqe_mpwrq_linear :
+			mlx5e_skb_from_cqe_mpwrq_nonlinear;
+		rq->mpwqe.log_stride_sz = mlx5e_mpwqe_get_log_stride_size(mdev, params);
+		rq->mpwqe.num_strides = BIT(mlx5e_mpwqe_get_log_num_strides(mdev, params));
+
+		err = mlx5e_create_rq_umr_mkey(mdev, rq);
+		if (err)
+			goto err_rq_wq_destroy;
+		rq->mkey_be = cpu_to_be32(rq->umr_mkey.key);
+
+		err = mlx5e_rq_alloc_mpwqe_info(rq, c);
+		if (err)
+			goto err_free;
+		break;
+	default: /* MLX5_WQ_TYPE_CYCLIC */
+		err = mlx5_wq_cyc_create(mdev, &rqp->wq, rqc_wq, &rq->wqe.wq,
+					 &rq->wq_ctrl);
+		if (err)
+			goto err_rq_wq_destroy;
+
+		rq->wqe.wq.db = &rq->wqe.wq.db[MLX5_RCV_DBR];
+
+		wq_sz = mlx5_wq_cyc_get_size(&rq->wqe.wq);
+
+		rq->wqe.info = rqp->frags_info;
+		rq->wqe.frags =
+			kvzalloc_node(array_size(sizeof(*rq->wqe.frags),
+					(wq_sz << rq->wqe.info.log_num_frags)),
+				      GFP_KERNEL, cpu_to_node(c->cpu));
+		if (!rq->wqe.frags) {
+			err = -ENOMEM;
+			goto err_free;
+		}
+
+		err = mlx5e_init_di_list(rq, params, wq_sz, c->cpu);
+		if (err)
+			goto err_free;
+		rq->post_wqes = mlx5e_post_rx_wqes;
+		rq->dealloc_wqe = mlx5e_dealloc_rx_wqe;
+
+#ifdef CONFIG_MLX5_EN_IPSEC
+		if (c->priv->ipsec)
+			rq->handle_rx_cqe = mlx5e_ipsec_handle_rx_cqe;
+		else
+#endif
+			rq->handle_rx_cqe = c->priv->profile->rx_handlers.handle_rx_cqe;
+		if (!rq->handle_rx_cqe) {
+			err = -EINVAL;
+			netdev_err(c->netdev, "RX handler of RQ is not set, err %d\n", err);
+			goto err_free;
+		}
+
+		rq->wqe.skb_from_cqe = mlx5e_rx_is_linear_skb(mdev, params) ?
+			mlx5e_skb_from_cqe_linear :
+			mlx5e_skb_from_cqe_nonlinear;
+		rq->mkey_be = c->mkey_be;
+	}
+
+	/* Create a page_pool and register it with rxq */
+	pp_params.order     = 0;
+	pp_params.flags     = 0; /* No-internal DMA mapping in page_pool */
+	pp_params.pool_size = pool_size;
+	pp_params.nid       = cpu_to_node(c->cpu);
+	pp_params.dev       = c->pdev;
+	pp_params.dma_dir   = rq->buff.map_dir;
+
+	/* page_pool can be used even when there is no rq->xdp_prog,
+	 * given page_pool does not handle DMA mapping there is no
+	 * required state to clear. And page_pool gracefully handle
+	 * elevated refcnt.
+	 */
+	rq->page_pool = page_pool_create(&pp_params);
+	if (IS_ERR(rq->page_pool)) {
+		err = PTR_ERR(rq->page_pool);
+		rq->page_pool = NULL;
+		goto err_free;
+	}
+	err = xdp_rxq_info_reg_mem_model(&rq->xdp_rxq,
+					 MEM_TYPE_PAGE_POOL, rq->page_pool);
+	if (err)
+		goto err_free;
+
+	for (i = 0; i < wq_sz; i++) {
+		if (rq->wq_type == MLX5_WQ_TYPE_LINKED_LIST_STRIDING_RQ) {
+			struct mlx5e_rx_wqe_ll *wqe =
+				mlx5_wq_ll_get_wqe(&rq->mpwqe.wq, i);
+			u32 byte_count =
+				rq->mpwqe.num_strides << rq->mpwqe.log_stride_sz;
+			u64 dma_offset = mlx5e_get_mpwqe_offset(rq, i);
+
+			wqe->data[0].addr = cpu_to_be64(dma_offset + rq->buff.headroom);
+			wqe->data[0].byte_count = cpu_to_be32(byte_count);
+			wqe->data[0].lkey = rq->mkey_be;
+		} else {
+			struct mlx5e_rx_wqe_cyc *wqe =
+				mlx5_wq_cyc_get_wqe(&rq->wqe.wq, i);
+			int f;
+
+			for (f = 0; f < rq->wqe.info.num_frags; f++) {
+				u32 frag_size = rq->wqe.info.arr[f].frag_size |
+					MLX5_HW_START_PADDING;
+
+				wqe->data[f].byte_count = cpu_to_be32(frag_size);
+				wqe->data[f].lkey = rq->mkey_be;
+			}
+			/* check if num_frags is not a pow of two */
+			if (rq->wqe.info.num_frags < (1 << rq->wqe.info.log_num_frags)) {
+				wqe->data[f].byte_count = 0;
+				wqe->data[f].lkey = cpu_to_be32(MLX5_INVALID_LKEY);
+				wqe->data[f].addr = 0;
+			}
+		}
+	}
+
+	INIT_WORK(&rq->dim.work, mlx5e_rx_dim_work);
+
+	switch (params->rx_cq_moderation.cq_period_mode) {
+	case MLX5_CQ_PERIOD_MODE_START_FROM_CQE:
+		rq->dim.mode = NET_DIM_CQ_PERIOD_MODE_START_FROM_CQE;
+		break;
+	case MLX5_CQ_PERIOD_MODE_START_FROM_EQE:
+	default:
+		rq->dim.mode = NET_DIM_CQ_PERIOD_MODE_START_FROM_EQE;
+	}
+
+	rq->page_cache.head = 0;
+	rq->page_cache.tail = 0;
+
+	return 0;
+
+err_free:
+	switch (rq->wq_type) {
+	case MLX5_WQ_TYPE_LINKED_LIST_STRIDING_RQ:
+		kvfree(rq->mpwqe.info);
+		mlx5_core_destroy_mkey(mdev, &rq->umr_mkey);
+		break;
+	default: /* MLX5_WQ_TYPE_CYCLIC */
+		kvfree(rq->wqe.frags);
+		mlx5e_free_di_list(rq);
+	}
+
+err_rq_wq_destroy:
+	if (rq->xdp_prog)
+		bpf_prog_put(rq->xdp_prog);
+	xdp_rxq_info_unreg(&rq->xdp_rxq);
+	if (rq->page_pool)
+		page_pool_destroy(rq->page_pool);
+	mlx5_wq_destroy(&rq->wq_ctrl);
+
+	return err;
+}
+
+static void mlx5e_free_rq(struct mlx5e_rq *rq)
+{
+	int i;
+
+	if (rq->xdp_prog)
+		bpf_prog_put(rq->xdp_prog);
+
+	xdp_rxq_info_unreg(&rq->xdp_rxq);
+	if (rq->page_pool)
+		page_pool_destroy(rq->page_pool);
+
+	switch (rq->wq_type) {
+	case MLX5_WQ_TYPE_LINKED_LIST_STRIDING_RQ:
+		kvfree(rq->mpwqe.info);
+		mlx5_core_destroy_mkey(rq->mdev, &rq->umr_mkey);
+		break;
+	default: /* MLX5_WQ_TYPE_CYCLIC */
+		kvfree(rq->wqe.frags);
+		mlx5e_free_di_list(rq);
+	}
+
+	for (i = rq->page_cache.head; i != rq->page_cache.tail;
+	     i = (i + 1) & (MLX5E_CACHE_SIZE - 1)) {
+		struct mlx5e_dma_info *dma_info = &rq->page_cache.page_cache[i];
+
+		mlx5e_page_release(rq, dma_info, false);
+	}
+	mlx5_wq_destroy(&rq->wq_ctrl);
+}
+
+static int mlx5e_create_rq(struct mlx5e_rq *rq,
+			   struct mlx5e_rq_param *param)
+{
+	struct mlx5_core_dev *mdev = rq->mdev;
+
+	void *in;
+	void *rqc;
+	void *wq;
+	int inlen;
+	int err;
+
+	inlen = MLX5_ST_SZ_BYTES(create_rq_in) +
+		sizeof(u64) * rq->wq_ctrl.buf.npages;
+	in = kvzalloc(inlen, GFP_KERNEL);
+	if (!in)
+		return -ENOMEM;
+
+	rqc = MLX5_ADDR_OF(create_rq_in, in, ctx);
+	wq  = MLX5_ADDR_OF(rqc, rqc, wq);
+
+	memcpy(rqc, param->rqc, sizeof(param->rqc));
+
+	MLX5_SET(rqc,  rqc, cqn,		rq->cq.mcq.cqn);
+	MLX5_SET(rqc,  rqc, state,		MLX5_RQC_STATE_RST);
+	MLX5_SET(wq,   wq,  log_wq_pg_sz,	rq->wq_ctrl.buf.page_shift -
+						MLX5_ADAPTER_PAGE_SHIFT);
+	MLX5_SET64(wq, wq,  dbr_addr,		rq->wq_ctrl.db.dma);
+
+	mlx5_fill_page_frag_array(&rq->wq_ctrl.buf,
+				  (__be64 *)MLX5_ADDR_OF(wq, wq, pas));
+
+	err = mlx5_core_create_rq(mdev, in, inlen, &rq->rqn);
+
+	kvfree(in);
+
+	return err;
+}
+
+static int mlx5e_modify_rq_state(struct mlx5e_rq *rq, int curr_state,
+				 int next_state)
+{
+	struct mlx5_core_dev *mdev = rq->mdev;
+
+	void *in;
+	void *rqc;
+	int inlen;
+	int err;
+
+	inlen = MLX5_ST_SZ_BYTES(modify_rq_in);
+	in = kvzalloc(inlen, GFP_KERNEL);
+	if (!in)
+		return -ENOMEM;
+
+	rqc = MLX5_ADDR_OF(modify_rq_in, in, ctx);
+
+	MLX5_SET(modify_rq_in, in, rq_state, curr_state);
+	MLX5_SET(rqc, rqc, state, next_state);
+
+	err = mlx5_core_modify_rq(mdev, rq->rqn, in, inlen);
+
+	kvfree(in);
+
+	return err;
+}
+
+static int mlx5e_modify_rq_scatter_fcs(struct mlx5e_rq *rq, bool enable)
+{
+	struct mlx5e_channel *c = rq->channel;
+	struct mlx5e_priv *priv = c->priv;
+	struct mlx5_core_dev *mdev = priv->mdev;
+
+	void *in;
+	void *rqc;
+	int inlen;
+	int err;
+
+	inlen = MLX5_ST_SZ_BYTES(modify_rq_in);
+	in = kvzalloc(inlen, GFP_KERNEL);
+	if (!in)
+		return -ENOMEM;
+
+	rqc = MLX5_ADDR_OF(modify_rq_in, in, ctx);
+
+	MLX5_SET(modify_rq_in, in, rq_state, MLX5_RQC_STATE_RDY);
+	MLX5_SET64(modify_rq_in, in, modify_bitmask,
+		   MLX5_MODIFY_RQ_IN_MODIFY_BITMASK_SCATTER_FCS);
+	MLX5_SET(rqc, rqc, scatter_fcs, enable);
+	MLX5_SET(rqc, rqc, state, MLX5_RQC_STATE_RDY);
+
+	err = mlx5_core_modify_rq(mdev, rq->rqn, in, inlen);
+
+	kvfree(in);
+
+	return err;
+}
+
+static int mlx5e_modify_rq_vsd(struct mlx5e_rq *rq, bool vsd)
+{
+	struct mlx5e_channel *c = rq->channel;
+	struct mlx5_core_dev *mdev = c->mdev;
+	void *in;
+	void *rqc;
+	int inlen;
+	int err;
+
+	inlen = MLX5_ST_SZ_BYTES(modify_rq_in);
+	in = kvzalloc(inlen, GFP_KERNEL);
+	if (!in)
+		return -ENOMEM;
+
+	rqc = MLX5_ADDR_OF(modify_rq_in, in, ctx);
+
+	MLX5_SET(modify_rq_in, in, rq_state, MLX5_RQC_STATE_RDY);
+	MLX5_SET64(modify_rq_in, in, modify_bitmask,
+		   MLX5_MODIFY_RQ_IN_MODIFY_BITMASK_VSD);
+	MLX5_SET(rqc, rqc, vsd, vsd);
+	MLX5_SET(rqc, rqc, state, MLX5_RQC_STATE_RDY);
+
+	err = mlx5_core_modify_rq(mdev, rq->rqn, in, inlen);
+
+	kvfree(in);
+
+	return err;
+}
+
+static void mlx5e_destroy_rq(struct mlx5e_rq *rq)
+{
+	mlx5_core_destroy_rq(rq->mdev, rq->rqn);
+}
+
+static int mlx5e_wait_for_min_rx_wqes(struct mlx5e_rq *rq, int wait_time)
+{
+	unsigned long exp_time = jiffies + msecs_to_jiffies(wait_time);
+	struct mlx5e_channel *c = rq->channel;
+
+	u16 min_wqes = mlx5_min_rx_wqes(rq->wq_type, mlx5e_rqwq_get_size(rq));
+
+	do {
+		if (mlx5e_rqwq_get_cur_sz(rq) >= min_wqes)
+			return 0;
+
+		msleep(20);
+	} while (time_before(jiffies, exp_time));
+
+	netdev_warn(c->netdev, "Failed to get min RX wqes on Channel[%d] RQN[0x%x] wq cur_sz(%d) min_rx_wqes(%d)\n",
+		    c->ix, rq->rqn, mlx5e_rqwq_get_cur_sz(rq), min_wqes);
+
+	return -ETIMEDOUT;
+}
+
+static void mlx5e_free_rx_descs(struct mlx5e_rq *rq)
+{
+	__be16 wqe_ix_be;
+	u16 wqe_ix;
+
+	if (rq->wq_type == MLX5_WQ_TYPE_LINKED_LIST_STRIDING_RQ) {
+		struct mlx5_wq_ll *wq = &rq->mpwqe.wq;
+
+		/* UMR WQE (if in progress) is always at wq->head */
+		if (rq->mpwqe.umr_in_progress)
+			rq->dealloc_wqe(rq, wq->head);
+
+		while (!mlx5_wq_ll_is_empty(wq)) {
+			struct mlx5e_rx_wqe_ll *wqe;
+
+			wqe_ix_be = *wq->tail_next;
+			wqe_ix    = be16_to_cpu(wqe_ix_be);
+			wqe       = mlx5_wq_ll_get_wqe(wq, wqe_ix);
+			rq->dealloc_wqe(rq, wqe_ix);
+			mlx5_wq_ll_pop(wq, wqe_ix_be,
+				       &wqe->next.next_wqe_index);
+		}
+	} else {
+		struct mlx5_wq_cyc *wq = &rq->wqe.wq;
+
+		while (!mlx5_wq_cyc_is_empty(wq)) {
+			wqe_ix = mlx5_wq_cyc_get_tail(wq);
+			rq->dealloc_wqe(rq, wqe_ix);
+			mlx5_wq_cyc_pop(wq);
+		}
+	}
+
+}
+
+static int mlx5e_open_rq(struct mlx5e_channel *c,
+			 struct mlx5e_params *params,
+			 struct mlx5e_rq_param *param,
+			 struct mlx5e_rq *rq)
+{
+	int err;
+
+	err = mlx5e_alloc_rq(c, params, param, rq);
+	if (err)
+		return err;
+
+	err = mlx5e_create_rq(rq, param);
+	if (err)
+		goto err_free_rq;
+
+	err = mlx5e_modify_rq_state(rq, MLX5_RQC_STATE_RST, MLX5_RQC_STATE_RDY);
+	if (err)
+		goto err_destroy_rq;
+
+	if (params->rx_dim_enabled)
+		__set_bit(MLX5E_RQ_STATE_AM, &c->rq.state);
+
+	/* We disable csum_complete when XDP is enabled since
+	 * XDP programs might manipulate packets which will render
+	 * skb->checksum incorrect.
+	 */
+	if (MLX5E_GET_PFLAG(params, MLX5E_PFLAG_RX_NO_CSUM_COMPLETE) || c->xdp)
+		__set_bit(MLX5E_RQ_STATE_NO_CSUM_COMPLETE, &c->rq.state);
+
+	return 0;
+
+err_destroy_rq:
+	mlx5e_destroy_rq(rq);
+err_free_rq:
+	mlx5e_free_rq(rq);
+
+	return err;
+}
+
+static void mlx5e_activate_rq(struct mlx5e_rq *rq)
+{
+	struct mlx5e_icosq *sq = &rq->channel->icosq;
+	struct mlx5_wq_cyc *wq = &sq->wq;
+	struct mlx5e_tx_wqe *nopwqe;
+
+	u16 pi = mlx5_wq_cyc_ctr2ix(wq, sq->pc);
+
+	set_bit(MLX5E_RQ_STATE_ENABLED, &rq->state);
+	sq->db.ico_wqe[pi].opcode     = MLX5_OPCODE_NOP;
+	nopwqe = mlx5e_post_nop(wq, sq->sqn, &sq->pc);
+	mlx5e_notify_hw(wq, sq->pc, sq->uar_map, &nopwqe->ctrl);
+}
+
+static void mlx5e_deactivate_rq(struct mlx5e_rq *rq)
+{
+	clear_bit(MLX5E_RQ_STATE_ENABLED, &rq->state);
+	napi_synchronize(&rq->channel->napi); /* prevent mlx5e_post_rx_wqes */
+}
+
+static void mlx5e_close_rq(struct mlx5e_rq *rq)
+{
+	cancel_work_sync(&rq->dim.work);
+	mlx5e_destroy_rq(rq);
+	mlx5e_free_rx_descs(rq);
+	mlx5e_free_rq(rq);
+}
+
+static void mlx5e_free_xdpsq_db(struct mlx5e_xdpsq *sq)
+{
+	kvfree(sq->db.xdpi);
+}
+
+static int mlx5e_alloc_xdpsq_db(struct mlx5e_xdpsq *sq, int numa)
+{
+	int wq_sz = mlx5_wq_cyc_get_size(&sq->wq);
+
+	sq->db.xdpi = kvzalloc_node(array_size(wq_sz, sizeof(*sq->db.xdpi)),
+				    GFP_KERNEL, numa);
+	if (!sq->db.xdpi) {
+		mlx5e_free_xdpsq_db(sq);
+		return -ENOMEM;
+	}
+
+	return 0;
+}
+
+static int mlx5e_alloc_xdpsq(struct mlx5e_channel *c,
+			     struct mlx5e_params *params,
+			     struct mlx5e_sq_param *param,
+			     struct mlx5e_xdpsq *sq,
+			     bool is_redirect)
+{
+	void *sqc_wq               = MLX5_ADDR_OF(sqc, param->sqc, wq);
+	struct mlx5_core_dev *mdev = c->mdev;
+	struct mlx5_wq_cyc *wq = &sq->wq;
+	int err;
+
+	sq->pdev      = c->pdev;
+	sq->mkey_be   = c->mkey_be;
+	sq->channel   = c;
+	sq->uar_map   = mdev->mlx5e_res.bfreg.map;
+	sq->min_inline_mode = params->tx_min_inline_mode;
+	sq->hw_mtu    = MLX5E_SW2HW_MTU(params, params->sw_mtu);
+	sq->stats     = is_redirect ?
+		&c->priv->channel_stats[c->ix].xdpsq :
+		&c->priv->channel_stats[c->ix].rq_xdpsq;
+
+	param->wq.db_numa_node = cpu_to_node(c->cpu);
+	err = mlx5_wq_cyc_create(mdev, &param->wq, sqc_wq, wq, &sq->wq_ctrl);
+	if (err)
+		return err;
+	wq->db = &wq->db[MLX5_SND_DBR];
+
+	err = mlx5e_alloc_xdpsq_db(sq, cpu_to_node(c->cpu));
+	if (err)
+		goto err_sq_wq_destroy;
+
+	return 0;
+
+err_sq_wq_destroy:
+	mlx5_wq_destroy(&sq->wq_ctrl);
+
+	return err;
+}
+
+static void mlx5e_free_xdpsq(struct mlx5e_xdpsq *sq)
+{
+	mlx5e_free_xdpsq_db(sq);
+	mlx5_wq_destroy(&sq->wq_ctrl);
+}
+
+static void mlx5e_free_icosq_db(struct mlx5e_icosq *sq)
+{
+	kvfree(sq->db.ico_wqe);
+}
+
+static int mlx5e_alloc_icosq_db(struct mlx5e_icosq *sq, int numa)
+{
+	u8 wq_sz = mlx5_wq_cyc_get_size(&sq->wq);
+
+	sq->db.ico_wqe = kvzalloc_node(array_size(wq_sz,
+						  sizeof(*sq->db.ico_wqe)),
+				       GFP_KERNEL, numa);
+	if (!sq->db.ico_wqe)
+		return -ENOMEM;
+
+	return 0;
+}
+
+static int mlx5e_alloc_icosq(struct mlx5e_channel *c,
+			     struct mlx5e_sq_param *param,
+			     struct mlx5e_icosq *sq)
+{
+	void *sqc_wq               = MLX5_ADDR_OF(sqc, param->sqc, wq);
+	struct mlx5_core_dev *mdev = c->mdev;
+	struct mlx5_wq_cyc *wq = &sq->wq;
+	int err;
+
+	sq->channel   = c;
+	sq->uar_map   = mdev->mlx5e_res.bfreg.map;
+
+	param->wq.db_numa_node = cpu_to_node(c->cpu);
+	err = mlx5_wq_cyc_create(mdev, &param->wq, sqc_wq, wq, &sq->wq_ctrl);
+	if (err)
+		return err;
+	wq->db = &wq->db[MLX5_SND_DBR];
+
+	err = mlx5e_alloc_icosq_db(sq, cpu_to_node(c->cpu));
+	if (err)
+		goto err_sq_wq_destroy;
+
+	return 0;
+
+err_sq_wq_destroy:
+	mlx5_wq_destroy(&sq->wq_ctrl);
+
+	return err;
+}
+
+static void mlx5e_free_icosq(struct mlx5e_icosq *sq)
+{
+	mlx5e_free_icosq_db(sq);
+	mlx5_wq_destroy(&sq->wq_ctrl);
+}
+
+static void mlx5e_free_txqsq_db(struct mlx5e_txqsq *sq)
+{
+	kvfree(sq->db.wqe_info);
+	kvfree(sq->db.dma_fifo);
+}
+
+static int mlx5e_alloc_txqsq_db(struct mlx5e_txqsq *sq, int numa)
+{
+	int wq_sz = mlx5_wq_cyc_get_size(&sq->wq);
+	int df_sz = wq_sz * MLX5_SEND_WQEBB_NUM_DS;
+
+	sq->db.dma_fifo = kvzalloc_node(array_size(df_sz,
+						   sizeof(*sq->db.dma_fifo)),
+					GFP_KERNEL, numa);
+	sq->db.wqe_info = kvzalloc_node(array_size(wq_sz,
+						   sizeof(*sq->db.wqe_info)),
+					GFP_KERNEL, numa);
+	if (!sq->db.dma_fifo || !sq->db.wqe_info) {
+		mlx5e_free_txqsq_db(sq);
+		return -ENOMEM;
+	}
+
+	sq->dma_fifo_mask = df_sz - 1;
+
+	return 0;
+}
+
+static void mlx5e_sq_recover(struct work_struct *work);
+static int mlx5e_alloc_txqsq(struct mlx5e_channel *c,
+			     int txq_ix,
+			     struct mlx5e_params *params,
+			     struct mlx5e_sq_param *param,
+			     struct mlx5e_txqsq *sq,
+			     int tc)
+{
+	void *sqc_wq               = MLX5_ADDR_OF(sqc, param->sqc, wq);
+	struct mlx5_core_dev *mdev = c->mdev;
+	struct mlx5_wq_cyc *wq = &sq->wq;
+	int err;
+
+	sq->pdev      = c->pdev;
+	sq->tstamp    = c->tstamp;
+	sq->clock     = &mdev->clock;
+	sq->mkey_be   = c->mkey_be;
+	sq->channel   = c;
+	sq->txq_ix    = txq_ix;
+	sq->uar_map   = mdev->mlx5e_res.bfreg.map;
+	sq->min_inline_mode = params->tx_min_inline_mode;
+	sq->stats     = &c->priv->channel_stats[c->ix].sq[tc];
+	INIT_WORK(&sq->recover.recover_work, mlx5e_sq_recover);
+	if (MLX5_IPSEC_DEV(c->priv->mdev))
+		set_bit(MLX5E_SQ_STATE_IPSEC, &sq->state);
+	if (mlx5_accel_is_tls_device(c->priv->mdev))
+		set_bit(MLX5E_SQ_STATE_TLS, &sq->state);
+
+	param->wq.db_numa_node = cpu_to_node(c->cpu);
+	err = mlx5_wq_cyc_create(mdev, &param->wq, sqc_wq, wq, &sq->wq_ctrl);
+	if (err)
+		return err;
+	wq->db    = &wq->db[MLX5_SND_DBR];
+
+	err = mlx5e_alloc_txqsq_db(sq, cpu_to_node(c->cpu));
+	if (err)
+		goto err_sq_wq_destroy;
+
+	INIT_WORK(&sq->dim.work, mlx5e_tx_dim_work);
+	sq->dim.mode = params->tx_cq_moderation.cq_period_mode;
+
+	return 0;
+
+err_sq_wq_destroy:
+	mlx5_wq_destroy(&sq->wq_ctrl);
+
+	return err;
+}
+
+static void mlx5e_free_txqsq(struct mlx5e_txqsq *sq)
+{
+	mlx5e_free_txqsq_db(sq);
+	mlx5_wq_destroy(&sq->wq_ctrl);
+}
+
+struct mlx5e_create_sq_param {
+	struct mlx5_wq_ctrl        *wq_ctrl;
+	u32                         cqn;
+	u32                         tisn;
+	u8                          tis_lst_sz;
+	u8                          min_inline_mode;
+};
+
+static int mlx5e_create_sq(struct mlx5_core_dev *mdev,
+			   struct mlx5e_sq_param *param,
+			   struct mlx5e_create_sq_param *csp,
+			   u32 *sqn)
+{
+	void *in;
+	void *sqc;
+	void *wq;
+	int inlen;
+	int err;
+
+	inlen = MLX5_ST_SZ_BYTES(create_sq_in) +
+		sizeof(u64) * csp->wq_ctrl->buf.npages;
+	in = kvzalloc(inlen, GFP_KERNEL);
+	if (!in)
+		return -ENOMEM;
+
+	sqc = MLX5_ADDR_OF(create_sq_in, in, ctx);
+	wq = MLX5_ADDR_OF(sqc, sqc, wq);
+
+	memcpy(sqc, param->sqc, sizeof(param->sqc));
+	MLX5_SET(sqc,  sqc, tis_lst_sz, csp->tis_lst_sz);
+	MLX5_SET(sqc,  sqc, tis_num_0, csp->tisn);
+	MLX5_SET(sqc,  sqc, cqn, csp->cqn);
+
+	if (MLX5_CAP_ETH(mdev, wqe_inline_mode) == MLX5_CAP_INLINE_MODE_VPORT_CONTEXT)
+		MLX5_SET(sqc,  sqc, min_wqe_inline_mode, csp->min_inline_mode);
+
+	MLX5_SET(sqc,  sqc, state, MLX5_SQC_STATE_RST);
+	MLX5_SET(sqc,  sqc, flush_in_error_en, 1);
+
+	MLX5_SET(wq,   wq, wq_type,       MLX5_WQ_TYPE_CYCLIC);
+	MLX5_SET(wq,   wq, uar_page,      mdev->mlx5e_res.bfreg.index);
+	MLX5_SET(wq,   wq, log_wq_pg_sz,  csp->wq_ctrl->buf.page_shift -
+					  MLX5_ADAPTER_PAGE_SHIFT);
+	MLX5_SET64(wq, wq, dbr_addr,      csp->wq_ctrl->db.dma);
+
+	mlx5_fill_page_frag_array(&csp->wq_ctrl->buf,
+				  (__be64 *)MLX5_ADDR_OF(wq, wq, pas));
+
+	err = mlx5_core_create_sq(mdev, in, inlen, sqn);
+
+	kvfree(in);
+
+	return err;
+}
+
+struct mlx5e_modify_sq_param {
+	int curr_state;
+	int next_state;
+	bool rl_update;
+	int rl_index;
+};
+
+static int mlx5e_modify_sq(struct mlx5_core_dev *mdev, u32 sqn,
+			   struct mlx5e_modify_sq_param *p)
+{
+	void *in;
+	void *sqc;
+	int inlen;
+	int err;
+
+	inlen = MLX5_ST_SZ_BYTES(modify_sq_in);
+	in = kvzalloc(inlen, GFP_KERNEL);
+	if (!in)
+		return -ENOMEM;
+
+	sqc = MLX5_ADDR_OF(modify_sq_in, in, ctx);
+
+	MLX5_SET(modify_sq_in, in, sq_state, p->curr_state);
+	MLX5_SET(sqc, sqc, state, p->next_state);
+	if (p->rl_update && p->next_state == MLX5_SQC_STATE_RDY) {
+		MLX5_SET64(modify_sq_in, in, modify_bitmask, 1);
+		MLX5_SET(sqc,  sqc, packet_pacing_rate_limit_index, p->rl_index);
+	}
+
+	err = mlx5_core_modify_sq(mdev, sqn, in, inlen);
+
+	kvfree(in);
+
+	return err;
+}
+
+static void mlx5e_destroy_sq(struct mlx5_core_dev *mdev, u32 sqn)
+{
+	mlx5_core_destroy_sq(mdev, sqn);
+}
+
+static int mlx5e_create_sq_rdy(struct mlx5_core_dev *mdev,
+			       struct mlx5e_sq_param *param,
+			       struct mlx5e_create_sq_param *csp,
+			       u32 *sqn)
+{
+	struct mlx5e_modify_sq_param msp = {0};
+	int err;
+
+	err = mlx5e_create_sq(mdev, param, csp, sqn);
+	if (err)
+		return err;
+
+	msp.curr_state = MLX5_SQC_STATE_RST;
+	msp.next_state = MLX5_SQC_STATE_RDY;
+	err = mlx5e_modify_sq(mdev, *sqn, &msp);
+	if (err)
+		mlx5e_destroy_sq(mdev, *sqn);
+
+	return err;
+}
+
+static int mlx5e_set_sq_maxrate(struct net_device *dev,
+				struct mlx5e_txqsq *sq, u32 rate);
+
+static int mlx5e_open_txqsq(struct mlx5e_channel *c,
+			    u32 tisn,
+			    int txq_ix,
+			    struct mlx5e_params *params,
+			    struct mlx5e_sq_param *param,
+			    struct mlx5e_txqsq *sq,
+			    int tc)
+{
+	struct mlx5e_create_sq_param csp = {};
+	u32 tx_rate;
+	int err;
+
+	err = mlx5e_alloc_txqsq(c, txq_ix, params, param, sq, tc);
+	if (err)
+		return err;
+
+	csp.tisn            = tisn;
+	csp.tis_lst_sz      = 1;
+	csp.cqn             = sq->cq.mcq.cqn;
+	csp.wq_ctrl         = &sq->wq_ctrl;
+	csp.min_inline_mode = sq->min_inline_mode;
+	err = mlx5e_create_sq_rdy(c->mdev, param, &csp, &sq->sqn);
+	if (err)
+		goto err_free_txqsq;
+
+	tx_rate = c->priv->tx_rates[sq->txq_ix];
+	if (tx_rate)
+		mlx5e_set_sq_maxrate(c->netdev, sq, tx_rate);
+
+	if (params->tx_dim_enabled)
+		sq->state |= BIT(MLX5E_SQ_STATE_AM);
+
+	return 0;
+
+err_free_txqsq:
+	clear_bit(MLX5E_SQ_STATE_ENABLED, &sq->state);
+	mlx5e_free_txqsq(sq);
+
+	return err;
+}
+
+static void mlx5e_reset_txqsq_cc_pc(struct mlx5e_txqsq *sq)
+{
+	WARN_ONCE(sq->cc != sq->pc,
+		  "SQ 0x%x: cc (0x%x) != pc (0x%x)\n",
+		  sq->sqn, sq->cc, sq->pc);
+	sq->cc = 0;
+	sq->dma_fifo_cc = 0;
+	sq->pc = 0;
+}
+
+static void mlx5e_activate_txqsq(struct mlx5e_txqsq *sq)
+{
+	sq->txq = netdev_get_tx_queue(sq->channel->netdev, sq->txq_ix);
+	clear_bit(MLX5E_SQ_STATE_RECOVERING, &sq->state);
+	set_bit(MLX5E_SQ_STATE_ENABLED, &sq->state);
+	netdev_tx_reset_queue(sq->txq);
+	netif_tx_start_queue(sq->txq);
+}
+
+static inline void netif_tx_disable_queue(struct netdev_queue *txq)
+{
+	__netif_tx_lock_bh(txq);
+	netif_tx_stop_queue(txq);
+	__netif_tx_unlock_bh(txq);
+}
+
+static void mlx5e_deactivate_txqsq(struct mlx5e_txqsq *sq)
+{
+	struct mlx5e_channel *c = sq->channel;
+	struct mlx5_wq_cyc *wq = &sq->wq;
+
+	clear_bit(MLX5E_SQ_STATE_ENABLED, &sq->state);
+	/* prevent netif_tx_wake_queue */
+	napi_synchronize(&c->napi);
+
+	netif_tx_disable_queue(sq->txq);
+
+	/* last doorbell out, godspeed .. */
+	if (mlx5e_wqc_has_room_for(wq, sq->cc, sq->pc, 1)) {
+		u16 pi = mlx5_wq_cyc_ctr2ix(wq, sq->pc);
+		struct mlx5e_tx_wqe *nop;
+
+		sq->db.wqe_info[pi].skb = NULL;
+		nop = mlx5e_post_nop(wq, sq->sqn, &sq->pc);
+		mlx5e_notify_hw(wq, sq->pc, sq->uar_map, &nop->ctrl);
+	}
+}
+
+static void mlx5e_close_txqsq(struct mlx5e_txqsq *sq)
+{
+	struct mlx5e_channel *c = sq->channel;
+	struct mlx5_core_dev *mdev = c->mdev;
+	struct mlx5_rate_limit rl = {0};
+
+	cancel_work_sync(&sq->dim.work);
+	mlx5e_destroy_sq(mdev, sq->sqn);
+	if (sq->rate_limit) {
+		rl.rate = sq->rate_limit;
+		mlx5_rl_remove_rate(mdev, &rl);
+	}
+	mlx5e_free_txqsq_descs(sq);
+	mlx5e_free_txqsq(sq);
+}
+
+static int mlx5e_wait_for_sq_flush(struct mlx5e_txqsq *sq)
+{
+	unsigned long exp_time = jiffies + msecs_to_jiffies(2000);
+
+	while (time_before(jiffies, exp_time)) {
+		if (sq->cc == sq->pc)
+			return 0;
+
+		msleep(20);
+	}
+
+	netdev_err(sq->channel->netdev,
+		   "Wait for SQ 0x%x flush timeout (sq cc = 0x%x, sq pc = 0x%x)\n",
+		   sq->sqn, sq->cc, sq->pc);
+
+	return -ETIMEDOUT;
+}
+
+static int mlx5e_sq_to_ready(struct mlx5e_txqsq *sq, int curr_state)
+{
+	struct mlx5_core_dev *mdev = sq->channel->mdev;
+	struct net_device *dev = sq->channel->netdev;
+	struct mlx5e_modify_sq_param msp = {0};
+	int err;
+
+	msp.curr_state = curr_state;
+	msp.next_state = MLX5_SQC_STATE_RST;
+
+	err = mlx5e_modify_sq(mdev, sq->sqn, &msp);
+	if (err) {
+		netdev_err(dev, "Failed to move sq 0x%x to reset\n", sq->sqn);
+		return err;
+	}
+
+	memset(&msp, 0, sizeof(msp));
+	msp.curr_state = MLX5_SQC_STATE_RST;
+	msp.next_state = MLX5_SQC_STATE_RDY;
+
+	err = mlx5e_modify_sq(mdev, sq->sqn, &msp);
+	if (err) {
+		netdev_err(dev, "Failed to move sq 0x%x to ready\n", sq->sqn);
+		return err;
+	}
+
+	return 0;
+}
+
+static void mlx5e_sq_recover(struct work_struct *work)
+{
+	struct mlx5e_txqsq_recover *recover =
+		container_of(work, struct mlx5e_txqsq_recover,
+			     recover_work);
+	struct mlx5e_txqsq *sq = container_of(recover, struct mlx5e_txqsq,
+					      recover);
+	struct mlx5_core_dev *mdev = sq->channel->mdev;
+	struct net_device *dev = sq->channel->netdev;
+	u8 state;
+	int err;
+
+	err = mlx5_core_query_sq_state(mdev, sq->sqn, &state);
+	if (err) {
+		netdev_err(dev, "Failed to query SQ 0x%x state. err = %d\n",
+			   sq->sqn, err);
+		return;
+	}
+
+	if (state != MLX5_RQC_STATE_ERR) {
+		netdev_err(dev, "SQ 0x%x not in ERROR state\n", sq->sqn);
+		return;
+	}
+
+	netif_tx_disable_queue(sq->txq);
+
+	if (mlx5e_wait_for_sq_flush(sq))
+		return;
+
+	/* If the interval between two consecutive recovers per SQ is too
+	 * short, don't recover to avoid infinite loop of ERR_CQE -> recover.
+	 * If we reached this state, there is probably a bug that needs to be
+	 * fixed. let's keep the queue close and let tx timeout cleanup.
+	 */
+	if (jiffies_to_msecs(jiffies - recover->last_recover) <
+	    MLX5E_SQ_RECOVER_MIN_INTERVAL) {
+		netdev_err(dev, "Recover SQ 0x%x canceled, too many error CQEs\n",
+			   sq->sqn);
+		return;
+	}
+
+	/* At this point, no new packets will arrive from the stack as TXQ is
+	 * marked with QUEUE_STATE_DRV_XOFF. In addition, NAPI cleared all
+	 * pending WQEs.  SQ can safely reset the SQ.
+	 */
+	if (mlx5e_sq_to_ready(sq, state))
+		return;
+
+	mlx5e_reset_txqsq_cc_pc(sq);
+	sq->stats->recover++;
+	recover->last_recover = jiffies;
+	mlx5e_activate_txqsq(sq);
+}
+
+static int mlx5e_open_icosq(struct mlx5e_channel *c,
+			    struct mlx5e_params *params,
+			    struct mlx5e_sq_param *param,
+			    struct mlx5e_icosq *sq)
+{
+	struct mlx5e_create_sq_param csp = {};
+	int err;
+
+	err = mlx5e_alloc_icosq(c, param, sq);
+	if (err)
+		return err;
+
+	csp.cqn             = sq->cq.mcq.cqn;
+	csp.wq_ctrl         = &sq->wq_ctrl;
+	csp.min_inline_mode = params->tx_min_inline_mode;
+	set_bit(MLX5E_SQ_STATE_ENABLED, &sq->state);
+	err = mlx5e_create_sq_rdy(c->mdev, param, &csp, &sq->sqn);
+	if (err)
+		goto err_free_icosq;
+
+	return 0;
+
+err_free_icosq:
+	clear_bit(MLX5E_SQ_STATE_ENABLED, &sq->state);
+	mlx5e_free_icosq(sq);
+
+	return err;
+}
+
+static void mlx5e_close_icosq(struct mlx5e_icosq *sq)
+{
+	struct mlx5e_channel *c = sq->channel;
+
+	clear_bit(MLX5E_SQ_STATE_ENABLED, &sq->state);
+	napi_synchronize(&c->napi);
+
+	mlx5e_destroy_sq(c->mdev, sq->sqn);
+	mlx5e_free_icosq(sq);
+}
+
+static int mlx5e_open_xdpsq(struct mlx5e_channel *c,
+			    struct mlx5e_params *params,
+			    struct mlx5e_sq_param *param,
+			    struct mlx5e_xdpsq *sq,
+			    bool is_redirect)
+{
+	unsigned int ds_cnt = MLX5E_XDP_TX_DS_COUNT;
+	struct mlx5e_create_sq_param csp = {};
+	unsigned int inline_hdr_sz = 0;
+	int err;
+	int i;
+
+	err = mlx5e_alloc_xdpsq(c, params, param, sq, is_redirect);
+	if (err)
+		return err;
+
+	csp.tis_lst_sz      = 1;
+	csp.tisn            = c->priv->tisn[0]; /* tc = 0 */
+	csp.cqn             = sq->cq.mcq.cqn;
+	csp.wq_ctrl         = &sq->wq_ctrl;
+	csp.min_inline_mode = sq->min_inline_mode;
+	if (is_redirect)
+		set_bit(MLX5E_SQ_STATE_REDIRECT, &sq->state);
+	set_bit(MLX5E_SQ_STATE_ENABLED, &sq->state);
+	err = mlx5e_create_sq_rdy(c->mdev, param, &csp, &sq->sqn);
+	if (err)
+		goto err_free_xdpsq;
+
+	if (sq->min_inline_mode != MLX5_INLINE_MODE_NONE) {
+		inline_hdr_sz = MLX5E_XDP_MIN_INLINE;
+		ds_cnt++;
+	}
+
+	/* Pre initialize fixed WQE fields */
+	for (i = 0; i < mlx5_wq_cyc_get_size(&sq->wq); i++) {
+		struct mlx5e_tx_wqe      *wqe  = mlx5_wq_cyc_get_wqe(&sq->wq, i);
+		struct mlx5_wqe_ctrl_seg *cseg = &wqe->ctrl;
+		struct mlx5_wqe_eth_seg  *eseg = &wqe->eth;
+		struct mlx5_wqe_data_seg *dseg;
+
+		cseg->qpn_ds = cpu_to_be32((sq->sqn << 8) | ds_cnt);
+		eseg->inline_hdr.sz = cpu_to_be16(inline_hdr_sz);
+
+		dseg = (struct mlx5_wqe_data_seg *)cseg + (ds_cnt - 1);
+		dseg->lkey = sq->mkey_be;
+	}
+
+	return 0;
+
+err_free_xdpsq:
+	clear_bit(MLX5E_SQ_STATE_ENABLED, &sq->state);
+	mlx5e_free_xdpsq(sq);
+
+	return err;
+}
+
+static void mlx5e_close_xdpsq(struct mlx5e_xdpsq *sq)
+{
+	struct mlx5e_channel *c = sq->channel;
+
+	clear_bit(MLX5E_SQ_STATE_ENABLED, &sq->state);
+	napi_synchronize(&c->napi);
+
+	mlx5e_destroy_sq(c->mdev, sq->sqn);
+	mlx5e_free_xdpsq_descs(sq);
+	mlx5e_free_xdpsq(sq);
+}
+
+static int mlx5e_alloc_cq_common(struct mlx5_core_dev *mdev,
+				 struct mlx5e_cq_param *param,
+				 struct mlx5e_cq *cq)
+{
+	struct mlx5_core_cq *mcq = &cq->mcq;
+	int eqn_not_used;
+	unsigned int irqn;
+	int err;
+	u32 i;
+
+	err = mlx5_vector2eqn(mdev, param->eq_ix, &eqn_not_used, &irqn);
+	if (err)
+		return err;
+
+	err = mlx5_cqwq_create(mdev, &param->wq, param->cqc, &cq->wq,
+			       &cq->wq_ctrl);
+	if (err)
+		return err;
+
+	mcq->cqe_sz     = 64;
+	mcq->set_ci_db  = cq->wq_ctrl.db.db;
+	mcq->arm_db     = cq->wq_ctrl.db.db + 1;
+	*mcq->set_ci_db = 0;
+	*mcq->arm_db    = 0;
+	mcq->vector     = param->eq_ix;
+	mcq->comp       = mlx5e_completion_event;
+	mcq->event      = mlx5e_cq_error_event;
+	mcq->irqn       = irqn;
+
+	for (i = 0; i < mlx5_cqwq_get_size(&cq->wq); i++) {
+		struct mlx5_cqe64 *cqe = mlx5_cqwq_get_wqe(&cq->wq, i);
+
+		cqe->op_own = 0xf1;
+	}
+
+	cq->mdev = mdev;
+
+	return 0;
+}
+
+static int mlx5e_alloc_cq(struct mlx5e_channel *c,
+			  struct mlx5e_cq_param *param,
+			  struct mlx5e_cq *cq)
+{
+	struct mlx5_core_dev *mdev = c->priv->mdev;
+	int err;
+
+	param->wq.buf_numa_node = cpu_to_node(c->cpu);
+	param->wq.db_numa_node  = cpu_to_node(c->cpu);
+	param->eq_ix   = c->ix;
+
+	err = mlx5e_alloc_cq_common(mdev, param, cq);
+
+	cq->napi    = &c->napi;
+	cq->channel = c;
+
+	return err;
+}
+
+static void mlx5e_free_cq(struct mlx5e_cq *cq)
+{
+	mlx5_wq_destroy(&cq->wq_ctrl);
+}
+
+static int mlx5e_create_cq(struct mlx5e_cq *cq, struct mlx5e_cq_param *param)
+{
+	struct mlx5_core_dev *mdev = cq->mdev;
+	struct mlx5_core_cq *mcq = &cq->mcq;
+
+	void *in;
+	void *cqc;
+	int inlen;
+	unsigned int irqn_not_used;
+	int eqn;
+	int err;
+
+	err = mlx5_vector2eqn(mdev, param->eq_ix, &eqn, &irqn_not_used);
+	if (err)
+		return err;
+
+	inlen = MLX5_ST_SZ_BYTES(create_cq_in) +
+		sizeof(u64) * cq->wq_ctrl.buf.npages;
+	in = kvzalloc(inlen, GFP_KERNEL);
+	if (!in)
+		return -ENOMEM;
+
+	cqc = MLX5_ADDR_OF(create_cq_in, in, cq_context);
+
+	memcpy(cqc, param->cqc, sizeof(param->cqc));
+
+	mlx5_fill_page_frag_array(&cq->wq_ctrl.buf,
+				  (__be64 *)MLX5_ADDR_OF(create_cq_in, in, pas));
+
+	MLX5_SET(cqc,   cqc, cq_period_mode, param->cq_period_mode);
+	MLX5_SET(cqc,   cqc, c_eqn,         eqn);
+	MLX5_SET(cqc,   cqc, uar_page,      mdev->priv.uar->index);
+	MLX5_SET(cqc,   cqc, log_page_size, cq->wq_ctrl.buf.page_shift -
+					    MLX5_ADAPTER_PAGE_SHIFT);
+	MLX5_SET64(cqc, cqc, dbr_addr,      cq->wq_ctrl.db.dma);
+
+	err = mlx5_core_create_cq(mdev, mcq, in, inlen);
+
+	kvfree(in);
+
+	if (err)
+		return err;
+
+	mlx5e_cq_arm(cq);
+
+	return 0;
+}
+
+static void mlx5e_destroy_cq(struct mlx5e_cq *cq)
+{
+	mlx5_core_destroy_cq(cq->mdev, &cq->mcq);
+}
+
+static int mlx5e_open_cq(struct mlx5e_channel *c,
+			 struct net_dim_cq_moder moder,
+			 struct mlx5e_cq_param *param,
+			 struct mlx5e_cq *cq)
+{
+	struct mlx5_core_dev *mdev = c->mdev;
+	int err;
+
+	err = mlx5e_alloc_cq(c, param, cq);
+	if (err)
+		return err;
+
+	err = mlx5e_create_cq(cq, param);
+	if (err)
+		goto err_free_cq;
+
+	if (MLX5_CAP_GEN(mdev, cq_moderation))
+		mlx5_core_modify_cq_moderation(mdev, &cq->mcq, moder.usec, moder.pkts);
+	return 0;
+
+err_free_cq:
+	mlx5e_free_cq(cq);
+
+	return err;
+}
+
+static void mlx5e_close_cq(struct mlx5e_cq *cq)
+{
+	mlx5e_destroy_cq(cq);
+	mlx5e_free_cq(cq);
+}
+
+static int mlx5e_get_cpu(struct mlx5e_priv *priv, int ix)
+{
+	return cpumask_first(priv->mdev->priv.irq_info[ix + MLX5_EQ_VEC_COMP_BASE].mask);
+}
+
+static int mlx5e_open_tx_cqs(struct mlx5e_channel *c,
+			     struct mlx5e_params *params,
+			     struct mlx5e_channel_param *cparam)
+{
+	int err;
+	int tc;
+
+	for (tc = 0; tc < c->num_tc; tc++) {
+		err = mlx5e_open_cq(c, params->tx_cq_moderation,
+				    &cparam->tx_cq, &c->sq[tc].cq);
+		if (err)
+			goto err_close_tx_cqs;
+	}
+
+	return 0;
+
+err_close_tx_cqs:
+	for (tc--; tc >= 0; tc--)
+		mlx5e_close_cq(&c->sq[tc].cq);
+
+	return err;
+}
+
+static void mlx5e_close_tx_cqs(struct mlx5e_channel *c)
+{
+	int tc;
+
+	for (tc = 0; tc < c->num_tc; tc++)
+		mlx5e_close_cq(&c->sq[tc].cq);
+}
+
+static int mlx5e_open_sqs(struct mlx5e_channel *c,
+			  struct mlx5e_params *params,
+			  struct mlx5e_channel_param *cparam)
+{
+	struct mlx5e_priv *priv = c->priv;
+	int err, tc, max_nch = priv->profile->max_nch(priv->mdev);
+
+	for (tc = 0; tc < params->num_tc; tc++) {
+		int txq_ix = c->ix + tc * max_nch;
+
+		err = mlx5e_open_txqsq(c, c->priv->tisn[tc], txq_ix,
+				       params, &cparam->sq, &c->sq[tc], tc);
+		if (err)
+			goto err_close_sqs;
+	}
+
+	return 0;
+
+err_close_sqs:
+	for (tc--; tc >= 0; tc--)
+		mlx5e_close_txqsq(&c->sq[tc]);
+
+	return err;
+}
+
+static void mlx5e_close_sqs(struct mlx5e_channel *c)
+{
+	int tc;
+
+	for (tc = 0; tc < c->num_tc; tc++)
+		mlx5e_close_txqsq(&c->sq[tc]);
+}
+
+static int mlx5e_set_sq_maxrate(struct net_device *dev,
+				struct mlx5e_txqsq *sq, u32 rate)
+{
+	struct mlx5e_priv *priv = netdev_priv(dev);
+	struct mlx5_core_dev *mdev = priv->mdev;
+	struct mlx5e_modify_sq_param msp = {0};
+	struct mlx5_rate_limit rl = {0};
+	u16 rl_index = 0;
+	int err;
+
+	if (rate == sq->rate_limit)
+		/* nothing to do */
+		return 0;
+
+	if (sq->rate_limit) {
+		rl.rate = sq->rate_limit;
+		/* remove current rl index to free space to next ones */
+		mlx5_rl_remove_rate(mdev, &rl);
+	}
+
+	sq->rate_limit = 0;
+
+	if (rate) {
+		rl.rate = rate;
+		err = mlx5_rl_add_rate(mdev, &rl_index, &rl);
+		if (err) {
+			netdev_err(dev, "Failed configuring rate %u: %d\n",
+				   rate, err);
+			return err;
+		}
+	}
+
+	msp.curr_state = MLX5_SQC_STATE_RDY;
+	msp.next_state = MLX5_SQC_STATE_RDY;
+	msp.rl_index   = rl_index;
+	msp.rl_update  = true;
+	err = mlx5e_modify_sq(mdev, sq->sqn, &msp);
+	if (err) {
+		netdev_err(dev, "Failed configuring rate %u: %d\n",
+			   rate, err);
+		/* remove the rate from the table */
+		if (rate)
+			mlx5_rl_remove_rate(mdev, &rl);
+		return err;
+	}
+
+	sq->rate_limit = rate;
+	return 0;
+}
+
+static int mlx5e_set_tx_maxrate(struct net_device *dev, int index, u32 rate)
+{
+	struct mlx5e_priv *priv = netdev_priv(dev);
+	struct mlx5_core_dev *mdev = priv->mdev;
+	struct mlx5e_txqsq *sq = priv->txq2sq[index];
+	int err = 0;
+
+	if (!mlx5_rl_is_supported(mdev)) {
+		netdev_err(dev, "Rate limiting is not supported on this device\n");
+		return -EINVAL;
+	}
+
+	/* rate is given in Mb/sec, HW config is in Kb/sec */
+	rate = rate << 10;
+
+	/* Check whether rate in valid range, 0 is always valid */
+	if (rate && !mlx5_rl_is_in_range(mdev, rate)) {
+		netdev_err(dev, "TX rate %u, is not in range\n", rate);
+		return -ERANGE;
+	}
+
+	mutex_lock(&priv->state_lock);
+	if (test_bit(MLX5E_STATE_OPENED, &priv->state))
+		err = mlx5e_set_sq_maxrate(dev, sq, rate);
+	if (!err)
+		priv->tx_rates[index] = rate;
+	mutex_unlock(&priv->state_lock);
+
+	return err;
+}
+
+static int mlx5e_open_channel(struct mlx5e_priv *priv, int ix,
+			      struct mlx5e_params *params,
+			      struct mlx5e_channel_param *cparam,
+			      struct mlx5e_channel **cp)
+{
+	struct net_dim_cq_moder icocq_moder = {0, 0};
+	struct net_device *netdev = priv->netdev;
+	int cpu = mlx5e_get_cpu(priv, ix);
+	struct mlx5e_channel *c;
+	unsigned int irq;
+	int err;
+	int eqn;
+
+	err = mlx5_vector2eqn(priv->mdev, ix, &eqn, &irq);
+	if (err)
+		return err;
+
+	c = kvzalloc_node(sizeof(*c), GFP_KERNEL, cpu_to_node(cpu));
+	if (!c)
+		return -ENOMEM;
+
+	c->priv     = priv;
+	c->mdev     = priv->mdev;
+	c->tstamp   = &priv->tstamp;
+	c->ix       = ix;
+	c->cpu      = cpu;
+	c->pdev     = &priv->mdev->pdev->dev;
+	c->netdev   = priv->netdev;
+	c->mkey_be  = cpu_to_be32(priv->mdev->mlx5e_res.mkey.key);
+	c->num_tc   = params->num_tc;
+	c->xdp      = !!params->xdp_prog;
+	c->stats    = &priv->channel_stats[ix].ch;
+
+	c->irq_desc = irq_to_desc(irq);
+
+	netif_napi_add(netdev, &c->napi, mlx5e_napi_poll, 64);
+
+	err = mlx5e_open_cq(c, icocq_moder, &cparam->icosq_cq, &c->icosq.cq);
+	if (err)
+		goto err_napi_del;
+
+	err = mlx5e_open_tx_cqs(c, params, cparam);
+	if (err)
+		goto err_close_icosq_cq;
+
+	err = mlx5e_open_cq(c, params->tx_cq_moderation, &cparam->tx_cq, &c->xdpsq.cq);
+	if (err)
+		goto err_close_tx_cqs;
+
+	err = mlx5e_open_cq(c, params->rx_cq_moderation, &cparam->rx_cq, &c->rq.cq);
+	if (err)
+		goto err_close_xdp_tx_cqs;
+
+	/* XDP SQ CQ params are same as normal TXQ sq CQ params */
+	err = c->xdp ? mlx5e_open_cq(c, params->tx_cq_moderation,
+				     &cparam->tx_cq, &c->rq.xdpsq.cq) : 0;
+	if (err)
+		goto err_close_rx_cq;
+
+	napi_enable(&c->napi);
+
+	err = mlx5e_open_icosq(c, params, &cparam->icosq, &c->icosq);
+	if (err)
+		goto err_disable_napi;
+
+	err = mlx5e_open_sqs(c, params, cparam);
+	if (err)
+		goto err_close_icosq;
+
+	err = c->xdp ? mlx5e_open_xdpsq(c, params, &cparam->xdp_sq, &c->rq.xdpsq, false) : 0;
+	if (err)
+		goto err_close_sqs;
+
+	err = mlx5e_open_rq(c, params, &cparam->rq, &c->rq);
+	if (err)
+		goto err_close_xdp_sq;
+
+	err = mlx5e_open_xdpsq(c, params, &cparam->xdp_sq, &c->xdpsq, true);
+	if (err)
+		goto err_close_rq;
+
+	*cp = c;
+
+	return 0;
+
+err_close_rq:
+	mlx5e_close_rq(&c->rq);
+
+err_close_xdp_sq:
+	if (c->xdp)
+		mlx5e_close_xdpsq(&c->rq.xdpsq);
+
+err_close_sqs:
+	mlx5e_close_sqs(c);
+
+err_close_icosq:
+	mlx5e_close_icosq(&c->icosq);
+
+err_disable_napi:
+	napi_disable(&c->napi);
+	if (c->xdp)
+		mlx5e_close_cq(&c->rq.xdpsq.cq);
+
+err_close_rx_cq:
+	mlx5e_close_cq(&c->rq.cq);
+
+err_close_xdp_tx_cqs:
+	mlx5e_close_cq(&c->xdpsq.cq);
+
+err_close_tx_cqs:
+	mlx5e_close_tx_cqs(c);
+
+err_close_icosq_cq:
+	mlx5e_close_cq(&c->icosq.cq);
+
+err_napi_del:
+	netif_napi_del(&c->napi);
+	kvfree(c);
+
+	return err;
+}
+
+static void mlx5e_activate_channel(struct mlx5e_channel *c)
+{
+	int tc;
+
+	for (tc = 0; tc < c->num_tc; tc++)
+		mlx5e_activate_txqsq(&c->sq[tc]);
+	mlx5e_activate_rq(&c->rq);
+	netif_set_xps_queue(c->netdev, get_cpu_mask(c->cpu), c->ix);
+}
+
+static void mlx5e_deactivate_channel(struct mlx5e_channel *c)
+{
+	int tc;
+
+	mlx5e_deactivate_rq(&c->rq);
+	for (tc = 0; tc < c->num_tc; tc++)
+		mlx5e_deactivate_txqsq(&c->sq[tc]);
+}
+
+static void mlx5e_close_channel(struct mlx5e_channel *c)
+{
+	mlx5e_close_xdpsq(&c->xdpsq);
+	mlx5e_close_rq(&c->rq);
+	if (c->xdp)
+		mlx5e_close_xdpsq(&c->rq.xdpsq);
+	mlx5e_close_sqs(c);
+	mlx5e_close_icosq(&c->icosq);
+	napi_disable(&c->napi);
+	if (c->xdp)
+		mlx5e_close_cq(&c->rq.xdpsq.cq);
+	mlx5e_close_cq(&c->rq.cq);
+	mlx5e_close_cq(&c->xdpsq.cq);
+	mlx5e_close_tx_cqs(c);
+	mlx5e_close_cq(&c->icosq.cq);
+	netif_napi_del(&c->napi);
+
+	kvfree(c);
+}
+
+#define DEFAULT_FRAG_SIZE (2048)
+
+static void mlx5e_build_rq_frags_info(struct mlx5_core_dev *mdev,
+				      struct mlx5e_params *params,
+				      struct mlx5e_rq_frags_info *info)
+{
+	u32 byte_count = MLX5E_SW2HW_MTU(params, params->sw_mtu);
+	int frag_size_max = DEFAULT_FRAG_SIZE;
+	u32 buf_size = 0;
+	int i;
+
+#ifdef CONFIG_MLX5_EN_IPSEC
+	if (MLX5_IPSEC_DEV(mdev))
+		byte_count += MLX5E_METADATA_ETHER_LEN;
+#endif
+
+	if (mlx5e_rx_is_linear_skb(mdev, params)) {
+		int frag_stride;
+
+		frag_stride = mlx5e_rx_get_linear_frag_sz(params);
+		frag_stride = roundup_pow_of_two(frag_stride);
+
+		info->arr[0].frag_size = byte_count;
+		info->arr[0].frag_stride = frag_stride;
+		info->num_frags = 1;
+		info->wqe_bulk = PAGE_SIZE / frag_stride;
+		goto out;
+	}
+
+	if (byte_count > PAGE_SIZE +
+	    (MLX5E_MAX_RX_FRAGS - 1) * frag_size_max)
+		frag_size_max = PAGE_SIZE;
+
+	i = 0;
+	while (buf_size < byte_count) {
+		int frag_size = byte_count - buf_size;
+
+		if (i < MLX5E_MAX_RX_FRAGS - 1)
+			frag_size = min(frag_size, frag_size_max);
+
+		info->arr[i].frag_size = frag_size;
+		info->arr[i].frag_stride = roundup_pow_of_two(frag_size);
+
+		buf_size += frag_size;
+		i++;
+	}
+	info->num_frags = i;
+	/* number of different wqes sharing a page */
+	info->wqe_bulk = 1 + (info->num_frags % 2);
+
+out:
+	info->wqe_bulk = max_t(u8, info->wqe_bulk, 8);
+	info->log_num_frags = order_base_2(info->num_frags);
+}
+
+static inline u8 mlx5e_get_rqwq_log_stride(u8 wq_type, int ndsegs)
+{
+	int sz = sizeof(struct mlx5_wqe_data_seg) * ndsegs;
+
+	switch (wq_type) {
+	case MLX5_WQ_TYPE_LINKED_LIST_STRIDING_RQ:
+		sz += sizeof(struct mlx5e_rx_wqe_ll);
+		break;
+	default: /* MLX5_WQ_TYPE_CYCLIC */
+		sz += sizeof(struct mlx5e_rx_wqe_cyc);
+	}
+
+	return order_base_2(sz);
+}
+
+static void mlx5e_build_rq_param(struct mlx5e_priv *priv,
+				 struct mlx5e_params *params,
+				 struct mlx5e_rq_param *param)
+{
+	struct mlx5_core_dev *mdev = priv->mdev;
+	void *rqc = param->rqc;
+	void *wq = MLX5_ADDR_OF(rqc, rqc, wq);
+	int ndsegs = 1;
+
+	switch (params->rq_wq_type) {
+	case MLX5_WQ_TYPE_LINKED_LIST_STRIDING_RQ:
+		MLX5_SET(wq, wq, log_wqe_num_of_strides,
+			 mlx5e_mpwqe_get_log_num_strides(mdev, params) -
+			 MLX5_MPWQE_LOG_NUM_STRIDES_BASE);
+		MLX5_SET(wq, wq, log_wqe_stride_size,
+			 mlx5e_mpwqe_get_log_stride_size(mdev, params) -
+			 MLX5_MPWQE_LOG_STRIDE_SZ_BASE);
+		MLX5_SET(wq, wq, log_wq_sz, mlx5e_mpwqe_get_log_rq_size(params));
+		break;
+	default: /* MLX5_WQ_TYPE_CYCLIC */
+		MLX5_SET(wq, wq, log_wq_sz, params->log_rq_mtu_frames);
+		mlx5e_build_rq_frags_info(mdev, params, &param->frags_info);
+		ndsegs = param->frags_info.num_frags;
+	}
+
+	MLX5_SET(wq, wq, wq_type,          params->rq_wq_type);
+	MLX5_SET(wq, wq, end_padding_mode, MLX5_WQ_END_PAD_MODE_ALIGN);
+	MLX5_SET(wq, wq, log_wq_stride,
+		 mlx5e_get_rqwq_log_stride(params->rq_wq_type, ndsegs));
+	MLX5_SET(wq, wq, pd,               mdev->mlx5e_res.pdn);
+	MLX5_SET(rqc, rqc, counter_set_id, priv->q_counter);
+	MLX5_SET(rqc, rqc, vsd,            params->vlan_strip_disable);
+	MLX5_SET(rqc, rqc, scatter_fcs,    params->scatter_fcs_en);
+
+	param->wq.buf_numa_node = dev_to_node(&mdev->pdev->dev);
+}
+
+static void mlx5e_build_drop_rq_param(struct mlx5e_priv *priv,
+				      struct mlx5e_rq_param *param)
+{
+	struct mlx5_core_dev *mdev = priv->mdev;
+	void *rqc = param->rqc;
+	void *wq = MLX5_ADDR_OF(rqc, rqc, wq);
+
+	MLX5_SET(wq, wq, wq_type, MLX5_WQ_TYPE_CYCLIC);
+	MLX5_SET(wq, wq, log_wq_stride,
+		 mlx5e_get_rqwq_log_stride(MLX5_WQ_TYPE_CYCLIC, 1));
+	MLX5_SET(rqc, rqc, counter_set_id, priv->drop_rq_q_counter);
+
+	param->wq.buf_numa_node = dev_to_node(&mdev->pdev->dev);
+}
+
+static void mlx5e_build_sq_param_common(struct mlx5e_priv *priv,
+					struct mlx5e_sq_param *param)
+{
+	void *sqc = param->sqc;
+	void *wq = MLX5_ADDR_OF(sqc, sqc, wq);
+
+	MLX5_SET(wq, wq, log_wq_stride, ilog2(MLX5_SEND_WQE_BB));
+	MLX5_SET(wq, wq, pd,            priv->mdev->mlx5e_res.pdn);
+
+	param->wq.buf_numa_node = dev_to_node(&priv->mdev->pdev->dev);
+}
+
+static void mlx5e_build_sq_param(struct mlx5e_priv *priv,
+				 struct mlx5e_params *params,
+				 struct mlx5e_sq_param *param)
+{
+	void *sqc = param->sqc;
+	void *wq = MLX5_ADDR_OF(sqc, sqc, wq);
+
+	mlx5e_build_sq_param_common(priv, param);
+	MLX5_SET(wq, wq, log_wq_sz, params->log_sq_size);
+	MLX5_SET(sqc, sqc, allow_swp, !!MLX5_IPSEC_DEV(priv->mdev));
+}
+
+static void mlx5e_build_common_cq_param(struct mlx5e_priv *priv,
+					struct mlx5e_cq_param *param)
+{
+	void *cqc = param->cqc;
+
+	MLX5_SET(cqc, cqc, uar_page, priv->mdev->priv.uar->index);
+}
+
+static void mlx5e_build_rx_cq_param(struct mlx5e_priv *priv,
+				    struct mlx5e_params *params,
+				    struct mlx5e_cq_param *param)
+{
+	struct mlx5_core_dev *mdev = priv->mdev;
+	void *cqc = param->cqc;
+	u8 log_cq_size;
+
+	switch (params->rq_wq_type) {
+	case MLX5_WQ_TYPE_LINKED_LIST_STRIDING_RQ:
+		log_cq_size = mlx5e_mpwqe_get_log_rq_size(params) +
+			mlx5e_mpwqe_get_log_num_strides(mdev, params);
+		break;
+	default: /* MLX5_WQ_TYPE_CYCLIC */
+		log_cq_size = params->log_rq_mtu_frames;
+	}
+
+	MLX5_SET(cqc, cqc, log_cq_size, log_cq_size);
+	if (MLX5E_GET_PFLAG(params, MLX5E_PFLAG_RX_CQE_COMPRESS)) {
+		MLX5_SET(cqc, cqc, mini_cqe_res_format, MLX5_CQE_FORMAT_CSUM);
+		MLX5_SET(cqc, cqc, cqe_comp_en, 1);
+	}
+
+	mlx5e_build_common_cq_param(priv, param);
+	param->cq_period_mode = params->rx_cq_moderation.cq_period_mode;
+}
+
+static void mlx5e_build_tx_cq_param(struct mlx5e_priv *priv,
+				    struct mlx5e_params *params,
+				    struct mlx5e_cq_param *param)
+{
+	void *cqc = param->cqc;
+
+	MLX5_SET(cqc, cqc, log_cq_size, params->log_sq_size);
+
+	mlx5e_build_common_cq_param(priv, param);
+	param->cq_period_mode = params->tx_cq_moderation.cq_period_mode;
+}
+
+static void mlx5e_build_ico_cq_param(struct mlx5e_priv *priv,
+				     u8 log_wq_size,
+				     struct mlx5e_cq_param *param)
+{
+	void *cqc = param->cqc;
+
+	MLX5_SET(cqc, cqc, log_cq_size, log_wq_size);
+
+	mlx5e_build_common_cq_param(priv, param);
+
+	param->cq_period_mode = NET_DIM_CQ_PERIOD_MODE_START_FROM_EQE;
+}
+
+static void mlx5e_build_icosq_param(struct mlx5e_priv *priv,
+				    u8 log_wq_size,
+				    struct mlx5e_sq_param *param)
+{
+	void *sqc = param->sqc;
+	void *wq = MLX5_ADDR_OF(sqc, sqc, wq);
+
+	mlx5e_build_sq_param_common(priv, param);
+
+	MLX5_SET(wq, wq, log_wq_sz, log_wq_size);
+	MLX5_SET(sqc, sqc, reg_umr, MLX5_CAP_ETH(priv->mdev, reg_umr_sq));
+}
+
+static void mlx5e_build_xdpsq_param(struct mlx5e_priv *priv,
+				    struct mlx5e_params *params,
+				    struct mlx5e_sq_param *param)
+{
+	void *sqc = param->sqc;
+	void *wq = MLX5_ADDR_OF(sqc, sqc, wq);
+
+	mlx5e_build_sq_param_common(priv, param);
+	MLX5_SET(wq, wq, log_wq_sz, params->log_sq_size);
+}
+
+static void mlx5e_build_channel_param(struct mlx5e_priv *priv,
+				      struct mlx5e_params *params,
+				      struct mlx5e_channel_param *cparam)
+{
+	u8 icosq_log_wq_sz = MLX5E_PARAMS_MINIMUM_LOG_SQ_SIZE;
+
+	mlx5e_build_rq_param(priv, params, &cparam->rq);
+	mlx5e_build_sq_param(priv, params, &cparam->sq);
+	mlx5e_build_xdpsq_param(priv, params, &cparam->xdp_sq);
+	mlx5e_build_icosq_param(priv, icosq_log_wq_sz, &cparam->icosq);
+	mlx5e_build_rx_cq_param(priv, params, &cparam->rx_cq);
+	mlx5e_build_tx_cq_param(priv, params, &cparam->tx_cq);
+	mlx5e_build_ico_cq_param(priv, icosq_log_wq_sz, &cparam->icosq_cq);
+}
+
+int mlx5e_open_channels(struct mlx5e_priv *priv,
+			struct mlx5e_channels *chs)
+{
+	struct mlx5e_channel_param *cparam;
+	int err = -ENOMEM;
+	int i;
+
+	chs->num = chs->params.num_channels;
+
+	chs->c = kcalloc(chs->num, sizeof(struct mlx5e_channel *), GFP_KERNEL);
+	cparam = kvzalloc(sizeof(struct mlx5e_channel_param), GFP_KERNEL);
+	if (!chs->c || !cparam)
+		goto err_free;
+
+	mlx5e_build_channel_param(priv, &chs->params, cparam);
+	for (i = 0; i < chs->num; i++) {
+		err = mlx5e_open_channel(priv, i, &chs->params, cparam, &chs->c[i]);
+		if (err)
+			goto err_close_channels;
+	}
+
+	kvfree(cparam);
+	return 0;
+
+err_close_channels:
+	for (i--; i >= 0; i--)
+		mlx5e_close_channel(chs->c[i]);
+
+err_free:
+	kfree(chs->c);
+	kvfree(cparam);
+	chs->num = 0;
+	return err;
+}
+
+static void mlx5e_activate_channels(struct mlx5e_channels *chs)
+{
+	int i;
+
+	for (i = 0; i < chs->num; i++)
+		mlx5e_activate_channel(chs->c[i]);
+}
+
+static int mlx5e_wait_channels_min_rx_wqes(struct mlx5e_channels *chs)
+{
+	int err = 0;
+	int i;
+
+	for (i = 0; i < chs->num; i++)
+		err |= mlx5e_wait_for_min_rx_wqes(&chs->c[i]->rq,
+						  err ? 0 : 20000);
+
+	return err ? -ETIMEDOUT : 0;
+}
+
+static void mlx5e_deactivate_channels(struct mlx5e_channels *chs)
+{
+	int i;
+
+	for (i = 0; i < chs->num; i++)
+		mlx5e_deactivate_channel(chs->c[i]);
+}
+
+void mlx5e_close_channels(struct mlx5e_channels *chs)
+{
+	int i;
+
+	for (i = 0; i < chs->num; i++)
+		mlx5e_close_channel(chs->c[i]);
+
+	kfree(chs->c);
+	chs->num = 0;
+}
+
+static int
+mlx5e_create_rqt(struct mlx5e_priv *priv, int sz, struct mlx5e_rqt *rqt)
+{
+	struct mlx5_core_dev *mdev = priv->mdev;
+	void *rqtc;
+	int inlen;
+	int err;
+	u32 *in;
+	int i;
+
+	inlen = MLX5_ST_SZ_BYTES(create_rqt_in) + sizeof(u32) * sz;
+	in = kvzalloc(inlen, GFP_KERNEL);
+	if (!in)
+		return -ENOMEM;
+
+	rqtc = MLX5_ADDR_OF(create_rqt_in, in, rqt_context);
+
+	MLX5_SET(rqtc, rqtc, rqt_actual_size, sz);
+	MLX5_SET(rqtc, rqtc, rqt_max_size, sz);
+
+	for (i = 0; i < sz; i++)
+		MLX5_SET(rqtc, rqtc, rq_num[i], priv->drop_rq.rqn);
+
+	err = mlx5_core_create_rqt(mdev, in, inlen, &rqt->rqtn);
+	if (!err)
+		rqt->enabled = true;
+
+	kvfree(in);
+	return err;
+}
+
+void mlx5e_destroy_rqt(struct mlx5e_priv *priv, struct mlx5e_rqt *rqt)
+{
+	rqt->enabled = false;
+	mlx5_core_destroy_rqt(priv->mdev, rqt->rqtn);
+}
+
+int mlx5e_create_indirect_rqt(struct mlx5e_priv *priv)
+{
+	struct mlx5e_rqt *rqt = &priv->indir_rqt;
+	int err;
+
+	err = mlx5e_create_rqt(priv, MLX5E_INDIR_RQT_SIZE, rqt);
+	if (err)
+		mlx5_core_warn(priv->mdev, "create indirect rqts failed, %d\n", err);
+	return err;
+}
+
+int mlx5e_create_direct_rqts(struct mlx5e_priv *priv)
+{
+	struct mlx5e_rqt *rqt;
+	int err;
+	int ix;
+
+	for (ix = 0; ix < priv->profile->max_nch(priv->mdev); ix++) {
+		rqt = &priv->direct_tir[ix].rqt;
+		err = mlx5e_create_rqt(priv, 1 /*size */, rqt);
+		if (err)
+			goto err_destroy_rqts;
+	}
+
+	return 0;
+
+err_destroy_rqts:
+	mlx5_core_warn(priv->mdev, "create direct rqts failed, %d\n", err);
+	for (ix--; ix >= 0; ix--)
+		mlx5e_destroy_rqt(priv, &priv->direct_tir[ix].rqt);
+
+	return err;
+}
+
+void mlx5e_destroy_direct_rqts(struct mlx5e_priv *priv)
+{
+	int i;
+
+	for (i = 0; i < priv->profile->max_nch(priv->mdev); i++)
+		mlx5e_destroy_rqt(priv, &priv->direct_tir[i].rqt);
+}
+
+static int mlx5e_rx_hash_fn(int hfunc)
+{
+	return (hfunc == ETH_RSS_HASH_TOP) ?
+	       MLX5_RX_HASH_FN_TOEPLITZ :
+	       MLX5_RX_HASH_FN_INVERTED_XOR8;
+}
+
+int mlx5e_bits_invert(unsigned long a, int size)
+{
+	int inv = 0;
+	int i;
+
+	for (i = 0; i < size; i++)
+		inv |= (test_bit(size - i - 1, &a) ? 1 : 0) << i;
+
+	return inv;
+}
+
+static void mlx5e_fill_rqt_rqns(struct mlx5e_priv *priv, int sz,
+				struct mlx5e_redirect_rqt_param rrp, void *rqtc)
+{
+	int i;
+
+	for (i = 0; i < sz; i++) {
+		u32 rqn;
+
+		if (rrp.is_rss) {
+			int ix = i;
+
+			if (rrp.rss.hfunc == ETH_RSS_HASH_XOR)
+				ix = mlx5e_bits_invert(i, ilog2(sz));
+
+			ix = priv->channels.params.indirection_rqt[ix];
+			rqn = rrp.rss.channels->c[ix]->rq.rqn;
+		} else {
+			rqn = rrp.rqn;
+		}
+		MLX5_SET(rqtc, rqtc, rq_num[i], rqn);
+	}
+}
+
+int mlx5e_redirect_rqt(struct mlx5e_priv *priv, u32 rqtn, int sz,
+		       struct mlx5e_redirect_rqt_param rrp)
+{
+	struct mlx5_core_dev *mdev = priv->mdev;
+	void *rqtc;
+	int inlen;
+	u32 *in;
+	int err;
+
+	inlen = MLX5_ST_SZ_BYTES(modify_rqt_in) + sizeof(u32) * sz;
+	in = kvzalloc(inlen, GFP_KERNEL);
+	if (!in)
+		return -ENOMEM;
+
+	rqtc = MLX5_ADDR_OF(modify_rqt_in, in, ctx);
+
+	MLX5_SET(rqtc, rqtc, rqt_actual_size, sz);
+	MLX5_SET(modify_rqt_in, in, bitmask.rqn_list, 1);
+	mlx5e_fill_rqt_rqns(priv, sz, rrp, rqtc);
+	err = mlx5_core_modify_rqt(mdev, rqtn, in, inlen);
+
+	kvfree(in);
+	return err;
+}
+
+static u32 mlx5e_get_direct_rqn(struct mlx5e_priv *priv, int ix,
+				struct mlx5e_redirect_rqt_param rrp)
+{
+	if (!rrp.is_rss)
+		return rrp.rqn;
+
+	if (ix >= rrp.rss.channels->num)
+		return priv->drop_rq.rqn;
+
+	return rrp.rss.channels->c[ix]->rq.rqn;
+}
+
+static void mlx5e_redirect_rqts(struct mlx5e_priv *priv,
+				struct mlx5e_redirect_rqt_param rrp)
+{
+	u32 rqtn;
+	int ix;
+
+	if (priv->indir_rqt.enabled) {
+		/* RSS RQ table */
+		rqtn = priv->indir_rqt.rqtn;
+		mlx5e_redirect_rqt(priv, rqtn, MLX5E_INDIR_RQT_SIZE, rrp);
+	}
+
+	for (ix = 0; ix < priv->profile->max_nch(priv->mdev); ix++) {
+		struct mlx5e_redirect_rqt_param direct_rrp = {
+			.is_rss = false,
+			{
+				.rqn    = mlx5e_get_direct_rqn(priv, ix, rrp)
+			},
+		};
+
+		/* Direct RQ Tables */
+		if (!priv->direct_tir[ix].rqt.enabled)
+			continue;
+
+		rqtn = priv->direct_tir[ix].rqt.rqtn;
+		mlx5e_redirect_rqt(priv, rqtn, 1, direct_rrp);
+	}
+}
+
+static void mlx5e_redirect_rqts_to_channels(struct mlx5e_priv *priv,
+					    struct mlx5e_channels *chs)
+{
+	struct mlx5e_redirect_rqt_param rrp = {
+		.is_rss        = true,
+		{
+			.rss = {
+				.channels  = chs,
+				.hfunc     = chs->params.rss_hfunc,
+			}
+		},
+	};
+
+	mlx5e_redirect_rqts(priv, rrp);
+}
+
+static void mlx5e_redirect_rqts_to_drop(struct mlx5e_priv *priv)
+{
+	struct mlx5e_redirect_rqt_param drop_rrp = {
+		.is_rss = false,
+		{
+			.rqn = priv->drop_rq.rqn,
+		},
+	};
+
+	mlx5e_redirect_rqts(priv, drop_rrp);
+}
+
+static void mlx5e_build_tir_ctx_lro(struct mlx5e_params *params, void *tirc)
+{
+	if (!params->lro_en)
+		return;
+
+#define ROUGH_MAX_L2_L3_HDR_SZ 256
+
+	MLX5_SET(tirc, tirc, lro_enable_mask,
+		 MLX5_TIRC_LRO_ENABLE_MASK_IPV4_LRO |
+		 MLX5_TIRC_LRO_ENABLE_MASK_IPV6_LRO);
+	MLX5_SET(tirc, tirc, lro_max_ip_payload_size,
+		 (params->lro_wqe_sz - ROUGH_MAX_L2_L3_HDR_SZ) >> 8);
+	MLX5_SET(tirc, tirc, lro_timeout_period_usecs, params->lro_timeout);
+}
+
+void mlx5e_build_indir_tir_ctx_hash(struct mlx5e_params *params,
+				    enum mlx5e_traffic_types tt,
+				    void *tirc, bool inner)
+{
+	void *hfso = inner ? MLX5_ADDR_OF(tirc, tirc, rx_hash_field_selector_inner) :
+			     MLX5_ADDR_OF(tirc, tirc, rx_hash_field_selector_outer);
+
+#define MLX5_HASH_IP            (MLX5_HASH_FIELD_SEL_SRC_IP   |\
+				 MLX5_HASH_FIELD_SEL_DST_IP)
+
+#define MLX5_HASH_IP_L4PORTS    (MLX5_HASH_FIELD_SEL_SRC_IP   |\
+				 MLX5_HASH_FIELD_SEL_DST_IP   |\
+				 MLX5_HASH_FIELD_SEL_L4_SPORT |\
+				 MLX5_HASH_FIELD_SEL_L4_DPORT)
+
+#define MLX5_HASH_IP_IPSEC_SPI  (MLX5_HASH_FIELD_SEL_SRC_IP   |\
+				 MLX5_HASH_FIELD_SEL_DST_IP   |\
+				 MLX5_HASH_FIELD_SEL_IPSEC_SPI)
+
+	MLX5_SET(tirc, tirc, rx_hash_fn, mlx5e_rx_hash_fn(params->rss_hfunc));
+	if (params->rss_hfunc == ETH_RSS_HASH_TOP) {
+		void *rss_key = MLX5_ADDR_OF(tirc, tirc,
+					     rx_hash_toeplitz_key);
+		size_t len = MLX5_FLD_SZ_BYTES(tirc,
+					       rx_hash_toeplitz_key);
+
+		MLX5_SET(tirc, tirc, rx_hash_symmetric, 1);
+		memcpy(rss_key, params->toeplitz_hash_key, len);
+	}
+
+	switch (tt) {
+	case MLX5E_TT_IPV4_TCP:
+		MLX5_SET(rx_hash_field_select, hfso, l3_prot_type,
+			 MLX5_L3_PROT_TYPE_IPV4);
+		MLX5_SET(rx_hash_field_select, hfso, l4_prot_type,
+			 MLX5_L4_PROT_TYPE_TCP);
+		MLX5_SET(rx_hash_field_select, hfso, selected_fields,
+			 MLX5_HASH_IP_L4PORTS);
+		break;
+
+	case MLX5E_TT_IPV6_TCP:
+		MLX5_SET(rx_hash_field_select, hfso, l3_prot_type,
+			 MLX5_L3_PROT_TYPE_IPV6);
+		MLX5_SET(rx_hash_field_select, hfso, l4_prot_type,
+			 MLX5_L4_PROT_TYPE_TCP);
+		MLX5_SET(rx_hash_field_select, hfso, selected_fields,
+			 MLX5_HASH_IP_L4PORTS);
+		break;
+
+	case MLX5E_TT_IPV4_UDP:
+		MLX5_SET(rx_hash_field_select, hfso, l3_prot_type,
+			 MLX5_L3_PROT_TYPE_IPV4);
+		MLX5_SET(rx_hash_field_select, hfso, l4_prot_type,
+			 MLX5_L4_PROT_TYPE_UDP);
+		MLX5_SET(rx_hash_field_select, hfso, selected_fields,
+			 MLX5_HASH_IP_L4PORTS);
+		break;
+
+	case MLX5E_TT_IPV6_UDP:
+		MLX5_SET(rx_hash_field_select, hfso, l3_prot_type,
+			 MLX5_L3_PROT_TYPE_IPV6);
+		MLX5_SET(rx_hash_field_select, hfso, l4_prot_type,
+			 MLX5_L4_PROT_TYPE_UDP);
+		MLX5_SET(rx_hash_field_select, hfso, selected_fields,
+			 MLX5_HASH_IP_L4PORTS);
+		break;
+
+	case MLX5E_TT_IPV4_IPSEC_AH:
+		MLX5_SET(rx_hash_field_select, hfso, l3_prot_type,
+			 MLX5_L3_PROT_TYPE_IPV4);
+		MLX5_SET(rx_hash_field_select, hfso, selected_fields,
+			 MLX5_HASH_IP_IPSEC_SPI);
+		break;
+
+	case MLX5E_TT_IPV6_IPSEC_AH:
+		MLX5_SET(rx_hash_field_select, hfso, l3_prot_type,
+			 MLX5_L3_PROT_TYPE_IPV6);
+		MLX5_SET(rx_hash_field_select, hfso, selected_fields,
+			 MLX5_HASH_IP_IPSEC_SPI);
+		break;
+
+	case MLX5E_TT_IPV4_IPSEC_ESP:
+		MLX5_SET(rx_hash_field_select, hfso, l3_prot_type,
+			 MLX5_L3_PROT_TYPE_IPV4);
+		MLX5_SET(rx_hash_field_select, hfso, selected_fields,
+			 MLX5_HASH_IP_IPSEC_SPI);
+		break;
+
+	case MLX5E_TT_IPV6_IPSEC_ESP:
+		MLX5_SET(rx_hash_field_select, hfso, l3_prot_type,
+			 MLX5_L3_PROT_TYPE_IPV6);
+		MLX5_SET(rx_hash_field_select, hfso, selected_fields,
+			 MLX5_HASH_IP_IPSEC_SPI);
+		break;
+
+	case MLX5E_TT_IPV4:
+		MLX5_SET(rx_hash_field_select, hfso, l3_prot_type,
+			 MLX5_L3_PROT_TYPE_IPV4);
+		MLX5_SET(rx_hash_field_select, hfso, selected_fields,
+			 MLX5_HASH_IP);
+		break;
+
+	case MLX5E_TT_IPV6:
+		MLX5_SET(rx_hash_field_select, hfso, l3_prot_type,
+			 MLX5_L3_PROT_TYPE_IPV6);
+		MLX5_SET(rx_hash_field_select, hfso, selected_fields,
+			 MLX5_HASH_IP);
+		break;
+	default:
+		WARN_ONCE(true, "%s: bad traffic type!\n", __func__);
+	}
+}
+
+static int mlx5e_modify_tirs_lro(struct mlx5e_priv *priv)
+{
+	struct mlx5_core_dev *mdev = priv->mdev;
+
+	void *in;
+	void *tirc;
+	int inlen;
+	int err;
+	int tt;
+	int ix;
+
+	inlen = MLX5_ST_SZ_BYTES(modify_tir_in);
+	in = kvzalloc(inlen, GFP_KERNEL);
+	if (!in)
+		return -ENOMEM;
+
+	MLX5_SET(modify_tir_in, in, bitmask.lro, 1);
+	tirc = MLX5_ADDR_OF(modify_tir_in, in, ctx);
+
+	mlx5e_build_tir_ctx_lro(&priv->channels.params, tirc);
+
+	for (tt = 0; tt < MLX5E_NUM_INDIR_TIRS; tt++) {
+		err = mlx5_core_modify_tir(mdev, priv->indir_tir[tt].tirn, in,
+					   inlen);
+		if (err)
+			goto free_in;
+	}
+
+	for (ix = 0; ix < priv->profile->max_nch(priv->mdev); ix++) {
+		err = mlx5_core_modify_tir(mdev, priv->direct_tir[ix].tirn,
+					   in, inlen);
+		if (err)
+			goto free_in;
+	}
+
+free_in:
+	kvfree(in);
+
+	return err;
+}
+
+static void mlx5e_build_inner_indir_tir_ctx(struct mlx5e_priv *priv,
+					    enum mlx5e_traffic_types tt,
+					    u32 *tirc)
+{
+	MLX5_SET(tirc, tirc, transport_domain, priv->mdev->mlx5e_res.td.tdn);
+
+	mlx5e_build_tir_ctx_lro(&priv->channels.params, tirc);
+
+	MLX5_SET(tirc, tirc, disp_type, MLX5_TIRC_DISP_TYPE_INDIRECT);
+	MLX5_SET(tirc, tirc, indirect_table, priv->indir_rqt.rqtn);
+	MLX5_SET(tirc, tirc, tunneled_offload_en, 0x1);
+
+	mlx5e_build_indir_tir_ctx_hash(&priv->channels.params, tt, tirc, true);
+}
+
+static int mlx5e_set_mtu(struct mlx5_core_dev *mdev,
+			 struct mlx5e_params *params, u16 mtu)
+{
+	u16 hw_mtu = MLX5E_SW2HW_MTU(params, mtu);
+	int err;
+
+	err = mlx5_set_port_mtu(mdev, hw_mtu, 1);
+	if (err)
+		return err;
+
+	/* Update vport context MTU */
+	mlx5_modify_nic_vport_mtu(mdev, hw_mtu);
+	return 0;
+}
+
+static void mlx5e_query_mtu(struct mlx5_core_dev *mdev,
+			    struct mlx5e_params *params, u16 *mtu)
+{
+	u16 hw_mtu = 0;
+	int err;
+
+	err = mlx5_query_nic_vport_mtu(mdev, &hw_mtu);
+	if (err || !hw_mtu) /* fallback to port oper mtu */
+		mlx5_query_port_oper_mtu(mdev, &hw_mtu, 1);
+
+	*mtu = MLX5E_HW2SW_MTU(params, hw_mtu);
+}
+
+static int mlx5e_set_dev_port_mtu(struct mlx5e_priv *priv)
+{
+	struct mlx5e_params *params = &priv->channels.params;
+	struct net_device *netdev = priv->netdev;
+	struct mlx5_core_dev *mdev = priv->mdev;
+	u16 mtu;
+	int err;
+
+	err = mlx5e_set_mtu(mdev, params, params->sw_mtu);
+	if (err)
+		return err;
+
+	mlx5e_query_mtu(mdev, params, &mtu);
+	if (mtu != params->sw_mtu)
+		netdev_warn(netdev, "%s: VPort MTU %d is different than netdev mtu %d\n",
+			    __func__, mtu, params->sw_mtu);
+
+	params->sw_mtu = mtu;
+	return 0;
+}
+
+static void mlx5e_netdev_set_tcs(struct net_device *netdev)
+{
+	struct mlx5e_priv *priv = netdev_priv(netdev);
+	int nch = priv->channels.params.num_channels;
+	int ntc = priv->channels.params.num_tc;
+	int tc;
+
+	netdev_reset_tc(netdev);
+
+	if (ntc == 1)
+		return;
+
+	netdev_set_num_tc(netdev, ntc);
+
+	/* Map netdev TCs to offset 0
+	 * We have our own UP to TXQ mapping for QoS
+	 */
+	for (tc = 0; tc < ntc; tc++)
+		netdev_set_tc_queue(netdev, tc, nch, 0);
+}
+
+static void mlx5e_build_tc2txq_maps(struct mlx5e_priv *priv)
+{
+	int max_nch = priv->profile->max_nch(priv->mdev);
+	int i, tc;
+
+	for (i = 0; i < max_nch; i++)
+		for (tc = 0; tc < priv->profile->max_tc; tc++)
+			priv->channel_tc2txq[i][tc] = i + tc * max_nch;
+}
+
+static void mlx5e_build_tx2sq_maps(struct mlx5e_priv *priv)
+{
+	struct mlx5e_channel *c;
+	struct mlx5e_txqsq *sq;
+	int i, tc;
+
+	for (i = 0; i < priv->channels.num; i++) {
+		c = priv->channels.c[i];
+		for (tc = 0; tc < c->num_tc; tc++) {
+			sq = &c->sq[tc];
+			priv->txq2sq[sq->txq_ix] = sq;
+		}
+	}
+}
+
+void mlx5e_activate_priv_channels(struct mlx5e_priv *priv)
+{
+	int num_txqs = priv->channels.num * priv->channels.params.num_tc;
+	struct net_device *netdev = priv->netdev;
+
+	mlx5e_netdev_set_tcs(netdev);
+	netif_set_real_num_tx_queues(netdev, num_txqs);
+	netif_set_real_num_rx_queues(netdev, priv->channels.num);
+
+	mlx5e_build_tx2sq_maps(priv);
+	mlx5e_activate_channels(&priv->channels);
+	mlx5e_xdp_tx_enable(priv);
+	netif_tx_start_all_queues(priv->netdev);
+
+	if (MLX5_ESWITCH_MANAGER(priv->mdev))
+		mlx5e_add_sqs_fwd_rules(priv);
+
+	mlx5e_wait_channels_min_rx_wqes(&priv->channels);
+	mlx5e_redirect_rqts_to_channels(priv, &priv->channels);
+}
+
+void mlx5e_deactivate_priv_channels(struct mlx5e_priv *priv)
+{
+	mlx5e_redirect_rqts_to_drop(priv);
+
+	if (MLX5_ESWITCH_MANAGER(priv->mdev))
+		mlx5e_remove_sqs_fwd_rules(priv);
+
+	/* FIXME: This is a W/A only for tx timeout watch dog false alarm when
+	 * polling for inactive tx queues.
+	 */
+	netif_tx_stop_all_queues(priv->netdev);
+	netif_tx_disable(priv->netdev);
+	mlx5e_xdp_tx_disable(priv);
+	mlx5e_deactivate_channels(&priv->channels);
+}
+
+void mlx5e_switch_priv_channels(struct mlx5e_priv *priv,
+				struct mlx5e_channels *new_chs,
+				mlx5e_fp_hw_modify hw_modify)
+{
+	struct net_device *netdev = priv->netdev;
+	int new_num_txqs;
+	int carrier_ok;
+	new_num_txqs = new_chs->num * new_chs->params.num_tc;
+
+	carrier_ok = netif_carrier_ok(netdev);
+	netif_carrier_off(netdev);
+
+	if (new_num_txqs < netdev->real_num_tx_queues)
+		netif_set_real_num_tx_queues(netdev, new_num_txqs);
+
+	mlx5e_deactivate_priv_channels(priv);
+	mlx5e_close_channels(&priv->channels);
+
+	priv->channels = *new_chs;
+
+	/* New channels are ready to roll, modify HW settings if needed */
+	if (hw_modify)
+		hw_modify(priv);
+
+	mlx5e_refresh_tirs(priv, false);
+	mlx5e_activate_priv_channels(priv);
+
+	/* return carrier back if needed */
+	if (carrier_ok)
+		netif_carrier_on(netdev);
+}
+
+void mlx5e_timestamp_init(struct mlx5e_priv *priv)
+{
+	priv->tstamp.tx_type   = HWTSTAMP_TX_OFF;
+	priv->tstamp.rx_filter = HWTSTAMP_FILTER_NONE;
+}
+
+int mlx5e_open_locked(struct net_device *netdev)
+{
+	struct mlx5e_priv *priv = netdev_priv(netdev);
+	int err;
+
+	set_bit(MLX5E_STATE_OPENED, &priv->state);
+
+	err = mlx5e_open_channels(priv, &priv->channels);
+	if (err)
+		goto err_clear_state_opened_flag;
+
+	mlx5e_refresh_tirs(priv, false);
+	mlx5e_activate_priv_channels(priv);
+	if (priv->profile->update_carrier)
+		priv->profile->update_carrier(priv);
+
+	if (priv->profile->update_stats)
+		queue_delayed_work(priv->wq, &priv->update_stats_work, 0);
+
+	return 0;
+
+err_clear_state_opened_flag:
+	clear_bit(MLX5E_STATE_OPENED, &priv->state);
+	return err;
+}
+
+int mlx5e_open(struct net_device *netdev)
+{
+	struct mlx5e_priv *priv = netdev_priv(netdev);
+	int err;
+
+	mutex_lock(&priv->state_lock);
+	err = mlx5e_open_locked(netdev);
+	if (!err)
+		mlx5_set_port_admin_status(priv->mdev, MLX5_PORT_UP);
+	mutex_unlock(&priv->state_lock);
+
+	if (mlx5_vxlan_allowed(priv->mdev->vxlan))
+		udp_tunnel_get_rx_info(netdev);
+
+	return err;
+}
+
+int mlx5e_close_locked(struct net_device *netdev)
+{
+	struct mlx5e_priv *priv = netdev_priv(netdev);
+
+	/* May already be CLOSED in case a previous configuration operation
+	 * (e.g RX/TX queue size change) that involves close&open failed.
+	 */
+	if (!test_bit(MLX5E_STATE_OPENED, &priv->state))
+		return 0;
+
+	clear_bit(MLX5E_STATE_OPENED, &priv->state);
+
+	netif_carrier_off(priv->netdev);
+	mlx5e_deactivate_priv_channels(priv);
+	mlx5e_close_channels(&priv->channels);
+
+	return 0;
+}
+
+int mlx5e_close(struct net_device *netdev)
+{
+	struct mlx5e_priv *priv = netdev_priv(netdev);
+	int err;
+
+	if (!netif_device_present(netdev))
+		return -ENODEV;
+
+	mutex_lock(&priv->state_lock);
+	mlx5_set_port_admin_status(priv->mdev, MLX5_PORT_DOWN);
+	err = mlx5e_close_locked(netdev);
+	mutex_unlock(&priv->state_lock);
+
+	return err;
+}
+
+static int mlx5e_alloc_drop_rq(struct mlx5_core_dev *mdev,
+			       struct mlx5e_rq *rq,
+			       struct mlx5e_rq_param *param)
+{
+	void *rqc = param->rqc;
+	void *rqc_wq = MLX5_ADDR_OF(rqc, rqc, wq);
+	int err;
+
+	param->wq.db_numa_node = param->wq.buf_numa_node;
+
+	err = mlx5_wq_cyc_create(mdev, &param->wq, rqc_wq, &rq->wqe.wq,
+				 &rq->wq_ctrl);
+	if (err)
+		return err;
+
+	/* Mark as unused given "Drop-RQ" packets never reach XDP */
+	xdp_rxq_info_unused(&rq->xdp_rxq);
+
+	rq->mdev = mdev;
+
+	return 0;
+}
+
+static int mlx5e_alloc_drop_cq(struct mlx5_core_dev *mdev,
+			       struct mlx5e_cq *cq,
+			       struct mlx5e_cq_param *param)
+{
+	param->wq.buf_numa_node = dev_to_node(&mdev->pdev->dev);
+	param->wq.db_numa_node  = dev_to_node(&mdev->pdev->dev);
+
+	return mlx5e_alloc_cq_common(mdev, param, cq);
+}
+
+static int mlx5e_open_drop_rq(struct mlx5e_priv *priv,
+			      struct mlx5e_rq *drop_rq)
+{
+	struct mlx5_core_dev *mdev = priv->mdev;
+	struct mlx5e_cq_param cq_param = {};
+	struct mlx5e_rq_param rq_param = {};
+	struct mlx5e_cq *cq = &drop_rq->cq;
+	int err;
+
+	mlx5e_build_drop_rq_param(priv, &rq_param);
+
+	err = mlx5e_alloc_drop_cq(mdev, cq, &cq_param);
+	if (err)
+		return err;
+
+	err = mlx5e_create_cq(cq, &cq_param);
+	if (err)
+		goto err_free_cq;
+
+	err = mlx5e_alloc_drop_rq(mdev, drop_rq, &rq_param);
+	if (err)
+		goto err_destroy_cq;
+
+	err = mlx5e_create_rq(drop_rq, &rq_param);
+	if (err)
+		goto err_free_rq;
+
+	err = mlx5e_modify_rq_state(drop_rq, MLX5_RQC_STATE_RST, MLX5_RQC_STATE_RDY);
+	if (err)
+		mlx5_core_warn(priv->mdev, "modify_rq_state failed, rx_if_down_packets won't be counted %d\n", err);
+
+	return 0;
+
+err_free_rq:
+	mlx5e_free_rq(drop_rq);
+
+err_destroy_cq:
+	mlx5e_destroy_cq(cq);
+
+err_free_cq:
+	mlx5e_free_cq(cq);
+
+	return err;
+}
+
+static void mlx5e_close_drop_rq(struct mlx5e_rq *drop_rq)
+{
+	mlx5e_destroy_rq(drop_rq);
+	mlx5e_free_rq(drop_rq);
+	mlx5e_destroy_cq(&drop_rq->cq);
+	mlx5e_free_cq(&drop_rq->cq);
+}
+
+int mlx5e_create_tis(struct mlx5_core_dev *mdev, int tc,
+		     u32 underlay_qpn, u32 *tisn)
+{
+	u32 in[MLX5_ST_SZ_DW(create_tis_in)] = {0};
+	void *tisc = MLX5_ADDR_OF(create_tis_in, in, ctx);
+
+	MLX5_SET(tisc, tisc, prio, tc << 1);
+	MLX5_SET(tisc, tisc, underlay_qpn, underlay_qpn);
+	MLX5_SET(tisc, tisc, transport_domain, mdev->mlx5e_res.td.tdn);
+
+	if (mlx5_lag_is_lacp_owner(mdev))
+		MLX5_SET(tisc, tisc, strict_lag_tx_port_affinity, 1);
+
+	return mlx5_core_create_tis(mdev, in, sizeof(in), tisn);
+}
+
+void mlx5e_destroy_tis(struct mlx5_core_dev *mdev, u32 tisn)
+{
+	mlx5_core_destroy_tis(mdev, tisn);
+}
+
+int mlx5e_create_tises(struct mlx5e_priv *priv)
+{
+	int err;
+	int tc;
+
+	for (tc = 0; tc < priv->profile->max_tc; tc++) {
+		err = mlx5e_create_tis(priv->mdev, tc, 0, &priv->tisn[tc]);
+		if (err)
+			goto err_close_tises;
+	}
+
+	return 0;
+
+err_close_tises:
+	for (tc--; tc >= 0; tc--)
+		mlx5e_destroy_tis(priv->mdev, priv->tisn[tc]);
+
+	return err;
+}
+
+void mlx5e_cleanup_nic_tx(struct mlx5e_priv *priv)
+{
+	int tc;
+
+	for (tc = 0; tc < priv->profile->max_tc; tc++)
+		mlx5e_destroy_tis(priv->mdev, priv->tisn[tc]);
+}
+
+static void mlx5e_build_indir_tir_ctx(struct mlx5e_priv *priv,
+				      enum mlx5e_traffic_types tt,
+				      u32 *tirc)
+{
+	MLX5_SET(tirc, tirc, transport_domain, priv->mdev->mlx5e_res.td.tdn);
+
+	mlx5e_build_tir_ctx_lro(&priv->channels.params, tirc);
+
+	MLX5_SET(tirc, tirc, disp_type, MLX5_TIRC_DISP_TYPE_INDIRECT);
+	MLX5_SET(tirc, tirc, indirect_table, priv->indir_rqt.rqtn);
+	mlx5e_build_indir_tir_ctx_hash(&priv->channels.params, tt, tirc, false);
+}
+
+static void mlx5e_build_direct_tir_ctx(struct mlx5e_priv *priv, u32 rqtn, u32 *tirc)
+{
+	MLX5_SET(tirc, tirc, transport_domain, priv->mdev->mlx5e_res.td.tdn);
+
+	mlx5e_build_tir_ctx_lro(&priv->channels.params, tirc);
+
+	MLX5_SET(tirc, tirc, disp_type, MLX5_TIRC_DISP_TYPE_INDIRECT);
+	MLX5_SET(tirc, tirc, indirect_table, rqtn);
+	MLX5_SET(tirc, tirc, rx_hash_fn, MLX5_RX_HASH_FN_INVERTED_XOR8);
+}
+
+int mlx5e_create_indirect_tirs(struct mlx5e_priv *priv)
+{
+	struct mlx5e_tir *tir;
+	void *tirc;
+	int inlen;
+	int i = 0;
+	int err;
+	u32 *in;
+	int tt;
+
+	inlen = MLX5_ST_SZ_BYTES(create_tir_in);
+	in = kvzalloc(inlen, GFP_KERNEL);
+	if (!in)
+		return -ENOMEM;
+
+	for (tt = 0; tt < MLX5E_NUM_INDIR_TIRS; tt++) {
+		memset(in, 0, inlen);
+		tir = &priv->indir_tir[tt];
+		tirc = MLX5_ADDR_OF(create_tir_in, in, ctx);
+		mlx5e_build_indir_tir_ctx(priv, tt, tirc);
+		err = mlx5e_create_tir(priv->mdev, tir, in, inlen);
+		if (err) {
+			mlx5_core_warn(priv->mdev, "create indirect tirs failed, %d\n", err);
+			goto err_destroy_inner_tirs;
+		}
+	}
+
+	if (!mlx5e_tunnel_inner_ft_supported(priv->mdev))
+		goto out;
+
+	for (i = 0; i < MLX5E_NUM_INDIR_TIRS; i++) {
+		memset(in, 0, inlen);
+		tir = &priv->inner_indir_tir[i];
+		tirc = MLX5_ADDR_OF(create_tir_in, in, ctx);
+		mlx5e_build_inner_indir_tir_ctx(priv, i, tirc);
+		err = mlx5e_create_tir(priv->mdev, tir, in, inlen);
+		if (err) {
+			mlx5_core_warn(priv->mdev, "create inner indirect tirs failed, %d\n", err);
+			goto err_destroy_inner_tirs;
+		}
+	}
+
+out:
+	kvfree(in);
+
+	return 0;
+
+err_destroy_inner_tirs:
+	for (i--; i >= 0; i--)
+		mlx5e_destroy_tir(priv->mdev, &priv->inner_indir_tir[i]);
+
+	for (tt--; tt >= 0; tt--)
+		mlx5e_destroy_tir(priv->mdev, &priv->indir_tir[tt]);
+
+	kvfree(in);
+
+	return err;
+}
+
+int mlx5e_create_direct_tirs(struct mlx5e_priv *priv)
+{
+	int nch = priv->profile->max_nch(priv->mdev);
+	struct mlx5e_tir *tir;
+	void *tirc;
+	int inlen;
+	int err;
+	u32 *in;
+	int ix;
+
+	inlen = MLX5_ST_SZ_BYTES(create_tir_in);
+	in = kvzalloc(inlen, GFP_KERNEL);
+	if (!in)
+		return -ENOMEM;
+
+	for (ix = 0; ix < nch; ix++) {
+		memset(in, 0, inlen);
+		tir = &priv->direct_tir[ix];
+		tirc = MLX5_ADDR_OF(create_tir_in, in, ctx);
+		mlx5e_build_direct_tir_ctx(priv, priv->direct_tir[ix].rqt.rqtn, tirc);
+		err = mlx5e_create_tir(priv->mdev, tir, in, inlen);
+		if (err)
+			goto err_destroy_ch_tirs;
+	}
+
+	kvfree(in);
+
+	return 0;
+
+err_destroy_ch_tirs:
+	mlx5_core_warn(priv->mdev, "create direct tirs failed, %d\n", err);
+	for (ix--; ix >= 0; ix--)
+		mlx5e_destroy_tir(priv->mdev, &priv->direct_tir[ix]);
+
+	kvfree(in);
+
+	return err;
+}
+
+void mlx5e_destroy_indirect_tirs(struct mlx5e_priv *priv)
+{
+	int i;
+
+	for (i = 0; i < MLX5E_NUM_INDIR_TIRS; i++)
+		mlx5e_destroy_tir(priv->mdev, &priv->indir_tir[i]);
+
+	if (!mlx5e_tunnel_inner_ft_supported(priv->mdev))
+		return;
+
+	for (i = 0; i < MLX5E_NUM_INDIR_TIRS; i++)
+		mlx5e_destroy_tir(priv->mdev, &priv->inner_indir_tir[i]);
+}
+
+void mlx5e_destroy_direct_tirs(struct mlx5e_priv *priv)
+{
+	int nch = priv->profile->max_nch(priv->mdev);
+	int i;
+
+	for (i = 0; i < nch; i++)
+		mlx5e_destroy_tir(priv->mdev, &priv->direct_tir[i]);
+}
+
+static int mlx5e_modify_channels_scatter_fcs(struct mlx5e_channels *chs, bool enable)
+{
+	int err = 0;
+	int i;
+
+	for (i = 0; i < chs->num; i++) {
+		err = mlx5e_modify_rq_scatter_fcs(&chs->c[i]->rq, enable);
+		if (err)
+			return err;
+	}
+
+	return 0;
+}
+
+static int mlx5e_modify_channels_vsd(struct mlx5e_channels *chs, bool vsd)
+{
+	int err = 0;
+	int i;
+
+	for (i = 0; i < chs->num; i++) {
+		err = mlx5e_modify_rq_vsd(&chs->c[i]->rq, vsd);
+		if (err)
+			return err;
+	}
+
+	return 0;
+}
+
+static int mlx5e_setup_tc_mqprio(struct net_device *netdev,
+				 struct tc_mqprio_qopt *mqprio)
+{
+	struct mlx5e_priv *priv = netdev_priv(netdev);
+	struct mlx5e_channels new_channels = {};
+	u8 tc = mqprio->num_tc;
+	int err = 0;
+
+	mqprio->hw = TC_MQPRIO_HW_OFFLOAD_TCS;
+
+	if (tc && tc != MLX5E_MAX_NUM_TC)
+		return -EINVAL;
+
+	mutex_lock(&priv->state_lock);
+
+	new_channels.params = priv->channels.params;
+	new_channels.params.num_tc = tc ? tc : 1;
+
+	if (!test_bit(MLX5E_STATE_OPENED, &priv->state)) {
+		priv->channels.params = new_channels.params;
+		goto out;
+	}
+
+	err = mlx5e_open_channels(priv, &new_channels);
+	if (err)
+		goto out;
+
+	priv->max_opened_tc = max_t(u8, priv->max_opened_tc,
+				    new_channels.params.num_tc);
+	mlx5e_switch_priv_channels(priv, &new_channels, NULL);
+out:
+	mutex_unlock(&priv->state_lock);
+	return err;
+}
+
+#ifdef CONFIG_MLX5_ESWITCH
+static int mlx5e_setup_tc_cls_flower(struct mlx5e_priv *priv,
+				     struct tc_cls_flower_offload *cls_flower,
+				     int flags)
+{
+	switch (cls_flower->command) {
+	case TC_CLSFLOWER_REPLACE:
+		return mlx5e_configure_flower(priv, cls_flower, flags);
+	case TC_CLSFLOWER_DESTROY:
+		return mlx5e_delete_flower(priv, cls_flower, flags);
+	case TC_CLSFLOWER_STATS:
+		return mlx5e_stats_flower(priv, cls_flower, flags);
+	default:
+		return -EOPNOTSUPP;
+	}
+}
+
+static int mlx5e_setup_tc_block_cb(enum tc_setup_type type, void *type_data,
+				   void *cb_priv)
+{
+	struct mlx5e_priv *priv = cb_priv;
+
+	if (!tc_cls_can_offload_and_chain0(priv->netdev, type_data))
+		return -EOPNOTSUPP;
+
+	switch (type) {
+	case TC_SETUP_CLSFLOWER:
+		return mlx5e_setup_tc_cls_flower(priv, type_data, MLX5E_TC_INGRESS);
+	default:
+		return -EOPNOTSUPP;
+	}
+}
+
+static int mlx5e_setup_tc_block(struct net_device *dev,
+				struct tc_block_offload *f)
+{
+	struct mlx5e_priv *priv = netdev_priv(dev);
+
+	if (f->binder_type != TCF_BLOCK_BINDER_TYPE_CLSACT_INGRESS)
+		return -EOPNOTSUPP;
+
+	switch (f->command) {
+	case TC_BLOCK_BIND:
+		return tcf_block_cb_register(f->block, mlx5e_setup_tc_block_cb,
+					     priv, priv, f->extack);
+	case TC_BLOCK_UNBIND:
+		tcf_block_cb_unregister(f->block, mlx5e_setup_tc_block_cb,
+					priv);
+		return 0;
+	default:
+		return -EOPNOTSUPP;
+	}
+}
+#endif
+
+static int mlx5e_setup_tc(struct net_device *dev, enum tc_setup_type type,
+			  void *type_data)
+{
+	switch (type) {
+#ifdef CONFIG_MLX5_ESWITCH
+	case TC_SETUP_BLOCK:
+		return mlx5e_setup_tc_block(dev, type_data);
+#endif
+	case TC_SETUP_QDISC_MQPRIO:
+		return mlx5e_setup_tc_mqprio(dev, type_data);
+	default:
+		return -EOPNOTSUPP;
+	}
+}
+
+static void
+mlx5e_get_stats(struct net_device *dev, struct rtnl_link_stats64 *stats)
+{
+	struct mlx5e_priv *priv = netdev_priv(dev);
+	struct mlx5e_sw_stats *sstats = &priv->stats.sw;
+	struct mlx5e_vport_stats *vstats = &priv->stats.vport;
+	struct mlx5e_pport_stats *pstats = &priv->stats.pport;
+
+	/* update HW stats in background for next time */
+	queue_delayed_work(priv->wq, &priv->update_stats_work, 0);
+
+	if (mlx5e_is_uplink_rep(priv)) {
+		stats->rx_packets = PPORT_802_3_GET(pstats, a_frames_received_ok);
+		stats->rx_bytes   = PPORT_802_3_GET(pstats, a_octets_received_ok);
+		stats->tx_packets = PPORT_802_3_GET(pstats, a_frames_transmitted_ok);
+		stats->tx_bytes   = PPORT_802_3_GET(pstats, a_octets_transmitted_ok);
+	} else {
+		mlx5e_grp_sw_update_stats(priv);
+		stats->rx_packets = sstats->rx_packets;
+		stats->rx_bytes   = sstats->rx_bytes;
+		stats->tx_packets = sstats->tx_packets;
+		stats->tx_bytes   = sstats->tx_bytes;
+		stats->tx_dropped = sstats->tx_queue_dropped;
+	}
+
+	stats->rx_dropped = priv->stats.qcnt.rx_out_of_buffer;
+
+	stats->rx_length_errors =
+		PPORT_802_3_GET(pstats, a_in_range_length_errors) +
+		PPORT_802_3_GET(pstats, a_out_of_range_length_field) +
+		PPORT_802_3_GET(pstats, a_frame_too_long_errors);
+	stats->rx_crc_errors =
+		PPORT_802_3_GET(pstats, a_frame_check_sequence_errors);
+	stats->rx_frame_errors = PPORT_802_3_GET(pstats, a_alignment_errors);
+	stats->tx_aborted_errors = PPORT_2863_GET(pstats, if_out_discards);
+	stats->rx_errors = stats->rx_length_errors + stats->rx_crc_errors +
+			   stats->rx_frame_errors;
+	stats->tx_errors = stats->tx_aborted_errors + stats->tx_carrier_errors;
+
+	/* vport multicast also counts packets that are dropped due to steering
+	 * or rx out of buffer
+	 */
+	stats->multicast =
+		VPORT_COUNTER_GET(vstats, received_eth_multicast.packets);
+}
+
+static void mlx5e_set_rx_mode(struct net_device *dev)
+{
+	struct mlx5e_priv *priv = netdev_priv(dev);
+
+	queue_work(priv->wq, &priv->set_rx_mode_work);
+}
+
+static int mlx5e_set_mac(struct net_device *netdev, void *addr)
+{
+	struct mlx5e_priv *priv = netdev_priv(netdev);
+	struct sockaddr *saddr = addr;
+
+	if (!is_valid_ether_addr(saddr->sa_data))
+		return -EADDRNOTAVAIL;
+
+	netif_addr_lock_bh(netdev);
+	ether_addr_copy(netdev->dev_addr, saddr->sa_data);
+	netif_addr_unlock_bh(netdev);
+
+	queue_work(priv->wq, &priv->set_rx_mode_work);
+
+	return 0;
+}
+
+#define MLX5E_SET_FEATURE(features, feature, enable)	\
+	do {						\
+		if (enable)				\
+			*features |= feature;		\
+		else					\
+			*features &= ~feature;		\
+	} while (0)
+
+typedef int (*mlx5e_feature_handler)(struct net_device *netdev, bool enable);
+
+static int set_feature_lro(struct net_device *netdev, bool enable)
+{
+	struct mlx5e_priv *priv = netdev_priv(netdev);
+	struct mlx5_core_dev *mdev = priv->mdev;
+	struct mlx5e_channels new_channels = {};
+	struct mlx5e_params *old_params;
+	int err = 0;
+	bool reset;
+
+	mutex_lock(&priv->state_lock);
+
+	old_params = &priv->channels.params;
+	if (enable && !MLX5E_GET_PFLAG(old_params, MLX5E_PFLAG_RX_STRIDING_RQ)) {
+		netdev_warn(netdev, "can't set LRO with legacy RQ\n");
+		err = -EINVAL;
+		goto out;
+	}
+
+	reset = test_bit(MLX5E_STATE_OPENED, &priv->state);
+
+	new_channels.params = *old_params;
+	new_channels.params.lro_en = enable;
+
+	if (old_params->rq_wq_type != MLX5_WQ_TYPE_CYCLIC) {
+		if (mlx5e_rx_mpwqe_is_linear_skb(mdev, old_params) ==
+		    mlx5e_rx_mpwqe_is_linear_skb(mdev, &new_channels.params))
+			reset = false;
+	}
+
+	if (!reset) {
+		*old_params = new_channels.params;
+		err = mlx5e_modify_tirs_lro(priv);
+		goto out;
+	}
+
+	err = mlx5e_open_channels(priv, &new_channels);
+	if (err)
+		goto out;
+
+	mlx5e_switch_priv_channels(priv, &new_channels, mlx5e_modify_tirs_lro);
+out:
+	mutex_unlock(&priv->state_lock);
+	return err;
+}
+
+static int set_feature_cvlan_filter(struct net_device *netdev, bool enable)
+{
+	struct mlx5e_priv *priv = netdev_priv(netdev);
+
+	if (enable)
+		mlx5e_enable_cvlan_filter(priv);
+	else
+		mlx5e_disable_cvlan_filter(priv);
+
+	return 0;
+}
+
+#ifdef CONFIG_MLX5_ESWITCH
+static int set_feature_tc_num_filters(struct net_device *netdev, bool enable)
+{
+	struct mlx5e_priv *priv = netdev_priv(netdev);
+
+	if (!enable && mlx5e_tc_num_filters(priv)) {
+		netdev_err(netdev,
+			   "Active offloaded tc filters, can't turn hw_tc_offload off\n");
+		return -EINVAL;
+	}
+
+	return 0;
+}
+#endif
+
+static int set_feature_rx_all(struct net_device *netdev, bool enable)
+{
+	struct mlx5e_priv *priv = netdev_priv(netdev);
+	struct mlx5_core_dev *mdev = priv->mdev;
+
+	return mlx5_set_port_fcs(mdev, !enable);
+}
+
+static int set_feature_rx_fcs(struct net_device *netdev, bool enable)
+{
+	struct mlx5e_priv *priv = netdev_priv(netdev);
+	int err;
+
+	mutex_lock(&priv->state_lock);
+
+	priv->channels.params.scatter_fcs_en = enable;
+	err = mlx5e_modify_channels_scatter_fcs(&priv->channels, enable);
+	if (err)
+		priv->channels.params.scatter_fcs_en = !enable;
+
+	mutex_unlock(&priv->state_lock);
+
+	return err;
+}
+
+static int set_feature_rx_vlan(struct net_device *netdev, bool enable)
+{
+	struct mlx5e_priv *priv = netdev_priv(netdev);
+	int err = 0;
+
+	mutex_lock(&priv->state_lock);
+
+	priv->channels.params.vlan_strip_disable = !enable;
+	if (!test_bit(MLX5E_STATE_OPENED, &priv->state))
+		goto unlock;
+
+	err = mlx5e_modify_channels_vsd(&priv->channels, !enable);
+	if (err)
+		priv->channels.params.vlan_strip_disable = enable;
+
+unlock:
+	mutex_unlock(&priv->state_lock);
+
+	return err;
+}
+
+#ifdef CONFIG_MLX5_EN_ARFS
+static int set_feature_arfs(struct net_device *netdev, bool enable)
+{
+	struct mlx5e_priv *priv = netdev_priv(netdev);
+	int err;
+
+	if (enable)
+		err = mlx5e_arfs_enable(priv);
+	else
+		err = mlx5e_arfs_disable(priv);
+
+	return err;
+}
+#endif
+
+static int mlx5e_handle_feature(struct net_device *netdev,
+				netdev_features_t *features,
+				netdev_features_t feature,
+				mlx5e_feature_handler feature_handler)
+{
+	netdev_features_t changes = *features ^ netdev->features;
+	bool enable = !!(*features & feature);
+	int err;
+
+	if (!(changes & feature))
+		return 0;
+
+	err = feature_handler(netdev, enable);
+	if (err) {
+		MLX5E_SET_FEATURE(features, feature, !enable);
+		netdev_err(netdev, "%s feature %pNF failed, err %d\n",
+			   enable ? "Enable" : "Disable", &feature, err);
+		return err;
+	}
+
+	return 0;
+}
+
+static int mlx5e_set_features(struct net_device *netdev,
+			      netdev_features_t features)
+{
+	netdev_features_t oper_features = features;
+	int err = 0;
+
+#define MLX5E_HANDLE_FEATURE(feature, handler) \
+	mlx5e_handle_feature(netdev, &oper_features, feature, handler)
+
+	err |= MLX5E_HANDLE_FEATURE(NETIF_F_LRO, set_feature_lro);
+	err |= MLX5E_HANDLE_FEATURE(NETIF_F_HW_VLAN_CTAG_FILTER,
+				    set_feature_cvlan_filter);
+#ifdef CONFIG_MLX5_ESWITCH
+	err |= MLX5E_HANDLE_FEATURE(NETIF_F_HW_TC, set_feature_tc_num_filters);
+#endif
+	err |= MLX5E_HANDLE_FEATURE(NETIF_F_RXALL, set_feature_rx_all);
+	err |= MLX5E_HANDLE_FEATURE(NETIF_F_RXFCS, set_feature_rx_fcs);
+	err |= MLX5E_HANDLE_FEATURE(NETIF_F_HW_VLAN_CTAG_RX, set_feature_rx_vlan);
+#ifdef CONFIG_MLX5_EN_ARFS
+	err |= MLX5E_HANDLE_FEATURE(NETIF_F_NTUPLE, set_feature_arfs);
+#endif
+
+	if (err) {
+		netdev->features = oper_features;
+		return -EINVAL;
+	}
+
+	return 0;
+}
+
+static netdev_features_t mlx5e_fix_features(struct net_device *netdev,
+					    netdev_features_t features)
+{
+	struct mlx5e_priv *priv = netdev_priv(netdev);
+	struct mlx5e_params *params;
+
+	mutex_lock(&priv->state_lock);
+	params = &priv->channels.params;
+	if (!bitmap_empty(priv->fs.vlan.active_svlans, VLAN_N_VID)) {
+		/* HW strips the outer C-tag header, this is a problem
+		 * for S-tag traffic.
+		 */
+		features &= ~NETIF_F_HW_VLAN_CTAG_RX;
+		if (!params->vlan_strip_disable)
+			netdev_warn(netdev, "Dropping C-tag vlan stripping offload due to S-tag vlan\n");
+	}
+	if (!MLX5E_GET_PFLAG(params, MLX5E_PFLAG_RX_STRIDING_RQ)) {
+		features &= ~NETIF_F_LRO;
+		if (params->lro_en)
+			netdev_warn(netdev, "Disabling LRO, not supported in legacy RQ\n");
+	}
+
+	if (params->xdp_prog) {
+		if (features & NETIF_F_LRO) {
+			netdev_warn(netdev, "LRO is incompatible with XDP\n");
+			features &= ~NETIF_F_LRO;
+		}
+	}
+
+	if (MLX5E_GET_PFLAG(params, MLX5E_PFLAG_RX_CQE_COMPRESS)) {
+		features &= ~NETIF_F_RXHASH;
+		if (netdev->features & NETIF_F_RXHASH)
+			netdev_warn(netdev, "Disabling rxhash, not supported when CQE compress is active\n");
+	}
+
+	mutex_unlock(&priv->state_lock);
+
+	return features;
+}
+
+int mlx5e_change_mtu(struct net_device *netdev, int new_mtu,
+		     change_hw_mtu_cb set_mtu_cb)
+{
+	struct mlx5e_priv *priv = netdev_priv(netdev);
+	struct mlx5e_channels new_channels = {};
+	struct mlx5e_params *params;
+	int err = 0;
+	bool reset;
+
+	mutex_lock(&priv->state_lock);
+
+	params = &priv->channels.params;
+
+	reset = !params->lro_en;
+	reset = reset && test_bit(MLX5E_STATE_OPENED, &priv->state);
+
+	new_channels.params = *params;
+	new_channels.params.sw_mtu = new_mtu;
+
+	if (params->xdp_prog &&
+	    !mlx5e_rx_is_linear_skb(priv->mdev, &new_channels.params)) {
+		netdev_err(netdev, "MTU(%d) > %d is not allowed while XDP enabled\n",
+			   new_mtu, mlx5e_xdp_max_mtu(params));
+		err = -EINVAL;
+		goto out;
+	}
+
+	if (params->rq_wq_type == MLX5_WQ_TYPE_LINKED_LIST_STRIDING_RQ) {
+		bool is_linear = mlx5e_rx_mpwqe_is_linear_skb(priv->mdev, &new_channels.params);
+		u8 ppw_old = mlx5e_mpwqe_log_pkts_per_wqe(params);
+		u8 ppw_new = mlx5e_mpwqe_log_pkts_per_wqe(&new_channels.params);
+
+		reset = reset && (is_linear || (ppw_old != ppw_new));
+	}
+
+	if (!reset) {
+		params->sw_mtu = new_mtu;
+		if (set_mtu_cb)
+			set_mtu_cb(priv);
+		netdev->mtu = params->sw_mtu;
+		goto out;
+	}
+
+	err = mlx5e_open_channels(priv, &new_channels);
+	if (err)
+		goto out;
+
+	mlx5e_switch_priv_channels(priv, &new_channels, set_mtu_cb);
+	netdev->mtu = new_channels.params.sw_mtu;
+
+out:
+	mutex_unlock(&priv->state_lock);
+	return err;
+}
+
+static int mlx5e_change_nic_mtu(struct net_device *netdev, int new_mtu)
+{
+	return mlx5e_change_mtu(netdev, new_mtu, mlx5e_set_dev_port_mtu);
+}
+
+int mlx5e_hwstamp_set(struct mlx5e_priv *priv, struct ifreq *ifr)
+{
+	struct hwtstamp_config config;
+	int err;
+
+	if (!MLX5_CAP_GEN(priv->mdev, device_frequency_khz) ||
+	    (mlx5_clock_get_ptp_index(priv->mdev) == -1))
+		return -EOPNOTSUPP;
+
+	if (copy_from_user(&config, ifr->ifr_data, sizeof(config)))
+		return -EFAULT;
+
+	/* TX HW timestamp */
+	switch (config.tx_type) {
+	case HWTSTAMP_TX_OFF:
+	case HWTSTAMP_TX_ON:
+		break;
+	default:
+		return -ERANGE;
+	}
+
+	mutex_lock(&priv->state_lock);
+	/* RX HW timestamp */
+	switch (config.rx_filter) {
+	case HWTSTAMP_FILTER_NONE:
+		/* Reset CQE compression to Admin default */
+		mlx5e_modify_rx_cqe_compression_locked(priv, priv->channels.params.rx_cqe_compress_def);
+		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:
+		/* Disable CQE compression */
+		netdev_warn(priv->netdev, "Disabling cqe compression");
+		err = mlx5e_modify_rx_cqe_compression_locked(priv, false);
+		if (err) {
+			netdev_err(priv->netdev, "Failed disabling cqe compression err=%d\n", err);
+			mutex_unlock(&priv->state_lock);
+			return err;
+		}
+		config.rx_filter = HWTSTAMP_FILTER_ALL;
+		break;
+	default:
+		mutex_unlock(&priv->state_lock);
+		return -ERANGE;
+	}
+
+	memcpy(&priv->tstamp, &config, sizeof(config));
+	mutex_unlock(&priv->state_lock);
+
+	/* might need to fix some features */
+	netdev_update_features(priv->netdev);
+
+	return copy_to_user(ifr->ifr_data, &config,
+			    sizeof(config)) ? -EFAULT : 0;
+}
+
+int mlx5e_hwstamp_get(struct mlx5e_priv *priv, struct ifreq *ifr)
+{
+	struct hwtstamp_config *cfg = &priv->tstamp;
+
+	if (!MLX5_CAP_GEN(priv->mdev, device_frequency_khz))
+		return -EOPNOTSUPP;
+
+	return copy_to_user(ifr->ifr_data, cfg, sizeof(*cfg)) ? -EFAULT : 0;
+}
+
+static int mlx5e_ioctl(struct net_device *dev, struct ifreq *ifr, int cmd)
+{
+	struct mlx5e_priv *priv = netdev_priv(dev);
+
+	switch (cmd) {
+	case SIOCSHWTSTAMP:
+		return mlx5e_hwstamp_set(priv, ifr);
+	case SIOCGHWTSTAMP:
+		return mlx5e_hwstamp_get(priv, ifr);
+	default:
+		return -EOPNOTSUPP;
+	}
+}
+
+#ifdef CONFIG_MLX5_ESWITCH
+static int mlx5e_set_vf_mac(struct net_device *dev, int vf, u8 *mac)
+{
+	struct mlx5e_priv *priv = netdev_priv(dev);
+	struct mlx5_core_dev *mdev = priv->mdev;
+
+	return mlx5_eswitch_set_vport_mac(mdev->priv.eswitch, vf + 1, mac);
+}
+
+static int mlx5e_set_vf_vlan(struct net_device *dev, int vf, u16 vlan, u8 qos,
+			     __be16 vlan_proto)
+{
+	struct mlx5e_priv *priv = netdev_priv(dev);
+	struct mlx5_core_dev *mdev = priv->mdev;
+
+	if (vlan_proto != htons(ETH_P_8021Q))
+		return -EPROTONOSUPPORT;
+
+	return mlx5_eswitch_set_vport_vlan(mdev->priv.eswitch, vf + 1,
+					   vlan, qos);
+}
+
+static int mlx5e_set_vf_spoofchk(struct net_device *dev, int vf, bool setting)
+{
+	struct mlx5e_priv *priv = netdev_priv(dev);
+	struct mlx5_core_dev *mdev = priv->mdev;
+
+	return mlx5_eswitch_set_vport_spoofchk(mdev->priv.eswitch, vf + 1, setting);
+}
+
+static int mlx5e_set_vf_trust(struct net_device *dev, int vf, bool setting)
+{
+	struct mlx5e_priv *priv = netdev_priv(dev);
+	struct mlx5_core_dev *mdev = priv->mdev;
+
+	return mlx5_eswitch_set_vport_trust(mdev->priv.eswitch, vf + 1, setting);
+}
+
+static int mlx5e_set_vf_rate(struct net_device *dev, int vf, int min_tx_rate,
+			     int max_tx_rate)
+{
+	struct mlx5e_priv *priv = netdev_priv(dev);
+	struct mlx5_core_dev *mdev = priv->mdev;
+
+	return mlx5_eswitch_set_vport_rate(mdev->priv.eswitch, vf + 1,
+					   max_tx_rate, min_tx_rate);
+}
+
+static int mlx5_vport_link2ifla(u8 esw_link)
+{
+	switch (esw_link) {
+	case MLX5_VPORT_ADMIN_STATE_DOWN:
+		return IFLA_VF_LINK_STATE_DISABLE;
+	case MLX5_VPORT_ADMIN_STATE_UP:
+		return IFLA_VF_LINK_STATE_ENABLE;
+	}
+	return IFLA_VF_LINK_STATE_AUTO;
+}
+
+static int mlx5_ifla_link2vport(u8 ifla_link)
+{
+	switch (ifla_link) {
+	case IFLA_VF_LINK_STATE_DISABLE:
+		return MLX5_VPORT_ADMIN_STATE_DOWN;
+	case IFLA_VF_LINK_STATE_ENABLE:
+		return MLX5_VPORT_ADMIN_STATE_UP;
+	}
+	return MLX5_VPORT_ADMIN_STATE_AUTO;
+}
+
+static int mlx5e_set_vf_link_state(struct net_device *dev, int vf,
+				   int link_state)
+{
+	struct mlx5e_priv *priv = netdev_priv(dev);
+	struct mlx5_core_dev *mdev = priv->mdev;
+
+	return mlx5_eswitch_set_vport_state(mdev->priv.eswitch, vf + 1,
+					    mlx5_ifla_link2vport(link_state));
+}
+
+static int mlx5e_get_vf_config(struct net_device *dev,
+			       int vf, struct ifla_vf_info *ivi)
+{
+	struct mlx5e_priv *priv = netdev_priv(dev);
+	struct mlx5_core_dev *mdev = priv->mdev;
+	int err;
+
+	err = mlx5_eswitch_get_vport_config(mdev->priv.eswitch, vf + 1, ivi);
+	if (err)
+		return err;
+	ivi->linkstate = mlx5_vport_link2ifla(ivi->linkstate);
+	return 0;
+}
+
+static int mlx5e_get_vf_stats(struct net_device *dev,
+			      int vf, struct ifla_vf_stats *vf_stats)
+{
+	struct mlx5e_priv *priv = netdev_priv(dev);
+	struct mlx5_core_dev *mdev = priv->mdev;
+
+	return mlx5_eswitch_get_vport_stats(mdev->priv.eswitch, vf + 1,
+					    vf_stats);
+}
+#endif
+
+struct mlx5e_vxlan_work {
+	struct work_struct	work;
+	struct mlx5e_priv	*priv;
+	u16			port;
+};
+
+static void mlx5e_vxlan_add_work(struct work_struct *work)
+{
+	struct mlx5e_vxlan_work *vxlan_work =
+		container_of(work, struct mlx5e_vxlan_work, work);
+	struct mlx5e_priv *priv = vxlan_work->priv;
+	u16 port = vxlan_work->port;
+
+	mutex_lock(&priv->state_lock);
+	mlx5_vxlan_add_port(priv->mdev->vxlan, port);
+	mutex_unlock(&priv->state_lock);
+
+	kfree(vxlan_work);
+}
+
+static void mlx5e_vxlan_del_work(struct work_struct *work)
+{
+	struct mlx5e_vxlan_work *vxlan_work =
+		container_of(work, struct mlx5e_vxlan_work, work);
+	struct mlx5e_priv *priv         = vxlan_work->priv;
+	u16 port = vxlan_work->port;
+
+	mutex_lock(&priv->state_lock);
+	mlx5_vxlan_del_port(priv->mdev->vxlan, port);
+	mutex_unlock(&priv->state_lock);
+	kfree(vxlan_work);
+}
+
+static void mlx5e_vxlan_queue_work(struct mlx5e_priv *priv, u16 port, int add)
+{
+	struct mlx5e_vxlan_work *vxlan_work;
+
+	vxlan_work = kmalloc(sizeof(*vxlan_work), GFP_ATOMIC);
+	if (!vxlan_work)
+		return;
+
+	if (add)
+		INIT_WORK(&vxlan_work->work, mlx5e_vxlan_add_work);
+	else
+		INIT_WORK(&vxlan_work->work, mlx5e_vxlan_del_work);
+
+	vxlan_work->priv = priv;
+	vxlan_work->port = port;
+	queue_work(priv->wq, &vxlan_work->work);
+}
+
+static void mlx5e_add_vxlan_port(struct net_device *netdev,
+				 struct udp_tunnel_info *ti)
+{
+	struct mlx5e_priv *priv = netdev_priv(netdev);
+
+	if (ti->type != UDP_TUNNEL_TYPE_VXLAN)
+		return;
+
+	if (!mlx5_vxlan_allowed(priv->mdev->vxlan))
+		return;
+
+	mlx5e_vxlan_queue_work(priv, be16_to_cpu(ti->port), 1);
+}
+
+static void mlx5e_del_vxlan_port(struct net_device *netdev,
+				 struct udp_tunnel_info *ti)
+{
+	struct mlx5e_priv *priv = netdev_priv(netdev);
+
+	if (ti->type != UDP_TUNNEL_TYPE_VXLAN)
+		return;
+
+	if (!mlx5_vxlan_allowed(priv->mdev->vxlan))
+		return;
+
+	mlx5e_vxlan_queue_work(priv, be16_to_cpu(ti->port), 0);
+}
+
+static netdev_features_t mlx5e_tunnel_features_check(struct mlx5e_priv *priv,
+						     struct sk_buff *skb,
+						     netdev_features_t features)
+{
+	unsigned int offset = 0;
+	struct udphdr *udph;
+	u8 proto;
+	u16 port;
+
+	switch (vlan_get_protocol(skb)) {
+	case htons(ETH_P_IP):
+		proto = ip_hdr(skb)->protocol;
+		break;
+	case htons(ETH_P_IPV6):
+		proto = ipv6_find_hdr(skb, &offset, -1, NULL, NULL);
+		break;
+	default:
+		goto out;
+	}
+
+	switch (proto) {
+	case IPPROTO_GRE:
+		return features;
+	case IPPROTO_UDP:
+		udph = udp_hdr(skb);
+		port = be16_to_cpu(udph->dest);
+
+		/* Verify if UDP port is being offloaded by HW */
+		if (mlx5_vxlan_lookup_port(priv->mdev->vxlan, port))
+			return features;
+	}
+
+out:
+	/* Disable CSUM and GSO if the udp dport is not offloaded by HW */
+	return features & ~(NETIF_F_CSUM_MASK | NETIF_F_GSO_MASK);
+}
+
+static netdev_features_t mlx5e_features_check(struct sk_buff *skb,
+					      struct net_device *netdev,
+					      netdev_features_t features)
+{
+	struct mlx5e_priv *priv = netdev_priv(netdev);
+
+	features = vlan_features_check(skb, features);
+	features = vxlan_features_check(skb, features);
+
+#ifdef CONFIG_MLX5_EN_IPSEC
+	if (mlx5e_ipsec_feature_check(skb, netdev, features))
+		return features;
+#endif
+
+	/* Validate if the tunneled packet is being offloaded by HW */
+	if (skb->encapsulation &&
+	    (features & NETIF_F_CSUM_MASK || features & NETIF_F_GSO_MASK))
+		return mlx5e_tunnel_features_check(priv, skb, features);
+
+	return features;
+}
+
+static bool mlx5e_tx_timeout_eq_recover(struct net_device *dev,
+					struct mlx5e_txqsq *sq)
+{
+	struct mlx5_eq *eq = sq->cq.mcq.eq;
+	u32 eqe_count;
+
+	netdev_err(dev, "EQ 0x%x: Cons = 0x%x, irqn = 0x%x\n",
+		   eq->eqn, eq->cons_index, eq->irqn);
+
+	eqe_count = mlx5_eq_poll_irq_disabled(eq);
+	if (!eqe_count)
+		return false;
+
+	netdev_err(dev, "Recover %d eqes on EQ 0x%x\n", eqe_count, eq->eqn);
+	sq->channel->stats->eq_rearm++;
+	return true;
+}
+
+static void mlx5e_tx_timeout_work(struct work_struct *work)
+{
+	struct mlx5e_priv *priv = container_of(work, struct mlx5e_priv,
+					       tx_timeout_work);
+	struct net_device *dev = priv->netdev;
+	bool reopen_channels = false;
+	int i, err;
+
+	rtnl_lock();
+	mutex_lock(&priv->state_lock);
+
+	if (!test_bit(MLX5E_STATE_OPENED, &priv->state))
+		goto unlock;
+
+	for (i = 0; i < priv->channels.num * priv->channels.params.num_tc; i++) {
+		struct netdev_queue *dev_queue = netdev_get_tx_queue(dev, i);
+		struct mlx5e_txqsq *sq = priv->txq2sq[i];
+
+		if (!netif_xmit_stopped(dev_queue))
+			continue;
+
+		netdev_err(dev,
+			   "TX timeout on queue: %d, SQ: 0x%x, CQ: 0x%x, SQ Cons: 0x%x SQ Prod: 0x%x, usecs since last trans: %u\n",
+			   i, sq->sqn, sq->cq.mcq.cqn, sq->cc, sq->pc,
+			   jiffies_to_usecs(jiffies - dev_queue->trans_start));
+
+		/* If we recover a lost interrupt, most likely TX timeout will
+		 * be resolved, skip reopening channels
+		 */
+		if (!mlx5e_tx_timeout_eq_recover(dev, sq)) {
+			clear_bit(MLX5E_SQ_STATE_ENABLED, &sq->state);
+			reopen_channels = true;
+		}
+	}
+
+	if (!reopen_channels)
+		goto unlock;
+
+	mlx5e_close_locked(dev);
+	err = mlx5e_open_locked(dev);
+	if (err)
+		netdev_err(priv->netdev,
+			   "mlx5e_open_locked failed recovering from a tx_timeout, err(%d).\n",
+			   err);
+
+unlock:
+	mutex_unlock(&priv->state_lock);
+	rtnl_unlock();
+}
+
+static void mlx5e_tx_timeout(struct net_device *dev)
+{
+	struct mlx5e_priv *priv = netdev_priv(dev);
+
+	netdev_err(dev, "TX timeout detected\n");
+	queue_work(priv->wq, &priv->tx_timeout_work);
+}
+
+static int mlx5e_xdp_allowed(struct mlx5e_priv *priv, struct bpf_prog *prog)
+{
+	struct net_device *netdev = priv->netdev;
+	struct mlx5e_channels new_channels = {};
+
+	if (priv->channels.params.lro_en) {
+		netdev_warn(netdev, "can't set XDP while LRO is on, disable LRO first\n");
+		return -EINVAL;
+	}
+
+	if (MLX5_IPSEC_DEV(priv->mdev)) {
+		netdev_warn(netdev, "can't set XDP with IPSec offload\n");
+		return -EINVAL;
+	}
+
+	new_channels.params = priv->channels.params;
+	new_channels.params.xdp_prog = prog;
+
+	if (!mlx5e_rx_is_linear_skb(priv->mdev, &new_channels.params)) {
+		netdev_warn(netdev, "XDP is not allowed with MTU(%d) > %d\n",
+			    new_channels.params.sw_mtu,
+			    mlx5e_xdp_max_mtu(&new_channels.params));
+		return -EINVAL;
+	}
+
+	return 0;
+}
+
+static int mlx5e_xdp_set(struct net_device *netdev, struct bpf_prog *prog)
+{
+	struct mlx5e_priv *priv = netdev_priv(netdev);
+	struct bpf_prog *old_prog;
+	bool reset, was_opened;
+	int err = 0;
+	int i;
+
+	mutex_lock(&priv->state_lock);
+
+	if (prog) {
+		err = mlx5e_xdp_allowed(priv, prog);
+		if (err)
+			goto unlock;
+	}
+
+	was_opened = test_bit(MLX5E_STATE_OPENED, &priv->state);
+	/* no need for full reset when exchanging programs */
+	reset = (!priv->channels.params.xdp_prog || !prog);
+
+	if (was_opened && reset)
+		mlx5e_close_locked(netdev);
+	if (was_opened && !reset) {
+		/* num_channels is invariant here, so we can take the
+		 * batched reference right upfront.
+		 */
+		prog = bpf_prog_add(prog, priv->channels.num);
+		if (IS_ERR(prog)) {
+			err = PTR_ERR(prog);
+			goto unlock;
+		}
+	}
+
+	/* exchange programs, extra prog reference we got from caller
+	 * as long as we don't fail from this point onwards.
+	 */
+	old_prog = xchg(&priv->channels.params.xdp_prog, prog);
+	if (old_prog)
+		bpf_prog_put(old_prog);
+
+	if (reset) /* change RQ type according to priv->xdp_prog */
+		mlx5e_set_rq_type(priv->mdev, &priv->channels.params);
+
+	if (was_opened && reset)
+		mlx5e_open_locked(netdev);
+
+	if (!test_bit(MLX5E_STATE_OPENED, &priv->state) || reset)
+		goto unlock;
+
+	/* exchanging programs w/o reset, we update ref counts on behalf
+	 * of the channels RQs here.
+	 */
+	for (i = 0; i < priv->channels.num; i++) {
+		struct mlx5e_channel *c = priv->channels.c[i];
+
+		clear_bit(MLX5E_RQ_STATE_ENABLED, &c->rq.state);
+		napi_synchronize(&c->napi);
+		/* prevent mlx5e_poll_rx_cq from accessing rq->xdp_prog */
+
+		old_prog = xchg(&c->rq.xdp_prog, prog);
+
+		set_bit(MLX5E_RQ_STATE_ENABLED, &c->rq.state);
+		/* napi_schedule in case we have missed anything */
+		napi_schedule(&c->napi);
+
+		if (old_prog)
+			bpf_prog_put(old_prog);
+	}
+
+unlock:
+	mutex_unlock(&priv->state_lock);
+
+	/* Need to fix some features. */
+	if (!err)
+		netdev_update_features(netdev);
+
+	return err;
+}
+
+static u32 mlx5e_xdp_query(struct net_device *dev)
+{
+	struct mlx5e_priv *priv = netdev_priv(dev);
+	const struct bpf_prog *xdp_prog;
+	u32 prog_id = 0;
+
+	mutex_lock(&priv->state_lock);
+	xdp_prog = priv->channels.params.xdp_prog;
+	if (xdp_prog)
+		prog_id = xdp_prog->aux->id;
+	mutex_unlock(&priv->state_lock);
+
+	return prog_id;
+}
+
+static int mlx5e_xdp(struct net_device *dev, struct netdev_bpf *xdp)
+{
+	switch (xdp->command) {
+	case XDP_SETUP_PROG:
+		return mlx5e_xdp_set(dev, xdp->prog);
+	case XDP_QUERY_PROG:
+		xdp->prog_id = mlx5e_xdp_query(dev);
+		return 0;
+	default:
+		return -EINVAL;
+	}
+}
+
+const struct net_device_ops mlx5e_netdev_ops = {
+	.ndo_open                = mlx5e_open,
+	.ndo_stop                = mlx5e_close,
+	.ndo_start_xmit          = mlx5e_xmit,
+	.ndo_setup_tc            = mlx5e_setup_tc,
+	.ndo_select_queue        = mlx5e_select_queue,
+	.ndo_get_stats64         = mlx5e_get_stats,
+	.ndo_set_rx_mode         = mlx5e_set_rx_mode,
+	.ndo_set_mac_address     = mlx5e_set_mac,
+	.ndo_vlan_rx_add_vid     = mlx5e_vlan_rx_add_vid,
+	.ndo_vlan_rx_kill_vid    = mlx5e_vlan_rx_kill_vid,
+	.ndo_set_features        = mlx5e_set_features,
+	.ndo_fix_features        = mlx5e_fix_features,
+	.ndo_change_mtu          = mlx5e_change_nic_mtu,
+	.ndo_do_ioctl            = mlx5e_ioctl,
+	.ndo_set_tx_maxrate      = mlx5e_set_tx_maxrate,
+	.ndo_udp_tunnel_add      = mlx5e_add_vxlan_port,
+	.ndo_udp_tunnel_del      = mlx5e_del_vxlan_port,
+	.ndo_features_check      = mlx5e_features_check,
+	.ndo_tx_timeout          = mlx5e_tx_timeout,
+	.ndo_bpf		 = mlx5e_xdp,
+	.ndo_xdp_xmit            = mlx5e_xdp_xmit,
+#ifdef CONFIG_MLX5_EN_ARFS
+	.ndo_rx_flow_steer	 = mlx5e_rx_flow_steer,
+#endif
+#ifdef CONFIG_MLX5_ESWITCH
+	/* SRIOV E-Switch NDOs */
+	.ndo_set_vf_mac          = mlx5e_set_vf_mac,
+	.ndo_set_vf_vlan         = mlx5e_set_vf_vlan,
+	.ndo_set_vf_spoofchk     = mlx5e_set_vf_spoofchk,
+	.ndo_set_vf_trust        = mlx5e_set_vf_trust,
+	.ndo_set_vf_rate         = mlx5e_set_vf_rate,
+	.ndo_get_vf_config       = mlx5e_get_vf_config,
+	.ndo_set_vf_link_state   = mlx5e_set_vf_link_state,
+	.ndo_get_vf_stats        = mlx5e_get_vf_stats,
+	.ndo_has_offload_stats	 = mlx5e_has_offload_stats,
+	.ndo_get_offload_stats	 = mlx5e_get_offload_stats,
+#endif
+};
+
+static int mlx5e_check_required_hca_cap(struct mlx5_core_dev *mdev)
+{
+	if (MLX5_CAP_GEN(mdev, port_type) != MLX5_CAP_PORT_TYPE_ETH)
+		return -EOPNOTSUPP;
+	if (!MLX5_CAP_GEN(mdev, eth_net_offloads) ||
+	    !MLX5_CAP_GEN(mdev, nic_flow_table) ||
+	    !MLX5_CAP_ETH(mdev, csum_cap) ||
+	    !MLX5_CAP_ETH(mdev, max_lso_cap) ||
+	    !MLX5_CAP_ETH(mdev, vlan_cap) ||
+	    !MLX5_CAP_ETH(mdev, rss_ind_tbl_cap) ||
+	    MLX5_CAP_FLOWTABLE(mdev,
+			       flow_table_properties_nic_receive.max_ft_level)
+			       < 3) {
+		mlx5_core_warn(mdev,
+			       "Not creating net device, some required device capabilities are missing\n");
+		return -EOPNOTSUPP;
+	}
+	if (!MLX5_CAP_ETH(mdev, self_lb_en_modifiable))
+		mlx5_core_warn(mdev, "Self loop back prevention is not supported\n");
+	if (!MLX5_CAP_GEN(mdev, cq_moderation))
+		mlx5_core_warn(mdev, "CQ moderation is not supported\n");
+
+	return 0;
+}
+
+void mlx5e_build_default_indir_rqt(u32 *indirection_rqt, int len,
+				   int num_channels)
+{
+	int i;
+
+	for (i = 0; i < len; i++)
+		indirection_rqt[i] = i % num_channels;
+}
+
+static bool slow_pci_heuristic(struct mlx5_core_dev *mdev)
+{
+	u32 link_speed = 0;
+	u32 pci_bw = 0;
+
+	mlx5e_port_max_linkspeed(mdev, &link_speed);
+	pci_bw = pcie_bandwidth_available(mdev->pdev, NULL, NULL, NULL);
+	mlx5_core_dbg_once(mdev, "Max link speed = %d, PCI BW = %d\n",
+			   link_speed, pci_bw);
+
+#define MLX5E_SLOW_PCI_RATIO (2)
+
+	return link_speed && pci_bw &&
+		link_speed > MLX5E_SLOW_PCI_RATIO * pci_bw;
+}
+
+static struct net_dim_cq_moder mlx5e_get_def_tx_moderation(u8 cq_period_mode)
+{
+	struct net_dim_cq_moder moder;
+
+	moder.cq_period_mode = cq_period_mode;
+	moder.pkts = MLX5E_PARAMS_DEFAULT_TX_CQ_MODERATION_PKTS;
+	moder.usec = MLX5E_PARAMS_DEFAULT_TX_CQ_MODERATION_USEC;
+	if (cq_period_mode == MLX5_CQ_PERIOD_MODE_START_FROM_CQE)
+		moder.usec = MLX5E_PARAMS_DEFAULT_TX_CQ_MODERATION_USEC_FROM_CQE;
+
+	return moder;
+}
+
+static struct net_dim_cq_moder mlx5e_get_def_rx_moderation(u8 cq_period_mode)
+{
+	struct net_dim_cq_moder moder;
+
+	moder.cq_period_mode = cq_period_mode;
+	moder.pkts = MLX5E_PARAMS_DEFAULT_RX_CQ_MODERATION_PKTS;
+	moder.usec = MLX5E_PARAMS_DEFAULT_RX_CQ_MODERATION_USEC;
+	if (cq_period_mode == MLX5_CQ_PERIOD_MODE_START_FROM_CQE)
+		moder.usec = MLX5E_PARAMS_DEFAULT_RX_CQ_MODERATION_USEC_FROM_CQE;
+
+	return moder;
+}
+
+static u8 mlx5_to_net_dim_cq_period_mode(u8 cq_period_mode)
+{
+	return cq_period_mode == MLX5_CQ_PERIOD_MODE_START_FROM_CQE ?
+		NET_DIM_CQ_PERIOD_MODE_START_FROM_CQE :
+		NET_DIM_CQ_PERIOD_MODE_START_FROM_EQE;
+}
+
+void mlx5e_set_tx_cq_mode_params(struct mlx5e_params *params, u8 cq_period_mode)
+{
+	if (params->tx_dim_enabled) {
+		u8 dim_period_mode = mlx5_to_net_dim_cq_period_mode(cq_period_mode);
+
+		params->tx_cq_moderation = net_dim_get_def_tx_moderation(dim_period_mode);
+	} else {
+		params->tx_cq_moderation = mlx5e_get_def_tx_moderation(cq_period_mode);
+	}
+
+	MLX5E_SET_PFLAG(params, MLX5E_PFLAG_TX_CQE_BASED_MODER,
+			params->tx_cq_moderation.cq_period_mode ==
+				MLX5_CQ_PERIOD_MODE_START_FROM_CQE);
+}
+
+void mlx5e_set_rx_cq_mode_params(struct mlx5e_params *params, u8 cq_period_mode)
+{
+	if (params->rx_dim_enabled) {
+		u8 dim_period_mode = mlx5_to_net_dim_cq_period_mode(cq_period_mode);
+
+		params->rx_cq_moderation = net_dim_get_def_rx_moderation(dim_period_mode);
+	} else {
+		params->rx_cq_moderation = mlx5e_get_def_rx_moderation(cq_period_mode);
+	}
+
+	MLX5E_SET_PFLAG(params, MLX5E_PFLAG_RX_CQE_BASED_MODER,
+			params->rx_cq_moderation.cq_period_mode ==
+				MLX5_CQ_PERIOD_MODE_START_FROM_CQE);
+}
+
+static u32 mlx5e_choose_lro_timeout(struct mlx5_core_dev *mdev, u32 wanted_timeout)
+{
+	int i;
+
+	/* The supported periods are organized in ascending order */
+	for (i = 0; i < MLX5E_LRO_TIMEOUT_ARR_SIZE - 1; i++)
+		if (MLX5_CAP_ETH(mdev, lro_timer_supported_periods[i]) >= wanted_timeout)
+			break;
+
+	return MLX5_CAP_ETH(mdev, lro_timer_supported_periods[i]);
+}
+
+void mlx5e_build_nic_params(struct mlx5_core_dev *mdev,
+			    struct mlx5e_params *params,
+			    u16 max_channels, u16 mtu)
+{
+	u8 rx_cq_period_mode;
+
+	params->sw_mtu = mtu;
+	params->hard_mtu = MLX5E_ETH_HARD_MTU;
+	params->num_channels = max_channels;
+	params->num_tc       = 1;
+
+	/* SQ */
+	params->log_sq_size = is_kdump_kernel() ?
+		MLX5E_PARAMS_MINIMUM_LOG_SQ_SIZE :
+		MLX5E_PARAMS_DEFAULT_LOG_SQ_SIZE;
+
+	/* set CQE compression */
+	params->rx_cqe_compress_def = false;
+	if (MLX5_CAP_GEN(mdev, cqe_compression) &&
+	    MLX5_CAP_GEN(mdev, vport_group_manager))
+		params->rx_cqe_compress_def = slow_pci_heuristic(mdev);
+
+	MLX5E_SET_PFLAG(params, MLX5E_PFLAG_RX_CQE_COMPRESS, params->rx_cqe_compress_def);
+	MLX5E_SET_PFLAG(params, MLX5E_PFLAG_RX_NO_CSUM_COMPLETE, false);
+
+	/* RQ */
+	/* Prefer Striding RQ, unless any of the following holds:
+	 * - Striding RQ configuration is not possible/supported.
+	 * - Slow PCI heuristic.
+	 * - Legacy RQ would use linear SKB while Striding RQ would use non-linear.
+	 */
+	if (!slow_pci_heuristic(mdev) &&
+	    mlx5e_striding_rq_possible(mdev, params) &&
+	    (mlx5e_rx_mpwqe_is_linear_skb(mdev, params) ||
+	     !mlx5e_rx_is_linear_skb(mdev, params)))
+		MLX5E_SET_PFLAG(params, MLX5E_PFLAG_RX_STRIDING_RQ, true);
+	mlx5e_set_rq_type(mdev, params);
+	mlx5e_init_rq_type_params(mdev, params);
+
+	/* HW LRO */
+
+	/* TODO: && MLX5_CAP_ETH(mdev, lro_cap) */
+	if (params->rq_wq_type == MLX5_WQ_TYPE_LINKED_LIST_STRIDING_RQ)
+		if (!mlx5e_rx_mpwqe_is_linear_skb(mdev, params))
+			params->lro_en = !slow_pci_heuristic(mdev);
+	params->lro_timeout = mlx5e_choose_lro_timeout(mdev, MLX5E_DEFAULT_LRO_TIMEOUT);
+
+	/* CQ moderation params */
+	rx_cq_period_mode = MLX5_CAP_GEN(mdev, cq_period_start_from_cqe) ?
+			MLX5_CQ_PERIOD_MODE_START_FROM_CQE :
+			MLX5_CQ_PERIOD_MODE_START_FROM_EQE;
+	params->rx_dim_enabled = MLX5_CAP_GEN(mdev, cq_moderation);
+	params->tx_dim_enabled = MLX5_CAP_GEN(mdev, cq_moderation);
+	mlx5e_set_rx_cq_mode_params(params, rx_cq_period_mode);
+	mlx5e_set_tx_cq_mode_params(params, MLX5_CQ_PERIOD_MODE_START_FROM_EQE);
+
+	/* TX inline */
+	params->tx_min_inline_mode = mlx5e_params_calculate_tx_min_inline(mdev);
+
+	/* RSS */
+	params->rss_hfunc = ETH_RSS_HASH_XOR;
+	netdev_rss_key_fill(params->toeplitz_hash_key, sizeof(params->toeplitz_hash_key));
+	mlx5e_build_default_indir_rqt(params->indirection_rqt,
+				      MLX5E_INDIR_RQT_SIZE, max_channels);
+}
+
+static void mlx5e_build_nic_netdev_priv(struct mlx5_core_dev *mdev,
+					struct net_device *netdev,
+					const struct mlx5e_profile *profile,
+					void *ppriv)
+{
+	struct mlx5e_priv *priv = netdev_priv(netdev);
+
+	priv->mdev        = mdev;
+	priv->netdev      = netdev;
+	priv->profile     = profile;
+	priv->ppriv       = ppriv;
+	priv->msglevel    = MLX5E_MSG_LEVEL;
+	priv->max_opened_tc = 1;
+
+	mlx5e_build_nic_params(mdev, &priv->channels.params,
+			       profile->max_nch(mdev), netdev->mtu);
+
+	mutex_init(&priv->state_lock);
+
+	INIT_WORK(&priv->update_carrier_work, mlx5e_update_carrier_work);
+	INIT_WORK(&priv->set_rx_mode_work, mlx5e_set_rx_mode_work);
+	INIT_WORK(&priv->tx_timeout_work, mlx5e_tx_timeout_work);
+	INIT_DELAYED_WORK(&priv->update_stats_work, mlx5e_update_stats_work);
+
+	mlx5e_timestamp_init(priv);
+}
+
+static void mlx5e_set_netdev_dev_addr(struct net_device *netdev)
+{
+	struct mlx5e_priv *priv = netdev_priv(netdev);
+
+	mlx5_query_nic_vport_mac_address(priv->mdev, 0, netdev->dev_addr);
+	if (is_zero_ether_addr(netdev->dev_addr) &&
+	    !MLX5_CAP_GEN(priv->mdev, vport_group_manager)) {
+		eth_hw_addr_random(netdev);
+		mlx5_core_info(priv->mdev, "Assigned random MAC address %pM\n", netdev->dev_addr);
+	}
+}
+
+#if IS_ENABLED(CONFIG_MLX5_ESWITCH)
+static const struct switchdev_ops mlx5e_switchdev_ops = {
+	.switchdev_port_attr_get	= mlx5e_attr_get,
+};
+#endif
+
+static void mlx5e_build_nic_netdev(struct net_device *netdev)
+{
+	struct mlx5e_priv *priv = netdev_priv(netdev);
+	struct mlx5_core_dev *mdev = priv->mdev;
+	bool fcs_supported;
+	bool fcs_enabled;
+
+	SET_NETDEV_DEV(netdev, &mdev->pdev->dev);
+
+	netdev->netdev_ops = &mlx5e_netdev_ops;
+
+#ifdef CONFIG_MLX5_CORE_EN_DCB
+	if (MLX5_CAP_GEN(mdev, vport_group_manager) && MLX5_CAP_GEN(mdev, qos))
+		netdev->dcbnl_ops = &mlx5e_dcbnl_ops;
+#endif
+
+	netdev->watchdog_timeo    = 15 * HZ;
+
+	netdev->ethtool_ops	  = &mlx5e_ethtool_ops;
+
+	netdev->vlan_features    |= NETIF_F_SG;
+	netdev->vlan_features    |= NETIF_F_IP_CSUM;
+	netdev->vlan_features    |= NETIF_F_IPV6_CSUM;
+	netdev->vlan_features    |= NETIF_F_GRO;
+	netdev->vlan_features    |= NETIF_F_TSO;
+	netdev->vlan_features    |= NETIF_F_TSO6;
+	netdev->vlan_features    |= NETIF_F_RXCSUM;
+	netdev->vlan_features    |= NETIF_F_RXHASH;
+
+	netdev->hw_enc_features  |= NETIF_F_HW_VLAN_CTAG_TX;
+	netdev->hw_enc_features  |= NETIF_F_HW_VLAN_CTAG_RX;
+
+	if (!!MLX5_CAP_ETH(mdev, lro_cap) &&
+	    mlx5e_check_fragmented_striding_rq_cap(mdev))
+		netdev->vlan_features    |= NETIF_F_LRO;
+
+	netdev->hw_features       = netdev->vlan_features;
+	netdev->hw_features      |= NETIF_F_HW_VLAN_CTAG_TX;
+	netdev->hw_features      |= NETIF_F_HW_VLAN_CTAG_RX;
+	netdev->hw_features      |= NETIF_F_HW_VLAN_CTAG_FILTER;
+	netdev->hw_features      |= NETIF_F_HW_VLAN_STAG_TX;
+
+	if (mlx5_vxlan_allowed(mdev->vxlan) || MLX5_CAP_ETH(mdev, tunnel_stateless_gre)) {
+		netdev->hw_enc_features |= NETIF_F_IP_CSUM;
+		netdev->hw_enc_features |= NETIF_F_IPV6_CSUM;
+		netdev->hw_enc_features |= NETIF_F_TSO;
+		netdev->hw_enc_features |= NETIF_F_TSO6;
+		netdev->hw_enc_features |= NETIF_F_GSO_PARTIAL;
+	}
+
+	if (mlx5_vxlan_allowed(mdev->vxlan)) {
+		netdev->hw_features     |= NETIF_F_GSO_UDP_TUNNEL;
+		netdev->hw_enc_features |= NETIF_F_GSO_UDP_TUNNEL;
+	}
+
+	if (MLX5_CAP_ETH(mdev, tunnel_stateless_gre)) {
+		netdev->hw_features     |= NETIF_F_GSO_GRE |
+					   NETIF_F_GSO_GRE_CSUM;
+		netdev->hw_enc_features |= NETIF_F_GSO_GRE |
+					   NETIF_F_GSO_GRE_CSUM;
+		netdev->gso_partial_features |= NETIF_F_GSO_GRE |
+						NETIF_F_GSO_GRE_CSUM;
+	}
+
+	netdev->hw_features	                 |= NETIF_F_GSO_PARTIAL;
+	netdev->gso_partial_features             |= NETIF_F_GSO_UDP_L4;
+	netdev->hw_features                      |= NETIF_F_GSO_UDP_L4;
+	netdev->features                         |= NETIF_F_GSO_UDP_L4;
+
+	mlx5_query_port_fcs(mdev, &fcs_supported, &fcs_enabled);
+
+	if (fcs_supported)
+		netdev->hw_features |= NETIF_F_RXALL;
+
+	if (MLX5_CAP_ETH(mdev, scatter_fcs))
+		netdev->hw_features |= NETIF_F_RXFCS;
+
+	netdev->features          = netdev->hw_features;
+	if (!priv->channels.params.lro_en)
+		netdev->features  &= ~NETIF_F_LRO;
+
+	if (fcs_enabled)
+		netdev->features  &= ~NETIF_F_RXALL;
+
+	if (!priv->channels.params.scatter_fcs_en)
+		netdev->features  &= ~NETIF_F_RXFCS;
+
+	/* prefere CQE compression over rxhash */
+	if (MLX5E_GET_PFLAG(&priv->channels.params, MLX5E_PFLAG_RX_CQE_COMPRESS))
+		netdev->features &= ~NETIF_F_RXHASH;
+
+#define FT_CAP(f) MLX5_CAP_FLOWTABLE(mdev, flow_table_properties_nic_receive.f)
+	if (FT_CAP(flow_modify_en) &&
+	    FT_CAP(modify_root) &&
+	    FT_CAP(identified_miss_table_mode) &&
+	    FT_CAP(flow_table_modify)) {
+#ifdef CONFIG_MLX5_ESWITCH
+		netdev->hw_features      |= NETIF_F_HW_TC;
+#endif
+#ifdef CONFIG_MLX5_EN_ARFS
+		netdev->hw_features	 |= NETIF_F_NTUPLE;
+#endif
+	}
+
+	netdev->features         |= NETIF_F_HIGHDMA;
+	netdev->features         |= NETIF_F_HW_VLAN_STAG_FILTER;
+
+	netdev->priv_flags       |= IFF_UNICAST_FLT;
+
+	mlx5e_set_netdev_dev_addr(netdev);
+
+#if IS_ENABLED(CONFIG_MLX5_ESWITCH)
+	if (MLX5_ESWITCH_MANAGER(mdev))
+		netdev->switchdev_ops = &mlx5e_switchdev_ops;
+#endif
+
+	mlx5e_ipsec_build_netdev(priv);
+	mlx5e_tls_build_netdev(priv);
+}
+
+static void mlx5e_create_q_counters(struct mlx5e_priv *priv)
+{
+	struct mlx5_core_dev *mdev = priv->mdev;
+	int err;
+
+	err = mlx5_core_alloc_q_counter(mdev, &priv->q_counter);
+	if (err) {
+		mlx5_core_warn(mdev, "alloc queue counter failed, %d\n", err);
+		priv->q_counter = 0;
+	}
+
+	err = mlx5_core_alloc_q_counter(mdev, &priv->drop_rq_q_counter);
+	if (err) {
+		mlx5_core_warn(mdev, "alloc drop RQ counter failed, %d\n", err);
+		priv->drop_rq_q_counter = 0;
+	}
+}
+
+static void mlx5e_destroy_q_counters(struct mlx5e_priv *priv)
+{
+	if (priv->q_counter)
+		mlx5_core_dealloc_q_counter(priv->mdev, priv->q_counter);
+
+	if (priv->drop_rq_q_counter)
+		mlx5_core_dealloc_q_counter(priv->mdev, priv->drop_rq_q_counter);
+}
+
+static void mlx5e_nic_init(struct mlx5_core_dev *mdev,
+			   struct net_device *netdev,
+			   const struct mlx5e_profile *profile,
+			   void *ppriv)
+{
+	struct mlx5e_priv *priv = netdev_priv(netdev);
+	int err;
+
+	mlx5e_build_nic_netdev_priv(mdev, netdev, profile, ppriv);
+	err = mlx5e_ipsec_init(priv);
+	if (err)
+		mlx5_core_err(mdev, "IPSec initialization failed, %d\n", err);
+	err = mlx5e_tls_init(priv);
+	if (err)
+		mlx5_core_err(mdev, "TLS initialization failed, %d\n", err);
+	mlx5e_build_nic_netdev(netdev);
+	mlx5e_build_tc2txq_maps(priv);
+}
+
+static void mlx5e_nic_cleanup(struct mlx5e_priv *priv)
+{
+	mlx5e_tls_cleanup(priv);
+	mlx5e_ipsec_cleanup(priv);
+}
+
+static int mlx5e_init_nic_rx(struct mlx5e_priv *priv)
+{
+	struct mlx5_core_dev *mdev = priv->mdev;
+	int err;
+
+	err = mlx5e_create_indirect_rqt(priv);
+	if (err)
+		return err;
+
+	err = mlx5e_create_direct_rqts(priv);
+	if (err)
+		goto err_destroy_indirect_rqts;
+
+	err = mlx5e_create_indirect_tirs(priv);
+	if (err)
+		goto err_destroy_direct_rqts;
+
+	err = mlx5e_create_direct_tirs(priv);
+	if (err)
+		goto err_destroy_indirect_tirs;
+
+	err = mlx5e_create_flow_steering(priv);
+	if (err) {
+		mlx5_core_warn(mdev, "create flow steering failed, %d\n", err);
+		goto err_destroy_direct_tirs;
+	}
+
+	err = mlx5e_tc_nic_init(priv);
+	if (err)
+		goto err_destroy_flow_steering;
+
+	return 0;
+
+err_destroy_flow_steering:
+	mlx5e_destroy_flow_steering(priv);
+err_destroy_direct_tirs:
+	mlx5e_destroy_direct_tirs(priv);
+err_destroy_indirect_tirs:
+	mlx5e_destroy_indirect_tirs(priv);
+err_destroy_direct_rqts:
+	mlx5e_destroy_direct_rqts(priv);
+err_destroy_indirect_rqts:
+	mlx5e_destroy_rqt(priv, &priv->indir_rqt);
+	return err;
+}
+
+static void mlx5e_cleanup_nic_rx(struct mlx5e_priv *priv)
+{
+	mlx5e_tc_nic_cleanup(priv);
+	mlx5e_destroy_flow_steering(priv);
+	mlx5e_destroy_direct_tirs(priv);
+	mlx5e_destroy_indirect_tirs(priv);
+	mlx5e_destroy_direct_rqts(priv);
+	mlx5e_destroy_rqt(priv, &priv->indir_rqt);
+}
+
+static int mlx5e_init_nic_tx(struct mlx5e_priv *priv)
+{
+	int err;
+
+	err = mlx5e_create_tises(priv);
+	if (err) {
+		mlx5_core_warn(priv->mdev, "create tises failed, %d\n", err);
+		return err;
+	}
+
+#ifdef CONFIG_MLX5_CORE_EN_DCB
+	mlx5e_dcbnl_initialize(priv);
+#endif
+	return 0;
+}
+
+static void mlx5e_nic_enable(struct mlx5e_priv *priv)
+{
+	struct net_device *netdev = priv->netdev;
+	struct mlx5_core_dev *mdev = priv->mdev;
+	u16 max_mtu;
+
+	mlx5e_init_l2_addr(priv);
+
+	/* Marking the link as currently not needed by the Driver */
+	if (!netif_running(netdev))
+		mlx5_set_port_admin_status(mdev, MLX5_PORT_DOWN);
+
+	/* MTU range: 68 - hw-specific max */
+	netdev->min_mtu = ETH_MIN_MTU;
+	mlx5_query_port_max_mtu(priv->mdev, &max_mtu, 1);
+	netdev->max_mtu = MLX5E_HW2SW_MTU(&priv->channels.params, max_mtu);
+	mlx5e_set_dev_port_mtu(priv);
+
+	mlx5_lag_add(mdev, netdev);
+
+	mlx5e_enable_async_events(priv);
+
+	if (MLX5_ESWITCH_MANAGER(priv->mdev))
+		mlx5e_register_vport_reps(priv);
+
+	if (netdev->reg_state != NETREG_REGISTERED)
+		return;
+#ifdef CONFIG_MLX5_CORE_EN_DCB
+	mlx5e_dcbnl_init_app(priv);
+#endif
+
+	queue_work(priv->wq, &priv->set_rx_mode_work);
+
+	rtnl_lock();
+	if (netif_running(netdev))
+		mlx5e_open(netdev);
+	netif_device_attach(netdev);
+	rtnl_unlock();
+}
+
+static void mlx5e_nic_disable(struct mlx5e_priv *priv)
+{
+	struct mlx5_core_dev *mdev = priv->mdev;
+
+#ifdef CONFIG_MLX5_CORE_EN_DCB
+	if (priv->netdev->reg_state == NETREG_REGISTERED)
+		mlx5e_dcbnl_delete_app(priv);
+#endif
+
+	rtnl_lock();
+	if (netif_running(priv->netdev))
+		mlx5e_close(priv->netdev);
+	netif_device_detach(priv->netdev);
+	rtnl_unlock();
+
+	queue_work(priv->wq, &priv->set_rx_mode_work);
+
+	if (MLX5_ESWITCH_MANAGER(priv->mdev))
+		mlx5e_unregister_vport_reps(priv);
+
+	mlx5e_disable_async_events(priv);
+	mlx5_lag_remove(mdev);
+}
+
+static const struct mlx5e_profile mlx5e_nic_profile = {
+	.init		   = mlx5e_nic_init,
+	.cleanup	   = mlx5e_nic_cleanup,
+	.init_rx	   = mlx5e_init_nic_rx,
+	.cleanup_rx	   = mlx5e_cleanup_nic_rx,
+	.init_tx	   = mlx5e_init_nic_tx,
+	.cleanup_tx	   = mlx5e_cleanup_nic_tx,
+	.enable		   = mlx5e_nic_enable,
+	.disable	   = mlx5e_nic_disable,
+	.update_stats	   = mlx5e_update_ndo_stats,
+	.max_nch	   = mlx5e_get_max_num_channels,
+	.update_carrier	   = mlx5e_update_carrier,
+	.rx_handlers.handle_rx_cqe       = mlx5e_handle_rx_cqe,
+	.rx_handlers.handle_rx_cqe_mpwqe = mlx5e_handle_rx_cqe_mpwrq,
+	.max_tc		   = MLX5E_MAX_NUM_TC,
+};
+
+/* mlx5e generic netdev management API (move to en_common.c) */
+
+struct net_device *mlx5e_create_netdev(struct mlx5_core_dev *mdev,
+				       const struct mlx5e_profile *profile,
+				       void *ppriv)
+{
+	int nch = profile->max_nch(mdev);
+	struct net_device *netdev;
+	struct mlx5e_priv *priv;
+
+	netdev = alloc_etherdev_mqs(sizeof(struct mlx5e_priv),
+				    nch * profile->max_tc,
+				    nch);
+	if (!netdev) {
+		mlx5_core_err(mdev, "alloc_etherdev_mqs() failed\n");
+		return NULL;
+	}
+
+#ifdef CONFIG_MLX5_EN_ARFS
+	netdev->rx_cpu_rmap = mdev->rmap;
+#endif
+
+	profile->init(mdev, netdev, profile, ppriv);
+
+	netif_carrier_off(netdev);
+
+	priv = netdev_priv(netdev);
+
+	priv->wq = create_singlethread_workqueue("mlx5e");
+	if (!priv->wq)
+		goto err_cleanup_nic;
+
+	return netdev;
+
+err_cleanup_nic:
+	if (profile->cleanup)
+		profile->cleanup(priv);
+	free_netdev(netdev);
+
+	return NULL;
+}
+
+int mlx5e_attach_netdev(struct mlx5e_priv *priv)
+{
+	struct mlx5_core_dev *mdev = priv->mdev;
+	const struct mlx5e_profile *profile;
+	int max_nch;
+	int err;
+
+	profile = priv->profile;
+	clear_bit(MLX5E_STATE_DESTROYING, &priv->state);
+
+	/* max number of channels may have changed */
+	max_nch = mlx5e_get_max_num_channels(priv->mdev);
+	if (priv->channels.params.num_channels > max_nch) {
+		mlx5_core_warn(priv->mdev, "MLX5E: Reducing number of channels to %d\n", max_nch);
+		priv->channels.params.num_channels = max_nch;
+		mlx5e_build_default_indir_rqt(priv->channels.params.indirection_rqt,
+					      MLX5E_INDIR_RQT_SIZE, max_nch);
+	}
+
+	err = profile->init_tx(priv);
+	if (err)
+		goto out;
+
+	mlx5e_create_q_counters(priv);
+
+	err = mlx5e_open_drop_rq(priv, &priv->drop_rq);
+	if (err) {
+		mlx5_core_err(mdev, "open drop rq failed, %d\n", err);
+		goto err_destroy_q_counters;
+	}
+
+	err = profile->init_rx(priv);
+	if (err)
+		goto err_close_drop_rq;
+
+	if (profile->enable)
+		profile->enable(priv);
+
+	return 0;
+
+err_close_drop_rq:
+	mlx5e_close_drop_rq(&priv->drop_rq);
+
+err_destroy_q_counters:
+	mlx5e_destroy_q_counters(priv);
+	profile->cleanup_tx(priv);
+
+out:
+	return err;
+}
+
+void mlx5e_detach_netdev(struct mlx5e_priv *priv)
+{
+	const struct mlx5e_profile *profile = priv->profile;
+
+	set_bit(MLX5E_STATE_DESTROYING, &priv->state);
+
+	if (profile->disable)
+		profile->disable(priv);
+	flush_workqueue(priv->wq);
+
+	profile->cleanup_rx(priv);
+	mlx5e_close_drop_rq(&priv->drop_rq);
+	mlx5e_destroy_q_counters(priv);
+	profile->cleanup_tx(priv);
+	cancel_delayed_work_sync(&priv->update_stats_work);
+}
+
+void mlx5e_destroy_netdev(struct mlx5e_priv *priv)
+{
+	const struct mlx5e_profile *profile = priv->profile;
+	struct net_device *netdev = priv->netdev;
+
+	destroy_workqueue(priv->wq);
+	if (profile->cleanup)
+		profile->cleanup(priv);
+	free_netdev(netdev);
+}
+
+/* mlx5e_attach and mlx5e_detach scope should be only creating/destroying
+ * hardware contexts and to connect it to the current netdev.
+ */
+static int mlx5e_attach(struct mlx5_core_dev *mdev, void *vpriv)
+{
+	struct mlx5e_priv *priv = vpriv;
+	struct net_device *netdev = priv->netdev;
+	int err;
+
+	if (netif_device_present(netdev))
+		return 0;
+
+	err = mlx5e_create_mdev_resources(mdev);
+	if (err)
+		return err;
+
+	err = mlx5e_attach_netdev(priv);
+	if (err) {
+		mlx5e_destroy_mdev_resources(mdev);
+		return err;
+	}
+
+	return 0;
+}
+
+static void mlx5e_detach(struct mlx5_core_dev *mdev, void *vpriv)
+{
+	struct mlx5e_priv *priv = vpriv;
+	struct net_device *netdev = priv->netdev;
+
+	if (!netif_device_present(netdev))
+		return;
+
+	mlx5e_detach_netdev(priv);
+	mlx5e_destroy_mdev_resources(mdev);
+}
+
+static void *mlx5e_add(struct mlx5_core_dev *mdev)
+{
+	struct net_device *netdev;
+	void *rpriv = NULL;
+	void *priv;
+	int err;
+
+	err = mlx5e_check_required_hca_cap(mdev);
+	if (err)
+		return NULL;
+
+#ifdef CONFIG_MLX5_ESWITCH
+	if (MLX5_ESWITCH_MANAGER(mdev)) {
+		rpriv = mlx5e_alloc_nic_rep_priv(mdev);
+		if (!rpriv) {
+			mlx5_core_warn(mdev, "Failed to alloc NIC rep priv data\n");
+			return NULL;
+		}
+	}
+#endif
+
+	netdev = mlx5e_create_netdev(mdev, &mlx5e_nic_profile, rpriv);
+	if (!netdev) {
+		mlx5_core_err(mdev, "mlx5e_create_netdev failed\n");
+		goto err_free_rpriv;
+	}
+
+	priv = netdev_priv(netdev);
+
+	err = mlx5e_attach(mdev, priv);
+	if (err) {
+		mlx5_core_err(mdev, "mlx5e_attach failed, %d\n", err);
+		goto err_destroy_netdev;
+	}
+
+	err = register_netdev(netdev);
+	if (err) {
+		mlx5_core_err(mdev, "register_netdev failed, %d\n", err);
+		goto err_detach;
+	}
+
+#ifdef CONFIG_MLX5_CORE_EN_DCB
+	mlx5e_dcbnl_init_app(priv);
+#endif
+	return priv;
+
+err_detach:
+	mlx5e_detach(mdev, priv);
+err_destroy_netdev:
+	mlx5e_destroy_netdev(priv);
+err_free_rpriv:
+	kfree(rpriv);
+	return NULL;
+}
+
+static void mlx5e_remove(struct mlx5_core_dev *mdev, void *vpriv)
+{
+	struct mlx5e_priv *priv = vpriv;
+	void *ppriv = priv->ppriv;
+
+#ifdef CONFIG_MLX5_CORE_EN_DCB
+	mlx5e_dcbnl_delete_app(priv);
+#endif
+	unregister_netdev(priv->netdev);
+	mlx5e_detach(mdev, vpriv);
+	mlx5e_destroy_netdev(priv);
+	kfree(ppriv);
+}
+
+static void *mlx5e_get_netdev(void *vpriv)
+{
+	struct mlx5e_priv *priv = vpriv;
+
+	return priv->netdev;
+}
+
+static struct mlx5_interface mlx5e_interface = {
+	.add       = mlx5e_add,
+	.remove    = mlx5e_remove,
+	.attach    = mlx5e_attach,
+	.detach    = mlx5e_detach,
+	.event     = mlx5e_async_event,
+	.protocol  = MLX5_INTERFACE_PROTOCOL_ETH,
+	.get_dev   = mlx5e_get_netdev,
+};
+
+void mlx5e_init(void)
+{
+	mlx5e_ipsec_build_inverse_table();
+	mlx5e_build_ptys2ethtool_map();
+	mlx5_register_interface(&mlx5e_interface);
+}
+
+void mlx5e_cleanup(void)
+{
+	mlx5_unregister_interface(&mlx5e_interface);
+}
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_rep.c b/drivers/net/ethernet/mellanox/mlx5/core/en_rep.c
new file mode 100644
index 000000000..1ab40d622
--- /dev/null
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_rep.c
@@ -0,0 +1,1295 @@
+/*
+ * Copyright (c) 2016, 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 <generated/utsrelease.h>
+#include <linux/mlx5/fs.h>
+#include <net/switchdev.h>
+#include <net/pkt_cls.h>
+#include <net/act_api.h>
+#include <net/netevent.h>
+#include <net/arp.h>
+
+#include "eswitch.h"
+#include "en.h"
+#include "en_rep.h"
+#include "en_tc.h"
+#include "fs_core.h"
+
+#define MLX5E_REP_PARAMS_LOG_SQ_SIZE \
+	max(0x6, MLX5E_PARAMS_MINIMUM_LOG_SQ_SIZE)
+#define MLX5E_REP_PARAMS_LOG_RQ_SIZE \
+	max(0x6, MLX5E_PARAMS_MINIMUM_LOG_RQ_SIZE)
+
+static const char mlx5e_rep_driver_name[] = "mlx5e_rep";
+
+static void mlx5e_rep_get_drvinfo(struct net_device *dev,
+				  struct ethtool_drvinfo *drvinfo)
+{
+	strlcpy(drvinfo->driver, mlx5e_rep_driver_name,
+		sizeof(drvinfo->driver));
+	strlcpy(drvinfo->version, UTS_RELEASE, sizeof(drvinfo->version));
+}
+
+static const struct counter_desc sw_rep_stats_desc[] = {
+	{ MLX5E_DECLARE_STAT(struct mlx5e_sw_stats, rx_packets) },
+	{ MLX5E_DECLARE_STAT(struct mlx5e_sw_stats, rx_bytes) },
+	{ MLX5E_DECLARE_STAT(struct mlx5e_sw_stats, tx_packets) },
+	{ MLX5E_DECLARE_STAT(struct mlx5e_sw_stats, tx_bytes) },
+};
+
+struct vport_stats {
+	u64 vport_rx_packets;
+	u64 vport_tx_packets;
+	u64 vport_rx_bytes;
+	u64 vport_tx_bytes;
+};
+
+static const struct counter_desc vport_rep_stats_desc[] = {
+	{ MLX5E_DECLARE_STAT(struct vport_stats, vport_rx_packets) },
+	{ MLX5E_DECLARE_STAT(struct vport_stats, vport_rx_bytes) },
+	{ MLX5E_DECLARE_STAT(struct vport_stats, vport_tx_packets) },
+	{ MLX5E_DECLARE_STAT(struct vport_stats, vport_tx_bytes) },
+};
+
+#define NUM_VPORT_REP_SW_COUNTERS ARRAY_SIZE(sw_rep_stats_desc)
+#define NUM_VPORT_REP_HW_COUNTERS ARRAY_SIZE(vport_rep_stats_desc)
+
+static void mlx5e_rep_get_strings(struct net_device *dev,
+				  u32 stringset, uint8_t *data)
+{
+	int i, j;
+
+	switch (stringset) {
+	case ETH_SS_STATS:
+		for (i = 0; i < NUM_VPORT_REP_SW_COUNTERS; i++)
+			strcpy(data + (i * ETH_GSTRING_LEN),
+			       sw_rep_stats_desc[i].format);
+		for (j = 0; j < NUM_VPORT_REP_HW_COUNTERS; j++, i++)
+			strcpy(data + (i * ETH_GSTRING_LEN),
+			       vport_rep_stats_desc[j].format);
+		break;
+	}
+}
+
+static void mlx5e_rep_update_hw_counters(struct mlx5e_priv *priv)
+{
+	struct mlx5_eswitch *esw = priv->mdev->priv.eswitch;
+	struct mlx5e_rep_priv *rpriv = priv->ppriv;
+	struct mlx5_eswitch_rep *rep = rpriv->rep;
+	struct rtnl_link_stats64 *vport_stats;
+	struct ifla_vf_stats vf_stats;
+	int err;
+
+	err = mlx5_eswitch_get_vport_stats(esw, rep->vport, &vf_stats);
+	if (err) {
+		pr_warn("vport %d error %d reading stats\n", rep->vport, err);
+		return;
+	}
+
+	vport_stats = &priv->stats.vf_vport;
+	/* flip tx/rx as we are reporting the counters for the switch vport */
+	vport_stats->rx_packets = vf_stats.tx_packets;
+	vport_stats->rx_bytes   = vf_stats.tx_bytes;
+	vport_stats->tx_packets = vf_stats.rx_packets;
+	vport_stats->tx_bytes   = vf_stats.rx_bytes;
+}
+
+static void mlx5e_rep_update_sw_counters(struct mlx5e_priv *priv)
+{
+	struct mlx5e_sw_stats *s = &priv->stats.sw;
+	struct mlx5e_rq_stats *rq_stats;
+	struct mlx5e_sq_stats *sq_stats;
+	int i, j;
+
+	memset(s, 0, sizeof(*s));
+	for (i = 0; i < priv->channels.num; i++) {
+		struct mlx5e_channel *c = priv->channels.c[i];
+
+		rq_stats = c->rq.stats;
+
+		s->rx_packets	+= rq_stats->packets;
+		s->rx_bytes	+= rq_stats->bytes;
+
+		for (j = 0; j < priv->channels.params.num_tc; j++) {
+			sq_stats = c->sq[j].stats;
+
+			s->tx_packets		+= sq_stats->packets;
+			s->tx_bytes		+= sq_stats->bytes;
+			s->tx_queue_dropped	+= sq_stats->dropped;
+		}
+	}
+}
+
+static void mlx5e_rep_get_ethtool_stats(struct net_device *dev,
+					struct ethtool_stats *stats, u64 *data)
+{
+	struct mlx5e_priv *priv = netdev_priv(dev);
+	int i, j;
+
+	if (!data)
+		return;
+
+	mutex_lock(&priv->state_lock);
+	if (test_bit(MLX5E_STATE_OPENED, &priv->state))
+		mlx5e_rep_update_sw_counters(priv);
+	mlx5e_rep_update_hw_counters(priv);
+	mutex_unlock(&priv->state_lock);
+
+	for (i = 0; i < NUM_VPORT_REP_SW_COUNTERS; i++)
+		data[i] = MLX5E_READ_CTR64_CPU(&priv->stats.sw,
+					       sw_rep_stats_desc, i);
+
+	for (j = 0; j < NUM_VPORT_REP_HW_COUNTERS; j++, i++)
+		data[i] = MLX5E_READ_CTR64_CPU(&priv->stats.vf_vport,
+					       vport_rep_stats_desc, j);
+}
+
+static int mlx5e_rep_get_sset_count(struct net_device *dev, int sset)
+{
+	switch (sset) {
+	case ETH_SS_STATS:
+		return NUM_VPORT_REP_SW_COUNTERS + NUM_VPORT_REP_HW_COUNTERS;
+	default:
+		return -EOPNOTSUPP;
+	}
+}
+
+static const struct ethtool_ops mlx5e_rep_ethtool_ops = {
+	.get_drvinfo	   = mlx5e_rep_get_drvinfo,
+	.get_link	   = ethtool_op_get_link,
+	.get_strings       = mlx5e_rep_get_strings,
+	.get_sset_count    = mlx5e_rep_get_sset_count,
+	.get_ethtool_stats = mlx5e_rep_get_ethtool_stats,
+};
+
+int mlx5e_attr_get(struct net_device *dev, struct switchdev_attr *attr)
+{
+	struct mlx5e_priv *priv = netdev_priv(dev);
+	struct mlx5e_rep_priv *rpriv = priv->ppriv;
+	struct mlx5_eswitch_rep *rep = rpriv->rep;
+	struct mlx5_eswitch *esw = priv->mdev->priv.eswitch;
+
+	if (esw->mode != SRIOV_OFFLOADS)
+		return -EOPNOTSUPP;
+
+	switch (attr->id) {
+	case SWITCHDEV_ATTR_ID_PORT_PARENT_ID:
+		attr->u.ppid.id_len = ETH_ALEN;
+		ether_addr_copy(attr->u.ppid.id, rep->hw_id);
+		break;
+	default:
+		return -EOPNOTSUPP;
+	}
+
+	return 0;
+}
+
+static void mlx5e_sqs2vport_stop(struct mlx5_eswitch *esw,
+				 struct mlx5_eswitch_rep *rep)
+{
+	struct mlx5e_rep_sq *rep_sq, *tmp;
+	struct mlx5e_rep_priv *rpriv;
+
+	if (esw->mode != SRIOV_OFFLOADS)
+		return;
+
+	rpriv = mlx5e_rep_to_rep_priv(rep);
+	list_for_each_entry_safe(rep_sq, tmp, &rpriv->vport_sqs_list, list) {
+		mlx5_eswitch_del_send_to_vport_rule(rep_sq->send_to_vport_rule);
+		list_del(&rep_sq->list);
+		kfree(rep_sq);
+	}
+}
+
+static int mlx5e_sqs2vport_start(struct mlx5_eswitch *esw,
+				 struct mlx5_eswitch_rep *rep,
+				 u32 *sqns_array, int sqns_num)
+{
+	struct mlx5_flow_handle *flow_rule;
+	struct mlx5e_rep_priv *rpriv;
+	struct mlx5e_rep_sq *rep_sq;
+	int err;
+	int i;
+
+	if (esw->mode != SRIOV_OFFLOADS)
+		return 0;
+
+	rpriv = mlx5e_rep_to_rep_priv(rep);
+	for (i = 0; i < sqns_num; i++) {
+		rep_sq = kzalloc(sizeof(*rep_sq), GFP_KERNEL);
+		if (!rep_sq) {
+			err = -ENOMEM;
+			goto out_err;
+		}
+
+		/* Add re-inject rule to the PF/representor sqs */
+		flow_rule = mlx5_eswitch_add_send_to_vport_rule(esw,
+								rep->vport,
+								sqns_array[i]);
+		if (IS_ERR(flow_rule)) {
+			err = PTR_ERR(flow_rule);
+			kfree(rep_sq);
+			goto out_err;
+		}
+		rep_sq->send_to_vport_rule = flow_rule;
+		list_add(&rep_sq->list, &rpriv->vport_sqs_list);
+	}
+	return 0;
+
+out_err:
+	mlx5e_sqs2vport_stop(esw, rep);
+	return err;
+}
+
+int mlx5e_add_sqs_fwd_rules(struct mlx5e_priv *priv)
+{
+	struct mlx5_eswitch *esw = priv->mdev->priv.eswitch;
+	struct mlx5e_rep_priv *rpriv = priv->ppriv;
+	struct mlx5_eswitch_rep *rep = rpriv->rep;
+	struct mlx5e_channel *c;
+	int n, tc, num_sqs = 0;
+	int err = -ENOMEM;
+	u32 *sqs;
+
+	sqs = kcalloc(priv->channels.num * priv->channels.params.num_tc, sizeof(*sqs), GFP_KERNEL);
+	if (!sqs)
+		goto out;
+
+	for (n = 0; n < priv->channels.num; n++) {
+		c = priv->channels.c[n];
+		for (tc = 0; tc < c->num_tc; tc++)
+			sqs[num_sqs++] = c->sq[tc].sqn;
+	}
+
+	err = mlx5e_sqs2vport_start(esw, rep, sqs, num_sqs);
+	kfree(sqs);
+
+out:
+	if (err)
+		netdev_warn(priv->netdev, "Failed to add SQs FWD rules %d\n", err);
+	return err;
+}
+
+void mlx5e_remove_sqs_fwd_rules(struct mlx5e_priv *priv)
+{
+	struct mlx5_eswitch *esw = priv->mdev->priv.eswitch;
+	struct mlx5e_rep_priv *rpriv = priv->ppriv;
+	struct mlx5_eswitch_rep *rep = rpriv->rep;
+
+	mlx5e_sqs2vport_stop(esw, rep);
+}
+
+static void mlx5e_rep_neigh_update_init_interval(struct mlx5e_rep_priv *rpriv)
+{
+#if IS_ENABLED(CONFIG_IPV6)
+	unsigned long ipv6_interval = NEIGH_VAR(&nd_tbl.parms,
+						DELAY_PROBE_TIME);
+#else
+	unsigned long ipv6_interval = ~0UL;
+#endif
+	unsigned long ipv4_interval = NEIGH_VAR(&arp_tbl.parms,
+						DELAY_PROBE_TIME);
+	struct net_device *netdev = rpriv->netdev;
+	struct mlx5e_priv *priv = netdev_priv(netdev);
+
+	rpriv->neigh_update.min_interval = min_t(unsigned long, ipv6_interval, ipv4_interval);
+	mlx5_fc_update_sampling_interval(priv->mdev, rpriv->neigh_update.min_interval);
+}
+
+void mlx5e_rep_queue_neigh_stats_work(struct mlx5e_priv *priv)
+{
+	struct mlx5e_rep_priv *rpriv = priv->ppriv;
+	struct mlx5e_neigh_update_table *neigh_update = &rpriv->neigh_update;
+
+	mlx5_fc_queue_stats_work(priv->mdev,
+				 &neigh_update->neigh_stats_work,
+				 neigh_update->min_interval);
+}
+
+static void mlx5e_rep_neigh_stats_work(struct work_struct *work)
+{
+	struct mlx5e_rep_priv *rpriv = container_of(work, struct mlx5e_rep_priv,
+						    neigh_update.neigh_stats_work.work);
+	struct net_device *netdev = rpriv->netdev;
+	struct mlx5e_priv *priv = netdev_priv(netdev);
+	struct mlx5e_neigh_hash_entry *nhe;
+
+	rtnl_lock();
+	if (!list_empty(&rpriv->neigh_update.neigh_list))
+		mlx5e_rep_queue_neigh_stats_work(priv);
+
+	list_for_each_entry(nhe, &rpriv->neigh_update.neigh_list, neigh_list)
+		mlx5e_tc_update_neigh_used_value(nhe);
+
+	rtnl_unlock();
+}
+
+static void mlx5e_rep_neigh_entry_hold(struct mlx5e_neigh_hash_entry *nhe)
+{
+	refcount_inc(&nhe->refcnt);
+}
+
+static void mlx5e_rep_neigh_entry_release(struct mlx5e_neigh_hash_entry *nhe)
+{
+	if (refcount_dec_and_test(&nhe->refcnt))
+		kfree(nhe);
+}
+
+static void mlx5e_rep_update_flows(struct mlx5e_priv *priv,
+				   struct mlx5e_encap_entry *e,
+				   bool neigh_connected,
+				   unsigned char ha[ETH_ALEN])
+{
+	struct ethhdr *eth = (struct ethhdr *)e->encap_header;
+
+	ASSERT_RTNL();
+
+	if ((!neigh_connected && (e->flags & MLX5_ENCAP_ENTRY_VALID)) ||
+	    !ether_addr_equal(e->h_dest, ha))
+		mlx5e_tc_encap_flows_del(priv, e);
+
+	if (neigh_connected && !(e->flags & MLX5_ENCAP_ENTRY_VALID)) {
+		ether_addr_copy(e->h_dest, ha);
+		ether_addr_copy(eth->h_dest, ha);
+
+		mlx5e_tc_encap_flows_add(priv, e);
+	}
+}
+
+static void mlx5e_rep_neigh_update(struct work_struct *work)
+{
+	struct mlx5e_neigh_hash_entry *nhe =
+		container_of(work, struct mlx5e_neigh_hash_entry, neigh_update_work);
+	struct neighbour *n = nhe->n;
+	struct mlx5e_encap_entry *e;
+	unsigned char ha[ETH_ALEN];
+	struct mlx5e_priv *priv;
+	bool neigh_connected;
+	bool encap_connected;
+	u8 nud_state, dead;
+
+	rtnl_lock();
+
+	/* If these parameters are changed after we release the lock,
+	 * we'll receive another event letting us know about it.
+	 * We use this lock to avoid inconsistency between the neigh validity
+	 * and it's hw address.
+	 */
+	read_lock_bh(&n->lock);
+	memcpy(ha, n->ha, ETH_ALEN);
+	nud_state = n->nud_state;
+	dead = n->dead;
+	read_unlock_bh(&n->lock);
+
+	neigh_connected = (nud_state & NUD_VALID) && !dead;
+
+	list_for_each_entry(e, &nhe->encap_list, encap_list) {
+		encap_connected = !!(e->flags & MLX5_ENCAP_ENTRY_VALID);
+		priv = netdev_priv(e->out_dev);
+
+		if (encap_connected != neigh_connected ||
+		    !ether_addr_equal(e->h_dest, ha))
+			mlx5e_rep_update_flows(priv, e, neigh_connected, ha);
+	}
+	mlx5e_rep_neigh_entry_release(nhe);
+	rtnl_unlock();
+	neigh_release(n);
+}
+
+static struct mlx5e_neigh_hash_entry *
+mlx5e_rep_neigh_entry_lookup(struct mlx5e_priv *priv,
+			     struct mlx5e_neigh *m_neigh);
+
+static int mlx5e_rep_netevent_event(struct notifier_block *nb,
+				    unsigned long event, void *ptr)
+{
+	struct mlx5e_rep_priv *rpriv = container_of(nb, struct mlx5e_rep_priv,
+						    neigh_update.netevent_nb);
+	struct mlx5e_neigh_update_table *neigh_update = &rpriv->neigh_update;
+	struct net_device *netdev = rpriv->netdev;
+	struct mlx5e_priv *priv = netdev_priv(netdev);
+	struct mlx5e_neigh_hash_entry *nhe = NULL;
+	struct mlx5e_neigh m_neigh = {};
+	struct neigh_parms *p;
+	struct neighbour *n;
+	bool found = false;
+
+	switch (event) {
+	case NETEVENT_NEIGH_UPDATE:
+		n = ptr;
+#if IS_ENABLED(CONFIG_IPV6)
+		if (n->tbl != &nd_tbl && n->tbl != &arp_tbl)
+#else
+		if (n->tbl != &arp_tbl)
+#endif
+			return NOTIFY_DONE;
+
+		m_neigh.dev = n->dev;
+		m_neigh.family = n->ops->family;
+		memcpy(&m_neigh.dst_ip, n->primary_key, n->tbl->key_len);
+
+		/* We are in atomic context and can't take RTNL mutex, so use
+		 * spin_lock_bh to lookup the neigh table. bh is used since
+		 * netevent can be called from a softirq context.
+		 */
+		spin_lock_bh(&neigh_update->encap_lock);
+		nhe = mlx5e_rep_neigh_entry_lookup(priv, &m_neigh);
+		if (!nhe) {
+			spin_unlock_bh(&neigh_update->encap_lock);
+			return NOTIFY_DONE;
+		}
+
+		/* This assignment is valid as long as the the neigh reference
+		 * is taken
+		 */
+		nhe->n = n;
+
+		/* Take a reference to ensure the neighbour and mlx5 encap
+		 * entry won't be destructed until we drop the reference in
+		 * delayed work.
+		 */
+		neigh_hold(n);
+		mlx5e_rep_neigh_entry_hold(nhe);
+
+		if (!queue_work(priv->wq, &nhe->neigh_update_work)) {
+			mlx5e_rep_neigh_entry_release(nhe);
+			neigh_release(n);
+		}
+		spin_unlock_bh(&neigh_update->encap_lock);
+		break;
+
+	case NETEVENT_DELAY_PROBE_TIME_UPDATE:
+		p = ptr;
+
+		/* We check the device is present since we don't care about
+		 * changes in the default table, we only care about changes
+		 * done per device delay prob time parameter.
+		 */
+#if IS_ENABLED(CONFIG_IPV6)
+		if (!p->dev || (p->tbl != &nd_tbl && p->tbl != &arp_tbl))
+#else
+		if (!p->dev || p->tbl != &arp_tbl)
+#endif
+			return NOTIFY_DONE;
+
+		/* We are in atomic context and can't take RTNL mutex,
+		 * so use spin_lock_bh to walk the neigh list and look for
+		 * the relevant device. bh is used since netevent can be
+		 * called from a softirq context.
+		 */
+		spin_lock_bh(&neigh_update->encap_lock);
+		list_for_each_entry(nhe, &neigh_update->neigh_list, neigh_list) {
+			if (p->dev == nhe->m_neigh.dev) {
+				found = true;
+				break;
+			}
+		}
+		spin_unlock_bh(&neigh_update->encap_lock);
+		if (!found)
+			return NOTIFY_DONE;
+
+		neigh_update->min_interval = min_t(unsigned long,
+						   NEIGH_VAR(p, DELAY_PROBE_TIME),
+						   neigh_update->min_interval);
+		mlx5_fc_update_sampling_interval(priv->mdev,
+						 neigh_update->min_interval);
+		break;
+	}
+	return NOTIFY_DONE;
+}
+
+static const struct rhashtable_params mlx5e_neigh_ht_params = {
+	.head_offset = offsetof(struct mlx5e_neigh_hash_entry, rhash_node),
+	.key_offset = offsetof(struct mlx5e_neigh_hash_entry, m_neigh),
+	.key_len = sizeof(struct mlx5e_neigh),
+	.automatic_shrinking = true,
+};
+
+static int mlx5e_rep_neigh_init(struct mlx5e_rep_priv *rpriv)
+{
+	struct mlx5e_neigh_update_table *neigh_update = &rpriv->neigh_update;
+	int err;
+
+	err = rhashtable_init(&neigh_update->neigh_ht, &mlx5e_neigh_ht_params);
+	if (err)
+		return err;
+
+	INIT_LIST_HEAD(&neigh_update->neigh_list);
+	spin_lock_init(&neigh_update->encap_lock);
+	INIT_DELAYED_WORK(&neigh_update->neigh_stats_work,
+			  mlx5e_rep_neigh_stats_work);
+	mlx5e_rep_neigh_update_init_interval(rpriv);
+
+	rpriv->neigh_update.netevent_nb.notifier_call = mlx5e_rep_netevent_event;
+	err = register_netevent_notifier(&rpriv->neigh_update.netevent_nb);
+	if (err)
+		goto out_err;
+	return 0;
+
+out_err:
+	rhashtable_destroy(&neigh_update->neigh_ht);
+	return err;
+}
+
+static void mlx5e_rep_neigh_cleanup(struct mlx5e_rep_priv *rpriv)
+{
+	struct mlx5e_neigh_update_table *neigh_update = &rpriv->neigh_update;
+	struct mlx5e_priv *priv = netdev_priv(rpriv->netdev);
+
+	unregister_netevent_notifier(&neigh_update->netevent_nb);
+
+	flush_workqueue(priv->wq); /* flush neigh update works */
+
+	cancel_delayed_work_sync(&rpriv->neigh_update.neigh_stats_work);
+
+	rhashtable_destroy(&neigh_update->neigh_ht);
+}
+
+static int mlx5e_rep_neigh_entry_insert(struct mlx5e_priv *priv,
+					struct mlx5e_neigh_hash_entry *nhe)
+{
+	struct mlx5e_rep_priv *rpriv = priv->ppriv;
+	int err;
+
+	err = rhashtable_insert_fast(&rpriv->neigh_update.neigh_ht,
+				     &nhe->rhash_node,
+				     mlx5e_neigh_ht_params);
+	if (err)
+		return err;
+
+	list_add(&nhe->neigh_list, &rpriv->neigh_update.neigh_list);
+
+	return err;
+}
+
+static void mlx5e_rep_neigh_entry_remove(struct mlx5e_priv *priv,
+					 struct mlx5e_neigh_hash_entry *nhe)
+{
+	struct mlx5e_rep_priv *rpriv = priv->ppriv;
+
+	spin_lock_bh(&rpriv->neigh_update.encap_lock);
+
+	list_del(&nhe->neigh_list);
+
+	rhashtable_remove_fast(&rpriv->neigh_update.neigh_ht,
+			       &nhe->rhash_node,
+			       mlx5e_neigh_ht_params);
+	spin_unlock_bh(&rpriv->neigh_update.encap_lock);
+}
+
+/* This function must only be called under RTNL lock or under the
+ * representor's encap_lock in case RTNL mutex can't be held.
+ */
+static struct mlx5e_neigh_hash_entry *
+mlx5e_rep_neigh_entry_lookup(struct mlx5e_priv *priv,
+			     struct mlx5e_neigh *m_neigh)
+{
+	struct mlx5e_rep_priv *rpriv = priv->ppriv;
+	struct mlx5e_neigh_update_table *neigh_update = &rpriv->neigh_update;
+
+	return rhashtable_lookup_fast(&neigh_update->neigh_ht, m_neigh,
+				      mlx5e_neigh_ht_params);
+}
+
+static int mlx5e_rep_neigh_entry_create(struct mlx5e_priv *priv,
+					struct mlx5e_encap_entry *e,
+					struct mlx5e_neigh_hash_entry **nhe)
+{
+	int err;
+
+	*nhe = kzalloc(sizeof(**nhe), GFP_KERNEL);
+	if (!*nhe)
+		return -ENOMEM;
+
+	memcpy(&(*nhe)->m_neigh, &e->m_neigh, sizeof(e->m_neigh));
+	INIT_WORK(&(*nhe)->neigh_update_work, mlx5e_rep_neigh_update);
+	INIT_LIST_HEAD(&(*nhe)->encap_list);
+	refcount_set(&(*nhe)->refcnt, 1);
+
+	err = mlx5e_rep_neigh_entry_insert(priv, *nhe);
+	if (err)
+		goto out_free;
+	return 0;
+
+out_free:
+	kfree(*nhe);
+	return err;
+}
+
+static void mlx5e_rep_neigh_entry_destroy(struct mlx5e_priv *priv,
+					  struct mlx5e_neigh_hash_entry *nhe)
+{
+	/* The neigh hash entry must be removed from the hash table regardless
+	 * of the reference count value, so it won't be found by the next
+	 * neigh notification call. The neigh hash entry reference count is
+	 * incremented only during creation and neigh notification calls and
+	 * protects from freeing the nhe struct.
+	 */
+	mlx5e_rep_neigh_entry_remove(priv, nhe);
+	mlx5e_rep_neigh_entry_release(nhe);
+}
+
+int mlx5e_rep_encap_entry_attach(struct mlx5e_priv *priv,
+				 struct mlx5e_encap_entry *e)
+{
+	struct mlx5e_neigh_hash_entry *nhe;
+	int err;
+
+	nhe = mlx5e_rep_neigh_entry_lookup(priv, &e->m_neigh);
+	if (!nhe) {
+		err = mlx5e_rep_neigh_entry_create(priv, e, &nhe);
+		if (err)
+			return err;
+	}
+	list_add(&e->encap_list, &nhe->encap_list);
+	return 0;
+}
+
+void mlx5e_rep_encap_entry_detach(struct mlx5e_priv *priv,
+				  struct mlx5e_encap_entry *e)
+{
+	struct mlx5e_neigh_hash_entry *nhe;
+
+	list_del(&e->encap_list);
+	nhe = mlx5e_rep_neigh_entry_lookup(priv, &e->m_neigh);
+
+	if (list_empty(&nhe->encap_list))
+		mlx5e_rep_neigh_entry_destroy(priv, nhe);
+}
+
+static int mlx5e_rep_open(struct net_device *dev)
+{
+	struct mlx5e_priv *priv = netdev_priv(dev);
+	struct mlx5e_rep_priv *rpriv = priv->ppriv;
+	struct mlx5_eswitch_rep *rep = rpriv->rep;
+	int err;
+
+	mutex_lock(&priv->state_lock);
+	err = mlx5e_open_locked(dev);
+	if (err)
+		goto unlock;
+
+	if (!mlx5_modify_vport_admin_state(priv->mdev,
+					   MLX5_VPORT_STATE_OP_MOD_ESW_VPORT,
+					   rep->vport, MLX5_VPORT_ADMIN_STATE_UP))
+		netif_carrier_on(dev);
+
+unlock:
+	mutex_unlock(&priv->state_lock);
+	return err;
+}
+
+static int mlx5e_rep_close(struct net_device *dev)
+{
+	struct mlx5e_priv *priv = netdev_priv(dev);
+	struct mlx5e_rep_priv *rpriv = priv->ppriv;
+	struct mlx5_eswitch_rep *rep = rpriv->rep;
+	int ret;
+
+	mutex_lock(&priv->state_lock);
+	mlx5_modify_vport_admin_state(priv->mdev,
+				      MLX5_VPORT_STATE_OP_MOD_ESW_VPORT,
+				      rep->vport, MLX5_VPORT_ADMIN_STATE_DOWN);
+	ret = mlx5e_close_locked(dev);
+	mutex_unlock(&priv->state_lock);
+	return ret;
+}
+
+static int mlx5e_rep_get_phys_port_name(struct net_device *dev,
+					char *buf, size_t len)
+{
+	struct mlx5e_priv *priv = netdev_priv(dev);
+	struct mlx5e_rep_priv *rpriv = priv->ppriv;
+	struct mlx5_eswitch_rep *rep = rpriv->rep;
+	int ret;
+
+	ret = snprintf(buf, len, "%d", rep->vport - 1);
+	if (ret >= len)
+		return -EOPNOTSUPP;
+
+	return 0;
+}
+
+static int
+mlx5e_rep_setup_tc_cls_flower(struct mlx5e_priv *priv,
+			      struct tc_cls_flower_offload *cls_flower, int flags)
+{
+	switch (cls_flower->command) {
+	case TC_CLSFLOWER_REPLACE:
+		return mlx5e_configure_flower(priv, cls_flower, flags);
+	case TC_CLSFLOWER_DESTROY:
+		return mlx5e_delete_flower(priv, cls_flower, flags);
+	case TC_CLSFLOWER_STATS:
+		return mlx5e_stats_flower(priv, cls_flower, flags);
+	default:
+		return -EOPNOTSUPP;
+	}
+}
+
+static int mlx5e_rep_setup_tc_cb_egdev(enum tc_setup_type type, void *type_data,
+				       void *cb_priv)
+{
+	struct mlx5e_priv *priv = cb_priv;
+
+	if (!tc_cls_can_offload_and_chain0(priv->netdev, type_data))
+		return -EOPNOTSUPP;
+
+	switch (type) {
+	case TC_SETUP_CLSFLOWER:
+		return mlx5e_rep_setup_tc_cls_flower(priv, type_data, MLX5E_TC_EGRESS);
+	default:
+		return -EOPNOTSUPP;
+	}
+}
+
+static int mlx5e_rep_setup_tc_cb(enum tc_setup_type type, void *type_data,
+				 void *cb_priv)
+{
+	struct mlx5e_priv *priv = cb_priv;
+
+	if (!tc_cls_can_offload_and_chain0(priv->netdev, type_data))
+		return -EOPNOTSUPP;
+
+	switch (type) {
+	case TC_SETUP_CLSFLOWER:
+		return mlx5e_rep_setup_tc_cls_flower(priv, type_data, MLX5E_TC_INGRESS);
+	default:
+		return -EOPNOTSUPP;
+	}
+}
+
+static int mlx5e_rep_setup_tc_block(struct net_device *dev,
+				    struct tc_block_offload *f)
+{
+	struct mlx5e_priv *priv = netdev_priv(dev);
+
+	if (f->binder_type != TCF_BLOCK_BINDER_TYPE_CLSACT_INGRESS)
+		return -EOPNOTSUPP;
+
+	switch (f->command) {
+	case TC_BLOCK_BIND:
+		return tcf_block_cb_register(f->block, mlx5e_rep_setup_tc_cb,
+					     priv, priv, f->extack);
+	case TC_BLOCK_UNBIND:
+		tcf_block_cb_unregister(f->block, mlx5e_rep_setup_tc_cb, priv);
+		return 0;
+	default:
+		return -EOPNOTSUPP;
+	}
+}
+
+static int mlx5e_rep_setup_tc(struct net_device *dev, enum tc_setup_type type,
+			      void *type_data)
+{
+	switch (type) {
+	case TC_SETUP_BLOCK:
+		return mlx5e_rep_setup_tc_block(dev, type_data);
+	default:
+		return -EOPNOTSUPP;
+	}
+}
+
+bool mlx5e_is_uplink_rep(struct mlx5e_priv *priv)
+{
+	struct mlx5_eswitch *esw = priv->mdev->priv.eswitch;
+	struct mlx5e_rep_priv *rpriv = priv->ppriv;
+	struct mlx5_eswitch_rep *rep;
+
+	if (!MLX5_ESWITCH_MANAGER(priv->mdev))
+		return false;
+
+	rep = rpriv->rep;
+	if (esw->mode == SRIOV_OFFLOADS &&
+	    rep && rep->vport == FDB_UPLINK_VPORT)
+		return true;
+
+	return false;
+}
+
+static bool mlx5e_is_vf_vport_rep(struct mlx5e_priv *priv)
+{
+	struct mlx5e_rep_priv *rpriv = priv->ppriv;
+	struct mlx5_eswitch_rep *rep;
+
+	if (!MLX5_ESWITCH_MANAGER(priv->mdev))
+		return false;
+
+	rep = rpriv->rep;
+	if (rep && rep->vport != FDB_UPLINK_VPORT)
+		return true;
+
+	return false;
+}
+
+bool mlx5e_has_offload_stats(const struct net_device *dev, int attr_id)
+{
+	struct mlx5e_priv *priv = netdev_priv(dev);
+
+	switch (attr_id) {
+	case IFLA_OFFLOAD_XSTATS_CPU_HIT:
+		if (mlx5e_is_vf_vport_rep(priv) || mlx5e_is_uplink_rep(priv))
+			return true;
+	}
+
+	return false;
+}
+
+static int
+mlx5e_get_sw_stats64(const struct net_device *dev,
+		     struct rtnl_link_stats64 *stats)
+{
+	struct mlx5e_priv *priv = netdev_priv(dev);
+	struct mlx5e_sw_stats *sstats = &priv->stats.sw;
+
+	mlx5e_rep_update_sw_counters(priv);
+
+	stats->rx_packets = sstats->rx_packets;
+	stats->rx_bytes   = sstats->rx_bytes;
+	stats->tx_packets = sstats->tx_packets;
+	stats->tx_bytes   = sstats->tx_bytes;
+
+	stats->tx_dropped = sstats->tx_queue_dropped;
+
+	return 0;
+}
+
+int mlx5e_get_offload_stats(int attr_id, const struct net_device *dev,
+			    void *sp)
+{
+	switch (attr_id) {
+	case IFLA_OFFLOAD_XSTATS_CPU_HIT:
+		return mlx5e_get_sw_stats64(dev, sp);
+	}
+
+	return -EINVAL;
+}
+
+static void
+mlx5e_rep_get_stats(struct net_device *dev, struct rtnl_link_stats64 *stats)
+{
+	struct mlx5e_priv *priv = netdev_priv(dev);
+
+	/* update HW stats in background for next time */
+	queue_delayed_work(priv->wq, &priv->update_stats_work, 0);
+
+	memcpy(stats, &priv->stats.vf_vport, sizeof(*stats));
+}
+
+static const struct switchdev_ops mlx5e_rep_switchdev_ops = {
+	.switchdev_port_attr_get	= mlx5e_attr_get,
+};
+
+static int mlx5e_change_rep_mtu(struct net_device *netdev, int new_mtu)
+{
+	return mlx5e_change_mtu(netdev, new_mtu, NULL);
+}
+
+static const struct net_device_ops mlx5e_netdev_ops_rep = {
+	.ndo_open                = mlx5e_rep_open,
+	.ndo_stop                = mlx5e_rep_close,
+	.ndo_start_xmit          = mlx5e_xmit,
+	.ndo_get_phys_port_name  = mlx5e_rep_get_phys_port_name,
+	.ndo_setup_tc            = mlx5e_rep_setup_tc,
+	.ndo_get_stats64         = mlx5e_rep_get_stats,
+	.ndo_has_offload_stats	 = mlx5e_has_offload_stats,
+	.ndo_get_offload_stats	 = mlx5e_get_offload_stats,
+	.ndo_change_mtu          = mlx5e_change_rep_mtu,
+};
+
+static void mlx5e_build_rep_params(struct mlx5_core_dev *mdev,
+				   struct mlx5e_params *params, u16 mtu)
+{
+	u8 cq_period_mode = MLX5_CAP_GEN(mdev, cq_period_start_from_cqe) ?
+					 MLX5_CQ_PERIOD_MODE_START_FROM_CQE :
+					 MLX5_CQ_PERIOD_MODE_START_FROM_EQE;
+
+	params->hard_mtu    = MLX5E_ETH_HARD_MTU;
+	params->sw_mtu      = mtu;
+	params->log_sq_size = MLX5E_REP_PARAMS_LOG_SQ_SIZE;
+	params->rq_wq_type  = MLX5_WQ_TYPE_CYCLIC;
+	params->log_rq_mtu_frames = MLX5E_REP_PARAMS_LOG_RQ_SIZE;
+
+	params->rx_dim_enabled = MLX5_CAP_GEN(mdev, cq_moderation);
+	mlx5e_set_rx_cq_mode_params(params, cq_period_mode);
+
+	params->num_tc                = 1;
+	params->lro_wqe_sz            = MLX5E_PARAMS_DEFAULT_LRO_WQE_SZ;
+
+	mlx5_query_min_inline(mdev, &params->tx_min_inline_mode);
+}
+
+static void mlx5e_build_rep_netdev(struct net_device *netdev)
+{
+	struct mlx5e_priv *priv = netdev_priv(netdev);
+	struct mlx5_core_dev *mdev = priv->mdev;
+	u16 max_mtu;
+
+	netdev->netdev_ops = &mlx5e_netdev_ops_rep;
+
+	netdev->watchdog_timeo    = 15 * HZ;
+
+	netdev->ethtool_ops	  = &mlx5e_rep_ethtool_ops;
+
+	netdev->switchdev_ops = &mlx5e_rep_switchdev_ops;
+
+	netdev->features	 |= NETIF_F_VLAN_CHALLENGED | NETIF_F_HW_TC | NETIF_F_NETNS_LOCAL;
+	netdev->hw_features      |= NETIF_F_HW_TC;
+
+	eth_hw_addr_random(netdev);
+
+	netdev->min_mtu = ETH_MIN_MTU;
+	mlx5_query_port_max_mtu(mdev, &max_mtu, 1);
+	netdev->max_mtu = MLX5E_HW2SW_MTU(&priv->channels.params, max_mtu);
+}
+
+static void mlx5e_init_rep(struct mlx5_core_dev *mdev,
+			   struct net_device *netdev,
+			   const struct mlx5e_profile *profile,
+			   void *ppriv)
+{
+	struct mlx5e_priv *priv = netdev_priv(netdev);
+
+	priv->mdev                         = mdev;
+	priv->netdev                       = netdev;
+	priv->profile                      = profile;
+	priv->ppriv                        = ppriv;
+
+	mutex_init(&priv->state_lock);
+
+	INIT_DELAYED_WORK(&priv->update_stats_work, mlx5e_update_stats_work);
+
+	priv->channels.params.num_channels = profile->max_nch(mdev);
+
+	mlx5e_build_rep_params(mdev, &priv->channels.params, netdev->mtu);
+	mlx5e_build_rep_netdev(netdev);
+
+	mlx5e_timestamp_init(priv);
+}
+
+static int mlx5e_init_rep_rx(struct mlx5e_priv *priv)
+{
+	struct mlx5_eswitch *esw = priv->mdev->priv.eswitch;
+	struct mlx5e_rep_priv *rpriv = priv->ppriv;
+	struct mlx5_eswitch_rep *rep = rpriv->rep;
+	struct mlx5_flow_handle *flow_rule;
+	int err;
+
+	mlx5e_init_l2_addr(priv);
+
+	err = mlx5e_create_direct_rqts(priv);
+	if (err)
+		return err;
+
+	err = mlx5e_create_direct_tirs(priv);
+	if (err)
+		goto err_destroy_direct_rqts;
+
+	flow_rule = mlx5_eswitch_create_vport_rx_rule(esw,
+						      rep->vport,
+						      priv->direct_tir[0].tirn);
+	if (IS_ERR(flow_rule)) {
+		err = PTR_ERR(flow_rule);
+		goto err_destroy_direct_tirs;
+	}
+	rpriv->vport_rx_rule = flow_rule;
+
+	return 0;
+
+err_destroy_direct_tirs:
+	mlx5e_destroy_direct_tirs(priv);
+err_destroy_direct_rqts:
+	mlx5e_destroy_direct_rqts(priv);
+	return err;
+}
+
+static void mlx5e_cleanup_rep_rx(struct mlx5e_priv *priv)
+{
+	struct mlx5e_rep_priv *rpriv = priv->ppriv;
+
+	mlx5_del_flow_rules(rpriv->vport_rx_rule);
+	mlx5e_destroy_direct_tirs(priv);
+	mlx5e_destroy_direct_rqts(priv);
+}
+
+static int mlx5e_init_rep_tx(struct mlx5e_priv *priv)
+{
+	int err;
+
+	err = mlx5e_create_tises(priv);
+	if (err) {
+		mlx5_core_warn(priv->mdev, "create tises failed, %d\n", err);
+		return err;
+	}
+	return 0;
+}
+
+static int mlx5e_get_rep_max_num_channels(struct mlx5_core_dev *mdev)
+{
+#define	MLX5E_PORT_REPRESENTOR_NCH 1
+	return MLX5E_PORT_REPRESENTOR_NCH;
+}
+
+static const struct mlx5e_profile mlx5e_rep_profile = {
+	.init			= mlx5e_init_rep,
+	.init_rx		= mlx5e_init_rep_rx,
+	.cleanup_rx		= mlx5e_cleanup_rep_rx,
+	.init_tx		= mlx5e_init_rep_tx,
+	.cleanup_tx		= mlx5e_cleanup_nic_tx,
+	.update_stats           = mlx5e_rep_update_hw_counters,
+	.max_nch		= mlx5e_get_rep_max_num_channels,
+	.update_carrier		= NULL,
+	.rx_handlers.handle_rx_cqe       = mlx5e_handle_rx_cqe_rep,
+	.rx_handlers.handle_rx_cqe_mpwqe = NULL /* Not supported */,
+	.max_tc			= 1,
+};
+
+/* e-Switch vport representors */
+
+static int
+mlx5e_nic_rep_load(struct mlx5_core_dev *dev, struct mlx5_eswitch_rep *rep)
+{
+	struct mlx5e_rep_priv *rpriv = mlx5e_rep_to_rep_priv(rep);
+	struct mlx5e_priv *priv = netdev_priv(rpriv->netdev);
+
+	int err;
+
+	if (test_bit(MLX5E_STATE_OPENED, &priv->state)) {
+		err = mlx5e_add_sqs_fwd_rules(priv);
+		if (err)
+			return err;
+	}
+
+	err = mlx5e_rep_neigh_init(rpriv);
+	if (err)
+		goto err_remove_sqs;
+
+	/* init shared tc flow table */
+	err = mlx5e_tc_esw_init(&rpriv->tc_ht);
+	if (err)
+		goto  err_neigh_cleanup;
+
+	return 0;
+
+err_neigh_cleanup:
+	mlx5e_rep_neigh_cleanup(rpriv);
+err_remove_sqs:
+	mlx5e_remove_sqs_fwd_rules(priv);
+	return err;
+}
+
+static void
+mlx5e_nic_rep_unload(struct mlx5_eswitch_rep *rep)
+{
+	struct mlx5e_rep_priv *rpriv = mlx5e_rep_to_rep_priv(rep);
+	struct mlx5e_priv *priv = netdev_priv(rpriv->netdev);
+
+	if (test_bit(MLX5E_STATE_OPENED, &priv->state))
+		mlx5e_remove_sqs_fwd_rules(priv);
+
+	/* clean uplink offloaded TC rules, delete shared tc flow table */
+	mlx5e_tc_esw_cleanup(&rpriv->tc_ht);
+
+	mlx5e_rep_neigh_cleanup(rpriv);
+}
+
+static int
+mlx5e_vport_rep_load(struct mlx5_core_dev *dev, struct mlx5_eswitch_rep *rep)
+{
+	struct mlx5e_rep_priv *uplink_rpriv;
+	struct mlx5e_rep_priv *rpriv;
+	struct net_device *netdev;
+	struct mlx5e_priv *upriv;
+	int err;
+
+	rpriv = kzalloc(sizeof(*rpriv), GFP_KERNEL);
+	if (!rpriv)
+		return -ENOMEM;
+
+	netdev = mlx5e_create_netdev(dev, &mlx5e_rep_profile, rpriv);
+	if (!netdev) {
+		pr_warn("Failed to create representor netdev for vport %d\n",
+			rep->vport);
+		kfree(rpriv);
+		return -EINVAL;
+	}
+
+	rpriv->netdev = netdev;
+	rpriv->rep = rep;
+	rep->rep_if[REP_ETH].priv = rpriv;
+	INIT_LIST_HEAD(&rpriv->vport_sqs_list);
+
+	err = mlx5e_attach_netdev(netdev_priv(netdev));
+	if (err) {
+		pr_warn("Failed to attach representor netdev for vport %d\n",
+			rep->vport);
+		goto err_destroy_netdev;
+	}
+
+	err = mlx5e_rep_neigh_init(rpriv);
+	if (err) {
+		pr_warn("Failed to initialized neighbours handling for vport %d\n",
+			rep->vport);
+		goto err_detach_netdev;
+	}
+
+	uplink_rpriv = mlx5_eswitch_get_uplink_priv(dev->priv.eswitch, REP_ETH);
+	upriv = netdev_priv(uplink_rpriv->netdev);
+	err = tc_setup_cb_egdev_register(netdev, mlx5e_rep_setup_tc_cb_egdev,
+					 upriv);
+	if (err)
+		goto err_neigh_cleanup;
+
+	err = register_netdev(netdev);
+	if (err) {
+		pr_warn("Failed to register representor netdev for vport %d\n",
+			rep->vport);
+		goto err_egdev_cleanup;
+	}
+
+	return 0;
+
+err_egdev_cleanup:
+	tc_setup_cb_egdev_unregister(netdev, mlx5e_rep_setup_tc_cb_egdev,
+				     upriv);
+
+err_neigh_cleanup:
+	mlx5e_rep_neigh_cleanup(rpriv);
+
+err_detach_netdev:
+	mlx5e_detach_netdev(netdev_priv(netdev));
+
+err_destroy_netdev:
+	mlx5e_destroy_netdev(netdev_priv(netdev));
+	kfree(rpriv);
+	return err;
+}
+
+static void
+mlx5e_vport_rep_unload(struct mlx5_eswitch_rep *rep)
+{
+	struct mlx5e_rep_priv *rpriv = mlx5e_rep_to_rep_priv(rep);
+	struct net_device *netdev = rpriv->netdev;
+	struct mlx5e_priv *priv = netdev_priv(netdev);
+	struct mlx5e_rep_priv *uplink_rpriv;
+	void *ppriv = priv->ppriv;
+	struct mlx5e_priv *upriv;
+
+	unregister_netdev(netdev);
+	uplink_rpriv = mlx5_eswitch_get_uplink_priv(priv->mdev->priv.eswitch,
+						    REP_ETH);
+	upriv = netdev_priv(uplink_rpriv->netdev);
+	tc_setup_cb_egdev_unregister(netdev, mlx5e_rep_setup_tc_cb_egdev,
+				     upriv);
+	mlx5e_rep_neigh_cleanup(rpriv);
+	mlx5e_detach_netdev(priv);
+	mlx5e_destroy_netdev(priv);
+	kfree(ppriv); /* mlx5e_rep_priv */
+}
+
+static void *mlx5e_vport_rep_get_proto_dev(struct mlx5_eswitch_rep *rep)
+{
+	struct mlx5e_rep_priv *rpriv;
+
+	rpriv = mlx5e_rep_to_rep_priv(rep);
+
+	return rpriv->netdev;
+}
+
+static void mlx5e_rep_register_vf_vports(struct mlx5e_priv *priv)
+{
+	struct mlx5_core_dev *mdev = priv->mdev;
+	struct mlx5_eswitch *esw   = mdev->priv.eswitch;
+	int total_vfs = MLX5_TOTAL_VPORTS(mdev);
+	int vport;
+
+	for (vport = 1; vport < total_vfs; vport++) {
+		struct mlx5_eswitch_rep_if rep_if = {};
+
+		rep_if.load = mlx5e_vport_rep_load;
+		rep_if.unload = mlx5e_vport_rep_unload;
+		rep_if.get_proto_dev = mlx5e_vport_rep_get_proto_dev;
+		mlx5_eswitch_register_vport_rep(esw, vport, &rep_if, REP_ETH);
+	}
+}
+
+static void mlx5e_rep_unregister_vf_vports(struct mlx5e_priv *priv)
+{
+	struct mlx5_core_dev *mdev = priv->mdev;
+	struct mlx5_eswitch *esw = mdev->priv.eswitch;
+	int total_vfs = MLX5_TOTAL_VPORTS(mdev);
+	int vport;
+
+	for (vport = 1; vport < total_vfs; vport++)
+		mlx5_eswitch_unregister_vport_rep(esw, vport, REP_ETH);
+}
+
+void mlx5e_register_vport_reps(struct mlx5e_priv *priv)
+{
+	struct mlx5_core_dev *mdev = priv->mdev;
+	struct mlx5_eswitch *esw   = mdev->priv.eswitch;
+	struct mlx5_eswitch_rep_if rep_if;
+	struct mlx5e_rep_priv *rpriv;
+
+	rpriv = priv->ppriv;
+	rpriv->netdev = priv->netdev;
+
+	rep_if.load = mlx5e_nic_rep_load;
+	rep_if.unload = mlx5e_nic_rep_unload;
+	rep_if.get_proto_dev = mlx5e_vport_rep_get_proto_dev;
+	rep_if.priv = rpriv;
+	INIT_LIST_HEAD(&rpriv->vport_sqs_list);
+	mlx5_eswitch_register_vport_rep(esw, 0, &rep_if, REP_ETH); /* UPLINK PF vport*/
+
+	mlx5e_rep_register_vf_vports(priv); /* VFs vports */
+}
+
+void mlx5e_unregister_vport_reps(struct mlx5e_priv *priv)
+{
+	struct mlx5_core_dev *mdev = priv->mdev;
+	struct mlx5_eswitch *esw   = mdev->priv.eswitch;
+
+	mlx5e_rep_unregister_vf_vports(priv); /* VFs vports */
+	mlx5_eswitch_unregister_vport_rep(esw, 0, REP_ETH); /* UPLINK PF*/
+}
+
+void *mlx5e_alloc_nic_rep_priv(struct mlx5_core_dev *mdev)
+{
+	struct mlx5_eswitch *esw = mdev->priv.eswitch;
+	struct mlx5e_rep_priv *rpriv;
+
+	rpriv = kzalloc(sizeof(*rpriv), GFP_KERNEL);
+	if (!rpriv)
+		return NULL;
+
+	rpriv->rep = &esw->offloads.vport_reps[0];
+	return rpriv;
+}
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_rep.h b/drivers/net/ethernet/mellanox/mlx5/core/en_rep.h
new file mode 100644
index 000000000..844d32d5c
--- /dev/null
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_rep.h
@@ -0,0 +1,169 @@
+/*
+ * Copyright (c) 2017, 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 __MLX5E_REP_H__
+#define __MLX5E_REP_H__
+
+#include <net/ip_tunnels.h>
+#include <linux/rhashtable.h>
+#include "eswitch.h"
+#include "en.h"
+
+#ifdef CONFIG_MLX5_ESWITCH
+struct mlx5e_neigh_update_table {
+	struct rhashtable       neigh_ht;
+	/* Save the neigh hash entries in a list in addition to the hash table
+	 * (neigh_ht). In order to iterate easily over the neigh entries.
+	 * Used for stats query.
+	 */
+	struct list_head	neigh_list;
+	/* protect lookup/remove operations */
+	spinlock_t              encap_lock;
+	struct notifier_block   netevent_nb;
+	struct delayed_work     neigh_stats_work;
+	unsigned long           min_interval; /* jiffies */
+};
+
+struct mlx5e_rep_priv {
+	struct mlx5_eswitch_rep *rep;
+	struct mlx5e_neigh_update_table neigh_update;
+	struct net_device      *netdev;
+	struct mlx5_flow_handle *vport_rx_rule;
+	struct list_head       vport_sqs_list;
+	struct rhashtable      tc_ht; /* valid for uplink rep */
+};
+
+static inline
+struct mlx5e_rep_priv *mlx5e_rep_to_rep_priv(struct mlx5_eswitch_rep *rep)
+{
+	return (struct mlx5e_rep_priv *)rep->rep_if[REP_ETH].priv;
+}
+
+struct mlx5e_neigh {
+	struct net_device *dev;
+	union {
+		__be32	v4;
+		struct in6_addr v6;
+	} dst_ip;
+	int family;
+};
+
+struct mlx5e_neigh_hash_entry {
+	struct rhash_head rhash_node;
+	struct mlx5e_neigh m_neigh;
+
+	/* Save the neigh hash entry in a list on the representor in
+	 * addition to the hash table. In order to iterate easily over the
+	 * neighbour entries. Used for stats query.
+	 */
+	struct list_head neigh_list;
+
+	/* encap list sharing the same neigh */
+	struct list_head encap_list;
+
+	/* valid only when the neigh reference is taken during
+	 * neigh_update_work workqueue callback.
+	 */
+	struct neighbour *n;
+	struct work_struct neigh_update_work;
+
+	/* neigh hash entry can be deleted only when the refcount is zero.
+	 * refcount is needed to avoid neigh hash entry removal by TC, while
+	 * it's used by the neigh notification call.
+	 */
+	refcount_t refcnt;
+
+	/* Save the last reported time offloaded trafic pass over one of the
+	 * neigh hash entry flows. Use it to periodically update the neigh
+	 * 'used' value and avoid neigh deleting by the kernel.
+	 */
+	unsigned long reported_lastuse;
+};
+
+enum {
+	/* set when the encap entry is successfully offloaded into HW */
+	MLX5_ENCAP_ENTRY_VALID     = BIT(0),
+};
+
+struct mlx5e_encap_entry {
+	/* neigh hash entry list of encaps sharing the same neigh */
+	struct list_head encap_list;
+	struct mlx5e_neigh m_neigh;
+	/* a node of the eswitch encap hash table which keeping all the encap
+	 * entries
+	 */
+	struct hlist_node encap_hlist;
+	struct list_head flows;
+	u32 encap_id;
+	struct ip_tunnel_info tun_info;
+	unsigned char h_dest[ETH_ALEN];	/* destination eth addr	*/
+
+	struct net_device *out_dev;
+	int tunnel_type;
+	u8 flags;
+	char *encap_header;
+	int encap_size;
+};
+
+struct mlx5e_rep_sq {
+	struct mlx5_flow_handle	*send_to_vport_rule;
+	struct list_head	 list;
+};
+
+void *mlx5e_alloc_nic_rep_priv(struct mlx5_core_dev *mdev);
+void mlx5e_register_vport_reps(struct mlx5e_priv *priv);
+void mlx5e_unregister_vport_reps(struct mlx5e_priv *priv);
+bool mlx5e_is_uplink_rep(struct mlx5e_priv *priv);
+int mlx5e_add_sqs_fwd_rules(struct mlx5e_priv *priv);
+void mlx5e_remove_sqs_fwd_rules(struct mlx5e_priv *priv);
+
+int mlx5e_get_offload_stats(int attr_id, const struct net_device *dev, void *sp);
+bool mlx5e_has_offload_stats(const struct net_device *dev, int attr_id);
+
+int mlx5e_attr_get(struct net_device *dev, struct switchdev_attr *attr);
+void mlx5e_handle_rx_cqe_rep(struct mlx5e_rq *rq, struct mlx5_cqe64 *cqe);
+
+int mlx5e_rep_encap_entry_attach(struct mlx5e_priv *priv,
+				 struct mlx5e_encap_entry *e);
+void mlx5e_rep_encap_entry_detach(struct mlx5e_priv *priv,
+				  struct mlx5e_encap_entry *e);
+
+void mlx5e_rep_queue_neigh_stats_work(struct mlx5e_priv *priv);
+#else /* CONFIG_MLX5_ESWITCH */
+static inline void mlx5e_register_vport_reps(struct mlx5e_priv *priv) {}
+static inline void mlx5e_unregister_vport_reps(struct mlx5e_priv *priv) {}
+static inline bool mlx5e_is_uplink_rep(struct mlx5e_priv *priv) { return false; }
+static inline int mlx5e_add_sqs_fwd_rules(struct mlx5e_priv *priv) { return 0; }
+static inline void mlx5e_remove_sqs_fwd_rules(struct mlx5e_priv *priv) {}
+#endif
+
+#endif /* __MLX5E_REP_H__ */
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_rx.c b/drivers/net/ethernet/mellanox/mlx5/core/en_rx.c
new file mode 100644
index 000000000..9d86e49a7
--- /dev/null
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_rx.c
@@ -0,0 +1,1470 @@
+/*
+ * Copyright (c) 2015, 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/prefetch.h>
+#include <linux/ip.h>
+#include <linux/ipv6.h>
+#include <linux/tcp.h>
+#include <net/busy_poll.h>
+#include <net/ip6_checksum.h>
+#include <net/page_pool.h>
+#include <net/inet_ecn.h>
+#include "en.h"
+#include "en_tc.h"
+#include "eswitch.h"
+#include "en_rep.h"
+#include "ipoib/ipoib.h"
+#include "en_accel/ipsec_rxtx.h"
+#include "en_accel/tls_rxtx.h"
+#include "lib/clock.h"
+#include "en/xdp.h"
+
+static inline bool mlx5e_rx_hw_stamp(struct hwtstamp_config *config)
+{
+	return config->rx_filter == HWTSTAMP_FILTER_ALL;
+}
+
+static inline void mlx5e_read_cqe_slot(struct mlx5e_cq *cq, u32 cqcc,
+				       void *data)
+{
+	u32 ci = mlx5_cqwq_ctr2ix(&cq->wq, cqcc);
+
+	memcpy(data, mlx5_cqwq_get_wqe(&cq->wq, ci), sizeof(struct mlx5_cqe64));
+}
+
+static inline void mlx5e_read_title_slot(struct mlx5e_rq *rq,
+					 struct mlx5e_cq *cq, u32 cqcc)
+{
+	mlx5e_read_cqe_slot(cq, cqcc, &cq->title);
+	cq->decmprs_left        = be32_to_cpu(cq->title.byte_cnt);
+	cq->decmprs_wqe_counter = be16_to_cpu(cq->title.wqe_counter);
+	rq->stats->cqe_compress_blks++;
+}
+
+static inline void mlx5e_read_mini_arr_slot(struct mlx5e_cq *cq, u32 cqcc)
+{
+	mlx5e_read_cqe_slot(cq, cqcc, cq->mini_arr);
+	cq->mini_arr_idx = 0;
+}
+
+static inline void mlx5e_cqes_update_owner(struct mlx5e_cq *cq, u32 cqcc, int n)
+{
+	struct mlx5_cqwq *wq = &cq->wq;
+
+	u8  op_own = mlx5_cqwq_get_ctr_wrap_cnt(wq, cqcc) & 1;
+	u32 ci     = mlx5_cqwq_ctr2ix(wq, cqcc);
+	u32 wq_sz  = mlx5_cqwq_get_size(wq);
+	u32 ci_top = min_t(u32, wq_sz, ci + n);
+
+	for (; ci < ci_top; ci++, n--) {
+		struct mlx5_cqe64 *cqe = mlx5_cqwq_get_wqe(&cq->wq, ci);
+
+		cqe->op_own = op_own;
+	}
+
+	if (unlikely(ci == wq_sz)) {
+		op_own = !op_own;
+		for (ci = 0; ci < n; ci++) {
+			struct mlx5_cqe64 *cqe = mlx5_cqwq_get_wqe(&cq->wq, ci);
+
+			cqe->op_own = op_own;
+		}
+	}
+}
+
+static inline void mlx5e_decompress_cqe(struct mlx5e_rq *rq,
+					struct mlx5e_cq *cq, u32 cqcc)
+{
+	cq->title.byte_cnt     = cq->mini_arr[cq->mini_arr_idx].byte_cnt;
+	cq->title.check_sum    = cq->mini_arr[cq->mini_arr_idx].checksum;
+	cq->title.op_own      &= 0xf0;
+	cq->title.op_own      |= 0x01 & (cqcc >> cq->wq.fbc.log_sz);
+	cq->title.wqe_counter  = cpu_to_be16(cq->decmprs_wqe_counter);
+
+	if (rq->wq_type == MLX5_WQ_TYPE_LINKED_LIST_STRIDING_RQ)
+		cq->decmprs_wqe_counter +=
+			mpwrq_get_cqe_consumed_strides(&cq->title);
+	else
+		cq->decmprs_wqe_counter =
+			mlx5_wq_cyc_ctr2ix(&rq->wqe.wq, cq->decmprs_wqe_counter + 1);
+}
+
+static inline void mlx5e_decompress_cqe_no_hash(struct mlx5e_rq *rq,
+						struct mlx5e_cq *cq, u32 cqcc)
+{
+	mlx5e_decompress_cqe(rq, cq, cqcc);
+	cq->title.rss_hash_type   = 0;
+	cq->title.rss_hash_result = 0;
+}
+
+static inline u32 mlx5e_decompress_cqes_cont(struct mlx5e_rq *rq,
+					     struct mlx5e_cq *cq,
+					     int update_owner_only,
+					     int budget_rem)
+{
+	u32 cqcc = cq->wq.cc + update_owner_only;
+	u32 cqe_count;
+	u32 i;
+
+	cqe_count = min_t(u32, cq->decmprs_left, budget_rem);
+
+	for (i = update_owner_only; i < cqe_count;
+	     i++, cq->mini_arr_idx++, cqcc++) {
+		if (cq->mini_arr_idx == MLX5_MINI_CQE_ARRAY_SIZE)
+			mlx5e_read_mini_arr_slot(cq, cqcc);
+
+		mlx5e_decompress_cqe_no_hash(rq, cq, cqcc);
+		rq->handle_rx_cqe(rq, &cq->title);
+	}
+	mlx5e_cqes_update_owner(cq, cq->wq.cc, cqcc - cq->wq.cc);
+	cq->wq.cc = cqcc;
+	cq->decmprs_left -= cqe_count;
+	rq->stats->cqe_compress_pkts += cqe_count;
+
+	return cqe_count;
+}
+
+static inline u32 mlx5e_decompress_cqes_start(struct mlx5e_rq *rq,
+					      struct mlx5e_cq *cq,
+					      int budget_rem)
+{
+	mlx5e_read_title_slot(rq, cq, cq->wq.cc);
+	mlx5e_read_mini_arr_slot(cq, cq->wq.cc + 1);
+	mlx5e_decompress_cqe(rq, cq, cq->wq.cc);
+	rq->handle_rx_cqe(rq, &cq->title);
+	cq->mini_arr_idx++;
+
+	return mlx5e_decompress_cqes_cont(rq, cq, 1, budget_rem) - 1;
+}
+
+static inline bool mlx5e_page_is_reserved(struct page *page)
+{
+	return page_is_pfmemalloc(page) || page_to_nid(page) != numa_mem_id();
+}
+
+static inline bool mlx5e_rx_cache_put(struct mlx5e_rq *rq,
+				      struct mlx5e_dma_info *dma_info)
+{
+	struct mlx5e_page_cache *cache = &rq->page_cache;
+	u32 tail_next = (cache->tail + 1) & (MLX5E_CACHE_SIZE - 1);
+	struct mlx5e_rq_stats *stats = rq->stats;
+
+	if (tail_next == cache->head) {
+		stats->cache_full++;
+		return false;
+	}
+
+	if (unlikely(mlx5e_page_is_reserved(dma_info->page))) {
+		stats->cache_waive++;
+		return false;
+	}
+
+	cache->page_cache[cache->tail] = *dma_info;
+	cache->tail = tail_next;
+	return true;
+}
+
+static inline bool mlx5e_rx_cache_get(struct mlx5e_rq *rq,
+				      struct mlx5e_dma_info *dma_info)
+{
+	struct mlx5e_page_cache *cache = &rq->page_cache;
+	struct mlx5e_rq_stats *stats = rq->stats;
+
+	if (unlikely(cache->head == cache->tail)) {
+		stats->cache_empty++;
+		return false;
+	}
+
+	if (page_ref_count(cache->page_cache[cache->head].page) != 1) {
+		stats->cache_busy++;
+		return false;
+	}
+
+	*dma_info = cache->page_cache[cache->head];
+	cache->head = (cache->head + 1) & (MLX5E_CACHE_SIZE - 1);
+	stats->cache_reuse++;
+
+	dma_sync_single_for_device(rq->pdev, dma_info->addr,
+				   PAGE_SIZE,
+				   DMA_FROM_DEVICE);
+	return true;
+}
+
+static inline int mlx5e_page_alloc_mapped(struct mlx5e_rq *rq,
+					  struct mlx5e_dma_info *dma_info)
+{
+	if (mlx5e_rx_cache_get(rq, dma_info))
+		return 0;
+
+	dma_info->page = page_pool_dev_alloc_pages(rq->page_pool);
+	if (unlikely(!dma_info->page))
+		return -ENOMEM;
+
+	dma_info->addr = dma_map_page(rq->pdev, dma_info->page, 0,
+				      PAGE_SIZE, rq->buff.map_dir);
+	if (unlikely(dma_mapping_error(rq->pdev, dma_info->addr))) {
+		put_page(dma_info->page);
+		dma_info->page = NULL;
+		return -ENOMEM;
+	}
+
+	return 0;
+}
+
+void mlx5e_page_dma_unmap(struct mlx5e_rq *rq, struct mlx5e_dma_info *dma_info)
+{
+	dma_unmap_page(rq->pdev, dma_info->addr, PAGE_SIZE, rq->buff.map_dir);
+}
+
+void mlx5e_page_release(struct mlx5e_rq *rq, struct mlx5e_dma_info *dma_info,
+			bool recycle)
+{
+	if (likely(recycle)) {
+		if (mlx5e_rx_cache_put(rq, dma_info))
+			return;
+
+		mlx5e_page_dma_unmap(rq, dma_info);
+		page_pool_recycle_direct(rq->page_pool, dma_info->page);
+	} else {
+		mlx5e_page_dma_unmap(rq, dma_info);
+		put_page(dma_info->page);
+	}
+}
+
+static inline int mlx5e_get_rx_frag(struct mlx5e_rq *rq,
+				    struct mlx5e_wqe_frag_info *frag)
+{
+	int err = 0;
+
+	if (!frag->offset)
+		/* On first frag (offset == 0), replenish page (dma_info actually).
+		 * Other frags that point to the same dma_info (with a different
+		 * offset) should just use the new one without replenishing again
+		 * by themselves.
+		 */
+		err = mlx5e_page_alloc_mapped(rq, frag->di);
+
+	return err;
+}
+
+static inline void mlx5e_put_rx_frag(struct mlx5e_rq *rq,
+				     struct mlx5e_wqe_frag_info *frag,
+				     bool recycle)
+{
+	if (frag->last_in_page)
+		mlx5e_page_release(rq, frag->di, recycle);
+}
+
+static inline struct mlx5e_wqe_frag_info *get_frag(struct mlx5e_rq *rq, u16 ix)
+{
+	return &rq->wqe.frags[ix << rq->wqe.info.log_num_frags];
+}
+
+static int mlx5e_alloc_rx_wqe(struct mlx5e_rq *rq, struct mlx5e_rx_wqe_cyc *wqe,
+			      u16 ix)
+{
+	struct mlx5e_wqe_frag_info *frag = get_frag(rq, ix);
+	int err;
+	int i;
+
+	for (i = 0; i < rq->wqe.info.num_frags; i++, frag++) {
+		err = mlx5e_get_rx_frag(rq, frag);
+		if (unlikely(err))
+			goto free_frags;
+
+		wqe->data[i].addr = cpu_to_be64(frag->di->addr +
+						frag->offset + rq->buff.headroom);
+	}
+
+	return 0;
+
+free_frags:
+	while (--i >= 0)
+		mlx5e_put_rx_frag(rq, --frag, true);
+
+	return err;
+}
+
+static inline void mlx5e_free_rx_wqe(struct mlx5e_rq *rq,
+				     struct mlx5e_wqe_frag_info *wi,
+				     bool recycle)
+{
+	int i;
+
+	for (i = 0; i < rq->wqe.info.num_frags; i++, wi++)
+		mlx5e_put_rx_frag(rq, wi, recycle);
+}
+
+void mlx5e_dealloc_rx_wqe(struct mlx5e_rq *rq, u16 ix)
+{
+	struct mlx5e_wqe_frag_info *wi = get_frag(rq, ix);
+
+	mlx5e_free_rx_wqe(rq, wi, false);
+}
+
+static int mlx5e_alloc_rx_wqes(struct mlx5e_rq *rq, u16 ix, u8 wqe_bulk)
+{
+	struct mlx5_wq_cyc *wq = &rq->wqe.wq;
+	int err;
+	int i;
+
+	for (i = 0; i < wqe_bulk; i++) {
+		struct mlx5e_rx_wqe_cyc *wqe = mlx5_wq_cyc_get_wqe(wq, ix + i);
+
+		err = mlx5e_alloc_rx_wqe(rq, wqe, ix + i);
+		if (unlikely(err))
+			goto free_wqes;
+	}
+
+	return 0;
+
+free_wqes:
+	while (--i >= 0)
+		mlx5e_dealloc_rx_wqe(rq, ix + i);
+
+	return err;
+}
+
+static inline void
+mlx5e_add_skb_frag(struct mlx5e_rq *rq, struct sk_buff *skb,
+		   struct mlx5e_dma_info *di, u32 frag_offset, u32 len,
+		   unsigned int truesize)
+{
+	dma_sync_single_for_cpu(rq->pdev,
+				di->addr + frag_offset,
+				len, DMA_FROM_DEVICE);
+	page_ref_inc(di->page);
+	skb_add_rx_frag(skb, skb_shinfo(skb)->nr_frags,
+			di->page, frag_offset, len, truesize);
+}
+
+static inline void
+mlx5e_copy_skb_header(struct device *pdev, struct sk_buff *skb,
+		      struct mlx5e_dma_info *dma_info,
+		      int offset_from, int offset_to, u32 headlen)
+{
+	const void *from = page_address(dma_info->page) + offset_from;
+	/* Aligning len to sizeof(long) optimizes memcpy performance */
+	unsigned int len = ALIGN(headlen, sizeof(long));
+
+	dma_sync_single_for_cpu(pdev, dma_info->addr + offset_from, len,
+				DMA_FROM_DEVICE);
+	skb_copy_to_linear_data_offset(skb, offset_to, from, len);
+}
+
+static inline void
+mlx5e_copy_skb_header_mpwqe(struct device *pdev,
+			    struct sk_buff *skb,
+			    struct mlx5e_dma_info *dma_info,
+			    u32 offset, u32 headlen)
+{
+	u16 headlen_pg = min_t(u32, headlen, PAGE_SIZE - offset);
+
+	mlx5e_copy_skb_header(pdev, skb, dma_info, offset, 0, headlen_pg);
+
+	if (unlikely(offset + headlen > PAGE_SIZE)) {
+		dma_info++;
+		mlx5e_copy_skb_header(pdev, skb, dma_info, 0, headlen_pg,
+				      headlen - headlen_pg);
+	}
+}
+
+static void
+mlx5e_free_rx_mpwqe(struct mlx5e_rq *rq, struct mlx5e_mpw_info *wi, bool recycle)
+{
+	const bool no_xdp_xmit =
+		bitmap_empty(wi->xdp_xmit_bitmap, MLX5_MPWRQ_PAGES_PER_WQE);
+	struct mlx5e_dma_info *dma_info = wi->umr.dma_info;
+	int i;
+
+	for (i = 0; i < MLX5_MPWRQ_PAGES_PER_WQE; i++)
+		if (no_xdp_xmit || !test_bit(i, wi->xdp_xmit_bitmap))
+			mlx5e_page_release(rq, &dma_info[i], recycle);
+}
+
+static void mlx5e_post_rx_mpwqe(struct mlx5e_rq *rq)
+{
+	struct mlx5_wq_ll *wq = &rq->mpwqe.wq;
+	struct mlx5e_rx_wqe_ll *wqe = mlx5_wq_ll_get_wqe(wq, wq->head);
+
+	rq->mpwqe.umr_in_progress = false;
+
+	mlx5_wq_ll_push(wq, be16_to_cpu(wqe->next.next_wqe_index));
+
+	/* ensure wqes are visible to device before updating doorbell record */
+	dma_wmb();
+
+	mlx5_wq_ll_update_db_record(wq);
+}
+
+static inline u16 mlx5e_icosq_wrap_cnt(struct mlx5e_icosq *sq)
+{
+	return sq->pc >> MLX5E_PARAMS_MINIMUM_LOG_SQ_SIZE;
+}
+
+static inline void mlx5e_fill_icosq_frag_edge(struct mlx5e_icosq *sq,
+					      struct mlx5_wq_cyc *wq,
+					      u16 pi, u16 nnops)
+{
+	struct mlx5e_sq_wqe_info *edge_wi, *wi = &sq->db.ico_wqe[pi];
+
+	edge_wi = wi + nnops;
+
+	/* fill sq frag edge with nops to avoid wqe wrapping two pages */
+	for (; wi < edge_wi; wi++) {
+		wi->opcode = MLX5_OPCODE_NOP;
+		mlx5e_post_nop(wq, sq->sqn, &sq->pc);
+	}
+}
+
+static int mlx5e_alloc_rx_mpwqe(struct mlx5e_rq *rq, u16 ix)
+{
+	struct mlx5e_mpw_info *wi = &rq->mpwqe.info[ix];
+	struct mlx5e_dma_info *dma_info = &wi->umr.dma_info[0];
+	struct mlx5e_icosq *sq = &rq->channel->icosq;
+	struct mlx5_wq_cyc *wq = &sq->wq;
+	struct mlx5e_umr_wqe *umr_wqe;
+	u16 xlt_offset = ix << (MLX5E_LOG_ALIGNED_MPWQE_PPW - 1);
+	u16 pi, contig_wqebbs_room;
+	int err;
+	int i;
+
+	pi = mlx5_wq_cyc_ctr2ix(wq, sq->pc);
+	contig_wqebbs_room = mlx5_wq_cyc_get_contig_wqebbs(wq, pi);
+	if (unlikely(contig_wqebbs_room < MLX5E_UMR_WQEBBS)) {
+		mlx5e_fill_icosq_frag_edge(sq, wq, pi, contig_wqebbs_room);
+		pi = mlx5_wq_cyc_ctr2ix(wq, sq->pc);
+	}
+
+	umr_wqe = mlx5_wq_cyc_get_wqe(wq, pi);
+	if (unlikely(mlx5e_icosq_wrap_cnt(sq) < 2))
+		memcpy(umr_wqe, &rq->mpwqe.umr_wqe,
+		       offsetof(struct mlx5e_umr_wqe, inline_mtts));
+
+	for (i = 0; i < MLX5_MPWRQ_PAGES_PER_WQE; i++, dma_info++) {
+		err = mlx5e_page_alloc_mapped(rq, dma_info);
+		if (unlikely(err))
+			goto err_unmap;
+		umr_wqe->inline_mtts[i].ptag = cpu_to_be64(dma_info->addr | MLX5_EN_WR);
+	}
+
+	bitmap_zero(wi->xdp_xmit_bitmap, MLX5_MPWRQ_PAGES_PER_WQE);
+	wi->consumed_strides = 0;
+
+	rq->mpwqe.umr_in_progress = true;
+
+	umr_wqe->ctrl.opmod_idx_opcode =
+		cpu_to_be32((sq->pc << MLX5_WQE_CTRL_WQE_INDEX_SHIFT) |
+			    MLX5_OPCODE_UMR);
+	umr_wqe->uctrl.xlt_offset = cpu_to_be16(xlt_offset);
+
+	sq->db.ico_wqe[pi].opcode = MLX5_OPCODE_UMR;
+	sq->pc += MLX5E_UMR_WQEBBS;
+	mlx5e_notify_hw(wq, sq->pc, sq->uar_map, &umr_wqe->ctrl);
+
+	return 0;
+
+err_unmap:
+	while (--i >= 0) {
+		dma_info--;
+		mlx5e_page_release(rq, dma_info, true);
+	}
+	rq->stats->buff_alloc_err++;
+
+	return err;
+}
+
+void mlx5e_dealloc_rx_mpwqe(struct mlx5e_rq *rq, u16 ix)
+{
+	struct mlx5e_mpw_info *wi = &rq->mpwqe.info[ix];
+	/* Don't recycle, this function is called on rq/netdev close */
+	mlx5e_free_rx_mpwqe(rq, wi, false);
+}
+
+bool mlx5e_post_rx_wqes(struct mlx5e_rq *rq)
+{
+	struct mlx5_wq_cyc *wq = &rq->wqe.wq;
+	u8 wqe_bulk;
+	int err;
+
+	if (unlikely(!test_bit(MLX5E_RQ_STATE_ENABLED, &rq->state)))
+		return false;
+
+	wqe_bulk = rq->wqe.info.wqe_bulk;
+
+	if (mlx5_wq_cyc_missing(wq) < wqe_bulk)
+		return false;
+
+	do {
+		u16 head = mlx5_wq_cyc_get_head(wq);
+
+		err = mlx5e_alloc_rx_wqes(rq, head, wqe_bulk);
+		if (unlikely(err)) {
+			rq->stats->buff_alloc_err++;
+			break;
+		}
+
+		mlx5_wq_cyc_push_n(wq, wqe_bulk);
+	} while (mlx5_wq_cyc_missing(wq) >= wqe_bulk);
+
+	/* ensure wqes are visible to device before updating doorbell record */
+	dma_wmb();
+
+	mlx5_wq_cyc_update_db_record(wq);
+
+	return !!err;
+}
+
+static inline void mlx5e_poll_ico_single_cqe(struct mlx5e_cq *cq,
+					     struct mlx5e_icosq *sq,
+					     struct mlx5e_rq *rq,
+					     struct mlx5_cqe64 *cqe)
+{
+	struct mlx5_wq_cyc *wq = &sq->wq;
+	u16 ci = mlx5_wq_cyc_ctr2ix(wq, be16_to_cpu(cqe->wqe_counter));
+	struct mlx5e_sq_wqe_info *icowi = &sq->db.ico_wqe[ci];
+
+	mlx5_cqwq_pop(&cq->wq);
+
+	if (unlikely((cqe->op_own >> 4) != MLX5_CQE_REQ)) {
+		netdev_WARN_ONCE(cq->channel->netdev,
+				 "Bad OP in ICOSQ CQE: 0x%x\n", cqe->op_own);
+		return;
+	}
+
+	if (likely(icowi->opcode == MLX5_OPCODE_UMR)) {
+		mlx5e_post_rx_mpwqe(rq);
+		return;
+	}
+
+	if (unlikely(icowi->opcode != MLX5_OPCODE_NOP))
+		netdev_WARN_ONCE(cq->channel->netdev,
+				 "Bad OPCODE in ICOSQ WQE info: 0x%x\n", icowi->opcode);
+}
+
+static void mlx5e_poll_ico_cq(struct mlx5e_cq *cq, struct mlx5e_rq *rq)
+{
+	struct mlx5e_icosq *sq = container_of(cq, struct mlx5e_icosq, cq);
+	struct mlx5_cqe64 *cqe;
+
+	if (unlikely(!test_bit(MLX5E_SQ_STATE_ENABLED, &sq->state)))
+		return;
+
+	cqe = mlx5_cqwq_get_cqe(&cq->wq);
+	if (likely(!cqe))
+		return;
+
+	/* by design, there's only a single cqe */
+	mlx5e_poll_ico_single_cqe(cq, sq, rq, cqe);
+
+	mlx5_cqwq_update_db_record(&cq->wq);
+}
+
+bool mlx5e_post_rx_mpwqes(struct mlx5e_rq *rq)
+{
+	struct mlx5_wq_ll *wq = &rq->mpwqe.wq;
+
+	if (unlikely(!test_bit(MLX5E_RQ_STATE_ENABLED, &rq->state)))
+		return false;
+
+	mlx5e_poll_ico_cq(&rq->channel->icosq.cq, rq);
+
+	if (mlx5_wq_ll_is_full(wq))
+		return false;
+
+	if (!rq->mpwqe.umr_in_progress)
+		mlx5e_alloc_rx_mpwqe(rq, wq->head);
+	else
+		rq->stats->congst_umr += mlx5_wq_ll_missing(wq) > 2;
+
+	return false;
+}
+
+static void mlx5e_lro_update_tcp_hdr(struct mlx5_cqe64 *cqe, struct tcphdr *tcp)
+{
+	u8 l4_hdr_type = get_cqe_l4_hdr_type(cqe);
+	u8 tcp_ack     = (l4_hdr_type == CQE_L4_HDR_TYPE_TCP_ACK_NO_DATA) ||
+			 (l4_hdr_type == CQE_L4_HDR_TYPE_TCP_ACK_AND_DATA);
+
+	tcp->check                      = 0;
+	tcp->psh                        = get_cqe_lro_tcppsh(cqe);
+
+	if (tcp_ack) {
+		tcp->ack                = 1;
+		tcp->ack_seq            = cqe->lro_ack_seq_num;
+		tcp->window             = cqe->lro_tcp_win;
+	}
+}
+
+static void mlx5e_lro_update_hdr(struct sk_buff *skb, struct mlx5_cqe64 *cqe,
+				 u32 cqe_bcnt)
+{
+	struct ethhdr	*eth = (struct ethhdr *)(skb->data);
+	struct tcphdr	*tcp;
+	int network_depth = 0;
+	__wsum check;
+	__be16 proto;
+	u16 tot_len;
+	void *ip_p;
+
+	proto = __vlan_get_protocol(skb, eth->h_proto, &network_depth);
+
+	tot_len = cqe_bcnt - network_depth;
+	ip_p = skb->data + network_depth;
+
+	if (proto == htons(ETH_P_IP)) {
+		struct iphdr *ipv4 = ip_p;
+
+		tcp = ip_p + sizeof(struct iphdr);
+		skb_shinfo(skb)->gso_type = SKB_GSO_TCPV4;
+
+		ipv4->ttl               = cqe->lro_min_ttl;
+		ipv4->tot_len           = cpu_to_be16(tot_len);
+		ipv4->check             = 0;
+		ipv4->check             = ip_fast_csum((unsigned char *)ipv4,
+						       ipv4->ihl);
+
+		mlx5e_lro_update_tcp_hdr(cqe, tcp);
+		check = csum_partial(tcp, tcp->doff * 4,
+				     csum_unfold((__force __sum16)cqe->check_sum));
+		/* Almost done, don't forget the pseudo header */
+		tcp->check = csum_tcpudp_magic(ipv4->saddr, ipv4->daddr,
+					       tot_len - sizeof(struct iphdr),
+					       IPPROTO_TCP, check);
+	} else {
+		u16 payload_len = tot_len - sizeof(struct ipv6hdr);
+		struct ipv6hdr *ipv6 = ip_p;
+
+		tcp = ip_p + sizeof(struct ipv6hdr);
+		skb_shinfo(skb)->gso_type = SKB_GSO_TCPV6;
+
+		ipv6->hop_limit         = cqe->lro_min_ttl;
+		ipv6->payload_len       = cpu_to_be16(payload_len);
+
+		mlx5e_lro_update_tcp_hdr(cqe, tcp);
+		check = csum_partial(tcp, tcp->doff * 4,
+				     csum_unfold((__force __sum16)cqe->check_sum));
+		/* Almost done, don't forget the pseudo header */
+		tcp->check = csum_ipv6_magic(&ipv6->saddr, &ipv6->daddr, payload_len,
+					     IPPROTO_TCP, check);
+	}
+}
+
+static inline void mlx5e_skb_set_hash(struct mlx5_cqe64 *cqe,
+				      struct sk_buff *skb)
+{
+	u8 cht = cqe->rss_hash_type;
+	int ht = (cht & CQE_RSS_HTYPE_L4) ? PKT_HASH_TYPE_L4 :
+		 (cht & CQE_RSS_HTYPE_IP) ? PKT_HASH_TYPE_L3 :
+					    PKT_HASH_TYPE_NONE;
+	skb_set_hash(skb, be32_to_cpu(cqe->rss_hash_result), ht);
+}
+
+static inline bool is_last_ethertype_ip(struct sk_buff *skb, int *network_depth,
+					__be16 *proto)
+{
+	*proto = ((struct ethhdr *)skb->data)->h_proto;
+	*proto = __vlan_get_protocol(skb, *proto, network_depth);
+
+	if (*proto == htons(ETH_P_IP))
+		return pskb_may_pull(skb, *network_depth + sizeof(struct iphdr));
+
+	if (*proto == htons(ETH_P_IPV6))
+		return pskb_may_pull(skb, *network_depth + sizeof(struct ipv6hdr));
+
+	return false;
+}
+
+static inline void mlx5e_enable_ecn(struct mlx5e_rq *rq, struct sk_buff *skb)
+{
+	int network_depth = 0;
+	__be16 proto;
+	void *ip;
+	int rc;
+
+	if (unlikely(!is_last_ethertype_ip(skb, &network_depth, &proto)))
+		return;
+
+	ip = skb->data + network_depth;
+	rc = ((proto == htons(ETH_P_IP)) ? IP_ECN_set_ce((struct iphdr *)ip) :
+					 IP6_ECN_set_ce(skb, (struct ipv6hdr *)ip));
+
+	rq->stats->ecn_mark += !!rc;
+}
+
+static u8 get_ip_proto(struct sk_buff *skb, int network_depth, __be16 proto)
+{
+	void *ip_p = skb->data + network_depth;
+
+	return (proto == htons(ETH_P_IP)) ? ((struct iphdr *)ip_p)->protocol :
+					    ((struct ipv6hdr *)ip_p)->nexthdr;
+}
+
+#define short_frame(size) ((size) <= ETH_ZLEN + ETH_FCS_LEN)
+
+#define MAX_PADDING 8
+
+static void
+tail_padding_csum_slow(struct sk_buff *skb, int offset, int len,
+		       struct mlx5e_rq_stats *stats)
+{
+	stats->csum_complete_tail_slow++;
+	skb->csum = csum_block_add(skb->csum,
+				   skb_checksum(skb, offset, len, 0),
+				   offset);
+}
+
+static void
+tail_padding_csum(struct sk_buff *skb, int offset,
+		  struct mlx5e_rq_stats *stats)
+{
+	u8 tail_padding[MAX_PADDING];
+	int len = skb->len - offset;
+	void *tail;
+
+	if (unlikely(len > MAX_PADDING)) {
+		tail_padding_csum_slow(skb, offset, len, stats);
+		return;
+	}
+
+	tail = skb_header_pointer(skb, offset, len, tail_padding);
+	if (unlikely(!tail)) {
+		tail_padding_csum_slow(skb, offset, len, stats);
+		return;
+	}
+
+	stats->csum_complete_tail++;
+	skb->csum = csum_block_add(skb->csum, csum_partial(tail, len, 0), offset);
+}
+
+static void
+mlx5e_skb_padding_csum(struct sk_buff *skb, int network_depth, __be16 proto,
+		       struct mlx5e_rq_stats *stats)
+{
+	struct ipv6hdr *ip6;
+	struct iphdr   *ip4;
+	int pkt_len;
+
+	switch (proto) {
+	case htons(ETH_P_IP):
+		ip4 = (struct iphdr *)(skb->data + network_depth);
+		pkt_len = network_depth + ntohs(ip4->tot_len);
+		break;
+	case htons(ETH_P_IPV6):
+		ip6 = (struct ipv6hdr *)(skb->data + network_depth);
+		pkt_len = network_depth + sizeof(*ip6) + ntohs(ip6->payload_len);
+		break;
+	default:
+		return;
+	}
+
+	if (likely(pkt_len >= skb->len))
+		return;
+
+	tail_padding_csum(skb, pkt_len, stats);
+}
+
+static inline void mlx5e_handle_csum(struct net_device *netdev,
+				     struct mlx5_cqe64 *cqe,
+				     struct mlx5e_rq *rq,
+				     struct sk_buff *skb,
+				     bool   lro)
+{
+	struct mlx5e_rq_stats *stats = rq->stats;
+	int network_depth = 0;
+	__be16 proto;
+
+	if (unlikely(!(netdev->features & NETIF_F_RXCSUM)))
+		goto csum_none;
+
+	if (lro) {
+		skb->ip_summed = CHECKSUM_UNNECESSARY;
+		stats->csum_unnecessary++;
+		return;
+	}
+
+	/* True when explicitly set via priv flag, or XDP prog is loaded */
+	if (test_bit(MLX5E_RQ_STATE_NO_CSUM_COMPLETE, &rq->state))
+		goto csum_unnecessary;
+
+	/* 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 be
+	 * CHECKSUM_UNNECESSARY even if they are not padded.
+	 */
+	if (short_frame(skb->len))
+		goto csum_unnecessary;
+
+	if (likely(is_last_ethertype_ip(skb, &network_depth, &proto))) {
+		if (unlikely(get_ip_proto(skb, network_depth, proto) == IPPROTO_SCTP))
+			goto csum_unnecessary;
+
+		skb->ip_summed = CHECKSUM_COMPLETE;
+		skb->csum = csum_unfold((__force __sum16)cqe->check_sum);
+		if (network_depth > ETH_HLEN)
+			/* CQE csum is calculated from the IP header and does
+			 * not cover VLAN headers (if present). This will add
+			 * the checksum manually.
+			 */
+			skb->csum = csum_partial(skb->data + ETH_HLEN,
+						 network_depth - ETH_HLEN,
+						 skb->csum);
+
+		mlx5e_skb_padding_csum(skb, network_depth, proto, stats);
+		stats->csum_complete++;
+		return;
+	}
+
+csum_unnecessary:
+	if (likely((cqe->hds_ip_ext & CQE_L3_OK) &&
+		   (cqe->hds_ip_ext & CQE_L4_OK))) {
+		skb->ip_summed = CHECKSUM_UNNECESSARY;
+		if (cqe_is_tunneled(cqe)) {
+			skb->csum_level = 1;
+			skb->encapsulation = 1;
+			stats->csum_unnecessary_inner++;
+			return;
+		}
+		stats->csum_unnecessary++;
+		return;
+	}
+csum_none:
+	skb->ip_summed = CHECKSUM_NONE;
+	stats->csum_none++;
+}
+
+#define MLX5E_CE_BIT_MASK 0x80
+
+static inline void mlx5e_build_rx_skb(struct mlx5_cqe64 *cqe,
+				      u32 cqe_bcnt,
+				      struct mlx5e_rq *rq,
+				      struct sk_buff *skb)
+{
+	u8 lro_num_seg = be32_to_cpu(cqe->srqn) >> 24;
+	struct mlx5e_rq_stats *stats = rq->stats;
+	struct net_device *netdev = rq->netdev;
+
+	skb->mac_len = ETH_HLEN;
+
+#ifdef CONFIG_MLX5_EN_TLS
+	mlx5e_tls_handle_rx_skb(netdev, skb, &cqe_bcnt);
+#endif
+
+	if (lro_num_seg > 1) {
+		mlx5e_lro_update_hdr(skb, cqe, cqe_bcnt);
+		skb_shinfo(skb)->gso_size = DIV_ROUND_UP(cqe_bcnt, lro_num_seg);
+		/* Subtract one since we already counted this as one
+		 * "regular" packet in mlx5e_complete_rx_cqe()
+		 */
+		stats->packets += lro_num_seg - 1;
+		stats->lro_packets++;
+		stats->lro_bytes += cqe_bcnt;
+	}
+
+	if (unlikely(mlx5e_rx_hw_stamp(rq->tstamp)))
+		skb_hwtstamps(skb)->hwtstamp =
+				mlx5_timecounter_cyc2time(rq->clock, get_cqe_ts(cqe));
+
+	skb_record_rx_queue(skb, rq->ix);
+
+	if (likely(netdev->features & NETIF_F_RXHASH))
+		mlx5e_skb_set_hash(cqe, skb);
+
+	if (cqe_has_vlan(cqe)) {
+		__vlan_hwaccel_put_tag(skb, htons(ETH_P_8021Q),
+				       be16_to_cpu(cqe->vlan_info));
+		stats->removed_vlan_packets++;
+	}
+
+	skb->mark = be32_to_cpu(cqe->sop_drop_qpn) & MLX5E_TC_FLOW_ID_MASK;
+
+	mlx5e_handle_csum(netdev, cqe, rq, skb, !!lro_num_seg);
+	/* checking CE bit in cqe - MSB in ml_path field */
+	if (unlikely(cqe->ml_path & MLX5E_CE_BIT_MASK))
+		mlx5e_enable_ecn(rq, skb);
+
+	skb->protocol = eth_type_trans(skb, netdev);
+}
+
+static inline void mlx5e_complete_rx_cqe(struct mlx5e_rq *rq,
+					 struct mlx5_cqe64 *cqe,
+					 u32 cqe_bcnt,
+					 struct sk_buff *skb)
+{
+	struct mlx5e_rq_stats *stats = rq->stats;
+
+	stats->packets++;
+	stats->bytes += cqe_bcnt;
+	mlx5e_build_rx_skb(cqe, cqe_bcnt, rq, skb);
+}
+
+static inline
+struct sk_buff *mlx5e_build_linear_skb(struct mlx5e_rq *rq, void *va,
+				       u32 frag_size, u16 headroom,
+				       u32 cqe_bcnt)
+{
+	struct sk_buff *skb = build_skb(va, frag_size);
+
+	if (unlikely(!skb)) {
+		rq->stats->buff_alloc_err++;
+		return NULL;
+	}
+
+	skb_reserve(skb, headroom);
+	skb_put(skb, cqe_bcnt);
+
+	return skb;
+}
+
+struct sk_buff *
+mlx5e_skb_from_cqe_linear(struct mlx5e_rq *rq, struct mlx5_cqe64 *cqe,
+			  struct mlx5e_wqe_frag_info *wi, u32 cqe_bcnt)
+{
+	struct mlx5e_dma_info *di = wi->di;
+	u16 rx_headroom = rq->buff.headroom;
+	struct sk_buff *skb;
+	void *va, *data;
+	bool consumed;
+	u32 frag_size;
+
+	va             = page_address(di->page) + wi->offset;
+	data           = va + rx_headroom;
+	frag_size      = MLX5_SKB_FRAG_SZ(rx_headroom + cqe_bcnt);
+
+	dma_sync_single_range_for_cpu(rq->pdev, di->addr, wi->offset,
+				      frag_size, DMA_FROM_DEVICE);
+	prefetchw(va); /* xdp_frame data area */
+	prefetch(data);
+
+	if (unlikely((cqe->op_own >> 4) != MLX5_CQE_RESP_SEND)) {
+		rq->stats->wqe_err++;
+		return NULL;
+	}
+
+	rcu_read_lock();
+	consumed = mlx5e_xdp_handle(rq, di, va, &rx_headroom, &cqe_bcnt);
+	rcu_read_unlock();
+	if (consumed)
+		return NULL; /* page/packet was consumed by XDP */
+
+	skb = mlx5e_build_linear_skb(rq, va, frag_size, rx_headroom, cqe_bcnt);
+	if (unlikely(!skb))
+		return NULL;
+
+	/* queue up for recycling/reuse */
+	page_ref_inc(di->page);
+
+	return skb;
+}
+
+struct sk_buff *
+mlx5e_skb_from_cqe_nonlinear(struct mlx5e_rq *rq, struct mlx5_cqe64 *cqe,
+			     struct mlx5e_wqe_frag_info *wi, u32 cqe_bcnt)
+{
+	struct mlx5e_rq_frag_info *frag_info = &rq->wqe.info.arr[0];
+	struct mlx5e_wqe_frag_info *head_wi = wi;
+	u16 headlen      = min_t(u32, MLX5E_RX_MAX_HEAD, cqe_bcnt);
+	u16 frag_headlen = headlen;
+	u16 byte_cnt     = cqe_bcnt - headlen;
+	struct sk_buff *skb;
+
+	if (unlikely((cqe->op_own >> 4) != MLX5_CQE_RESP_SEND)) {
+		rq->stats->wqe_err++;
+		return NULL;
+	}
+
+	/* XDP is not supported in this configuration, as incoming packets
+	 * might spread among multiple pages.
+	 */
+	skb = napi_alloc_skb(rq->cq.napi,
+			     ALIGN(MLX5E_RX_MAX_HEAD, sizeof(long)));
+	if (unlikely(!skb)) {
+		rq->stats->buff_alloc_err++;
+		return NULL;
+	}
+
+	prefetchw(skb->data);
+
+	while (byte_cnt) {
+		u16 frag_consumed_bytes =
+			min_t(u16, frag_info->frag_size - frag_headlen, byte_cnt);
+
+		mlx5e_add_skb_frag(rq, skb, wi->di, wi->offset + frag_headlen,
+				   frag_consumed_bytes, frag_info->frag_stride);
+		byte_cnt -= frag_consumed_bytes;
+		frag_headlen = 0;
+		frag_info++;
+		wi++;
+	}
+
+	/* copy header */
+	mlx5e_copy_skb_header(rq->pdev, skb, head_wi->di, head_wi->offset,
+			      0, headlen);
+	/* skb linear part was allocated with headlen and aligned to long */
+	skb->tail += headlen;
+	skb->len  += headlen;
+
+	return skb;
+}
+
+void mlx5e_handle_rx_cqe(struct mlx5e_rq *rq, struct mlx5_cqe64 *cqe)
+{
+	struct mlx5_wq_cyc *wq = &rq->wqe.wq;
+	struct mlx5e_wqe_frag_info *wi;
+	struct sk_buff *skb;
+	u32 cqe_bcnt;
+	u16 ci;
+
+	ci       = mlx5_wq_cyc_ctr2ix(wq, be16_to_cpu(cqe->wqe_counter));
+	wi       = get_frag(rq, ci);
+	cqe_bcnt = be32_to_cpu(cqe->byte_cnt);
+
+	skb = rq->wqe.skb_from_cqe(rq, cqe, wi, cqe_bcnt);
+	if (!skb) {
+		/* probably for XDP */
+		if (__test_and_clear_bit(MLX5E_RQ_FLAG_XDP_XMIT, rq->flags)) {
+			/* do not return page to cache,
+			 * it will be returned on XDP_TX completion.
+			 */
+			goto wq_cyc_pop;
+		}
+		goto free_wqe;
+	}
+
+	mlx5e_complete_rx_cqe(rq, cqe, cqe_bcnt, skb);
+	napi_gro_receive(rq->cq.napi, skb);
+
+free_wqe:
+	mlx5e_free_rx_wqe(rq, wi, true);
+wq_cyc_pop:
+	mlx5_wq_cyc_pop(wq);
+}
+
+#ifdef CONFIG_MLX5_ESWITCH
+void mlx5e_handle_rx_cqe_rep(struct mlx5e_rq *rq, struct mlx5_cqe64 *cqe)
+{
+	struct net_device *netdev = rq->netdev;
+	struct mlx5e_priv *priv = netdev_priv(netdev);
+	struct mlx5e_rep_priv *rpriv  = priv->ppriv;
+	struct mlx5_eswitch_rep *rep = rpriv->rep;
+	struct mlx5_wq_cyc *wq = &rq->wqe.wq;
+	struct mlx5e_wqe_frag_info *wi;
+	struct sk_buff *skb;
+	u32 cqe_bcnt;
+	u16 ci;
+
+	ci       = mlx5_wq_cyc_ctr2ix(wq, be16_to_cpu(cqe->wqe_counter));
+	wi       = get_frag(rq, ci);
+	cqe_bcnt = be32_to_cpu(cqe->byte_cnt);
+
+	skb = rq->wqe.skb_from_cqe(rq, cqe, wi, cqe_bcnt);
+	if (!skb) {
+		/* probably for XDP */
+		if (__test_and_clear_bit(MLX5E_RQ_FLAG_XDP_XMIT, rq->flags)) {
+			/* do not return page to cache,
+			 * it will be returned on XDP_TX completion.
+			 */
+			goto wq_cyc_pop;
+		}
+		goto free_wqe;
+	}
+
+	mlx5e_complete_rx_cqe(rq, cqe, cqe_bcnt, skb);
+
+	if (rep->vlan && skb_vlan_tag_present(skb))
+		skb_vlan_pop(skb);
+
+	napi_gro_receive(rq->cq.napi, skb);
+
+free_wqe:
+	mlx5e_free_rx_wqe(rq, wi, true);
+wq_cyc_pop:
+	mlx5_wq_cyc_pop(wq);
+}
+#endif
+
+struct sk_buff *
+mlx5e_skb_from_cqe_mpwrq_nonlinear(struct mlx5e_rq *rq, struct mlx5e_mpw_info *wi,
+				   u16 cqe_bcnt, u32 head_offset, u32 page_idx)
+{
+	u16 headlen = min_t(u16, MLX5E_RX_MAX_HEAD, cqe_bcnt);
+	struct mlx5e_dma_info *di = &wi->umr.dma_info[page_idx];
+	u32 frag_offset    = head_offset + headlen;
+	u32 byte_cnt       = cqe_bcnt - headlen;
+	struct mlx5e_dma_info *head_di = di;
+	struct sk_buff *skb;
+
+	skb = napi_alloc_skb(rq->cq.napi,
+			     ALIGN(MLX5E_RX_MAX_HEAD, sizeof(long)));
+	if (unlikely(!skb)) {
+		rq->stats->buff_alloc_err++;
+		return NULL;
+	}
+
+	prefetchw(skb->data);
+
+	if (unlikely(frag_offset >= PAGE_SIZE)) {
+		di++;
+		frag_offset -= PAGE_SIZE;
+	}
+
+	while (byte_cnt) {
+		u32 pg_consumed_bytes =
+			min_t(u32, PAGE_SIZE - frag_offset, byte_cnt);
+		unsigned int truesize =
+			ALIGN(pg_consumed_bytes, BIT(rq->mpwqe.log_stride_sz));
+
+		mlx5e_add_skb_frag(rq, skb, di, frag_offset,
+				   pg_consumed_bytes, truesize);
+		byte_cnt -= pg_consumed_bytes;
+		frag_offset = 0;
+		di++;
+	}
+	/* copy header */
+	mlx5e_copy_skb_header_mpwqe(rq->pdev, skb, head_di,
+				    head_offset, headlen);
+	/* skb linear part was allocated with headlen and aligned to long */
+	skb->tail += headlen;
+	skb->len  += headlen;
+
+	return skb;
+}
+
+struct sk_buff *
+mlx5e_skb_from_cqe_mpwrq_linear(struct mlx5e_rq *rq, struct mlx5e_mpw_info *wi,
+				u16 cqe_bcnt, u32 head_offset, u32 page_idx)
+{
+	struct mlx5e_dma_info *di = &wi->umr.dma_info[page_idx];
+	u16 rx_headroom = rq->buff.headroom;
+	u32 cqe_bcnt32 = cqe_bcnt;
+	struct sk_buff *skb;
+	void *va, *data;
+	u32 frag_size;
+	bool consumed;
+
+	/* Check packet size. Note LRO doesn't use linear SKB */
+	if (unlikely(cqe_bcnt > rq->hw_mtu)) {
+		rq->stats->oversize_pkts_sw_drop++;
+		return NULL;
+	}
+
+	va             = page_address(di->page) + head_offset;
+	data           = va + rx_headroom;
+	frag_size      = MLX5_SKB_FRAG_SZ(rx_headroom + cqe_bcnt32);
+
+	dma_sync_single_range_for_cpu(rq->pdev, di->addr, head_offset,
+				      frag_size, DMA_FROM_DEVICE);
+	prefetchw(va); /* xdp_frame data area */
+	prefetch(data);
+
+	rcu_read_lock();
+	consumed = mlx5e_xdp_handle(rq, di, va, &rx_headroom, &cqe_bcnt32);
+	rcu_read_unlock();
+	if (consumed) {
+		if (__test_and_clear_bit(MLX5E_RQ_FLAG_XDP_XMIT, rq->flags))
+			__set_bit(page_idx, wi->xdp_xmit_bitmap); /* non-atomic */
+		return NULL; /* page/packet was consumed by XDP */
+	}
+
+	skb = mlx5e_build_linear_skb(rq, va, frag_size, rx_headroom, cqe_bcnt32);
+	if (unlikely(!skb))
+		return NULL;
+
+	/* queue up for recycling/reuse */
+	page_ref_inc(di->page);
+
+	return skb;
+}
+
+void mlx5e_handle_rx_cqe_mpwrq(struct mlx5e_rq *rq, struct mlx5_cqe64 *cqe)
+{
+	u16 cstrides       = mpwrq_get_cqe_consumed_strides(cqe);
+	u16 wqe_id         = be16_to_cpu(cqe->wqe_id);
+	struct mlx5e_mpw_info *wi = &rq->mpwqe.info[wqe_id];
+	u16 stride_ix      = mpwrq_get_cqe_stride_index(cqe);
+	u32 wqe_offset     = stride_ix << rq->mpwqe.log_stride_sz;
+	u32 head_offset    = wqe_offset & (PAGE_SIZE - 1);
+	u32 page_idx       = wqe_offset >> PAGE_SHIFT;
+	struct mlx5e_rx_wqe_ll *wqe;
+	struct mlx5_wq_ll *wq;
+	struct sk_buff *skb;
+	u16 cqe_bcnt;
+
+	wi->consumed_strides += cstrides;
+
+	if (unlikely((cqe->op_own >> 4) != MLX5_CQE_RESP_SEND)) {
+		rq->stats->wqe_err++;
+		goto mpwrq_cqe_out;
+	}
+
+	if (unlikely(mpwrq_is_filler_cqe(cqe))) {
+		struct mlx5e_rq_stats *stats = rq->stats;
+
+		stats->mpwqe_filler_cqes++;
+		stats->mpwqe_filler_strides += cstrides;
+		goto mpwrq_cqe_out;
+	}
+
+	cqe_bcnt = mpwrq_get_cqe_byte_cnt(cqe);
+
+	skb = rq->mpwqe.skb_from_cqe_mpwrq(rq, wi, cqe_bcnt, head_offset,
+					   page_idx);
+	if (!skb)
+		goto mpwrq_cqe_out;
+
+	mlx5e_complete_rx_cqe(rq, cqe, cqe_bcnt, skb);
+	napi_gro_receive(rq->cq.napi, skb);
+
+mpwrq_cqe_out:
+	if (likely(wi->consumed_strides < rq->mpwqe.num_strides))
+		return;
+
+	wq  = &rq->mpwqe.wq;
+	wqe = mlx5_wq_ll_get_wqe(wq, wqe_id);
+	mlx5e_free_rx_mpwqe(rq, wi, true);
+	mlx5_wq_ll_pop(wq, cqe->wqe_id, &wqe->next.next_wqe_index);
+}
+
+int mlx5e_poll_rx_cq(struct mlx5e_cq *cq, int budget)
+{
+	struct mlx5e_rq *rq = container_of(cq, struct mlx5e_rq, cq);
+	struct mlx5e_xdpsq *xdpsq = &rq->xdpsq;
+	struct mlx5_cqe64 *cqe;
+	int work_done = 0;
+
+	if (unlikely(!test_bit(MLX5E_RQ_STATE_ENABLED, &rq->state)))
+		return 0;
+
+	if (cq->decmprs_left) {
+		work_done += mlx5e_decompress_cqes_cont(rq, cq, 0, budget);
+		if (cq->decmprs_left || work_done >= budget)
+			goto out;
+	}
+
+	cqe = mlx5_cqwq_get_cqe(&cq->wq);
+	if (!cqe) {
+		if (unlikely(work_done))
+			goto out;
+		return 0;
+	}
+
+	do {
+		if (mlx5_get_cqe_format(cqe) == MLX5_COMPRESSED) {
+			work_done +=
+				mlx5e_decompress_cqes_start(rq, cq,
+							    budget - work_done);
+			continue;
+		}
+
+		mlx5_cqwq_pop(&cq->wq);
+
+		rq->handle_rx_cqe(rq, cqe);
+	} while ((++work_done < budget) && (cqe = mlx5_cqwq_get_cqe(&cq->wq)));
+
+out:
+	if (xdpsq->doorbell) {
+		mlx5e_xmit_xdp_doorbell(xdpsq);
+		xdpsq->doorbell = false;
+	}
+
+	if (xdpsq->redirect_flush) {
+		xdp_do_flush_map();
+		xdpsq->redirect_flush = false;
+	}
+
+	mlx5_cqwq_update_db_record(&cq->wq);
+
+	/* ensure cq space is freed before enabling more cqes */
+	wmb();
+
+	return work_done;
+}
+
+#ifdef CONFIG_MLX5_CORE_IPOIB
+
+#define MLX5_IB_GRH_SGID_OFFSET 8
+#define MLX5_IB_GRH_DGID_OFFSET 24
+#define MLX5_GID_SIZE           16
+
+static inline void mlx5i_complete_rx_cqe(struct mlx5e_rq *rq,
+					 struct mlx5_cqe64 *cqe,
+					 u32 cqe_bcnt,
+					 struct sk_buff *skb)
+{
+	struct mlx5e_rq_stats *stats = rq->stats;
+	struct hwtstamp_config *tstamp;
+	struct net_device *netdev;
+	struct mlx5e_priv *priv;
+	char *pseudo_header;
+	u32 flags_rqpn;
+	u32 qpn;
+	u8 *dgid;
+	u8 g;
+
+	qpn = be32_to_cpu(cqe->sop_drop_qpn) & 0xffffff;
+	netdev = mlx5i_pkey_get_netdev(rq->netdev, qpn);
+
+	/* No mapping present, cannot process SKB. This might happen if a child
+	 * interface is going down while having unprocessed CQEs on parent RQ
+	 */
+	if (unlikely(!netdev)) {
+		/* TODO: add drop counters support */
+		skb->dev = NULL;
+		pr_warn_once("Unable to map QPN %u to dev - dropping skb\n", qpn);
+		return;
+	}
+
+	priv = mlx5i_epriv(netdev);
+	tstamp = &priv->tstamp;
+
+	flags_rqpn = be32_to_cpu(cqe->flags_rqpn);
+	g = (flags_rqpn >> 28) & 3;
+	dgid = skb->data + MLX5_IB_GRH_DGID_OFFSET;
+	if ((!g) || dgid[0] != 0xff)
+		skb->pkt_type = PACKET_HOST;
+	else if (memcmp(dgid, netdev->broadcast + 4, MLX5_GID_SIZE) == 0)
+		skb->pkt_type = PACKET_BROADCAST;
+	else
+		skb->pkt_type = PACKET_MULTICAST;
+
+	/* Drop packets that this interface sent, ie multicast packets
+	 * that the HCA has replicated.
+	 */
+	if (g && (qpn == (flags_rqpn & 0xffffff)) &&
+	    (memcmp(netdev->dev_addr + 4, skb->data + MLX5_IB_GRH_SGID_OFFSET,
+		    MLX5_GID_SIZE) == 0)) {
+		skb->dev = NULL;
+		return;
+	}
+
+	skb_pull(skb, MLX5_IB_GRH_BYTES);
+
+	skb->protocol = *((__be16 *)(skb->data));
+
+	if (netdev->features & NETIF_F_RXCSUM) {
+		skb->ip_summed = CHECKSUM_COMPLETE;
+		skb->csum = csum_unfold((__force __sum16)cqe->check_sum);
+		stats->csum_complete++;
+	} else {
+		skb->ip_summed = CHECKSUM_NONE;
+		stats->csum_none++;
+	}
+
+	if (unlikely(mlx5e_rx_hw_stamp(tstamp)))
+		skb_hwtstamps(skb)->hwtstamp =
+				mlx5_timecounter_cyc2time(rq->clock, get_cqe_ts(cqe));
+
+	skb_record_rx_queue(skb, rq->ix);
+
+	if (likely(netdev->features & NETIF_F_RXHASH))
+		mlx5e_skb_set_hash(cqe, skb);
+
+	/* 20 bytes of ipoib header and 4 for encap existing */
+	pseudo_header = skb_push(skb, MLX5_IPOIB_PSEUDO_LEN);
+	memset(pseudo_header, 0, MLX5_IPOIB_PSEUDO_LEN);
+	skb_reset_mac_header(skb);
+	skb_pull(skb, MLX5_IPOIB_HARD_LEN);
+
+	skb->dev = netdev;
+
+	stats->packets++;
+	stats->bytes += cqe_bcnt;
+}
+
+void mlx5i_handle_rx_cqe(struct mlx5e_rq *rq, struct mlx5_cqe64 *cqe)
+{
+	struct mlx5_wq_cyc *wq = &rq->wqe.wq;
+	struct mlx5e_wqe_frag_info *wi;
+	struct sk_buff *skb;
+	u32 cqe_bcnt;
+	u16 ci;
+
+	ci       = mlx5_wq_cyc_ctr2ix(wq, be16_to_cpu(cqe->wqe_counter));
+	wi       = get_frag(rq, ci);
+	cqe_bcnt = be32_to_cpu(cqe->byte_cnt);
+
+	skb = rq->wqe.skb_from_cqe(rq, cqe, wi, cqe_bcnt);
+	if (!skb)
+		goto wq_free_wqe;
+
+	mlx5i_complete_rx_cqe(rq, cqe, cqe_bcnt, skb);
+	if (unlikely(!skb->dev)) {
+		dev_kfree_skb_any(skb);
+		goto wq_free_wqe;
+	}
+	napi_gro_receive(rq->cq.napi, skb);
+
+wq_free_wqe:
+	mlx5e_free_rx_wqe(rq, wi, true);
+	mlx5_wq_cyc_pop(wq);
+}
+
+#endif /* CONFIG_MLX5_CORE_IPOIB */
+
+#ifdef CONFIG_MLX5_EN_IPSEC
+
+void mlx5e_ipsec_handle_rx_cqe(struct mlx5e_rq *rq, struct mlx5_cqe64 *cqe)
+{
+	struct mlx5_wq_cyc *wq = &rq->wqe.wq;
+	struct mlx5e_wqe_frag_info *wi;
+	struct sk_buff *skb;
+	u32 cqe_bcnt;
+	u16 ci;
+
+	ci       = mlx5_wq_cyc_ctr2ix(wq, be16_to_cpu(cqe->wqe_counter));
+	wi       = get_frag(rq, ci);
+	cqe_bcnt = be32_to_cpu(cqe->byte_cnt);
+
+	skb = rq->wqe.skb_from_cqe(rq, cqe, wi, cqe_bcnt);
+	if (unlikely(!skb)) {
+		/* a DROP, save the page-reuse checks */
+		mlx5e_free_rx_wqe(rq, wi, true);
+		goto wq_cyc_pop;
+	}
+	skb = mlx5e_ipsec_handle_rx_skb(rq->netdev, skb, &cqe_bcnt);
+	if (unlikely(!skb)) {
+		mlx5e_free_rx_wqe(rq, wi, true);
+		goto wq_cyc_pop;
+	}
+
+	mlx5e_complete_rx_cqe(rq, cqe, cqe_bcnt, skb);
+	napi_gro_receive(rq->cq.napi, skb);
+
+	mlx5e_free_rx_wqe(rq, wi, true);
+wq_cyc_pop:
+	mlx5_wq_cyc_pop(wq);
+}
+
+#endif /* CONFIG_MLX5_EN_IPSEC */
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_selftest.c b/drivers/net/ethernet/mellanox/mlx5/core/en_selftest.c
new file mode 100644
index 000000000..5fb088b54
--- /dev/null
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_selftest.c
@@ -0,0 +1,355 @@
+/*
+ * Copyright (c) 2016, Mellanox Technologies, Ltd.  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/prefetch.h>
+#include <linux/ip.h>
+#include <linux/udp.h>
+#include <net/udp.h>
+#include "en.h"
+#include "en/port.h"
+
+enum {
+	MLX5E_ST_LINK_STATE,
+	MLX5E_ST_LINK_SPEED,
+	MLX5E_ST_HEALTH_INFO,
+#ifdef CONFIG_INET
+	MLX5E_ST_LOOPBACK,
+#endif
+	MLX5E_ST_NUM,
+};
+
+const char mlx5e_self_tests[MLX5E_ST_NUM][ETH_GSTRING_LEN] = {
+	"Link Test",
+	"Speed Test",
+	"Health Test",
+#ifdef CONFIG_INET
+	"Loopback Test",
+#endif
+};
+
+int mlx5e_self_test_num(struct mlx5e_priv *priv)
+{
+	return ARRAY_SIZE(mlx5e_self_tests);
+}
+
+static int mlx5e_test_health_info(struct mlx5e_priv *priv)
+{
+	struct mlx5_core_health *health = &priv->mdev->priv.health;
+
+	return health->sick ? 1 : 0;
+}
+
+static int mlx5e_test_link_state(struct mlx5e_priv *priv)
+{
+	u8 port_state;
+
+	if (!netif_carrier_ok(priv->netdev))
+		return 1;
+
+	port_state = mlx5_query_vport_state(priv->mdev, MLX5_VPORT_STATE_OP_MOD_VNIC_VPORT, 0);
+	return port_state == VPORT_STATE_UP ? 0 : 1;
+}
+
+static int mlx5e_test_link_speed(struct mlx5e_priv *priv)
+{
+	u32 speed;
+
+	if (!netif_carrier_ok(priv->netdev))
+		return 1;
+
+	return mlx5e_port_linkspeed(priv->mdev, &speed);
+}
+
+struct mlx5ehdr {
+	__be32 version;
+	__be64 magic;
+};
+
+#ifdef CONFIG_INET
+/* loopback test */
+#define MLX5E_TEST_PKT_SIZE (sizeof(struct ethhdr) + sizeof(struct iphdr) +\
+			     sizeof(struct udphdr) + sizeof(struct mlx5ehdr))
+#define MLX5E_TEST_MAGIC 0x5AEED15C001ULL
+
+static struct sk_buff *mlx5e_test_get_udp_skb(struct mlx5e_priv *priv)
+{
+	struct sk_buff *skb = NULL;
+	struct mlx5ehdr *mlxh;
+	struct ethhdr *ethh;
+	struct udphdr *udph;
+	struct iphdr *iph;
+	int    iplen;
+
+	skb = netdev_alloc_skb(priv->netdev, MLX5E_TEST_PKT_SIZE);
+	if (!skb) {
+		netdev_err(priv->netdev, "\tFailed to alloc loopback skb\n");
+		return NULL;
+	}
+
+	prefetchw(skb->data);
+	skb_reserve(skb, NET_IP_ALIGN);
+
+	/*  Reserve for ethernet and IP header  */
+	ethh = skb_push(skb, ETH_HLEN);
+	skb_reset_mac_header(skb);
+
+	skb_set_network_header(skb, skb->len);
+	iph = skb_put(skb, sizeof(struct iphdr));
+
+	skb_set_transport_header(skb, skb->len);
+	udph = skb_put(skb, sizeof(struct udphdr));
+
+	/* Fill ETH header */
+	ether_addr_copy(ethh->h_dest, priv->netdev->dev_addr);
+	eth_zero_addr(ethh->h_source);
+	ethh->h_proto = htons(ETH_P_IP);
+
+	/* Fill UDP header */
+	udph->source = htons(9);
+	udph->dest = htons(9); /* Discard Protocol */
+	udph->len = htons(sizeof(struct mlx5ehdr) + sizeof(struct udphdr));
+	udph->check = 0;
+
+	/* Fill IP header */
+	iph->ihl = 5;
+	iph->ttl = 32;
+	iph->version = 4;
+	iph->protocol = IPPROTO_UDP;
+	iplen = sizeof(struct iphdr) + sizeof(struct udphdr) +
+		sizeof(struct mlx5ehdr);
+	iph->tot_len = htons(iplen);
+	iph->frag_off = 0;
+	iph->saddr = 0;
+	iph->daddr = 0;
+	iph->tos = 0;
+	iph->id = 0;
+	ip_send_check(iph);
+
+	/* Fill test header and data */
+	mlxh = skb_put(skb, sizeof(*mlxh));
+	mlxh->version = 0;
+	mlxh->magic = cpu_to_be64(MLX5E_TEST_MAGIC);
+
+	skb->csum = 0;
+	skb->ip_summed = CHECKSUM_PARTIAL;
+	udp4_hwcsum(skb, iph->saddr, iph->daddr);
+
+	skb->protocol = htons(ETH_P_IP);
+	skb->pkt_type = PACKET_HOST;
+	skb->dev = priv->netdev;
+
+	return skb;
+}
+
+struct mlx5e_lbt_priv {
+	struct packet_type pt;
+	struct completion comp;
+	bool loopback_ok;
+	bool local_lb;
+};
+
+static int
+mlx5e_test_loopback_validate(struct sk_buff *skb,
+			     struct net_device *ndev,
+			     struct packet_type *pt,
+			     struct net_device *orig_ndev)
+{
+	struct mlx5e_lbt_priv *lbtp = pt->af_packet_priv;
+	struct mlx5ehdr *mlxh;
+	struct ethhdr *ethh;
+	struct udphdr *udph;
+	struct iphdr *iph;
+
+	/* We are only going to peek, no need to clone the SKB */
+	if (MLX5E_TEST_PKT_SIZE - ETH_HLEN > skb_headlen(skb))
+		goto out;
+
+	ethh = (struct ethhdr *)skb_mac_header(skb);
+	if (!ether_addr_equal(ethh->h_dest, orig_ndev->dev_addr))
+		goto out;
+
+	iph = ip_hdr(skb);
+	if (iph->protocol != IPPROTO_UDP)
+		goto out;
+
+	/* Don't assume skb_transport_header() was set */
+	udph = (struct udphdr *)((u8 *)iph + 4 * iph->ihl);
+	if (udph->dest != htons(9))
+		goto out;
+
+	mlxh = (struct mlx5ehdr *)((char *)udph + sizeof(*udph));
+	if (mlxh->magic != cpu_to_be64(MLX5E_TEST_MAGIC))
+		goto out; /* so close ! */
+
+	/* bingo */
+	lbtp->loopback_ok = true;
+	complete(&lbtp->comp);
+out:
+	kfree_skb(skb);
+	return 0;
+}
+
+static int mlx5e_test_loopback_setup(struct mlx5e_priv *priv,
+				     struct mlx5e_lbt_priv *lbtp)
+{
+	int err = 0;
+
+	/* Temporarily enable local_lb */
+	err = mlx5_nic_vport_query_local_lb(priv->mdev, &lbtp->local_lb);
+	if (err)
+		return err;
+
+	if (!lbtp->local_lb) {
+		err = mlx5_nic_vport_update_local_lb(priv->mdev, true);
+		if (err)
+			return err;
+	}
+
+	err = mlx5e_refresh_tirs(priv, true);
+	if (err)
+		goto out;
+
+	lbtp->loopback_ok = false;
+	init_completion(&lbtp->comp);
+
+	lbtp->pt.type = htons(ETH_P_IP);
+	lbtp->pt.func = mlx5e_test_loopback_validate;
+	lbtp->pt.dev = priv->netdev;
+	lbtp->pt.af_packet_priv = lbtp;
+	dev_add_pack(&lbtp->pt);
+
+	return 0;
+
+out:
+	if (!lbtp->local_lb)
+		mlx5_nic_vport_update_local_lb(priv->mdev, false);
+
+	return err;
+}
+
+static void mlx5e_test_loopback_cleanup(struct mlx5e_priv *priv,
+					struct mlx5e_lbt_priv *lbtp)
+{
+	if (!lbtp->local_lb)
+		mlx5_nic_vport_update_local_lb(priv->mdev, false);
+
+	dev_remove_pack(&lbtp->pt);
+	mlx5e_refresh_tirs(priv, false);
+}
+
+#define MLX5E_LB_VERIFY_TIMEOUT (msecs_to_jiffies(200))
+static int mlx5e_test_loopback(struct mlx5e_priv *priv)
+{
+	struct mlx5e_lbt_priv *lbtp;
+	struct sk_buff *skb = NULL;
+	int err;
+
+	if (!test_bit(MLX5E_STATE_OPENED, &priv->state)) {
+		netdev_err(priv->netdev,
+			   "\tCan't perform loopback test while device is down\n");
+		return -ENODEV;
+	}
+
+	lbtp = kzalloc(sizeof(*lbtp), GFP_KERNEL);
+	if (!lbtp)
+		return -ENOMEM;
+	lbtp->loopback_ok = false;
+
+	err = mlx5e_test_loopback_setup(priv, lbtp);
+	if (err)
+		goto out;
+
+	skb = mlx5e_test_get_udp_skb(priv);
+	if (!skb) {
+		err = -ENOMEM;
+		goto cleanup;
+	}
+
+	skb_set_queue_mapping(skb, 0);
+	err = dev_queue_xmit(skb);
+	if (err) {
+		netdev_err(priv->netdev,
+			   "\tFailed to xmit loopback packet err(%d)\n",
+			   err);
+		goto cleanup;
+	}
+
+	wait_for_completion_timeout(&lbtp->comp, MLX5E_LB_VERIFY_TIMEOUT);
+	err = !lbtp->loopback_ok;
+
+cleanup:
+	mlx5e_test_loopback_cleanup(priv, lbtp);
+out:
+	kfree(lbtp);
+	return err;
+}
+#endif
+
+static int (*mlx5e_st_func[MLX5E_ST_NUM])(struct mlx5e_priv *) = {
+	mlx5e_test_link_state,
+	mlx5e_test_link_speed,
+	mlx5e_test_health_info,
+#ifdef CONFIG_INET
+	mlx5e_test_loopback,
+#endif
+};
+
+void mlx5e_self_test(struct net_device *ndev, struct ethtool_test *etest,
+		     u64 *buf)
+{
+	struct mlx5e_priv *priv = netdev_priv(ndev);
+	int i;
+
+	memset(buf, 0, sizeof(u64) * MLX5E_ST_NUM);
+
+	mutex_lock(&priv->state_lock);
+	netdev_info(ndev, "Self test begin..\n");
+
+	for (i = 0; i < MLX5E_ST_NUM; i++) {
+		netdev_info(ndev, "\t[%d] %s start..\n",
+			    i, mlx5e_self_tests[i]);
+		buf[i] = mlx5e_st_func[i](priv);
+		netdev_info(ndev, "\t[%d] %s end: result(%lld)\n",
+			    i, mlx5e_self_tests[i], buf[i]);
+	}
+
+	mutex_unlock(&priv->state_lock);
+
+	for (i = 0; i < MLX5E_ST_NUM; i++) {
+		if (buf[i]) {
+			etest->flags |= ETH_TEST_FL_FAILED;
+			break;
+		}
+	}
+	netdev_info(ndev, "Self test out: status flags(0x%x)\n",
+		    etest->flags);
+}
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_stats.c b/drivers/net/ethernet/mellanox/mlx5/core/en_stats.c
new file mode 100644
index 000000000..9a68dee58
--- /dev/null
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_stats.c
@@ -0,0 +1,1404 @@
+/*
+ * Copyright (c) 2017, Mellanox Technologies, Ltd.  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 "en.h"
+#include "en_accel/ipsec.h"
+#include "en_accel/tls.h"
+
+static const struct counter_desc sw_stats_desc[] = {
+	{ MLX5E_DECLARE_STAT(struct mlx5e_sw_stats, rx_packets) },
+	{ MLX5E_DECLARE_STAT(struct mlx5e_sw_stats, rx_bytes) },
+	{ MLX5E_DECLARE_STAT(struct mlx5e_sw_stats, tx_packets) },
+	{ MLX5E_DECLARE_STAT(struct mlx5e_sw_stats, tx_bytes) },
+	{ MLX5E_DECLARE_STAT(struct mlx5e_sw_stats, tx_tso_packets) },
+	{ MLX5E_DECLARE_STAT(struct mlx5e_sw_stats, tx_tso_bytes) },
+	{ MLX5E_DECLARE_STAT(struct mlx5e_sw_stats, tx_tso_inner_packets) },
+	{ MLX5E_DECLARE_STAT(struct mlx5e_sw_stats, tx_tso_inner_bytes) },
+	{ MLX5E_DECLARE_STAT(struct mlx5e_sw_stats, tx_added_vlan_packets) },
+	{ MLX5E_DECLARE_STAT(struct mlx5e_sw_stats, tx_nop) },
+
+#ifdef CONFIG_MLX5_EN_TLS
+	{ MLX5E_DECLARE_STAT(struct mlx5e_sw_stats, tx_tls_ooo) },
+	{ MLX5E_DECLARE_STAT(struct mlx5e_sw_stats, tx_tls_resync_bytes) },
+#endif
+
+	{ MLX5E_DECLARE_STAT(struct mlx5e_sw_stats, rx_lro_packets) },
+	{ MLX5E_DECLARE_STAT(struct mlx5e_sw_stats, rx_lro_bytes) },
+	{ MLX5E_DECLARE_STAT(struct mlx5e_sw_stats, rx_ecn_mark) },
+	{ MLX5E_DECLARE_STAT(struct mlx5e_sw_stats, rx_removed_vlan_packets) },
+	{ MLX5E_DECLARE_STAT(struct mlx5e_sw_stats, rx_csum_unnecessary) },
+	{ MLX5E_DECLARE_STAT(struct mlx5e_sw_stats, rx_csum_none) },
+	{ MLX5E_DECLARE_STAT(struct mlx5e_sw_stats, rx_csum_complete) },
+	{ MLX5E_DECLARE_STAT(struct mlx5e_sw_stats, rx_csum_complete_tail) },
+	{ MLX5E_DECLARE_STAT(struct mlx5e_sw_stats, rx_csum_complete_tail_slow) },
+	{ MLX5E_DECLARE_STAT(struct mlx5e_sw_stats, rx_csum_unnecessary_inner) },
+	{ MLX5E_DECLARE_STAT(struct mlx5e_sw_stats, rx_xdp_drop) },
+	{ MLX5E_DECLARE_STAT(struct mlx5e_sw_stats, rx_xdp_redirect) },
+	{ MLX5E_DECLARE_STAT(struct mlx5e_sw_stats, rx_xdp_tx_xmit) },
+	{ MLX5E_DECLARE_STAT(struct mlx5e_sw_stats, rx_xdp_tx_full) },
+	{ MLX5E_DECLARE_STAT(struct mlx5e_sw_stats, rx_xdp_tx_err) },
+	{ MLX5E_DECLARE_STAT(struct mlx5e_sw_stats, rx_xdp_tx_cqe) },
+	{ MLX5E_DECLARE_STAT(struct mlx5e_sw_stats, tx_csum_none) },
+	{ MLX5E_DECLARE_STAT(struct mlx5e_sw_stats, tx_csum_partial) },
+	{ MLX5E_DECLARE_STAT(struct mlx5e_sw_stats, tx_csum_partial_inner) },
+	{ MLX5E_DECLARE_STAT(struct mlx5e_sw_stats, tx_queue_stopped) },
+	{ MLX5E_DECLARE_STAT(struct mlx5e_sw_stats, tx_queue_dropped) },
+	{ MLX5E_DECLARE_STAT(struct mlx5e_sw_stats, tx_xmit_more) },
+	{ MLX5E_DECLARE_STAT(struct mlx5e_sw_stats, tx_recover) },
+	{ MLX5E_DECLARE_STAT(struct mlx5e_sw_stats, tx_cqes) },
+	{ MLX5E_DECLARE_STAT(struct mlx5e_sw_stats, tx_queue_wake) },
+	{ MLX5E_DECLARE_STAT(struct mlx5e_sw_stats, tx_cqe_err) },
+	{ MLX5E_DECLARE_STAT(struct mlx5e_sw_stats, tx_xdp_xmit) },
+	{ MLX5E_DECLARE_STAT(struct mlx5e_sw_stats, tx_xdp_full) },
+	{ MLX5E_DECLARE_STAT(struct mlx5e_sw_stats, tx_xdp_err) },
+	{ MLX5E_DECLARE_STAT(struct mlx5e_sw_stats, tx_xdp_cqes) },
+	{ MLX5E_DECLARE_STAT(struct mlx5e_sw_stats, rx_wqe_err) },
+	{ MLX5E_DECLARE_STAT(struct mlx5e_sw_stats, rx_mpwqe_filler_cqes) },
+	{ MLX5E_DECLARE_STAT(struct mlx5e_sw_stats, rx_mpwqe_filler_strides) },
+	{ MLX5E_DECLARE_STAT(struct mlx5e_sw_stats, rx_oversize_pkts_sw_drop) },
+	{ MLX5E_DECLARE_STAT(struct mlx5e_sw_stats, rx_buff_alloc_err) },
+	{ MLX5E_DECLARE_STAT(struct mlx5e_sw_stats, rx_cqe_compress_blks) },
+	{ MLX5E_DECLARE_STAT(struct mlx5e_sw_stats, rx_cqe_compress_pkts) },
+	{ MLX5E_DECLARE_STAT(struct mlx5e_sw_stats, rx_page_reuse) },
+	{ MLX5E_DECLARE_STAT(struct mlx5e_sw_stats, rx_cache_reuse) },
+	{ MLX5E_DECLARE_STAT(struct mlx5e_sw_stats, rx_cache_full) },
+	{ MLX5E_DECLARE_STAT(struct mlx5e_sw_stats, rx_cache_empty) },
+	{ MLX5E_DECLARE_STAT(struct mlx5e_sw_stats, rx_cache_busy) },
+	{ MLX5E_DECLARE_STAT(struct mlx5e_sw_stats, rx_cache_waive) },
+	{ MLX5E_DECLARE_STAT(struct mlx5e_sw_stats, rx_congst_umr) },
+	{ MLX5E_DECLARE_STAT(struct mlx5e_sw_stats, ch_events) },
+	{ MLX5E_DECLARE_STAT(struct mlx5e_sw_stats, ch_poll) },
+	{ MLX5E_DECLARE_STAT(struct mlx5e_sw_stats, ch_arm) },
+	{ MLX5E_DECLARE_STAT(struct mlx5e_sw_stats, ch_aff_change) },
+	{ MLX5E_DECLARE_STAT(struct mlx5e_sw_stats, ch_eq_rearm) },
+};
+
+#define NUM_SW_COUNTERS			ARRAY_SIZE(sw_stats_desc)
+
+static int mlx5e_grp_sw_get_num_stats(struct mlx5e_priv *priv)
+{
+	return NUM_SW_COUNTERS;
+}
+
+static int mlx5e_grp_sw_fill_strings(struct mlx5e_priv *priv, u8 *data, int idx)
+{
+	int i;
+
+	for (i = 0; i < NUM_SW_COUNTERS; i++)
+		strcpy(data + (idx++) * ETH_GSTRING_LEN, sw_stats_desc[i].format);
+	return idx;
+}
+
+static int mlx5e_grp_sw_fill_stats(struct mlx5e_priv *priv, u64 *data, int idx)
+{
+	int i;
+
+	for (i = 0; i < NUM_SW_COUNTERS; i++)
+		data[idx++] = MLX5E_READ_CTR64_CPU(&priv->stats.sw, sw_stats_desc, i);
+	return idx;
+}
+
+void mlx5e_grp_sw_update_stats(struct mlx5e_priv *priv)
+{
+	struct mlx5e_sw_stats temp, *s = &temp;
+	int i;
+
+	memset(s, 0, sizeof(*s));
+
+	for (i = 0; i < priv->profile->max_nch(priv->mdev); i++) {
+		struct mlx5e_channel_stats *channel_stats =
+			&priv->channel_stats[i];
+		struct mlx5e_xdpsq_stats *xdpsq_red_stats = &channel_stats->xdpsq;
+		struct mlx5e_xdpsq_stats *xdpsq_stats = &channel_stats->rq_xdpsq;
+		struct mlx5e_rq_stats *rq_stats = &channel_stats->rq;
+		struct mlx5e_ch_stats *ch_stats = &channel_stats->ch;
+		int j;
+
+		s->rx_packets	+= rq_stats->packets;
+		s->rx_bytes	+= rq_stats->bytes;
+		s->rx_lro_packets += rq_stats->lro_packets;
+		s->rx_lro_bytes	+= rq_stats->lro_bytes;
+		s->rx_ecn_mark	+= rq_stats->ecn_mark;
+		s->rx_removed_vlan_packets += rq_stats->removed_vlan_packets;
+		s->rx_csum_none	+= rq_stats->csum_none;
+		s->rx_csum_complete += rq_stats->csum_complete;
+		s->rx_csum_complete_tail += rq_stats->csum_complete_tail;
+		s->rx_csum_complete_tail_slow += rq_stats->csum_complete_tail_slow;
+		s->rx_csum_unnecessary += rq_stats->csum_unnecessary;
+		s->rx_csum_unnecessary_inner += rq_stats->csum_unnecessary_inner;
+		s->rx_xdp_drop     += rq_stats->xdp_drop;
+		s->rx_xdp_redirect += rq_stats->xdp_redirect;
+		s->rx_xdp_tx_xmit  += xdpsq_stats->xmit;
+		s->rx_xdp_tx_full  += xdpsq_stats->full;
+		s->rx_xdp_tx_err   += xdpsq_stats->err;
+		s->rx_xdp_tx_cqe   += xdpsq_stats->cqes;
+		s->rx_wqe_err   += rq_stats->wqe_err;
+		s->rx_mpwqe_filler_cqes    += rq_stats->mpwqe_filler_cqes;
+		s->rx_mpwqe_filler_strides += rq_stats->mpwqe_filler_strides;
+		s->rx_oversize_pkts_sw_drop += rq_stats->oversize_pkts_sw_drop;
+		s->rx_buff_alloc_err += rq_stats->buff_alloc_err;
+		s->rx_cqe_compress_blks += rq_stats->cqe_compress_blks;
+		s->rx_cqe_compress_pkts += rq_stats->cqe_compress_pkts;
+		s->rx_page_reuse  += rq_stats->page_reuse;
+		s->rx_cache_reuse += rq_stats->cache_reuse;
+		s->rx_cache_full  += rq_stats->cache_full;
+		s->rx_cache_empty += rq_stats->cache_empty;
+		s->rx_cache_busy  += rq_stats->cache_busy;
+		s->rx_cache_waive += rq_stats->cache_waive;
+		s->rx_congst_umr  += rq_stats->congst_umr;
+		s->ch_events      += ch_stats->events;
+		s->ch_poll        += ch_stats->poll;
+		s->ch_arm         += ch_stats->arm;
+		s->ch_aff_change  += ch_stats->aff_change;
+		s->ch_eq_rearm    += ch_stats->eq_rearm;
+		/* xdp redirect */
+		s->tx_xdp_xmit    += xdpsq_red_stats->xmit;
+		s->tx_xdp_full    += xdpsq_red_stats->full;
+		s->tx_xdp_err     += xdpsq_red_stats->err;
+		s->tx_xdp_cqes    += xdpsq_red_stats->cqes;
+
+		for (j = 0; j < priv->max_opened_tc; j++) {
+			struct mlx5e_sq_stats *sq_stats = &channel_stats->sq[j];
+
+			s->tx_packets		+= sq_stats->packets;
+			s->tx_bytes		+= sq_stats->bytes;
+			s->tx_tso_packets	+= sq_stats->tso_packets;
+			s->tx_tso_bytes		+= sq_stats->tso_bytes;
+			s->tx_tso_inner_packets	+= sq_stats->tso_inner_packets;
+			s->tx_tso_inner_bytes	+= sq_stats->tso_inner_bytes;
+			s->tx_added_vlan_packets += sq_stats->added_vlan_packets;
+			s->tx_nop               += sq_stats->nop;
+			s->tx_queue_stopped	+= sq_stats->stopped;
+			s->tx_queue_wake	+= sq_stats->wake;
+			s->tx_queue_dropped	+= sq_stats->dropped;
+			s->tx_cqe_err		+= sq_stats->cqe_err;
+			s->tx_recover		+= sq_stats->recover;
+			s->tx_xmit_more		+= sq_stats->xmit_more;
+			s->tx_csum_partial_inner += sq_stats->csum_partial_inner;
+			s->tx_csum_none		+= sq_stats->csum_none;
+			s->tx_csum_partial	+= sq_stats->csum_partial;
+#ifdef CONFIG_MLX5_EN_TLS
+			s->tx_tls_ooo		+= sq_stats->tls_ooo;
+			s->tx_tls_resync_bytes	+= sq_stats->tls_resync_bytes;
+#endif
+			s->tx_cqes		+= sq_stats->cqes;
+
+			/* https://gcc.gnu.org/bugzilla/show_bug.cgi?id=92657 */
+			barrier();
+		}
+	}
+
+	memcpy(&priv->stats.sw, s, sizeof(*s));
+}
+
+static const struct counter_desc q_stats_desc[] = {
+	{ MLX5E_DECLARE_STAT(struct mlx5e_qcounter_stats, rx_out_of_buffer) },
+};
+
+static const struct counter_desc drop_rq_stats_desc[] = {
+	{ MLX5E_DECLARE_STAT(struct mlx5e_qcounter_stats, rx_if_down_packets) },
+};
+
+#define NUM_Q_COUNTERS			ARRAY_SIZE(q_stats_desc)
+#define NUM_DROP_RQ_COUNTERS		ARRAY_SIZE(drop_rq_stats_desc)
+
+static int mlx5e_grp_q_get_num_stats(struct mlx5e_priv *priv)
+{
+	int num_stats = 0;
+
+	if (priv->q_counter)
+		num_stats += NUM_Q_COUNTERS;
+
+	if (priv->drop_rq_q_counter)
+		num_stats += NUM_DROP_RQ_COUNTERS;
+
+	return num_stats;
+}
+
+static int mlx5e_grp_q_fill_strings(struct mlx5e_priv *priv, u8 *data, int idx)
+{
+	int i;
+
+	for (i = 0; i < NUM_Q_COUNTERS && priv->q_counter; i++)
+		strcpy(data + (idx++) * ETH_GSTRING_LEN,
+		       q_stats_desc[i].format);
+
+	for (i = 0; i < NUM_DROP_RQ_COUNTERS && priv->drop_rq_q_counter; i++)
+		strcpy(data + (idx++) * ETH_GSTRING_LEN,
+		       drop_rq_stats_desc[i].format);
+
+	return idx;
+}
+
+static int mlx5e_grp_q_fill_stats(struct mlx5e_priv *priv, u64 *data, int idx)
+{
+	int i;
+
+	for (i = 0; i < NUM_Q_COUNTERS && priv->q_counter; i++)
+		data[idx++] = MLX5E_READ_CTR32_CPU(&priv->stats.qcnt,
+						   q_stats_desc, i);
+	for (i = 0; i < NUM_DROP_RQ_COUNTERS && priv->drop_rq_q_counter; i++)
+		data[idx++] = MLX5E_READ_CTR32_CPU(&priv->stats.qcnt,
+						   drop_rq_stats_desc, i);
+	return idx;
+}
+
+static void mlx5e_grp_q_update_stats(struct mlx5e_priv *priv)
+{
+	struct mlx5e_qcounter_stats *qcnt = &priv->stats.qcnt;
+	u32 out[MLX5_ST_SZ_DW(query_q_counter_out)];
+
+	if (priv->q_counter &&
+	    !mlx5_core_query_q_counter(priv->mdev, priv->q_counter, 0, out,
+				       sizeof(out)))
+		qcnt->rx_out_of_buffer = MLX5_GET(query_q_counter_out,
+						  out, out_of_buffer);
+	if (priv->drop_rq_q_counter &&
+	    !mlx5_core_query_q_counter(priv->mdev, priv->drop_rq_q_counter, 0,
+				       out, sizeof(out)))
+		qcnt->rx_if_down_packets = MLX5_GET(query_q_counter_out, out,
+						    out_of_buffer);
+}
+
+#define VNIC_ENV_OFF(c) MLX5_BYTE_OFF(query_vnic_env_out, c)
+static const struct counter_desc vnic_env_stats_desc[] = {
+	{ "rx_steer_missed_packets",
+		VNIC_ENV_OFF(vport_env.nic_receive_steering_discard) },
+};
+
+#define NUM_VNIC_ENV_COUNTERS		ARRAY_SIZE(vnic_env_stats_desc)
+
+static int mlx5e_grp_vnic_env_get_num_stats(struct mlx5e_priv *priv)
+{
+	return MLX5_CAP_GEN(priv->mdev, nic_receive_steering_discard) ?
+		NUM_VNIC_ENV_COUNTERS : 0;
+}
+
+static int mlx5e_grp_vnic_env_fill_strings(struct mlx5e_priv *priv, u8 *data,
+					   int idx)
+{
+	int i;
+
+	if (!MLX5_CAP_GEN(priv->mdev, nic_receive_steering_discard))
+		return idx;
+
+	for (i = 0; i < NUM_VNIC_ENV_COUNTERS; i++)
+		strcpy(data + (idx++) * ETH_GSTRING_LEN,
+		       vnic_env_stats_desc[i].format);
+	return idx;
+}
+
+static int mlx5e_grp_vnic_env_fill_stats(struct mlx5e_priv *priv, u64 *data,
+					 int idx)
+{
+	int i;
+
+	if (!MLX5_CAP_GEN(priv->mdev, nic_receive_steering_discard))
+		return idx;
+
+	for (i = 0; i < NUM_VNIC_ENV_COUNTERS; i++)
+		data[idx++] = MLX5E_READ_CTR64_BE(priv->stats.vnic.query_vnic_env_out,
+						  vnic_env_stats_desc, i);
+	return idx;
+}
+
+static void mlx5e_grp_vnic_env_update_stats(struct mlx5e_priv *priv)
+{
+	u32 *out = (u32 *)priv->stats.vnic.query_vnic_env_out;
+	int outlen = MLX5_ST_SZ_BYTES(query_vnic_env_out);
+	u32 in[MLX5_ST_SZ_DW(query_vnic_env_in)] = {0};
+	struct mlx5_core_dev *mdev = priv->mdev;
+
+	if (!MLX5_CAP_GEN(priv->mdev, nic_receive_steering_discard))
+		return;
+
+	MLX5_SET(query_vnic_env_in, in, opcode,
+		 MLX5_CMD_OP_QUERY_VNIC_ENV);
+	MLX5_SET(query_vnic_env_in, in, op_mod, 0);
+	MLX5_SET(query_vnic_env_in, in, other_vport, 0);
+	mlx5_cmd_exec(mdev, in, sizeof(in), out, outlen);
+}
+
+#define VPORT_COUNTER_OFF(c) MLX5_BYTE_OFF(query_vport_counter_out, c)
+static const struct counter_desc vport_stats_desc[] = {
+	{ "rx_vport_unicast_packets",
+		VPORT_COUNTER_OFF(received_eth_unicast.packets) },
+	{ "rx_vport_unicast_bytes",
+		VPORT_COUNTER_OFF(received_eth_unicast.octets) },
+	{ "tx_vport_unicast_packets",
+		VPORT_COUNTER_OFF(transmitted_eth_unicast.packets) },
+	{ "tx_vport_unicast_bytes",
+		VPORT_COUNTER_OFF(transmitted_eth_unicast.octets) },
+	{ "rx_vport_multicast_packets",
+		VPORT_COUNTER_OFF(received_eth_multicast.packets) },
+	{ "rx_vport_multicast_bytes",
+		VPORT_COUNTER_OFF(received_eth_multicast.octets) },
+	{ "tx_vport_multicast_packets",
+		VPORT_COUNTER_OFF(transmitted_eth_multicast.packets) },
+	{ "tx_vport_multicast_bytes",
+		VPORT_COUNTER_OFF(transmitted_eth_multicast.octets) },
+	{ "rx_vport_broadcast_packets",
+		VPORT_COUNTER_OFF(received_eth_broadcast.packets) },
+	{ "rx_vport_broadcast_bytes",
+		VPORT_COUNTER_OFF(received_eth_broadcast.octets) },
+	{ "tx_vport_broadcast_packets",
+		VPORT_COUNTER_OFF(transmitted_eth_broadcast.packets) },
+	{ "tx_vport_broadcast_bytes",
+		VPORT_COUNTER_OFF(transmitted_eth_broadcast.octets) },
+	{ "rx_vport_rdma_unicast_packets",
+		VPORT_COUNTER_OFF(received_ib_unicast.packets) },
+	{ "rx_vport_rdma_unicast_bytes",
+		VPORT_COUNTER_OFF(received_ib_unicast.octets) },
+	{ "tx_vport_rdma_unicast_packets",
+		VPORT_COUNTER_OFF(transmitted_ib_unicast.packets) },
+	{ "tx_vport_rdma_unicast_bytes",
+		VPORT_COUNTER_OFF(transmitted_ib_unicast.octets) },
+	{ "rx_vport_rdma_multicast_packets",
+		VPORT_COUNTER_OFF(received_ib_multicast.packets) },
+	{ "rx_vport_rdma_multicast_bytes",
+		VPORT_COUNTER_OFF(received_ib_multicast.octets) },
+	{ "tx_vport_rdma_multicast_packets",
+		VPORT_COUNTER_OFF(transmitted_ib_multicast.packets) },
+	{ "tx_vport_rdma_multicast_bytes",
+		VPORT_COUNTER_OFF(transmitted_ib_multicast.octets) },
+};
+
+#define NUM_VPORT_COUNTERS		ARRAY_SIZE(vport_stats_desc)
+
+static int mlx5e_grp_vport_get_num_stats(struct mlx5e_priv *priv)
+{
+	return NUM_VPORT_COUNTERS;
+}
+
+static int mlx5e_grp_vport_fill_strings(struct mlx5e_priv *priv, u8 *data,
+					int idx)
+{
+	int i;
+
+	for (i = 0; i < NUM_VPORT_COUNTERS; i++)
+		strcpy(data + (idx++) * ETH_GSTRING_LEN, vport_stats_desc[i].format);
+	return idx;
+}
+
+static int mlx5e_grp_vport_fill_stats(struct mlx5e_priv *priv, u64 *data,
+				      int idx)
+{
+	int i;
+
+	for (i = 0; i < NUM_VPORT_COUNTERS; i++)
+		data[idx++] = MLX5E_READ_CTR64_BE(priv->stats.vport.query_vport_out,
+						  vport_stats_desc, i);
+	return idx;
+}
+
+static void mlx5e_grp_vport_update_stats(struct mlx5e_priv *priv)
+{
+	int outlen = MLX5_ST_SZ_BYTES(query_vport_counter_out);
+	u32 *out = (u32 *)priv->stats.vport.query_vport_out;
+	u32 in[MLX5_ST_SZ_DW(query_vport_counter_in)] = {0};
+	struct mlx5_core_dev *mdev = priv->mdev;
+
+	MLX5_SET(query_vport_counter_in, in, opcode, MLX5_CMD_OP_QUERY_VPORT_COUNTER);
+	MLX5_SET(query_vport_counter_in, in, op_mod, 0);
+	MLX5_SET(query_vport_counter_in, in, other_vport, 0);
+	mlx5_cmd_exec(mdev, in, sizeof(in), out, outlen);
+}
+
+#define PPORT_802_3_OFF(c) \
+	MLX5_BYTE_OFF(ppcnt_reg, \
+		      counter_set.eth_802_3_cntrs_grp_data_layout.c##_high)
+static const struct counter_desc pport_802_3_stats_desc[] = {
+	{ "tx_packets_phy", PPORT_802_3_OFF(a_frames_transmitted_ok) },
+	{ "rx_packets_phy", PPORT_802_3_OFF(a_frames_received_ok) },
+	{ "rx_crc_errors_phy", PPORT_802_3_OFF(a_frame_check_sequence_errors) },
+	{ "tx_bytes_phy", PPORT_802_3_OFF(a_octets_transmitted_ok) },
+	{ "rx_bytes_phy", PPORT_802_3_OFF(a_octets_received_ok) },
+	{ "tx_multicast_phy", PPORT_802_3_OFF(a_multicast_frames_xmitted_ok) },
+	{ "tx_broadcast_phy", PPORT_802_3_OFF(a_broadcast_frames_xmitted_ok) },
+	{ "rx_multicast_phy", PPORT_802_3_OFF(a_multicast_frames_received_ok) },
+	{ "rx_broadcast_phy", PPORT_802_3_OFF(a_broadcast_frames_received_ok) },
+	{ "rx_in_range_len_errors_phy", PPORT_802_3_OFF(a_in_range_length_errors) },
+	{ "rx_out_of_range_len_phy", PPORT_802_3_OFF(a_out_of_range_length_field) },
+	{ "rx_oversize_pkts_phy", PPORT_802_3_OFF(a_frame_too_long_errors) },
+	{ "rx_symbol_err_phy", PPORT_802_3_OFF(a_symbol_error_during_carrier) },
+	{ "tx_mac_control_phy", PPORT_802_3_OFF(a_mac_control_frames_transmitted) },
+	{ "rx_mac_control_phy", PPORT_802_3_OFF(a_mac_control_frames_received) },
+	{ "rx_unsupported_op_phy", PPORT_802_3_OFF(a_unsupported_opcodes_received) },
+	{ "rx_pause_ctrl_phy", PPORT_802_3_OFF(a_pause_mac_ctrl_frames_received) },
+	{ "tx_pause_ctrl_phy", PPORT_802_3_OFF(a_pause_mac_ctrl_frames_transmitted) },
+};
+
+#define NUM_PPORT_802_3_COUNTERS	ARRAY_SIZE(pport_802_3_stats_desc)
+
+static int mlx5e_grp_802_3_get_num_stats(struct mlx5e_priv *priv)
+{
+	return NUM_PPORT_802_3_COUNTERS;
+}
+
+static int mlx5e_grp_802_3_fill_strings(struct mlx5e_priv *priv, u8 *data,
+					int idx)
+{
+	int i;
+
+	for (i = 0; i < NUM_PPORT_802_3_COUNTERS; i++)
+		strcpy(data + (idx++) * ETH_GSTRING_LEN, pport_802_3_stats_desc[i].format);
+	return idx;
+}
+
+static int mlx5e_grp_802_3_fill_stats(struct mlx5e_priv *priv, u64 *data,
+				      int idx)
+{
+	int i;
+
+	for (i = 0; i < NUM_PPORT_802_3_COUNTERS; i++)
+		data[idx++] = MLX5E_READ_CTR64_BE(&priv->stats.pport.IEEE_802_3_counters,
+						  pport_802_3_stats_desc, i);
+	return idx;
+}
+
+static void mlx5e_grp_802_3_update_stats(struct mlx5e_priv *priv)
+{
+	struct mlx5e_pport_stats *pstats = &priv->stats.pport;
+	struct mlx5_core_dev *mdev = priv->mdev;
+	u32 in[MLX5_ST_SZ_DW(ppcnt_reg)] = {0};
+	int sz = MLX5_ST_SZ_BYTES(ppcnt_reg);
+	void *out;
+
+	MLX5_SET(ppcnt_reg, in, local_port, 1);
+	out = pstats->IEEE_802_3_counters;
+	MLX5_SET(ppcnt_reg, in, grp, MLX5_IEEE_802_3_COUNTERS_GROUP);
+	mlx5_core_access_reg(mdev, in, sz, out, sz, MLX5_REG_PPCNT, 0, 0);
+}
+
+#define PPORT_2863_OFF(c) \
+	MLX5_BYTE_OFF(ppcnt_reg, \
+		      counter_set.eth_2863_cntrs_grp_data_layout.c##_high)
+static const struct counter_desc pport_2863_stats_desc[] = {
+	{ "rx_discards_phy", PPORT_2863_OFF(if_in_discards) },
+	{ "tx_discards_phy", PPORT_2863_OFF(if_out_discards) },
+	{ "tx_errors_phy", PPORT_2863_OFF(if_out_errors) },
+};
+
+#define NUM_PPORT_2863_COUNTERS		ARRAY_SIZE(pport_2863_stats_desc)
+
+static int mlx5e_grp_2863_get_num_stats(struct mlx5e_priv *priv)
+{
+	return NUM_PPORT_2863_COUNTERS;
+}
+
+static int mlx5e_grp_2863_fill_strings(struct mlx5e_priv *priv, u8 *data,
+				       int idx)
+{
+	int i;
+
+	for (i = 0; i < NUM_PPORT_2863_COUNTERS; i++)
+		strcpy(data + (idx++) * ETH_GSTRING_LEN, pport_2863_stats_desc[i].format);
+	return idx;
+}
+
+static int mlx5e_grp_2863_fill_stats(struct mlx5e_priv *priv, u64 *data,
+				     int idx)
+{
+	int i;
+
+	for (i = 0; i < NUM_PPORT_2863_COUNTERS; i++)
+		data[idx++] = MLX5E_READ_CTR64_BE(&priv->stats.pport.RFC_2863_counters,
+						  pport_2863_stats_desc, i);
+	return idx;
+}
+
+static void mlx5e_grp_2863_update_stats(struct mlx5e_priv *priv)
+{
+	struct mlx5e_pport_stats *pstats = &priv->stats.pport;
+	struct mlx5_core_dev *mdev = priv->mdev;
+	u32 in[MLX5_ST_SZ_DW(ppcnt_reg)] = {0};
+	int sz = MLX5_ST_SZ_BYTES(ppcnt_reg);
+	void *out;
+
+	MLX5_SET(ppcnt_reg, in, local_port, 1);
+	out = pstats->RFC_2863_counters;
+	MLX5_SET(ppcnt_reg, in, grp, MLX5_RFC_2863_COUNTERS_GROUP);
+	mlx5_core_access_reg(mdev, in, sz, out, sz, MLX5_REG_PPCNT, 0, 0);
+}
+
+#define PPORT_2819_OFF(c) \
+	MLX5_BYTE_OFF(ppcnt_reg, \
+		      counter_set.eth_2819_cntrs_grp_data_layout.c##_high)
+static const struct counter_desc pport_2819_stats_desc[] = {
+	{ "rx_undersize_pkts_phy", PPORT_2819_OFF(ether_stats_undersize_pkts) },
+	{ "rx_fragments_phy", PPORT_2819_OFF(ether_stats_fragments) },
+	{ "rx_jabbers_phy", PPORT_2819_OFF(ether_stats_jabbers) },
+	{ "rx_64_bytes_phy", PPORT_2819_OFF(ether_stats_pkts64octets) },
+	{ "rx_65_to_127_bytes_phy", PPORT_2819_OFF(ether_stats_pkts65to127octets) },
+	{ "rx_128_to_255_bytes_phy", PPORT_2819_OFF(ether_stats_pkts128to255octets) },
+	{ "rx_256_to_511_bytes_phy", PPORT_2819_OFF(ether_stats_pkts256to511octets) },
+	{ "rx_512_to_1023_bytes_phy", PPORT_2819_OFF(ether_stats_pkts512to1023octets) },
+	{ "rx_1024_to_1518_bytes_phy", PPORT_2819_OFF(ether_stats_pkts1024to1518octets) },
+	{ "rx_1519_to_2047_bytes_phy", PPORT_2819_OFF(ether_stats_pkts1519to2047octets) },
+	{ "rx_2048_to_4095_bytes_phy", PPORT_2819_OFF(ether_stats_pkts2048to4095octets) },
+	{ "rx_4096_to_8191_bytes_phy", PPORT_2819_OFF(ether_stats_pkts4096to8191octets) },
+	{ "rx_8192_to_10239_bytes_phy", PPORT_2819_OFF(ether_stats_pkts8192to10239octets) },
+};
+
+#define NUM_PPORT_2819_COUNTERS		ARRAY_SIZE(pport_2819_stats_desc)
+
+static int mlx5e_grp_2819_get_num_stats(struct mlx5e_priv *priv)
+{
+	return NUM_PPORT_2819_COUNTERS;
+}
+
+static int mlx5e_grp_2819_fill_strings(struct mlx5e_priv *priv, u8 *data,
+				       int idx)
+{
+	int i;
+
+	for (i = 0; i < NUM_PPORT_2819_COUNTERS; i++)
+		strcpy(data + (idx++) * ETH_GSTRING_LEN, pport_2819_stats_desc[i].format);
+	return idx;
+}
+
+static int mlx5e_grp_2819_fill_stats(struct mlx5e_priv *priv, u64 *data,
+				     int idx)
+{
+	int i;
+
+	for (i = 0; i < NUM_PPORT_2819_COUNTERS; i++)
+		data[idx++] = MLX5E_READ_CTR64_BE(&priv->stats.pport.RFC_2819_counters,
+						  pport_2819_stats_desc, i);
+	return idx;
+}
+
+static void mlx5e_grp_2819_update_stats(struct mlx5e_priv *priv)
+{
+	struct mlx5e_pport_stats *pstats = &priv->stats.pport;
+	struct mlx5_core_dev *mdev = priv->mdev;
+	u32 in[MLX5_ST_SZ_DW(ppcnt_reg)] = {0};
+	int sz = MLX5_ST_SZ_BYTES(ppcnt_reg);
+	void *out;
+
+	MLX5_SET(ppcnt_reg, in, local_port, 1);
+	out = pstats->RFC_2819_counters;
+	MLX5_SET(ppcnt_reg, in, grp, MLX5_RFC_2819_COUNTERS_GROUP);
+	mlx5_core_access_reg(mdev, in, sz, out, sz, MLX5_REG_PPCNT, 0, 0);
+}
+
+#define PPORT_PHY_STATISTICAL_OFF(c) \
+	MLX5_BYTE_OFF(ppcnt_reg, \
+		      counter_set.phys_layer_statistical_cntrs.c##_high)
+static const struct counter_desc pport_phy_statistical_stats_desc[] = {
+	{ "rx_pcs_symbol_err_phy", PPORT_PHY_STATISTICAL_OFF(phy_symbol_errors) },
+	{ "rx_corrected_bits_phy", PPORT_PHY_STATISTICAL_OFF(phy_corrected_bits) },
+};
+
+#define NUM_PPORT_PHY_STATISTICAL_COUNTERS ARRAY_SIZE(pport_phy_statistical_stats_desc)
+
+static int mlx5e_grp_phy_get_num_stats(struct mlx5e_priv *priv)
+{
+	/* "1" for link_down_events special counter */
+	return MLX5_CAP_PCAM_FEATURE((priv)->mdev, ppcnt_statistical_group) ?
+		NUM_PPORT_PHY_STATISTICAL_COUNTERS + 1 : 1;
+}
+
+static int mlx5e_grp_phy_fill_strings(struct mlx5e_priv *priv, u8 *data,
+				      int idx)
+{
+	int i;
+
+	strcpy(data + (idx++) * ETH_GSTRING_LEN, "link_down_events_phy");
+
+	if (!MLX5_CAP_PCAM_FEATURE((priv)->mdev, ppcnt_statistical_group))
+		return idx;
+
+	for (i = 0; i < NUM_PPORT_PHY_STATISTICAL_COUNTERS; i++)
+		strcpy(data + (idx++) * ETH_GSTRING_LEN,
+		       pport_phy_statistical_stats_desc[i].format);
+	return idx;
+}
+
+static int mlx5e_grp_phy_fill_stats(struct mlx5e_priv *priv, u64 *data, int idx)
+{
+	int i;
+
+	/* link_down_events_phy has special handling since it is not stored in __be64 format */
+	data[idx++] = MLX5_GET(ppcnt_reg, priv->stats.pport.phy_counters,
+			       counter_set.phys_layer_cntrs.link_down_events);
+
+	if (!MLX5_CAP_PCAM_FEATURE((priv)->mdev, ppcnt_statistical_group))
+		return idx;
+
+	for (i = 0; i < NUM_PPORT_PHY_STATISTICAL_COUNTERS; i++)
+		data[idx++] =
+			MLX5E_READ_CTR64_BE(&priv->stats.pport.phy_statistical_counters,
+					    pport_phy_statistical_stats_desc, i);
+	return idx;
+}
+
+static void mlx5e_grp_phy_update_stats(struct mlx5e_priv *priv)
+{
+	struct mlx5e_pport_stats *pstats = &priv->stats.pport;
+	struct mlx5_core_dev *mdev = priv->mdev;
+	u32 in[MLX5_ST_SZ_DW(ppcnt_reg)] = {0};
+	int sz = MLX5_ST_SZ_BYTES(ppcnt_reg);
+	void *out;
+
+	MLX5_SET(ppcnt_reg, in, local_port, 1);
+	out = pstats->phy_counters;
+	MLX5_SET(ppcnt_reg, in, grp, MLX5_PHYSICAL_LAYER_COUNTERS_GROUP);
+	mlx5_core_access_reg(mdev, in, sz, out, sz, MLX5_REG_PPCNT, 0, 0);
+
+	if (!MLX5_CAP_PCAM_FEATURE(mdev, ppcnt_statistical_group))
+		return;
+
+	out = pstats->phy_statistical_counters;
+	MLX5_SET(ppcnt_reg, in, grp, MLX5_PHYSICAL_LAYER_STATISTICAL_GROUP);
+	mlx5_core_access_reg(mdev, in, sz, out, sz, MLX5_REG_PPCNT, 0, 0);
+}
+
+#define PPORT_ETH_EXT_OFF(c) \
+	MLX5_BYTE_OFF(ppcnt_reg, \
+		      counter_set.eth_extended_cntrs_grp_data_layout.c##_high)
+static const struct counter_desc pport_eth_ext_stats_desc[] = {
+	{ "rx_buffer_passed_thres_phy", PPORT_ETH_EXT_OFF(rx_buffer_almost_full) },
+};
+
+#define NUM_PPORT_ETH_EXT_COUNTERS	ARRAY_SIZE(pport_eth_ext_stats_desc)
+
+static int mlx5e_grp_eth_ext_get_num_stats(struct mlx5e_priv *priv)
+{
+	if (MLX5_CAP_PCAM_FEATURE((priv)->mdev, rx_buffer_fullness_counters))
+		return NUM_PPORT_ETH_EXT_COUNTERS;
+
+	return 0;
+}
+
+static int mlx5e_grp_eth_ext_fill_strings(struct mlx5e_priv *priv, u8 *data,
+					  int idx)
+{
+	int i;
+
+	if (MLX5_CAP_PCAM_FEATURE((priv)->mdev, rx_buffer_fullness_counters))
+		for (i = 0; i < NUM_PPORT_ETH_EXT_COUNTERS; i++)
+			strcpy(data + (idx++) * ETH_GSTRING_LEN,
+			       pport_eth_ext_stats_desc[i].format);
+	return idx;
+}
+
+static int mlx5e_grp_eth_ext_fill_stats(struct mlx5e_priv *priv, u64 *data,
+					int idx)
+{
+	int i;
+
+	if (MLX5_CAP_PCAM_FEATURE((priv)->mdev, rx_buffer_fullness_counters))
+		for (i = 0; i < NUM_PPORT_ETH_EXT_COUNTERS; i++)
+			data[idx++] =
+				MLX5E_READ_CTR64_BE(&priv->stats.pport.eth_ext_counters,
+						    pport_eth_ext_stats_desc, i);
+	return idx;
+}
+
+static void mlx5e_grp_eth_ext_update_stats(struct mlx5e_priv *priv)
+{
+	struct mlx5e_pport_stats *pstats = &priv->stats.pport;
+	struct mlx5_core_dev *mdev = priv->mdev;
+	u32 in[MLX5_ST_SZ_DW(ppcnt_reg)] = {0};
+	int sz = MLX5_ST_SZ_BYTES(ppcnt_reg);
+	void *out;
+
+	if (!MLX5_CAP_PCAM_FEATURE(mdev, rx_buffer_fullness_counters))
+		return;
+
+	MLX5_SET(ppcnt_reg, in, local_port, 1);
+	out = pstats->eth_ext_counters;
+	MLX5_SET(ppcnt_reg, in, grp, MLX5_ETHERNET_EXTENDED_COUNTERS_GROUP);
+	mlx5_core_access_reg(mdev, in, sz, out, sz, MLX5_REG_PPCNT, 0, 0);
+}
+
+#define PCIE_PERF_OFF(c) \
+	MLX5_BYTE_OFF(mpcnt_reg, counter_set.pcie_perf_cntrs_grp_data_layout.c)
+static const struct counter_desc pcie_perf_stats_desc[] = {
+	{ "rx_pci_signal_integrity", PCIE_PERF_OFF(rx_errors) },
+	{ "tx_pci_signal_integrity", PCIE_PERF_OFF(tx_errors) },
+};
+
+#define PCIE_PERF_OFF64(c) \
+	MLX5_BYTE_OFF(mpcnt_reg, counter_set.pcie_perf_cntrs_grp_data_layout.c##_high)
+static const struct counter_desc pcie_perf_stats_desc64[] = {
+	{ "outbound_pci_buffer_overflow", PCIE_PERF_OFF64(tx_overflow_buffer_pkt) },
+};
+
+static const struct counter_desc pcie_perf_stall_stats_desc[] = {
+	{ "outbound_pci_stalled_rd", PCIE_PERF_OFF(outbound_stalled_reads) },
+	{ "outbound_pci_stalled_wr", PCIE_PERF_OFF(outbound_stalled_writes) },
+	{ "outbound_pci_stalled_rd_events", PCIE_PERF_OFF(outbound_stalled_reads_events) },
+	{ "outbound_pci_stalled_wr_events", PCIE_PERF_OFF(outbound_stalled_writes_events) },
+};
+
+#define NUM_PCIE_PERF_COUNTERS		ARRAY_SIZE(pcie_perf_stats_desc)
+#define NUM_PCIE_PERF_COUNTERS64	ARRAY_SIZE(pcie_perf_stats_desc64)
+#define NUM_PCIE_PERF_STALL_COUNTERS	ARRAY_SIZE(pcie_perf_stall_stats_desc)
+
+static int mlx5e_grp_pcie_get_num_stats(struct mlx5e_priv *priv)
+{
+	int num_stats = 0;
+
+	if (MLX5_CAP_MCAM_FEATURE((priv)->mdev, pcie_performance_group))
+		num_stats += NUM_PCIE_PERF_COUNTERS;
+
+	if (MLX5_CAP_MCAM_FEATURE((priv)->mdev, tx_overflow_buffer_pkt))
+		num_stats += NUM_PCIE_PERF_COUNTERS64;
+
+	if (MLX5_CAP_MCAM_FEATURE((priv)->mdev, pcie_outbound_stalled))
+		num_stats += NUM_PCIE_PERF_STALL_COUNTERS;
+
+	return num_stats;
+}
+
+static int mlx5e_grp_pcie_fill_strings(struct mlx5e_priv *priv, u8 *data,
+				       int idx)
+{
+	int i;
+
+	if (MLX5_CAP_MCAM_FEATURE((priv)->mdev, pcie_performance_group))
+		for (i = 0; i < NUM_PCIE_PERF_COUNTERS; i++)
+			strcpy(data + (idx++) * ETH_GSTRING_LEN,
+			       pcie_perf_stats_desc[i].format);
+
+	if (MLX5_CAP_MCAM_FEATURE((priv)->mdev, tx_overflow_buffer_pkt))
+		for (i = 0; i < NUM_PCIE_PERF_COUNTERS64; i++)
+			strcpy(data + (idx++) * ETH_GSTRING_LEN,
+			       pcie_perf_stats_desc64[i].format);
+
+	if (MLX5_CAP_MCAM_FEATURE((priv)->mdev, pcie_outbound_stalled))
+		for (i = 0; i < NUM_PCIE_PERF_STALL_COUNTERS; i++)
+			strcpy(data + (idx++) * ETH_GSTRING_LEN,
+			       pcie_perf_stall_stats_desc[i].format);
+	return idx;
+}
+
+static int mlx5e_grp_pcie_fill_stats(struct mlx5e_priv *priv, u64 *data,
+				     int idx)
+{
+	int i;
+
+	if (MLX5_CAP_MCAM_FEATURE((priv)->mdev, pcie_performance_group))
+		for (i = 0; i < NUM_PCIE_PERF_COUNTERS; i++)
+			data[idx++] =
+				MLX5E_READ_CTR32_BE(&priv->stats.pcie.pcie_perf_counters,
+						    pcie_perf_stats_desc, i);
+
+	if (MLX5_CAP_MCAM_FEATURE((priv)->mdev, tx_overflow_buffer_pkt))
+		for (i = 0; i < NUM_PCIE_PERF_COUNTERS64; i++)
+			data[idx++] =
+				MLX5E_READ_CTR64_BE(&priv->stats.pcie.pcie_perf_counters,
+						    pcie_perf_stats_desc64, i);
+
+	if (MLX5_CAP_MCAM_FEATURE((priv)->mdev, pcie_outbound_stalled))
+		for (i = 0; i < NUM_PCIE_PERF_STALL_COUNTERS; i++)
+			data[idx++] =
+				MLX5E_READ_CTR32_BE(&priv->stats.pcie.pcie_perf_counters,
+						    pcie_perf_stall_stats_desc, i);
+	return idx;
+}
+
+static void mlx5e_grp_pcie_update_stats(struct mlx5e_priv *priv)
+{
+	struct mlx5e_pcie_stats *pcie_stats = &priv->stats.pcie;
+	struct mlx5_core_dev *mdev = priv->mdev;
+	u32 in[MLX5_ST_SZ_DW(mpcnt_reg)] = {0};
+	int sz = MLX5_ST_SZ_BYTES(mpcnt_reg);
+	void *out;
+
+	if (!MLX5_CAP_MCAM_FEATURE(mdev, pcie_performance_group))
+		return;
+
+	out = pcie_stats->pcie_perf_counters;
+	MLX5_SET(mpcnt_reg, in, grp, MLX5_PCIE_PERFORMANCE_COUNTERS_GROUP);
+	mlx5_core_access_reg(mdev, in, sz, out, sz, MLX5_REG_MPCNT, 0, 0);
+}
+
+#define PPORT_PER_PRIO_OFF(c) \
+	MLX5_BYTE_OFF(ppcnt_reg, \
+		      counter_set.eth_per_prio_grp_data_layout.c##_high)
+static const struct counter_desc pport_per_prio_traffic_stats_desc[] = {
+	{ "rx_prio%d_bytes", PPORT_PER_PRIO_OFF(rx_octets) },
+	{ "rx_prio%d_packets", PPORT_PER_PRIO_OFF(rx_frames) },
+	{ "tx_prio%d_bytes", PPORT_PER_PRIO_OFF(tx_octets) },
+	{ "tx_prio%d_packets", PPORT_PER_PRIO_OFF(tx_frames) },
+};
+
+#define NUM_PPORT_PER_PRIO_TRAFFIC_COUNTERS	ARRAY_SIZE(pport_per_prio_traffic_stats_desc)
+
+static int mlx5e_grp_per_prio_traffic_get_num_stats(void)
+{
+	return NUM_PPORT_PER_PRIO_TRAFFIC_COUNTERS * NUM_PPORT_PRIO;
+}
+
+static int mlx5e_grp_per_prio_traffic_fill_strings(struct mlx5e_priv *priv,
+						   u8 *data,
+						   int idx)
+{
+	int i, prio;
+
+	for (prio = 0; prio < NUM_PPORT_PRIO; prio++) {
+		for (i = 0; i < NUM_PPORT_PER_PRIO_TRAFFIC_COUNTERS; i++)
+			sprintf(data + (idx++) * ETH_GSTRING_LEN,
+				pport_per_prio_traffic_stats_desc[i].format, prio);
+	}
+
+	return idx;
+}
+
+static int mlx5e_grp_per_prio_traffic_fill_stats(struct mlx5e_priv *priv,
+						 u64 *data,
+						 int idx)
+{
+	int i, prio;
+
+	for (prio = 0; prio < NUM_PPORT_PRIO; prio++) {
+		for (i = 0; i < NUM_PPORT_PER_PRIO_TRAFFIC_COUNTERS; i++)
+			data[idx++] =
+				MLX5E_READ_CTR64_BE(&priv->stats.pport.per_prio_counters[prio],
+						    pport_per_prio_traffic_stats_desc, i);
+	}
+
+	return idx;
+}
+
+static const struct counter_desc pport_per_prio_pfc_stats_desc[] = {
+	/* %s is "global" or "prio{i}" */
+	{ "rx_%s_pause", PPORT_PER_PRIO_OFF(rx_pause) },
+	{ "rx_%s_pause_duration", PPORT_PER_PRIO_OFF(rx_pause_duration) },
+	{ "tx_%s_pause", PPORT_PER_PRIO_OFF(tx_pause) },
+	{ "tx_%s_pause_duration", PPORT_PER_PRIO_OFF(tx_pause_duration) },
+	{ "rx_%s_pause_transition", PPORT_PER_PRIO_OFF(rx_pause_transition) },
+};
+
+static const struct counter_desc pport_pfc_stall_stats_desc[] = {
+	{ "tx_pause_storm_warning_events ", PPORT_PER_PRIO_OFF(device_stall_minor_watermark_cnt) },
+	{ "tx_pause_storm_error_events", PPORT_PER_PRIO_OFF(device_stall_critical_watermark_cnt) },
+};
+
+#define NUM_PPORT_PER_PRIO_PFC_COUNTERS		ARRAY_SIZE(pport_per_prio_pfc_stats_desc)
+#define NUM_PPORT_PFC_STALL_COUNTERS(priv)	(ARRAY_SIZE(pport_pfc_stall_stats_desc) * \
+						 MLX5_CAP_PCAM_FEATURE((priv)->mdev, pfcc_mask) * \
+						 MLX5_CAP_DEBUG((priv)->mdev, stall_detect))
+
+static unsigned long mlx5e_query_pfc_combined(struct mlx5e_priv *priv)
+{
+	struct mlx5_core_dev *mdev = priv->mdev;
+	u8 pfc_en_tx;
+	u8 pfc_en_rx;
+	int err;
+
+	if (MLX5_CAP_GEN(mdev, port_type) != MLX5_CAP_PORT_TYPE_ETH)
+		return 0;
+
+	err = mlx5_query_port_pfc(mdev, &pfc_en_tx, &pfc_en_rx);
+
+	return err ? 0 : pfc_en_tx | pfc_en_rx;
+}
+
+static bool mlx5e_query_global_pause_combined(struct mlx5e_priv *priv)
+{
+	struct mlx5_core_dev *mdev = priv->mdev;
+	u32 rx_pause;
+	u32 tx_pause;
+	int err;
+
+	if (MLX5_CAP_GEN(mdev, port_type) != MLX5_CAP_PORT_TYPE_ETH)
+		return false;
+
+	err = mlx5_query_port_pause(mdev, &rx_pause, &tx_pause);
+
+	return err ? false : rx_pause | tx_pause;
+}
+
+static int mlx5e_grp_per_prio_pfc_get_num_stats(struct mlx5e_priv *priv)
+{
+	return (mlx5e_query_global_pause_combined(priv) +
+		hweight8(mlx5e_query_pfc_combined(priv))) *
+		NUM_PPORT_PER_PRIO_PFC_COUNTERS +
+		NUM_PPORT_PFC_STALL_COUNTERS(priv);
+}
+
+static int mlx5e_grp_per_prio_pfc_fill_strings(struct mlx5e_priv *priv,
+					       u8 *data,
+					       int idx)
+{
+	unsigned long pfc_combined;
+	int i, prio;
+
+	pfc_combined = mlx5e_query_pfc_combined(priv);
+	for_each_set_bit(prio, &pfc_combined, NUM_PPORT_PRIO) {
+		for (i = 0; i < NUM_PPORT_PER_PRIO_PFC_COUNTERS; i++) {
+			char pfc_string[ETH_GSTRING_LEN];
+
+			snprintf(pfc_string, sizeof(pfc_string), "prio%d", prio);
+			sprintf(data + (idx++) * ETH_GSTRING_LEN,
+				pport_per_prio_pfc_stats_desc[i].format, pfc_string);
+		}
+	}
+
+	if (mlx5e_query_global_pause_combined(priv)) {
+		for (i = 0; i < NUM_PPORT_PER_PRIO_PFC_COUNTERS; i++) {
+			sprintf(data + (idx++) * ETH_GSTRING_LEN,
+				pport_per_prio_pfc_stats_desc[i].format, "global");
+		}
+	}
+
+	for (i = 0; i < NUM_PPORT_PFC_STALL_COUNTERS(priv); i++)
+		strcpy(data + (idx++) * ETH_GSTRING_LEN,
+		       pport_pfc_stall_stats_desc[i].format);
+
+	return idx;
+}
+
+static int mlx5e_grp_per_prio_pfc_fill_stats(struct mlx5e_priv *priv,
+					     u64 *data,
+					     int idx)
+{
+	unsigned long pfc_combined;
+	int i, prio;
+
+	pfc_combined = mlx5e_query_pfc_combined(priv);
+	for_each_set_bit(prio, &pfc_combined, NUM_PPORT_PRIO) {
+		for (i = 0; i < NUM_PPORT_PER_PRIO_PFC_COUNTERS; i++) {
+			data[idx++] =
+				MLX5E_READ_CTR64_BE(&priv->stats.pport.per_prio_counters[prio],
+						    pport_per_prio_pfc_stats_desc, i);
+		}
+	}
+
+	if (mlx5e_query_global_pause_combined(priv)) {
+		for (i = 0; i < NUM_PPORT_PER_PRIO_PFC_COUNTERS; i++) {
+			data[idx++] =
+				MLX5E_READ_CTR64_BE(&priv->stats.pport.per_prio_counters[0],
+						    pport_per_prio_pfc_stats_desc, i);
+		}
+	}
+
+	for (i = 0; i < NUM_PPORT_PFC_STALL_COUNTERS(priv); i++)
+		data[idx++] = MLX5E_READ_CTR64_BE(&priv->stats.pport.per_prio_counters[0],
+						  pport_pfc_stall_stats_desc, i);
+
+	return idx;
+}
+
+static int mlx5e_grp_per_prio_get_num_stats(struct mlx5e_priv *priv)
+{
+	return mlx5e_grp_per_prio_traffic_get_num_stats() +
+		mlx5e_grp_per_prio_pfc_get_num_stats(priv);
+}
+
+static int mlx5e_grp_per_prio_fill_strings(struct mlx5e_priv *priv, u8 *data,
+					   int idx)
+{
+	idx = mlx5e_grp_per_prio_traffic_fill_strings(priv, data, idx);
+	idx = mlx5e_grp_per_prio_pfc_fill_strings(priv, data, idx);
+	return idx;
+}
+
+static int mlx5e_grp_per_prio_fill_stats(struct mlx5e_priv *priv, u64 *data,
+					 int idx)
+{
+	idx = mlx5e_grp_per_prio_traffic_fill_stats(priv, data, idx);
+	idx = mlx5e_grp_per_prio_pfc_fill_stats(priv, data, idx);
+	return idx;
+}
+
+static void mlx5e_grp_per_prio_update_stats(struct mlx5e_priv *priv)
+{
+	struct mlx5e_pport_stats *pstats = &priv->stats.pport;
+	struct mlx5_core_dev *mdev = priv->mdev;
+	u32 in[MLX5_ST_SZ_DW(ppcnt_reg)] = {0};
+	int sz = MLX5_ST_SZ_BYTES(ppcnt_reg);
+	int prio;
+	void *out;
+
+	MLX5_SET(ppcnt_reg, in, local_port, 1);
+	MLX5_SET(ppcnt_reg, in, grp, MLX5_PER_PRIORITY_COUNTERS_GROUP);
+	for (prio = 0; prio < NUM_PPORT_PRIO; prio++) {
+		out = pstats->per_prio_counters[prio];
+		MLX5_SET(ppcnt_reg, in, prio_tc, prio);
+		mlx5_core_access_reg(mdev, in, sz, out, sz,
+				     MLX5_REG_PPCNT, 0, 0);
+	}
+}
+
+static const struct counter_desc mlx5e_pme_status_desc[] = {
+	{ "module_unplug", 8 },
+};
+
+static const struct counter_desc mlx5e_pme_error_desc[] = {
+	{ "module_bus_stuck", 16 },       /* bus stuck (I2C or data shorted) */
+	{ "module_high_temp", 48 },       /* high temperature */
+	{ "module_bad_shorted", 56 },    /* bad or shorted cable/module */
+};
+
+#define NUM_PME_STATUS_STATS		ARRAY_SIZE(mlx5e_pme_status_desc)
+#define NUM_PME_ERR_STATS		ARRAY_SIZE(mlx5e_pme_error_desc)
+
+static int mlx5e_grp_pme_get_num_stats(struct mlx5e_priv *priv)
+{
+	return NUM_PME_STATUS_STATS + NUM_PME_ERR_STATS;
+}
+
+static int mlx5e_grp_pme_fill_strings(struct mlx5e_priv *priv, u8 *data,
+				      int idx)
+{
+	int i;
+
+	for (i = 0; i < NUM_PME_STATUS_STATS; i++)
+		strcpy(data + (idx++) * ETH_GSTRING_LEN, mlx5e_pme_status_desc[i].format);
+
+	for (i = 0; i < NUM_PME_ERR_STATS; i++)
+		strcpy(data + (idx++) * ETH_GSTRING_LEN, mlx5e_pme_error_desc[i].format);
+
+	return idx;
+}
+
+static int mlx5e_grp_pme_fill_stats(struct mlx5e_priv *priv, u64 *data,
+				    int idx)
+{
+	struct mlx5_priv *mlx5_priv = &priv->mdev->priv;
+	int i;
+
+	for (i = 0; i < NUM_PME_STATUS_STATS; i++)
+		data[idx++] = MLX5E_READ_CTR64_CPU(mlx5_priv->pme_stats.status_counters,
+						   mlx5e_pme_status_desc, i);
+
+	for (i = 0; i < NUM_PME_ERR_STATS; i++)
+		data[idx++] = MLX5E_READ_CTR64_CPU(mlx5_priv->pme_stats.error_counters,
+						   mlx5e_pme_error_desc, i);
+
+	return idx;
+}
+
+static int mlx5e_grp_ipsec_get_num_stats(struct mlx5e_priv *priv)
+{
+	return mlx5e_ipsec_get_count(priv);
+}
+
+static int mlx5e_grp_ipsec_fill_strings(struct mlx5e_priv *priv, u8 *data,
+					int idx)
+{
+	return idx + mlx5e_ipsec_get_strings(priv,
+					     data + idx * ETH_GSTRING_LEN);
+}
+
+static int mlx5e_grp_ipsec_fill_stats(struct mlx5e_priv *priv, u64 *data,
+				      int idx)
+{
+	return idx + mlx5e_ipsec_get_stats(priv, data + idx);
+}
+
+static void mlx5e_grp_ipsec_update_stats(struct mlx5e_priv *priv)
+{
+	mlx5e_ipsec_update_stats(priv);
+}
+
+static int mlx5e_grp_tls_get_num_stats(struct mlx5e_priv *priv)
+{
+	return mlx5e_tls_get_count(priv);
+}
+
+static int mlx5e_grp_tls_fill_strings(struct mlx5e_priv *priv, u8 *data,
+				      int idx)
+{
+	return idx + mlx5e_tls_get_strings(priv, data + idx * ETH_GSTRING_LEN);
+}
+
+static int mlx5e_grp_tls_fill_stats(struct mlx5e_priv *priv, u64 *data, int idx)
+{
+	return idx + mlx5e_tls_get_stats(priv, data + idx);
+}
+
+static const struct counter_desc rq_stats_desc[] = {
+	{ MLX5E_DECLARE_RX_STAT(struct mlx5e_rq_stats, packets) },
+	{ MLX5E_DECLARE_RX_STAT(struct mlx5e_rq_stats, bytes) },
+	{ MLX5E_DECLARE_RX_STAT(struct mlx5e_rq_stats, csum_complete) },
+	{ MLX5E_DECLARE_RX_STAT(struct mlx5e_rq_stats, csum_complete_tail) },
+	{ MLX5E_DECLARE_RX_STAT(struct mlx5e_rq_stats, csum_complete_tail_slow) },
+	{ MLX5E_DECLARE_RX_STAT(struct mlx5e_rq_stats, csum_unnecessary) },
+	{ MLX5E_DECLARE_RX_STAT(struct mlx5e_rq_stats, csum_unnecessary_inner) },
+	{ MLX5E_DECLARE_RX_STAT(struct mlx5e_rq_stats, csum_none) },
+	{ MLX5E_DECLARE_RX_STAT(struct mlx5e_rq_stats, xdp_drop) },
+	{ MLX5E_DECLARE_RX_STAT(struct mlx5e_rq_stats, xdp_redirect) },
+	{ MLX5E_DECLARE_RX_STAT(struct mlx5e_rq_stats, lro_packets) },
+	{ MLX5E_DECLARE_RX_STAT(struct mlx5e_rq_stats, lro_bytes) },
+	{ MLX5E_DECLARE_RX_STAT(struct mlx5e_rq_stats, ecn_mark) },
+	{ MLX5E_DECLARE_RX_STAT(struct mlx5e_rq_stats, removed_vlan_packets) },
+	{ MLX5E_DECLARE_RX_STAT(struct mlx5e_rq_stats, wqe_err) },
+	{ MLX5E_DECLARE_RX_STAT(struct mlx5e_rq_stats, mpwqe_filler_cqes) },
+	{ MLX5E_DECLARE_RX_STAT(struct mlx5e_rq_stats, mpwqe_filler_strides) },
+	{ MLX5E_DECLARE_RX_STAT(struct mlx5e_rq_stats, oversize_pkts_sw_drop) },
+	{ MLX5E_DECLARE_RX_STAT(struct mlx5e_rq_stats, buff_alloc_err) },
+	{ MLX5E_DECLARE_RX_STAT(struct mlx5e_rq_stats, cqe_compress_blks) },
+	{ MLX5E_DECLARE_RX_STAT(struct mlx5e_rq_stats, cqe_compress_pkts) },
+	{ MLX5E_DECLARE_RX_STAT(struct mlx5e_rq_stats, page_reuse) },
+	{ MLX5E_DECLARE_RX_STAT(struct mlx5e_rq_stats, cache_reuse) },
+	{ MLX5E_DECLARE_RX_STAT(struct mlx5e_rq_stats, cache_full) },
+	{ MLX5E_DECLARE_RX_STAT(struct mlx5e_rq_stats, cache_empty) },
+	{ MLX5E_DECLARE_RX_STAT(struct mlx5e_rq_stats, cache_busy) },
+	{ MLX5E_DECLARE_RX_STAT(struct mlx5e_rq_stats, cache_waive) },
+	{ MLX5E_DECLARE_RX_STAT(struct mlx5e_rq_stats, congst_umr) },
+};
+
+static const struct counter_desc sq_stats_desc[] = {
+	{ MLX5E_DECLARE_TX_STAT(struct mlx5e_sq_stats, packets) },
+	{ MLX5E_DECLARE_TX_STAT(struct mlx5e_sq_stats, bytes) },
+	{ MLX5E_DECLARE_TX_STAT(struct mlx5e_sq_stats, tso_packets) },
+	{ MLX5E_DECLARE_TX_STAT(struct mlx5e_sq_stats, tso_bytes) },
+	{ MLX5E_DECLARE_TX_STAT(struct mlx5e_sq_stats, tso_inner_packets) },
+	{ MLX5E_DECLARE_TX_STAT(struct mlx5e_sq_stats, tso_inner_bytes) },
+	{ MLX5E_DECLARE_TX_STAT(struct mlx5e_sq_stats, csum_partial) },
+	{ MLX5E_DECLARE_TX_STAT(struct mlx5e_sq_stats, csum_partial_inner) },
+	{ MLX5E_DECLARE_TX_STAT(struct mlx5e_sq_stats, added_vlan_packets) },
+	{ MLX5E_DECLARE_TX_STAT(struct mlx5e_sq_stats, nop) },
+	{ MLX5E_DECLARE_TX_STAT(struct mlx5e_sq_stats, csum_none) },
+	{ MLX5E_DECLARE_TX_STAT(struct mlx5e_sq_stats, stopped) },
+	{ MLX5E_DECLARE_TX_STAT(struct mlx5e_sq_stats, dropped) },
+	{ MLX5E_DECLARE_TX_STAT(struct mlx5e_sq_stats, xmit_more) },
+	{ MLX5E_DECLARE_TX_STAT(struct mlx5e_sq_stats, recover) },
+	{ MLX5E_DECLARE_TX_STAT(struct mlx5e_sq_stats, cqes) },
+	{ MLX5E_DECLARE_TX_STAT(struct mlx5e_sq_stats, wake) },
+	{ MLX5E_DECLARE_TX_STAT(struct mlx5e_sq_stats, cqe_err) },
+};
+
+static const struct counter_desc rq_xdpsq_stats_desc[] = {
+	{ MLX5E_DECLARE_RQ_XDPSQ_STAT(struct mlx5e_xdpsq_stats, xmit) },
+	{ MLX5E_DECLARE_RQ_XDPSQ_STAT(struct mlx5e_xdpsq_stats, full) },
+	{ MLX5E_DECLARE_RQ_XDPSQ_STAT(struct mlx5e_xdpsq_stats, err) },
+	{ MLX5E_DECLARE_RQ_XDPSQ_STAT(struct mlx5e_xdpsq_stats, cqes) },
+};
+
+static const struct counter_desc xdpsq_stats_desc[] = {
+	{ MLX5E_DECLARE_XDPSQ_STAT(struct mlx5e_xdpsq_stats, xmit) },
+	{ MLX5E_DECLARE_XDPSQ_STAT(struct mlx5e_xdpsq_stats, full) },
+	{ MLX5E_DECLARE_XDPSQ_STAT(struct mlx5e_xdpsq_stats, err) },
+	{ MLX5E_DECLARE_XDPSQ_STAT(struct mlx5e_xdpsq_stats, cqes) },
+};
+
+static const struct counter_desc ch_stats_desc[] = {
+	{ MLX5E_DECLARE_CH_STAT(struct mlx5e_ch_stats, events) },
+	{ MLX5E_DECLARE_CH_STAT(struct mlx5e_ch_stats, poll) },
+	{ MLX5E_DECLARE_CH_STAT(struct mlx5e_ch_stats, arm) },
+	{ MLX5E_DECLARE_CH_STAT(struct mlx5e_ch_stats, aff_change) },
+	{ MLX5E_DECLARE_CH_STAT(struct mlx5e_ch_stats, eq_rearm) },
+};
+
+#define NUM_RQ_STATS			ARRAY_SIZE(rq_stats_desc)
+#define NUM_SQ_STATS			ARRAY_SIZE(sq_stats_desc)
+#define NUM_XDPSQ_STATS			ARRAY_SIZE(xdpsq_stats_desc)
+#define NUM_RQ_XDPSQ_STATS		ARRAY_SIZE(rq_xdpsq_stats_desc)
+#define NUM_CH_STATS			ARRAY_SIZE(ch_stats_desc)
+
+static int mlx5e_grp_channels_get_num_stats(struct mlx5e_priv *priv)
+{
+	int max_nch = priv->profile->max_nch(priv->mdev);
+
+	return (NUM_RQ_STATS * max_nch) +
+	       (NUM_CH_STATS * max_nch) +
+	       (NUM_SQ_STATS * max_nch * priv->max_opened_tc) +
+	       (NUM_RQ_XDPSQ_STATS * max_nch) +
+	       (NUM_XDPSQ_STATS * max_nch);
+}
+
+static int mlx5e_grp_channels_fill_strings(struct mlx5e_priv *priv, u8 *data,
+					   int idx)
+{
+	int max_nch = priv->profile->max_nch(priv->mdev);
+	int i, j, tc;
+
+	for (i = 0; i < max_nch; i++)
+		for (j = 0; j < NUM_CH_STATS; j++)
+			sprintf(data + (idx++) * ETH_GSTRING_LEN,
+				ch_stats_desc[j].format, i);
+
+	for (i = 0; i < max_nch; i++) {
+		for (j = 0; j < NUM_RQ_STATS; j++)
+			sprintf(data + (idx++) * ETH_GSTRING_LEN,
+				rq_stats_desc[j].format, i);
+		for (j = 0; j < NUM_RQ_XDPSQ_STATS; j++)
+			sprintf(data + (idx++) * ETH_GSTRING_LEN,
+				rq_xdpsq_stats_desc[j].format, i);
+	}
+
+	for (tc = 0; tc < priv->max_opened_tc; tc++)
+		for (i = 0; i < max_nch; i++)
+			for (j = 0; j < NUM_SQ_STATS; j++)
+				sprintf(data + (idx++) * ETH_GSTRING_LEN,
+					sq_stats_desc[j].format,
+					priv->channel_tc2txq[i][tc]);
+
+	for (i = 0; i < max_nch; i++)
+		for (j = 0; j < NUM_XDPSQ_STATS; j++)
+			sprintf(data + (idx++) * ETH_GSTRING_LEN,
+				xdpsq_stats_desc[j].format, i);
+
+	return idx;
+}
+
+static int mlx5e_grp_channels_fill_stats(struct mlx5e_priv *priv, u64 *data,
+					 int idx)
+{
+	int max_nch = priv->profile->max_nch(priv->mdev);
+	int i, j, tc;
+
+	for (i = 0; i < max_nch; i++)
+		for (j = 0; j < NUM_CH_STATS; j++)
+			data[idx++] =
+				MLX5E_READ_CTR64_CPU(&priv->channel_stats[i].ch,
+						     ch_stats_desc, j);
+
+	for (i = 0; i < max_nch; i++) {
+		for (j = 0; j < NUM_RQ_STATS; j++)
+			data[idx++] =
+				MLX5E_READ_CTR64_CPU(&priv->channel_stats[i].rq,
+						     rq_stats_desc, j);
+		for (j = 0; j < NUM_RQ_XDPSQ_STATS; j++)
+			data[idx++] =
+				MLX5E_READ_CTR64_CPU(&priv->channel_stats[i].rq_xdpsq,
+						     rq_xdpsq_stats_desc, j);
+	}
+
+	for (tc = 0; tc < priv->max_opened_tc; tc++)
+		for (i = 0; i < max_nch; i++)
+			for (j = 0; j < NUM_SQ_STATS; j++)
+				data[idx++] =
+					MLX5E_READ_CTR64_CPU(&priv->channel_stats[i].sq[tc],
+							     sq_stats_desc, j);
+
+	for (i = 0; i < max_nch; i++)
+		for (j = 0; j < NUM_XDPSQ_STATS; j++)
+			data[idx++] =
+				MLX5E_READ_CTR64_CPU(&priv->channel_stats[i].xdpsq,
+						     xdpsq_stats_desc, j);
+
+	return idx;
+}
+
+/* The stats groups order is opposite to the update_stats() order calls */
+const struct mlx5e_stats_grp mlx5e_stats_grps[] = {
+	{
+		.get_num_stats = mlx5e_grp_sw_get_num_stats,
+		.fill_strings = mlx5e_grp_sw_fill_strings,
+		.fill_stats = mlx5e_grp_sw_fill_stats,
+		.update_stats = mlx5e_grp_sw_update_stats,
+	},
+	{
+		.get_num_stats = mlx5e_grp_q_get_num_stats,
+		.fill_strings = mlx5e_grp_q_fill_strings,
+		.fill_stats = mlx5e_grp_q_fill_stats,
+		.update_stats_mask = MLX5E_NDO_UPDATE_STATS,
+		.update_stats = mlx5e_grp_q_update_stats,
+	},
+	{
+		.get_num_stats = mlx5e_grp_vnic_env_get_num_stats,
+		.fill_strings = mlx5e_grp_vnic_env_fill_strings,
+		.fill_stats = mlx5e_grp_vnic_env_fill_stats,
+		.update_stats = mlx5e_grp_vnic_env_update_stats,
+	},
+	{
+		.get_num_stats = mlx5e_grp_vport_get_num_stats,
+		.fill_strings = mlx5e_grp_vport_fill_strings,
+		.fill_stats = mlx5e_grp_vport_fill_stats,
+		.update_stats_mask = MLX5E_NDO_UPDATE_STATS,
+		.update_stats = mlx5e_grp_vport_update_stats,
+	},
+	{
+		.get_num_stats = mlx5e_grp_802_3_get_num_stats,
+		.fill_strings = mlx5e_grp_802_3_fill_strings,
+		.fill_stats = mlx5e_grp_802_3_fill_stats,
+		.update_stats_mask = MLX5E_NDO_UPDATE_STATS,
+		.update_stats = mlx5e_grp_802_3_update_stats,
+	},
+	{
+		.get_num_stats = mlx5e_grp_2863_get_num_stats,
+		.fill_strings = mlx5e_grp_2863_fill_strings,
+		.fill_stats = mlx5e_grp_2863_fill_stats,
+		.update_stats = mlx5e_grp_2863_update_stats,
+	},
+	{
+		.get_num_stats = mlx5e_grp_2819_get_num_stats,
+		.fill_strings = mlx5e_grp_2819_fill_strings,
+		.fill_stats = mlx5e_grp_2819_fill_stats,
+		.update_stats = mlx5e_grp_2819_update_stats,
+	},
+	{
+		.get_num_stats = mlx5e_grp_phy_get_num_stats,
+		.fill_strings = mlx5e_grp_phy_fill_strings,
+		.fill_stats = mlx5e_grp_phy_fill_stats,
+		.update_stats = mlx5e_grp_phy_update_stats,
+	},
+	{
+		.get_num_stats = mlx5e_grp_eth_ext_get_num_stats,
+		.fill_strings = mlx5e_grp_eth_ext_fill_strings,
+		.fill_stats = mlx5e_grp_eth_ext_fill_stats,
+		.update_stats = mlx5e_grp_eth_ext_update_stats,
+	},
+	{
+		.get_num_stats = mlx5e_grp_pcie_get_num_stats,
+		.fill_strings = mlx5e_grp_pcie_fill_strings,
+		.fill_stats = mlx5e_grp_pcie_fill_stats,
+		.update_stats = mlx5e_grp_pcie_update_stats,
+	},
+	{
+		.get_num_stats = mlx5e_grp_per_prio_get_num_stats,
+		.fill_strings = mlx5e_grp_per_prio_fill_strings,
+		.fill_stats = mlx5e_grp_per_prio_fill_stats,
+		.update_stats = mlx5e_grp_per_prio_update_stats,
+	},
+	{
+		.get_num_stats = mlx5e_grp_pme_get_num_stats,
+		.fill_strings = mlx5e_grp_pme_fill_strings,
+		.fill_stats = mlx5e_grp_pme_fill_stats,
+	},
+	{
+		.get_num_stats = mlx5e_grp_ipsec_get_num_stats,
+		.fill_strings = mlx5e_grp_ipsec_fill_strings,
+		.fill_stats = mlx5e_grp_ipsec_fill_stats,
+		.update_stats = mlx5e_grp_ipsec_update_stats,
+	},
+	{
+		.get_num_stats = mlx5e_grp_tls_get_num_stats,
+		.fill_strings = mlx5e_grp_tls_fill_strings,
+		.fill_stats = mlx5e_grp_tls_fill_stats,
+	},
+	{
+		.get_num_stats = mlx5e_grp_channels_get_num_stats,
+		.fill_strings = mlx5e_grp_channels_fill_strings,
+		.fill_stats = mlx5e_grp_channels_fill_stats,
+	}
+};
+
+const int mlx5e_num_stats_grps = ARRAY_SIZE(mlx5e_stats_grps);
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_stats.h b/drivers/net/ethernet/mellanox/mlx5/core/en_stats.h
new file mode 100644
index 000000000..3ea8033ed
--- /dev/null
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_stats.h
@@ -0,0 +1,284 @@
+/*
+ * Copyright (c) 2015-2016, 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 __MLX5_EN_STATS_H__
+#define __MLX5_EN_STATS_H__
+
+#define MLX5E_READ_CTR64_CPU(ptr, dsc, i) \
+	(*(u64 *)((char *)ptr + dsc[i].offset))
+#define MLX5E_READ_CTR64_BE(ptr, dsc, i) \
+	be64_to_cpu(*(__be64 *)((char *)ptr + dsc[i].offset))
+#define MLX5E_READ_CTR32_CPU(ptr, dsc, i) \
+	(*(u32 *)((char *)ptr + dsc[i].offset))
+#define MLX5E_READ_CTR32_BE(ptr, dsc, i) \
+	be32_to_cpu(*(__be32 *)((char *)ptr + dsc[i].offset))
+
+#define MLX5E_DECLARE_STAT(type, fld) #fld, offsetof(type, fld)
+#define MLX5E_DECLARE_RX_STAT(type, fld) "rx%d_"#fld, offsetof(type, fld)
+#define MLX5E_DECLARE_TX_STAT(type, fld) "tx%d_"#fld, offsetof(type, fld)
+#define MLX5E_DECLARE_XDPSQ_STAT(type, fld) "tx%d_xdp_"#fld, offsetof(type, fld)
+#define MLX5E_DECLARE_RQ_XDPSQ_STAT(type, fld) "rx%d_xdp_tx_"#fld, offsetof(type, fld)
+#define MLX5E_DECLARE_CH_STAT(type, fld) "ch%d_"#fld, offsetof(type, fld)
+
+struct counter_desc {
+	char		format[ETH_GSTRING_LEN];
+	size_t		offset; /* Byte offset */
+};
+
+struct mlx5e_sw_stats {
+	u64 rx_packets;
+	u64 rx_bytes;
+	u64 tx_packets;
+	u64 tx_bytes;
+	u64 tx_tso_packets;
+	u64 tx_tso_bytes;
+	u64 tx_tso_inner_packets;
+	u64 tx_tso_inner_bytes;
+	u64 tx_added_vlan_packets;
+	u64 tx_nop;
+	u64 rx_lro_packets;
+	u64 rx_lro_bytes;
+	u64 rx_ecn_mark;
+	u64 rx_removed_vlan_packets;
+	u64 rx_csum_unnecessary;
+	u64 rx_csum_none;
+	u64 rx_csum_complete;
+	u64 rx_csum_complete_tail;
+	u64 rx_csum_complete_tail_slow;
+	u64 rx_csum_unnecessary_inner;
+	u64 rx_xdp_drop;
+	u64 rx_xdp_redirect;
+	u64 rx_xdp_tx_xmit;
+	u64 rx_xdp_tx_full;
+	u64 rx_xdp_tx_err;
+	u64 rx_xdp_tx_cqe;
+	u64 tx_csum_none;
+	u64 tx_csum_partial;
+	u64 tx_csum_partial_inner;
+	u64 tx_queue_stopped;
+	u64 tx_queue_dropped;
+	u64 tx_xmit_more;
+	u64 tx_recover;
+	u64 tx_cqes;
+	u64 tx_queue_wake;
+	u64 tx_cqe_err;
+	u64 tx_xdp_xmit;
+	u64 tx_xdp_full;
+	u64 tx_xdp_err;
+	u64 tx_xdp_cqes;
+	u64 rx_wqe_err;
+	u64 rx_mpwqe_filler_cqes;
+	u64 rx_mpwqe_filler_strides;
+	u64 rx_oversize_pkts_sw_drop;
+	u64 rx_buff_alloc_err;
+	u64 rx_cqe_compress_blks;
+	u64 rx_cqe_compress_pkts;
+	u64 rx_page_reuse;
+	u64 rx_cache_reuse;
+	u64 rx_cache_full;
+	u64 rx_cache_empty;
+	u64 rx_cache_busy;
+	u64 rx_cache_waive;
+	u64 rx_congst_umr;
+	u64 ch_events;
+	u64 ch_poll;
+	u64 ch_arm;
+	u64 ch_aff_change;
+	u64 ch_eq_rearm;
+
+#ifdef CONFIG_MLX5_EN_TLS
+	u64 tx_tls_ooo;
+	u64 tx_tls_resync_bytes;
+#endif
+};
+
+struct mlx5e_qcounter_stats {
+	u32 rx_out_of_buffer;
+	u32 rx_if_down_packets;
+};
+
+struct mlx5e_vnic_env_stats {
+	__be64 query_vnic_env_out[MLX5_ST_SZ_QW(query_vnic_env_out)];
+};
+
+#define VPORT_COUNTER_GET(vstats, c) MLX5_GET64(query_vport_counter_out, \
+						vstats->query_vport_out, c)
+
+struct mlx5e_vport_stats {
+	__be64 query_vport_out[MLX5_ST_SZ_QW(query_vport_counter_out)];
+};
+
+#define PPORT_802_3_GET(pstats, c) \
+	MLX5_GET64(ppcnt_reg, pstats->IEEE_802_3_counters, \
+		   counter_set.eth_802_3_cntrs_grp_data_layout.c##_high)
+#define PPORT_2863_GET(pstats, c) \
+	MLX5_GET64(ppcnt_reg, pstats->RFC_2863_counters, \
+		   counter_set.eth_2863_cntrs_grp_data_layout.c##_high)
+#define PPORT_2819_GET(pstats, c) \
+	MLX5_GET64(ppcnt_reg, pstats->RFC_2819_counters, \
+		   counter_set.eth_2819_cntrs_grp_data_layout.c##_high)
+#define PPORT_PHY_STATISTICAL_GET(pstats, c) \
+	MLX5_GET64(ppcnt_reg, (pstats)->phy_statistical_counters, \
+		   counter_set.phys_layer_statistical_cntrs.c##_high)
+#define PPORT_PER_PRIO_GET(pstats, prio, c) \
+	MLX5_GET64(ppcnt_reg, pstats->per_prio_counters[prio], \
+		   counter_set.eth_per_prio_grp_data_layout.c##_high)
+#define NUM_PPORT_PRIO				8
+#define PPORT_ETH_EXT_GET(pstats, c) \
+	MLX5_GET64(ppcnt_reg, (pstats)->eth_ext_counters, \
+		   counter_set.eth_extended_cntrs_grp_data_layout.c##_high)
+
+struct mlx5e_pport_stats {
+	__be64 IEEE_802_3_counters[MLX5_ST_SZ_QW(ppcnt_reg)];
+	__be64 RFC_2863_counters[MLX5_ST_SZ_QW(ppcnt_reg)];
+	__be64 RFC_2819_counters[MLX5_ST_SZ_QW(ppcnt_reg)];
+	__be64 per_prio_counters[NUM_PPORT_PRIO][MLX5_ST_SZ_QW(ppcnt_reg)];
+	__be64 phy_counters[MLX5_ST_SZ_QW(ppcnt_reg)];
+	__be64 phy_statistical_counters[MLX5_ST_SZ_QW(ppcnt_reg)];
+	__be64 eth_ext_counters[MLX5_ST_SZ_QW(ppcnt_reg)];
+};
+
+#define PCIE_PERF_GET(pcie_stats, c) \
+	MLX5_GET(mpcnt_reg, (pcie_stats)->pcie_perf_counters, \
+		 counter_set.pcie_perf_cntrs_grp_data_layout.c)
+
+#define PCIE_PERF_GET64(pcie_stats, c) \
+	MLX5_GET64(mpcnt_reg, (pcie_stats)->pcie_perf_counters, \
+		   counter_set.pcie_perf_cntrs_grp_data_layout.c##_high)
+
+struct mlx5e_pcie_stats {
+	__be64 pcie_perf_counters[MLX5_ST_SZ_QW(mpcnt_reg)];
+};
+
+struct mlx5e_rq_stats {
+	u64 packets;
+	u64 bytes;
+	u64 csum_complete;
+	u64 csum_complete_tail;
+	u64 csum_complete_tail_slow;
+	u64 csum_unnecessary;
+	u64 csum_unnecessary_inner;
+	u64 csum_none;
+	u64 lro_packets;
+	u64 lro_bytes;
+	u64 ecn_mark;
+	u64 removed_vlan_packets;
+	u64 xdp_drop;
+	u64 xdp_redirect;
+	u64 wqe_err;
+	u64 mpwqe_filler_cqes;
+	u64 mpwqe_filler_strides;
+	u64 oversize_pkts_sw_drop;
+	u64 buff_alloc_err;
+	u64 cqe_compress_blks;
+	u64 cqe_compress_pkts;
+	u64 page_reuse;
+	u64 cache_reuse;
+	u64 cache_full;
+	u64 cache_empty;
+	u64 cache_busy;
+	u64 cache_waive;
+	u64 congst_umr;
+};
+
+struct mlx5e_sq_stats {
+	/* commonly accessed in data path */
+	u64 packets;
+	u64 bytes;
+	u64 xmit_more;
+	u64 tso_packets;
+	u64 tso_bytes;
+	u64 tso_inner_packets;
+	u64 tso_inner_bytes;
+	u64 csum_partial;
+	u64 csum_partial_inner;
+	u64 added_vlan_packets;
+	u64 nop;
+#ifdef CONFIG_MLX5_EN_TLS
+	u64 tls_ooo;
+	u64 tls_resync_bytes;
+#endif
+	/* less likely accessed in data path */
+	u64 csum_none;
+	u64 stopped;
+	u64 dropped;
+	u64 recover;
+	/* dirtied @completion */
+	u64 cqes ____cacheline_aligned_in_smp;
+	u64 wake;
+	u64 cqe_err;
+};
+
+struct mlx5e_xdpsq_stats {
+	u64 xmit;
+	u64 full;
+	u64 err;
+	/* dirtied @completion */
+	u64 cqes ____cacheline_aligned_in_smp;
+};
+
+struct mlx5e_ch_stats {
+	u64 events;
+	u64 poll;
+	u64 arm;
+	u64 aff_change;
+	u64 eq_rearm;
+};
+
+struct mlx5e_stats {
+	struct mlx5e_sw_stats sw;
+	struct mlx5e_qcounter_stats qcnt;
+	struct mlx5e_vnic_env_stats vnic;
+	struct mlx5e_vport_stats vport;
+	struct mlx5e_pport_stats pport;
+	struct rtnl_link_stats64 vf_vport;
+	struct mlx5e_pcie_stats pcie;
+};
+
+enum {
+	MLX5E_NDO_UPDATE_STATS = BIT(0x1),
+};
+
+struct mlx5e_priv;
+struct mlx5e_stats_grp {
+	u16 update_stats_mask;
+	int (*get_num_stats)(struct mlx5e_priv *priv);
+	int (*fill_strings)(struct mlx5e_priv *priv, u8 *data, int idx);
+	int (*fill_stats)(struct mlx5e_priv *priv, u64 *data, int idx);
+	void (*update_stats)(struct mlx5e_priv *priv);
+};
+
+extern const struct mlx5e_stats_grp mlx5e_stats_grps[];
+extern const int mlx5e_num_stats_grps;
+
+void mlx5e_grp_sw_update_stats(struct mlx5e_priv *priv);
+
+#endif /* __MLX5_EN_STATS_H__ */
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c b/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c
new file mode 100644
index 000000000..305085377
--- /dev/null
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c
@@ -0,0 +1,3063 @@
+/*
+ * Copyright (c) 2016, 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/flow_dissector.h>
+#include <net/sch_generic.h>
+#include <net/pkt_cls.h>
+#include <net/tc_act/tc_gact.h>
+#include <net/tc_act/tc_skbedit.h>
+#include <linux/mlx5/fs.h>
+#include <linux/mlx5/device.h>
+#include <linux/rhashtable.h>
+#include <net/switchdev.h>
+#include <net/tc_act/tc_mirred.h>
+#include <net/tc_act/tc_vlan.h>
+#include <net/tc_act/tc_tunnel_key.h>
+#include <net/tc_act/tc_pedit.h>
+#include <net/tc_act/tc_csum.h>
+#include <net/vxlan.h>
+#include <net/arp.h>
+#include "en.h"
+#include "en_rep.h"
+#include "en_tc.h"
+#include "eswitch.h"
+#include "lib/vxlan.h"
+#include "fs_core.h"
+#include "en/port.h"
+
+struct mlx5_nic_flow_attr {
+	u32 action;
+	u32 flow_tag;
+	u32 mod_hdr_id;
+	u32 hairpin_tirn;
+	u8 match_level;
+	struct mlx5_flow_table	*hairpin_ft;
+};
+
+#define MLX5E_TC_FLOW_BASE (MLX5E_TC_LAST_EXPORTED_BIT + 1)
+
+enum {
+	MLX5E_TC_FLOW_INGRESS	= MLX5E_TC_INGRESS,
+	MLX5E_TC_FLOW_EGRESS	= MLX5E_TC_EGRESS,
+	MLX5E_TC_FLOW_ESWITCH	= BIT(MLX5E_TC_FLOW_BASE),
+	MLX5E_TC_FLOW_NIC	= BIT(MLX5E_TC_FLOW_BASE + 1),
+	MLX5E_TC_FLOW_OFFLOADED	= BIT(MLX5E_TC_FLOW_BASE + 2),
+	MLX5E_TC_FLOW_HAIRPIN	= BIT(MLX5E_TC_FLOW_BASE + 3),
+	MLX5E_TC_FLOW_HAIRPIN_RSS = BIT(MLX5E_TC_FLOW_BASE + 4),
+};
+
+#define MLX5E_TC_MAX_SPLITS 1
+
+struct mlx5e_tc_flow {
+	struct rhash_head	node;
+	struct mlx5e_priv	*priv;
+	u64			cookie;
+	u8			flags;
+	struct mlx5_flow_handle *rule[MLX5E_TC_MAX_SPLITS + 1];
+	struct list_head	encap;   /* flows sharing the same encap ID */
+	struct list_head	mod_hdr; /* flows sharing the same mod hdr ID */
+	struct list_head	hairpin; /* flows sharing the same hairpin */
+	union {
+		struct mlx5_esw_flow_attr esw_attr[0];
+		struct mlx5_nic_flow_attr nic_attr[0];
+	};
+};
+
+struct mlx5e_tc_flow_parse_attr {
+	struct ip_tunnel_info tun_info;
+	struct mlx5_flow_spec spec;
+	int num_mod_hdr_actions;
+	int max_mod_hdr_actions;
+	void *mod_hdr_actions;
+	int mirred_ifindex;
+};
+
+enum {
+	MLX5_HEADER_TYPE_VXLAN = 0x0,
+	MLX5_HEADER_TYPE_NVGRE = 0x1,
+};
+
+#define MLX5E_TC_TABLE_NUM_GROUPS 4
+#define MLX5E_TC_TABLE_MAX_GROUP_SIZE BIT(16)
+
+struct mlx5e_hairpin {
+	struct mlx5_hairpin *pair;
+
+	struct mlx5_core_dev *func_mdev;
+	struct mlx5e_priv *func_priv;
+	u32 tdn;
+	u32 tirn;
+
+	int num_channels;
+	struct mlx5e_rqt indir_rqt;
+	u32 indir_tirn[MLX5E_NUM_INDIR_TIRS];
+	struct mlx5e_ttc_table ttc;
+};
+
+struct mlx5e_hairpin_entry {
+	/* a node of a hash table which keeps all the  hairpin entries */
+	struct hlist_node hairpin_hlist;
+
+	/* flows sharing the same hairpin */
+	struct list_head flows;
+
+	u16 peer_vhca_id;
+	u8 prio;
+	struct mlx5e_hairpin *hp;
+};
+
+struct mod_hdr_key {
+	int num_actions;
+	void *actions;
+};
+
+struct mlx5e_mod_hdr_entry {
+	/* a node of a hash table which keeps all the mod_hdr entries */
+	struct hlist_node mod_hdr_hlist;
+
+	/* flows sharing the same mod_hdr entry */
+	struct list_head flows;
+
+	struct mod_hdr_key key;
+
+	u32 mod_hdr_id;
+};
+
+#define MLX5_MH_ACT_SZ MLX5_UN_SZ_BYTES(set_action_in_add_action_in_auto)
+
+static inline u32 hash_mod_hdr_info(struct mod_hdr_key *key)
+{
+	return jhash(key->actions,
+		     key->num_actions * MLX5_MH_ACT_SZ, 0);
+}
+
+static inline int cmp_mod_hdr_info(struct mod_hdr_key *a,
+				   struct mod_hdr_key *b)
+{
+	if (a->num_actions != b->num_actions)
+		return 1;
+
+	return memcmp(a->actions, b->actions, a->num_actions * MLX5_MH_ACT_SZ);
+}
+
+static int mlx5e_attach_mod_hdr(struct mlx5e_priv *priv,
+				struct mlx5e_tc_flow *flow,
+				struct mlx5e_tc_flow_parse_attr *parse_attr)
+{
+	struct mlx5_eswitch *esw = priv->mdev->priv.eswitch;
+	int num_actions, actions_size, namespace, err;
+	struct mlx5e_mod_hdr_entry *mh;
+	struct mod_hdr_key key;
+	bool found = false;
+	u32 hash_key;
+
+	num_actions  = parse_attr->num_mod_hdr_actions;
+	actions_size = MLX5_MH_ACT_SZ * num_actions;
+
+	key.actions = parse_attr->mod_hdr_actions;
+	key.num_actions = num_actions;
+
+	hash_key = hash_mod_hdr_info(&key);
+
+	if (flow->flags & MLX5E_TC_FLOW_ESWITCH) {
+		namespace = MLX5_FLOW_NAMESPACE_FDB;
+		hash_for_each_possible(esw->offloads.mod_hdr_tbl, mh,
+				       mod_hdr_hlist, hash_key) {
+			if (!cmp_mod_hdr_info(&mh->key, &key)) {
+				found = true;
+				break;
+			}
+		}
+	} else {
+		namespace = MLX5_FLOW_NAMESPACE_KERNEL;
+		hash_for_each_possible(priv->fs.tc.mod_hdr_tbl, mh,
+				       mod_hdr_hlist, hash_key) {
+			if (!cmp_mod_hdr_info(&mh->key, &key)) {
+				found = true;
+				break;
+			}
+		}
+	}
+
+	if (found)
+		goto attach_flow;
+
+	mh = kzalloc(sizeof(*mh) + actions_size, GFP_KERNEL);
+	if (!mh)
+		return -ENOMEM;
+
+	mh->key.actions = (void *)mh + sizeof(*mh);
+	memcpy(mh->key.actions, key.actions, actions_size);
+	mh->key.num_actions = num_actions;
+	INIT_LIST_HEAD(&mh->flows);
+
+	err = mlx5_modify_header_alloc(priv->mdev, namespace,
+				       mh->key.num_actions,
+				       mh->key.actions,
+				       &mh->mod_hdr_id);
+	if (err)
+		goto out_err;
+
+	if (flow->flags & MLX5E_TC_FLOW_ESWITCH)
+		hash_add(esw->offloads.mod_hdr_tbl, &mh->mod_hdr_hlist, hash_key);
+	else
+		hash_add(priv->fs.tc.mod_hdr_tbl, &mh->mod_hdr_hlist, hash_key);
+
+attach_flow:
+	list_add(&flow->mod_hdr, &mh->flows);
+	if (flow->flags & MLX5E_TC_FLOW_ESWITCH)
+		flow->esw_attr->mod_hdr_id = mh->mod_hdr_id;
+	else
+		flow->nic_attr->mod_hdr_id = mh->mod_hdr_id;
+
+	return 0;
+
+out_err:
+	kfree(mh);
+	return err;
+}
+
+static void mlx5e_detach_mod_hdr(struct mlx5e_priv *priv,
+				 struct mlx5e_tc_flow *flow)
+{
+	struct list_head *next = flow->mod_hdr.next;
+
+	list_del(&flow->mod_hdr);
+
+	if (list_empty(next)) {
+		struct mlx5e_mod_hdr_entry *mh;
+
+		mh = list_entry(next, struct mlx5e_mod_hdr_entry, flows);
+
+		mlx5_modify_header_dealloc(priv->mdev, mh->mod_hdr_id);
+		hash_del(&mh->mod_hdr_hlist);
+		kfree(mh);
+	}
+}
+
+static
+struct mlx5_core_dev *mlx5e_hairpin_get_mdev(struct net *net, int ifindex)
+{
+	struct net_device *netdev;
+	struct mlx5e_priv *priv;
+
+	netdev = __dev_get_by_index(net, ifindex);
+	priv = netdev_priv(netdev);
+	return priv->mdev;
+}
+
+static int mlx5e_hairpin_create_transport(struct mlx5e_hairpin *hp)
+{
+	u32 in[MLX5_ST_SZ_DW(create_tir_in)] = {0};
+	void *tirc;
+	int err;
+
+	err = mlx5_core_alloc_transport_domain(hp->func_mdev, &hp->tdn);
+	if (err)
+		goto alloc_tdn_err;
+
+	tirc = MLX5_ADDR_OF(create_tir_in, in, ctx);
+
+	MLX5_SET(tirc, tirc, disp_type, MLX5_TIRC_DISP_TYPE_DIRECT);
+	MLX5_SET(tirc, tirc, inline_rqn, hp->pair->rqn[0]);
+	MLX5_SET(tirc, tirc, transport_domain, hp->tdn);
+
+	err = mlx5_core_create_tir(hp->func_mdev, in, MLX5_ST_SZ_BYTES(create_tir_in), &hp->tirn);
+	if (err)
+		goto create_tir_err;
+
+	return 0;
+
+create_tir_err:
+	mlx5_core_dealloc_transport_domain(hp->func_mdev, hp->tdn);
+alloc_tdn_err:
+	return err;
+}
+
+static void mlx5e_hairpin_destroy_transport(struct mlx5e_hairpin *hp)
+{
+	mlx5_core_destroy_tir(hp->func_mdev, hp->tirn);
+	mlx5_core_dealloc_transport_domain(hp->func_mdev, hp->tdn);
+}
+
+static void mlx5e_hairpin_fill_rqt_rqns(struct mlx5e_hairpin *hp, void *rqtc)
+{
+	u32 indirection_rqt[MLX5E_INDIR_RQT_SIZE], rqn;
+	struct mlx5e_priv *priv = hp->func_priv;
+	int i, ix, sz = MLX5E_INDIR_RQT_SIZE;
+
+	mlx5e_build_default_indir_rqt(indirection_rqt, sz,
+				      hp->num_channels);
+
+	for (i = 0; i < sz; i++) {
+		ix = i;
+		if (priv->channels.params.rss_hfunc == ETH_RSS_HASH_XOR)
+			ix = mlx5e_bits_invert(i, ilog2(sz));
+		ix = indirection_rqt[ix];
+		rqn = hp->pair->rqn[ix];
+		MLX5_SET(rqtc, rqtc, rq_num[i], rqn);
+	}
+}
+
+static int mlx5e_hairpin_create_indirect_rqt(struct mlx5e_hairpin *hp)
+{
+	int inlen, err, sz = MLX5E_INDIR_RQT_SIZE;
+	struct mlx5e_priv *priv = hp->func_priv;
+	struct mlx5_core_dev *mdev = priv->mdev;
+	void *rqtc;
+	u32 *in;
+
+	inlen = MLX5_ST_SZ_BYTES(create_rqt_in) + sizeof(u32) * sz;
+	in = kvzalloc(inlen, GFP_KERNEL);
+	if (!in)
+		return -ENOMEM;
+
+	rqtc = MLX5_ADDR_OF(create_rqt_in, in, rqt_context);
+
+	MLX5_SET(rqtc, rqtc, rqt_actual_size, sz);
+	MLX5_SET(rqtc, rqtc, rqt_max_size, sz);
+
+	mlx5e_hairpin_fill_rqt_rqns(hp, rqtc);
+
+	err = mlx5_core_create_rqt(mdev, in, inlen, &hp->indir_rqt.rqtn);
+	if (!err)
+		hp->indir_rqt.enabled = true;
+
+	kvfree(in);
+	return err;
+}
+
+static int mlx5e_hairpin_create_indirect_tirs(struct mlx5e_hairpin *hp)
+{
+	struct mlx5e_priv *priv = hp->func_priv;
+	u32 in[MLX5_ST_SZ_DW(create_tir_in)];
+	int tt, i, err;
+	void *tirc;
+
+	for (tt = 0; tt < MLX5E_NUM_INDIR_TIRS; tt++) {
+		memset(in, 0, MLX5_ST_SZ_BYTES(create_tir_in));
+		tirc = MLX5_ADDR_OF(create_tir_in, in, ctx);
+
+		MLX5_SET(tirc, tirc, transport_domain, hp->tdn);
+		MLX5_SET(tirc, tirc, disp_type, MLX5_TIRC_DISP_TYPE_INDIRECT);
+		MLX5_SET(tirc, tirc, indirect_table, hp->indir_rqt.rqtn);
+		mlx5e_build_indir_tir_ctx_hash(&priv->channels.params, tt, tirc, false);
+
+		err = mlx5_core_create_tir(hp->func_mdev, in,
+					   MLX5_ST_SZ_BYTES(create_tir_in), &hp->indir_tirn[tt]);
+		if (err) {
+			mlx5_core_warn(hp->func_mdev, "create indirect tirs failed, %d\n", err);
+			goto err_destroy_tirs;
+		}
+	}
+	return 0;
+
+err_destroy_tirs:
+	for (i = 0; i < tt; i++)
+		mlx5_core_destroy_tir(hp->func_mdev, hp->indir_tirn[i]);
+	return err;
+}
+
+static void mlx5e_hairpin_destroy_indirect_tirs(struct mlx5e_hairpin *hp)
+{
+	int tt;
+
+	for (tt = 0; tt < MLX5E_NUM_INDIR_TIRS; tt++)
+		mlx5_core_destroy_tir(hp->func_mdev, hp->indir_tirn[tt]);
+}
+
+static void mlx5e_hairpin_set_ttc_params(struct mlx5e_hairpin *hp,
+					 struct ttc_params *ttc_params)
+{
+	struct mlx5_flow_table_attr *ft_attr = &ttc_params->ft_attr;
+	int tt;
+
+	memset(ttc_params, 0, sizeof(*ttc_params));
+
+	ttc_params->any_tt_tirn = hp->tirn;
+
+	for (tt = 0; tt < MLX5E_NUM_INDIR_TIRS; tt++)
+		ttc_params->indir_tirn[tt] = hp->indir_tirn[tt];
+
+	ft_attr->max_fte = MLX5E_NUM_TT;
+	ft_attr->level = MLX5E_TC_TTC_FT_LEVEL;
+	ft_attr->prio = MLX5E_TC_PRIO;
+}
+
+static int mlx5e_hairpin_rss_init(struct mlx5e_hairpin *hp)
+{
+	struct mlx5e_priv *priv = hp->func_priv;
+	struct ttc_params ttc_params;
+	int err;
+
+	err = mlx5e_hairpin_create_indirect_rqt(hp);
+	if (err)
+		return err;
+
+	err = mlx5e_hairpin_create_indirect_tirs(hp);
+	if (err)
+		goto err_create_indirect_tirs;
+
+	mlx5e_hairpin_set_ttc_params(hp, &ttc_params);
+	err = mlx5e_create_ttc_table(priv, &ttc_params, &hp->ttc);
+	if (err)
+		goto err_create_ttc_table;
+
+	netdev_dbg(priv->netdev, "add hairpin: using %d channels rss ttc table id %x\n",
+		   hp->num_channels, hp->ttc.ft.t->id);
+
+	return 0;
+
+err_create_ttc_table:
+	mlx5e_hairpin_destroy_indirect_tirs(hp);
+err_create_indirect_tirs:
+	mlx5e_destroy_rqt(priv, &hp->indir_rqt);
+
+	return err;
+}
+
+static void mlx5e_hairpin_rss_cleanup(struct mlx5e_hairpin *hp)
+{
+	struct mlx5e_priv *priv = hp->func_priv;
+
+	mlx5e_destroy_ttc_table(priv, &hp->ttc);
+	mlx5e_hairpin_destroy_indirect_tirs(hp);
+	mlx5e_destroy_rqt(priv, &hp->indir_rqt);
+}
+
+static struct mlx5e_hairpin *
+mlx5e_hairpin_create(struct mlx5e_priv *priv, struct mlx5_hairpin_params *params,
+		     int peer_ifindex)
+{
+	struct mlx5_core_dev *func_mdev, *peer_mdev;
+	struct mlx5e_hairpin *hp;
+	struct mlx5_hairpin *pair;
+	int err;
+
+	hp = kzalloc(sizeof(*hp), GFP_KERNEL);
+	if (!hp)
+		return ERR_PTR(-ENOMEM);
+
+	func_mdev = priv->mdev;
+	peer_mdev = mlx5e_hairpin_get_mdev(dev_net(priv->netdev), peer_ifindex);
+
+	pair = mlx5_core_hairpin_create(func_mdev, peer_mdev, params);
+	if (IS_ERR(pair)) {
+		err = PTR_ERR(pair);
+		goto create_pair_err;
+	}
+	hp->pair = pair;
+	hp->func_mdev = func_mdev;
+	hp->func_priv = priv;
+	hp->num_channels = params->num_channels;
+
+	err = mlx5e_hairpin_create_transport(hp);
+	if (err)
+		goto create_transport_err;
+
+	if (hp->num_channels > 1) {
+		err = mlx5e_hairpin_rss_init(hp);
+		if (err)
+			goto rss_init_err;
+	}
+
+	return hp;
+
+rss_init_err:
+	mlx5e_hairpin_destroy_transport(hp);
+create_transport_err:
+	mlx5_core_hairpin_destroy(hp->pair);
+create_pair_err:
+	kfree(hp);
+	return ERR_PTR(err);
+}
+
+static void mlx5e_hairpin_destroy(struct mlx5e_hairpin *hp)
+{
+	if (hp->num_channels > 1)
+		mlx5e_hairpin_rss_cleanup(hp);
+	mlx5e_hairpin_destroy_transport(hp);
+	mlx5_core_hairpin_destroy(hp->pair);
+	kvfree(hp);
+}
+
+static inline u32 hash_hairpin_info(u16 peer_vhca_id, u8 prio)
+{
+	return (peer_vhca_id << 16 | prio);
+}
+
+static struct mlx5e_hairpin_entry *mlx5e_hairpin_get(struct mlx5e_priv *priv,
+						     u16 peer_vhca_id, u8 prio)
+{
+	struct mlx5e_hairpin_entry *hpe;
+	u32 hash_key = hash_hairpin_info(peer_vhca_id, prio);
+
+	hash_for_each_possible(priv->fs.tc.hairpin_tbl, hpe,
+			       hairpin_hlist, hash_key) {
+		if (hpe->peer_vhca_id == peer_vhca_id && hpe->prio == prio)
+			return hpe;
+	}
+
+	return NULL;
+}
+
+#define UNKNOWN_MATCH_PRIO 8
+
+static int mlx5e_hairpin_get_prio(struct mlx5e_priv *priv,
+				  struct mlx5_flow_spec *spec, u8 *match_prio)
+{
+	void *headers_c, *headers_v;
+	u8 prio_val, prio_mask = 0;
+	bool vlan_present;
+
+#ifdef CONFIG_MLX5_CORE_EN_DCB
+	if (priv->dcbx_dp.trust_state != MLX5_QPTS_TRUST_PCP) {
+		netdev_warn(priv->netdev,
+			    "only PCP trust state supported for hairpin\n");
+		return -EOPNOTSUPP;
+	}
+#endif
+	headers_c = MLX5_ADDR_OF(fte_match_param, spec->match_criteria, outer_headers);
+	headers_v = MLX5_ADDR_OF(fte_match_param, spec->match_value, outer_headers);
+
+	vlan_present = MLX5_GET(fte_match_set_lyr_2_4, headers_v, cvlan_tag);
+	if (vlan_present) {
+		prio_mask = MLX5_GET(fte_match_set_lyr_2_4, headers_c, first_prio);
+		prio_val = MLX5_GET(fte_match_set_lyr_2_4, headers_v, first_prio);
+	}
+
+	if (!vlan_present || !prio_mask) {
+		prio_val = UNKNOWN_MATCH_PRIO;
+	} else if (prio_mask != 0x7) {
+		netdev_warn(priv->netdev,
+			    "masked priority match not supported for hairpin\n");
+		return -EOPNOTSUPP;
+	}
+
+	*match_prio = prio_val;
+	return 0;
+}
+
+static int mlx5e_hairpin_flow_add(struct mlx5e_priv *priv,
+				  struct mlx5e_tc_flow *flow,
+				  struct mlx5e_tc_flow_parse_attr *parse_attr)
+{
+	int peer_ifindex = parse_attr->mirred_ifindex;
+	struct mlx5_hairpin_params params;
+	struct mlx5_core_dev *peer_mdev;
+	struct mlx5e_hairpin_entry *hpe;
+	struct mlx5e_hairpin *hp;
+	u64 link_speed64;
+	u32 link_speed;
+	u8 match_prio;
+	u16 peer_id;
+	int err;
+
+	peer_mdev = mlx5e_hairpin_get_mdev(dev_net(priv->netdev), peer_ifindex);
+	if (!MLX5_CAP_GEN(priv->mdev, hairpin) || !MLX5_CAP_GEN(peer_mdev, hairpin)) {
+		netdev_warn(priv->netdev, "hairpin is not supported\n");
+		return -EOPNOTSUPP;
+	}
+
+	peer_id = MLX5_CAP_GEN(peer_mdev, vhca_id);
+	err = mlx5e_hairpin_get_prio(priv, &parse_attr->spec, &match_prio);
+	if (err)
+		return err;
+	hpe = mlx5e_hairpin_get(priv, peer_id, match_prio);
+	if (hpe)
+		goto attach_flow;
+
+	hpe = kzalloc(sizeof(*hpe), GFP_KERNEL);
+	if (!hpe)
+		return -ENOMEM;
+
+	INIT_LIST_HEAD(&hpe->flows);
+	hpe->peer_vhca_id = peer_id;
+	hpe->prio = match_prio;
+
+	params.log_data_size = 15;
+	params.log_data_size = min_t(u8, params.log_data_size,
+				     MLX5_CAP_GEN(priv->mdev, log_max_hairpin_wq_data_sz));
+	params.log_data_size = max_t(u8, params.log_data_size,
+				     MLX5_CAP_GEN(priv->mdev, log_min_hairpin_wq_data_sz));
+
+	params.log_num_packets = params.log_data_size -
+				 MLX5_MPWRQ_MIN_LOG_STRIDE_SZ(priv->mdev);
+	params.log_num_packets = min_t(u8, params.log_num_packets,
+				       MLX5_CAP_GEN(priv->mdev, log_max_hairpin_num_packets));
+
+	params.q_counter = priv->q_counter;
+	/* set hairpin pair per each 50Gbs share of the link */
+	mlx5e_port_max_linkspeed(priv->mdev, &link_speed);
+	link_speed = max_t(u32, link_speed, 50000);
+	link_speed64 = link_speed;
+	do_div(link_speed64, 50000);
+	params.num_channels = link_speed64;
+
+	hp = mlx5e_hairpin_create(priv, &params, peer_ifindex);
+	if (IS_ERR(hp)) {
+		err = PTR_ERR(hp);
+		goto create_hairpin_err;
+	}
+
+	netdev_dbg(priv->netdev, "add hairpin: tirn %x rqn %x peer %s sqn %x prio %d (log) data %d packets %d\n",
+		   hp->tirn, hp->pair->rqn[0], hp->pair->peer_mdev->priv.name,
+		   hp->pair->sqn[0], match_prio, params.log_data_size, params.log_num_packets);
+
+	hpe->hp = hp;
+	hash_add(priv->fs.tc.hairpin_tbl, &hpe->hairpin_hlist,
+		 hash_hairpin_info(peer_id, match_prio));
+
+attach_flow:
+	if (hpe->hp->num_channels > 1) {
+		flow->flags |= MLX5E_TC_FLOW_HAIRPIN_RSS;
+		flow->nic_attr->hairpin_ft = hpe->hp->ttc.ft.t;
+	} else {
+		flow->nic_attr->hairpin_tirn = hpe->hp->tirn;
+	}
+	list_add(&flow->hairpin, &hpe->flows);
+
+	return 0;
+
+create_hairpin_err:
+	kfree(hpe);
+	return err;
+}
+
+static void mlx5e_hairpin_flow_del(struct mlx5e_priv *priv,
+				   struct mlx5e_tc_flow *flow)
+{
+	struct list_head *next = flow->hairpin.next;
+
+	list_del(&flow->hairpin);
+
+	/* no more hairpin flows for us, release the hairpin pair */
+	if (list_empty(next)) {
+		struct mlx5e_hairpin_entry *hpe;
+
+		hpe = list_entry(next, struct mlx5e_hairpin_entry, flows);
+
+		netdev_dbg(priv->netdev, "del hairpin: peer %s\n",
+			   hpe->hp->pair->peer_mdev->priv.name);
+
+		mlx5e_hairpin_destroy(hpe->hp);
+		hash_del(&hpe->hairpin_hlist);
+		kfree(hpe);
+	}
+}
+
+static struct mlx5_flow_handle *
+mlx5e_tc_add_nic_flow(struct mlx5e_priv *priv,
+		      struct mlx5e_tc_flow_parse_attr *parse_attr,
+		      struct mlx5e_tc_flow *flow)
+{
+	struct mlx5_nic_flow_attr *attr = flow->nic_attr;
+	struct mlx5_core_dev *dev = priv->mdev;
+	struct mlx5_flow_destination dest[2] = {};
+	struct mlx5_flow_act flow_act = {
+		.action = attr->action,
+		.has_flow_tag = true,
+		.flow_tag = attr->flow_tag,
+		.encap_id = 0,
+	};
+	struct mlx5_fc *counter = NULL;
+	struct mlx5_flow_handle *rule;
+	bool table_created = false;
+	int err, dest_ix = 0;
+
+	if (flow->flags & MLX5E_TC_FLOW_HAIRPIN) {
+		err = mlx5e_hairpin_flow_add(priv, flow, parse_attr);
+		if (err) {
+			rule = ERR_PTR(err);
+			goto err_add_hairpin_flow;
+		}
+		if (flow->flags & MLX5E_TC_FLOW_HAIRPIN_RSS) {
+			dest[dest_ix].type = MLX5_FLOW_DESTINATION_TYPE_FLOW_TABLE;
+			dest[dest_ix].ft = attr->hairpin_ft;
+		} else {
+			dest[dest_ix].type = MLX5_FLOW_DESTINATION_TYPE_TIR;
+			dest[dest_ix].tir_num = attr->hairpin_tirn;
+		}
+		dest_ix++;
+	} else if (attr->action & MLX5_FLOW_CONTEXT_ACTION_FWD_DEST) {
+		dest[dest_ix].type = MLX5_FLOW_DESTINATION_TYPE_FLOW_TABLE;
+		dest[dest_ix].ft = priv->fs.vlan.ft.t;
+		dest_ix++;
+	}
+
+	if (attr->action & MLX5_FLOW_CONTEXT_ACTION_COUNT) {
+		counter = mlx5_fc_create(dev, true);
+		if (IS_ERR(counter)) {
+			rule = ERR_CAST(counter);
+			goto err_fc_create;
+		}
+		dest[dest_ix].type = MLX5_FLOW_DESTINATION_TYPE_COUNTER;
+		dest[dest_ix].counter = counter;
+		dest_ix++;
+	}
+
+	if (attr->action & MLX5_FLOW_CONTEXT_ACTION_MOD_HDR) {
+		err = mlx5e_attach_mod_hdr(priv, flow, parse_attr);
+		flow_act.modify_id = attr->mod_hdr_id;
+		kfree(parse_attr->mod_hdr_actions);
+		if (err) {
+			rule = ERR_PTR(err);
+			goto err_create_mod_hdr_id;
+		}
+	}
+
+	if (IS_ERR_OR_NULL(priv->fs.tc.t)) {
+		int tc_grp_size, tc_tbl_size;
+		u32 max_flow_counter;
+
+		max_flow_counter = (MLX5_CAP_GEN(dev, max_flow_counter_31_16) << 16) |
+				    MLX5_CAP_GEN(dev, max_flow_counter_15_0);
+
+		tc_grp_size = min_t(int, max_flow_counter, MLX5E_TC_TABLE_MAX_GROUP_SIZE);
+
+		tc_tbl_size = min_t(int, tc_grp_size * MLX5E_TC_TABLE_NUM_GROUPS,
+				    BIT(MLX5_CAP_FLOWTABLE_NIC_RX(dev, log_max_ft_size)));
+
+		priv->fs.tc.t =
+			mlx5_create_auto_grouped_flow_table(priv->fs.ns,
+							    MLX5E_TC_PRIO,
+							    tc_tbl_size,
+							    MLX5E_TC_TABLE_NUM_GROUPS,
+							    MLX5E_TC_FT_LEVEL, 0);
+		if (IS_ERR(priv->fs.tc.t)) {
+			netdev_err(priv->netdev,
+				   "Failed to create tc offload table\n");
+			rule = ERR_CAST(priv->fs.tc.t);
+			goto err_create_ft;
+		}
+
+		table_created = true;
+	}
+
+	if (attr->match_level != MLX5_MATCH_NONE)
+		parse_attr->spec.match_criteria_enable = MLX5_MATCH_OUTER_HEADERS;
+
+	rule = mlx5_add_flow_rules(priv->fs.tc.t, &parse_attr->spec,
+				   &flow_act, dest, dest_ix);
+
+	if (IS_ERR(rule))
+		goto err_add_rule;
+
+	return rule;
+
+err_add_rule:
+	if (table_created) {
+		mlx5_destroy_flow_table(priv->fs.tc.t);
+		priv->fs.tc.t = NULL;
+	}
+err_create_ft:
+	if (attr->action & MLX5_FLOW_CONTEXT_ACTION_MOD_HDR)
+		mlx5e_detach_mod_hdr(priv, flow);
+err_create_mod_hdr_id:
+	mlx5_fc_destroy(dev, counter);
+err_fc_create:
+	if (flow->flags & MLX5E_TC_FLOW_HAIRPIN)
+		mlx5e_hairpin_flow_del(priv, flow);
+err_add_hairpin_flow:
+	return rule;
+}
+
+static void mlx5e_tc_del_nic_flow(struct mlx5e_priv *priv,
+				  struct mlx5e_tc_flow *flow)
+{
+	struct mlx5_nic_flow_attr *attr = flow->nic_attr;
+	struct mlx5_fc *counter = NULL;
+
+	counter = mlx5_flow_rule_counter(flow->rule[0]);
+	mlx5_del_flow_rules(flow->rule[0]);
+	mlx5_fc_destroy(priv->mdev, counter);
+
+	if (!mlx5e_tc_num_filters(priv) && priv->fs.tc.t) {
+		mlx5_destroy_flow_table(priv->fs.tc.t);
+		priv->fs.tc.t = NULL;
+	}
+
+	if (attr->action & MLX5_FLOW_CONTEXT_ACTION_MOD_HDR)
+		mlx5e_detach_mod_hdr(priv, flow);
+
+	if (flow->flags & MLX5E_TC_FLOW_HAIRPIN)
+		mlx5e_hairpin_flow_del(priv, flow);
+}
+
+static void mlx5e_detach_encap(struct mlx5e_priv *priv,
+			       struct mlx5e_tc_flow *flow);
+
+static int mlx5e_attach_encap(struct mlx5e_priv *priv,
+			      struct ip_tunnel_info *tun_info,
+			      struct net_device *mirred_dev,
+			      struct net_device **encap_dev,
+			      struct mlx5e_tc_flow *flow);
+
+static struct mlx5_flow_handle *
+mlx5e_tc_add_fdb_flow(struct mlx5e_priv *priv,
+		      struct mlx5e_tc_flow_parse_attr *parse_attr,
+		      struct mlx5e_tc_flow *flow)
+{
+	struct mlx5_eswitch *esw = priv->mdev->priv.eswitch;
+	struct mlx5_esw_flow_attr *attr = flow->esw_attr;
+	struct net_device *out_dev, *encap_dev = NULL;
+	struct mlx5_flow_handle *rule = NULL;
+	struct mlx5e_rep_priv *rpriv;
+	struct mlx5e_priv *out_priv;
+	int err;
+
+	if (attr->action & MLX5_FLOW_CONTEXT_ACTION_ENCAP) {
+		out_dev = __dev_get_by_index(dev_net(priv->netdev),
+					     attr->parse_attr->mirred_ifindex);
+		err = mlx5e_attach_encap(priv, &parse_attr->tun_info,
+					 out_dev, &encap_dev, flow);
+		if (err) {
+			rule = ERR_PTR(err);
+			if (err != -EAGAIN)
+				goto err_attach_encap;
+		}
+		out_priv = netdev_priv(encap_dev);
+		rpriv = out_priv->ppriv;
+		attr->out_rep[attr->out_count] = rpriv->rep;
+		attr->out_mdev[attr->out_count++] = out_priv->mdev;
+	}
+
+	err = mlx5_eswitch_add_vlan_action(esw, attr);
+	if (err) {
+		rule = ERR_PTR(err);
+		goto err_add_vlan;
+	}
+
+	if (attr->action & MLX5_FLOW_CONTEXT_ACTION_MOD_HDR) {
+		err = mlx5e_attach_mod_hdr(priv, flow, parse_attr);
+		kfree(parse_attr->mod_hdr_actions);
+		if (err) {
+			rule = ERR_PTR(err);
+			goto err_mod_hdr;
+		}
+	}
+
+	/* we get here if (1) there's no error (rule being null) or when
+	 * (2) there's an encap action and we're on -EAGAIN (no valid neigh)
+	 */
+	if (rule != ERR_PTR(-EAGAIN)) {
+		rule = mlx5_eswitch_add_offloaded_rule(esw, &parse_attr->spec, attr);
+		if (IS_ERR(rule))
+			goto err_add_rule;
+
+		if (attr->mirror_count) {
+			flow->rule[1] = mlx5_eswitch_add_fwd_rule(esw, &parse_attr->spec, attr);
+			if (IS_ERR(flow->rule[1]))
+				goto err_fwd_rule;
+		}
+	}
+	return rule;
+
+err_fwd_rule:
+	mlx5_eswitch_del_offloaded_rule(esw, rule, attr);
+	rule = flow->rule[1];
+err_add_rule:
+	if (attr->action & MLX5_FLOW_CONTEXT_ACTION_MOD_HDR)
+		mlx5e_detach_mod_hdr(priv, flow);
+err_mod_hdr:
+	mlx5_eswitch_del_vlan_action(esw, attr);
+err_add_vlan:
+	if (attr->action & MLX5_FLOW_CONTEXT_ACTION_ENCAP)
+		mlx5e_detach_encap(priv, flow);
+err_attach_encap:
+	return rule;
+}
+
+static void mlx5e_tc_del_fdb_flow(struct mlx5e_priv *priv,
+				  struct mlx5e_tc_flow *flow)
+{
+	struct mlx5_eswitch *esw = priv->mdev->priv.eswitch;
+	struct mlx5_esw_flow_attr *attr = flow->esw_attr;
+
+	if (flow->flags & MLX5E_TC_FLOW_OFFLOADED) {
+		flow->flags &= ~MLX5E_TC_FLOW_OFFLOADED;
+		if (attr->mirror_count)
+			mlx5_eswitch_del_offloaded_rule(esw, flow->rule[1], attr);
+		mlx5_eswitch_del_offloaded_rule(esw, flow->rule[0], attr);
+	}
+
+	mlx5_eswitch_del_vlan_action(esw, attr);
+
+	if (attr->action & MLX5_FLOW_CONTEXT_ACTION_ENCAP) {
+		mlx5e_detach_encap(priv, flow);
+		kvfree(attr->parse_attr);
+	}
+
+	if (attr->action & MLX5_FLOW_CONTEXT_ACTION_MOD_HDR)
+		mlx5e_detach_mod_hdr(priv, flow);
+}
+
+void mlx5e_tc_encap_flows_add(struct mlx5e_priv *priv,
+			      struct mlx5e_encap_entry *e)
+{
+	struct mlx5_eswitch *esw = priv->mdev->priv.eswitch;
+	struct mlx5_esw_flow_attr *esw_attr;
+	struct mlx5e_tc_flow *flow;
+	int err;
+
+	err = mlx5_encap_alloc(priv->mdev, e->tunnel_type,
+			       e->encap_size, e->encap_header,
+			       &e->encap_id);
+	if (err) {
+		mlx5_core_warn(priv->mdev, "Failed to offload cached encapsulation header, %d\n",
+			       err);
+		return;
+	}
+	e->flags |= MLX5_ENCAP_ENTRY_VALID;
+	mlx5e_rep_queue_neigh_stats_work(priv);
+
+	list_for_each_entry(flow, &e->flows, encap) {
+		esw_attr = flow->esw_attr;
+		esw_attr->encap_id = e->encap_id;
+		flow->rule[0] = mlx5_eswitch_add_offloaded_rule(esw, &esw_attr->parse_attr->spec, esw_attr);
+		if (IS_ERR(flow->rule[0])) {
+			err = PTR_ERR(flow->rule[0]);
+			mlx5_core_warn(priv->mdev, "Failed to update cached encapsulation flow, %d\n",
+				       err);
+			continue;
+		}
+
+		if (esw_attr->mirror_count) {
+			flow->rule[1] = mlx5_eswitch_add_fwd_rule(esw, &esw_attr->parse_attr->spec, esw_attr);
+			if (IS_ERR(flow->rule[1])) {
+				mlx5_eswitch_del_offloaded_rule(esw, flow->rule[0], esw_attr);
+				err = PTR_ERR(flow->rule[1]);
+				mlx5_core_warn(priv->mdev, "Failed to update cached mirror flow, %d\n",
+					       err);
+				continue;
+			}
+		}
+
+		flow->flags |= MLX5E_TC_FLOW_OFFLOADED;
+	}
+}
+
+void mlx5e_tc_encap_flows_del(struct mlx5e_priv *priv,
+			      struct mlx5e_encap_entry *e)
+{
+	struct mlx5_eswitch *esw = priv->mdev->priv.eswitch;
+	struct mlx5e_tc_flow *flow;
+
+	list_for_each_entry(flow, &e->flows, encap) {
+		if (flow->flags & MLX5E_TC_FLOW_OFFLOADED) {
+			struct mlx5_esw_flow_attr *attr = flow->esw_attr;
+
+			flow->flags &= ~MLX5E_TC_FLOW_OFFLOADED;
+			if (attr->mirror_count)
+				mlx5_eswitch_del_offloaded_rule(esw, flow->rule[1], attr);
+			mlx5_eswitch_del_offloaded_rule(esw, flow->rule[0], attr);
+		}
+	}
+
+	if (e->flags & MLX5_ENCAP_ENTRY_VALID) {
+		e->flags &= ~MLX5_ENCAP_ENTRY_VALID;
+		mlx5_encap_dealloc(priv->mdev, e->encap_id);
+	}
+}
+
+void mlx5e_tc_update_neigh_used_value(struct mlx5e_neigh_hash_entry *nhe)
+{
+	struct mlx5e_neigh *m_neigh = &nhe->m_neigh;
+	struct mlx5e_tc_flow *flow;
+	struct mlx5e_encap_entry *e;
+	struct mlx5_fc *counter;
+	struct neigh_table *tbl;
+	bool neigh_used = false;
+	struct neighbour *n;
+	u64 lastuse;
+
+	if (m_neigh->family == AF_INET)
+		tbl = &arp_tbl;
+#if IS_ENABLED(CONFIG_IPV6)
+	else if (m_neigh->family == AF_INET6)
+		tbl = &nd_tbl;
+#endif
+	else
+		return;
+
+	list_for_each_entry(e, &nhe->encap_list, encap_list) {
+		if (!(e->flags & MLX5_ENCAP_ENTRY_VALID))
+			continue;
+		list_for_each_entry(flow, &e->flows, encap) {
+			if (flow->flags & MLX5E_TC_FLOW_OFFLOADED) {
+				counter = mlx5_flow_rule_counter(flow->rule[0]);
+				lastuse = mlx5_fc_query_lastuse(counter);
+				if (time_after((unsigned long)lastuse, nhe->reported_lastuse)) {
+					neigh_used = true;
+					break;
+				}
+			}
+		}
+		if (neigh_used)
+			break;
+	}
+
+	if (neigh_used) {
+		nhe->reported_lastuse = jiffies;
+
+		/* find the relevant neigh according to the cached device and
+		 * dst ip pair
+		 */
+		n = neigh_lookup(tbl, &m_neigh->dst_ip, m_neigh->dev);
+		if (!n)
+			return;
+
+		neigh_event_send(n, NULL);
+		neigh_release(n);
+	}
+}
+
+static void mlx5e_detach_encap(struct mlx5e_priv *priv,
+			       struct mlx5e_tc_flow *flow)
+{
+	struct list_head *next = flow->encap.next;
+
+	list_del(&flow->encap);
+	if (list_empty(next)) {
+		struct mlx5e_encap_entry *e;
+
+		e = list_entry(next, struct mlx5e_encap_entry, flows);
+		mlx5e_rep_encap_entry_detach(netdev_priv(e->out_dev), e);
+
+		if (e->flags & MLX5_ENCAP_ENTRY_VALID)
+			mlx5_encap_dealloc(priv->mdev, e->encap_id);
+
+		hash_del_rcu(&e->encap_hlist);
+		kfree(e->encap_header);
+		kfree(e);
+	}
+}
+
+static void mlx5e_tc_del_flow(struct mlx5e_priv *priv,
+			      struct mlx5e_tc_flow *flow)
+{
+	if (flow->flags & MLX5E_TC_FLOW_ESWITCH)
+		mlx5e_tc_del_fdb_flow(priv, flow);
+	else
+		mlx5e_tc_del_nic_flow(priv, flow);
+}
+
+static void parse_vxlan_attr(struct mlx5_flow_spec *spec,
+			     struct tc_cls_flower_offload *f)
+{
+	void *headers_c = MLX5_ADDR_OF(fte_match_param, spec->match_criteria,
+				       outer_headers);
+	void *headers_v = MLX5_ADDR_OF(fte_match_param, spec->match_value,
+				       outer_headers);
+	void *misc_c = MLX5_ADDR_OF(fte_match_param, spec->match_criteria,
+				    misc_parameters);
+	void *misc_v = MLX5_ADDR_OF(fte_match_param, spec->match_value,
+				    misc_parameters);
+
+	MLX5_SET_TO_ONES(fte_match_set_lyr_2_4, headers_c, ip_protocol);
+	MLX5_SET(fte_match_set_lyr_2_4, headers_v, ip_protocol, IPPROTO_UDP);
+
+	if (dissector_uses_key(f->dissector, FLOW_DISSECTOR_KEY_ENC_KEYID)) {
+		struct flow_dissector_key_keyid *key =
+			skb_flow_dissector_target(f->dissector,
+						  FLOW_DISSECTOR_KEY_ENC_KEYID,
+						  f->key);
+		struct flow_dissector_key_keyid *mask =
+			skb_flow_dissector_target(f->dissector,
+						  FLOW_DISSECTOR_KEY_ENC_KEYID,
+						  f->mask);
+		MLX5_SET(fte_match_set_misc, misc_c, vxlan_vni,
+			 be32_to_cpu(mask->keyid));
+		MLX5_SET(fte_match_set_misc, misc_v, vxlan_vni,
+			 be32_to_cpu(key->keyid));
+	}
+}
+
+static int parse_tunnel_attr(struct mlx5e_priv *priv,
+			     struct mlx5_flow_spec *spec,
+			     struct tc_cls_flower_offload *f)
+{
+	void *headers_c = MLX5_ADDR_OF(fte_match_param, spec->match_criteria,
+				       outer_headers);
+	void *headers_v = MLX5_ADDR_OF(fte_match_param, spec->match_value,
+				       outer_headers);
+
+	struct flow_dissector_key_control *enc_control =
+		skb_flow_dissector_target(f->dissector,
+					  FLOW_DISSECTOR_KEY_ENC_CONTROL,
+					  f->key);
+
+	if (dissector_uses_key(f->dissector, FLOW_DISSECTOR_KEY_ENC_PORTS)) {
+		struct flow_dissector_key_ports *key =
+			skb_flow_dissector_target(f->dissector,
+						  FLOW_DISSECTOR_KEY_ENC_PORTS,
+						  f->key);
+		struct flow_dissector_key_ports *mask =
+			skb_flow_dissector_target(f->dissector,
+						  FLOW_DISSECTOR_KEY_ENC_PORTS,
+						  f->mask);
+
+		/* Full udp dst port must be given */
+		if (memchr_inv(&mask->dst, 0xff, sizeof(mask->dst)))
+			goto vxlan_match_offload_err;
+
+		if (mlx5_vxlan_lookup_port(priv->mdev->vxlan, be16_to_cpu(key->dst)) &&
+		    MLX5_CAP_ESW(priv->mdev, vxlan_encap_decap))
+			parse_vxlan_attr(spec, f);
+		else {
+			netdev_warn(priv->netdev,
+				    "%d isn't an offloaded vxlan udp dport\n", be16_to_cpu(key->dst));
+			return -EOPNOTSUPP;
+		}
+
+		MLX5_SET(fte_match_set_lyr_2_4, headers_c,
+			 udp_dport, ntohs(mask->dst));
+		MLX5_SET(fte_match_set_lyr_2_4, headers_v,
+			 udp_dport, ntohs(key->dst));
+
+		MLX5_SET(fte_match_set_lyr_2_4, headers_c,
+			 udp_sport, ntohs(mask->src));
+		MLX5_SET(fte_match_set_lyr_2_4, headers_v,
+			 udp_sport, ntohs(key->src));
+	} else { /* udp dst port must be given */
+vxlan_match_offload_err:
+		netdev_warn(priv->netdev,
+			    "IP tunnel decap offload supported only for vxlan, must set UDP dport\n");
+		return -EOPNOTSUPP;
+	}
+
+	if (enc_control->addr_type == FLOW_DISSECTOR_KEY_IPV4_ADDRS) {
+		struct flow_dissector_key_ipv4_addrs *key =
+			skb_flow_dissector_target(f->dissector,
+						  FLOW_DISSECTOR_KEY_ENC_IPV4_ADDRS,
+						  f->key);
+		struct flow_dissector_key_ipv4_addrs *mask =
+			skb_flow_dissector_target(f->dissector,
+						  FLOW_DISSECTOR_KEY_ENC_IPV4_ADDRS,
+						  f->mask);
+		MLX5_SET(fte_match_set_lyr_2_4, headers_c,
+			 src_ipv4_src_ipv6.ipv4_layout.ipv4,
+			 ntohl(mask->src));
+		MLX5_SET(fte_match_set_lyr_2_4, headers_v,
+			 src_ipv4_src_ipv6.ipv4_layout.ipv4,
+			 ntohl(key->src));
+
+		MLX5_SET(fte_match_set_lyr_2_4, headers_c,
+			 dst_ipv4_dst_ipv6.ipv4_layout.ipv4,
+			 ntohl(mask->dst));
+		MLX5_SET(fte_match_set_lyr_2_4, headers_v,
+			 dst_ipv4_dst_ipv6.ipv4_layout.ipv4,
+			 ntohl(key->dst));
+
+		MLX5_SET_TO_ONES(fte_match_set_lyr_2_4, headers_c, ethertype);
+		MLX5_SET(fte_match_set_lyr_2_4, headers_v, ethertype, ETH_P_IP);
+	} else if (enc_control->addr_type == FLOW_DISSECTOR_KEY_IPV6_ADDRS) {
+		struct flow_dissector_key_ipv6_addrs *key =
+			skb_flow_dissector_target(f->dissector,
+						  FLOW_DISSECTOR_KEY_ENC_IPV6_ADDRS,
+						  f->key);
+		struct flow_dissector_key_ipv6_addrs *mask =
+			skb_flow_dissector_target(f->dissector,
+						  FLOW_DISSECTOR_KEY_ENC_IPV6_ADDRS,
+						  f->mask);
+
+		memcpy(MLX5_ADDR_OF(fte_match_set_lyr_2_4, headers_c,
+				    src_ipv4_src_ipv6.ipv6_layout.ipv6),
+		       &mask->src, MLX5_FLD_SZ_BYTES(ipv6_layout, ipv6));
+		memcpy(MLX5_ADDR_OF(fte_match_set_lyr_2_4, headers_v,
+				    src_ipv4_src_ipv6.ipv6_layout.ipv6),
+		       &key->src, MLX5_FLD_SZ_BYTES(ipv6_layout, ipv6));
+
+		memcpy(MLX5_ADDR_OF(fte_match_set_lyr_2_4, headers_c,
+				    dst_ipv4_dst_ipv6.ipv6_layout.ipv6),
+		       &mask->dst, MLX5_FLD_SZ_BYTES(ipv6_layout, ipv6));
+		memcpy(MLX5_ADDR_OF(fte_match_set_lyr_2_4, headers_v,
+				    dst_ipv4_dst_ipv6.ipv6_layout.ipv6),
+		       &key->dst, MLX5_FLD_SZ_BYTES(ipv6_layout, ipv6));
+
+		MLX5_SET_TO_ONES(fte_match_set_lyr_2_4, headers_c, ethertype);
+		MLX5_SET(fte_match_set_lyr_2_4, headers_v, ethertype, ETH_P_IPV6);
+	}
+
+	if (dissector_uses_key(f->dissector, FLOW_DISSECTOR_KEY_ENC_IP)) {
+		struct flow_dissector_key_ip *key =
+			skb_flow_dissector_target(f->dissector,
+						  FLOW_DISSECTOR_KEY_ENC_IP,
+						  f->key);
+		struct flow_dissector_key_ip *mask =
+			skb_flow_dissector_target(f->dissector,
+						  FLOW_DISSECTOR_KEY_ENC_IP,
+						  f->mask);
+
+		MLX5_SET(fte_match_set_lyr_2_4, headers_c, ip_ecn, mask->tos & 0x3);
+		MLX5_SET(fte_match_set_lyr_2_4, headers_v, ip_ecn, key->tos & 0x3);
+
+		MLX5_SET(fte_match_set_lyr_2_4, headers_c, ip_dscp, mask->tos >> 2);
+		MLX5_SET(fte_match_set_lyr_2_4, headers_v, ip_dscp, key->tos  >> 2);
+
+		MLX5_SET(fte_match_set_lyr_2_4, headers_c, ttl_hoplimit, mask->ttl);
+		MLX5_SET(fte_match_set_lyr_2_4, headers_v, ttl_hoplimit, key->ttl);
+	}
+
+	/* Enforce DMAC when offloading incoming tunneled flows.
+	 * Flow counters require a match on the DMAC.
+	 */
+	MLX5_SET_TO_ONES(fte_match_set_lyr_2_4, headers_c, dmac_47_16);
+	MLX5_SET_TO_ONES(fte_match_set_lyr_2_4, headers_c, dmac_15_0);
+	ether_addr_copy(MLX5_ADDR_OF(fte_match_set_lyr_2_4, headers_v,
+				     dmac_47_16), priv->netdev->dev_addr);
+
+	/* let software handle IP fragments */
+	MLX5_SET(fte_match_set_lyr_2_4, headers_c, frag, 1);
+	MLX5_SET(fte_match_set_lyr_2_4, headers_v, frag, 0);
+
+	return 0;
+}
+
+static int __parse_cls_flower(struct mlx5e_priv *priv,
+			      struct mlx5_flow_spec *spec,
+			      struct tc_cls_flower_offload *f,
+			      u8 *match_level)
+{
+	void *headers_c = MLX5_ADDR_OF(fte_match_param, spec->match_criteria,
+				       outer_headers);
+	void *headers_v = MLX5_ADDR_OF(fte_match_param, spec->match_value,
+				       outer_headers);
+	void *misc_c = MLX5_ADDR_OF(fte_match_param, spec->match_criteria,
+				    misc_parameters);
+	void *misc_v = MLX5_ADDR_OF(fte_match_param, spec->match_value,
+				    misc_parameters);
+	u16 addr_type = 0;
+	u8 ip_proto = 0;
+
+	*match_level = MLX5_MATCH_NONE;
+
+	if (f->dissector->used_keys &
+	    ~(BIT(FLOW_DISSECTOR_KEY_CONTROL) |
+	      BIT(FLOW_DISSECTOR_KEY_BASIC) |
+	      BIT(FLOW_DISSECTOR_KEY_ETH_ADDRS) |
+	      BIT(FLOW_DISSECTOR_KEY_VLAN) |
+	      BIT(FLOW_DISSECTOR_KEY_CVLAN) |
+	      BIT(FLOW_DISSECTOR_KEY_IPV4_ADDRS) |
+	      BIT(FLOW_DISSECTOR_KEY_IPV6_ADDRS) |
+	      BIT(FLOW_DISSECTOR_KEY_PORTS) |
+	      BIT(FLOW_DISSECTOR_KEY_ENC_KEYID) |
+	      BIT(FLOW_DISSECTOR_KEY_ENC_IPV4_ADDRS) |
+	      BIT(FLOW_DISSECTOR_KEY_ENC_IPV6_ADDRS) |
+	      BIT(FLOW_DISSECTOR_KEY_ENC_PORTS)	|
+	      BIT(FLOW_DISSECTOR_KEY_ENC_CONTROL) |
+	      BIT(FLOW_DISSECTOR_KEY_TCP) |
+	      BIT(FLOW_DISSECTOR_KEY_IP)  |
+	      BIT(FLOW_DISSECTOR_KEY_ENC_IP))) {
+		netdev_warn(priv->netdev, "Unsupported key used: 0x%x\n",
+			    f->dissector->used_keys);
+		return -EOPNOTSUPP;
+	}
+
+	if ((dissector_uses_key(f->dissector,
+				FLOW_DISSECTOR_KEY_ENC_IPV4_ADDRS) ||
+	     dissector_uses_key(f->dissector, FLOW_DISSECTOR_KEY_ENC_KEYID) ||
+	     dissector_uses_key(f->dissector, FLOW_DISSECTOR_KEY_ENC_PORTS)) &&
+	    dissector_uses_key(f->dissector, FLOW_DISSECTOR_KEY_ENC_CONTROL)) {
+		struct flow_dissector_key_control *key =
+			skb_flow_dissector_target(f->dissector,
+						  FLOW_DISSECTOR_KEY_ENC_CONTROL,
+						  f->key);
+		switch (key->addr_type) {
+		case FLOW_DISSECTOR_KEY_IPV4_ADDRS:
+		case FLOW_DISSECTOR_KEY_IPV6_ADDRS:
+			if (parse_tunnel_attr(priv, spec, f))
+				return -EOPNOTSUPP;
+			break;
+		default:
+			return -EOPNOTSUPP;
+		}
+
+		/* In decap flow, header pointers should point to the inner
+		 * headers, outer header were already set by parse_tunnel_attr
+		 */
+		headers_c = MLX5_ADDR_OF(fte_match_param, spec->match_criteria,
+					 inner_headers);
+		headers_v = MLX5_ADDR_OF(fte_match_param, spec->match_value,
+					 inner_headers);
+	}
+
+	if (dissector_uses_key(f->dissector, FLOW_DISSECTOR_KEY_BASIC)) {
+		struct flow_dissector_key_basic *key =
+			skb_flow_dissector_target(f->dissector,
+						  FLOW_DISSECTOR_KEY_BASIC,
+						  f->key);
+		struct flow_dissector_key_basic *mask =
+			skb_flow_dissector_target(f->dissector,
+						  FLOW_DISSECTOR_KEY_BASIC,
+						  f->mask);
+		MLX5_SET(fte_match_set_lyr_2_4, headers_c, ethertype,
+			 ntohs(mask->n_proto));
+		MLX5_SET(fte_match_set_lyr_2_4, headers_v, ethertype,
+			 ntohs(key->n_proto));
+
+		if (mask->n_proto)
+			*match_level = MLX5_MATCH_L2;
+	}
+
+	if (dissector_uses_key(f->dissector, FLOW_DISSECTOR_KEY_VLAN)) {
+		struct flow_dissector_key_vlan *key =
+			skb_flow_dissector_target(f->dissector,
+						  FLOW_DISSECTOR_KEY_VLAN,
+						  f->key);
+		struct flow_dissector_key_vlan *mask =
+			skb_flow_dissector_target(f->dissector,
+						  FLOW_DISSECTOR_KEY_VLAN,
+						  f->mask);
+		if (mask->vlan_id || mask->vlan_priority || mask->vlan_tpid) {
+			if (key->vlan_tpid == htons(ETH_P_8021AD)) {
+				MLX5_SET(fte_match_set_lyr_2_4, headers_c,
+					 svlan_tag, 1);
+				MLX5_SET(fte_match_set_lyr_2_4, headers_v,
+					 svlan_tag, 1);
+			} else {
+				MLX5_SET(fte_match_set_lyr_2_4, headers_c,
+					 cvlan_tag, 1);
+				MLX5_SET(fte_match_set_lyr_2_4, headers_v,
+					 cvlan_tag, 1);
+			}
+
+			MLX5_SET(fte_match_set_lyr_2_4, headers_c, first_vid, mask->vlan_id);
+			MLX5_SET(fte_match_set_lyr_2_4, headers_v, first_vid, key->vlan_id);
+
+			MLX5_SET(fte_match_set_lyr_2_4, headers_c, first_prio, mask->vlan_priority);
+			MLX5_SET(fte_match_set_lyr_2_4, headers_v, first_prio, key->vlan_priority);
+
+			*match_level = MLX5_MATCH_L2;
+		}
+	} else if (*match_level != MLX5_MATCH_NONE) {
+		MLX5_SET(fte_match_set_lyr_2_4, headers_c, svlan_tag, 1);
+		MLX5_SET(fte_match_set_lyr_2_4, headers_c, cvlan_tag, 1);
+		*match_level = MLX5_MATCH_L2;
+	}
+
+	if (dissector_uses_key(f->dissector, FLOW_DISSECTOR_KEY_CVLAN)) {
+		struct flow_dissector_key_vlan *key =
+			skb_flow_dissector_target(f->dissector,
+						  FLOW_DISSECTOR_KEY_CVLAN,
+						  f->key);
+		struct flow_dissector_key_vlan *mask =
+			skb_flow_dissector_target(f->dissector,
+						  FLOW_DISSECTOR_KEY_CVLAN,
+						  f->mask);
+		if (mask->vlan_id || mask->vlan_priority || mask->vlan_tpid) {
+			if (key->vlan_tpid == htons(ETH_P_8021AD)) {
+				MLX5_SET(fte_match_set_misc, misc_c,
+					 outer_second_svlan_tag, 1);
+				MLX5_SET(fte_match_set_misc, misc_v,
+					 outer_second_svlan_tag, 1);
+			} else {
+				MLX5_SET(fte_match_set_misc, misc_c,
+					 outer_second_cvlan_tag, 1);
+				MLX5_SET(fte_match_set_misc, misc_v,
+					 outer_second_cvlan_tag, 1);
+			}
+
+			MLX5_SET(fte_match_set_misc, misc_c, outer_second_vid,
+				 mask->vlan_id);
+			MLX5_SET(fte_match_set_misc, misc_v, outer_second_vid,
+				 key->vlan_id);
+			MLX5_SET(fte_match_set_misc, misc_c, outer_second_prio,
+				 mask->vlan_priority);
+			MLX5_SET(fte_match_set_misc, misc_v, outer_second_prio,
+				 key->vlan_priority);
+
+			*match_level = MLX5_MATCH_L2;
+		}
+	}
+
+	if (dissector_uses_key(f->dissector, FLOW_DISSECTOR_KEY_ETH_ADDRS)) {
+		struct flow_dissector_key_eth_addrs *key =
+			skb_flow_dissector_target(f->dissector,
+						  FLOW_DISSECTOR_KEY_ETH_ADDRS,
+						  f->key);
+		struct flow_dissector_key_eth_addrs *mask =
+			skb_flow_dissector_target(f->dissector,
+						  FLOW_DISSECTOR_KEY_ETH_ADDRS,
+						  f->mask);
+
+		ether_addr_copy(MLX5_ADDR_OF(fte_match_set_lyr_2_4, headers_c,
+					     dmac_47_16),
+				mask->dst);
+		ether_addr_copy(MLX5_ADDR_OF(fte_match_set_lyr_2_4, headers_v,
+					     dmac_47_16),
+				key->dst);
+
+		ether_addr_copy(MLX5_ADDR_OF(fte_match_set_lyr_2_4, headers_c,
+					     smac_47_16),
+				mask->src);
+		ether_addr_copy(MLX5_ADDR_OF(fte_match_set_lyr_2_4, headers_v,
+					     smac_47_16),
+				key->src);
+
+		if (!is_zero_ether_addr(mask->src) || !is_zero_ether_addr(mask->dst))
+			*match_level = MLX5_MATCH_L2;
+	}
+
+	if (dissector_uses_key(f->dissector, FLOW_DISSECTOR_KEY_CONTROL)) {
+		struct flow_dissector_key_control *key =
+			skb_flow_dissector_target(f->dissector,
+						  FLOW_DISSECTOR_KEY_CONTROL,
+						  f->key);
+
+		struct flow_dissector_key_control *mask =
+			skb_flow_dissector_target(f->dissector,
+						  FLOW_DISSECTOR_KEY_CONTROL,
+						  f->mask);
+		addr_type = key->addr_type;
+
+		/* the HW doesn't support frag first/later */
+		if (mask->flags & FLOW_DIS_FIRST_FRAG)
+			return -EOPNOTSUPP;
+
+		if (mask->flags & FLOW_DIS_IS_FRAGMENT) {
+			MLX5_SET(fte_match_set_lyr_2_4, headers_c, frag, 1);
+			MLX5_SET(fte_match_set_lyr_2_4, headers_v, frag,
+				 key->flags & FLOW_DIS_IS_FRAGMENT);
+
+			/* the HW doesn't need L3 inline to match on frag=no */
+			if (!(key->flags & FLOW_DIS_IS_FRAGMENT))
+				*match_level = MLX5_MATCH_L2;
+	/* ***  L2 attributes parsing up to here *** */
+			else
+				*match_level = MLX5_MATCH_L3;
+		}
+	}
+
+	if (dissector_uses_key(f->dissector, FLOW_DISSECTOR_KEY_BASIC)) {
+		struct flow_dissector_key_basic *key =
+			skb_flow_dissector_target(f->dissector,
+						  FLOW_DISSECTOR_KEY_BASIC,
+						  f->key);
+		struct flow_dissector_key_basic *mask =
+			skb_flow_dissector_target(f->dissector,
+						  FLOW_DISSECTOR_KEY_BASIC,
+						  f->mask);
+		ip_proto = key->ip_proto;
+
+		MLX5_SET(fte_match_set_lyr_2_4, headers_c, ip_protocol,
+			 mask->ip_proto);
+		MLX5_SET(fte_match_set_lyr_2_4, headers_v, ip_protocol,
+			 key->ip_proto);
+
+		if (mask->ip_proto)
+			*match_level = MLX5_MATCH_L3;
+	}
+
+	if (addr_type == FLOW_DISSECTOR_KEY_IPV4_ADDRS) {
+		struct flow_dissector_key_ipv4_addrs *key =
+			skb_flow_dissector_target(f->dissector,
+						  FLOW_DISSECTOR_KEY_IPV4_ADDRS,
+						  f->key);
+		struct flow_dissector_key_ipv4_addrs *mask =
+			skb_flow_dissector_target(f->dissector,
+						  FLOW_DISSECTOR_KEY_IPV4_ADDRS,
+						  f->mask);
+
+		memcpy(MLX5_ADDR_OF(fte_match_set_lyr_2_4, headers_c,
+				    src_ipv4_src_ipv6.ipv4_layout.ipv4),
+		       &mask->src, sizeof(mask->src));
+		memcpy(MLX5_ADDR_OF(fte_match_set_lyr_2_4, headers_v,
+				    src_ipv4_src_ipv6.ipv4_layout.ipv4),
+		       &key->src, sizeof(key->src));
+		memcpy(MLX5_ADDR_OF(fte_match_set_lyr_2_4, headers_c,
+				    dst_ipv4_dst_ipv6.ipv4_layout.ipv4),
+		       &mask->dst, sizeof(mask->dst));
+		memcpy(MLX5_ADDR_OF(fte_match_set_lyr_2_4, headers_v,
+				    dst_ipv4_dst_ipv6.ipv4_layout.ipv4),
+		       &key->dst, sizeof(key->dst));
+
+		if (mask->src || mask->dst)
+			*match_level = MLX5_MATCH_L3;
+	}
+
+	if (addr_type == FLOW_DISSECTOR_KEY_IPV6_ADDRS) {
+		struct flow_dissector_key_ipv6_addrs *key =
+			skb_flow_dissector_target(f->dissector,
+						  FLOW_DISSECTOR_KEY_IPV6_ADDRS,
+						  f->key);
+		struct flow_dissector_key_ipv6_addrs *mask =
+			skb_flow_dissector_target(f->dissector,
+						  FLOW_DISSECTOR_KEY_IPV6_ADDRS,
+						  f->mask);
+
+		memcpy(MLX5_ADDR_OF(fte_match_set_lyr_2_4, headers_c,
+				    src_ipv4_src_ipv6.ipv6_layout.ipv6),
+		       &mask->src, sizeof(mask->src));
+		memcpy(MLX5_ADDR_OF(fte_match_set_lyr_2_4, headers_v,
+				    src_ipv4_src_ipv6.ipv6_layout.ipv6),
+		       &key->src, sizeof(key->src));
+
+		memcpy(MLX5_ADDR_OF(fte_match_set_lyr_2_4, headers_c,
+				    dst_ipv4_dst_ipv6.ipv6_layout.ipv6),
+		       &mask->dst, sizeof(mask->dst));
+		memcpy(MLX5_ADDR_OF(fte_match_set_lyr_2_4, headers_v,
+				    dst_ipv4_dst_ipv6.ipv6_layout.ipv6),
+		       &key->dst, sizeof(key->dst));
+
+		if (ipv6_addr_type(&mask->src) != IPV6_ADDR_ANY ||
+		    ipv6_addr_type(&mask->dst) != IPV6_ADDR_ANY)
+			*match_level = MLX5_MATCH_L3;
+	}
+
+	if (dissector_uses_key(f->dissector, FLOW_DISSECTOR_KEY_IP)) {
+		struct flow_dissector_key_ip *key =
+			skb_flow_dissector_target(f->dissector,
+						  FLOW_DISSECTOR_KEY_IP,
+						  f->key);
+		struct flow_dissector_key_ip *mask =
+			skb_flow_dissector_target(f->dissector,
+						  FLOW_DISSECTOR_KEY_IP,
+						  f->mask);
+
+		MLX5_SET(fte_match_set_lyr_2_4, headers_c, ip_ecn, mask->tos & 0x3);
+		MLX5_SET(fte_match_set_lyr_2_4, headers_v, ip_ecn, key->tos & 0x3);
+
+		MLX5_SET(fte_match_set_lyr_2_4, headers_c, ip_dscp, mask->tos >> 2);
+		MLX5_SET(fte_match_set_lyr_2_4, headers_v, ip_dscp, key->tos  >> 2);
+
+		MLX5_SET(fte_match_set_lyr_2_4, headers_c, ttl_hoplimit, mask->ttl);
+		MLX5_SET(fte_match_set_lyr_2_4, headers_v, ttl_hoplimit, key->ttl);
+
+		if (mask->ttl &&
+		    !MLX5_CAP_ESW_FLOWTABLE_FDB(priv->mdev,
+						ft_field_support.outer_ipv4_ttl))
+			return -EOPNOTSUPP;
+
+		if (mask->tos || mask->ttl)
+			*match_level = MLX5_MATCH_L3;
+	}
+
+	/* ***  L3 attributes parsing up to here *** */
+
+	if (dissector_uses_key(f->dissector, FLOW_DISSECTOR_KEY_PORTS)) {
+		struct flow_dissector_key_ports *key =
+			skb_flow_dissector_target(f->dissector,
+						  FLOW_DISSECTOR_KEY_PORTS,
+						  f->key);
+		struct flow_dissector_key_ports *mask =
+			skb_flow_dissector_target(f->dissector,
+						  FLOW_DISSECTOR_KEY_PORTS,
+						  f->mask);
+		switch (ip_proto) {
+		case IPPROTO_TCP:
+			MLX5_SET(fte_match_set_lyr_2_4, headers_c,
+				 tcp_sport, ntohs(mask->src));
+			MLX5_SET(fte_match_set_lyr_2_4, headers_v,
+				 tcp_sport, ntohs(key->src));
+
+			MLX5_SET(fte_match_set_lyr_2_4, headers_c,
+				 tcp_dport, ntohs(mask->dst));
+			MLX5_SET(fte_match_set_lyr_2_4, headers_v,
+				 tcp_dport, ntohs(key->dst));
+			break;
+
+		case IPPROTO_UDP:
+			MLX5_SET(fte_match_set_lyr_2_4, headers_c,
+				 udp_sport, ntohs(mask->src));
+			MLX5_SET(fte_match_set_lyr_2_4, headers_v,
+				 udp_sport, ntohs(key->src));
+
+			MLX5_SET(fte_match_set_lyr_2_4, headers_c,
+				 udp_dport, ntohs(mask->dst));
+			MLX5_SET(fte_match_set_lyr_2_4, headers_v,
+				 udp_dport, ntohs(key->dst));
+			break;
+		default:
+			netdev_err(priv->netdev,
+				   "Only UDP and TCP transport are supported\n");
+			return -EINVAL;
+		}
+
+		if (mask->src || mask->dst)
+			*match_level = MLX5_MATCH_L4;
+	}
+
+	if (dissector_uses_key(f->dissector, FLOW_DISSECTOR_KEY_TCP)) {
+		struct flow_dissector_key_tcp *key =
+			skb_flow_dissector_target(f->dissector,
+						  FLOW_DISSECTOR_KEY_TCP,
+						  f->key);
+		struct flow_dissector_key_tcp *mask =
+			skb_flow_dissector_target(f->dissector,
+						  FLOW_DISSECTOR_KEY_TCP,
+						  f->mask);
+
+		MLX5_SET(fte_match_set_lyr_2_4, headers_c, tcp_flags,
+			 ntohs(mask->flags));
+		MLX5_SET(fte_match_set_lyr_2_4, headers_v, tcp_flags,
+			 ntohs(key->flags));
+
+		if (mask->flags)
+			*match_level = MLX5_MATCH_L4;
+	}
+
+	return 0;
+}
+
+static int parse_cls_flower(struct mlx5e_priv *priv,
+			    struct mlx5e_tc_flow *flow,
+			    struct mlx5_flow_spec *spec,
+			    struct tc_cls_flower_offload *f)
+{
+	struct mlx5_core_dev *dev = priv->mdev;
+	struct mlx5_eswitch *esw = dev->priv.eswitch;
+	struct mlx5e_rep_priv *rpriv = priv->ppriv;
+	struct mlx5_eswitch_rep *rep;
+	u8 match_level;
+	int err;
+
+	err = __parse_cls_flower(priv, spec, f, &match_level);
+
+	if (!err && (flow->flags & MLX5E_TC_FLOW_ESWITCH)) {
+		rep = rpriv->rep;
+		if (rep->vport != FDB_UPLINK_VPORT &&
+		    (esw->offloads.inline_mode != MLX5_INLINE_MODE_NONE &&
+		    esw->offloads.inline_mode < match_level)) {
+			netdev_warn(priv->netdev,
+				    "Flow is not offloaded due to min inline setting, required %d actual %d\n",
+				    match_level, esw->offloads.inline_mode);
+			return -EOPNOTSUPP;
+		}
+	}
+
+	if (flow->flags & MLX5E_TC_FLOW_ESWITCH)
+		flow->esw_attr->match_level = match_level;
+	else
+		flow->nic_attr->match_level = match_level;
+
+	return err;
+}
+
+struct pedit_headers {
+	struct ethhdr  eth;
+	struct iphdr   ip4;
+	struct ipv6hdr ip6;
+	struct tcphdr  tcp;
+	struct udphdr  udp;
+};
+
+static int pedit_header_offsets[] = {
+	[TCA_PEDIT_KEY_EX_HDR_TYPE_ETH] = offsetof(struct pedit_headers, eth),
+	[TCA_PEDIT_KEY_EX_HDR_TYPE_IP4] = offsetof(struct pedit_headers, ip4),
+	[TCA_PEDIT_KEY_EX_HDR_TYPE_IP6] = offsetof(struct pedit_headers, ip6),
+	[TCA_PEDIT_KEY_EX_HDR_TYPE_TCP] = offsetof(struct pedit_headers, tcp),
+	[TCA_PEDIT_KEY_EX_HDR_TYPE_UDP] = offsetof(struct pedit_headers, udp),
+};
+
+#define pedit_header(_ph, _htype) ((void *)(_ph) + pedit_header_offsets[_htype])
+
+static int set_pedit_val(u8 hdr_type, u32 mask, u32 val, u32 offset,
+			 struct pedit_headers *masks,
+			 struct pedit_headers *vals)
+{
+	u32 *curr_pmask, *curr_pval;
+
+	if (hdr_type >= __PEDIT_HDR_TYPE_MAX)
+		goto out_err;
+
+	curr_pmask = (u32 *)(pedit_header(masks, hdr_type) + offset);
+	curr_pval  = (u32 *)(pedit_header(vals, hdr_type) + offset);
+
+	if (*curr_pmask & mask)  /* disallow acting twice on the same location */
+		goto out_err;
+
+	*curr_pmask |= mask;
+	*curr_pval  |= (val & mask);
+
+	return 0;
+
+out_err:
+	return -EOPNOTSUPP;
+}
+
+struct mlx5_fields {
+	u8  field;
+	u8  size;
+	u32 offset;
+};
+
+#define OFFLOAD(fw_field, size, field, off) \
+		{MLX5_ACTION_IN_FIELD_OUT_ ## fw_field, size, offsetof(struct pedit_headers, field) + (off)}
+
+static struct mlx5_fields fields[] = {
+	OFFLOAD(DMAC_47_16, 4, eth.h_dest[0], 0),
+	OFFLOAD(DMAC_15_0,  2, eth.h_dest[4], 0),
+	OFFLOAD(SMAC_47_16, 4, eth.h_source[0], 0),
+	OFFLOAD(SMAC_15_0,  2, eth.h_source[4], 0),
+	OFFLOAD(ETHERTYPE,  2, eth.h_proto, 0),
+
+	OFFLOAD(IP_TTL, 1, ip4.ttl,   0),
+	OFFLOAD(SIPV4,  4, ip4.saddr, 0),
+	OFFLOAD(DIPV4,  4, ip4.daddr, 0),
+
+	OFFLOAD(SIPV6_127_96, 4, ip6.saddr.s6_addr32[0], 0),
+	OFFLOAD(SIPV6_95_64,  4, ip6.saddr.s6_addr32[1], 0),
+	OFFLOAD(SIPV6_63_32,  4, ip6.saddr.s6_addr32[2], 0),
+	OFFLOAD(SIPV6_31_0,   4, ip6.saddr.s6_addr32[3], 0),
+	OFFLOAD(DIPV6_127_96, 4, ip6.daddr.s6_addr32[0], 0),
+	OFFLOAD(DIPV6_95_64,  4, ip6.daddr.s6_addr32[1], 0),
+	OFFLOAD(DIPV6_63_32,  4, ip6.daddr.s6_addr32[2], 0),
+	OFFLOAD(DIPV6_31_0,   4, ip6.daddr.s6_addr32[3], 0),
+	OFFLOAD(IPV6_HOPLIMIT, 1, ip6.hop_limit, 0),
+
+	OFFLOAD(TCP_SPORT, 2, tcp.source,  0),
+	OFFLOAD(TCP_DPORT, 2, tcp.dest,    0),
+	OFFLOAD(TCP_FLAGS, 1, tcp.ack_seq, 5),
+
+	OFFLOAD(UDP_SPORT, 2, udp.source, 0),
+	OFFLOAD(UDP_DPORT, 2, udp.dest,   0),
+};
+
+/* On input attr->max_mod_hdr_actions tells how many HW actions can be parsed at
+ * max from the SW pedit action. On success, attr->num_mod_hdr_actions
+ * says how many HW actions were actually parsed.
+ */
+static int offload_pedit_fields(struct pedit_headers *masks,
+				struct pedit_headers *vals,
+				struct mlx5e_tc_flow_parse_attr *parse_attr)
+{
+	struct pedit_headers *set_masks, *add_masks, *set_vals, *add_vals;
+	int i, action_size, nactions, max_actions, first, last, next_z;
+	void *s_masks_p, *a_masks_p, *vals_p;
+	struct mlx5_fields *f;
+	u8 cmd, field_bsize;
+	u32 s_mask, a_mask;
+	unsigned long mask;
+	__be32 mask_be32;
+	__be16 mask_be16;
+	void *action;
+
+	set_masks = &masks[TCA_PEDIT_KEY_EX_CMD_SET];
+	add_masks = &masks[TCA_PEDIT_KEY_EX_CMD_ADD];
+	set_vals = &vals[TCA_PEDIT_KEY_EX_CMD_SET];
+	add_vals = &vals[TCA_PEDIT_KEY_EX_CMD_ADD];
+
+	action_size = MLX5_UN_SZ_BYTES(set_action_in_add_action_in_auto);
+	action = parse_attr->mod_hdr_actions +
+		 parse_attr->num_mod_hdr_actions * action_size;
+
+	max_actions = parse_attr->max_mod_hdr_actions;
+	nactions = parse_attr->num_mod_hdr_actions;
+
+	for (i = 0; i < ARRAY_SIZE(fields); i++) {
+		f = &fields[i];
+		/* avoid seeing bits set from previous iterations */
+		s_mask = 0;
+		a_mask = 0;
+
+		s_masks_p = (void *)set_masks + f->offset;
+		a_masks_p = (void *)add_masks + f->offset;
+
+		memcpy(&s_mask, s_masks_p, f->size);
+		memcpy(&a_mask, a_masks_p, f->size);
+
+		if (!s_mask && !a_mask) /* nothing to offload here */
+			continue;
+
+		if (s_mask && a_mask) {
+			printk(KERN_WARNING "mlx5: can't set and add to the same HW field (%x)\n", f->field);
+			return -EOPNOTSUPP;
+		}
+
+		if (nactions == max_actions) {
+			printk(KERN_WARNING "mlx5: parsed %d pedit actions, can't do more\n", nactions);
+			return -EOPNOTSUPP;
+		}
+
+		if (s_mask) {
+			cmd  = MLX5_ACTION_TYPE_SET;
+			mask = s_mask;
+			vals_p = (void *)set_vals + f->offset;
+			/* clear to denote we consumed this field */
+			memset(s_masks_p, 0, f->size);
+		} else {
+			cmd  = MLX5_ACTION_TYPE_ADD;
+			mask = a_mask;
+			vals_p = (void *)add_vals + f->offset;
+			/* clear to denote we consumed this field */
+			memset(a_masks_p, 0, f->size);
+		}
+
+		field_bsize = f->size * BITS_PER_BYTE;
+
+		if (field_bsize == 32) {
+			mask_be32 = *(__be32 *)&mask;
+			mask = (__force unsigned long)cpu_to_le32(be32_to_cpu(mask_be32));
+		} else if (field_bsize == 16) {
+			mask_be16 = *(__be16 *)&mask;
+			mask = (__force unsigned long)cpu_to_le16(be16_to_cpu(mask_be16));
+		}
+
+		first = find_first_bit(&mask, field_bsize);
+		next_z = find_next_zero_bit(&mask, field_bsize, first);
+		last  = find_last_bit(&mask, field_bsize);
+		if (first < next_z && next_z < last) {
+			printk(KERN_WARNING "mlx5: rewrite of few sub-fields (mask %lx) isn't offloaded\n",
+			       mask);
+			return -EOPNOTSUPP;
+		}
+
+		MLX5_SET(set_action_in, action, action_type, cmd);
+		MLX5_SET(set_action_in, action, field, f->field);
+
+		if (cmd == MLX5_ACTION_TYPE_SET) {
+			MLX5_SET(set_action_in, action, offset, first);
+			/* length is num of bits to be written, zero means length of 32 */
+			MLX5_SET(set_action_in, action, length, (last - first + 1));
+		}
+
+		if (field_bsize == 32)
+			MLX5_SET(set_action_in, action, data, ntohl(*(__be32 *)vals_p) >> first);
+		else if (field_bsize == 16)
+			MLX5_SET(set_action_in, action, data, ntohs(*(__be16 *)vals_p) >> first);
+		else if (field_bsize == 8)
+			MLX5_SET(set_action_in, action, data, *(u8 *)vals_p >> first);
+
+		action += action_size;
+		nactions++;
+	}
+
+	parse_attr->num_mod_hdr_actions = nactions;
+	return 0;
+}
+
+static int alloc_mod_hdr_actions(struct mlx5e_priv *priv,
+				 const struct tc_action *a, int namespace,
+				 struct mlx5e_tc_flow_parse_attr *parse_attr)
+{
+	int nkeys, action_size, max_actions;
+
+	nkeys = tcf_pedit_nkeys(a);
+	action_size = MLX5_UN_SZ_BYTES(set_action_in_add_action_in_auto);
+
+	if (namespace == MLX5_FLOW_NAMESPACE_FDB) /* FDB offloading */
+		max_actions = MLX5_CAP_ESW_FLOWTABLE_FDB(priv->mdev, max_modify_header_actions);
+	else /* namespace is MLX5_FLOW_NAMESPACE_KERNEL - NIC offloading */
+		max_actions = MLX5_CAP_FLOWTABLE_NIC_RX(priv->mdev, max_modify_header_actions);
+
+	/* can get up to crazingly 16 HW actions in 32 bits pedit SW key */
+	max_actions = min(max_actions, nkeys * 16);
+
+	parse_attr->mod_hdr_actions = kcalloc(max_actions, action_size, GFP_KERNEL);
+	if (!parse_attr->mod_hdr_actions)
+		return -ENOMEM;
+
+	parse_attr->max_mod_hdr_actions = max_actions;
+	return 0;
+}
+
+static const struct pedit_headers zero_masks = {};
+
+static int parse_tc_pedit_action(struct mlx5e_priv *priv,
+				 const struct tc_action *a, int namespace,
+				 struct mlx5e_tc_flow_parse_attr *parse_attr)
+{
+	struct pedit_headers masks[__PEDIT_CMD_MAX], vals[__PEDIT_CMD_MAX], *cmd_masks;
+	int nkeys, i, err = -EOPNOTSUPP;
+	u32 mask, val, offset;
+	u8 cmd, htype;
+
+	nkeys = tcf_pedit_nkeys(a);
+
+	memset(masks, 0, sizeof(struct pedit_headers) * __PEDIT_CMD_MAX);
+	memset(vals,  0, sizeof(struct pedit_headers) * __PEDIT_CMD_MAX);
+
+	for (i = 0; i < nkeys; i++) {
+		htype = tcf_pedit_htype(a, i);
+		cmd = tcf_pedit_cmd(a, i);
+		err = -EOPNOTSUPP; /* can't be all optimistic */
+
+		if (htype == TCA_PEDIT_KEY_EX_HDR_TYPE_NETWORK) {
+			netdev_warn(priv->netdev, "legacy pedit isn't offloaded\n");
+			goto out_err;
+		}
+
+		if (cmd != TCA_PEDIT_KEY_EX_CMD_SET && cmd != TCA_PEDIT_KEY_EX_CMD_ADD) {
+			netdev_warn(priv->netdev, "pedit cmd %d isn't offloaded\n", cmd);
+			goto out_err;
+		}
+
+		mask = tcf_pedit_mask(a, i);
+		val = tcf_pedit_val(a, i);
+		offset = tcf_pedit_offset(a, i);
+
+		err = set_pedit_val(htype, ~mask, val, offset, &masks[cmd], &vals[cmd]);
+		if (err)
+			goto out_err;
+	}
+
+	if (!parse_attr->mod_hdr_actions) {
+		err = alloc_mod_hdr_actions(priv, a, namespace, parse_attr);
+		if (err)
+			goto out_err;
+	}
+
+	err = offload_pedit_fields(masks, vals, parse_attr);
+	if (err < 0)
+		goto out_dealloc_parsed_actions;
+
+	for (cmd = 0; cmd < __PEDIT_CMD_MAX; cmd++) {
+		cmd_masks = &masks[cmd];
+		if (memcmp(cmd_masks, &zero_masks, sizeof(zero_masks))) {
+			netdev_warn(priv->netdev, "attempt to offload an unsupported field (cmd %d)\n", cmd);
+			print_hex_dump(KERN_WARNING, "mask: ", DUMP_PREFIX_ADDRESS,
+				       16, 1, cmd_masks, sizeof(zero_masks), true);
+			err = -EOPNOTSUPP;
+			goto out_dealloc_parsed_actions;
+		}
+	}
+
+	return 0;
+
+out_dealloc_parsed_actions:
+	kfree(parse_attr->mod_hdr_actions);
+out_err:
+	return err;
+}
+
+static bool csum_offload_supported(struct mlx5e_priv *priv, u32 action, u32 update_flags)
+{
+	u32 prot_flags = TCA_CSUM_UPDATE_FLAG_IPV4HDR | TCA_CSUM_UPDATE_FLAG_TCP |
+			 TCA_CSUM_UPDATE_FLAG_UDP;
+
+	/*  The HW recalcs checksums only if re-writing headers */
+	if (!(action & MLX5_FLOW_CONTEXT_ACTION_MOD_HDR)) {
+		netdev_warn(priv->netdev,
+			    "TC csum action is only offloaded with pedit\n");
+		return false;
+	}
+
+	if (update_flags & ~prot_flags) {
+		netdev_warn(priv->netdev,
+			    "can't offload TC csum action for some header/s - flags %#x\n",
+			    update_flags);
+		return false;
+	}
+
+	return true;
+}
+
+static bool modify_header_match_supported(struct mlx5_flow_spec *spec,
+					  struct tcf_exts *exts)
+{
+	const struct tc_action *a;
+	bool modify_ip_header;
+	LIST_HEAD(actions);
+	u8 htype, ip_proto;
+	void *headers_v;
+	u16 ethertype;
+	int nkeys, i;
+
+	headers_v = MLX5_ADDR_OF(fte_match_param, spec->match_value, outer_headers);
+	ethertype = MLX5_GET(fte_match_set_lyr_2_4, headers_v, ethertype);
+
+	/* for non-IP we only re-write MACs, so we're okay */
+	if (ethertype != ETH_P_IP && ethertype != ETH_P_IPV6)
+		goto out_ok;
+
+	modify_ip_header = false;
+	tcf_exts_for_each_action(i, a, exts) {
+		int k;
+
+		if (!is_tcf_pedit(a))
+			continue;
+
+		nkeys = tcf_pedit_nkeys(a);
+		for (k = 0; k < nkeys; k++) {
+			htype = tcf_pedit_htype(a, k);
+			if (htype == TCA_PEDIT_KEY_EX_HDR_TYPE_IP4 ||
+			    htype == TCA_PEDIT_KEY_EX_HDR_TYPE_IP6) {
+				modify_ip_header = true;
+				break;
+			}
+		}
+	}
+
+	ip_proto = MLX5_GET(fte_match_set_lyr_2_4, headers_v, ip_protocol);
+	if (modify_ip_header && ip_proto != IPPROTO_TCP &&
+	    ip_proto != IPPROTO_UDP && ip_proto != IPPROTO_ICMP) {
+		pr_info("can't offload re-write of ip proto %d\n", ip_proto);
+		return false;
+	}
+
+out_ok:
+	return true;
+}
+
+static bool actions_match_supported(struct mlx5e_priv *priv,
+				    struct tcf_exts *exts,
+				    struct mlx5e_tc_flow_parse_attr *parse_attr,
+				    struct mlx5e_tc_flow *flow)
+{
+	u32 actions;
+
+	if (flow->flags & MLX5E_TC_FLOW_ESWITCH)
+		actions = flow->esw_attr->action;
+	else
+		actions = flow->nic_attr->action;
+
+	if (flow->flags & MLX5E_TC_FLOW_EGRESS &&
+	    !(actions & MLX5_FLOW_CONTEXT_ACTION_DECAP))
+		return false;
+
+	if (actions & MLX5_FLOW_CONTEXT_ACTION_MOD_HDR)
+		return modify_header_match_supported(&parse_attr->spec, exts);
+
+	return true;
+}
+
+static bool same_hw_devs(struct mlx5e_priv *priv, struct mlx5e_priv *peer_priv)
+{
+	struct mlx5_core_dev *fmdev, *pmdev;
+	u64 fsystem_guid, psystem_guid;
+
+	fmdev = priv->mdev;
+	pmdev = peer_priv->mdev;
+
+	mlx5_query_nic_vport_system_image_guid(fmdev, &fsystem_guid);
+	mlx5_query_nic_vport_system_image_guid(pmdev, &psystem_guid);
+
+	return (fsystem_guid == psystem_guid);
+}
+
+static int parse_tc_nic_actions(struct mlx5e_priv *priv, struct tcf_exts *exts,
+				struct mlx5e_tc_flow_parse_attr *parse_attr,
+				struct mlx5e_tc_flow *flow)
+{
+	struct mlx5_nic_flow_attr *attr = flow->nic_attr;
+	const struct tc_action *a;
+	LIST_HEAD(actions);
+	u32 action = 0;
+	int err, i;
+
+	if (!tcf_exts_has_actions(exts))
+		return -EINVAL;
+
+	attr->flow_tag = MLX5_FS_DEFAULT_FLOW_TAG;
+
+	tcf_exts_for_each_action(i, a, exts) {
+		if (is_tcf_gact_shot(a)) {
+			action |= MLX5_FLOW_CONTEXT_ACTION_DROP;
+			if (MLX5_CAP_FLOWTABLE(priv->mdev,
+					       flow_table_properties_nic_receive.flow_counter))
+				action |= MLX5_FLOW_CONTEXT_ACTION_COUNT;
+			continue;
+		}
+
+		if (is_tcf_pedit(a)) {
+			err = parse_tc_pedit_action(priv, a, MLX5_FLOW_NAMESPACE_KERNEL,
+						    parse_attr);
+			if (err)
+				return err;
+
+			action |= MLX5_FLOW_CONTEXT_ACTION_MOD_HDR |
+				  MLX5_FLOW_CONTEXT_ACTION_FWD_DEST;
+			continue;
+		}
+
+		if (is_tcf_csum(a)) {
+			if (csum_offload_supported(priv, action,
+						   tcf_csum_update_flags(a)))
+				continue;
+
+			return -EOPNOTSUPP;
+		}
+
+		if (is_tcf_mirred_egress_redirect(a)) {
+			struct net_device *peer_dev = tcf_mirred_dev(a);
+
+			if (priv->netdev->netdev_ops == peer_dev->netdev_ops &&
+			    same_hw_devs(priv, netdev_priv(peer_dev))) {
+				parse_attr->mirred_ifindex = peer_dev->ifindex;
+				flow->flags |= MLX5E_TC_FLOW_HAIRPIN;
+				action |= MLX5_FLOW_CONTEXT_ACTION_FWD_DEST |
+					  MLX5_FLOW_CONTEXT_ACTION_COUNT;
+			} else {
+				netdev_warn(priv->netdev, "device %s not on same HW, can't offload\n",
+					    peer_dev->name);
+				return -EINVAL;
+			}
+			continue;
+		}
+
+		if (is_tcf_skbedit_mark(a)) {
+			u32 mark = tcf_skbedit_mark(a);
+
+			if (mark & ~MLX5E_TC_FLOW_ID_MASK) {
+				netdev_warn(priv->netdev, "Bad flow mark - only 16 bit is supported: 0x%x\n",
+					    mark);
+				return -EINVAL;
+			}
+
+			attr->flow_tag = mark;
+			action |= MLX5_FLOW_CONTEXT_ACTION_FWD_DEST;
+			continue;
+		}
+
+		return -EINVAL;
+	}
+
+	attr->action = action;
+	if (!actions_match_supported(priv, exts, parse_attr, flow))
+		return -EOPNOTSUPP;
+
+	return 0;
+}
+
+static inline int cmp_encap_info(struct ip_tunnel_key *a,
+				 struct ip_tunnel_key *b)
+{
+	return memcmp(a, b, sizeof(*a));
+}
+
+static inline int hash_encap_info(struct ip_tunnel_key *key)
+{
+	return jhash(key, sizeof(*key), 0);
+}
+
+static int mlx5e_route_lookup_ipv4(struct mlx5e_priv *priv,
+				   struct net_device *mirred_dev,
+				   struct net_device **out_dev,
+				   struct flowi4 *fl4,
+				   struct neighbour **out_n,
+				   u8 *out_ttl)
+{
+	struct mlx5_eswitch *esw = priv->mdev->priv.eswitch;
+	struct mlx5e_rep_priv *uplink_rpriv;
+	struct rtable *rt;
+	struct neighbour *n = NULL;
+
+#if IS_ENABLED(CONFIG_INET)
+	int ret;
+
+	rt = ip_route_output_key(dev_net(mirred_dev), fl4);
+	ret = PTR_ERR_OR_ZERO(rt);
+	if (ret)
+		return ret;
+#else
+	return -EOPNOTSUPP;
+#endif
+	uplink_rpriv = mlx5_eswitch_get_uplink_priv(esw, REP_ETH);
+	/* if the egress device isn't on the same HW e-switch, we use the uplink */
+	if (!switchdev_port_same_parent_id(priv->netdev, rt->dst.dev))
+		*out_dev = uplink_rpriv->netdev;
+	else
+		*out_dev = rt->dst.dev;
+
+	if (!(*out_ttl))
+		*out_ttl = ip4_dst_hoplimit(&rt->dst);
+	n = dst_neigh_lookup(&rt->dst, &fl4->daddr);
+	ip_rt_put(rt);
+	if (!n)
+		return -ENOMEM;
+
+	*out_n = n;
+	return 0;
+}
+
+static bool is_merged_eswitch_dev(struct mlx5e_priv *priv,
+				  struct net_device *peer_netdev)
+{
+	struct mlx5e_priv *peer_priv;
+
+	peer_priv = netdev_priv(peer_netdev);
+
+	return (MLX5_CAP_ESW(priv->mdev, merged_eswitch) &&
+		(priv->netdev->netdev_ops == peer_netdev->netdev_ops) &&
+		same_hw_devs(priv, peer_priv) &&
+		MLX5_VPORT_MANAGER(peer_priv->mdev) &&
+		(peer_priv->mdev->priv.eswitch->mode == SRIOV_OFFLOADS));
+}
+
+static int mlx5e_route_lookup_ipv6(struct mlx5e_priv *priv,
+				   struct net_device *mirred_dev,
+				   struct net_device **out_dev,
+				   struct flowi6 *fl6,
+				   struct neighbour **out_n,
+				   u8 *out_ttl)
+{
+	struct neighbour *n = NULL;
+	struct dst_entry *dst;
+
+#if IS_ENABLED(CONFIG_INET) && IS_ENABLED(CONFIG_IPV6)
+	struct mlx5e_rep_priv *uplink_rpriv;
+	struct mlx5_eswitch *esw = priv->mdev->priv.eswitch;
+
+	dst = ipv6_stub->ipv6_dst_lookup_flow(dev_net(mirred_dev), NULL, fl6,
+					      NULL);
+	if (IS_ERR(dst))
+		return PTR_ERR(dst);
+
+	if (!(*out_ttl))
+		*out_ttl = ip6_dst_hoplimit(dst);
+
+	uplink_rpriv = mlx5_eswitch_get_uplink_priv(esw, REP_ETH);
+	/* if the egress device isn't on the same HW e-switch, we use the uplink */
+	if (!switchdev_port_same_parent_id(priv->netdev, dst->dev))
+		*out_dev = uplink_rpriv->netdev;
+	else
+		*out_dev = dst->dev;
+#else
+	return -EOPNOTSUPP;
+#endif
+
+	n = dst_neigh_lookup(dst, &fl6->daddr);
+	dst_release(dst);
+	if (!n)
+		return -ENOMEM;
+
+	*out_n = n;
+	return 0;
+}
+
+static void gen_vxlan_header_ipv4(struct net_device *out_dev,
+				  char buf[], int encap_size,
+				  unsigned char h_dest[ETH_ALEN],
+				  u8 tos, u8 ttl,
+				  __be32 daddr,
+				  __be32 saddr,
+				  __be16 udp_dst_port,
+				  __be32 vx_vni)
+{
+	struct ethhdr *eth = (struct ethhdr *)buf;
+	struct iphdr  *ip = (struct iphdr *)((char *)eth + sizeof(struct ethhdr));
+	struct udphdr *udp = (struct udphdr *)((char *)ip + sizeof(struct iphdr));
+	struct vxlanhdr *vxh = (struct vxlanhdr *)((char *)udp + sizeof(struct udphdr));
+
+	memset(buf, 0, encap_size);
+
+	ether_addr_copy(eth->h_dest, h_dest);
+	ether_addr_copy(eth->h_source, out_dev->dev_addr);
+	eth->h_proto = htons(ETH_P_IP);
+
+	ip->daddr = daddr;
+	ip->saddr = saddr;
+
+	ip->tos = tos;
+	ip->ttl = ttl;
+	ip->protocol = IPPROTO_UDP;
+	ip->version = 0x4;
+	ip->ihl = 0x5;
+
+	udp->dest = udp_dst_port;
+	vxh->vx_flags = VXLAN_HF_VNI;
+	vxh->vx_vni = vxlan_vni_field(vx_vni);
+}
+
+static void gen_vxlan_header_ipv6(struct net_device *out_dev,
+				  char buf[], int encap_size,
+				  unsigned char h_dest[ETH_ALEN],
+				  u8 tos, u8 ttl,
+				  struct in6_addr *daddr,
+				  struct in6_addr *saddr,
+				  __be16 udp_dst_port,
+				  __be32 vx_vni)
+{
+	struct ethhdr *eth = (struct ethhdr *)buf;
+	struct ipv6hdr *ip6h = (struct ipv6hdr *)((char *)eth + sizeof(struct ethhdr));
+	struct udphdr *udp = (struct udphdr *)((char *)ip6h + sizeof(struct ipv6hdr));
+	struct vxlanhdr *vxh = (struct vxlanhdr *)((char *)udp + sizeof(struct udphdr));
+
+	memset(buf, 0, encap_size);
+
+	ether_addr_copy(eth->h_dest, h_dest);
+	ether_addr_copy(eth->h_source, out_dev->dev_addr);
+	eth->h_proto = htons(ETH_P_IPV6);
+
+	ip6_flow_hdr(ip6h, tos, 0);
+	/* the HW fills up ipv6 payload len */
+	ip6h->nexthdr     = IPPROTO_UDP;
+	ip6h->hop_limit   = ttl;
+	ip6h->daddr	  = *daddr;
+	ip6h->saddr	  = *saddr;
+
+	udp->dest = udp_dst_port;
+	vxh->vx_flags = VXLAN_HF_VNI;
+	vxh->vx_vni = vxlan_vni_field(vx_vni);
+}
+
+static int mlx5e_create_encap_header_ipv4(struct mlx5e_priv *priv,
+					  struct net_device *mirred_dev,
+					  struct mlx5e_encap_entry *e)
+{
+	int max_encap_size = MLX5_CAP_ESW(priv->mdev, max_encap_header_size);
+	int ipv4_encap_size = ETH_HLEN + sizeof(struct iphdr) + VXLAN_HLEN;
+	struct ip_tunnel_key *tun_key = &e->tun_info.key;
+	struct net_device *out_dev;
+	struct neighbour *n = NULL;
+	struct flowi4 fl4 = {};
+	u8 nud_state, tos, ttl;
+	char *encap_header;
+	int err;
+
+	if (max_encap_size < ipv4_encap_size) {
+		mlx5_core_warn(priv->mdev, "encap size %d too big, max supported is %d\n",
+			       ipv4_encap_size, max_encap_size);
+		return -EOPNOTSUPP;
+	}
+
+	encap_header = kzalloc(ipv4_encap_size, GFP_KERNEL);
+	if (!encap_header)
+		return -ENOMEM;
+
+	switch (e->tunnel_type) {
+	case MLX5_HEADER_TYPE_VXLAN:
+		fl4.flowi4_proto = IPPROTO_UDP;
+		fl4.fl4_dport = tun_key->tp_dst;
+		break;
+	default:
+		err = -EOPNOTSUPP;
+		goto free_encap;
+	}
+
+	tos = tun_key->tos;
+	ttl = tun_key->ttl;
+
+	fl4.flowi4_tos = tun_key->tos;
+	fl4.daddr = tun_key->u.ipv4.dst;
+	fl4.saddr = tun_key->u.ipv4.src;
+
+	err = mlx5e_route_lookup_ipv4(priv, mirred_dev, &out_dev,
+				      &fl4, &n, &ttl);
+	if (err)
+		goto free_encap;
+
+	/* used by mlx5e_detach_encap to lookup a neigh hash table
+	 * entry in the neigh hash table when a user deletes a rule
+	 */
+	e->m_neigh.dev = n->dev;
+	e->m_neigh.family = n->ops->family;
+	memcpy(&e->m_neigh.dst_ip, n->primary_key, n->tbl->key_len);
+	e->out_dev = out_dev;
+
+	/* It's importent to add the neigh to the hash table before checking
+	 * the neigh validity state. So if we'll get a notification, in case the
+	 * neigh changes it's validity state, we would find the relevant neigh
+	 * in the hash.
+	 */
+	err = mlx5e_rep_encap_entry_attach(netdev_priv(out_dev), e);
+	if (err)
+		goto free_encap;
+
+	read_lock_bh(&n->lock);
+	nud_state = n->nud_state;
+	ether_addr_copy(e->h_dest, n->ha);
+	read_unlock_bh(&n->lock);
+
+	switch (e->tunnel_type) {
+	case MLX5_HEADER_TYPE_VXLAN:
+		gen_vxlan_header_ipv4(out_dev, encap_header,
+				      ipv4_encap_size, e->h_dest, tos, ttl,
+				      fl4.daddr,
+				      fl4.saddr, tun_key->tp_dst,
+				      tunnel_id_to_key32(tun_key->tun_id));
+		break;
+	default:
+		err = -EOPNOTSUPP;
+		goto destroy_neigh_entry;
+	}
+	e->encap_size = ipv4_encap_size;
+	e->encap_header = encap_header;
+
+	if (!(nud_state & NUD_VALID)) {
+		neigh_event_send(n, NULL);
+		err = -EAGAIN;
+		goto out;
+	}
+
+	err = mlx5_encap_alloc(priv->mdev, e->tunnel_type,
+			       ipv4_encap_size, encap_header, &e->encap_id);
+	if (err)
+		goto destroy_neigh_entry;
+
+	e->flags |= MLX5_ENCAP_ENTRY_VALID;
+	mlx5e_rep_queue_neigh_stats_work(netdev_priv(out_dev));
+	neigh_release(n);
+	return err;
+
+destroy_neigh_entry:
+	mlx5e_rep_encap_entry_detach(netdev_priv(e->out_dev), e);
+free_encap:
+	kfree(encap_header);
+out:
+	if (n)
+		neigh_release(n);
+	return err;
+}
+
+static int mlx5e_create_encap_header_ipv6(struct mlx5e_priv *priv,
+					  struct net_device *mirred_dev,
+					  struct mlx5e_encap_entry *e)
+{
+	int max_encap_size = MLX5_CAP_ESW(priv->mdev, max_encap_header_size);
+	int ipv6_encap_size = ETH_HLEN + sizeof(struct ipv6hdr) + VXLAN_HLEN;
+	struct ip_tunnel_key *tun_key = &e->tun_info.key;
+	struct net_device *out_dev = NULL;
+	struct neighbour *n = NULL;
+	struct flowi6 fl6 = {};
+	u8 nud_state, tos, ttl;
+	char *encap_header;
+	int err;
+
+	if (max_encap_size < ipv6_encap_size) {
+		mlx5_core_warn(priv->mdev, "encap size %d too big, max supported is %d\n",
+			       ipv6_encap_size, max_encap_size);
+		return -EOPNOTSUPP;
+	}
+
+	encap_header = kzalloc(ipv6_encap_size, GFP_KERNEL);
+	if (!encap_header)
+		return -ENOMEM;
+
+	switch (e->tunnel_type) {
+	case MLX5_HEADER_TYPE_VXLAN:
+		fl6.flowi6_proto = IPPROTO_UDP;
+		fl6.fl6_dport = tun_key->tp_dst;
+		break;
+	default:
+		err = -EOPNOTSUPP;
+		goto free_encap;
+	}
+
+	tos = tun_key->tos;
+	ttl = tun_key->ttl;
+
+	fl6.flowlabel = ip6_make_flowinfo(RT_TOS(tun_key->tos), tun_key->label);
+	fl6.daddr = tun_key->u.ipv6.dst;
+	fl6.saddr = tun_key->u.ipv6.src;
+
+	err = mlx5e_route_lookup_ipv6(priv, mirred_dev, &out_dev,
+				      &fl6, &n, &ttl);
+	if (err)
+		goto free_encap;
+
+	/* used by mlx5e_detach_encap to lookup a neigh hash table
+	 * entry in the neigh hash table when a user deletes a rule
+	 */
+	e->m_neigh.dev = n->dev;
+	e->m_neigh.family = n->ops->family;
+	memcpy(&e->m_neigh.dst_ip, n->primary_key, n->tbl->key_len);
+	e->out_dev = out_dev;
+
+	/* It's importent to add the neigh to the hash table before checking
+	 * the neigh validity state. So if we'll get a notification, in case the
+	 * neigh changes it's validity state, we would find the relevant neigh
+	 * in the hash.
+	 */
+	err = mlx5e_rep_encap_entry_attach(netdev_priv(out_dev), e);
+	if (err)
+		goto free_encap;
+
+	read_lock_bh(&n->lock);
+	nud_state = n->nud_state;
+	ether_addr_copy(e->h_dest, n->ha);
+	read_unlock_bh(&n->lock);
+
+	switch (e->tunnel_type) {
+	case MLX5_HEADER_TYPE_VXLAN:
+		gen_vxlan_header_ipv6(out_dev, encap_header,
+				      ipv6_encap_size, e->h_dest, tos, ttl,
+				      &fl6.daddr,
+				      &fl6.saddr, tun_key->tp_dst,
+				      tunnel_id_to_key32(tun_key->tun_id));
+		break;
+	default:
+		err = -EOPNOTSUPP;
+		goto destroy_neigh_entry;
+	}
+
+	e->encap_size = ipv6_encap_size;
+	e->encap_header = encap_header;
+
+	if (!(nud_state & NUD_VALID)) {
+		neigh_event_send(n, NULL);
+		err = -EAGAIN;
+		goto out;
+	}
+
+	err = mlx5_encap_alloc(priv->mdev, e->tunnel_type,
+			       ipv6_encap_size, encap_header, &e->encap_id);
+	if (err)
+		goto destroy_neigh_entry;
+
+	e->flags |= MLX5_ENCAP_ENTRY_VALID;
+	mlx5e_rep_queue_neigh_stats_work(netdev_priv(out_dev));
+	neigh_release(n);
+	return err;
+
+destroy_neigh_entry:
+	mlx5e_rep_encap_entry_detach(netdev_priv(e->out_dev), e);
+free_encap:
+	kfree(encap_header);
+out:
+	if (n)
+		neigh_release(n);
+	return err;
+}
+
+static int mlx5e_attach_encap(struct mlx5e_priv *priv,
+			      struct ip_tunnel_info *tun_info,
+			      struct net_device *mirred_dev,
+			      struct net_device **encap_dev,
+			      struct mlx5e_tc_flow *flow)
+{
+	struct mlx5_eswitch *esw = priv->mdev->priv.eswitch;
+	unsigned short family = ip_tunnel_info_af(tun_info);
+	struct mlx5_esw_flow_attr *attr = flow->esw_attr;
+	struct ip_tunnel_key *key = &tun_info->key;
+	struct mlx5e_encap_entry *e;
+	int tunnel_type, err = 0;
+	uintptr_t hash_key;
+	bool found = false;
+
+	/* udp dst port must be set */
+	if (!memchr_inv(&key->tp_dst, 0, sizeof(key->tp_dst)))
+		goto vxlan_encap_offload_err;
+
+	/* setting udp src port isn't supported */
+	if (memchr_inv(&key->tp_src, 0, sizeof(key->tp_src))) {
+vxlan_encap_offload_err:
+		netdev_warn(priv->netdev,
+			    "must set udp dst port and not set udp src port\n");
+		return -EOPNOTSUPP;
+	}
+
+	if (mlx5_vxlan_lookup_port(priv->mdev->vxlan, be16_to_cpu(key->tp_dst)) &&
+	    MLX5_CAP_ESW(priv->mdev, vxlan_encap_decap)) {
+		tunnel_type = MLX5_HEADER_TYPE_VXLAN;
+	} else {
+		netdev_warn(priv->netdev,
+			    "%d isn't an offloaded vxlan udp dport\n", be16_to_cpu(key->tp_dst));
+		return -EOPNOTSUPP;
+	}
+
+	hash_key = hash_encap_info(key);
+
+	hash_for_each_possible_rcu(esw->offloads.encap_tbl, e,
+				   encap_hlist, hash_key) {
+		if (!cmp_encap_info(&e->tun_info.key, key)) {
+			found = true;
+			break;
+		}
+	}
+
+	/* must verify if encap is valid or not */
+	if (found)
+		goto attach_flow;
+
+	e = kzalloc(sizeof(*e), GFP_KERNEL);
+	if (!e)
+		return -ENOMEM;
+
+	e->tun_info = *tun_info;
+	e->tunnel_type = tunnel_type;
+	INIT_LIST_HEAD(&e->flows);
+
+	if (family == AF_INET)
+		err = mlx5e_create_encap_header_ipv4(priv, mirred_dev, e);
+	else if (family == AF_INET6)
+		err = mlx5e_create_encap_header_ipv6(priv, mirred_dev, e);
+
+	if (err && err != -EAGAIN)
+		goto out_err;
+
+	hash_add_rcu(esw->offloads.encap_tbl, &e->encap_hlist, hash_key);
+
+attach_flow:
+	list_add(&flow->encap, &e->flows);
+	*encap_dev = e->out_dev;
+	if (e->flags & MLX5_ENCAP_ENTRY_VALID)
+		attr->encap_id = e->encap_id;
+	else
+		err = -EAGAIN;
+
+	return err;
+
+out_err:
+	kfree(e);
+	return err;
+}
+
+static int parse_tc_vlan_action(struct mlx5e_priv *priv,
+				const struct tc_action *a,
+				struct mlx5_esw_flow_attr *attr,
+				u32 *action)
+{
+	u8 vlan_idx = attr->total_vlan;
+
+	if (vlan_idx >= MLX5_FS_VLAN_DEPTH)
+		return -EOPNOTSUPP;
+
+	if (tcf_vlan_action(a) == TCA_VLAN_ACT_POP) {
+		if (vlan_idx) {
+			if (!mlx5_eswitch_vlan_actions_supported(priv->mdev,
+								 MLX5_FS_VLAN_DEPTH))
+				return -EOPNOTSUPP;
+
+			*action |= MLX5_FLOW_CONTEXT_ACTION_VLAN_POP_2;
+		} else {
+			*action |= MLX5_FLOW_CONTEXT_ACTION_VLAN_POP;
+		}
+	} else if (tcf_vlan_action(a) == TCA_VLAN_ACT_PUSH) {
+		attr->vlan_vid[vlan_idx] = tcf_vlan_push_vid(a);
+		attr->vlan_prio[vlan_idx] = tcf_vlan_push_prio(a);
+		attr->vlan_proto[vlan_idx] = tcf_vlan_push_proto(a);
+		if (!attr->vlan_proto[vlan_idx])
+			attr->vlan_proto[vlan_idx] = htons(ETH_P_8021Q);
+
+		if (vlan_idx) {
+			if (!mlx5_eswitch_vlan_actions_supported(priv->mdev,
+								 MLX5_FS_VLAN_DEPTH))
+				return -EOPNOTSUPP;
+
+			*action |= MLX5_FLOW_CONTEXT_ACTION_VLAN_PUSH_2;
+		} else {
+			if (!mlx5_eswitch_vlan_actions_supported(priv->mdev, 1) &&
+			    (tcf_vlan_push_proto(a) != htons(ETH_P_8021Q) ||
+			     tcf_vlan_push_prio(a)))
+				return -EOPNOTSUPP;
+
+			*action |= MLX5_FLOW_CONTEXT_ACTION_VLAN_PUSH;
+		}
+	} else { /* action is TCA_VLAN_ACT_MODIFY */
+		return -EOPNOTSUPP;
+	}
+
+	attr->total_vlan = vlan_idx + 1;
+
+	return 0;
+}
+
+static int parse_tc_fdb_actions(struct mlx5e_priv *priv, struct tcf_exts *exts,
+				struct mlx5e_tc_flow_parse_attr *parse_attr,
+				struct mlx5e_tc_flow *flow)
+{
+	struct mlx5_esw_flow_attr *attr = flow->esw_attr;
+	struct mlx5e_rep_priv *rpriv = priv->ppriv;
+	struct ip_tunnel_info *info = NULL;
+	const struct tc_action *a;
+	LIST_HEAD(actions);
+	bool encap = false;
+	u32 action = 0;
+	int err, i;
+
+	if (!tcf_exts_has_actions(exts))
+		return -EINVAL;
+
+	attr->in_rep = rpriv->rep;
+	attr->in_mdev = priv->mdev;
+
+	tcf_exts_for_each_action(i, a, exts) {
+		if (is_tcf_gact_shot(a)) {
+			action |= MLX5_FLOW_CONTEXT_ACTION_DROP |
+				  MLX5_FLOW_CONTEXT_ACTION_COUNT;
+			continue;
+		}
+
+		if (is_tcf_pedit(a)) {
+			err = parse_tc_pedit_action(priv, a, MLX5_FLOW_NAMESPACE_FDB,
+						    parse_attr);
+			if (err)
+				return err;
+
+			action |= MLX5_FLOW_CONTEXT_ACTION_MOD_HDR;
+			attr->mirror_count = attr->out_count;
+			continue;
+		}
+
+		if (is_tcf_csum(a)) {
+			if (csum_offload_supported(priv, action,
+						   tcf_csum_update_flags(a)))
+				continue;
+
+			return -EOPNOTSUPP;
+		}
+
+		if (is_tcf_mirred_egress_redirect(a) || is_tcf_mirred_egress_mirror(a)) {
+			struct mlx5e_priv *out_priv;
+			struct net_device *out_dev;
+
+			out_dev = tcf_mirred_dev(a);
+
+			if (attr->out_count >= MLX5_MAX_FLOW_FWD_VPORTS) {
+				pr_err("can't support more than %d output ports, can't offload forwarding\n",
+				       attr->out_count);
+				return -EOPNOTSUPP;
+			}
+
+			if (switchdev_port_same_parent_id(priv->netdev,
+							  out_dev) ||
+			    is_merged_eswitch_dev(priv, out_dev)) {
+				action |= MLX5_FLOW_CONTEXT_ACTION_FWD_DEST |
+					  MLX5_FLOW_CONTEXT_ACTION_COUNT;
+				out_priv = netdev_priv(out_dev);
+				rpriv = out_priv->ppriv;
+				attr->out_rep[attr->out_count] = rpriv->rep;
+				attr->out_mdev[attr->out_count++] = out_priv->mdev;
+			} else if (encap) {
+				parse_attr->mirred_ifindex = out_dev->ifindex;
+				parse_attr->tun_info = *info;
+				attr->parse_attr = parse_attr;
+				action |= MLX5_FLOW_CONTEXT_ACTION_ENCAP |
+					  MLX5_FLOW_CONTEXT_ACTION_FWD_DEST |
+					  MLX5_FLOW_CONTEXT_ACTION_COUNT;
+				/* attr->out_rep is resolved when we handle encap */
+			} else {
+				pr_err("devices %s %s not on same switch HW, can't offload forwarding\n",
+				       priv->netdev->name, out_dev->name);
+				return -EINVAL;
+			}
+			continue;
+		}
+
+		if (is_tcf_tunnel_set(a)) {
+			info = tcf_tunnel_info(a);
+			if (info)
+				encap = true;
+			else
+				return -EOPNOTSUPP;
+			attr->mirror_count = attr->out_count;
+			continue;
+		}
+
+		if (is_tcf_vlan(a)) {
+			err = parse_tc_vlan_action(priv, a, attr, &action);
+
+			if (err)
+				return err;
+
+			attr->mirror_count = attr->out_count;
+			continue;
+		}
+
+		if (is_tcf_tunnel_release(a)) {
+			action |= MLX5_FLOW_CONTEXT_ACTION_DECAP;
+			continue;
+		}
+
+		return -EINVAL;
+	}
+
+	attr->action = action;
+	if (!actions_match_supported(priv, exts, parse_attr, flow))
+		return -EOPNOTSUPP;
+
+	if (attr->out_count > 1 && !mlx5_esw_has_fwd_fdb(priv->mdev)) {
+		netdev_warn_once(priv->netdev, "current firmware doesn't support split rule for port mirroring\n");
+		return -EOPNOTSUPP;
+	}
+
+	return 0;
+}
+
+static void get_flags(int flags, u8 *flow_flags)
+{
+	u8 __flow_flags = 0;
+
+	if (flags & MLX5E_TC_INGRESS)
+		__flow_flags |= MLX5E_TC_FLOW_INGRESS;
+	if (flags & MLX5E_TC_EGRESS)
+		__flow_flags |= MLX5E_TC_FLOW_EGRESS;
+
+	*flow_flags = __flow_flags;
+}
+
+static const struct rhashtable_params tc_ht_params = {
+	.head_offset = offsetof(struct mlx5e_tc_flow, node),
+	.key_offset = offsetof(struct mlx5e_tc_flow, cookie),
+	.key_len = sizeof(((struct mlx5e_tc_flow *)0)->cookie),
+	.automatic_shrinking = true,
+};
+
+static struct rhashtable *get_tc_ht(struct mlx5e_priv *priv)
+{
+	struct mlx5_eswitch *esw = priv->mdev->priv.eswitch;
+	struct mlx5e_rep_priv *uplink_rpriv;
+
+	if (MLX5_VPORT_MANAGER(priv->mdev) && esw->mode == SRIOV_OFFLOADS) {
+		uplink_rpriv = mlx5_eswitch_get_uplink_priv(esw, REP_ETH);
+		return &uplink_rpriv->tc_ht;
+	} else
+		return &priv->fs.tc.ht;
+}
+
+int mlx5e_configure_flower(struct mlx5e_priv *priv,
+			   struct tc_cls_flower_offload *f, int flags)
+{
+	struct mlx5_eswitch *esw = priv->mdev->priv.eswitch;
+	struct mlx5e_tc_flow_parse_attr *parse_attr;
+	struct rhashtable *tc_ht = get_tc_ht(priv);
+	struct mlx5e_tc_flow *flow;
+	int attr_size, err = 0;
+	u8 flow_flags = 0;
+
+	get_flags(flags, &flow_flags);
+
+	flow = rhashtable_lookup_fast(tc_ht, &f->cookie, tc_ht_params);
+	if (flow) {
+		netdev_warn_once(priv->netdev, "flow cookie %lx already exists, ignoring\n", f->cookie);
+		return 0;
+	}
+
+	if (esw && esw->mode == SRIOV_OFFLOADS) {
+		flow_flags |= MLX5E_TC_FLOW_ESWITCH;
+		attr_size  = sizeof(struct mlx5_esw_flow_attr);
+	} else {
+		flow_flags |= MLX5E_TC_FLOW_NIC;
+		attr_size  = sizeof(struct mlx5_nic_flow_attr);
+	}
+
+	flow = kzalloc(sizeof(*flow) + attr_size, GFP_KERNEL);
+	parse_attr = kvzalloc(sizeof(*parse_attr), GFP_KERNEL);
+	if (!parse_attr || !flow) {
+		err = -ENOMEM;
+		goto err_free;
+	}
+
+	flow->cookie = f->cookie;
+	flow->flags = flow_flags;
+	flow->priv = priv;
+
+	err = parse_cls_flower(priv, flow, &parse_attr->spec, f);
+	if (err < 0)
+		goto err_free;
+
+	if (flow->flags & MLX5E_TC_FLOW_ESWITCH) {
+		err = parse_tc_fdb_actions(priv, f->exts, parse_attr, flow);
+		if (err < 0)
+			goto err_free;
+		flow->rule[0] = mlx5e_tc_add_fdb_flow(priv, parse_attr, flow);
+	} else {
+		err = parse_tc_nic_actions(priv, f->exts, parse_attr, flow);
+		if (err < 0)
+			goto err_free;
+		flow->rule[0] = mlx5e_tc_add_nic_flow(priv, parse_attr, flow);
+	}
+
+	if (IS_ERR(flow->rule[0])) {
+		err = PTR_ERR(flow->rule[0]);
+		if (err != -EAGAIN)
+			goto err_free;
+	}
+
+	if (err != -EAGAIN)
+		flow->flags |= MLX5E_TC_FLOW_OFFLOADED;
+
+	if (!(flow->flags & MLX5E_TC_FLOW_ESWITCH) ||
+	    !(flow->esw_attr->action & MLX5_FLOW_CONTEXT_ACTION_ENCAP))
+		kvfree(parse_attr);
+
+	err = rhashtable_insert_fast(tc_ht, &flow->node, tc_ht_params);
+	if (err) {
+		mlx5e_tc_del_flow(priv, flow);
+		kfree(flow);
+	}
+
+	return err;
+
+err_free:
+	kvfree(parse_attr);
+	kfree(flow);
+	return err;
+}
+
+#define DIRECTION_MASK (MLX5E_TC_INGRESS | MLX5E_TC_EGRESS)
+#define FLOW_DIRECTION_MASK (MLX5E_TC_FLOW_INGRESS | MLX5E_TC_FLOW_EGRESS)
+
+static bool same_flow_direction(struct mlx5e_tc_flow *flow, int flags)
+{
+	if ((flow->flags & FLOW_DIRECTION_MASK) == (flags & DIRECTION_MASK))
+		return true;
+
+	return false;
+}
+
+int mlx5e_delete_flower(struct mlx5e_priv *priv,
+			struct tc_cls_flower_offload *f, int flags)
+{
+	struct rhashtable *tc_ht = get_tc_ht(priv);
+	struct mlx5e_tc_flow *flow;
+
+	flow = rhashtable_lookup_fast(tc_ht, &f->cookie, tc_ht_params);
+	if (!flow || !same_flow_direction(flow, flags))
+		return -EINVAL;
+
+	rhashtable_remove_fast(tc_ht, &flow->node, tc_ht_params);
+
+	mlx5e_tc_del_flow(priv, flow);
+
+	kfree(flow);
+
+	return 0;
+}
+
+int mlx5e_stats_flower(struct mlx5e_priv *priv,
+		       struct tc_cls_flower_offload *f, int flags)
+{
+	struct rhashtable *tc_ht = get_tc_ht(priv);
+	struct mlx5e_tc_flow *flow;
+	struct mlx5_fc *counter;
+	u64 bytes;
+	u64 packets;
+	u64 lastuse;
+
+	flow = rhashtable_lookup_fast(tc_ht, &f->cookie, tc_ht_params);
+	if (!flow || !same_flow_direction(flow, flags))
+		return -EINVAL;
+
+	if (!(flow->flags & MLX5E_TC_FLOW_OFFLOADED))
+		return 0;
+
+	counter = mlx5_flow_rule_counter(flow->rule[0]);
+	if (!counter)
+		return 0;
+
+	mlx5_fc_query_cached(counter, &bytes, &packets, &lastuse);
+
+	tcf_exts_stats_update(f->exts, bytes, packets, lastuse);
+
+	return 0;
+}
+
+static void mlx5e_tc_hairpin_update_dead_peer(struct mlx5e_priv *priv,
+					      struct mlx5e_priv *peer_priv)
+{
+	struct mlx5_core_dev *peer_mdev = peer_priv->mdev;
+	struct mlx5e_hairpin_entry *hpe;
+	u16 peer_vhca_id;
+	int bkt;
+
+	if (!same_hw_devs(priv, peer_priv))
+		return;
+
+	peer_vhca_id = MLX5_CAP_GEN(peer_mdev, vhca_id);
+
+	hash_for_each(priv->fs.tc.hairpin_tbl, bkt, hpe, hairpin_hlist) {
+		if (hpe->peer_vhca_id == peer_vhca_id)
+			hpe->hp->pair->peer_gone = true;
+	}
+}
+
+static int mlx5e_tc_netdev_event(struct notifier_block *this,
+				 unsigned long event, void *ptr)
+{
+	struct net_device *ndev = netdev_notifier_info_to_dev(ptr);
+	struct mlx5e_flow_steering *fs;
+	struct mlx5e_priv *peer_priv;
+	struct mlx5e_tc_table *tc;
+	struct mlx5e_priv *priv;
+
+	if (ndev->netdev_ops != &mlx5e_netdev_ops ||
+	    event != NETDEV_UNREGISTER ||
+	    ndev->reg_state == NETREG_REGISTERED)
+		return NOTIFY_DONE;
+
+	tc = container_of(this, struct mlx5e_tc_table, netdevice_nb);
+	fs = container_of(tc, struct mlx5e_flow_steering, tc);
+	priv = container_of(fs, struct mlx5e_priv, fs);
+	peer_priv = netdev_priv(ndev);
+	if (priv == peer_priv ||
+	    !(priv->netdev->features & NETIF_F_HW_TC))
+		return NOTIFY_DONE;
+
+	mlx5e_tc_hairpin_update_dead_peer(priv, peer_priv);
+
+	return NOTIFY_DONE;
+}
+
+int mlx5e_tc_nic_init(struct mlx5e_priv *priv)
+{
+	struct mlx5e_tc_table *tc = &priv->fs.tc;
+	int err;
+
+	hash_init(tc->mod_hdr_tbl);
+	hash_init(tc->hairpin_tbl);
+
+	err = rhashtable_init(&tc->ht, &tc_ht_params);
+	if (err)
+		return err;
+
+	tc->netdevice_nb.notifier_call = mlx5e_tc_netdev_event;
+	if (register_netdevice_notifier(&tc->netdevice_nb)) {
+		tc->netdevice_nb.notifier_call = NULL;
+		mlx5_core_warn(priv->mdev, "Failed to register netdev notifier\n");
+	}
+
+	return err;
+}
+
+static void _mlx5e_tc_del_flow(void *ptr, void *arg)
+{
+	struct mlx5e_tc_flow *flow = ptr;
+	struct mlx5e_priv *priv = flow->priv;
+
+	mlx5e_tc_del_flow(priv, flow);
+	kfree(flow);
+}
+
+void mlx5e_tc_nic_cleanup(struct mlx5e_priv *priv)
+{
+	struct mlx5e_tc_table *tc = &priv->fs.tc;
+
+	if (tc->netdevice_nb.notifier_call)
+		unregister_netdevice_notifier(&tc->netdevice_nb);
+
+	rhashtable_free_and_destroy(&tc->ht, _mlx5e_tc_del_flow, NULL);
+
+	if (!IS_ERR_OR_NULL(tc->t)) {
+		mlx5_destroy_flow_table(tc->t);
+		tc->t = NULL;
+	}
+}
+
+int mlx5e_tc_esw_init(struct rhashtable *tc_ht)
+{
+	return rhashtable_init(tc_ht, &tc_ht_params);
+}
+
+void mlx5e_tc_esw_cleanup(struct rhashtable *tc_ht)
+{
+	rhashtable_free_and_destroy(tc_ht, _mlx5e_tc_del_flow, NULL);
+}
+
+int mlx5e_tc_num_filters(struct mlx5e_priv *priv)
+{
+	struct rhashtable *tc_ht = get_tc_ht(priv);
+
+	return atomic_read(&tc_ht->nelems);
+}
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_tc.h b/drivers/net/ethernet/mellanox/mlx5/core/en_tc.h
new file mode 100644
index 000000000..49436bf3b
--- /dev/null
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_tc.h
@@ -0,0 +1,79 @@
+/*
+ * Copyright (c) 2016, 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 __MLX5_EN_TC_H__
+#define __MLX5_EN_TC_H__
+
+#include <net/pkt_cls.h>
+
+#define MLX5E_TC_FLOW_ID_MASK 0x0000ffff
+
+#ifdef CONFIG_MLX5_ESWITCH
+
+enum {
+	MLX5E_TC_INGRESS = BIT(0),
+	MLX5E_TC_EGRESS  = BIT(1),
+	MLX5E_TC_LAST_EXPORTED_BIT = 1,
+};
+
+int mlx5e_tc_nic_init(struct mlx5e_priv *priv);
+void mlx5e_tc_nic_cleanup(struct mlx5e_priv *priv);
+
+int mlx5e_tc_esw_init(struct rhashtable *tc_ht);
+void mlx5e_tc_esw_cleanup(struct rhashtable *tc_ht);
+
+int mlx5e_configure_flower(struct mlx5e_priv *priv,
+			   struct tc_cls_flower_offload *f, int flags);
+int mlx5e_delete_flower(struct mlx5e_priv *priv,
+			struct tc_cls_flower_offload *f, int flags);
+
+int mlx5e_stats_flower(struct mlx5e_priv *priv,
+		       struct tc_cls_flower_offload *f, int flags);
+
+struct mlx5e_encap_entry;
+void mlx5e_tc_encap_flows_add(struct mlx5e_priv *priv,
+			      struct mlx5e_encap_entry *e);
+void mlx5e_tc_encap_flows_del(struct mlx5e_priv *priv,
+			      struct mlx5e_encap_entry *e);
+
+struct mlx5e_neigh_hash_entry;
+void mlx5e_tc_update_neigh_used_value(struct mlx5e_neigh_hash_entry *nhe);
+
+int mlx5e_tc_num_filters(struct mlx5e_priv *priv);
+
+#else /* CONFIG_MLX5_ESWITCH */
+static inline int  mlx5e_tc_nic_init(struct mlx5e_priv *priv) { return 0; }
+static inline void mlx5e_tc_nic_cleanup(struct mlx5e_priv *priv) {}
+static inline int  mlx5e_tc_num_filters(struct mlx5e_priv *priv) { return 0; }
+#endif
+
+#endif /* __MLX5_EN_TC_H__ */
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_tx.c b/drivers/net/ethernet/mellanox/mlx5/core/en_tx.c
new file mode 100644
index 000000000..52d3989bb
--- /dev/null
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_tx.c
@@ -0,0 +1,730 @@
+/*
+ * Copyright (c) 2015-2016, 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/tcp.h>
+#include <linux/if_vlan.h>
+#include <net/dsfield.h>
+#include "en.h"
+#include "ipoib/ipoib.h"
+#include "en_accel/en_accel.h"
+#include "lib/clock.h"
+
+#define MLX5E_SQ_NOPS_ROOM  MLX5_SEND_WQE_MAX_WQEBBS
+
+#ifndef CONFIG_MLX5_EN_TLS
+#define MLX5E_SQ_STOP_ROOM (MLX5_SEND_WQE_MAX_WQEBBS +\
+			    MLX5E_SQ_NOPS_ROOM)
+#else
+/* TLS offload requires MLX5E_SQ_STOP_ROOM to have
+ * enough room for a resync SKB, a normal SKB and a NOP
+ */
+#define MLX5E_SQ_STOP_ROOM (2 * MLX5_SEND_WQE_MAX_WQEBBS +\
+			    MLX5E_SQ_NOPS_ROOM)
+#endif
+
+static inline void mlx5e_tx_dma_unmap(struct device *pdev,
+				      struct mlx5e_sq_dma *dma)
+{
+	switch (dma->type) {
+	case MLX5E_DMA_MAP_SINGLE:
+		dma_unmap_single(pdev, dma->addr, dma->size, DMA_TO_DEVICE);
+		break;
+	case MLX5E_DMA_MAP_PAGE:
+		dma_unmap_page(pdev, dma->addr, dma->size, DMA_TO_DEVICE);
+		break;
+	default:
+		WARN_ONCE(true, "mlx5e_tx_dma_unmap unknown DMA type!\n");
+	}
+}
+
+static inline struct mlx5e_sq_dma *mlx5e_dma_get(struct mlx5e_txqsq *sq, u32 i)
+{
+	return &sq->db.dma_fifo[i & sq->dma_fifo_mask];
+}
+
+static inline void mlx5e_dma_push(struct mlx5e_txqsq *sq,
+				  dma_addr_t addr,
+				  u32 size,
+				  enum mlx5e_dma_map_type map_type)
+{
+	struct mlx5e_sq_dma *dma = mlx5e_dma_get(sq, sq->dma_fifo_pc++);
+
+	dma->addr = addr;
+	dma->size = size;
+	dma->type = map_type;
+}
+
+static void mlx5e_dma_unmap_wqe_err(struct mlx5e_txqsq *sq, u8 num_dma)
+{
+	int i;
+
+	for (i = 0; i < num_dma; i++) {
+		struct mlx5e_sq_dma *last_pushed_dma =
+			mlx5e_dma_get(sq, --sq->dma_fifo_pc);
+
+		mlx5e_tx_dma_unmap(sq->pdev, last_pushed_dma);
+	}
+}
+
+#ifdef CONFIG_MLX5_CORE_EN_DCB
+static inline int mlx5e_get_dscp_up(struct mlx5e_priv *priv, struct sk_buff *skb)
+{
+	int dscp_cp = 0;
+
+	if (skb->protocol == htons(ETH_P_IP))
+		dscp_cp = ipv4_get_dsfield(ip_hdr(skb)) >> 2;
+	else if (skb->protocol == htons(ETH_P_IPV6))
+		dscp_cp = ipv6_get_dsfield(ipv6_hdr(skb)) >> 2;
+
+	return priv->dcbx_dp.dscp2prio[dscp_cp];
+}
+#endif
+
+u16 mlx5e_select_queue(struct net_device *dev, struct sk_buff *skb,
+		       struct net_device *sb_dev,
+		       select_queue_fallback_t fallback)
+{
+	struct mlx5e_priv *priv = netdev_priv(dev);
+	int channel_ix = fallback(dev, skb, NULL);
+	u16 num_channels;
+	int up = 0;
+
+	if (!netdev_get_num_tc(dev))
+		return channel_ix;
+
+#ifdef CONFIG_MLX5_CORE_EN_DCB
+	if (priv->dcbx_dp.trust_state == MLX5_QPTS_TRUST_DSCP)
+		up = mlx5e_get_dscp_up(priv, skb);
+	else
+#endif
+		if (skb_vlan_tag_present(skb))
+			up = skb->vlan_tci >> VLAN_PRIO_SHIFT;
+
+	/* channel_ix can be larger than num_channels since
+	 * dev->num_real_tx_queues = num_channels * num_tc
+	 */
+	num_channels = priv->channels.params.num_channels;
+	if (channel_ix >= num_channels)
+		channel_ix = reciprocal_scale(channel_ix, num_channels);
+
+	return priv->channel_tc2txq[channel_ix][up];
+}
+
+static inline int mlx5e_skb_l2_header_offset(struct sk_buff *skb)
+{
+#define MLX5E_MIN_INLINE (ETH_HLEN + VLAN_HLEN)
+
+	return max(skb_network_offset(skb), MLX5E_MIN_INLINE);
+}
+
+static inline int mlx5e_skb_l3_header_offset(struct sk_buff *skb)
+{
+	struct flow_keys keys;
+
+	if (skb_transport_header_was_set(skb))
+		return skb_transport_offset(skb);
+	else if (skb_flow_dissect_flow_keys(skb, &keys, 0))
+		return keys.control.thoff;
+	else
+		return mlx5e_skb_l2_header_offset(skb);
+}
+
+static inline u16 mlx5e_calc_min_inline(enum mlx5_inline_modes mode,
+					struct sk_buff *skb)
+{
+	u16 hlen;
+
+	switch (mode) {
+	case MLX5_INLINE_MODE_NONE:
+		return 0;
+	case MLX5_INLINE_MODE_TCP_UDP:
+		hlen = eth_get_headlen(skb->data, skb_headlen(skb));
+		if (hlen == ETH_HLEN && !skb_vlan_tag_present(skb))
+			hlen += VLAN_HLEN;
+		break;
+	case MLX5_INLINE_MODE_IP:
+		/* When transport header is set to zero, it means no transport
+		 * header. When transport header is set to 0xff's, it means
+		 * transport header wasn't set.
+		 */
+		if (skb_transport_offset(skb)) {
+			hlen = mlx5e_skb_l3_header_offset(skb);
+			break;
+		}
+		/* fall through */
+	case MLX5_INLINE_MODE_L2:
+	default:
+		hlen = mlx5e_skb_l2_header_offset(skb);
+	}
+	return min_t(u16, hlen, skb_headlen(skb));
+}
+
+static inline void mlx5e_insert_vlan(void *start, struct sk_buff *skb, u16 ihs)
+{
+	struct vlan_ethhdr *vhdr = (struct vlan_ethhdr *)start;
+	int cpy1_sz = 2 * ETH_ALEN;
+	int cpy2_sz = ihs - cpy1_sz;
+
+	memcpy(vhdr, skb->data, cpy1_sz);
+	vhdr->h_vlan_proto = skb->vlan_proto;
+	vhdr->h_vlan_TCI = cpu_to_be16(skb_vlan_tag_get(skb));
+	memcpy(&vhdr->h_vlan_encapsulated_proto, skb->data + cpy1_sz, cpy2_sz);
+}
+
+static inline void
+mlx5e_txwqe_build_eseg_csum(struct mlx5e_txqsq *sq, struct sk_buff *skb, struct mlx5_wqe_eth_seg *eseg)
+{
+	if (likely(skb->ip_summed == CHECKSUM_PARTIAL)) {
+		eseg->cs_flags = MLX5_ETH_WQE_L3_CSUM;
+		if (skb->encapsulation) {
+			eseg->cs_flags |= MLX5_ETH_WQE_L3_INNER_CSUM |
+					  MLX5_ETH_WQE_L4_INNER_CSUM;
+			sq->stats->csum_partial_inner++;
+		} else {
+			eseg->cs_flags |= MLX5_ETH_WQE_L4_CSUM;
+			sq->stats->csum_partial++;
+		}
+	} else
+		sq->stats->csum_none++;
+}
+
+static inline u16
+mlx5e_tx_get_gso_ihs(struct mlx5e_txqsq *sq, struct sk_buff *skb)
+{
+	struct mlx5e_sq_stats *stats = sq->stats;
+	u16 ihs;
+
+	if (skb->encapsulation) {
+		ihs = skb_inner_transport_offset(skb) + inner_tcp_hdrlen(skb);
+		stats->tso_inner_packets++;
+		stats->tso_inner_bytes += skb->len - ihs;
+	} else {
+		if (skb_shinfo(skb)->gso_type & SKB_GSO_UDP_L4)
+			ihs = skb_transport_offset(skb) + sizeof(struct udphdr);
+		else
+			ihs = skb_transport_offset(skb) + tcp_hdrlen(skb);
+		stats->tso_packets++;
+		stats->tso_bytes += skb->len - ihs;
+	}
+
+	return ihs;
+}
+
+static inline int
+mlx5e_txwqe_build_dsegs(struct mlx5e_txqsq *sq, struct sk_buff *skb,
+			unsigned char *skb_data, u16 headlen,
+			struct mlx5_wqe_data_seg *dseg)
+{
+	dma_addr_t dma_addr = 0;
+	u8 num_dma          = 0;
+	int i;
+
+	if (headlen) {
+		dma_addr = dma_map_single(sq->pdev, skb_data, headlen,
+					  DMA_TO_DEVICE);
+		if (unlikely(dma_mapping_error(sq->pdev, dma_addr)))
+			goto dma_unmap_wqe_err;
+
+		dseg->addr       = cpu_to_be64(dma_addr);
+		dseg->lkey       = sq->mkey_be;
+		dseg->byte_count = cpu_to_be32(headlen);
+
+		mlx5e_dma_push(sq, dma_addr, headlen, MLX5E_DMA_MAP_SINGLE);
+		num_dma++;
+		dseg++;
+	}
+
+	for (i = 0; i < skb_shinfo(skb)->nr_frags; i++) {
+		struct skb_frag_struct *frag = &skb_shinfo(skb)->frags[i];
+		int fsz = skb_frag_size(frag);
+
+		dma_addr = skb_frag_dma_map(sq->pdev, frag, 0, fsz,
+					    DMA_TO_DEVICE);
+		if (unlikely(dma_mapping_error(sq->pdev, dma_addr)))
+			goto dma_unmap_wqe_err;
+
+		dseg->addr       = cpu_to_be64(dma_addr);
+		dseg->lkey       = sq->mkey_be;
+		dseg->byte_count = cpu_to_be32(fsz);
+
+		mlx5e_dma_push(sq, dma_addr, fsz, MLX5E_DMA_MAP_PAGE);
+		num_dma++;
+		dseg++;
+	}
+
+	return num_dma;
+
+dma_unmap_wqe_err:
+	mlx5e_dma_unmap_wqe_err(sq, num_dma);
+	return -ENOMEM;
+}
+
+static inline void mlx5e_fill_sq_frag_edge(struct mlx5e_txqsq *sq,
+					   struct mlx5_wq_cyc *wq,
+					   u16 pi, u16 nnops)
+{
+	struct mlx5e_tx_wqe_info *edge_wi, *wi = &sq->db.wqe_info[pi];
+
+	edge_wi = wi + nnops;
+
+	/* fill sq frag edge with nops to avoid wqe wrapping two pages */
+	for (; wi < edge_wi; wi++) {
+		wi->skb        = NULL;
+		wi->num_wqebbs = 1;
+		mlx5e_post_nop(wq, sq->sqn, &sq->pc);
+	}
+	sq->stats->nop += nnops;
+}
+
+static inline void
+mlx5e_txwqe_complete(struct mlx5e_txqsq *sq, struct sk_buff *skb,
+		     u8 opcode, u16 ds_cnt, u8 num_wqebbs, u32 num_bytes, u8 num_dma,
+		     struct mlx5e_tx_wqe_info *wi, struct mlx5_wqe_ctrl_seg *cseg)
+{
+	struct mlx5_wq_cyc *wq = &sq->wq;
+
+	wi->num_bytes = num_bytes;
+	wi->num_dma = num_dma;
+	wi->num_wqebbs = num_wqebbs;
+	wi->skb = skb;
+
+	cseg->opmod_idx_opcode = cpu_to_be32((sq->pc << 8) | opcode);
+	cseg->qpn_ds           = cpu_to_be32((sq->sqn << 8) | ds_cnt);
+
+	netdev_tx_sent_queue(sq->txq, num_bytes);
+
+	if (unlikely(skb_shinfo(skb)->tx_flags & SKBTX_HW_TSTAMP))
+		skb_shinfo(skb)->tx_flags |= SKBTX_IN_PROGRESS;
+
+	sq->pc += wi->num_wqebbs;
+	if (unlikely(!mlx5e_wqc_has_room_for(wq, sq->cc, sq->pc, MLX5E_SQ_STOP_ROOM))) {
+		netif_tx_stop_queue(sq->txq);
+		sq->stats->stopped++;
+	}
+
+	if (!skb->xmit_more || netif_xmit_stopped(sq->txq))
+		mlx5e_notify_hw(wq, sq->pc, sq->uar_map, cseg);
+}
+
+#define INL_HDR_START_SZ (sizeof(((struct mlx5_wqe_eth_seg *)NULL)->inline_hdr.start))
+
+netdev_tx_t mlx5e_sq_xmit(struct mlx5e_txqsq *sq, struct sk_buff *skb,
+			  struct mlx5e_tx_wqe *wqe, u16 pi)
+{
+	struct mlx5_wq_cyc *wq = &sq->wq;
+	struct mlx5_wqe_ctrl_seg *cseg;
+	struct mlx5_wqe_eth_seg  *eseg;
+	struct mlx5_wqe_data_seg *dseg;
+	struct mlx5e_tx_wqe_info *wi;
+
+	struct mlx5e_sq_stats *stats = sq->stats;
+	u16 headlen, ihs, contig_wqebbs_room;
+	u16 ds_cnt, ds_cnt_inl = 0;
+	u8 num_wqebbs, opcode;
+	u32 num_bytes;
+	int num_dma;
+	__be16 mss;
+
+	/* Calc ihs and ds cnt, no writes to wqe yet */
+	ds_cnt = sizeof(*wqe) / MLX5_SEND_WQE_DS;
+	if (skb_is_gso(skb)) {
+		opcode    = MLX5_OPCODE_LSO;
+		mss       = cpu_to_be16(skb_shinfo(skb)->gso_size);
+		ihs       = mlx5e_tx_get_gso_ihs(sq, skb);
+		num_bytes = skb->len + (skb_shinfo(skb)->gso_segs - 1) * ihs;
+		stats->packets += skb_shinfo(skb)->gso_segs;
+	} else {
+		opcode    = MLX5_OPCODE_SEND;
+		mss       = 0;
+		ihs       = mlx5e_calc_min_inline(sq->min_inline_mode, skb);
+		num_bytes = max_t(unsigned int, skb->len, ETH_ZLEN);
+		stats->packets++;
+	}
+
+	stats->bytes     += num_bytes;
+	stats->xmit_more += skb->xmit_more;
+
+	headlen = skb->len - ihs - skb->data_len;
+	ds_cnt += !!headlen;
+	ds_cnt += skb_shinfo(skb)->nr_frags;
+
+	if (ihs) {
+		ihs += !!skb_vlan_tag_present(skb) * VLAN_HLEN;
+
+		ds_cnt_inl = DIV_ROUND_UP(ihs - INL_HDR_START_SZ, MLX5_SEND_WQE_DS);
+		ds_cnt += ds_cnt_inl;
+	}
+
+	num_wqebbs = DIV_ROUND_UP(ds_cnt, MLX5_SEND_WQEBB_NUM_DS);
+	contig_wqebbs_room = mlx5_wq_cyc_get_contig_wqebbs(wq, pi);
+	if (unlikely(contig_wqebbs_room < num_wqebbs)) {
+#ifdef CONFIG_MLX5_EN_IPSEC
+		struct mlx5_wqe_eth_seg cur_eth = wqe->eth;
+#endif
+		mlx5e_fill_sq_frag_edge(sq, wq, pi, contig_wqebbs_room);
+		mlx5e_sq_fetch_wqe(sq, &wqe, &pi);
+#ifdef CONFIG_MLX5_EN_IPSEC
+		wqe->eth = cur_eth;
+#endif
+	}
+
+	/* fill wqe */
+	wi   = &sq->db.wqe_info[pi];
+	cseg = &wqe->ctrl;
+	eseg = &wqe->eth;
+	dseg =  wqe->data;
+
+	mlx5e_txwqe_build_eseg_csum(sq, skb, eseg);
+
+	eseg->mss = mss;
+
+	if (ihs) {
+		eseg->inline_hdr.sz = cpu_to_be16(ihs);
+		if (skb_vlan_tag_present(skb)) {
+			ihs -= VLAN_HLEN;
+			mlx5e_insert_vlan(eseg->inline_hdr.start, skb, ihs);
+			stats->added_vlan_packets++;
+		} else {
+			memcpy(eseg->inline_hdr.start, skb->data, ihs);
+		}
+		dseg += ds_cnt_inl;
+	} else if (skb_vlan_tag_present(skb)) {
+		eseg->insert.type = cpu_to_be16(MLX5_ETH_WQE_INSERT_VLAN);
+		if (skb->vlan_proto == cpu_to_be16(ETH_P_8021AD))
+			eseg->insert.type |= cpu_to_be16(MLX5_ETH_WQE_SVLAN);
+		eseg->insert.vlan_tci = cpu_to_be16(skb_vlan_tag_get(skb));
+		stats->added_vlan_packets++;
+	}
+
+	num_dma = mlx5e_txwqe_build_dsegs(sq, skb, skb->data + ihs, headlen, dseg);
+	if (unlikely(num_dma < 0))
+		goto err_drop;
+
+	mlx5e_txwqe_complete(sq, skb, opcode, ds_cnt, num_wqebbs, num_bytes,
+			     num_dma, wi, cseg);
+
+	return NETDEV_TX_OK;
+
+err_drop:
+	stats->dropped++;
+	dev_kfree_skb_any(skb);
+
+	return NETDEV_TX_OK;
+}
+
+netdev_tx_t mlx5e_xmit(struct sk_buff *skb, struct net_device *dev)
+{
+	struct mlx5e_priv *priv = netdev_priv(dev);
+	struct mlx5e_tx_wqe *wqe;
+	struct mlx5e_txqsq *sq;
+	u16 pi;
+
+	sq = priv->txq2sq[skb_get_queue_mapping(skb)];
+	mlx5e_sq_fetch_wqe(sq, &wqe, &pi);
+
+	/* might send skbs and update wqe and pi */
+	skb = mlx5e_accel_handle_tx(skb, sq, dev, &wqe, &pi);
+	if (unlikely(!skb))
+		return NETDEV_TX_OK;
+
+	return mlx5e_sq_xmit(sq, skb, wqe, pi);
+}
+
+static void mlx5e_dump_error_cqe(struct mlx5e_txqsq *sq,
+				 struct mlx5_err_cqe *err_cqe)
+{
+	struct mlx5_cqwq *wq = &sq->cq.wq;
+	u32 ci;
+
+	ci = mlx5_cqwq_ctr2ix(wq, wq->cc - 1);
+
+	netdev_err(sq->channel->netdev,
+		   "Error cqe on cqn 0x%x, ci 0x%x, sqn 0x%x, syndrome 0x%x, vendor syndrome 0x%x\n",
+		   sq->cq.mcq.cqn, ci, sq->sqn, err_cqe->syndrome,
+		   err_cqe->vendor_err_synd);
+	mlx5_dump_err_cqe(sq->cq.mdev, err_cqe);
+}
+
+bool mlx5e_poll_tx_cq(struct mlx5e_cq *cq, int napi_budget)
+{
+	struct mlx5e_sq_stats *stats;
+	struct mlx5e_txqsq *sq;
+	struct mlx5_cqe64 *cqe;
+	u32 dma_fifo_cc;
+	u32 nbytes;
+	u16 npkts;
+	u16 sqcc;
+	int i;
+
+	sq = container_of(cq, struct mlx5e_txqsq, cq);
+
+	if (unlikely(!test_bit(MLX5E_SQ_STATE_ENABLED, &sq->state)))
+		return false;
+
+	cqe = mlx5_cqwq_get_cqe(&cq->wq);
+	if (!cqe)
+		return false;
+
+	stats = sq->stats;
+
+	npkts = 0;
+	nbytes = 0;
+
+	/* sq->cc must be updated only after mlx5_cqwq_update_db_record(),
+	 * otherwise a cq overrun may occur
+	 */
+	sqcc = sq->cc;
+
+	/* avoid dirtying sq cache line every cqe */
+	dma_fifo_cc = sq->dma_fifo_cc;
+
+	i = 0;
+	do {
+		u16 wqe_counter;
+		bool last_wqe;
+
+		mlx5_cqwq_pop(&cq->wq);
+
+		wqe_counter = be16_to_cpu(cqe->wqe_counter);
+
+		if (unlikely(cqe->op_own >> 4 == MLX5_CQE_REQ_ERR)) {
+			if (!test_and_set_bit(MLX5E_SQ_STATE_RECOVERING,
+					      &sq->state)) {
+				mlx5e_dump_error_cqe(sq,
+						     (struct mlx5_err_cqe *)cqe);
+				queue_work(cq->channel->priv->wq,
+					   &sq->recover.recover_work);
+			}
+			stats->cqe_err++;
+		}
+
+		do {
+			struct mlx5e_tx_wqe_info *wi;
+			struct sk_buff *skb;
+			u16 ci;
+			int j;
+
+			last_wqe = (sqcc == wqe_counter);
+
+			ci = mlx5_wq_cyc_ctr2ix(&sq->wq, sqcc);
+			wi = &sq->db.wqe_info[ci];
+			skb = wi->skb;
+
+			if (unlikely(!skb)) { /* nop */
+				sqcc++;
+				continue;
+			}
+
+			if (unlikely(skb_shinfo(skb)->tx_flags &
+				     SKBTX_HW_TSTAMP)) {
+				struct skb_shared_hwtstamps hwts = {};
+
+				hwts.hwtstamp =
+					mlx5_timecounter_cyc2time(sq->clock,
+								  get_cqe_ts(cqe));
+				skb_tstamp_tx(skb, &hwts);
+			}
+
+			for (j = 0; j < wi->num_dma; j++) {
+				struct mlx5e_sq_dma *dma =
+					mlx5e_dma_get(sq, dma_fifo_cc++);
+
+				mlx5e_tx_dma_unmap(sq->pdev, dma);
+			}
+
+			npkts++;
+			nbytes += wi->num_bytes;
+			sqcc += wi->num_wqebbs;
+			napi_consume_skb(skb, napi_budget);
+		} while (!last_wqe);
+
+	} while ((++i < MLX5E_TX_CQ_POLL_BUDGET) && (cqe = mlx5_cqwq_get_cqe(&cq->wq)));
+
+	stats->cqes += i;
+
+	mlx5_cqwq_update_db_record(&cq->wq);
+
+	/* ensure cq space is freed before enabling more cqes */
+	wmb();
+
+	sq->dma_fifo_cc = dma_fifo_cc;
+	sq->cc = sqcc;
+
+	netdev_tx_completed_queue(sq->txq, npkts, nbytes);
+
+	if (netif_tx_queue_stopped(sq->txq) &&
+	    mlx5e_wqc_has_room_for(&sq->wq, sq->cc, sq->pc,
+				   MLX5E_SQ_STOP_ROOM) &&
+	    !test_bit(MLX5E_SQ_STATE_RECOVERING, &sq->state)) {
+		netif_tx_wake_queue(sq->txq);
+		stats->wake++;
+	}
+
+	return (i == MLX5E_TX_CQ_POLL_BUDGET);
+}
+
+void mlx5e_free_txqsq_descs(struct mlx5e_txqsq *sq)
+{
+	struct mlx5e_tx_wqe_info *wi;
+	u32 nbytes = 0;
+	u16 ci, npkts = 0;
+	struct sk_buff *skb;
+	int i;
+
+	while (sq->cc != sq->pc) {
+		ci = mlx5_wq_cyc_ctr2ix(&sq->wq, sq->cc);
+		wi = &sq->db.wqe_info[ci];
+		skb = wi->skb;
+
+		if (!skb) { /* nop */
+			sq->cc++;
+			continue;
+		}
+
+		for (i = 0; i < wi->num_dma; i++) {
+			struct mlx5e_sq_dma *dma =
+				mlx5e_dma_get(sq, sq->dma_fifo_cc++);
+
+			mlx5e_tx_dma_unmap(sq->pdev, dma);
+		}
+
+		dev_kfree_skb_any(skb);
+		npkts++;
+		nbytes += wi->num_bytes;
+		sq->cc += wi->num_wqebbs;
+	}
+	netdev_tx_completed_queue(sq->txq, npkts, nbytes);
+}
+
+#ifdef CONFIG_MLX5_CORE_IPOIB
+static inline void
+mlx5i_txwqe_build_datagram(struct mlx5_av *av, u32 dqpn, u32 dqkey,
+			   struct mlx5_wqe_datagram_seg *dseg)
+{
+	memcpy(&dseg->av, av, sizeof(struct mlx5_av));
+	dseg->av.dqp_dct = cpu_to_be32(dqpn | MLX5_EXTENDED_UD_AV);
+	dseg->av.key.qkey.qkey = cpu_to_be32(dqkey);
+}
+
+netdev_tx_t mlx5i_sq_xmit(struct mlx5e_txqsq *sq, struct sk_buff *skb,
+			  struct mlx5_av *av, u32 dqpn, u32 dqkey)
+{
+	struct mlx5_wq_cyc *wq = &sq->wq;
+	struct mlx5i_tx_wqe *wqe;
+
+	struct mlx5_wqe_datagram_seg *datagram;
+	struct mlx5_wqe_ctrl_seg *cseg;
+	struct mlx5_wqe_eth_seg  *eseg;
+	struct mlx5_wqe_data_seg *dseg;
+	struct mlx5e_tx_wqe_info *wi;
+
+	struct mlx5e_sq_stats *stats = sq->stats;
+	u16 headlen, ihs, pi, contig_wqebbs_room;
+	u16 ds_cnt, ds_cnt_inl = 0;
+	u8 num_wqebbs, opcode;
+	u32 num_bytes;
+	int num_dma;
+	__be16 mss;
+
+	/* Calc ihs and ds cnt, no writes to wqe yet */
+	ds_cnt = sizeof(*wqe) / MLX5_SEND_WQE_DS;
+	if (skb_is_gso(skb)) {
+		opcode    = MLX5_OPCODE_LSO;
+		mss       = cpu_to_be16(skb_shinfo(skb)->gso_size);
+		ihs       = mlx5e_tx_get_gso_ihs(sq, skb);
+		num_bytes = skb->len + (skb_shinfo(skb)->gso_segs - 1) * ihs;
+		stats->packets += skb_shinfo(skb)->gso_segs;
+	} else {
+		opcode    = MLX5_OPCODE_SEND;
+		mss       = 0;
+		ihs       = mlx5e_calc_min_inline(sq->min_inline_mode, skb);
+		num_bytes = max_t(unsigned int, skb->len, ETH_ZLEN);
+		stats->packets++;
+	}
+
+	stats->bytes     += num_bytes;
+	stats->xmit_more += skb->xmit_more;
+
+	headlen = skb->len - ihs - skb->data_len;
+	ds_cnt += !!headlen;
+	ds_cnt += skb_shinfo(skb)->nr_frags;
+
+	if (ihs) {
+		ds_cnt_inl = DIV_ROUND_UP(ihs - INL_HDR_START_SZ, MLX5_SEND_WQE_DS);
+		ds_cnt += ds_cnt_inl;
+	}
+
+	num_wqebbs = DIV_ROUND_UP(ds_cnt, MLX5_SEND_WQEBB_NUM_DS);
+	pi = mlx5_wq_cyc_ctr2ix(wq, sq->pc);
+	contig_wqebbs_room = mlx5_wq_cyc_get_contig_wqebbs(wq, pi);
+	if (unlikely(contig_wqebbs_room < num_wqebbs)) {
+		mlx5e_fill_sq_frag_edge(sq, wq, pi, contig_wqebbs_room);
+		pi = mlx5_wq_cyc_ctr2ix(wq, sq->pc);
+	}
+
+	mlx5i_sq_fetch_wqe(sq, &wqe, pi);
+
+	/* fill wqe */
+	wi       = &sq->db.wqe_info[pi];
+	cseg     = &wqe->ctrl;
+	datagram = &wqe->datagram;
+	eseg     = &wqe->eth;
+	dseg     =  wqe->data;
+
+	mlx5i_txwqe_build_datagram(av, dqpn, dqkey, datagram);
+
+	mlx5e_txwqe_build_eseg_csum(sq, skb, eseg);
+
+	eseg->mss = mss;
+
+	if (ihs) {
+		memcpy(eseg->inline_hdr.start, skb->data, ihs);
+		eseg->inline_hdr.sz = cpu_to_be16(ihs);
+		dseg += ds_cnt_inl;
+	}
+
+	num_dma = mlx5e_txwqe_build_dsegs(sq, skb, skb->data + ihs, headlen, dseg);
+	if (unlikely(num_dma < 0))
+		goto err_drop;
+
+	mlx5e_txwqe_complete(sq, skb, opcode, ds_cnt, num_wqebbs, num_bytes,
+			     num_dma, wi, cseg);
+
+	return NETDEV_TX_OK;
+
+err_drop:
+	stats->dropped++;
+	dev_kfree_skb_any(skb);
+
+	return NETDEV_TX_OK;
+}
+#endif
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_txrx.c b/drivers/net/ethernet/mellanox/mlx5/core/en_txrx.c
new file mode 100644
index 000000000..85d517360
--- /dev/null
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_txrx.c
@@ -0,0 +1,144 @@
+/*
+ * Copyright (c) 2015, 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/irq.h>
+#include "en.h"
+#include "en/xdp.h"
+
+static inline bool mlx5e_channel_no_affinity_change(struct mlx5e_channel *c)
+{
+	int current_cpu = smp_processor_id();
+	const struct cpumask *aff;
+	struct irq_data *idata;
+
+	idata = irq_desc_get_irq_data(c->irq_desc);
+	aff = irq_data_get_affinity_mask(idata);
+	return cpumask_test_cpu(current_cpu, aff);
+}
+
+static void mlx5e_handle_tx_dim(struct mlx5e_txqsq *sq)
+{
+	struct mlx5e_sq_stats *stats = sq->stats;
+	struct net_dim_sample dim_sample;
+
+	if (unlikely(!test_bit(MLX5E_SQ_STATE_AM, &sq->state)))
+		return;
+
+	net_dim_sample(sq->cq.event_ctr, stats->packets, stats->bytes,
+		       &dim_sample);
+	net_dim(&sq->dim, dim_sample);
+}
+
+static void mlx5e_handle_rx_dim(struct mlx5e_rq *rq)
+{
+	struct mlx5e_rq_stats *stats = rq->stats;
+	struct net_dim_sample dim_sample;
+
+	if (unlikely(!test_bit(MLX5E_RQ_STATE_AM, &rq->state)))
+		return;
+
+	net_dim_sample(rq->cq.event_ctr, stats->packets, stats->bytes,
+		       &dim_sample);
+	net_dim(&rq->dim, dim_sample);
+}
+
+int mlx5e_napi_poll(struct napi_struct *napi, int budget)
+{
+	struct mlx5e_channel *c = container_of(napi, struct mlx5e_channel,
+					       napi);
+	struct mlx5e_ch_stats *ch_stats = c->stats;
+	bool busy = false;
+	int work_done = 0;
+	int i;
+
+	ch_stats->poll++;
+
+	for (i = 0; i < c->num_tc; i++)
+		busy |= mlx5e_poll_tx_cq(&c->sq[i].cq, budget);
+
+	busy |= mlx5e_poll_xdpsq_cq(&c->xdpsq.cq);
+
+	if (c->xdp)
+		busy |= mlx5e_poll_xdpsq_cq(&c->rq.xdpsq.cq);
+
+	if (likely(budget)) { /* budget=0 means: don't poll rx rings */
+		work_done = mlx5e_poll_rx_cq(&c->rq.cq, budget);
+		busy |= work_done == budget;
+	}
+
+	busy |= c->rq.post_wqes(&c->rq);
+
+	if (busy) {
+		if (likely(mlx5e_channel_no_affinity_change(c)))
+			return budget;
+		ch_stats->aff_change++;
+		if (budget && work_done == budget)
+			work_done--;
+	}
+
+	if (unlikely(!napi_complete_done(napi, work_done)))
+		return work_done;
+
+	ch_stats->arm++;
+
+	for (i = 0; i < c->num_tc; i++) {
+		mlx5e_handle_tx_dim(&c->sq[i]);
+		mlx5e_cq_arm(&c->sq[i].cq);
+	}
+
+	mlx5e_handle_rx_dim(&c->rq);
+
+	mlx5e_cq_arm(&c->rq.cq);
+	mlx5e_cq_arm(&c->icosq.cq);
+	mlx5e_cq_arm(&c->xdpsq.cq);
+
+	return work_done;
+}
+
+void mlx5e_completion_event(struct mlx5_core_cq *mcq)
+{
+	struct mlx5e_cq *cq = container_of(mcq, struct mlx5e_cq, mcq);
+
+	napi_schedule(cq->napi);
+	cq->event_ctr++;
+	cq->channel->stats->events++;
+}
+
+void mlx5e_cq_error_event(struct mlx5_core_cq *mcq, enum mlx5_event event)
+{
+	struct mlx5e_cq *cq = container_of(mcq, struct mlx5e_cq, mcq);
+	struct mlx5e_channel *c = cq->channel;
+	struct net_device *netdev = c->netdev;
+
+	netdev_err(netdev, "%s: cqn=0x%.6x event=0x%.2x\n",
+		   __func__, mcq->cqn, event);
+}
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/eq.c b/drivers/net/ethernet/mellanox/mlx5/core/eq.c
new file mode 100644
index 000000000..c1e1a16a9
--- /dev/null
+++ b/drivers/net/ethernet/mellanox/mlx5/core/eq.c
@@ -0,0 +1,991 @@
+/*
+ * Copyright (c) 2013-2015, 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/interrupt.h>
+#include <linux/module.h>
+#include <linux/mlx5/driver.h>
+#include <linux/mlx5/cmd.h>
+#ifdef CONFIG_RFS_ACCEL
+#include <linux/cpu_rmap.h>
+#endif
+#include "mlx5_core.h"
+#include "fpga/core.h"
+#include "eswitch.h"
+#include "lib/clock.h"
+#include "diag/fw_tracer.h"
+
+enum {
+	MLX5_EQE_SIZE		= sizeof(struct mlx5_eqe),
+	MLX5_EQE_OWNER_INIT_VAL	= 0x1,
+};
+
+enum {
+	MLX5_EQ_STATE_ARMED		= 0x9,
+	MLX5_EQ_STATE_FIRED		= 0xa,
+	MLX5_EQ_STATE_ALWAYS_ARMED	= 0xb,
+};
+
+enum {
+	MLX5_NUM_SPARE_EQE	= 0x80,
+	MLX5_NUM_ASYNC_EQE	= 0x1000,
+	MLX5_NUM_CMD_EQE	= 32,
+	MLX5_NUM_PF_DRAIN	= 64,
+};
+
+enum {
+	MLX5_EQ_DOORBEL_OFFSET	= 0x40,
+};
+
+#define MLX5_ASYNC_EVENT_MASK ((1ull << MLX5_EVENT_TYPE_PATH_MIG)	    | \
+			       (1ull << MLX5_EVENT_TYPE_COMM_EST)	    | \
+			       (1ull << MLX5_EVENT_TYPE_SQ_DRAINED)	    | \
+			       (1ull << MLX5_EVENT_TYPE_CQ_ERROR)	    | \
+			       (1ull << MLX5_EVENT_TYPE_WQ_CATAS_ERROR)	    | \
+			       (1ull << MLX5_EVENT_TYPE_PATH_MIG_FAILED)    | \
+			       (1ull << MLX5_EVENT_TYPE_WQ_INVAL_REQ_ERROR) | \
+			       (1ull << MLX5_EVENT_TYPE_WQ_ACCESS_ERROR)    | \
+			       (1ull << MLX5_EVENT_TYPE_PORT_CHANGE)	    | \
+			       (1ull << MLX5_EVENT_TYPE_SRQ_CATAS_ERROR)    | \
+			       (1ull << MLX5_EVENT_TYPE_SRQ_LAST_WQE)	    | \
+			       (1ull << MLX5_EVENT_TYPE_SRQ_RQ_LIMIT))
+
+struct map_eq_in {
+	u64	mask;
+	u32	reserved;
+	u32	unmap_eqn;
+};
+
+struct cre_des_eq {
+	u8	reserved[15];
+	u8	eqn;
+};
+
+static int mlx5_cmd_destroy_eq(struct mlx5_core_dev *dev, u8 eqn)
+{
+	u32 out[MLX5_ST_SZ_DW(destroy_eq_out)] = {0};
+	u32 in[MLX5_ST_SZ_DW(destroy_eq_in)]   = {0};
+
+	MLX5_SET(destroy_eq_in, in, opcode, MLX5_CMD_OP_DESTROY_EQ);
+	MLX5_SET(destroy_eq_in, in, eq_number, eqn);
+	return mlx5_cmd_exec(dev, in, sizeof(in), out, sizeof(out));
+}
+
+static struct mlx5_eqe *get_eqe(struct mlx5_eq *eq, u32 entry)
+{
+	return mlx5_buf_offset(&eq->buf, entry * MLX5_EQE_SIZE);
+}
+
+static struct mlx5_eqe *next_eqe_sw(struct mlx5_eq *eq)
+{
+	struct mlx5_eqe *eqe = get_eqe(eq, eq->cons_index & (eq->nent - 1));
+
+	return ((eqe->owner & 1) ^ !!(eq->cons_index & eq->nent)) ? NULL : eqe;
+}
+
+static const char *eqe_type_str(u8 type)
+{
+	switch (type) {
+	case MLX5_EVENT_TYPE_COMP:
+		return "MLX5_EVENT_TYPE_COMP";
+	case MLX5_EVENT_TYPE_PATH_MIG:
+		return "MLX5_EVENT_TYPE_PATH_MIG";
+	case MLX5_EVENT_TYPE_COMM_EST:
+		return "MLX5_EVENT_TYPE_COMM_EST";
+	case MLX5_EVENT_TYPE_SQ_DRAINED:
+		return "MLX5_EVENT_TYPE_SQ_DRAINED";
+	case MLX5_EVENT_TYPE_SRQ_LAST_WQE:
+		return "MLX5_EVENT_TYPE_SRQ_LAST_WQE";
+	case MLX5_EVENT_TYPE_SRQ_RQ_LIMIT:
+		return "MLX5_EVENT_TYPE_SRQ_RQ_LIMIT";
+	case MLX5_EVENT_TYPE_CQ_ERROR:
+		return "MLX5_EVENT_TYPE_CQ_ERROR";
+	case MLX5_EVENT_TYPE_WQ_CATAS_ERROR:
+		return "MLX5_EVENT_TYPE_WQ_CATAS_ERROR";
+	case MLX5_EVENT_TYPE_PATH_MIG_FAILED:
+		return "MLX5_EVENT_TYPE_PATH_MIG_FAILED";
+	case MLX5_EVENT_TYPE_WQ_INVAL_REQ_ERROR:
+		return "MLX5_EVENT_TYPE_WQ_INVAL_REQ_ERROR";
+	case MLX5_EVENT_TYPE_WQ_ACCESS_ERROR:
+		return "MLX5_EVENT_TYPE_WQ_ACCESS_ERROR";
+	case MLX5_EVENT_TYPE_SRQ_CATAS_ERROR:
+		return "MLX5_EVENT_TYPE_SRQ_CATAS_ERROR";
+	case MLX5_EVENT_TYPE_INTERNAL_ERROR:
+		return "MLX5_EVENT_TYPE_INTERNAL_ERROR";
+	case MLX5_EVENT_TYPE_PORT_CHANGE:
+		return "MLX5_EVENT_TYPE_PORT_CHANGE";
+	case MLX5_EVENT_TYPE_GPIO_EVENT:
+		return "MLX5_EVENT_TYPE_GPIO_EVENT";
+	case MLX5_EVENT_TYPE_PORT_MODULE_EVENT:
+		return "MLX5_EVENT_TYPE_PORT_MODULE_EVENT";
+	case MLX5_EVENT_TYPE_TEMP_WARN_EVENT:
+		return "MLX5_EVENT_TYPE_TEMP_WARN_EVENT";
+	case MLX5_EVENT_TYPE_REMOTE_CONFIG:
+		return "MLX5_EVENT_TYPE_REMOTE_CONFIG";
+	case MLX5_EVENT_TYPE_DB_BF_CONGESTION:
+		return "MLX5_EVENT_TYPE_DB_BF_CONGESTION";
+	case MLX5_EVENT_TYPE_STALL_EVENT:
+		return "MLX5_EVENT_TYPE_STALL_EVENT";
+	case MLX5_EVENT_TYPE_CMD:
+		return "MLX5_EVENT_TYPE_CMD";
+	case MLX5_EVENT_TYPE_PAGE_REQUEST:
+		return "MLX5_EVENT_TYPE_PAGE_REQUEST";
+	case MLX5_EVENT_TYPE_PAGE_FAULT:
+		return "MLX5_EVENT_TYPE_PAGE_FAULT";
+	case MLX5_EVENT_TYPE_PPS_EVENT:
+		return "MLX5_EVENT_TYPE_PPS_EVENT";
+	case MLX5_EVENT_TYPE_NIC_VPORT_CHANGE:
+		return "MLX5_EVENT_TYPE_NIC_VPORT_CHANGE";
+	case MLX5_EVENT_TYPE_FPGA_ERROR:
+		return "MLX5_EVENT_TYPE_FPGA_ERROR";
+	case MLX5_EVENT_TYPE_FPGA_QP_ERROR:
+		return "MLX5_EVENT_TYPE_FPGA_QP_ERROR";
+	case MLX5_EVENT_TYPE_GENERAL_EVENT:
+		return "MLX5_EVENT_TYPE_GENERAL_EVENT";
+	case MLX5_EVENT_TYPE_DEVICE_TRACER:
+		return "MLX5_EVENT_TYPE_DEVICE_TRACER";
+	default:
+		return "Unrecognized event";
+	}
+}
+
+static enum mlx5_dev_event port_subtype_event(u8 subtype)
+{
+	switch (subtype) {
+	case MLX5_PORT_CHANGE_SUBTYPE_DOWN:
+		return MLX5_DEV_EVENT_PORT_DOWN;
+	case MLX5_PORT_CHANGE_SUBTYPE_ACTIVE:
+		return MLX5_DEV_EVENT_PORT_UP;
+	case MLX5_PORT_CHANGE_SUBTYPE_INITIALIZED:
+		return MLX5_DEV_EVENT_PORT_INITIALIZED;
+	case MLX5_PORT_CHANGE_SUBTYPE_LID:
+		return MLX5_DEV_EVENT_LID_CHANGE;
+	case MLX5_PORT_CHANGE_SUBTYPE_PKEY:
+		return MLX5_DEV_EVENT_PKEY_CHANGE;
+	case MLX5_PORT_CHANGE_SUBTYPE_GUID:
+		return MLX5_DEV_EVENT_GUID_CHANGE;
+	case MLX5_PORT_CHANGE_SUBTYPE_CLIENT_REREG:
+		return MLX5_DEV_EVENT_CLIENT_REREG;
+	}
+	return -1;
+}
+
+static void eq_update_ci(struct mlx5_eq *eq, int arm)
+{
+	__be32 __iomem *addr = eq->doorbell + (arm ? 0 : 2);
+	u32 val = (eq->cons_index & 0xffffff) | (eq->eqn << 24);
+
+	__raw_writel((__force u32)cpu_to_be32(val), addr);
+	/* We still want ordering, just not swabbing, so add a barrier */
+	mb();
+}
+
+#ifdef CONFIG_INFINIBAND_ON_DEMAND_PAGING
+static void eqe_pf_action(struct work_struct *work)
+{
+	struct mlx5_pagefault *pfault = container_of(work,
+						     struct mlx5_pagefault,
+						     work);
+	struct mlx5_eq *eq = pfault->eq;
+
+	mlx5_core_page_fault(eq->dev, pfault);
+	mempool_free(pfault, eq->pf_ctx.pool);
+}
+
+static void eq_pf_process(struct mlx5_eq *eq)
+{
+	struct mlx5_core_dev *dev = eq->dev;
+	struct mlx5_eqe_page_fault *pf_eqe;
+	struct mlx5_pagefault *pfault;
+	struct mlx5_eqe *eqe;
+	int set_ci = 0;
+
+	while ((eqe = next_eqe_sw(eq))) {
+		pfault = mempool_alloc(eq->pf_ctx.pool, GFP_ATOMIC);
+		if (!pfault) {
+			schedule_work(&eq->pf_ctx.work);
+			break;
+		}
+
+		dma_rmb();
+		pf_eqe = &eqe->data.page_fault;
+		pfault->event_subtype = eqe->sub_type;
+		pfault->bytes_committed = be32_to_cpu(pf_eqe->bytes_committed);
+
+		mlx5_core_dbg(dev,
+			      "PAGE_FAULT: subtype: 0x%02x, bytes_committed: 0x%06x\n",
+			      eqe->sub_type, pfault->bytes_committed);
+
+		switch (eqe->sub_type) {
+		case MLX5_PFAULT_SUBTYPE_RDMA:
+			/* RDMA based event */
+			pfault->type =
+				be32_to_cpu(pf_eqe->rdma.pftype_token) >> 24;
+			pfault->token =
+				be32_to_cpu(pf_eqe->rdma.pftype_token) &
+				MLX5_24BIT_MASK;
+			pfault->rdma.r_key =
+				be32_to_cpu(pf_eqe->rdma.r_key);
+			pfault->rdma.packet_size =
+				be16_to_cpu(pf_eqe->rdma.packet_length);
+			pfault->rdma.rdma_op_len =
+				be32_to_cpu(pf_eqe->rdma.rdma_op_len);
+			pfault->rdma.rdma_va =
+				be64_to_cpu(pf_eqe->rdma.rdma_va);
+			mlx5_core_dbg(dev,
+				      "PAGE_FAULT: type:0x%x, token: 0x%06x, r_key: 0x%08x\n",
+				      pfault->type, pfault->token,
+				      pfault->rdma.r_key);
+			mlx5_core_dbg(dev,
+				      "PAGE_FAULT: rdma_op_len: 0x%08x, rdma_va: 0x%016llx\n",
+				      pfault->rdma.rdma_op_len,
+				      pfault->rdma.rdma_va);
+			break;
+
+		case MLX5_PFAULT_SUBTYPE_WQE:
+			/* WQE based event */
+			pfault->type =
+				(be32_to_cpu(pf_eqe->wqe.pftype_wq) >> 24) & 0x7;
+			pfault->token =
+				be32_to_cpu(pf_eqe->wqe.token);
+			pfault->wqe.wq_num =
+				be32_to_cpu(pf_eqe->wqe.pftype_wq) &
+				MLX5_24BIT_MASK;
+			pfault->wqe.wqe_index =
+				be16_to_cpu(pf_eqe->wqe.wqe_index);
+			pfault->wqe.packet_size =
+				be16_to_cpu(pf_eqe->wqe.packet_length);
+			mlx5_core_dbg(dev,
+				      "PAGE_FAULT: type:0x%x, token: 0x%06x, wq_num: 0x%06x, wqe_index: 0x%04x\n",
+				      pfault->type, pfault->token,
+				      pfault->wqe.wq_num,
+				      pfault->wqe.wqe_index);
+			break;
+
+		default:
+			mlx5_core_warn(dev,
+				       "Unsupported page fault event sub-type: 0x%02hhx\n",
+				       eqe->sub_type);
+			/* Unsupported page faults should still be
+			 * resolved by the page fault handler
+			 */
+		}
+
+		pfault->eq = eq;
+		INIT_WORK(&pfault->work, eqe_pf_action);
+		queue_work(eq->pf_ctx.wq, &pfault->work);
+
+		++eq->cons_index;
+		++set_ci;
+
+		if (unlikely(set_ci >= MLX5_NUM_SPARE_EQE)) {
+			eq_update_ci(eq, 0);
+			set_ci = 0;
+		}
+	}
+
+	eq_update_ci(eq, 1);
+}
+
+static irqreturn_t mlx5_eq_pf_int(int irq, void *eq_ptr)
+{
+	struct mlx5_eq *eq = eq_ptr;
+	unsigned long flags;
+
+	if (spin_trylock_irqsave(&eq->pf_ctx.lock, flags)) {
+		eq_pf_process(eq);
+		spin_unlock_irqrestore(&eq->pf_ctx.lock, flags);
+	} else {
+		schedule_work(&eq->pf_ctx.work);
+	}
+
+	return IRQ_HANDLED;
+}
+
+/* mempool_refill() was proposed but unfortunately wasn't accepted
+ * http://lkml.iu.edu/hypermail/linux/kernel/1512.1/05073.html
+ * Chip workaround.
+ */
+static void mempool_refill(mempool_t *pool)
+{
+	while (pool->curr_nr < pool->min_nr)
+		mempool_free(mempool_alloc(pool, GFP_KERNEL), pool);
+}
+
+static void eq_pf_action(struct work_struct *work)
+{
+	struct mlx5_eq *eq = container_of(work, struct mlx5_eq, pf_ctx.work);
+
+	mempool_refill(eq->pf_ctx.pool);
+
+	spin_lock_irq(&eq->pf_ctx.lock);
+	eq_pf_process(eq);
+	spin_unlock_irq(&eq->pf_ctx.lock);
+}
+
+static int init_pf_ctx(struct mlx5_eq_pagefault *pf_ctx, const char *name)
+{
+	spin_lock_init(&pf_ctx->lock);
+	INIT_WORK(&pf_ctx->work, eq_pf_action);
+
+	pf_ctx->wq = alloc_ordered_workqueue(name,
+					     WQ_MEM_RECLAIM);
+	if (!pf_ctx->wq)
+		return -ENOMEM;
+
+	pf_ctx->pool = mempool_create_kmalloc_pool
+		(MLX5_NUM_PF_DRAIN, sizeof(struct mlx5_pagefault));
+	if (!pf_ctx->pool)
+		goto err_wq;
+
+	return 0;
+err_wq:
+	destroy_workqueue(pf_ctx->wq);
+	return -ENOMEM;
+}
+
+int mlx5_core_page_fault_resume(struct mlx5_core_dev *dev, u32 token,
+				u32 wq_num, u8 type, int error)
+{
+	u32 out[MLX5_ST_SZ_DW(page_fault_resume_out)] = {0};
+	u32 in[MLX5_ST_SZ_DW(page_fault_resume_in)]   = {0};
+
+	MLX5_SET(page_fault_resume_in, in, opcode,
+		 MLX5_CMD_OP_PAGE_FAULT_RESUME);
+	MLX5_SET(page_fault_resume_in, in, error, !!error);
+	MLX5_SET(page_fault_resume_in, in, page_fault_type, type);
+	MLX5_SET(page_fault_resume_in, in, wq_number, wq_num);
+	MLX5_SET(page_fault_resume_in, in, token, token);
+
+	return mlx5_cmd_exec(dev, in, sizeof(in), out, sizeof(out));
+}
+EXPORT_SYMBOL_GPL(mlx5_core_page_fault_resume);
+#endif
+
+static void general_event_handler(struct mlx5_core_dev *dev,
+				  struct mlx5_eqe *eqe)
+{
+	switch (eqe->sub_type) {
+	case MLX5_GENERAL_SUBTYPE_DELAY_DROP_TIMEOUT:
+		if (dev->event)
+			dev->event(dev, MLX5_DEV_EVENT_DELAY_DROP_TIMEOUT, 0);
+		break;
+	default:
+		mlx5_core_dbg(dev, "General event with unrecognized subtype: sub_type %d\n",
+			      eqe->sub_type);
+	}
+}
+
+static void mlx5_temp_warning_event(struct mlx5_core_dev *dev,
+				    struct mlx5_eqe *eqe)
+{
+	u64 value_lsb;
+	u64 value_msb;
+
+	value_lsb = be64_to_cpu(eqe->data.temp_warning.sensor_warning_lsb);
+	value_msb = be64_to_cpu(eqe->data.temp_warning.sensor_warning_msb);
+
+	mlx5_core_warn(dev,
+		       "High temperature on sensors with bit set %llx %llx",
+		       value_msb, value_lsb);
+}
+
+/* caller must eventually call mlx5_cq_put on the returned cq */
+static struct mlx5_core_cq *mlx5_eq_cq_get(struct mlx5_eq *eq, u32 cqn)
+{
+	struct mlx5_cq_table *table = &eq->cq_table;
+	struct mlx5_core_cq *cq = NULL;
+
+	spin_lock(&table->lock);
+	cq = radix_tree_lookup(&table->tree, cqn);
+	if (likely(cq))
+		mlx5_cq_hold(cq);
+	spin_unlock(&table->lock);
+
+	return cq;
+}
+
+static void mlx5_eq_cq_completion(struct mlx5_eq *eq, u32 cqn)
+{
+	struct mlx5_core_cq *cq = mlx5_eq_cq_get(eq, cqn);
+
+	if (unlikely(!cq)) {
+		mlx5_core_warn(eq->dev, "Completion event for bogus CQ 0x%x\n", cqn);
+		return;
+	}
+
+	++cq->arm_sn;
+
+	cq->comp(cq);
+
+	mlx5_cq_put(cq);
+}
+
+static void mlx5_eq_cq_event(struct mlx5_eq *eq, u32 cqn, int event_type)
+{
+	struct mlx5_core_cq *cq = mlx5_eq_cq_get(eq, cqn);
+
+	if (unlikely(!cq)) {
+		mlx5_core_warn(eq->dev, "Async event for bogus CQ 0x%x\n", cqn);
+		return;
+	}
+
+	cq->event(cq, event_type);
+
+	mlx5_cq_put(cq);
+}
+
+static irqreturn_t mlx5_eq_int(int irq, void *eq_ptr)
+{
+	struct mlx5_eq *eq = eq_ptr;
+	struct mlx5_core_dev *dev = eq->dev;
+	struct mlx5_eqe *eqe;
+	int set_ci = 0;
+	u32 cqn = -1;
+	u32 rsn;
+	u8 port;
+
+	while ((eqe = next_eqe_sw(eq))) {
+		/*
+		 * Make sure we read EQ entry contents after we've
+		 * checked the ownership bit.
+		 */
+		dma_rmb();
+
+		mlx5_core_dbg(eq->dev, "eqn %d, eqe type %s\n",
+			      eq->eqn, eqe_type_str(eqe->type));
+		switch (eqe->type) {
+		case MLX5_EVENT_TYPE_COMP:
+			cqn = be32_to_cpu(eqe->data.comp.cqn) & 0xffffff;
+			mlx5_eq_cq_completion(eq, cqn);
+			break;
+		case MLX5_EVENT_TYPE_DCT_DRAINED:
+			rsn = be32_to_cpu(eqe->data.dct.dctn) & 0xffffff;
+			rsn |= (MLX5_RES_DCT << MLX5_USER_INDEX_LEN);
+			mlx5_rsc_event(dev, rsn, eqe->type);
+			break;
+		case MLX5_EVENT_TYPE_PATH_MIG:
+		case MLX5_EVENT_TYPE_COMM_EST:
+		case MLX5_EVENT_TYPE_SQ_DRAINED:
+		case MLX5_EVENT_TYPE_SRQ_LAST_WQE:
+		case MLX5_EVENT_TYPE_WQ_CATAS_ERROR:
+		case MLX5_EVENT_TYPE_PATH_MIG_FAILED:
+		case MLX5_EVENT_TYPE_WQ_INVAL_REQ_ERROR:
+		case MLX5_EVENT_TYPE_WQ_ACCESS_ERROR:
+			rsn = be32_to_cpu(eqe->data.qp_srq.qp_srq_n) & 0xffffff;
+			rsn |= (eqe->data.qp_srq.type << MLX5_USER_INDEX_LEN);
+			mlx5_core_dbg(dev, "event %s(%d) arrived on resource 0x%x\n",
+				      eqe_type_str(eqe->type), eqe->type, rsn);
+			mlx5_rsc_event(dev, rsn, eqe->type);
+			break;
+
+		case MLX5_EVENT_TYPE_SRQ_RQ_LIMIT:
+		case MLX5_EVENT_TYPE_SRQ_CATAS_ERROR:
+			rsn = be32_to_cpu(eqe->data.qp_srq.qp_srq_n) & 0xffffff;
+			mlx5_core_dbg(dev, "SRQ event %s(%d): srqn 0x%x\n",
+				      eqe_type_str(eqe->type), eqe->type, rsn);
+			mlx5_srq_event(dev, rsn, eqe->type);
+			break;
+
+		case MLX5_EVENT_TYPE_CMD:
+			mlx5_cmd_comp_handler(dev, be32_to_cpu(eqe->data.cmd.vector), false);
+			break;
+
+		case MLX5_EVENT_TYPE_PORT_CHANGE:
+			port = (eqe->data.port.port >> 4) & 0xf;
+			switch (eqe->sub_type) {
+			case MLX5_PORT_CHANGE_SUBTYPE_DOWN:
+			case MLX5_PORT_CHANGE_SUBTYPE_ACTIVE:
+			case MLX5_PORT_CHANGE_SUBTYPE_LID:
+			case MLX5_PORT_CHANGE_SUBTYPE_PKEY:
+			case MLX5_PORT_CHANGE_SUBTYPE_GUID:
+			case MLX5_PORT_CHANGE_SUBTYPE_CLIENT_REREG:
+			case MLX5_PORT_CHANGE_SUBTYPE_INITIALIZED:
+				if (dev->event)
+					dev->event(dev, port_subtype_event(eqe->sub_type),
+						   (unsigned long)port);
+				break;
+			default:
+				mlx5_core_warn(dev, "Port event with unrecognized subtype: port %d, sub_type %d\n",
+					       port, eqe->sub_type);
+			}
+			break;
+		case MLX5_EVENT_TYPE_CQ_ERROR:
+			cqn = be32_to_cpu(eqe->data.cq_err.cqn) & 0xffffff;
+			mlx5_core_warn(dev, "CQ error on CQN 0x%x, syndrome 0x%x\n",
+				       cqn, eqe->data.cq_err.syndrome);
+			mlx5_eq_cq_event(eq, cqn, eqe->type);
+			break;
+
+		case MLX5_EVENT_TYPE_PAGE_REQUEST:
+			{
+				u16 func_id = be16_to_cpu(eqe->data.req_pages.func_id);
+				s32 npages = be32_to_cpu(eqe->data.req_pages.num_pages);
+
+				mlx5_core_dbg(dev, "page request for func 0x%x, npages %d\n",
+					      func_id, npages);
+				mlx5_core_req_pages_handler(dev, func_id, npages);
+			}
+			break;
+
+		case MLX5_EVENT_TYPE_NIC_VPORT_CHANGE:
+			mlx5_eswitch_vport_event(dev->priv.eswitch, eqe);
+			break;
+
+		case MLX5_EVENT_TYPE_PORT_MODULE_EVENT:
+			mlx5_port_module_event(dev, eqe);
+			break;
+
+		case MLX5_EVENT_TYPE_PPS_EVENT:
+			mlx5_pps_event(dev, eqe);
+			break;
+
+		case MLX5_EVENT_TYPE_FPGA_ERROR:
+		case MLX5_EVENT_TYPE_FPGA_QP_ERROR:
+			mlx5_fpga_event(dev, eqe->type, &eqe->data.raw);
+			break;
+
+		case MLX5_EVENT_TYPE_TEMP_WARN_EVENT:
+			mlx5_temp_warning_event(dev, eqe);
+			break;
+
+		case MLX5_EVENT_TYPE_GENERAL_EVENT:
+			general_event_handler(dev, eqe);
+			break;
+
+		case MLX5_EVENT_TYPE_DEVICE_TRACER:
+			mlx5_fw_tracer_event(dev, eqe);
+			break;
+
+		default:
+			mlx5_core_warn(dev, "Unhandled event 0x%x on EQ 0x%x\n",
+				       eqe->type, eq->eqn);
+			break;
+		}
+
+		++eq->cons_index;
+		++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 MLX5_NUM_SPARE_EQE extra
+		 * entries, so we must update our consumer index at
+		 * least that often.
+		 */
+		if (unlikely(set_ci >= MLX5_NUM_SPARE_EQE)) {
+			eq_update_ci(eq, 0);
+			set_ci = 0;
+		}
+	}
+
+	eq_update_ci(eq, 1);
+
+	if (cqn != -1)
+		tasklet_schedule(&eq->tasklet_ctx.task);
+
+	return IRQ_HANDLED;
+}
+
+/* Some architectures don't latch interrupts when they are disabled, so using
+ * mlx5_eq_poll_irq_disabled could end up losing interrupts while trying to
+ * avoid losing them.  It is not recommended to use it, unless this is the last
+ * resort.
+ */
+u32 mlx5_eq_poll_irq_disabled(struct mlx5_eq *eq)
+{
+	u32 count_eqe;
+
+	disable_irq(eq->irqn);
+	count_eqe = eq->cons_index;
+	mlx5_eq_int(eq->irqn, eq);
+	count_eqe = eq->cons_index - count_eqe;
+	enable_irq(eq->irqn);
+
+	return count_eqe;
+}
+
+static void init_eq_buf(struct mlx5_eq *eq)
+{
+	struct mlx5_eqe *eqe;
+	int i;
+
+	for (i = 0; i < eq->nent; i++) {
+		eqe = get_eqe(eq, i);
+		eqe->owner = MLX5_EQE_OWNER_INIT_VAL;
+	}
+}
+
+int mlx5_create_map_eq(struct mlx5_core_dev *dev, struct mlx5_eq *eq, u8 vecidx,
+		       int nent, u64 mask, const char *name,
+		       enum mlx5_eq_type type)
+{
+	struct mlx5_cq_table *cq_table = &eq->cq_table;
+	u32 out[MLX5_ST_SZ_DW(create_eq_out)] = {0};
+	struct mlx5_priv *priv = &dev->priv;
+	irq_handler_t handler;
+	__be64 *pas;
+	void *eqc;
+	int inlen;
+	u32 *in;
+	int err;
+
+	/* Init CQ table */
+	memset(cq_table, 0, sizeof(*cq_table));
+	spin_lock_init(&cq_table->lock);
+	INIT_RADIX_TREE(&cq_table->tree, GFP_ATOMIC);
+
+	eq->type = type;
+	eq->nent = roundup_pow_of_two(nent + MLX5_NUM_SPARE_EQE);
+	eq->cons_index = 0;
+	err = mlx5_buf_alloc(dev, eq->nent * MLX5_EQE_SIZE, &eq->buf);
+	if (err)
+		return err;
+
+#ifdef CONFIG_INFINIBAND_ON_DEMAND_PAGING
+	if (type == MLX5_EQ_TYPE_PF)
+		handler = mlx5_eq_pf_int;
+	else
+#endif
+		handler = mlx5_eq_int;
+
+	init_eq_buf(eq);
+
+	inlen = MLX5_ST_SZ_BYTES(create_eq_in) +
+		MLX5_FLD_SZ_BYTES(create_eq_in, pas[0]) * eq->buf.npages;
+
+	in = kvzalloc(inlen, GFP_KERNEL);
+	if (!in) {
+		err = -ENOMEM;
+		goto err_buf;
+	}
+
+	pas = (__be64 *)MLX5_ADDR_OF(create_eq_in, in, pas);
+	mlx5_fill_page_array(&eq->buf, pas);
+
+	MLX5_SET(create_eq_in, in, opcode, MLX5_CMD_OP_CREATE_EQ);
+	MLX5_SET64(create_eq_in, in, event_bitmask, mask);
+
+	eqc = MLX5_ADDR_OF(create_eq_in, in, eq_context_entry);
+	MLX5_SET(eqc, eqc, log_eq_size, ilog2(eq->nent));
+	MLX5_SET(eqc, eqc, uar_page, priv->uar->index);
+	MLX5_SET(eqc, eqc, intr, vecidx);
+	MLX5_SET(eqc, eqc, log_page_size,
+		 eq->buf.page_shift - MLX5_ADAPTER_PAGE_SHIFT);
+
+	err = mlx5_cmd_exec(dev, in, inlen, out, sizeof(out));
+	if (err)
+		goto err_in;
+
+	snprintf(priv->irq_info[vecidx].name, MLX5_MAX_IRQ_NAME, "%s@pci:%s",
+		 name, pci_name(dev->pdev));
+
+	eq->eqn = MLX5_GET(create_eq_out, out, eq_number);
+	eq->irqn = pci_irq_vector(dev->pdev, vecidx);
+	eq->dev = dev;
+	eq->doorbell = priv->uar->map + MLX5_EQ_DOORBEL_OFFSET;
+	err = request_irq(eq->irqn, handler, 0,
+			  priv->irq_info[vecidx].name, eq);
+	if (err)
+		goto err_eq;
+
+	err = mlx5_debug_eq_add(dev, eq);
+	if (err)
+		goto err_irq;
+
+#ifdef CONFIG_INFINIBAND_ON_DEMAND_PAGING
+	if (type == MLX5_EQ_TYPE_PF) {
+		err = init_pf_ctx(&eq->pf_ctx, name);
+		if (err)
+			goto err_irq;
+	} else
+#endif
+	{
+		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, mlx5_cq_tasklet_cb,
+			     (unsigned long)&eq->tasklet_ctx);
+	}
+
+	/* EQs are created in ARMED state
+	 */
+	eq_update_ci(eq, 1);
+
+	kvfree(in);
+	return 0;
+
+err_irq:
+	free_irq(eq->irqn, eq);
+
+err_eq:
+	mlx5_cmd_destroy_eq(dev, eq->eqn);
+
+err_in:
+	kvfree(in);
+
+err_buf:
+	mlx5_buf_free(dev, &eq->buf);
+	return err;
+}
+
+int mlx5_destroy_unmap_eq(struct mlx5_core_dev *dev, struct mlx5_eq *eq)
+{
+	int err;
+
+	mlx5_debug_eq_remove(dev, eq);
+	free_irq(eq->irqn, eq);
+	err = mlx5_cmd_destroy_eq(dev, eq->eqn);
+	if (err)
+		mlx5_core_warn(dev, "failed to destroy a previously created eq: eqn %d\n",
+			       eq->eqn);
+	synchronize_irq(eq->irqn);
+
+	if (eq->type == MLX5_EQ_TYPE_COMP) {
+		tasklet_disable(&eq->tasklet_ctx.task);
+#ifdef CONFIG_INFINIBAND_ON_DEMAND_PAGING
+	} else if (eq->type == MLX5_EQ_TYPE_PF) {
+		cancel_work_sync(&eq->pf_ctx.work);
+		destroy_workqueue(eq->pf_ctx.wq);
+		mempool_destroy(eq->pf_ctx.pool);
+#endif
+	}
+	mlx5_buf_free(dev, &eq->buf);
+
+	return err;
+}
+
+int mlx5_eq_add_cq(struct mlx5_eq *eq, struct mlx5_core_cq *cq)
+{
+	struct mlx5_cq_table *table = &eq->cq_table;
+	int err;
+
+	spin_lock_irq(&table->lock);
+	err = radix_tree_insert(&table->tree, cq->cqn, cq);
+	spin_unlock_irq(&table->lock);
+
+	return err;
+}
+
+int mlx5_eq_del_cq(struct mlx5_eq *eq, struct mlx5_core_cq *cq)
+{
+	struct mlx5_cq_table *table = &eq->cq_table;
+	struct mlx5_core_cq *tmp;
+
+	spin_lock_irq(&table->lock);
+	tmp = radix_tree_delete(&table->tree, cq->cqn);
+	spin_unlock_irq(&table->lock);
+
+	if (!tmp) {
+		mlx5_core_warn(eq->dev, "cq 0x%x not found in eq 0x%x tree\n", eq->eqn, cq->cqn);
+		return -ENOENT;
+	}
+
+	if (tmp != cq) {
+		mlx5_core_warn(eq->dev, "corruption on cqn 0x%x in eq 0x%x\n", eq->eqn, cq->cqn);
+		return -EINVAL;
+	}
+
+	return 0;
+}
+
+int mlx5_eq_init(struct mlx5_core_dev *dev)
+{
+	int err;
+
+	spin_lock_init(&dev->priv.eq_table.lock);
+
+	err = mlx5_eq_debugfs_init(dev);
+
+	return err;
+}
+
+void mlx5_eq_cleanup(struct mlx5_core_dev *dev)
+{
+	mlx5_eq_debugfs_cleanup(dev);
+}
+
+int mlx5_start_eqs(struct mlx5_core_dev *dev)
+{
+	struct mlx5_eq_table *table = &dev->priv.eq_table;
+	u64 async_event_mask = MLX5_ASYNC_EVENT_MASK;
+	int err;
+
+	if (MLX5_VPORT_MANAGER(dev))
+		async_event_mask |= (1ull << MLX5_EVENT_TYPE_NIC_VPORT_CHANGE);
+
+	if (MLX5_CAP_GEN(dev, port_type) == MLX5_CAP_PORT_TYPE_ETH &&
+	    MLX5_CAP_GEN(dev, general_notification_event))
+		async_event_mask |= (1ull << MLX5_EVENT_TYPE_GENERAL_EVENT);
+
+	if (MLX5_CAP_GEN(dev, port_module_event))
+		async_event_mask |= (1ull << MLX5_EVENT_TYPE_PORT_MODULE_EVENT);
+	else
+		mlx5_core_dbg(dev, "port_module_event is not set\n");
+
+	if (MLX5_PPS_CAP(dev))
+		async_event_mask |= (1ull << MLX5_EVENT_TYPE_PPS_EVENT);
+
+	if (MLX5_CAP_GEN(dev, fpga))
+		async_event_mask |= (1ull << MLX5_EVENT_TYPE_FPGA_ERROR) |
+				    (1ull << MLX5_EVENT_TYPE_FPGA_QP_ERROR);
+	if (MLX5_CAP_GEN_MAX(dev, dct))
+		async_event_mask |= (1ull << MLX5_EVENT_TYPE_DCT_DRAINED);
+
+	if (MLX5_CAP_GEN(dev, temp_warn_event))
+		async_event_mask |= (1ull << MLX5_EVENT_TYPE_TEMP_WARN_EVENT);
+
+	if (MLX5_CAP_MCAM_REG(dev, tracer_registers))
+		async_event_mask |= (1ull << MLX5_EVENT_TYPE_DEVICE_TRACER);
+
+	err = mlx5_create_map_eq(dev, &table->cmd_eq, MLX5_EQ_VEC_CMD,
+				 MLX5_NUM_CMD_EQE, 1ull << MLX5_EVENT_TYPE_CMD,
+				 "mlx5_cmd_eq", MLX5_EQ_TYPE_ASYNC);
+	if (err) {
+		mlx5_core_warn(dev, "failed to create cmd EQ %d\n", err);
+		return err;
+	}
+
+	mlx5_cmd_use_events(dev);
+
+	err = mlx5_create_map_eq(dev, &table->async_eq, MLX5_EQ_VEC_ASYNC,
+				 MLX5_NUM_ASYNC_EQE, async_event_mask,
+				 "mlx5_async_eq", MLX5_EQ_TYPE_ASYNC);
+	if (err) {
+		mlx5_core_warn(dev, "failed to create async EQ %d\n", err);
+		goto err1;
+	}
+
+	err = mlx5_create_map_eq(dev, &table->pages_eq,
+				 MLX5_EQ_VEC_PAGES,
+				 /* TODO: sriov max_vf + */ 1,
+				 1 << MLX5_EVENT_TYPE_PAGE_REQUEST, "mlx5_pages_eq",
+				 MLX5_EQ_TYPE_ASYNC);
+	if (err) {
+		mlx5_core_warn(dev, "failed to create pages EQ %d\n", err);
+		goto err2;
+	}
+
+#ifdef CONFIG_INFINIBAND_ON_DEMAND_PAGING
+	if (MLX5_CAP_GEN(dev, pg)) {
+		err = mlx5_create_map_eq(dev, &table->pfault_eq,
+					 MLX5_EQ_VEC_PFAULT,
+					 MLX5_NUM_ASYNC_EQE,
+					 1 << MLX5_EVENT_TYPE_PAGE_FAULT,
+					 "mlx5_page_fault_eq",
+					 MLX5_EQ_TYPE_PF);
+		if (err) {
+			mlx5_core_warn(dev, "failed to create page fault EQ %d\n",
+				       err);
+			goto err3;
+		}
+	}
+
+	return err;
+err3:
+	mlx5_destroy_unmap_eq(dev, &table->pages_eq);
+#else
+	return err;
+#endif
+
+err2:
+	mlx5_destroy_unmap_eq(dev, &table->async_eq);
+
+err1:
+	mlx5_cmd_use_polling(dev);
+	mlx5_destroy_unmap_eq(dev, &table->cmd_eq);
+	return err;
+}
+
+void mlx5_stop_eqs(struct mlx5_core_dev *dev)
+{
+	struct mlx5_eq_table *table = &dev->priv.eq_table;
+	int err;
+
+#ifdef CONFIG_INFINIBAND_ON_DEMAND_PAGING
+	if (MLX5_CAP_GEN(dev, pg)) {
+		err = mlx5_destroy_unmap_eq(dev, &table->pfault_eq);
+		if (err)
+			mlx5_core_err(dev, "failed to destroy page fault eq, err(%d)\n",
+				      err);
+	}
+#endif
+
+	err = mlx5_destroy_unmap_eq(dev, &table->pages_eq);
+	if (err)
+		mlx5_core_err(dev, "failed to destroy pages eq, err(%d)\n",
+			      err);
+
+	err = mlx5_destroy_unmap_eq(dev, &table->async_eq);
+	if (err)
+		mlx5_core_err(dev, "failed to destroy async eq, err(%d)\n",
+			      err);
+	mlx5_cmd_use_polling(dev);
+
+	err = mlx5_destroy_unmap_eq(dev, &table->cmd_eq);
+	if (err)
+		mlx5_core_err(dev, "failed to destroy command eq, err(%d)\n",
+			      err);
+}
+
+int mlx5_core_eq_query(struct mlx5_core_dev *dev, struct mlx5_eq *eq,
+		       u32 *out, int outlen)
+{
+	u32 in[MLX5_ST_SZ_DW(query_eq_in)] = {0};
+
+	MLX5_SET(query_eq_in, in, opcode, MLX5_CMD_OP_QUERY_EQ);
+	MLX5_SET(query_eq_in, in, eq_number, eq->eqn);
+	return mlx5_cmd_exec(dev, in, sizeof(in), out, outlen);
+}
+
+/* This function should only be called after mlx5_cmd_force_teardown_hca */
+void mlx5_core_eq_free_irqs(struct mlx5_core_dev *dev)
+{
+	struct mlx5_eq_table *table = &dev->priv.eq_table;
+	struct mlx5_eq *eq;
+
+#ifdef CONFIG_RFS_ACCEL
+	if (dev->rmap) {
+		free_irq_cpu_rmap(dev->rmap);
+		dev->rmap = NULL;
+	}
+#endif
+	list_for_each_entry(eq, &table->comp_eqs_list, list)
+		free_irq(eq->irqn, eq);
+
+	free_irq(table->pages_eq.irqn, &table->pages_eq);
+	free_irq(table->async_eq.irqn, &table->async_eq);
+	free_irq(table->cmd_eq.irqn, &table->cmd_eq);
+#ifdef CONFIG_INFINIBAND_ON_DEMAND_PAGING
+	if (MLX5_CAP_GEN(dev, pg))
+		free_irq(table->pfault_eq.irqn, &table->pfault_eq);
+#endif
+	pci_free_irq_vectors(dev->pdev);
+}
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/eswitch.c b/drivers/net/ethernet/mellanox/mlx5/core/eswitch.c
new file mode 100644
index 000000000..2190daace
--- /dev/null
+++ b/drivers/net/ethernet/mellanox/mlx5/core/eswitch.c
@@ -0,0 +1,2219 @@
+/*
+ * Copyright (c) 2015, 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/etherdevice.h>
+#include <linux/mlx5/driver.h>
+#include <linux/mlx5/mlx5_ifc.h>
+#include <linux/mlx5/vport.h>
+#include <linux/mlx5/fs.h>
+#include "mlx5_core.h"
+#include "eswitch.h"
+#include "fs_core.h"
+
+#define UPLINK_VPORT 0xFFFF
+
+enum {
+	MLX5_ACTION_NONE = 0,
+	MLX5_ACTION_ADD  = 1,
+	MLX5_ACTION_DEL  = 2,
+};
+
+/* Vport UC/MC hash node */
+struct vport_addr {
+	struct l2addr_node     node;
+	u8                     action;
+	u32                    vport;
+	struct mlx5_flow_handle *flow_rule;
+	bool mpfs; /* UC MAC was added to MPFs */
+	/* A flag indicating that mac was added due to mc promiscuous vport */
+	bool mc_promisc;
+};
+
+enum {
+	UC_ADDR_CHANGE = BIT(0),
+	MC_ADDR_CHANGE = BIT(1),
+	PROMISC_CHANGE = BIT(3),
+};
+
+/* Vport context events */
+#define SRIOV_VPORT_EVENTS (UC_ADDR_CHANGE | \
+			    MC_ADDR_CHANGE | \
+			    PROMISC_CHANGE)
+
+static int arm_vport_context_events_cmd(struct mlx5_core_dev *dev, u16 vport,
+					u32 events_mask)
+{
+	int in[MLX5_ST_SZ_DW(modify_nic_vport_context_in)]   = {0};
+	int out[MLX5_ST_SZ_DW(modify_nic_vport_context_out)] = {0};
+	void *nic_vport_ctx;
+
+	MLX5_SET(modify_nic_vport_context_in, in,
+		 opcode, MLX5_CMD_OP_MODIFY_NIC_VPORT_CONTEXT);
+	MLX5_SET(modify_nic_vport_context_in, in, field_select.change_event, 1);
+	MLX5_SET(modify_nic_vport_context_in, in, vport_number, vport);
+	MLX5_SET(modify_nic_vport_context_in, in, other_vport, 1);
+	nic_vport_ctx = MLX5_ADDR_OF(modify_nic_vport_context_in,
+				     in, nic_vport_context);
+
+	MLX5_SET(nic_vport_context, nic_vport_ctx, arm_change_event, 1);
+
+	if (events_mask & UC_ADDR_CHANGE)
+		MLX5_SET(nic_vport_context, nic_vport_ctx,
+			 event_on_uc_address_change, 1);
+	if (events_mask & MC_ADDR_CHANGE)
+		MLX5_SET(nic_vport_context, nic_vport_ctx,
+			 event_on_mc_address_change, 1);
+	if (events_mask & PROMISC_CHANGE)
+		MLX5_SET(nic_vport_context, nic_vport_ctx,
+			 event_on_promisc_change, 1);
+
+	return mlx5_cmd_exec(dev, in, sizeof(in), out, sizeof(out));
+}
+
+/* E-Switch vport context HW commands */
+static int modify_esw_vport_context_cmd(struct mlx5_core_dev *dev, u16 vport,
+					void *in, int inlen)
+{
+	u32 out[MLX5_ST_SZ_DW(modify_esw_vport_context_out)] = {0};
+
+	MLX5_SET(modify_esw_vport_context_in, in, opcode,
+		 MLX5_CMD_OP_MODIFY_ESW_VPORT_CONTEXT);
+	MLX5_SET(modify_esw_vport_context_in, in, vport_number, vport);
+	MLX5_SET(modify_esw_vport_context_in, in, other_vport, 1);
+	return mlx5_cmd_exec(dev, in, inlen, out, sizeof(out));
+}
+
+static int modify_esw_vport_cvlan(struct mlx5_core_dev *dev, u32 vport,
+				  u16 vlan, u8 qos, u8 set_flags)
+{
+	u32 in[MLX5_ST_SZ_DW(modify_esw_vport_context_in)] = {0};
+
+	if (!MLX5_CAP_ESW(dev, vport_cvlan_strip) ||
+	    !MLX5_CAP_ESW(dev, vport_cvlan_insert_if_not_exist))
+		return -EOPNOTSUPP;
+
+	esw_debug(dev, "Set Vport[%d] VLAN %d qos %d set=%x\n",
+		  vport, vlan, qos, set_flags);
+
+	if (set_flags & SET_VLAN_STRIP)
+		MLX5_SET(modify_esw_vport_context_in, in,
+			 esw_vport_context.vport_cvlan_strip, 1);
+
+	if (set_flags & SET_VLAN_INSERT) {
+		/* insert only if no vlan in packet */
+		MLX5_SET(modify_esw_vport_context_in, in,
+			 esw_vport_context.vport_cvlan_insert, 1);
+
+		MLX5_SET(modify_esw_vport_context_in, in,
+			 esw_vport_context.cvlan_pcp, qos);
+		MLX5_SET(modify_esw_vport_context_in, in,
+			 esw_vport_context.cvlan_id, vlan);
+	}
+
+	MLX5_SET(modify_esw_vport_context_in, in,
+		 field_select.vport_cvlan_strip, 1);
+	MLX5_SET(modify_esw_vport_context_in, in,
+		 field_select.vport_cvlan_insert, 1);
+
+	return modify_esw_vport_context_cmd(dev, vport, in, sizeof(in));
+}
+
+/* E-Switch FDB */
+static struct mlx5_flow_handle *
+__esw_fdb_set_vport_rule(struct mlx5_eswitch *esw, u32 vport, bool rx_rule,
+			 u8 mac_c[ETH_ALEN], u8 mac_v[ETH_ALEN])
+{
+	int match_header = (is_zero_ether_addr(mac_c) ? 0 :
+			    MLX5_MATCH_OUTER_HEADERS);
+	struct mlx5_flow_handle *flow_rule = NULL;
+	struct mlx5_flow_act flow_act = {0};
+	struct mlx5_flow_destination dest = {};
+	struct mlx5_flow_spec *spec;
+	void *mv_misc = NULL;
+	void *mc_misc = NULL;
+	u8 *dmac_v = NULL;
+	u8 *dmac_c = NULL;
+
+	if (rx_rule)
+		match_header |= MLX5_MATCH_MISC_PARAMETERS;
+
+	spec = kvzalloc(sizeof(*spec), GFP_KERNEL);
+	if (!spec)
+		return NULL;
+
+	dmac_v = MLX5_ADDR_OF(fte_match_param, spec->match_value,
+			      outer_headers.dmac_47_16);
+	dmac_c = MLX5_ADDR_OF(fte_match_param, spec->match_criteria,
+			      outer_headers.dmac_47_16);
+
+	if (match_header & MLX5_MATCH_OUTER_HEADERS) {
+		ether_addr_copy(dmac_v, mac_v);
+		ether_addr_copy(dmac_c, mac_c);
+	}
+
+	if (match_header & MLX5_MATCH_MISC_PARAMETERS) {
+		mv_misc  = MLX5_ADDR_OF(fte_match_param, spec->match_value,
+					misc_parameters);
+		mc_misc  = MLX5_ADDR_OF(fte_match_param, spec->match_criteria,
+					misc_parameters);
+		MLX5_SET(fte_match_set_misc, mv_misc, source_port, UPLINK_VPORT);
+		MLX5_SET_TO_ONES(fte_match_set_misc, mc_misc, source_port);
+	}
+
+	dest.type = MLX5_FLOW_DESTINATION_TYPE_VPORT;
+	dest.vport.num = vport;
+
+	esw_debug(esw->dev,
+		  "\tFDB add rule dmac_v(%pM) dmac_c(%pM) -> vport(%d)\n",
+		  dmac_v, dmac_c, vport);
+	spec->match_criteria_enable = match_header;
+	flow_act.action =  MLX5_FLOW_CONTEXT_ACTION_FWD_DEST;
+	flow_rule =
+		mlx5_add_flow_rules(esw->fdb_table.legacy.fdb, spec,
+				    &flow_act, &dest, 1);
+	if (IS_ERR(flow_rule)) {
+		esw_warn(esw->dev,
+			 "FDB: Failed to add flow rule: dmac_v(%pM) dmac_c(%pM) -> vport(%d), err(%ld)\n",
+			 dmac_v, dmac_c, vport, PTR_ERR(flow_rule));
+		flow_rule = NULL;
+	}
+
+	kvfree(spec);
+	return flow_rule;
+}
+
+static struct mlx5_flow_handle *
+esw_fdb_set_vport_rule(struct mlx5_eswitch *esw, u8 mac[ETH_ALEN], u32 vport)
+{
+	u8 mac_c[ETH_ALEN];
+
+	eth_broadcast_addr(mac_c);
+	return __esw_fdb_set_vport_rule(esw, vport, false, mac_c, mac);
+}
+
+static struct mlx5_flow_handle *
+esw_fdb_set_vport_allmulti_rule(struct mlx5_eswitch *esw, u32 vport)
+{
+	u8 mac_c[ETH_ALEN];
+	u8 mac_v[ETH_ALEN];
+
+	eth_zero_addr(mac_c);
+	eth_zero_addr(mac_v);
+	mac_c[0] = 0x01;
+	mac_v[0] = 0x01;
+	return __esw_fdb_set_vport_rule(esw, vport, false, mac_c, mac_v);
+}
+
+static struct mlx5_flow_handle *
+esw_fdb_set_vport_promisc_rule(struct mlx5_eswitch *esw, u32 vport)
+{
+	u8 mac_c[ETH_ALEN];
+	u8 mac_v[ETH_ALEN];
+
+	eth_zero_addr(mac_c);
+	eth_zero_addr(mac_v);
+	return __esw_fdb_set_vport_rule(esw, vport, true, mac_c, mac_v);
+}
+
+static int esw_create_legacy_fdb_table(struct mlx5_eswitch *esw)
+{
+	int inlen = MLX5_ST_SZ_BYTES(create_flow_group_in);
+	struct mlx5_flow_table_attr ft_attr = {};
+	struct mlx5_core_dev *dev = esw->dev;
+	struct mlx5_flow_namespace *root_ns;
+	struct mlx5_flow_table *fdb;
+	struct mlx5_flow_group *g;
+	void *match_criteria;
+	int table_size;
+	u32 *flow_group_in;
+	u8 *dmac;
+	int err = 0;
+
+	esw_debug(dev, "Create FDB log_max_size(%d)\n",
+		  MLX5_CAP_ESW_FLOWTABLE_FDB(dev, log_max_ft_size));
+
+	root_ns = mlx5_get_flow_namespace(dev, MLX5_FLOW_NAMESPACE_FDB);
+	if (!root_ns) {
+		esw_warn(dev, "Failed to get FDB flow namespace\n");
+		return -EOPNOTSUPP;
+	}
+
+	flow_group_in = kvzalloc(inlen, GFP_KERNEL);
+	if (!flow_group_in)
+		return -ENOMEM;
+
+	table_size = BIT(MLX5_CAP_ESW_FLOWTABLE_FDB(dev, log_max_ft_size));
+
+	ft_attr.max_fte = table_size;
+	fdb = mlx5_create_flow_table(root_ns, &ft_attr);
+	if (IS_ERR(fdb)) {
+		err = PTR_ERR(fdb);
+		esw_warn(dev, "Failed to create FDB Table err %d\n", err);
+		goto out;
+	}
+	esw->fdb_table.legacy.fdb = fdb;
+
+	/* Addresses group : Full match unicast/multicast addresses */
+	MLX5_SET(create_flow_group_in, flow_group_in, match_criteria_enable,
+		 MLX5_MATCH_OUTER_HEADERS);
+	match_criteria = MLX5_ADDR_OF(create_flow_group_in, flow_group_in, match_criteria);
+	dmac = MLX5_ADDR_OF(fte_match_param, match_criteria, outer_headers.dmac_47_16);
+	MLX5_SET(create_flow_group_in, flow_group_in, start_flow_index, 0);
+	/* Preserve 2 entries for allmulti and promisc rules*/
+	MLX5_SET(create_flow_group_in, flow_group_in, end_flow_index, table_size - 3);
+	eth_broadcast_addr(dmac);
+	g = mlx5_create_flow_group(fdb, flow_group_in);
+	if (IS_ERR(g)) {
+		err = PTR_ERR(g);
+		esw_warn(dev, "Failed to create flow group err(%d)\n", err);
+		goto out;
+	}
+	esw->fdb_table.legacy.addr_grp = g;
+
+	/* Allmulti group : One rule that forwards any mcast traffic */
+	MLX5_SET(create_flow_group_in, flow_group_in, match_criteria_enable,
+		 MLX5_MATCH_OUTER_HEADERS);
+	MLX5_SET(create_flow_group_in, flow_group_in, start_flow_index, table_size - 2);
+	MLX5_SET(create_flow_group_in, flow_group_in, end_flow_index, table_size - 2);
+	eth_zero_addr(dmac);
+	dmac[0] = 0x01;
+	g = mlx5_create_flow_group(fdb, flow_group_in);
+	if (IS_ERR(g)) {
+		err = PTR_ERR(g);
+		esw_warn(dev, "Failed to create allmulti flow group err(%d)\n", err);
+		goto out;
+	}
+	esw->fdb_table.legacy.allmulti_grp = g;
+
+	/* Promiscuous group :
+	 * One rule that forward all unmatched traffic from previous groups
+	 */
+	eth_zero_addr(dmac);
+	MLX5_SET(create_flow_group_in, flow_group_in, match_criteria_enable,
+		 MLX5_MATCH_MISC_PARAMETERS);
+	MLX5_SET_TO_ONES(fte_match_param, match_criteria, misc_parameters.source_port);
+	MLX5_SET(create_flow_group_in, flow_group_in, start_flow_index, table_size - 1);
+	MLX5_SET(create_flow_group_in, flow_group_in, end_flow_index, table_size - 1);
+	g = mlx5_create_flow_group(fdb, flow_group_in);
+	if (IS_ERR(g)) {
+		err = PTR_ERR(g);
+		esw_warn(dev, "Failed to create promisc flow group err(%d)\n", err);
+		goto out;
+	}
+	esw->fdb_table.legacy.promisc_grp = g;
+
+out:
+	if (err) {
+		if (!IS_ERR_OR_NULL(esw->fdb_table.legacy.allmulti_grp)) {
+			mlx5_destroy_flow_group(esw->fdb_table.legacy.allmulti_grp);
+			esw->fdb_table.legacy.allmulti_grp = NULL;
+		}
+		if (!IS_ERR_OR_NULL(esw->fdb_table.legacy.addr_grp)) {
+			mlx5_destroy_flow_group(esw->fdb_table.legacy.addr_grp);
+			esw->fdb_table.legacy.addr_grp = NULL;
+		}
+		if (!IS_ERR_OR_NULL(esw->fdb_table.legacy.fdb)) {
+			mlx5_destroy_flow_table(esw->fdb_table.legacy.fdb);
+			esw->fdb_table.legacy.fdb = NULL;
+		}
+	}
+
+	kvfree(flow_group_in);
+	return err;
+}
+
+static void esw_destroy_legacy_fdb_table(struct mlx5_eswitch *esw)
+{
+	if (!esw->fdb_table.legacy.fdb)
+		return;
+
+	esw_debug(esw->dev, "Destroy FDB Table\n");
+	mlx5_destroy_flow_group(esw->fdb_table.legacy.promisc_grp);
+	mlx5_destroy_flow_group(esw->fdb_table.legacy.allmulti_grp);
+	mlx5_destroy_flow_group(esw->fdb_table.legacy.addr_grp);
+	mlx5_destroy_flow_table(esw->fdb_table.legacy.fdb);
+	esw->fdb_table.legacy.fdb = NULL;
+	esw->fdb_table.legacy.addr_grp = NULL;
+	esw->fdb_table.legacy.allmulti_grp = NULL;
+	esw->fdb_table.legacy.promisc_grp = NULL;
+}
+
+/* E-Switch vport UC/MC lists management */
+typedef int (*vport_addr_action)(struct mlx5_eswitch *esw,
+				 struct vport_addr *vaddr);
+
+static int esw_add_uc_addr(struct mlx5_eswitch *esw, struct vport_addr *vaddr)
+{
+	u8 *mac = vaddr->node.addr;
+	u32 vport = vaddr->vport;
+	int err;
+
+	/* Skip mlx5_mpfs_add_mac for PFs,
+	 * it is already done by the PF netdev in mlx5e_execute_l2_action
+	 */
+	if (!vport)
+		goto fdb_add;
+
+	err = mlx5_mpfs_add_mac(esw->dev, mac);
+	if (err) {
+		esw_warn(esw->dev,
+			 "Failed to add L2 table mac(%pM) for vport(%d), err(%d)\n",
+			 mac, vport, err);
+		return err;
+	}
+	vaddr->mpfs = true;
+
+fdb_add:
+	/* SRIOV is enabled: Forward UC MAC to vport */
+	if (esw->fdb_table.legacy.fdb && esw->mode == SRIOV_LEGACY)
+		vaddr->flow_rule = esw_fdb_set_vport_rule(esw, mac, vport);
+
+	esw_debug(esw->dev, "\tADDED UC MAC: vport[%d] %pM fr(%p)\n",
+		  vport, mac, vaddr->flow_rule);
+
+	return 0;
+}
+
+static int esw_del_uc_addr(struct mlx5_eswitch *esw, struct vport_addr *vaddr)
+{
+	u8 *mac = vaddr->node.addr;
+	u32 vport = vaddr->vport;
+	int err = 0;
+
+	/* Skip mlx5_mpfs_del_mac for PFs,
+	 * it is already done by the PF netdev in mlx5e_execute_l2_action
+	 */
+	if (!vport || !vaddr->mpfs)
+		goto fdb_del;
+
+	err = mlx5_mpfs_del_mac(esw->dev, mac);
+	if (err)
+		esw_warn(esw->dev,
+			 "Failed to del L2 table mac(%pM) for vport(%d), err(%d)\n",
+			 mac, vport, err);
+	vaddr->mpfs = false;
+
+fdb_del:
+	if (vaddr->flow_rule)
+		mlx5_del_flow_rules(vaddr->flow_rule);
+	vaddr->flow_rule = NULL;
+
+	return 0;
+}
+
+static void update_allmulti_vports(struct mlx5_eswitch *esw,
+				   struct vport_addr *vaddr,
+				   struct esw_mc_addr *esw_mc)
+{
+	u8 *mac = vaddr->node.addr;
+	u32 vport_idx = 0;
+
+	for (vport_idx = 0; vport_idx < esw->total_vports; vport_idx++) {
+		struct mlx5_vport *vport = &esw->vports[vport_idx];
+		struct hlist_head *vport_hash = vport->mc_list;
+		struct vport_addr *iter_vaddr =
+					l2addr_hash_find(vport_hash,
+							 mac,
+							 struct vport_addr);
+		if (IS_ERR_OR_NULL(vport->allmulti_rule) ||
+		    vaddr->vport == vport_idx)
+			continue;
+		switch (vaddr->action) {
+		case MLX5_ACTION_ADD:
+			if (iter_vaddr)
+				continue;
+			iter_vaddr = l2addr_hash_add(vport_hash, mac,
+						     struct vport_addr,
+						     GFP_KERNEL);
+			if (!iter_vaddr) {
+				esw_warn(esw->dev,
+					 "ALL-MULTI: Failed to add MAC(%pM) to vport[%d] DB\n",
+					 mac, vport_idx);
+				continue;
+			}
+			iter_vaddr->vport = vport_idx;
+			iter_vaddr->flow_rule =
+					esw_fdb_set_vport_rule(esw,
+							       mac,
+							       vport_idx);
+			iter_vaddr->mc_promisc = true;
+			break;
+		case MLX5_ACTION_DEL:
+			if (!iter_vaddr)
+				continue;
+			mlx5_del_flow_rules(iter_vaddr->flow_rule);
+			l2addr_hash_del(iter_vaddr);
+			break;
+		}
+	}
+}
+
+static int esw_add_mc_addr(struct mlx5_eswitch *esw, struct vport_addr *vaddr)
+{
+	struct hlist_head *hash = esw->mc_table;
+	struct esw_mc_addr *esw_mc;
+	u8 *mac = vaddr->node.addr;
+	u32 vport = vaddr->vport;
+
+	if (!esw->fdb_table.legacy.fdb)
+		return 0;
+
+	esw_mc = l2addr_hash_find(hash, mac, struct esw_mc_addr);
+	if (esw_mc)
+		goto add;
+
+	esw_mc = l2addr_hash_add(hash, mac, struct esw_mc_addr, GFP_KERNEL);
+	if (!esw_mc)
+		return -ENOMEM;
+
+	esw_mc->uplink_rule = /* Forward MC MAC to Uplink */
+		esw_fdb_set_vport_rule(esw, mac, UPLINK_VPORT);
+
+	/* Add this multicast mac to all the mc promiscuous vports */
+	update_allmulti_vports(esw, vaddr, esw_mc);
+
+add:
+	/* If the multicast mac is added as a result of mc promiscuous vport,
+	 * don't increment the multicast ref count
+	 */
+	if (!vaddr->mc_promisc)
+		esw_mc->refcnt++;
+
+	/* Forward MC MAC to vport */
+	vaddr->flow_rule = esw_fdb_set_vport_rule(esw, mac, vport);
+	esw_debug(esw->dev,
+		  "\tADDED MC MAC: vport[%d] %pM fr(%p) refcnt(%d) uplinkfr(%p)\n",
+		  vport, mac, vaddr->flow_rule,
+		  esw_mc->refcnt, esw_mc->uplink_rule);
+	return 0;
+}
+
+static int esw_del_mc_addr(struct mlx5_eswitch *esw, struct vport_addr *vaddr)
+{
+	struct hlist_head *hash = esw->mc_table;
+	struct esw_mc_addr *esw_mc;
+	u8 *mac = vaddr->node.addr;
+	u32 vport = vaddr->vport;
+
+	if (!esw->fdb_table.legacy.fdb)
+		return 0;
+
+	esw_mc = l2addr_hash_find(hash, mac, struct esw_mc_addr);
+	if (!esw_mc) {
+		esw_warn(esw->dev,
+			 "Failed to find eswitch MC addr for MAC(%pM) vport(%d)",
+			 mac, vport);
+		return -EINVAL;
+	}
+	esw_debug(esw->dev,
+		  "\tDELETE MC MAC: vport[%d] %pM fr(%p) refcnt(%d) uplinkfr(%p)\n",
+		  vport, mac, vaddr->flow_rule, esw_mc->refcnt,
+		  esw_mc->uplink_rule);
+
+	if (vaddr->flow_rule)
+		mlx5_del_flow_rules(vaddr->flow_rule);
+	vaddr->flow_rule = NULL;
+
+	/* If the multicast mac is added as a result of mc promiscuous vport,
+	 * don't decrement the multicast ref count.
+	 */
+	if (vaddr->mc_promisc || (--esw_mc->refcnt > 0))
+		return 0;
+
+	/* Remove this multicast mac from all the mc promiscuous vports */
+	update_allmulti_vports(esw, vaddr, esw_mc);
+
+	if (esw_mc->uplink_rule)
+		mlx5_del_flow_rules(esw_mc->uplink_rule);
+
+	l2addr_hash_del(esw_mc);
+	return 0;
+}
+
+/* Apply vport UC/MC list to HW l2 table and FDB table */
+static void esw_apply_vport_addr_list(struct mlx5_eswitch *esw,
+				      u32 vport_num, int list_type)
+{
+	struct mlx5_vport *vport = &esw->vports[vport_num];
+	bool is_uc = list_type == MLX5_NVPRT_LIST_TYPE_UC;
+	vport_addr_action vport_addr_add;
+	vport_addr_action vport_addr_del;
+	struct vport_addr *addr;
+	struct l2addr_node *node;
+	struct hlist_head *hash;
+	struct hlist_node *tmp;
+	int hi;
+
+	vport_addr_add = is_uc ? esw_add_uc_addr :
+				 esw_add_mc_addr;
+	vport_addr_del = is_uc ? esw_del_uc_addr :
+				 esw_del_mc_addr;
+
+	hash = is_uc ? vport->uc_list : vport->mc_list;
+	for_each_l2hash_node(node, tmp, hash, hi) {
+		addr = container_of(node, struct vport_addr, node);
+		switch (addr->action) {
+		case MLX5_ACTION_ADD:
+			vport_addr_add(esw, addr);
+			addr->action = MLX5_ACTION_NONE;
+			break;
+		case MLX5_ACTION_DEL:
+			vport_addr_del(esw, addr);
+			l2addr_hash_del(addr);
+			break;
+		}
+	}
+}
+
+/* Sync vport UC/MC list from vport context */
+static void esw_update_vport_addr_list(struct mlx5_eswitch *esw,
+				       u32 vport_num, int list_type)
+{
+	struct mlx5_vport *vport = &esw->vports[vport_num];
+	bool is_uc = list_type == MLX5_NVPRT_LIST_TYPE_UC;
+	u8 (*mac_list)[ETH_ALEN];
+	struct l2addr_node *node;
+	struct vport_addr *addr;
+	struct hlist_head *hash;
+	struct hlist_node *tmp;
+	int size;
+	int err;
+	int hi;
+	int i;
+
+	size = is_uc ? MLX5_MAX_UC_PER_VPORT(esw->dev) :
+		       MLX5_MAX_MC_PER_VPORT(esw->dev);
+
+	mac_list = kcalloc(size, ETH_ALEN, GFP_KERNEL);
+	if (!mac_list)
+		return;
+
+	hash = is_uc ? vport->uc_list : vport->mc_list;
+
+	for_each_l2hash_node(node, tmp, hash, hi) {
+		addr = container_of(node, struct vport_addr, node);
+		addr->action = MLX5_ACTION_DEL;
+	}
+
+	if (!vport->enabled)
+		goto out;
+
+	err = mlx5_query_nic_vport_mac_list(esw->dev, vport_num, list_type,
+					    mac_list, &size);
+	if (err)
+		goto out;
+	esw_debug(esw->dev, "vport[%d] context update %s list size (%d)\n",
+		  vport_num, is_uc ? "UC" : "MC", size);
+
+	for (i = 0; i < size; i++) {
+		if (is_uc && !is_valid_ether_addr(mac_list[i]))
+			continue;
+
+		if (!is_uc && !is_multicast_ether_addr(mac_list[i]))
+			continue;
+
+		addr = l2addr_hash_find(hash, mac_list[i], struct vport_addr);
+		if (addr) {
+			addr->action = MLX5_ACTION_NONE;
+			/* If this mac was previously added because of allmulti
+			 * promiscuous rx mode, its now converted to be original
+			 * vport mac.
+			 */
+			if (addr->mc_promisc) {
+				struct esw_mc_addr *esw_mc =
+					l2addr_hash_find(esw->mc_table,
+							 mac_list[i],
+							 struct esw_mc_addr);
+				if (!esw_mc) {
+					esw_warn(esw->dev,
+						 "Failed to MAC(%pM) in mcast DB\n",
+						 mac_list[i]);
+					continue;
+				}
+				esw_mc->refcnt++;
+				addr->mc_promisc = false;
+			}
+			continue;
+		}
+
+		addr = l2addr_hash_add(hash, mac_list[i], struct vport_addr,
+				       GFP_KERNEL);
+		if (!addr) {
+			esw_warn(esw->dev,
+				 "Failed to add MAC(%pM) to vport[%d] DB\n",
+				 mac_list[i], vport_num);
+			continue;
+		}
+		addr->vport = vport_num;
+		addr->action = MLX5_ACTION_ADD;
+	}
+out:
+	kfree(mac_list);
+}
+
+/* Sync vport UC/MC list from vport context
+ * Must be called after esw_update_vport_addr_list
+ */
+static void esw_update_vport_mc_promisc(struct mlx5_eswitch *esw, u32 vport_num)
+{
+	struct mlx5_vport *vport = &esw->vports[vport_num];
+	struct l2addr_node *node;
+	struct vport_addr *addr;
+	struct hlist_head *hash;
+	struct hlist_node *tmp;
+	int hi;
+
+	hash = vport->mc_list;
+
+	for_each_l2hash_node(node, tmp, esw->mc_table, hi) {
+		u8 *mac = node->addr;
+
+		addr = l2addr_hash_find(hash, mac, struct vport_addr);
+		if (addr) {
+			if (addr->action == MLX5_ACTION_DEL)
+				addr->action = MLX5_ACTION_NONE;
+			continue;
+		}
+		addr = l2addr_hash_add(hash, mac, struct vport_addr,
+				       GFP_KERNEL);
+		if (!addr) {
+			esw_warn(esw->dev,
+				 "Failed to add allmulti MAC(%pM) to vport[%d] DB\n",
+				 mac, vport_num);
+			continue;
+		}
+		addr->vport = vport_num;
+		addr->action = MLX5_ACTION_ADD;
+		addr->mc_promisc = true;
+	}
+}
+
+/* Apply vport rx mode to HW FDB table */
+static void esw_apply_vport_rx_mode(struct mlx5_eswitch *esw, u32 vport_num,
+				    bool promisc, bool mc_promisc)
+{
+	struct esw_mc_addr *allmulti_addr = &esw->mc_promisc;
+	struct mlx5_vport *vport = &esw->vports[vport_num];
+
+	if (IS_ERR_OR_NULL(vport->allmulti_rule) != mc_promisc)
+		goto promisc;
+
+	if (mc_promisc) {
+		vport->allmulti_rule =
+				esw_fdb_set_vport_allmulti_rule(esw, vport_num);
+		if (!allmulti_addr->uplink_rule)
+			allmulti_addr->uplink_rule =
+				esw_fdb_set_vport_allmulti_rule(esw,
+								UPLINK_VPORT);
+		allmulti_addr->refcnt++;
+	} else if (vport->allmulti_rule) {
+		mlx5_del_flow_rules(vport->allmulti_rule);
+		vport->allmulti_rule = NULL;
+
+		if (--allmulti_addr->refcnt > 0)
+			goto promisc;
+
+		if (allmulti_addr->uplink_rule)
+			mlx5_del_flow_rules(allmulti_addr->uplink_rule);
+		allmulti_addr->uplink_rule = NULL;
+	}
+
+promisc:
+	if (IS_ERR_OR_NULL(vport->promisc_rule) != promisc)
+		return;
+
+	if (promisc) {
+		vport->promisc_rule = esw_fdb_set_vport_promisc_rule(esw,
+								     vport_num);
+	} else if (vport->promisc_rule) {
+		mlx5_del_flow_rules(vport->promisc_rule);
+		vport->promisc_rule = NULL;
+	}
+}
+
+/* Sync vport rx mode from vport context */
+static void esw_update_vport_rx_mode(struct mlx5_eswitch *esw, u32 vport_num)
+{
+	struct mlx5_vport *vport = &esw->vports[vport_num];
+	int promisc_all = 0;
+	int promisc_uc = 0;
+	int promisc_mc = 0;
+	int err;
+
+	err = mlx5_query_nic_vport_promisc(esw->dev,
+					   vport_num,
+					   &promisc_uc,
+					   &promisc_mc,
+					   &promisc_all);
+	if (err)
+		return;
+	esw_debug(esw->dev, "vport[%d] context update rx mode promisc_all=%d, all_multi=%d\n",
+		  vport_num, promisc_all, promisc_mc);
+
+	if (!vport->info.trusted || !vport->enabled) {
+		promisc_uc = 0;
+		promisc_mc = 0;
+		promisc_all = 0;
+	}
+
+	esw_apply_vport_rx_mode(esw, vport_num, promisc_all,
+				(promisc_all || promisc_mc));
+}
+
+static void esw_vport_change_handle_locked(struct mlx5_vport *vport)
+{
+	struct mlx5_core_dev *dev = vport->dev;
+	struct mlx5_eswitch *esw = dev->priv.eswitch;
+	u8 mac[ETH_ALEN];
+
+	mlx5_query_nic_vport_mac_address(dev, vport->vport, mac);
+	esw_debug(dev, "vport[%d] Context Changed: perm mac: %pM\n",
+		  vport->vport, mac);
+
+	if (vport->enabled_events & UC_ADDR_CHANGE) {
+		esw_update_vport_addr_list(esw, vport->vport,
+					   MLX5_NVPRT_LIST_TYPE_UC);
+		esw_apply_vport_addr_list(esw, vport->vport,
+					  MLX5_NVPRT_LIST_TYPE_UC);
+	}
+
+	if (vport->enabled_events & MC_ADDR_CHANGE) {
+		esw_update_vport_addr_list(esw, vport->vport,
+					   MLX5_NVPRT_LIST_TYPE_MC);
+	}
+
+	if (vport->enabled_events & PROMISC_CHANGE) {
+		esw_update_vport_rx_mode(esw, vport->vport);
+		if (!IS_ERR_OR_NULL(vport->allmulti_rule))
+			esw_update_vport_mc_promisc(esw, vport->vport);
+	}
+
+	if (vport->enabled_events & (PROMISC_CHANGE | MC_ADDR_CHANGE)) {
+		esw_apply_vport_addr_list(esw, vport->vport,
+					  MLX5_NVPRT_LIST_TYPE_MC);
+	}
+
+	esw_debug(esw->dev, "vport[%d] Context Changed: Done\n", vport->vport);
+	if (vport->enabled)
+		arm_vport_context_events_cmd(dev, vport->vport,
+					     vport->enabled_events);
+}
+
+static void esw_vport_change_handler(struct work_struct *work)
+{
+	struct mlx5_vport *vport =
+		container_of(work, struct mlx5_vport, vport_change_handler);
+	struct mlx5_eswitch *esw = vport->dev->priv.eswitch;
+
+	mutex_lock(&esw->state_lock);
+	esw_vport_change_handle_locked(vport);
+	mutex_unlock(&esw->state_lock);
+}
+
+static int esw_vport_enable_egress_acl(struct mlx5_eswitch *esw,
+				       struct mlx5_vport *vport)
+{
+	int inlen = MLX5_ST_SZ_BYTES(create_flow_group_in);
+	struct mlx5_flow_group *vlan_grp = NULL;
+	struct mlx5_flow_group *drop_grp = NULL;
+	struct mlx5_core_dev *dev = esw->dev;
+	struct mlx5_flow_namespace *root_ns;
+	struct mlx5_flow_table *acl;
+	void *match_criteria;
+	u32 *flow_group_in;
+	/* The egress acl table contains 2 rules:
+	 * 1)Allow traffic with vlan_tag=vst_vlan_id
+	 * 2)Drop all other traffic.
+	 */
+	int table_size = 2;
+	int err = 0;
+
+	if (!MLX5_CAP_ESW_EGRESS_ACL(dev, ft_support))
+		return -EOPNOTSUPP;
+
+	if (!IS_ERR_OR_NULL(vport->egress.acl))
+		return 0;
+
+	esw_debug(dev, "Create vport[%d] egress ACL log_max_size(%d)\n",
+		  vport->vport, MLX5_CAP_ESW_EGRESS_ACL(dev, log_max_ft_size));
+
+	root_ns = mlx5_get_flow_vport_acl_namespace(dev, MLX5_FLOW_NAMESPACE_ESW_EGRESS,
+						    vport->vport);
+	if (!root_ns) {
+		esw_warn(dev, "Failed to get E-Switch egress flow namespace for vport (%d)\n", vport->vport);
+		return -EOPNOTSUPP;
+	}
+
+	flow_group_in = kvzalloc(inlen, GFP_KERNEL);
+	if (!flow_group_in)
+		return -ENOMEM;
+
+	acl = mlx5_create_vport_flow_table(root_ns, 0, table_size, 0, vport->vport);
+	if (IS_ERR(acl)) {
+		err = PTR_ERR(acl);
+		esw_warn(dev, "Failed to create E-Switch vport[%d] egress flow Table, err(%d)\n",
+			 vport->vport, err);
+		goto out;
+	}
+
+	MLX5_SET(create_flow_group_in, flow_group_in, match_criteria_enable, MLX5_MATCH_OUTER_HEADERS);
+	match_criteria = MLX5_ADDR_OF(create_flow_group_in, flow_group_in, match_criteria);
+	MLX5_SET_TO_ONES(fte_match_param, match_criteria, outer_headers.cvlan_tag);
+	MLX5_SET_TO_ONES(fte_match_param, match_criteria, outer_headers.first_vid);
+	MLX5_SET(create_flow_group_in, flow_group_in, start_flow_index, 0);
+	MLX5_SET(create_flow_group_in, flow_group_in, end_flow_index, 0);
+
+	vlan_grp = mlx5_create_flow_group(acl, flow_group_in);
+	if (IS_ERR(vlan_grp)) {
+		err = PTR_ERR(vlan_grp);
+		esw_warn(dev, "Failed to create E-Switch vport[%d] egress allowed vlans flow group, err(%d)\n",
+			 vport->vport, err);
+		goto out;
+	}
+
+	memset(flow_group_in, 0, inlen);
+	MLX5_SET(create_flow_group_in, flow_group_in, start_flow_index, 1);
+	MLX5_SET(create_flow_group_in, flow_group_in, end_flow_index, 1);
+	drop_grp = mlx5_create_flow_group(acl, flow_group_in);
+	if (IS_ERR(drop_grp)) {
+		err = PTR_ERR(drop_grp);
+		esw_warn(dev, "Failed to create E-Switch vport[%d] egress drop flow group, err(%d)\n",
+			 vport->vport, err);
+		goto out;
+	}
+
+	vport->egress.acl = acl;
+	vport->egress.drop_grp = drop_grp;
+	vport->egress.allowed_vlans_grp = vlan_grp;
+out:
+	kvfree(flow_group_in);
+	if (err && !IS_ERR_OR_NULL(vlan_grp))
+		mlx5_destroy_flow_group(vlan_grp);
+	if (err && !IS_ERR_OR_NULL(acl))
+		mlx5_destroy_flow_table(acl);
+	return err;
+}
+
+static void esw_vport_cleanup_egress_rules(struct mlx5_eswitch *esw,
+					   struct mlx5_vport *vport)
+{
+	if (!IS_ERR_OR_NULL(vport->egress.allowed_vlan))
+		mlx5_del_flow_rules(vport->egress.allowed_vlan);
+
+	if (!IS_ERR_OR_NULL(vport->egress.drop_rule))
+		mlx5_del_flow_rules(vport->egress.drop_rule);
+
+	vport->egress.allowed_vlan = NULL;
+	vport->egress.drop_rule = NULL;
+}
+
+static void esw_vport_disable_egress_acl(struct mlx5_eswitch *esw,
+					 struct mlx5_vport *vport)
+{
+	if (IS_ERR_OR_NULL(vport->egress.acl))
+		return;
+
+	esw_debug(esw->dev, "Destroy vport[%d] E-Switch egress ACL\n", vport->vport);
+
+	esw_vport_cleanup_egress_rules(esw, vport);
+	mlx5_destroy_flow_group(vport->egress.allowed_vlans_grp);
+	mlx5_destroy_flow_group(vport->egress.drop_grp);
+	mlx5_destroy_flow_table(vport->egress.acl);
+	vport->egress.allowed_vlans_grp = NULL;
+	vport->egress.drop_grp = NULL;
+	vport->egress.acl = NULL;
+}
+
+static int esw_vport_enable_ingress_acl(struct mlx5_eswitch *esw,
+					struct mlx5_vport *vport)
+{
+	int inlen = MLX5_ST_SZ_BYTES(create_flow_group_in);
+	struct mlx5_core_dev *dev = esw->dev;
+	struct mlx5_flow_namespace *root_ns;
+	struct mlx5_flow_table *acl;
+	struct mlx5_flow_group *g;
+	void *match_criteria;
+	u32 *flow_group_in;
+	/* The ingress acl table contains 4 groups
+	 * (2 active rules at the same time -
+	 *      1 allow rule from one of the first 3 groups.
+	 *      1 drop rule from the last group):
+	 * 1)Allow untagged traffic with smac=original mac.
+	 * 2)Allow untagged traffic.
+	 * 3)Allow traffic with smac=original mac.
+	 * 4)Drop all other traffic.
+	 */
+	int table_size = 4;
+	int err = 0;
+
+	if (!MLX5_CAP_ESW_INGRESS_ACL(dev, ft_support))
+		return -EOPNOTSUPP;
+
+	if (!IS_ERR_OR_NULL(vport->ingress.acl))
+		return 0;
+
+	esw_debug(dev, "Create vport[%d] ingress ACL log_max_size(%d)\n",
+		  vport->vport, MLX5_CAP_ESW_INGRESS_ACL(dev, log_max_ft_size));
+
+	root_ns = mlx5_get_flow_vport_acl_namespace(dev, MLX5_FLOW_NAMESPACE_ESW_INGRESS,
+						    vport->vport);
+	if (!root_ns) {
+		esw_warn(dev, "Failed to get E-Switch ingress flow namespace for vport (%d)\n", vport->vport);
+		return -EOPNOTSUPP;
+	}
+
+	flow_group_in = kvzalloc(inlen, GFP_KERNEL);
+	if (!flow_group_in)
+		return -ENOMEM;
+
+	acl = mlx5_create_vport_flow_table(root_ns, 0, table_size, 0, vport->vport);
+	if (IS_ERR(acl)) {
+		err = PTR_ERR(acl);
+		esw_warn(dev, "Failed to create E-Switch vport[%d] ingress flow Table, err(%d)\n",
+			 vport->vport, err);
+		goto out;
+	}
+	vport->ingress.acl = acl;
+
+	match_criteria = MLX5_ADDR_OF(create_flow_group_in, flow_group_in, match_criteria);
+
+	MLX5_SET(create_flow_group_in, flow_group_in, match_criteria_enable, MLX5_MATCH_OUTER_HEADERS);
+	MLX5_SET_TO_ONES(fte_match_param, match_criteria, outer_headers.cvlan_tag);
+	MLX5_SET_TO_ONES(fte_match_param, match_criteria, outer_headers.smac_47_16);
+	MLX5_SET_TO_ONES(fte_match_param, match_criteria, outer_headers.smac_15_0);
+	MLX5_SET(create_flow_group_in, flow_group_in, start_flow_index, 0);
+	MLX5_SET(create_flow_group_in, flow_group_in, end_flow_index, 0);
+
+	g = mlx5_create_flow_group(acl, flow_group_in);
+	if (IS_ERR(g)) {
+		err = PTR_ERR(g);
+		esw_warn(dev, "Failed to create E-Switch vport[%d] ingress untagged spoofchk flow group, err(%d)\n",
+			 vport->vport, err);
+		goto out;
+	}
+	vport->ingress.allow_untagged_spoofchk_grp = g;
+
+	memset(flow_group_in, 0, inlen);
+	MLX5_SET(create_flow_group_in, flow_group_in, match_criteria_enable, MLX5_MATCH_OUTER_HEADERS);
+	MLX5_SET_TO_ONES(fte_match_param, match_criteria, outer_headers.cvlan_tag);
+	MLX5_SET(create_flow_group_in, flow_group_in, start_flow_index, 1);
+	MLX5_SET(create_flow_group_in, flow_group_in, end_flow_index, 1);
+
+	g = mlx5_create_flow_group(acl, flow_group_in);
+	if (IS_ERR(g)) {
+		err = PTR_ERR(g);
+		esw_warn(dev, "Failed to create E-Switch vport[%d] ingress untagged flow group, err(%d)\n",
+			 vport->vport, err);
+		goto out;
+	}
+	vport->ingress.allow_untagged_only_grp = g;
+
+	memset(flow_group_in, 0, inlen);
+	MLX5_SET(create_flow_group_in, flow_group_in, match_criteria_enable, MLX5_MATCH_OUTER_HEADERS);
+	MLX5_SET_TO_ONES(fte_match_param, match_criteria, outer_headers.smac_47_16);
+	MLX5_SET_TO_ONES(fte_match_param, match_criteria, outer_headers.smac_15_0);
+	MLX5_SET(create_flow_group_in, flow_group_in, start_flow_index, 2);
+	MLX5_SET(create_flow_group_in, flow_group_in, end_flow_index, 2);
+
+	g = mlx5_create_flow_group(acl, flow_group_in);
+	if (IS_ERR(g)) {
+		err = PTR_ERR(g);
+		esw_warn(dev, "Failed to create E-Switch vport[%d] ingress spoofchk flow group, err(%d)\n",
+			 vport->vport, err);
+		goto out;
+	}
+	vport->ingress.allow_spoofchk_only_grp = g;
+
+	memset(flow_group_in, 0, inlen);
+	MLX5_SET(create_flow_group_in, flow_group_in, start_flow_index, 3);
+	MLX5_SET(create_flow_group_in, flow_group_in, end_flow_index, 3);
+
+	g = mlx5_create_flow_group(acl, flow_group_in);
+	if (IS_ERR(g)) {
+		err = PTR_ERR(g);
+		esw_warn(dev, "Failed to create E-Switch vport[%d] ingress drop flow group, err(%d)\n",
+			 vport->vport, err);
+		goto out;
+	}
+	vport->ingress.drop_grp = g;
+
+out:
+	if (err) {
+		if (!IS_ERR_OR_NULL(vport->ingress.allow_spoofchk_only_grp))
+			mlx5_destroy_flow_group(
+					vport->ingress.allow_spoofchk_only_grp);
+		if (!IS_ERR_OR_NULL(vport->ingress.allow_untagged_only_grp))
+			mlx5_destroy_flow_group(
+					vport->ingress.allow_untagged_only_grp);
+		if (!IS_ERR_OR_NULL(vport->ingress.allow_untagged_spoofchk_grp))
+			mlx5_destroy_flow_group(
+				vport->ingress.allow_untagged_spoofchk_grp);
+		if (!IS_ERR_OR_NULL(vport->ingress.acl))
+			mlx5_destroy_flow_table(vport->ingress.acl);
+	}
+
+	kvfree(flow_group_in);
+	return err;
+}
+
+static void esw_vport_cleanup_ingress_rules(struct mlx5_eswitch *esw,
+					    struct mlx5_vport *vport)
+{
+	if (!IS_ERR_OR_NULL(vport->ingress.drop_rule))
+		mlx5_del_flow_rules(vport->ingress.drop_rule);
+
+	if (!IS_ERR_OR_NULL(vport->ingress.allow_rule))
+		mlx5_del_flow_rules(vport->ingress.allow_rule);
+
+	vport->ingress.drop_rule = NULL;
+	vport->ingress.allow_rule = NULL;
+}
+
+static void esw_vport_disable_ingress_acl(struct mlx5_eswitch *esw,
+					  struct mlx5_vport *vport)
+{
+	if (IS_ERR_OR_NULL(vport->ingress.acl))
+		return;
+
+	esw_debug(esw->dev, "Destroy vport[%d] E-Switch ingress ACL\n", vport->vport);
+
+	esw_vport_cleanup_ingress_rules(esw, vport);
+	mlx5_destroy_flow_group(vport->ingress.allow_spoofchk_only_grp);
+	mlx5_destroy_flow_group(vport->ingress.allow_untagged_only_grp);
+	mlx5_destroy_flow_group(vport->ingress.allow_untagged_spoofchk_grp);
+	mlx5_destroy_flow_group(vport->ingress.drop_grp);
+	mlx5_destroy_flow_table(vport->ingress.acl);
+	vport->ingress.acl = NULL;
+	vport->ingress.drop_grp = NULL;
+	vport->ingress.allow_spoofchk_only_grp = NULL;
+	vport->ingress.allow_untagged_only_grp = NULL;
+	vport->ingress.allow_untagged_spoofchk_grp = NULL;
+}
+
+static int esw_vport_ingress_config(struct mlx5_eswitch *esw,
+				    struct mlx5_vport *vport)
+{
+	struct mlx5_fc *counter = vport->ingress.drop_counter;
+	struct mlx5_flow_destination drop_ctr_dst = {0};
+	struct mlx5_flow_destination *dst = NULL;
+	struct mlx5_flow_act flow_act = {0};
+	struct mlx5_flow_spec *spec;
+	int dest_num = 0;
+	int err = 0;
+	u8 *smac_v;
+
+	esw_vport_cleanup_ingress_rules(esw, vport);
+
+	if (!vport->info.vlan && !vport->info.qos && !vport->info.spoofchk) {
+		esw_vport_disable_ingress_acl(esw, vport);
+		return 0;
+	}
+
+	err = esw_vport_enable_ingress_acl(esw, vport);
+	if (err) {
+		mlx5_core_warn(esw->dev,
+			       "failed to enable ingress acl (%d) on vport[%d]\n",
+			       err, vport->vport);
+		return err;
+	}
+
+	esw_debug(esw->dev,
+		  "vport[%d] configure ingress rules, vlan(%d) qos(%d)\n",
+		  vport->vport, vport->info.vlan, vport->info.qos);
+
+	spec = kvzalloc(sizeof(*spec), GFP_KERNEL);
+	if (!spec) {
+		err = -ENOMEM;
+		goto out;
+	}
+
+	if (vport->info.vlan || vport->info.qos)
+		MLX5_SET_TO_ONES(fte_match_param, spec->match_criteria, outer_headers.cvlan_tag);
+
+	if (vport->info.spoofchk) {
+		MLX5_SET_TO_ONES(fte_match_param, spec->match_criteria, outer_headers.smac_47_16);
+		MLX5_SET_TO_ONES(fte_match_param, spec->match_criteria, outer_headers.smac_15_0);
+		smac_v = MLX5_ADDR_OF(fte_match_param,
+				      spec->match_value,
+				      outer_headers.smac_47_16);
+		ether_addr_copy(smac_v, vport->info.mac);
+	}
+
+	spec->match_criteria_enable = MLX5_MATCH_OUTER_HEADERS;
+	flow_act.action = MLX5_FLOW_CONTEXT_ACTION_ALLOW;
+	vport->ingress.allow_rule =
+		mlx5_add_flow_rules(vport->ingress.acl, spec,
+				    &flow_act, NULL, 0);
+	if (IS_ERR(vport->ingress.allow_rule)) {
+		err = PTR_ERR(vport->ingress.allow_rule);
+		esw_warn(esw->dev,
+			 "vport[%d] configure ingress allow rule, err(%d)\n",
+			 vport->vport, err);
+		vport->ingress.allow_rule = NULL;
+		goto out;
+	}
+
+	memset(spec, 0, sizeof(*spec));
+	flow_act.action = MLX5_FLOW_CONTEXT_ACTION_DROP;
+
+	/* Attach drop flow counter */
+	if (counter) {
+		flow_act.action |= MLX5_FLOW_CONTEXT_ACTION_COUNT;
+		drop_ctr_dst.type = MLX5_FLOW_DESTINATION_TYPE_COUNTER;
+		drop_ctr_dst.counter = counter;
+		dst = &drop_ctr_dst;
+		dest_num++;
+	}
+	vport->ingress.drop_rule =
+		mlx5_add_flow_rules(vport->ingress.acl, spec,
+				    &flow_act, dst, dest_num);
+	if (IS_ERR(vport->ingress.drop_rule)) {
+		err = PTR_ERR(vport->ingress.drop_rule);
+		esw_warn(esw->dev,
+			 "vport[%d] configure ingress drop rule, err(%d)\n",
+			 vport->vport, err);
+		vport->ingress.drop_rule = NULL;
+		goto out;
+	}
+
+out:
+	if (err)
+		esw_vport_cleanup_ingress_rules(esw, vport);
+	kvfree(spec);
+	return err;
+}
+
+static int esw_vport_egress_config(struct mlx5_eswitch *esw,
+				   struct mlx5_vport *vport)
+{
+	struct mlx5_fc *counter = vport->egress.drop_counter;
+	struct mlx5_flow_destination drop_ctr_dst = {0};
+	struct mlx5_flow_destination *dst = NULL;
+	struct mlx5_flow_act flow_act = {0};
+	struct mlx5_flow_spec *spec;
+	int dest_num = 0;
+	int err = 0;
+
+	esw_vport_cleanup_egress_rules(esw, vport);
+
+	if (!vport->info.vlan && !vport->info.qos) {
+		esw_vport_disable_egress_acl(esw, vport);
+		return 0;
+	}
+
+	err = esw_vport_enable_egress_acl(esw, vport);
+	if (err) {
+		mlx5_core_warn(esw->dev,
+			       "failed to enable egress acl (%d) on vport[%d]\n",
+			       err, vport->vport);
+		return err;
+	}
+
+	esw_debug(esw->dev,
+		  "vport[%d] configure egress rules, vlan(%d) qos(%d)\n",
+		  vport->vport, vport->info.vlan, vport->info.qos);
+
+	spec = kvzalloc(sizeof(*spec), GFP_KERNEL);
+	if (!spec) {
+		err = -ENOMEM;
+		goto out;
+	}
+
+	/* Allowed vlan rule */
+	MLX5_SET_TO_ONES(fte_match_param, spec->match_criteria, outer_headers.cvlan_tag);
+	MLX5_SET_TO_ONES(fte_match_param, spec->match_value, outer_headers.cvlan_tag);
+	MLX5_SET_TO_ONES(fte_match_param, spec->match_criteria, outer_headers.first_vid);
+	MLX5_SET(fte_match_param, spec->match_value, outer_headers.first_vid, vport->info.vlan);
+
+	spec->match_criteria_enable = MLX5_MATCH_OUTER_HEADERS;
+	flow_act.action = MLX5_FLOW_CONTEXT_ACTION_ALLOW;
+	vport->egress.allowed_vlan =
+		mlx5_add_flow_rules(vport->egress.acl, spec,
+				    &flow_act, NULL, 0);
+	if (IS_ERR(vport->egress.allowed_vlan)) {
+		err = PTR_ERR(vport->egress.allowed_vlan);
+		esw_warn(esw->dev,
+			 "vport[%d] configure egress allowed vlan rule failed, err(%d)\n",
+			 vport->vport, err);
+		vport->egress.allowed_vlan = NULL;
+		goto out;
+	}
+
+	/* Drop others rule (star rule) */
+	memset(spec, 0, sizeof(*spec));
+	flow_act.action = MLX5_FLOW_CONTEXT_ACTION_DROP;
+
+	/* Attach egress drop flow counter */
+	if (counter) {
+		flow_act.action |= MLX5_FLOW_CONTEXT_ACTION_COUNT;
+		drop_ctr_dst.type = MLX5_FLOW_DESTINATION_TYPE_COUNTER;
+		drop_ctr_dst.counter = counter;
+		dst = &drop_ctr_dst;
+		dest_num++;
+	}
+	vport->egress.drop_rule =
+		mlx5_add_flow_rules(vport->egress.acl, spec,
+				    &flow_act, dst, dest_num);
+	if (IS_ERR(vport->egress.drop_rule)) {
+		err = PTR_ERR(vport->egress.drop_rule);
+		esw_warn(esw->dev,
+			 "vport[%d] configure egress drop rule failed, err(%d)\n",
+			 vport->vport, err);
+		vport->egress.drop_rule = NULL;
+	}
+out:
+	kvfree(spec);
+	return err;
+}
+
+/* Vport QoS management */
+static int esw_create_tsar(struct mlx5_eswitch *esw)
+{
+	u32 tsar_ctx[MLX5_ST_SZ_DW(scheduling_context)] = {0};
+	struct mlx5_core_dev *dev = esw->dev;
+	int err;
+
+	if (!MLX5_CAP_GEN(dev, qos) || !MLX5_CAP_QOS(dev, esw_scheduling))
+		return 0;
+
+	if (esw->qos.enabled)
+		return -EEXIST;
+
+	err = mlx5_create_scheduling_element_cmd(dev,
+						 SCHEDULING_HIERARCHY_E_SWITCH,
+						 tsar_ctx,
+						 &esw->qos.root_tsar_id);
+	if (err) {
+		esw_warn(esw->dev, "E-Switch create TSAR failed (%d)\n", err);
+		return err;
+	}
+
+	esw->qos.enabled = true;
+	return 0;
+}
+
+static void esw_destroy_tsar(struct mlx5_eswitch *esw)
+{
+	int err;
+
+	if (!esw->qos.enabled)
+		return;
+
+	err = mlx5_destroy_scheduling_element_cmd(esw->dev,
+						  SCHEDULING_HIERARCHY_E_SWITCH,
+						  esw->qos.root_tsar_id);
+	if (err)
+		esw_warn(esw->dev, "E-Switch destroy TSAR failed (%d)\n", err);
+
+	esw->qos.enabled = false;
+}
+
+static int esw_vport_enable_qos(struct mlx5_eswitch *esw, int vport_num,
+				u32 initial_max_rate, u32 initial_bw_share)
+{
+	u32 sched_ctx[MLX5_ST_SZ_DW(scheduling_context)] = {0};
+	struct mlx5_vport *vport = &esw->vports[vport_num];
+	struct mlx5_core_dev *dev = esw->dev;
+	void *vport_elem;
+	int err = 0;
+
+	if (!esw->qos.enabled || !MLX5_CAP_GEN(dev, qos) ||
+	    !MLX5_CAP_QOS(dev, esw_scheduling))
+		return 0;
+
+	if (vport->qos.enabled)
+		return -EEXIST;
+
+	MLX5_SET(scheduling_context, sched_ctx, element_type,
+		 SCHEDULING_CONTEXT_ELEMENT_TYPE_VPORT);
+	vport_elem = MLX5_ADDR_OF(scheduling_context, sched_ctx,
+				  element_attributes);
+	MLX5_SET(vport_element, vport_elem, vport_number, vport_num);
+	MLX5_SET(scheduling_context, sched_ctx, parent_element_id,
+		 esw->qos.root_tsar_id);
+	MLX5_SET(scheduling_context, sched_ctx, max_average_bw,
+		 initial_max_rate);
+	MLX5_SET(scheduling_context, sched_ctx, bw_share, initial_bw_share);
+
+	err = mlx5_create_scheduling_element_cmd(dev,
+						 SCHEDULING_HIERARCHY_E_SWITCH,
+						 sched_ctx,
+						 &vport->qos.esw_tsar_ix);
+	if (err) {
+		esw_warn(esw->dev, "E-Switch create TSAR vport element failed (vport=%d,err=%d)\n",
+			 vport_num, err);
+		return err;
+	}
+
+	vport->qos.enabled = true;
+	return 0;
+}
+
+static void esw_vport_disable_qos(struct mlx5_eswitch *esw, int vport_num)
+{
+	struct mlx5_vport *vport = &esw->vports[vport_num];
+	int err = 0;
+
+	if (!vport->qos.enabled)
+		return;
+
+	err = mlx5_destroy_scheduling_element_cmd(esw->dev,
+						  SCHEDULING_HIERARCHY_E_SWITCH,
+						  vport->qos.esw_tsar_ix);
+	if (err)
+		esw_warn(esw->dev, "E-Switch destroy TSAR vport element failed (vport=%d,err=%d)\n",
+			 vport_num, err);
+
+	vport->qos.enabled = false;
+}
+
+static int esw_vport_qos_config(struct mlx5_eswitch *esw, int vport_num,
+				u32 max_rate, u32 bw_share)
+{
+	u32 sched_ctx[MLX5_ST_SZ_DW(scheduling_context)] = {0};
+	struct mlx5_vport *vport = &esw->vports[vport_num];
+	struct mlx5_core_dev *dev = esw->dev;
+	void *vport_elem;
+	u32 bitmask = 0;
+	int err = 0;
+
+	if (!MLX5_CAP_GEN(dev, qos) || !MLX5_CAP_QOS(dev, esw_scheduling))
+		return -EOPNOTSUPP;
+
+	if (!vport->qos.enabled)
+		return -EIO;
+
+	MLX5_SET(scheduling_context, sched_ctx, element_type,
+		 SCHEDULING_CONTEXT_ELEMENT_TYPE_VPORT);
+	vport_elem = MLX5_ADDR_OF(scheduling_context, sched_ctx,
+				  element_attributes);
+	MLX5_SET(vport_element, vport_elem, vport_number, vport_num);
+	MLX5_SET(scheduling_context, sched_ctx, parent_element_id,
+		 esw->qos.root_tsar_id);
+	MLX5_SET(scheduling_context, sched_ctx, max_average_bw,
+		 max_rate);
+	MLX5_SET(scheduling_context, sched_ctx, bw_share, bw_share);
+	bitmask |= MODIFY_SCHEDULING_ELEMENT_IN_MODIFY_BITMASK_MAX_AVERAGE_BW;
+	bitmask |= MODIFY_SCHEDULING_ELEMENT_IN_MODIFY_BITMASK_BW_SHARE;
+
+	err = mlx5_modify_scheduling_element_cmd(dev,
+						 SCHEDULING_HIERARCHY_E_SWITCH,
+						 sched_ctx,
+						 vport->qos.esw_tsar_ix,
+						 bitmask);
+	if (err) {
+		esw_warn(esw->dev, "E-Switch modify TSAR vport element failed (vport=%d,err=%d)\n",
+			 vport_num, err);
+		return err;
+	}
+
+	return 0;
+}
+
+static void node_guid_gen_from_mac(u64 *node_guid, u8 mac[ETH_ALEN])
+{
+	((u8 *)node_guid)[7] = mac[0];
+	((u8 *)node_guid)[6] = mac[1];
+	((u8 *)node_guid)[5] = mac[2];
+	((u8 *)node_guid)[4] = 0xff;
+	((u8 *)node_guid)[3] = 0xfe;
+	((u8 *)node_guid)[2] = mac[3];
+	((u8 *)node_guid)[1] = mac[4];
+	((u8 *)node_guid)[0] = mac[5];
+}
+
+static void esw_apply_vport_conf(struct mlx5_eswitch *esw,
+				 struct mlx5_vport *vport)
+{
+	int vport_num = vport->vport;
+
+	if (!vport_num)
+		return;
+
+	mlx5_modify_vport_admin_state(esw->dev,
+				      MLX5_VPORT_STATE_OP_MOD_ESW_VPORT,
+				      vport_num,
+				      vport->info.link_state);
+	mlx5_modify_nic_vport_mac_address(esw->dev, vport_num, vport->info.mac);
+	mlx5_modify_nic_vport_node_guid(esw->dev, vport_num, vport->info.node_guid);
+	modify_esw_vport_cvlan(esw->dev, vport_num, vport->info.vlan, vport->info.qos,
+			       (vport->info.vlan || vport->info.qos));
+
+	/* Only legacy mode needs ACLs */
+	if (esw->mode == SRIOV_LEGACY) {
+		esw_vport_ingress_config(esw, vport);
+		esw_vport_egress_config(esw, vport);
+	}
+}
+
+static void esw_vport_create_drop_counters(struct mlx5_vport *vport)
+{
+	struct mlx5_core_dev *dev = vport->dev;
+
+	if (MLX5_CAP_ESW_INGRESS_ACL(dev, flow_counter)) {
+		vport->ingress.drop_counter = mlx5_fc_create(dev, false);
+		if (IS_ERR(vport->ingress.drop_counter)) {
+			esw_warn(dev,
+				 "vport[%d] configure ingress drop rule counter failed\n",
+				 vport->vport);
+			vport->ingress.drop_counter = NULL;
+		}
+	}
+
+	if (MLX5_CAP_ESW_EGRESS_ACL(dev, flow_counter)) {
+		vport->egress.drop_counter = mlx5_fc_create(dev, false);
+		if (IS_ERR(vport->egress.drop_counter)) {
+			esw_warn(dev,
+				 "vport[%d] configure egress drop rule counter failed\n",
+				 vport->vport);
+			vport->egress.drop_counter = NULL;
+		}
+	}
+}
+
+static void esw_vport_destroy_drop_counters(struct mlx5_vport *vport)
+{
+	struct mlx5_core_dev *dev = vport->dev;
+
+	if (vport->ingress.drop_counter)
+		mlx5_fc_destroy(dev, vport->ingress.drop_counter);
+	if (vport->egress.drop_counter)
+		mlx5_fc_destroy(dev, vport->egress.drop_counter);
+}
+
+static void esw_enable_vport(struct mlx5_eswitch *esw, int vport_num,
+			     int enable_events)
+{
+	struct mlx5_vport *vport = &esw->vports[vport_num];
+
+	mutex_lock(&esw->state_lock);
+	WARN_ON(vport->enabled);
+
+	esw_debug(esw->dev, "Enabling VPORT(%d)\n", vport_num);
+
+	/* Create steering drop counters for ingress and egress ACLs */
+	if (vport_num && esw->mode == SRIOV_LEGACY)
+		esw_vport_create_drop_counters(vport);
+
+	/* Restore old vport configuration */
+	esw_apply_vport_conf(esw, vport);
+
+	/* Attach vport to the eswitch rate limiter */
+	if (esw_vport_enable_qos(esw, vport_num, vport->info.max_rate,
+				 vport->qos.bw_share))
+		esw_warn(esw->dev, "Failed to attach vport %d to eswitch rate limiter", vport_num);
+
+	/* Sync with current vport context */
+	vport->enabled_events = enable_events;
+	vport->enabled = true;
+
+	/* only PF is trusted by default */
+	if (!vport_num)
+		vport->info.trusted = true;
+
+	esw_vport_change_handle_locked(vport);
+
+	esw->enabled_vports++;
+	esw_debug(esw->dev, "Enabled VPORT(%d)\n", vport_num);
+	mutex_unlock(&esw->state_lock);
+}
+
+static void esw_disable_vport(struct mlx5_eswitch *esw, int vport_num)
+{
+	struct mlx5_vport *vport = &esw->vports[vport_num];
+
+	if (!vport->enabled)
+		return;
+
+	esw_debug(esw->dev, "Disabling vport(%d)\n", vport_num);
+	/* Mark this vport as disabled to discard new events */
+	vport->enabled = false;
+
+	synchronize_irq(pci_irq_vector(esw->dev->pdev, MLX5_EQ_VEC_ASYNC));
+	/* Wait for current already scheduled events to complete */
+	flush_workqueue(esw->work_queue);
+	/* Disable events from this vport */
+	arm_vport_context_events_cmd(esw->dev, vport->vport, 0);
+	mutex_lock(&esw->state_lock);
+	/* We don't assume VFs will cleanup after themselves.
+	 * Calling vport change handler while vport is disabled will cleanup
+	 * the vport resources.
+	 */
+	esw_vport_change_handle_locked(vport);
+	vport->enabled_events = 0;
+	esw_vport_disable_qos(esw, vport_num);
+	if (vport_num && esw->mode == SRIOV_LEGACY) {
+		mlx5_modify_vport_admin_state(esw->dev,
+					      MLX5_VPORT_STATE_OP_MOD_ESW_VPORT,
+					      vport_num,
+					      MLX5_VPORT_ADMIN_STATE_DOWN);
+		esw_vport_disable_egress_acl(esw, vport);
+		esw_vport_disable_ingress_acl(esw, vport);
+		esw_vport_destroy_drop_counters(vport);
+	}
+	esw->enabled_vports--;
+	mutex_unlock(&esw->state_lock);
+}
+
+/* Public E-Switch API */
+#define ESW_ALLOWED(esw) ((esw) && MLX5_ESWITCH_MANAGER((esw)->dev))
+
+
+int mlx5_eswitch_enable_sriov(struct mlx5_eswitch *esw, int nvfs, int mode)
+{
+	int err;
+	int i, enabled_events;
+
+	if (!ESW_ALLOWED(esw) ||
+	    !MLX5_CAP_ESW_FLOWTABLE_FDB(esw->dev, ft_support)) {
+		esw_warn(esw->dev, "E-Switch FDB is not supported, aborting ...\n");
+		return -EOPNOTSUPP;
+	}
+
+	if (!MLX5_CAP_ESW_INGRESS_ACL(esw->dev, ft_support))
+		esw_warn(esw->dev, "E-Switch ingress ACL is not supported by FW\n");
+
+	if (!MLX5_CAP_ESW_EGRESS_ACL(esw->dev, ft_support))
+		esw_warn(esw->dev, "E-Switch engress ACL is not supported by FW\n");
+
+	esw_info(esw->dev, "E-Switch enable SRIOV: nvfs(%d) mode (%d)\n", nvfs, mode);
+	esw->mode = mode;
+
+	if (mode == SRIOV_LEGACY) {
+		err = esw_create_legacy_fdb_table(esw);
+	} else {
+		mlx5_reload_interface(esw->dev, MLX5_INTERFACE_PROTOCOL_IB);
+
+		err = esw_offloads_init(esw, nvfs + 1);
+	}
+
+	if (err)
+		goto abort;
+
+	err = esw_create_tsar(esw);
+	if (err)
+		esw_warn(esw->dev, "Failed to create eswitch TSAR");
+
+	/* Don't enable vport events when in SRIOV_OFFLOADS mode, since:
+	 * 1. L2 table (MPFS) is programmed by PF/VF representors netdevs set_rx_mode
+	 * 2. FDB/Eswitch is programmed by user space tools
+	 */
+	enabled_events = (mode == SRIOV_LEGACY) ? SRIOV_VPORT_EVENTS : 0;
+	for (i = 0; i <= nvfs; i++)
+		esw_enable_vport(esw, i, enabled_events);
+
+	esw_info(esw->dev, "SRIOV enabled: active vports(%d)\n",
+		 esw->enabled_vports);
+	return 0;
+
+abort:
+	esw->mode = SRIOV_NONE;
+
+	if (mode == SRIOV_OFFLOADS)
+		mlx5_reload_interface(esw->dev, MLX5_INTERFACE_PROTOCOL_IB);
+
+	return err;
+}
+
+void mlx5_eswitch_disable_sriov(struct mlx5_eswitch *esw)
+{
+	struct esw_mc_addr *mc_promisc;
+	int old_mode;
+	int nvports;
+	int i;
+
+	if (!ESW_ALLOWED(esw) || esw->mode == SRIOV_NONE)
+		return;
+
+	esw_info(esw->dev, "disable SRIOV: active vports(%d) mode(%d)\n",
+		 esw->enabled_vports, esw->mode);
+
+	mc_promisc = &esw->mc_promisc;
+	nvports = esw->enabled_vports;
+
+	for (i = 0; i < esw->total_vports; i++)
+		esw_disable_vport(esw, i);
+
+	if (mc_promisc && mc_promisc->uplink_rule)
+		mlx5_del_flow_rules(mc_promisc->uplink_rule);
+
+	esw_destroy_tsar(esw);
+
+	if (esw->mode == SRIOV_LEGACY)
+		esw_destroy_legacy_fdb_table(esw);
+	else if (esw->mode == SRIOV_OFFLOADS)
+		esw_offloads_cleanup(esw, nvports);
+
+	old_mode = esw->mode;
+	esw->mode = SRIOV_NONE;
+
+	if (old_mode == SRIOV_OFFLOADS)
+		mlx5_reload_interface(esw->dev, MLX5_INTERFACE_PROTOCOL_IB);
+}
+
+int mlx5_eswitch_init(struct mlx5_core_dev *dev)
+{
+	int total_vports = MLX5_TOTAL_VPORTS(dev);
+	struct mlx5_eswitch *esw;
+	int vport_num;
+	int err;
+
+	if (!MLX5_VPORT_MANAGER(dev))
+		return 0;
+
+	esw_info(dev,
+		 "Total vports %d, per vport: max uc(%d) max mc(%d)\n",
+		 total_vports,
+		 MLX5_MAX_UC_PER_VPORT(dev),
+		 MLX5_MAX_MC_PER_VPORT(dev));
+
+	esw = kzalloc(sizeof(*esw), GFP_KERNEL);
+	if (!esw)
+		return -ENOMEM;
+
+	esw->dev = dev;
+
+	esw->work_queue = create_singlethread_workqueue("mlx5_esw_wq");
+	if (!esw->work_queue) {
+		err = -ENOMEM;
+		goto abort;
+	}
+
+	esw->vports = kcalloc(total_vports, sizeof(struct mlx5_vport),
+			      GFP_KERNEL);
+	if (!esw->vports) {
+		err = -ENOMEM;
+		goto abort;
+	}
+
+	err = esw_offloads_init_reps(esw);
+	if (err)
+		goto abort;
+
+	hash_init(esw->offloads.encap_tbl);
+	hash_init(esw->offloads.mod_hdr_tbl);
+	mutex_init(&esw->state_lock);
+
+	for (vport_num = 0; vport_num < total_vports; vport_num++) {
+		struct mlx5_vport *vport = &esw->vports[vport_num];
+
+		vport->vport = vport_num;
+		vport->info.link_state = MLX5_VPORT_ADMIN_STATE_AUTO;
+		vport->dev = dev;
+		INIT_WORK(&vport->vport_change_handler,
+			  esw_vport_change_handler);
+	}
+
+	esw->total_vports = total_vports;
+	esw->enabled_vports = 0;
+	esw->mode = SRIOV_NONE;
+	esw->offloads.inline_mode = MLX5_INLINE_MODE_NONE;
+	if (MLX5_CAP_ESW_FLOWTABLE_FDB(dev, encap) &&
+	    MLX5_CAP_ESW_FLOWTABLE_FDB(dev, decap))
+		esw->offloads.encap = DEVLINK_ESWITCH_ENCAP_MODE_BASIC;
+	else
+		esw->offloads.encap = DEVLINK_ESWITCH_ENCAP_MODE_NONE;
+
+	dev->priv.eswitch = esw;
+	return 0;
+abort:
+	if (esw->work_queue)
+		destroy_workqueue(esw->work_queue);
+	esw_offloads_cleanup_reps(esw);
+	kfree(esw->vports);
+	kfree(esw);
+	return err;
+}
+
+void mlx5_eswitch_cleanup(struct mlx5_eswitch *esw)
+{
+	if (!esw || !MLX5_VPORT_MANAGER(esw->dev))
+		return;
+
+	esw_info(esw->dev, "cleanup\n");
+
+	esw->dev->priv.eswitch = NULL;
+	destroy_workqueue(esw->work_queue);
+	esw_offloads_cleanup_reps(esw);
+	kfree(esw->vports);
+	kfree(esw);
+}
+
+void mlx5_eswitch_vport_event(struct mlx5_eswitch *esw, struct mlx5_eqe *eqe)
+{
+	struct mlx5_eqe_vport_change *vc_eqe = &eqe->data.vport_change;
+	u16 vport_num = be16_to_cpu(vc_eqe->vport_num);
+	struct mlx5_vport *vport;
+
+	if (!esw) {
+		pr_warn("MLX5 E-Switch: vport %d got an event while eswitch is not initialized\n",
+			vport_num);
+		return;
+	}
+
+	vport = &esw->vports[vport_num];
+	if (vport->enabled)
+		queue_work(esw->work_queue, &vport->vport_change_handler);
+}
+
+/* Vport Administration */
+#define LEGAL_VPORT(esw, vport) (vport >= 0 && vport < esw->total_vports)
+
+int mlx5_eswitch_set_vport_mac(struct mlx5_eswitch *esw,
+			       int vport, u8 mac[ETH_ALEN])
+{
+	struct mlx5_vport *evport;
+	u64 node_guid;
+	int err = 0;
+
+	if (!esw || !MLX5_CAP_GEN(esw->dev, vport_group_manager))
+		return -EPERM;
+	if (!LEGAL_VPORT(esw, vport) || is_multicast_ether_addr(mac))
+		return -EINVAL;
+
+	mutex_lock(&esw->state_lock);
+	evport = &esw->vports[vport];
+
+	if (evport->info.spoofchk && !is_valid_ether_addr(mac))
+		mlx5_core_warn(esw->dev,
+			       "Set invalid MAC while spoofchk is on, vport(%d)\n",
+			       vport);
+
+	err = mlx5_modify_nic_vport_mac_address(esw->dev, vport, mac);
+	if (err) {
+		mlx5_core_warn(esw->dev,
+			       "Failed to mlx5_modify_nic_vport_mac vport(%d) err=(%d)\n",
+			       vport, err);
+		goto unlock;
+	}
+
+	node_guid_gen_from_mac(&node_guid, mac);
+	err = mlx5_modify_nic_vport_node_guid(esw->dev, vport, node_guid);
+	if (err)
+		mlx5_core_warn(esw->dev,
+			       "Failed to set vport %d node guid, err = %d. RDMA_CM will not function properly for this VF.\n",
+			       vport, err);
+
+	ether_addr_copy(evport->info.mac, mac);
+	evport->info.node_guid = node_guid;
+	if (evport->enabled && esw->mode == SRIOV_LEGACY)
+		err = esw_vport_ingress_config(esw, evport);
+
+unlock:
+	mutex_unlock(&esw->state_lock);
+	return err;
+}
+
+int mlx5_eswitch_set_vport_state(struct mlx5_eswitch *esw,
+				 int vport, int link_state)
+{
+	struct mlx5_vport *evport;
+	int err = 0;
+
+	if (!ESW_ALLOWED(esw))
+		return -EPERM;
+	if (!LEGAL_VPORT(esw, vport))
+		return -EINVAL;
+
+	mutex_lock(&esw->state_lock);
+	evport = &esw->vports[vport];
+
+	err = mlx5_modify_vport_admin_state(esw->dev,
+					    MLX5_VPORT_STATE_OP_MOD_ESW_VPORT,
+					    vport, link_state);
+	if (err) {
+		mlx5_core_warn(esw->dev,
+			       "Failed to set vport %d link state, err = %d",
+			       vport, err);
+		goto unlock;
+	}
+
+	evport->info.link_state = link_state;
+
+unlock:
+	mutex_unlock(&esw->state_lock);
+	return err;
+}
+
+int mlx5_eswitch_get_vport_config(struct mlx5_eswitch *esw,
+				  int vport, struct ifla_vf_info *ivi)
+{
+	struct mlx5_vport *evport;
+
+	if (!esw || !MLX5_CAP_GEN(esw->dev, vport_group_manager))
+		return -EPERM;
+	if (!LEGAL_VPORT(esw, vport))
+		return -EINVAL;
+
+	evport = &esw->vports[vport];
+
+	memset(ivi, 0, sizeof(*ivi));
+	ivi->vf = vport - 1;
+
+	mutex_lock(&esw->state_lock);
+	ether_addr_copy(ivi->mac, evport->info.mac);
+	ivi->linkstate = evport->info.link_state;
+	ivi->vlan = evport->info.vlan;
+	ivi->qos = evport->info.qos;
+	ivi->spoofchk = evport->info.spoofchk;
+	ivi->trusted = evport->info.trusted;
+	ivi->min_tx_rate = evport->info.min_rate;
+	ivi->max_tx_rate = evport->info.max_rate;
+	mutex_unlock(&esw->state_lock);
+
+	return 0;
+}
+
+int __mlx5_eswitch_set_vport_vlan(struct mlx5_eswitch *esw,
+				  int vport, u16 vlan, u8 qos, u8 set_flags)
+{
+	struct mlx5_vport *evport;
+	int err = 0;
+
+	if (!ESW_ALLOWED(esw))
+		return -EPERM;
+	if (!LEGAL_VPORT(esw, vport) || (vlan > 4095) || (qos > 7))
+		return -EINVAL;
+
+	mutex_lock(&esw->state_lock);
+	evport = &esw->vports[vport];
+
+	err = modify_esw_vport_cvlan(esw->dev, vport, vlan, qos, set_flags);
+	if (err)
+		goto unlock;
+
+	evport->info.vlan = vlan;
+	evport->info.qos = qos;
+	if (evport->enabled && esw->mode == SRIOV_LEGACY) {
+		err = esw_vport_ingress_config(esw, evport);
+		if (err)
+			goto unlock;
+		err = esw_vport_egress_config(esw, evport);
+	}
+
+unlock:
+	mutex_unlock(&esw->state_lock);
+	return err;
+}
+
+int mlx5_eswitch_set_vport_vlan(struct mlx5_eswitch *esw,
+				int vport, u16 vlan, u8 qos)
+{
+	u8 set_flags = 0;
+
+	if (vlan || qos)
+		set_flags = SET_VLAN_STRIP | SET_VLAN_INSERT;
+
+	return __mlx5_eswitch_set_vport_vlan(esw, vport, vlan, qos, set_flags);
+}
+
+int mlx5_eswitch_set_vport_spoofchk(struct mlx5_eswitch *esw,
+				    int vport, bool spoofchk)
+{
+	struct mlx5_vport *evport;
+	bool pschk;
+	int err = 0;
+
+	if (!ESW_ALLOWED(esw))
+		return -EPERM;
+	if (!LEGAL_VPORT(esw, vport))
+		return -EINVAL;
+
+	mutex_lock(&esw->state_lock);
+	evport = &esw->vports[vport];
+	pschk = evport->info.spoofchk;
+	evport->info.spoofchk = spoofchk;
+	if (pschk && !is_valid_ether_addr(evport->info.mac))
+		mlx5_core_warn(esw->dev,
+			       "Spoofchk in set while MAC is invalid, vport(%d)\n",
+			       evport->vport);
+	if (evport->enabled && esw->mode == SRIOV_LEGACY)
+		err = esw_vport_ingress_config(esw, evport);
+	if (err)
+		evport->info.spoofchk = pschk;
+	mutex_unlock(&esw->state_lock);
+
+	return err;
+}
+
+int mlx5_eswitch_set_vport_trust(struct mlx5_eswitch *esw,
+				 int vport, bool setting)
+{
+	struct mlx5_vport *evport;
+
+	if (!ESW_ALLOWED(esw))
+		return -EPERM;
+	if (!LEGAL_VPORT(esw, vport))
+		return -EINVAL;
+
+	mutex_lock(&esw->state_lock);
+	evport = &esw->vports[vport];
+	evport->info.trusted = setting;
+	if (evport->enabled)
+		esw_vport_change_handle_locked(evport);
+	mutex_unlock(&esw->state_lock);
+
+	return 0;
+}
+
+static u32 calculate_vports_min_rate_divider(struct mlx5_eswitch *esw)
+{
+	u32 fw_max_bw_share = MLX5_CAP_QOS(esw->dev, max_tsar_bw_share);
+	struct mlx5_vport *evport;
+	u32 max_guarantee = 0;
+	int i;
+
+	for (i = 0; i < esw->total_vports; i++) {
+		evport = &esw->vports[i];
+		if (!evport->enabled || evport->info.min_rate < max_guarantee)
+			continue;
+		max_guarantee = evport->info.min_rate;
+	}
+
+	if (max_guarantee)
+		return max_t(u32, max_guarantee / fw_max_bw_share, 1);
+	return 0;
+}
+
+static int normalize_vports_min_rate(struct mlx5_eswitch *esw)
+{
+	u32 fw_max_bw_share = MLX5_CAP_QOS(esw->dev, max_tsar_bw_share);
+	u32 divider = calculate_vports_min_rate_divider(esw);
+	struct mlx5_vport *evport;
+	u32 vport_max_rate;
+	u32 vport_min_rate;
+	u32 bw_share;
+	int err;
+	int i;
+
+	for (i = 0; i < esw->total_vports; i++) {
+		evport = &esw->vports[i];
+		if (!evport->enabled)
+			continue;
+		vport_min_rate = evport->info.min_rate;
+		vport_max_rate = evport->info.max_rate;
+		bw_share = 0;
+
+		if (divider)
+			bw_share = MLX5_RATE_TO_BW_SHARE(vport_min_rate,
+							 divider,
+							 fw_max_bw_share);
+
+		if (bw_share == evport->qos.bw_share)
+			continue;
+
+		err = esw_vport_qos_config(esw, i, vport_max_rate,
+					   bw_share);
+		if (!err)
+			evport->qos.bw_share = bw_share;
+		else
+			return err;
+	}
+
+	return 0;
+}
+
+int mlx5_eswitch_set_vport_rate(struct mlx5_eswitch *esw, int vport,
+				u32 max_rate, u32 min_rate)
+{
+	struct mlx5_vport *evport;
+	u32 fw_max_bw_share;
+	u32 previous_min_rate;
+	bool min_rate_supported;
+	bool max_rate_supported;
+	int err = 0;
+
+	if (!ESW_ALLOWED(esw))
+		return -EPERM;
+	if (!LEGAL_VPORT(esw, vport))
+		return -EINVAL;
+
+	fw_max_bw_share = MLX5_CAP_QOS(esw->dev, max_tsar_bw_share);
+	min_rate_supported = MLX5_CAP_QOS(esw->dev, esw_bw_share) &&
+				fw_max_bw_share >= MLX5_MIN_BW_SHARE;
+	max_rate_supported = MLX5_CAP_QOS(esw->dev, esw_rate_limit);
+
+	if ((min_rate && !min_rate_supported) || (max_rate && !max_rate_supported))
+		return -EOPNOTSUPP;
+
+	mutex_lock(&esw->state_lock);
+	evport = &esw->vports[vport];
+
+	if (min_rate == evport->info.min_rate)
+		goto set_max_rate;
+
+	previous_min_rate = evport->info.min_rate;
+	evport->info.min_rate = min_rate;
+	err = normalize_vports_min_rate(esw);
+	if (err) {
+		evport->info.min_rate = previous_min_rate;
+		goto unlock;
+	}
+
+set_max_rate:
+	if (max_rate == evport->info.max_rate)
+		goto unlock;
+
+	err = esw_vport_qos_config(esw, vport, max_rate, evport->qos.bw_share);
+	if (!err)
+		evport->info.max_rate = max_rate;
+
+unlock:
+	mutex_unlock(&esw->state_lock);
+	return err;
+}
+
+static int mlx5_eswitch_query_vport_drop_stats(struct mlx5_core_dev *dev,
+					       int vport_idx,
+					       struct mlx5_vport_drop_stats *stats)
+{
+	struct mlx5_eswitch *esw = dev->priv.eswitch;
+	struct mlx5_vport *vport = &esw->vports[vport_idx];
+	u64 rx_discard_vport_down, tx_discard_vport_down;
+	u64 bytes = 0;
+	int err = 0;
+
+	if (!vport->enabled || esw->mode != SRIOV_LEGACY)
+		return 0;
+
+	if (vport->egress.drop_counter)
+		mlx5_fc_query(dev, vport->egress.drop_counter,
+			      &stats->rx_dropped, &bytes);
+
+	if (vport->ingress.drop_counter)
+		mlx5_fc_query(dev, vport->ingress.drop_counter,
+			      &stats->tx_dropped, &bytes);
+
+	if (!MLX5_CAP_GEN(dev, receive_discard_vport_down) &&
+	    !MLX5_CAP_GEN(dev, transmit_discard_vport_down))
+		return 0;
+
+	err = mlx5_query_vport_down_stats(dev, vport_idx,
+					  &rx_discard_vport_down,
+					  &tx_discard_vport_down);
+	if (err)
+		return err;
+
+	if (MLX5_CAP_GEN(dev, receive_discard_vport_down))
+		stats->rx_dropped += rx_discard_vport_down;
+	if (MLX5_CAP_GEN(dev, transmit_discard_vport_down))
+		stats->tx_dropped += tx_discard_vport_down;
+
+	return 0;
+}
+
+int mlx5_eswitch_get_vport_stats(struct mlx5_eswitch *esw,
+				 int vport,
+				 struct ifla_vf_stats *vf_stats)
+{
+	int outlen = MLX5_ST_SZ_BYTES(query_vport_counter_out);
+	u32 in[MLX5_ST_SZ_DW(query_vport_counter_in)] = {0};
+	struct mlx5_vport_drop_stats stats = {0};
+	int err = 0;
+	u32 *out;
+
+	if (!ESW_ALLOWED(esw))
+		return -EPERM;
+	if (!LEGAL_VPORT(esw, vport))
+		return -EINVAL;
+
+	out = kvzalloc(outlen, GFP_KERNEL);
+	if (!out)
+		return -ENOMEM;
+
+	MLX5_SET(query_vport_counter_in, in, opcode,
+		 MLX5_CMD_OP_QUERY_VPORT_COUNTER);
+	MLX5_SET(query_vport_counter_in, in, op_mod, 0);
+	MLX5_SET(query_vport_counter_in, in, vport_number, vport);
+	if (vport)
+		MLX5_SET(query_vport_counter_in, in, other_vport, 1);
+
+	memset(out, 0, outlen);
+	err = mlx5_cmd_exec(esw->dev, in, sizeof(in), out, outlen);
+	if (err)
+		goto free_out;
+
+	#define MLX5_GET_CTR(p, x) \
+		MLX5_GET64(query_vport_counter_out, p, x)
+
+	memset(vf_stats, 0, sizeof(*vf_stats));
+	vf_stats->rx_packets =
+		MLX5_GET_CTR(out, received_eth_unicast.packets) +
+		MLX5_GET_CTR(out, received_ib_unicast.packets) +
+		MLX5_GET_CTR(out, received_eth_multicast.packets) +
+		MLX5_GET_CTR(out, received_ib_multicast.packets) +
+		MLX5_GET_CTR(out, received_eth_broadcast.packets);
+
+	vf_stats->rx_bytes =
+		MLX5_GET_CTR(out, received_eth_unicast.octets) +
+		MLX5_GET_CTR(out, received_ib_unicast.octets) +
+		MLX5_GET_CTR(out, received_eth_multicast.octets) +
+		MLX5_GET_CTR(out, received_ib_multicast.octets) +
+		MLX5_GET_CTR(out, received_eth_broadcast.octets);
+
+	vf_stats->tx_packets =
+		MLX5_GET_CTR(out, transmitted_eth_unicast.packets) +
+		MLX5_GET_CTR(out, transmitted_ib_unicast.packets) +
+		MLX5_GET_CTR(out, transmitted_eth_multicast.packets) +
+		MLX5_GET_CTR(out, transmitted_ib_multicast.packets) +
+		MLX5_GET_CTR(out, transmitted_eth_broadcast.packets);
+
+	vf_stats->tx_bytes =
+		MLX5_GET_CTR(out, transmitted_eth_unicast.octets) +
+		MLX5_GET_CTR(out, transmitted_ib_unicast.octets) +
+		MLX5_GET_CTR(out, transmitted_eth_multicast.octets) +
+		MLX5_GET_CTR(out, transmitted_ib_multicast.octets) +
+		MLX5_GET_CTR(out, transmitted_eth_broadcast.octets);
+
+	vf_stats->multicast =
+		MLX5_GET_CTR(out, received_eth_multicast.packets) +
+		MLX5_GET_CTR(out, received_ib_multicast.packets);
+
+	vf_stats->broadcast =
+		MLX5_GET_CTR(out, received_eth_broadcast.packets);
+
+	err = mlx5_eswitch_query_vport_drop_stats(esw->dev, vport, &stats);
+	if (err)
+		goto free_out;
+	vf_stats->rx_dropped = stats.rx_dropped;
+	vf_stats->tx_dropped = stats.tx_dropped;
+
+free_out:
+	kvfree(out);
+	return err;
+}
+
+u8 mlx5_eswitch_mode(struct mlx5_eswitch *esw)
+{
+	return ESW_ALLOWED(esw) ? esw->mode : SRIOV_NONE;
+}
+EXPORT_SYMBOL_GPL(mlx5_eswitch_mode);
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/eswitch.h b/drivers/net/ethernet/mellanox/mlx5/core/eswitch.h
new file mode 100644
index 000000000..c17bfcab5
--- /dev/null
+++ b/drivers/net/ethernet/mellanox/mlx5/core/eswitch.h
@@ -0,0 +1,319 @@
+/*
+ * Copyright (c) 2015, Mellanox Technologies, Ltd.  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 __MLX5_ESWITCH_H__
+#define __MLX5_ESWITCH_H__
+
+#include <linux/if_ether.h>
+#include <linux/if_link.h>
+#include <net/devlink.h>
+#include <linux/mlx5/device.h>
+#include <linux/mlx5/eswitch.h>
+#include <linux/mlx5/fs.h>
+#include "lib/mpfs.h"
+
+#ifdef CONFIG_MLX5_ESWITCH
+
+#define MLX5_MAX_UC_PER_VPORT(dev) \
+	(1 << MLX5_CAP_GEN(dev, log_max_current_uc_list))
+
+#define MLX5_MAX_MC_PER_VPORT(dev) \
+	(1 << MLX5_CAP_GEN(dev, log_max_current_mc_list))
+
+#define FDB_UPLINK_VPORT 0xffff
+
+#define MLX5_MIN_BW_SHARE 1
+
+#define MLX5_RATE_TO_BW_SHARE(rate, divider, limit) \
+	min_t(u32, max_t(u32, (rate) / (divider), MLX5_MIN_BW_SHARE), limit)
+
+#define mlx5_esw_has_fwd_fdb(dev) \
+	MLX5_CAP_ESW_FLOWTABLE(dev, fdb_multi_path_to_table)
+
+struct vport_ingress {
+	struct mlx5_flow_table *acl;
+	struct mlx5_flow_group *allow_untagged_spoofchk_grp;
+	struct mlx5_flow_group *allow_spoofchk_only_grp;
+	struct mlx5_flow_group *allow_untagged_only_grp;
+	struct mlx5_flow_group *drop_grp;
+	struct mlx5_flow_handle  *allow_rule;
+	struct mlx5_flow_handle  *drop_rule;
+	struct mlx5_fc           *drop_counter;
+};
+
+struct vport_egress {
+	struct mlx5_flow_table *acl;
+	struct mlx5_flow_group *allowed_vlans_grp;
+	struct mlx5_flow_group *drop_grp;
+	struct mlx5_flow_handle  *allowed_vlan;
+	struct mlx5_flow_handle  *drop_rule;
+	struct mlx5_fc           *drop_counter;
+};
+
+struct mlx5_vport_drop_stats {
+	u64 rx_dropped;
+	u64 tx_dropped;
+};
+
+struct mlx5_vport_info {
+	u8                      mac[ETH_ALEN];
+	u16                     vlan;
+	u8                      qos;
+	u64                     node_guid;
+	int                     link_state;
+	u32                     min_rate;
+	u32                     max_rate;
+	bool                    spoofchk;
+	bool                    trusted;
+};
+
+struct mlx5_vport {
+	struct mlx5_core_dev    *dev;
+	int                     vport;
+	struct hlist_head       uc_list[MLX5_L2_ADDR_HASH_SIZE];
+	struct hlist_head       mc_list[MLX5_L2_ADDR_HASH_SIZE];
+	struct mlx5_flow_handle *promisc_rule;
+	struct mlx5_flow_handle *allmulti_rule;
+	struct work_struct      vport_change_handler;
+
+	struct vport_ingress    ingress;
+	struct vport_egress     egress;
+
+	struct mlx5_vport_info  info;
+
+	struct {
+		bool            enabled;
+		u32             esw_tsar_ix;
+		u32             bw_share;
+	} qos;
+
+	bool                    enabled;
+	u16                     enabled_events;
+};
+
+struct mlx5_eswitch_fdb {
+	union {
+		struct legacy_fdb {
+			struct mlx5_flow_table *fdb;
+			struct mlx5_flow_group *addr_grp;
+			struct mlx5_flow_group *allmulti_grp;
+			struct mlx5_flow_group *promisc_grp;
+		} legacy;
+
+		struct offloads_fdb {
+			struct mlx5_flow_table *fast_fdb;
+			struct mlx5_flow_table *fwd_fdb;
+			struct mlx5_flow_table *slow_fdb;
+			struct mlx5_flow_group *send_to_vport_grp;
+			struct mlx5_flow_group *miss_grp;
+			struct mlx5_flow_handle *miss_rule_uni;
+			struct mlx5_flow_handle *miss_rule_multi;
+			int vlan_push_pop_refcount;
+		} offloads;
+	};
+};
+
+struct mlx5_esw_offload {
+	struct mlx5_flow_table *ft_offloads;
+	struct mlx5_flow_group *vport_rx_group;
+	struct mlx5_eswitch_rep *vport_reps;
+	DECLARE_HASHTABLE(encap_tbl, 8);
+	DECLARE_HASHTABLE(mod_hdr_tbl, 8);
+	u8 inline_mode;
+	u64 num_flows;
+	u8 encap;
+};
+
+/* E-Switch MC FDB table hash node */
+struct esw_mc_addr { /* SRIOV only */
+	struct l2addr_node     node;
+	struct mlx5_flow_handle *uplink_rule; /* Forward to uplink rule */
+	u32                    refcnt;
+};
+
+struct mlx5_eswitch {
+	struct mlx5_core_dev    *dev;
+	struct mlx5_eswitch_fdb fdb_table;
+	struct hlist_head       mc_table[MLX5_L2_ADDR_HASH_SIZE];
+	struct workqueue_struct *work_queue;
+	struct mlx5_vport       *vports;
+	int                     total_vports;
+	int                     enabled_vports;
+	/* Synchronize between vport change events
+	 * and async SRIOV admin state changes
+	 */
+	struct mutex            state_lock;
+	struct esw_mc_addr	mc_promisc;
+
+	struct {
+		bool            enabled;
+		u32             root_tsar_id;
+	} qos;
+
+	struct mlx5_esw_offload offloads;
+	int                     mode;
+};
+
+void esw_offloads_cleanup(struct mlx5_eswitch *esw, int nvports);
+int esw_offloads_init(struct mlx5_eswitch *esw, int nvports);
+void esw_offloads_cleanup_reps(struct mlx5_eswitch *esw);
+int esw_offloads_init_reps(struct mlx5_eswitch *esw);
+
+/* E-Switch API */
+int mlx5_eswitch_init(struct mlx5_core_dev *dev);
+void mlx5_eswitch_cleanup(struct mlx5_eswitch *esw);
+void mlx5_eswitch_vport_event(struct mlx5_eswitch *esw, struct mlx5_eqe *eqe);
+int mlx5_eswitch_enable_sriov(struct mlx5_eswitch *esw, int nvfs, int mode);
+void mlx5_eswitch_disable_sriov(struct mlx5_eswitch *esw);
+int mlx5_eswitch_set_vport_mac(struct mlx5_eswitch *esw,
+			       int vport, u8 mac[ETH_ALEN]);
+int mlx5_eswitch_set_vport_state(struct mlx5_eswitch *esw,
+				 int vport, int link_state);
+int mlx5_eswitch_set_vport_vlan(struct mlx5_eswitch *esw,
+				int vport, u16 vlan, u8 qos);
+int mlx5_eswitch_set_vport_spoofchk(struct mlx5_eswitch *esw,
+				    int vport, bool spoofchk);
+int mlx5_eswitch_set_vport_trust(struct mlx5_eswitch *esw,
+				 int vport_num, bool setting);
+int mlx5_eswitch_set_vport_rate(struct mlx5_eswitch *esw, int vport,
+				u32 max_rate, u32 min_rate);
+int mlx5_eswitch_get_vport_config(struct mlx5_eswitch *esw,
+				  int vport, struct ifla_vf_info *ivi);
+int mlx5_eswitch_get_vport_stats(struct mlx5_eswitch *esw,
+				 int vport,
+				 struct ifla_vf_stats *vf_stats);
+void mlx5_eswitch_del_send_to_vport_rule(struct mlx5_flow_handle *rule);
+
+struct mlx5_flow_spec;
+struct mlx5_esw_flow_attr;
+
+struct mlx5_flow_handle *
+mlx5_eswitch_add_offloaded_rule(struct mlx5_eswitch *esw,
+				struct mlx5_flow_spec *spec,
+				struct mlx5_esw_flow_attr *attr);
+struct mlx5_flow_handle *
+mlx5_eswitch_add_fwd_rule(struct mlx5_eswitch *esw,
+			  struct mlx5_flow_spec *spec,
+			  struct mlx5_esw_flow_attr *attr);
+void
+mlx5_eswitch_del_offloaded_rule(struct mlx5_eswitch *esw,
+				struct mlx5_flow_handle *rule,
+				struct mlx5_esw_flow_attr *attr);
+
+struct mlx5_flow_handle *
+mlx5_eswitch_create_vport_rx_rule(struct mlx5_eswitch *esw, int vport, u32 tirn);
+
+enum {
+	SET_VLAN_STRIP	= BIT(0),
+	SET_VLAN_INSERT	= BIT(1)
+};
+
+enum mlx5_flow_match_level {
+	MLX5_MATCH_NONE	= MLX5_INLINE_MODE_NONE,
+	MLX5_MATCH_L2	= MLX5_INLINE_MODE_L2,
+	MLX5_MATCH_L3	= MLX5_INLINE_MODE_IP,
+	MLX5_MATCH_L4	= MLX5_INLINE_MODE_TCP_UDP,
+};
+
+/* current maximum for flow based vport multicasting */
+#define MLX5_MAX_FLOW_FWD_VPORTS 2
+
+struct mlx5_esw_flow_attr {
+	struct mlx5_eswitch_rep *in_rep;
+	struct mlx5_eswitch_rep *out_rep[MLX5_MAX_FLOW_FWD_VPORTS];
+	struct mlx5_core_dev	*out_mdev[MLX5_MAX_FLOW_FWD_VPORTS];
+	struct mlx5_core_dev	*in_mdev;
+
+	int mirror_count;
+	int out_count;
+
+	int	action;
+	__be16	vlan_proto[MLX5_FS_VLAN_DEPTH];
+	u16	vlan_vid[MLX5_FS_VLAN_DEPTH];
+	u8	vlan_prio[MLX5_FS_VLAN_DEPTH];
+	u8	total_vlan;
+	bool	vlan_handled;
+	u32	encap_id;
+	u32	mod_hdr_id;
+	u8	match_level;
+	struct mlx5e_tc_flow_parse_attr *parse_attr;
+};
+
+int mlx5_devlink_eswitch_mode_set(struct devlink *devlink, u16 mode);
+int mlx5_devlink_eswitch_mode_get(struct devlink *devlink, u16 *mode);
+int mlx5_devlink_eswitch_inline_mode_set(struct devlink *devlink, u8 mode);
+int mlx5_devlink_eswitch_inline_mode_get(struct devlink *devlink, u8 *mode);
+int mlx5_eswitch_inline_mode_get(struct mlx5_eswitch *esw, int nvfs, u8 *mode);
+int mlx5_devlink_eswitch_encap_mode_set(struct devlink *devlink, u8 encap);
+int mlx5_devlink_eswitch_encap_mode_get(struct devlink *devlink, u8 *encap);
+void *mlx5_eswitch_get_uplink_priv(struct mlx5_eswitch *esw, u8 rep_type);
+
+int mlx5_eswitch_add_vlan_action(struct mlx5_eswitch *esw,
+				 struct mlx5_esw_flow_attr *attr);
+int mlx5_eswitch_del_vlan_action(struct mlx5_eswitch *esw,
+				 struct mlx5_esw_flow_attr *attr);
+int __mlx5_eswitch_set_vport_vlan(struct mlx5_eswitch *esw,
+				  int vport, u16 vlan, u8 qos, u8 set_flags);
+
+static inline bool mlx5_eswitch_vlan_actions_supported(struct mlx5_core_dev *dev,
+						       u8 vlan_depth)
+{
+	bool ret = MLX5_CAP_ESW_FLOWTABLE_FDB(dev, pop_vlan) &&
+		   MLX5_CAP_ESW_FLOWTABLE_FDB(dev, push_vlan);
+
+	if (vlan_depth == 1)
+		return ret;
+
+	return  ret && MLX5_CAP_ESW_FLOWTABLE_FDB(dev, pop_vlan_2) &&
+		MLX5_CAP_ESW_FLOWTABLE_FDB(dev, push_vlan_2);
+}
+
+#define MLX5_DEBUG_ESWITCH_MASK BIT(3)
+
+#define esw_info(dev, format, ...)				\
+	pr_info("(%s): E-Switch: " format, (dev)->priv.name, ##__VA_ARGS__)
+
+#define esw_warn(dev, format, ...)				\
+	pr_warn("(%s): E-Switch: " format, (dev)->priv.name, ##__VA_ARGS__)
+
+#define esw_debug(dev, format, ...)				\
+	mlx5_core_dbg_mask(dev, MLX5_DEBUG_ESWITCH_MASK, format, ##__VA_ARGS__)
+#else  /* CONFIG_MLX5_ESWITCH */
+/* eswitch API stubs */
+static inline int  mlx5_eswitch_init(struct mlx5_core_dev *dev) { return 0; }
+static inline void mlx5_eswitch_cleanup(struct mlx5_eswitch *esw) {}
+static inline void mlx5_eswitch_vport_event(struct mlx5_eswitch *esw, struct mlx5_eqe *eqe) {}
+static inline int  mlx5_eswitch_enable_sriov(struct mlx5_eswitch *esw, int nvfs, int mode) { return 0; }
+static inline void mlx5_eswitch_disable_sriov(struct mlx5_eswitch *esw) {}
+#endif /* CONFIG_MLX5_ESWITCH */
+
+#endif /* __MLX5_ESWITCH_H__ */
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c b/drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c
new file mode 100644
index 000000000..3028e8d90
--- /dev/null
+++ b/drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c
@@ -0,0 +1,1368 @@
+/*
+ * Copyright (c) 2016, 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/etherdevice.h>
+#include <linux/mlx5/driver.h>
+#include <linux/mlx5/mlx5_ifc.h>
+#include <linux/mlx5/vport.h>
+#include <linux/mlx5/fs.h>
+#include "mlx5_core.h"
+#include "eswitch.h"
+
+enum {
+	FDB_FAST_PATH = 0,
+	FDB_SLOW_PATH
+};
+
+struct mlx5_flow_handle *
+mlx5_eswitch_add_offloaded_rule(struct mlx5_eswitch *esw,
+				struct mlx5_flow_spec *spec,
+				struct mlx5_esw_flow_attr *attr)
+{
+	struct mlx5_flow_destination dest[MLX5_MAX_FLOW_FWD_VPORTS + 1] = {};
+	struct mlx5_flow_act flow_act = {0};
+	struct mlx5_flow_table *ft = NULL;
+	struct mlx5_fc *counter = NULL;
+	struct mlx5_flow_handle *rule;
+	int j, i = 0;
+	void *misc;
+
+	if (esw->mode != SRIOV_OFFLOADS)
+		return ERR_PTR(-EOPNOTSUPP);
+
+	if (attr->mirror_count)
+		ft = esw->fdb_table.offloads.fwd_fdb;
+	else
+		ft = esw->fdb_table.offloads.fast_fdb;
+
+	flow_act.action = attr->action;
+	/* if per flow vlan pop/push is emulated, don't set that into the firmware */
+	if (!mlx5_eswitch_vlan_actions_supported(esw->dev, 1))
+		flow_act.action &= ~(MLX5_FLOW_CONTEXT_ACTION_VLAN_PUSH |
+				     MLX5_FLOW_CONTEXT_ACTION_VLAN_POP);
+	else if (flow_act.action & MLX5_FLOW_CONTEXT_ACTION_VLAN_PUSH) {
+		flow_act.vlan[0].ethtype = ntohs(attr->vlan_proto[0]);
+		flow_act.vlan[0].vid = attr->vlan_vid[0];
+		flow_act.vlan[0].prio = attr->vlan_prio[0];
+		if (flow_act.action & MLX5_FLOW_CONTEXT_ACTION_VLAN_PUSH_2) {
+			flow_act.vlan[1].ethtype = ntohs(attr->vlan_proto[1]);
+			flow_act.vlan[1].vid = attr->vlan_vid[1];
+			flow_act.vlan[1].prio = attr->vlan_prio[1];
+		}
+	}
+
+	if (flow_act.action & MLX5_FLOW_CONTEXT_ACTION_FWD_DEST) {
+		for (j = attr->mirror_count; j < attr->out_count; j++) {
+			dest[i].type = MLX5_FLOW_DESTINATION_TYPE_VPORT;
+			dest[i].vport.num = attr->out_rep[j]->vport;
+			dest[i].vport.vhca_id =
+				MLX5_CAP_GEN(attr->out_mdev[j], vhca_id);
+			dest[i].vport.vhca_id_valid = !!MLX5_CAP_ESW(esw->dev, merged_eswitch);
+			i++;
+		}
+	}
+	if (flow_act.action & MLX5_FLOW_CONTEXT_ACTION_COUNT) {
+		counter = mlx5_fc_create(esw->dev, true);
+		if (IS_ERR(counter)) {
+			rule = ERR_CAST(counter);
+			goto err_counter_alloc;
+		}
+		dest[i].type = MLX5_FLOW_DESTINATION_TYPE_COUNTER;
+		dest[i].counter = counter;
+		i++;
+	}
+
+	misc = MLX5_ADDR_OF(fte_match_param, spec->match_value, misc_parameters);
+	MLX5_SET(fte_match_set_misc, misc, source_port, attr->in_rep->vport);
+
+	if (MLX5_CAP_ESW(esw->dev, merged_eswitch))
+		MLX5_SET(fte_match_set_misc, misc,
+			 source_eswitch_owner_vhca_id,
+			 MLX5_CAP_GEN(attr->in_mdev, vhca_id));
+
+	misc = MLX5_ADDR_OF(fte_match_param, spec->match_criteria, misc_parameters);
+	MLX5_SET_TO_ONES(fte_match_set_misc, misc, source_port);
+	if (MLX5_CAP_ESW(esw->dev, merged_eswitch))
+		MLX5_SET_TO_ONES(fte_match_set_misc, misc,
+				 source_eswitch_owner_vhca_id);
+
+	if (attr->match_level == MLX5_MATCH_NONE)
+		spec->match_criteria_enable = MLX5_MATCH_MISC_PARAMETERS;
+	else
+		spec->match_criteria_enable = MLX5_MATCH_OUTER_HEADERS |
+					      MLX5_MATCH_MISC_PARAMETERS;
+
+	if (flow_act.action & MLX5_FLOW_CONTEXT_ACTION_DECAP)
+		spec->match_criteria_enable |= MLX5_MATCH_INNER_HEADERS;
+
+	if (flow_act.action & MLX5_FLOW_CONTEXT_ACTION_MOD_HDR)
+		flow_act.modify_id = attr->mod_hdr_id;
+
+	if (flow_act.action & MLX5_FLOW_CONTEXT_ACTION_ENCAP)
+		flow_act.encap_id = attr->encap_id;
+
+	rule = mlx5_add_flow_rules(ft, spec, &flow_act, dest, i);
+	if (IS_ERR(rule))
+		goto err_add_rule;
+	else
+		esw->offloads.num_flows++;
+
+	return rule;
+
+err_add_rule:
+	mlx5_fc_destroy(esw->dev, counter);
+err_counter_alloc:
+	return rule;
+}
+
+struct mlx5_flow_handle *
+mlx5_eswitch_add_fwd_rule(struct mlx5_eswitch *esw,
+			  struct mlx5_flow_spec *spec,
+			  struct mlx5_esw_flow_attr *attr)
+{
+	struct mlx5_flow_destination dest[MLX5_MAX_FLOW_FWD_VPORTS + 1] = {};
+	struct mlx5_flow_act flow_act = {0};
+	struct mlx5_flow_handle *rule;
+	void *misc;
+	int i;
+
+	flow_act.action = MLX5_FLOW_CONTEXT_ACTION_FWD_DEST;
+	for (i = 0; i < attr->mirror_count; i++) {
+		dest[i].type = MLX5_FLOW_DESTINATION_TYPE_VPORT;
+		dest[i].vport.num = attr->out_rep[i]->vport;
+		dest[i].vport.vhca_id =
+			MLX5_CAP_GEN(attr->out_mdev[i], vhca_id);
+		dest[i].vport.vhca_id_valid = !!MLX5_CAP_ESW(esw->dev, merged_eswitch);
+	}
+	dest[i].type = MLX5_FLOW_DESTINATION_TYPE_FLOW_TABLE;
+	dest[i].ft = esw->fdb_table.offloads.fwd_fdb,
+	i++;
+
+	misc = MLX5_ADDR_OF(fte_match_param, spec->match_value, misc_parameters);
+	MLX5_SET(fte_match_set_misc, misc, source_port, attr->in_rep->vport);
+
+	if (MLX5_CAP_ESW(esw->dev, merged_eswitch))
+		MLX5_SET(fte_match_set_misc, misc,
+			 source_eswitch_owner_vhca_id,
+			 MLX5_CAP_GEN(attr->in_mdev, vhca_id));
+
+	misc = MLX5_ADDR_OF(fte_match_param, spec->match_criteria, misc_parameters);
+	MLX5_SET_TO_ONES(fte_match_set_misc, misc, source_port);
+	if (MLX5_CAP_ESW(esw->dev, merged_eswitch))
+		MLX5_SET_TO_ONES(fte_match_set_misc, misc,
+				 source_eswitch_owner_vhca_id);
+
+	if (attr->match_level == MLX5_MATCH_NONE)
+		spec->match_criteria_enable = MLX5_MATCH_MISC_PARAMETERS;
+	else
+		spec->match_criteria_enable = MLX5_MATCH_OUTER_HEADERS |
+					      MLX5_MATCH_MISC_PARAMETERS;
+
+	rule = mlx5_add_flow_rules(esw->fdb_table.offloads.fast_fdb, spec, &flow_act, dest, i);
+
+	if (!IS_ERR(rule))
+		esw->offloads.num_flows++;
+
+	return rule;
+}
+
+void
+mlx5_eswitch_del_offloaded_rule(struct mlx5_eswitch *esw,
+				struct mlx5_flow_handle *rule,
+				struct mlx5_esw_flow_attr *attr)
+{
+	struct mlx5_fc *counter = NULL;
+
+	counter = mlx5_flow_rule_counter(rule);
+	mlx5_del_flow_rules(rule);
+	mlx5_fc_destroy(esw->dev, counter);
+	esw->offloads.num_flows--;
+}
+
+static int esw_set_global_vlan_pop(struct mlx5_eswitch *esw, u8 val)
+{
+	struct mlx5_eswitch_rep *rep;
+	int vf_vport, err = 0;
+
+	esw_debug(esw->dev, "%s applying global %s policy\n", __func__, val ? "pop" : "none");
+	for (vf_vport = 1; vf_vport < esw->enabled_vports; vf_vport++) {
+		rep = &esw->offloads.vport_reps[vf_vport];
+		if (!rep->rep_if[REP_ETH].valid)
+			continue;
+
+		err = __mlx5_eswitch_set_vport_vlan(esw, rep->vport, 0, 0, val);
+		if (err)
+			goto out;
+	}
+
+out:
+	return err;
+}
+
+static struct mlx5_eswitch_rep *
+esw_vlan_action_get_vport(struct mlx5_esw_flow_attr *attr, bool push, bool pop)
+{
+	struct mlx5_eswitch_rep *in_rep, *out_rep, *vport = NULL;
+
+	in_rep  = attr->in_rep;
+	out_rep = attr->out_rep[0];
+
+	if (push)
+		vport = in_rep;
+	else if (pop)
+		vport = out_rep;
+	else
+		vport = in_rep;
+
+	return vport;
+}
+
+static int esw_add_vlan_action_check(struct mlx5_esw_flow_attr *attr,
+				     bool push, bool pop, bool fwd)
+{
+	struct mlx5_eswitch_rep *in_rep, *out_rep;
+
+	if ((push || pop) && !fwd)
+		goto out_notsupp;
+
+	in_rep  = attr->in_rep;
+	out_rep = attr->out_rep[0];
+
+	if (push && in_rep->vport == FDB_UPLINK_VPORT)
+		goto out_notsupp;
+
+	if (pop && out_rep->vport == FDB_UPLINK_VPORT)
+		goto out_notsupp;
+
+	/* vport has vlan push configured, can't offload VF --> wire rules w.o it */
+	if (!push && !pop && fwd)
+		if (in_rep->vlan && out_rep->vport == FDB_UPLINK_VPORT)
+			goto out_notsupp;
+
+	/* protects against (1) setting rules with different vlans to push and
+	 * (2) setting rules w.o vlans (attr->vlan = 0) && w. vlans to push (!= 0)
+	 */
+	if (push && in_rep->vlan_refcount && (in_rep->vlan != attr->vlan_vid[0]))
+		goto out_notsupp;
+
+	return 0;
+
+out_notsupp:
+	return -EOPNOTSUPP;
+}
+
+int mlx5_eswitch_add_vlan_action(struct mlx5_eswitch *esw,
+				 struct mlx5_esw_flow_attr *attr)
+{
+	struct offloads_fdb *offloads = &esw->fdb_table.offloads;
+	struct mlx5_eswitch_rep *vport = NULL;
+	bool push, pop, fwd;
+	int err = 0;
+
+	/* nop if we're on the vlan push/pop non emulation mode */
+	if (mlx5_eswitch_vlan_actions_supported(esw->dev, 1))
+		return 0;
+
+	push = !!(attr->action & MLX5_FLOW_CONTEXT_ACTION_VLAN_PUSH);
+	pop  = !!(attr->action & MLX5_FLOW_CONTEXT_ACTION_VLAN_POP);
+	fwd  = !!(attr->action & MLX5_FLOW_CONTEXT_ACTION_FWD_DEST);
+
+	err = esw_add_vlan_action_check(attr, push, pop, fwd);
+	if (err)
+		return err;
+
+	attr->vlan_handled = false;
+
+	vport = esw_vlan_action_get_vport(attr, push, pop);
+
+	if (!push && !pop && fwd) {
+		/* tracks VF --> wire rules without vlan push action */
+		if (attr->out_rep[0]->vport == FDB_UPLINK_VPORT) {
+			vport->vlan_refcount++;
+			attr->vlan_handled = true;
+		}
+
+		return 0;
+	}
+
+	if (!push && !pop)
+		return 0;
+
+	if (!(offloads->vlan_push_pop_refcount)) {
+		/* it's the 1st vlan rule, apply global vlan pop policy */
+		err = esw_set_global_vlan_pop(esw, SET_VLAN_STRIP);
+		if (err)
+			goto out;
+	}
+	offloads->vlan_push_pop_refcount++;
+
+	if (push) {
+		if (vport->vlan_refcount)
+			goto skip_set_push;
+
+		err = __mlx5_eswitch_set_vport_vlan(esw, vport->vport, attr->vlan_vid[0], 0,
+						    SET_VLAN_INSERT | SET_VLAN_STRIP);
+		if (err)
+			goto out;
+		vport->vlan = attr->vlan_vid[0];
+skip_set_push:
+		vport->vlan_refcount++;
+	}
+out:
+	if (!err)
+		attr->vlan_handled = true;
+	return err;
+}
+
+int mlx5_eswitch_del_vlan_action(struct mlx5_eswitch *esw,
+				 struct mlx5_esw_flow_attr *attr)
+{
+	struct offloads_fdb *offloads = &esw->fdb_table.offloads;
+	struct mlx5_eswitch_rep *vport = NULL;
+	bool push, pop, fwd;
+	int err = 0;
+
+	/* nop if we're on the vlan push/pop non emulation mode */
+	if (mlx5_eswitch_vlan_actions_supported(esw->dev, 1))
+		return 0;
+
+	if (!attr->vlan_handled)
+		return 0;
+
+	push = !!(attr->action & MLX5_FLOW_CONTEXT_ACTION_VLAN_PUSH);
+	pop  = !!(attr->action & MLX5_FLOW_CONTEXT_ACTION_VLAN_POP);
+	fwd  = !!(attr->action & MLX5_FLOW_CONTEXT_ACTION_FWD_DEST);
+
+	vport = esw_vlan_action_get_vport(attr, push, pop);
+
+	if (!push && !pop && fwd) {
+		/* tracks VF --> wire rules without vlan push action */
+		if (attr->out_rep[0]->vport == FDB_UPLINK_VPORT)
+			vport->vlan_refcount--;
+
+		return 0;
+	}
+
+	if (push) {
+		vport->vlan_refcount--;
+		if (vport->vlan_refcount)
+			goto skip_unset_push;
+
+		vport->vlan = 0;
+		err = __mlx5_eswitch_set_vport_vlan(esw, vport->vport,
+						    0, 0, SET_VLAN_STRIP);
+		if (err)
+			goto out;
+	}
+
+skip_unset_push:
+	offloads->vlan_push_pop_refcount--;
+	if (offloads->vlan_push_pop_refcount)
+		return 0;
+
+	/* no more vlan rules, stop global vlan pop policy */
+	err = esw_set_global_vlan_pop(esw, 0);
+
+out:
+	return err;
+}
+
+struct mlx5_flow_handle *
+mlx5_eswitch_add_send_to_vport_rule(struct mlx5_eswitch *esw, int vport, u32 sqn)
+{
+	struct mlx5_flow_act flow_act = {0};
+	struct mlx5_flow_destination dest = {};
+	struct mlx5_flow_handle *flow_rule;
+	struct mlx5_flow_spec *spec;
+	void *misc;
+
+	spec = kvzalloc(sizeof(*spec), GFP_KERNEL);
+	if (!spec) {
+		flow_rule = ERR_PTR(-ENOMEM);
+		goto out;
+	}
+
+	misc = MLX5_ADDR_OF(fte_match_param, spec->match_value, misc_parameters);
+	MLX5_SET(fte_match_set_misc, misc, source_sqn, sqn);
+	MLX5_SET(fte_match_set_misc, misc, source_port, 0x0); /* source vport is 0 */
+
+	misc = MLX5_ADDR_OF(fte_match_param, spec->match_criteria, misc_parameters);
+	MLX5_SET_TO_ONES(fte_match_set_misc, misc, source_sqn);
+	MLX5_SET_TO_ONES(fte_match_set_misc, misc, source_port);
+
+	spec->match_criteria_enable = MLX5_MATCH_MISC_PARAMETERS;
+	dest.type = MLX5_FLOW_DESTINATION_TYPE_VPORT;
+	dest.vport.num = vport;
+	flow_act.action = MLX5_FLOW_CONTEXT_ACTION_FWD_DEST;
+
+	flow_rule = mlx5_add_flow_rules(esw->fdb_table.offloads.slow_fdb, spec,
+					&flow_act, &dest, 1);
+	if (IS_ERR(flow_rule))
+		esw_warn(esw->dev, "FDB: Failed to add send to vport rule err %ld\n", PTR_ERR(flow_rule));
+out:
+	kvfree(spec);
+	return flow_rule;
+}
+EXPORT_SYMBOL(mlx5_eswitch_add_send_to_vport_rule);
+
+void mlx5_eswitch_del_send_to_vport_rule(struct mlx5_flow_handle *rule)
+{
+	mlx5_del_flow_rules(rule);
+}
+
+static int esw_add_fdb_miss_rule(struct mlx5_eswitch *esw)
+{
+	struct mlx5_flow_act flow_act = {0};
+	struct mlx5_flow_destination dest = {};
+	struct mlx5_flow_handle *flow_rule = NULL;
+	struct mlx5_flow_spec *spec;
+	void *headers_c;
+	void *headers_v;
+	int err = 0;
+	u8 *dmac_c;
+	u8 *dmac_v;
+
+	spec = kvzalloc(sizeof(*spec), GFP_KERNEL);
+	if (!spec) {
+		err = -ENOMEM;
+		goto out;
+	}
+
+	spec->match_criteria_enable = MLX5_MATCH_OUTER_HEADERS;
+	headers_c = MLX5_ADDR_OF(fte_match_param, spec->match_criteria,
+				 outer_headers);
+	dmac_c = MLX5_ADDR_OF(fte_match_param, headers_c,
+			      outer_headers.dmac_47_16);
+	dmac_c[0] = 0x01;
+
+	dest.type = MLX5_FLOW_DESTINATION_TYPE_VPORT;
+	dest.vport.num = 0;
+	flow_act.action = MLX5_FLOW_CONTEXT_ACTION_FWD_DEST;
+
+	flow_rule = mlx5_add_flow_rules(esw->fdb_table.offloads.slow_fdb, spec,
+					&flow_act, &dest, 1);
+	if (IS_ERR(flow_rule)) {
+		err = PTR_ERR(flow_rule);
+		esw_warn(esw->dev,  "FDB: Failed to add unicast miss flow rule err %d\n", err);
+		goto out;
+	}
+
+	esw->fdb_table.offloads.miss_rule_uni = flow_rule;
+
+	headers_v = MLX5_ADDR_OF(fte_match_param, spec->match_value,
+				 outer_headers);
+	dmac_v = MLX5_ADDR_OF(fte_match_param, headers_v,
+			      outer_headers.dmac_47_16);
+	dmac_v[0] = 0x01;
+	flow_rule = mlx5_add_flow_rules(esw->fdb_table.offloads.slow_fdb, spec,
+					&flow_act, &dest, 1);
+	if (IS_ERR(flow_rule)) {
+		err = PTR_ERR(flow_rule);
+		esw_warn(esw->dev, "FDB: Failed to add multicast miss flow rule err %d\n", err);
+		mlx5_del_flow_rules(esw->fdb_table.offloads.miss_rule_uni);
+		goto out;
+	}
+
+	esw->fdb_table.offloads.miss_rule_multi = flow_rule;
+
+out:
+	kvfree(spec);
+	return err;
+}
+
+#define ESW_OFFLOADS_NUM_GROUPS  4
+
+static int esw_create_offloads_fast_fdb_table(struct mlx5_eswitch *esw)
+{
+	struct mlx5_core_dev *dev = esw->dev;
+	struct mlx5_flow_namespace *root_ns;
+	struct mlx5_flow_table *fdb = NULL;
+	int esw_size, err = 0;
+	u32 flags = 0;
+	u32 max_flow_counter = (MLX5_CAP_GEN(dev, max_flow_counter_31_16) << 16) |
+				MLX5_CAP_GEN(dev, max_flow_counter_15_0);
+
+	root_ns = mlx5_get_flow_namespace(dev, MLX5_FLOW_NAMESPACE_FDB);
+	if (!root_ns) {
+		esw_warn(dev, "Failed to get FDB flow namespace\n");
+		err = -EOPNOTSUPP;
+		goto out_namespace;
+	}
+
+	esw_debug(dev, "Create offloads FDB table, min (max esw size(2^%d), max counters(%d)*groups(%d))\n",
+		  MLX5_CAP_ESW_FLOWTABLE_FDB(dev, log_max_ft_size),
+		  max_flow_counter, ESW_OFFLOADS_NUM_GROUPS);
+
+	esw_size = min_t(int, max_flow_counter * ESW_OFFLOADS_NUM_GROUPS,
+			 1 << MLX5_CAP_ESW_FLOWTABLE_FDB(dev, log_max_ft_size));
+
+	if (mlx5_esw_has_fwd_fdb(dev))
+		esw_size >>= 1;
+
+	if (esw->offloads.encap != DEVLINK_ESWITCH_ENCAP_MODE_NONE)
+		flags |= MLX5_FLOW_TABLE_TUNNEL_EN;
+
+	fdb = mlx5_create_auto_grouped_flow_table(root_ns, FDB_FAST_PATH,
+						  esw_size,
+						  ESW_OFFLOADS_NUM_GROUPS, 0,
+						  flags);
+	if (IS_ERR(fdb)) {
+		err = PTR_ERR(fdb);
+		esw_warn(dev, "Failed to create Fast path FDB Table err %d\n", err);
+		goto out_namespace;
+	}
+	esw->fdb_table.offloads.fast_fdb = fdb;
+
+	if (!mlx5_esw_has_fwd_fdb(dev))
+		goto out_namespace;
+
+	fdb = mlx5_create_auto_grouped_flow_table(root_ns, FDB_FAST_PATH,
+						  esw_size,
+						  ESW_OFFLOADS_NUM_GROUPS, 1,
+						  flags);
+	if (IS_ERR(fdb)) {
+		err = PTR_ERR(fdb);
+		esw_warn(dev, "Failed to create fwd table err %d\n", err);
+		goto out_ft;
+	}
+	esw->fdb_table.offloads.fwd_fdb = fdb;
+
+	return err;
+
+out_ft:
+	mlx5_destroy_flow_table(esw->fdb_table.offloads.fast_fdb);
+out_namespace:
+	return err;
+}
+
+static void esw_destroy_offloads_fast_fdb_table(struct mlx5_eswitch *esw)
+{
+	if (mlx5_esw_has_fwd_fdb(esw->dev))
+		mlx5_destroy_flow_table(esw->fdb_table.offloads.fwd_fdb);
+	mlx5_destroy_flow_table(esw->fdb_table.offloads.fast_fdb);
+}
+
+#define MAX_PF_SQ 256
+#define MAX_SQ_NVPORTS 32
+
+static int esw_create_offloads_fdb_tables(struct mlx5_eswitch *esw, int nvports)
+{
+	int inlen = MLX5_ST_SZ_BYTES(create_flow_group_in);
+	struct mlx5_flow_table_attr ft_attr = {};
+	struct mlx5_core_dev *dev = esw->dev;
+	struct mlx5_flow_namespace *root_ns;
+	struct mlx5_flow_table *fdb = NULL;
+	int table_size, ix, err = 0;
+	struct mlx5_flow_group *g;
+	void *match_criteria;
+	u32 *flow_group_in;
+	u8 *dmac;
+
+	esw_debug(esw->dev, "Create offloads FDB Tables\n");
+	flow_group_in = kvzalloc(inlen, GFP_KERNEL);
+	if (!flow_group_in)
+		return -ENOMEM;
+
+	root_ns = mlx5_get_flow_namespace(dev, MLX5_FLOW_NAMESPACE_FDB);
+	if (!root_ns) {
+		esw_warn(dev, "Failed to get FDB flow namespace\n");
+		err = -EOPNOTSUPP;
+		goto ns_err;
+	}
+
+	err = esw_create_offloads_fast_fdb_table(esw);
+	if (err)
+		goto fast_fdb_err;
+
+	table_size = nvports * MAX_SQ_NVPORTS + MAX_PF_SQ + 2;
+
+	ft_attr.max_fte = table_size;
+	ft_attr.prio = FDB_SLOW_PATH;
+
+	fdb = mlx5_create_flow_table(root_ns, &ft_attr);
+	if (IS_ERR(fdb)) {
+		err = PTR_ERR(fdb);
+		esw_warn(dev, "Failed to create slow path FDB Table err %d\n", err);
+		goto slow_fdb_err;
+	}
+	esw->fdb_table.offloads.slow_fdb = fdb;
+
+	/* create send-to-vport group */
+	memset(flow_group_in, 0, inlen);
+	MLX5_SET(create_flow_group_in, flow_group_in, match_criteria_enable,
+		 MLX5_MATCH_MISC_PARAMETERS);
+
+	match_criteria = MLX5_ADDR_OF(create_flow_group_in, flow_group_in, match_criteria);
+
+	MLX5_SET_TO_ONES(fte_match_param, match_criteria, misc_parameters.source_sqn);
+	MLX5_SET_TO_ONES(fte_match_param, match_criteria, misc_parameters.source_port);
+
+	ix = nvports * MAX_SQ_NVPORTS + MAX_PF_SQ;
+	MLX5_SET(create_flow_group_in, flow_group_in, start_flow_index, 0);
+	MLX5_SET(create_flow_group_in, flow_group_in, end_flow_index, ix - 1);
+
+	g = mlx5_create_flow_group(fdb, flow_group_in);
+	if (IS_ERR(g)) {
+		err = PTR_ERR(g);
+		esw_warn(dev, "Failed to create send-to-vport flow group err(%d)\n", err);
+		goto send_vport_err;
+	}
+	esw->fdb_table.offloads.send_to_vport_grp = g;
+
+	/* create miss group */
+	memset(flow_group_in, 0, inlen);
+	MLX5_SET(create_flow_group_in, flow_group_in, match_criteria_enable,
+		 MLX5_MATCH_OUTER_HEADERS);
+	match_criteria = MLX5_ADDR_OF(create_flow_group_in, flow_group_in,
+				      match_criteria);
+	dmac = MLX5_ADDR_OF(fte_match_param, match_criteria,
+			    outer_headers.dmac_47_16);
+	dmac[0] = 0x01;
+
+	MLX5_SET(create_flow_group_in, flow_group_in, start_flow_index, ix);
+	MLX5_SET(create_flow_group_in, flow_group_in, end_flow_index, ix + 2);
+
+	g = mlx5_create_flow_group(fdb, flow_group_in);
+	if (IS_ERR(g)) {
+		err = PTR_ERR(g);
+		esw_warn(dev, "Failed to create miss flow group err(%d)\n", err);
+		goto miss_err;
+	}
+	esw->fdb_table.offloads.miss_grp = g;
+
+	err = esw_add_fdb_miss_rule(esw);
+	if (err)
+		goto miss_rule_err;
+
+	kvfree(flow_group_in);
+	return 0;
+
+miss_rule_err:
+	mlx5_destroy_flow_group(esw->fdb_table.offloads.miss_grp);
+miss_err:
+	mlx5_destroy_flow_group(esw->fdb_table.offloads.send_to_vport_grp);
+send_vport_err:
+	mlx5_destroy_flow_table(esw->fdb_table.offloads.slow_fdb);
+slow_fdb_err:
+	esw_destroy_offloads_fast_fdb_table(esw);
+fast_fdb_err:
+ns_err:
+	kvfree(flow_group_in);
+	return err;
+}
+
+static void esw_destroy_offloads_fdb_tables(struct mlx5_eswitch *esw)
+{
+	if (!esw->fdb_table.offloads.fast_fdb)
+		return;
+
+	esw_debug(esw->dev, "Destroy offloads FDB Tables\n");
+	mlx5_del_flow_rules(esw->fdb_table.offloads.miss_rule_multi);
+	mlx5_del_flow_rules(esw->fdb_table.offloads.miss_rule_uni);
+	mlx5_destroy_flow_group(esw->fdb_table.offloads.send_to_vport_grp);
+	mlx5_destroy_flow_group(esw->fdb_table.offloads.miss_grp);
+
+	mlx5_destroy_flow_table(esw->fdb_table.offloads.slow_fdb);
+	esw_destroy_offloads_fast_fdb_table(esw);
+}
+
+static int esw_create_offloads_table(struct mlx5_eswitch *esw)
+{
+	struct mlx5_flow_table_attr ft_attr = {};
+	struct mlx5_core_dev *dev = esw->dev;
+	struct mlx5_flow_table *ft_offloads;
+	struct mlx5_flow_namespace *ns;
+	int err = 0;
+
+	ns = mlx5_get_flow_namespace(dev, MLX5_FLOW_NAMESPACE_OFFLOADS);
+	if (!ns) {
+		esw_warn(esw->dev, "Failed to get offloads flow namespace\n");
+		return -EOPNOTSUPP;
+	}
+
+	ft_attr.max_fte = dev->priv.sriov.num_vfs + 2;
+
+	ft_offloads = mlx5_create_flow_table(ns, &ft_attr);
+	if (IS_ERR(ft_offloads)) {
+		err = PTR_ERR(ft_offloads);
+		esw_warn(esw->dev, "Failed to create offloads table, err %d\n", err);
+		return err;
+	}
+
+	esw->offloads.ft_offloads = ft_offloads;
+	return 0;
+}
+
+static void esw_destroy_offloads_table(struct mlx5_eswitch *esw)
+{
+	struct mlx5_esw_offload *offloads = &esw->offloads;
+
+	mlx5_destroy_flow_table(offloads->ft_offloads);
+}
+
+static int esw_create_vport_rx_group(struct mlx5_eswitch *esw)
+{
+	int inlen = MLX5_ST_SZ_BYTES(create_flow_group_in);
+	struct mlx5_flow_group *g;
+	struct mlx5_priv *priv = &esw->dev->priv;
+	u32 *flow_group_in;
+	void *match_criteria, *misc;
+	int err = 0;
+	int nvports = priv->sriov.num_vfs + 2;
+
+	flow_group_in = kvzalloc(inlen, GFP_KERNEL);
+	if (!flow_group_in)
+		return -ENOMEM;
+
+	/* create vport rx group */
+	memset(flow_group_in, 0, inlen);
+	MLX5_SET(create_flow_group_in, flow_group_in, match_criteria_enable,
+		 MLX5_MATCH_MISC_PARAMETERS);
+
+	match_criteria = MLX5_ADDR_OF(create_flow_group_in, flow_group_in, match_criteria);
+	misc = MLX5_ADDR_OF(fte_match_param, match_criteria, misc_parameters);
+	MLX5_SET_TO_ONES(fte_match_set_misc, misc, source_port);
+
+	MLX5_SET(create_flow_group_in, flow_group_in, start_flow_index, 0);
+	MLX5_SET(create_flow_group_in, flow_group_in, end_flow_index, nvports - 1);
+
+	g = mlx5_create_flow_group(esw->offloads.ft_offloads, flow_group_in);
+
+	if (IS_ERR(g)) {
+		err = PTR_ERR(g);
+		mlx5_core_warn(esw->dev, "Failed to create vport rx group err %d\n", err);
+		goto out;
+	}
+
+	esw->offloads.vport_rx_group = g;
+out:
+	kvfree(flow_group_in);
+	return err;
+}
+
+static void esw_destroy_vport_rx_group(struct mlx5_eswitch *esw)
+{
+	mlx5_destroy_flow_group(esw->offloads.vport_rx_group);
+}
+
+struct mlx5_flow_handle *
+mlx5_eswitch_create_vport_rx_rule(struct mlx5_eswitch *esw, int vport, u32 tirn)
+{
+	struct mlx5_flow_act flow_act = {0};
+	struct mlx5_flow_destination dest = {};
+	struct mlx5_flow_handle *flow_rule;
+	struct mlx5_flow_spec *spec;
+	void *misc;
+
+	spec = kvzalloc(sizeof(*spec), GFP_KERNEL);
+	if (!spec) {
+		flow_rule = ERR_PTR(-ENOMEM);
+		goto out;
+	}
+
+	misc = MLX5_ADDR_OF(fte_match_param, spec->match_value, misc_parameters);
+	MLX5_SET(fte_match_set_misc, misc, source_port, vport);
+
+	misc = MLX5_ADDR_OF(fte_match_param, spec->match_criteria, misc_parameters);
+	MLX5_SET_TO_ONES(fte_match_set_misc, misc, source_port);
+
+	spec->match_criteria_enable = MLX5_MATCH_MISC_PARAMETERS;
+	dest.type = MLX5_FLOW_DESTINATION_TYPE_TIR;
+	dest.tir_num = tirn;
+
+	flow_act.action = MLX5_FLOW_CONTEXT_ACTION_FWD_DEST;
+	flow_rule = mlx5_add_flow_rules(esw->offloads.ft_offloads, spec,
+					&flow_act, &dest, 1);
+	if (IS_ERR(flow_rule)) {
+		esw_warn(esw->dev, "fs offloads: Failed to add vport rx rule err %ld\n", PTR_ERR(flow_rule));
+		goto out;
+	}
+
+out:
+	kvfree(spec);
+	return flow_rule;
+}
+
+static int esw_offloads_start(struct mlx5_eswitch *esw)
+{
+	int err, err1, num_vfs = esw->dev->priv.sriov.num_vfs;
+
+	if (esw->mode != SRIOV_LEGACY) {
+		esw_warn(esw->dev, "Can't set offloads mode, SRIOV legacy not enabled\n");
+		return -EINVAL;
+	}
+
+	mlx5_eswitch_disable_sriov(esw);
+	err = mlx5_eswitch_enable_sriov(esw, num_vfs, SRIOV_OFFLOADS);
+	if (err) {
+		esw_warn(esw->dev, "Failed setting eswitch to offloads, err %d\n", err);
+		err1 = mlx5_eswitch_enable_sriov(esw, num_vfs, SRIOV_LEGACY);
+		if (err1)
+			esw_warn(esw->dev, "Failed setting eswitch back to legacy, err %d\n", err1);
+	}
+	if (esw->offloads.inline_mode == MLX5_INLINE_MODE_NONE) {
+		if (mlx5_eswitch_inline_mode_get(esw,
+						 num_vfs,
+						 &esw->offloads.inline_mode)) {
+			esw->offloads.inline_mode = MLX5_INLINE_MODE_L2;
+			esw_warn(esw->dev, "Inline mode is different between vports\n");
+		}
+	}
+	return err;
+}
+
+void esw_offloads_cleanup_reps(struct mlx5_eswitch *esw)
+{
+	kfree(esw->offloads.vport_reps);
+}
+
+int esw_offloads_init_reps(struct mlx5_eswitch *esw)
+{
+	int total_vfs = MLX5_TOTAL_VPORTS(esw->dev);
+	struct mlx5_core_dev *dev = esw->dev;
+	struct mlx5_esw_offload *offloads;
+	struct mlx5_eswitch_rep *rep;
+	u8 hw_id[ETH_ALEN];
+	int vport;
+
+	esw->offloads.vport_reps = kcalloc(total_vfs,
+					   sizeof(struct mlx5_eswitch_rep),
+					   GFP_KERNEL);
+	if (!esw->offloads.vport_reps)
+		return -ENOMEM;
+
+	offloads = &esw->offloads;
+	mlx5_query_nic_vport_mac_address(dev, 0, hw_id);
+
+	for (vport = 0; vport < total_vfs; vport++) {
+		rep = &offloads->vport_reps[vport];
+
+		rep->vport = vport;
+		ether_addr_copy(rep->hw_id, hw_id);
+	}
+
+	offloads->vport_reps[0].vport = FDB_UPLINK_VPORT;
+
+	return 0;
+}
+
+static void esw_offloads_unload_reps_type(struct mlx5_eswitch *esw, int nvports,
+					  u8 rep_type)
+{
+	struct mlx5_eswitch_rep *rep;
+	int vport;
+
+	for (vport = nvports - 1; vport >= 0; vport--) {
+		rep = &esw->offloads.vport_reps[vport];
+		if (!rep->rep_if[rep_type].valid)
+			continue;
+
+		rep->rep_if[rep_type].unload(rep);
+	}
+}
+
+static void esw_offloads_unload_reps(struct mlx5_eswitch *esw, int nvports)
+{
+	u8 rep_type = NUM_REP_TYPES;
+
+	while (rep_type-- > 0)
+		esw_offloads_unload_reps_type(esw, nvports, rep_type);
+}
+
+static int esw_offloads_load_reps_type(struct mlx5_eswitch *esw, int nvports,
+				       u8 rep_type)
+{
+	struct mlx5_eswitch_rep *rep;
+	int vport;
+	int err;
+
+	for (vport = 0; vport < nvports; vport++) {
+		rep = &esw->offloads.vport_reps[vport];
+		if (!rep->rep_if[rep_type].valid)
+			continue;
+
+		err = rep->rep_if[rep_type].load(esw->dev, rep);
+		if (err)
+			goto err_reps;
+	}
+
+	return 0;
+
+err_reps:
+	esw_offloads_unload_reps_type(esw, vport, rep_type);
+	return err;
+}
+
+static int esw_offloads_load_reps(struct mlx5_eswitch *esw, int nvports)
+{
+	u8 rep_type = 0;
+	int err;
+
+	for (rep_type = 0; rep_type < NUM_REP_TYPES; rep_type++) {
+		err = esw_offloads_load_reps_type(esw, nvports, rep_type);
+		if (err)
+			goto err_reps;
+	}
+
+	return err;
+
+err_reps:
+	while (rep_type-- > 0)
+		esw_offloads_unload_reps_type(esw, nvports, rep_type);
+	return err;
+}
+
+int esw_offloads_init(struct mlx5_eswitch *esw, int nvports)
+{
+	int err;
+
+	err = esw_create_offloads_fdb_tables(esw, nvports);
+	if (err)
+		return err;
+
+	err = esw_create_offloads_table(esw);
+	if (err)
+		goto create_ft_err;
+
+	err = esw_create_vport_rx_group(esw);
+	if (err)
+		goto create_fg_err;
+
+	err = esw_offloads_load_reps(esw, nvports);
+	if (err)
+		goto err_reps;
+
+	return 0;
+
+err_reps:
+	esw_destroy_vport_rx_group(esw);
+
+create_fg_err:
+	esw_destroy_offloads_table(esw);
+
+create_ft_err:
+	esw_destroy_offloads_fdb_tables(esw);
+
+	return err;
+}
+
+static int esw_offloads_stop(struct mlx5_eswitch *esw)
+{
+	int err, err1, num_vfs = esw->dev->priv.sriov.num_vfs;
+
+	mlx5_eswitch_disable_sriov(esw);
+	err = mlx5_eswitch_enable_sriov(esw, num_vfs, SRIOV_LEGACY);
+	if (err) {
+		esw_warn(esw->dev, "Failed setting eswitch to legacy, err %d\n", err);
+		err1 = mlx5_eswitch_enable_sriov(esw, num_vfs, SRIOV_OFFLOADS);
+		if (err1)
+			esw_warn(esw->dev, "Failed setting eswitch back to offloads, err %d\n", err);
+	}
+
+	/* enable back PF RoCE */
+	mlx5_reload_interface(esw->dev, MLX5_INTERFACE_PROTOCOL_IB);
+
+	return err;
+}
+
+void esw_offloads_cleanup(struct mlx5_eswitch *esw, int nvports)
+{
+	esw_offloads_unload_reps(esw, nvports);
+	esw_destroy_vport_rx_group(esw);
+	esw_destroy_offloads_table(esw);
+	esw_destroy_offloads_fdb_tables(esw);
+}
+
+static int esw_mode_from_devlink(u16 mode, u16 *mlx5_mode)
+{
+	switch (mode) {
+	case DEVLINK_ESWITCH_MODE_LEGACY:
+		*mlx5_mode = SRIOV_LEGACY;
+		break;
+	case DEVLINK_ESWITCH_MODE_SWITCHDEV:
+		*mlx5_mode = SRIOV_OFFLOADS;
+		break;
+	default:
+		return -EINVAL;
+	}
+
+	return 0;
+}
+
+static int esw_mode_to_devlink(u16 mlx5_mode, u16 *mode)
+{
+	switch (mlx5_mode) {
+	case SRIOV_LEGACY:
+		*mode = DEVLINK_ESWITCH_MODE_LEGACY;
+		break;
+	case SRIOV_OFFLOADS:
+		*mode = DEVLINK_ESWITCH_MODE_SWITCHDEV;
+		break;
+	default:
+		return -EINVAL;
+	}
+
+	return 0;
+}
+
+static int esw_inline_mode_from_devlink(u8 mode, u8 *mlx5_mode)
+{
+	switch (mode) {
+	case DEVLINK_ESWITCH_INLINE_MODE_NONE:
+		*mlx5_mode = MLX5_INLINE_MODE_NONE;
+		break;
+	case DEVLINK_ESWITCH_INLINE_MODE_LINK:
+		*mlx5_mode = MLX5_INLINE_MODE_L2;
+		break;
+	case DEVLINK_ESWITCH_INLINE_MODE_NETWORK:
+		*mlx5_mode = MLX5_INLINE_MODE_IP;
+		break;
+	case DEVLINK_ESWITCH_INLINE_MODE_TRANSPORT:
+		*mlx5_mode = MLX5_INLINE_MODE_TCP_UDP;
+		break;
+	default:
+		return -EINVAL;
+	}
+
+	return 0;
+}
+
+static int esw_inline_mode_to_devlink(u8 mlx5_mode, u8 *mode)
+{
+	switch (mlx5_mode) {
+	case MLX5_INLINE_MODE_NONE:
+		*mode = DEVLINK_ESWITCH_INLINE_MODE_NONE;
+		break;
+	case MLX5_INLINE_MODE_L2:
+		*mode = DEVLINK_ESWITCH_INLINE_MODE_LINK;
+		break;
+	case MLX5_INLINE_MODE_IP:
+		*mode = DEVLINK_ESWITCH_INLINE_MODE_NETWORK;
+		break;
+	case MLX5_INLINE_MODE_TCP_UDP:
+		*mode = DEVLINK_ESWITCH_INLINE_MODE_TRANSPORT;
+		break;
+	default:
+		return -EINVAL;
+	}
+
+	return 0;
+}
+
+static int mlx5_devlink_eswitch_check(struct devlink *devlink)
+{
+	struct mlx5_core_dev *dev = devlink_priv(devlink);
+
+	if (MLX5_CAP_GEN(dev, port_type) != MLX5_CAP_PORT_TYPE_ETH)
+		return -EOPNOTSUPP;
+
+	if(!MLX5_ESWITCH_MANAGER(dev))
+		return -EPERM;
+
+	if (dev->priv.eswitch->mode == SRIOV_NONE)
+		return -EOPNOTSUPP;
+
+	return 0;
+}
+
+int mlx5_devlink_eswitch_mode_set(struct devlink *devlink, u16 mode)
+{
+	struct mlx5_core_dev *dev = devlink_priv(devlink);
+	u16 cur_mlx5_mode, mlx5_mode = 0;
+	int err;
+
+	err = mlx5_devlink_eswitch_check(devlink);
+	if (err)
+		return err;
+
+	cur_mlx5_mode = dev->priv.eswitch->mode;
+
+	if (esw_mode_from_devlink(mode, &mlx5_mode))
+		return -EINVAL;
+
+	if (cur_mlx5_mode == mlx5_mode)
+		return 0;
+
+	if (mode == DEVLINK_ESWITCH_MODE_SWITCHDEV)
+		return esw_offloads_start(dev->priv.eswitch);
+	else if (mode == DEVLINK_ESWITCH_MODE_LEGACY)
+		return esw_offloads_stop(dev->priv.eswitch);
+	else
+		return -EINVAL;
+}
+
+int mlx5_devlink_eswitch_mode_get(struct devlink *devlink, u16 *mode)
+{
+	struct mlx5_core_dev *dev = devlink_priv(devlink);
+	int err;
+
+	err = mlx5_devlink_eswitch_check(devlink);
+	if (err)
+		return err;
+
+	return esw_mode_to_devlink(dev->priv.eswitch->mode, mode);
+}
+
+int mlx5_devlink_eswitch_inline_mode_set(struct devlink *devlink, u8 mode)
+{
+	struct mlx5_core_dev *dev = devlink_priv(devlink);
+	struct mlx5_eswitch *esw = dev->priv.eswitch;
+	int err, vport;
+	u8 mlx5_mode;
+
+	err = mlx5_devlink_eswitch_check(devlink);
+	if (err)
+		return err;
+
+	switch (MLX5_CAP_ETH(dev, wqe_inline_mode)) {
+	case MLX5_CAP_INLINE_MODE_NOT_REQUIRED:
+		if (mode == DEVLINK_ESWITCH_INLINE_MODE_NONE)
+			return 0;
+		/* fall through */
+	case MLX5_CAP_INLINE_MODE_L2:
+		esw_warn(dev, "Inline mode can't be set\n");
+		return -EOPNOTSUPP;
+	case MLX5_CAP_INLINE_MODE_VPORT_CONTEXT:
+		break;
+	}
+
+	if (esw->offloads.num_flows > 0) {
+		esw_warn(dev, "Can't set inline mode when flows are configured\n");
+		return -EOPNOTSUPP;
+	}
+
+	err = esw_inline_mode_from_devlink(mode, &mlx5_mode);
+	if (err)
+		goto out;
+
+	for (vport = 1; vport < esw->enabled_vports; vport++) {
+		err = mlx5_modify_nic_vport_min_inline(dev, vport, mlx5_mode);
+		if (err) {
+			esw_warn(dev, "Failed to set min inline on vport %d\n",
+				 vport);
+			goto revert_inline_mode;
+		}
+	}
+
+	esw->offloads.inline_mode = mlx5_mode;
+	return 0;
+
+revert_inline_mode:
+	while (--vport > 0)
+		mlx5_modify_nic_vport_min_inline(dev,
+						 vport,
+						 esw->offloads.inline_mode);
+out:
+	return err;
+}
+
+int mlx5_devlink_eswitch_inline_mode_get(struct devlink *devlink, u8 *mode)
+{
+	struct mlx5_core_dev *dev = devlink_priv(devlink);
+	struct mlx5_eswitch *esw = dev->priv.eswitch;
+	int err;
+
+	err = mlx5_devlink_eswitch_check(devlink);
+	if (err)
+		return err;
+
+	return esw_inline_mode_to_devlink(esw->offloads.inline_mode, mode);
+}
+
+int mlx5_eswitch_inline_mode_get(struct mlx5_eswitch *esw, int nvfs, u8 *mode)
+{
+	u8 prev_mlx5_mode, mlx5_mode = MLX5_INLINE_MODE_L2;
+	struct mlx5_core_dev *dev = esw->dev;
+	int vport;
+
+	if (!MLX5_CAP_GEN(dev, vport_group_manager))
+		return -EOPNOTSUPP;
+
+	if (esw->mode == SRIOV_NONE)
+		return -EOPNOTSUPP;
+
+	switch (MLX5_CAP_ETH(dev, wqe_inline_mode)) {
+	case MLX5_CAP_INLINE_MODE_NOT_REQUIRED:
+		mlx5_mode = MLX5_INLINE_MODE_NONE;
+		goto out;
+	case MLX5_CAP_INLINE_MODE_L2:
+		mlx5_mode = MLX5_INLINE_MODE_L2;
+		goto out;
+	case MLX5_CAP_INLINE_MODE_VPORT_CONTEXT:
+		goto query_vports;
+	}
+
+query_vports:
+	for (vport = 1; vport <= nvfs; vport++) {
+		mlx5_query_nic_vport_min_inline(dev, vport, &mlx5_mode);
+		if (vport > 1 && prev_mlx5_mode != mlx5_mode)
+			return -EINVAL;
+		prev_mlx5_mode = mlx5_mode;
+	}
+
+out:
+	*mode = mlx5_mode;
+	return 0;
+}
+
+int mlx5_devlink_eswitch_encap_mode_set(struct devlink *devlink, u8 encap)
+{
+	struct mlx5_core_dev *dev = devlink_priv(devlink);
+	struct mlx5_eswitch *esw = dev->priv.eswitch;
+	int err;
+
+	err = mlx5_devlink_eswitch_check(devlink);
+	if (err)
+		return err;
+
+	if (encap != DEVLINK_ESWITCH_ENCAP_MODE_NONE &&
+	    (!MLX5_CAP_ESW_FLOWTABLE_FDB(dev, encap) ||
+	     !MLX5_CAP_ESW_FLOWTABLE_FDB(dev, decap)))
+		return -EOPNOTSUPP;
+
+	if (encap && encap != DEVLINK_ESWITCH_ENCAP_MODE_BASIC)
+		return -EOPNOTSUPP;
+
+	if (esw->mode == SRIOV_LEGACY) {
+		esw->offloads.encap = encap;
+		return 0;
+	}
+
+	if (esw->offloads.encap == encap)
+		return 0;
+
+	if (esw->offloads.num_flows > 0) {
+		esw_warn(dev, "Can't set encapsulation when flows are configured\n");
+		return -EOPNOTSUPP;
+	}
+
+	esw_destroy_offloads_fast_fdb_table(esw);
+
+	esw->offloads.encap = encap;
+	err = esw_create_offloads_fast_fdb_table(esw);
+	if (err) {
+		esw_warn(esw->dev, "Failed re-creating fast FDB table, err %d\n", err);
+		esw->offloads.encap = !encap;
+		(void)esw_create_offloads_fast_fdb_table(esw);
+	}
+	return err;
+}
+
+int mlx5_devlink_eswitch_encap_mode_get(struct devlink *devlink, u8 *encap)
+{
+	struct mlx5_core_dev *dev = devlink_priv(devlink);
+	struct mlx5_eswitch *esw = dev->priv.eswitch;
+	int err;
+
+	err = mlx5_devlink_eswitch_check(devlink);
+	if (err)
+		return err;
+
+	*encap = esw->offloads.encap;
+	return 0;
+}
+
+void mlx5_eswitch_register_vport_rep(struct mlx5_eswitch *esw,
+				     int vport_index,
+				     struct mlx5_eswitch_rep_if *__rep_if,
+				     u8 rep_type)
+{
+	struct mlx5_esw_offload *offloads = &esw->offloads;
+	struct mlx5_eswitch_rep_if *rep_if;
+
+	rep_if = &offloads->vport_reps[vport_index].rep_if[rep_type];
+
+	rep_if->load   = __rep_if->load;
+	rep_if->unload = __rep_if->unload;
+	rep_if->get_proto_dev = __rep_if->get_proto_dev;
+	rep_if->priv = __rep_if->priv;
+
+	rep_if->valid = true;
+}
+EXPORT_SYMBOL(mlx5_eswitch_register_vport_rep);
+
+void mlx5_eswitch_unregister_vport_rep(struct mlx5_eswitch *esw,
+				       int vport_index, u8 rep_type)
+{
+	struct mlx5_esw_offload *offloads = &esw->offloads;
+	struct mlx5_eswitch_rep *rep;
+
+	rep = &offloads->vport_reps[vport_index];
+
+	if (esw->mode == SRIOV_OFFLOADS && esw->vports[vport_index].enabled)
+		rep->rep_if[rep_type].unload(rep);
+
+	rep->rep_if[rep_type].valid = false;
+}
+EXPORT_SYMBOL(mlx5_eswitch_unregister_vport_rep);
+
+void *mlx5_eswitch_get_uplink_priv(struct mlx5_eswitch *esw, u8 rep_type)
+{
+#define UPLINK_REP_INDEX 0
+	struct mlx5_esw_offload *offloads = &esw->offloads;
+	struct mlx5_eswitch_rep *rep;
+
+	rep = &offloads->vport_reps[UPLINK_REP_INDEX];
+	return rep->rep_if[rep_type].priv;
+}
+
+void *mlx5_eswitch_get_proto_dev(struct mlx5_eswitch *esw,
+				 int vport,
+				 u8 rep_type)
+{
+	struct mlx5_esw_offload *offloads = &esw->offloads;
+	struct mlx5_eswitch_rep *rep;
+
+	if (vport == FDB_UPLINK_VPORT)
+		vport = UPLINK_REP_INDEX;
+
+	rep = &offloads->vport_reps[vport];
+
+	if (rep->rep_if[rep_type].valid &&
+	    rep->rep_if[rep_type].get_proto_dev)
+		return rep->rep_if[rep_type].get_proto_dev(rep);
+	return NULL;
+}
+EXPORT_SYMBOL(mlx5_eswitch_get_proto_dev);
+
+void *mlx5_eswitch_uplink_get_proto_dev(struct mlx5_eswitch *esw, u8 rep_type)
+{
+	return mlx5_eswitch_get_proto_dev(esw, UPLINK_REP_INDEX, rep_type);
+}
+EXPORT_SYMBOL(mlx5_eswitch_uplink_get_proto_dev);
+
+struct mlx5_eswitch_rep *mlx5_eswitch_vport_rep(struct mlx5_eswitch *esw,
+						int vport)
+{
+	return &esw->offloads.vport_reps[vport];
+}
+EXPORT_SYMBOL(mlx5_eswitch_vport_rep);
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/fpga/Makefile b/drivers/net/ethernet/mellanox/mlx5/core/fpga/Makefile
new file mode 100644
index 000000000..d8e17110f
--- /dev/null
+++ b/drivers/net/ethernet/mellanox/mlx5/core/fpga/Makefile
@@ -0,0 +1 @@
+subdir-ccflags-y += -I$(src)/..
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/fpga/cmd.c b/drivers/net/ethernet/mellanox/mlx5/core/fpga/cmd.c
new file mode 100644
index 000000000..c0fd2212e
--- /dev/null
+++ b/drivers/net/ethernet/mellanox/mlx5/core/fpga/cmd.c
@@ -0,0 +1,238 @@
+/*
+ * Copyright (c) 2017, 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/etherdevice.h>
+#include <linux/mlx5/cmd.h>
+#include <linux/mlx5/driver.h>
+#include <linux/mlx5/device.h>
+
+#include "mlx5_core.h"
+#include "fpga/cmd.h"
+
+#define MLX5_FPGA_ACCESS_REG_SZ (MLX5_ST_SZ_DW(fpga_access_reg) + \
+				 MLX5_FPGA_ACCESS_REG_SIZE_MAX)
+
+int mlx5_fpga_access_reg(struct mlx5_core_dev *dev, u8 size, u64 addr,
+			 void *buf, bool write)
+{
+	u32 in[MLX5_FPGA_ACCESS_REG_SZ] = {0};
+	u32 out[MLX5_FPGA_ACCESS_REG_SZ];
+	int err;
+
+	if (size & 3)
+		return -EINVAL;
+	if (addr & 3)
+		return -EINVAL;
+	if (size > MLX5_FPGA_ACCESS_REG_SIZE_MAX)
+		return -EINVAL;
+
+	MLX5_SET(fpga_access_reg, in, size, size);
+	MLX5_SET64(fpga_access_reg, in, address, addr);
+	if (write)
+		memcpy(MLX5_ADDR_OF(fpga_access_reg, in, data), buf, size);
+
+	err = mlx5_core_access_reg(dev, in, sizeof(in), out, sizeof(out),
+				   MLX5_REG_FPGA_ACCESS_REG, 0, write);
+	if (err)
+		return err;
+
+	if (!write)
+		memcpy(buf, MLX5_ADDR_OF(fpga_access_reg, out, data), size);
+
+	return 0;
+}
+
+int mlx5_fpga_caps(struct mlx5_core_dev *dev)
+{
+	u32 in[MLX5_ST_SZ_DW(fpga_cap)] = {0};
+
+	return mlx5_core_access_reg(dev, in, sizeof(in), dev->caps.fpga,
+				    MLX5_ST_SZ_BYTES(fpga_cap),
+				    MLX5_REG_FPGA_CAP, 0, 0);
+}
+
+int mlx5_fpga_ctrl_op(struct mlx5_core_dev *dev, u8 op)
+{
+	u32 in[MLX5_ST_SZ_DW(fpga_ctrl)] = {0};
+	u32 out[MLX5_ST_SZ_DW(fpga_ctrl)];
+
+	MLX5_SET(fpga_ctrl, in, operation, op);
+
+	return mlx5_core_access_reg(dev, in, sizeof(in), out, sizeof(out),
+				    MLX5_REG_FPGA_CTRL, 0, true);
+}
+
+int mlx5_fpga_sbu_caps(struct mlx5_core_dev *dev, void *caps, int size)
+{
+	unsigned int cap_size = MLX5_CAP_FPGA(dev, sandbox_extended_caps_len);
+	u64 addr = MLX5_CAP64_FPGA(dev, sandbox_extended_caps_addr);
+	unsigned int read;
+	int ret = 0;
+
+	if (cap_size > size) {
+		mlx5_core_warn(dev, "Not enough buffer %u for FPGA SBU caps %u",
+			       size, cap_size);
+		return -EINVAL;
+	}
+
+	while (cap_size > 0) {
+		read = min_t(unsigned int, cap_size,
+			     MLX5_FPGA_ACCESS_REG_SIZE_MAX);
+
+		ret = mlx5_fpga_access_reg(dev, read, addr, caps, false);
+		if (ret) {
+			mlx5_core_warn(dev, "Error reading FPGA SBU caps %u bytes at address 0x%llx: %d",
+				       read, addr, ret);
+			return ret;
+		}
+
+		cap_size -= read;
+		addr += read;
+		caps += read;
+	}
+
+	return ret;
+}
+
+int mlx5_fpga_query(struct mlx5_core_dev *dev, struct mlx5_fpga_query *query)
+{
+	u32 in[MLX5_ST_SZ_DW(fpga_ctrl)] = {0};
+	u32 out[MLX5_ST_SZ_DW(fpga_ctrl)];
+	int err;
+
+	err = mlx5_core_access_reg(dev, in, sizeof(in), out, sizeof(out),
+				   MLX5_REG_FPGA_CTRL, 0, false);
+	if (err)
+		return err;
+
+	query->status = MLX5_GET(fpga_ctrl, out, status);
+	query->admin_image = MLX5_GET(fpga_ctrl, out, flash_select_admin);
+	query->oper_image = MLX5_GET(fpga_ctrl, out, flash_select_oper);
+	return 0;
+}
+
+int mlx5_fpga_create_qp(struct mlx5_core_dev *dev, void *fpga_qpc,
+			u32 *fpga_qpn)
+{
+	u32 in[MLX5_ST_SZ_DW(fpga_create_qp_in)] = {0};
+	u32 out[MLX5_ST_SZ_DW(fpga_create_qp_out)];
+	int ret;
+
+	MLX5_SET(fpga_create_qp_in, in, opcode, MLX5_CMD_OP_FPGA_CREATE_QP);
+	memcpy(MLX5_ADDR_OF(fpga_create_qp_in, in, fpga_qpc), fpga_qpc,
+	       MLX5_FLD_SZ_BYTES(fpga_create_qp_in, fpga_qpc));
+
+	ret = mlx5_cmd_exec(dev, in, sizeof(in), out, sizeof(out));
+	if (ret)
+		return ret;
+
+	memcpy(fpga_qpc, MLX5_ADDR_OF(fpga_create_qp_out, out, fpga_qpc),
+	       MLX5_FLD_SZ_BYTES(fpga_create_qp_out, fpga_qpc));
+	*fpga_qpn = MLX5_GET(fpga_create_qp_out, out, fpga_qpn);
+	return ret;
+}
+
+int mlx5_fpga_modify_qp(struct mlx5_core_dev *dev, u32 fpga_qpn,
+			enum mlx5_fpga_qpc_field_select fields,
+			void *fpga_qpc)
+{
+	u32 in[MLX5_ST_SZ_DW(fpga_modify_qp_in)] = {0};
+	u32 out[MLX5_ST_SZ_DW(fpga_modify_qp_out)];
+
+	MLX5_SET(fpga_modify_qp_in, in, opcode, MLX5_CMD_OP_FPGA_MODIFY_QP);
+	MLX5_SET(fpga_modify_qp_in, in, field_select, fields);
+	MLX5_SET(fpga_modify_qp_in, in, fpga_qpn, fpga_qpn);
+	memcpy(MLX5_ADDR_OF(fpga_modify_qp_in, in, fpga_qpc), fpga_qpc,
+	       MLX5_FLD_SZ_BYTES(fpga_modify_qp_in, fpga_qpc));
+
+	return mlx5_cmd_exec(dev, in, sizeof(in), out, sizeof(out));
+}
+
+int mlx5_fpga_query_qp(struct mlx5_core_dev *dev,
+		       u32 fpga_qpn, void *fpga_qpc)
+{
+	u32 in[MLX5_ST_SZ_DW(fpga_query_qp_in)] = {0};
+	u32 out[MLX5_ST_SZ_DW(fpga_query_qp_out)];
+	int ret;
+
+	MLX5_SET(fpga_query_qp_in, in, opcode, MLX5_CMD_OP_FPGA_QUERY_QP);
+	MLX5_SET(fpga_query_qp_in, in, fpga_qpn, fpga_qpn);
+
+	ret = mlx5_cmd_exec(dev, in, sizeof(in), out, sizeof(out));
+	if (ret)
+		return ret;
+
+	memcpy(fpga_qpc, MLX5_ADDR_OF(fpga_query_qp_out, out, fpga_qpc),
+	       MLX5_FLD_SZ_BYTES(fpga_query_qp_out, fpga_qpc));
+	return ret;
+}
+
+int mlx5_fpga_destroy_qp(struct mlx5_core_dev *dev, u32 fpga_qpn)
+{
+	u32 in[MLX5_ST_SZ_DW(fpga_destroy_qp_in)] = {0};
+	u32 out[MLX5_ST_SZ_DW(fpga_destroy_qp_out)];
+
+	MLX5_SET(fpga_destroy_qp_in, in, opcode, MLX5_CMD_OP_FPGA_DESTROY_QP);
+	MLX5_SET(fpga_destroy_qp_in, in, fpga_qpn, fpga_qpn);
+
+	return mlx5_cmd_exec(dev, in, sizeof(in), out, sizeof(out));
+}
+
+int mlx5_fpga_query_qp_counters(struct mlx5_core_dev *dev, u32 fpga_qpn,
+				bool clear, struct mlx5_fpga_qp_counters *data)
+{
+	u32 in[MLX5_ST_SZ_DW(fpga_query_qp_counters_in)] = {0};
+	u32 out[MLX5_ST_SZ_DW(fpga_query_qp_counters_out)];
+	int ret;
+
+	MLX5_SET(fpga_query_qp_counters_in, in, opcode,
+		 MLX5_CMD_OP_FPGA_QUERY_QP_COUNTERS);
+	MLX5_SET(fpga_query_qp_counters_in, in, clear, clear);
+	MLX5_SET(fpga_query_qp_counters_in, in, fpga_qpn, fpga_qpn);
+
+	ret = mlx5_cmd_exec(dev, in, sizeof(in), out, sizeof(out));
+	if (ret)
+		return ret;
+
+	data->rx_ack_packets = MLX5_GET64(fpga_query_qp_counters_out, out,
+					  rx_ack_packets);
+	data->rx_send_packets = MLX5_GET64(fpga_query_qp_counters_out, out,
+					   rx_send_packets);
+	data->tx_ack_packets = MLX5_GET64(fpga_query_qp_counters_out, out,
+					  tx_ack_packets);
+	data->tx_send_packets = MLX5_GET64(fpga_query_qp_counters_out, out,
+					   tx_send_packets);
+	data->rx_total_drop = MLX5_GET64(fpga_query_qp_counters_out, out,
+					 rx_total_drop);
+
+	return ret;
+}
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/fpga/cmd.h b/drivers/net/ethernet/mellanox/mlx5/core/fpga/cmd.h
new file mode 100644
index 000000000..eb8b0fe0b
--- /dev/null
+++ b/drivers/net/ethernet/mellanox/mlx5/core/fpga/cmd.h
@@ -0,0 +1,91 @@
+/*
+ * Copyright (c) 2017, Mellanox Technologies, Ltd.  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 __MLX5_FPGA_H__
+#define __MLX5_FPGA_H__
+
+#include <linux/mlx5/driver.h>
+
+enum mlx5_fpga_device_id {
+	MLX5_FPGA_DEVICE_UNKNOWN = 0,
+	MLX5_FPGA_DEVICE_KU040 = 1,
+	MLX5_FPGA_DEVICE_KU060 = 2,
+	MLX5_FPGA_DEVICE_KU060_2 = 3,
+};
+
+enum mlx5_fpga_image {
+	MLX5_FPGA_IMAGE_USER = 0,
+	MLX5_FPGA_IMAGE_FACTORY,
+};
+
+enum mlx5_fpga_status {
+	MLX5_FPGA_STATUS_SUCCESS = 0,
+	MLX5_FPGA_STATUS_FAILURE = 1,
+	MLX5_FPGA_STATUS_IN_PROGRESS = 2,
+	MLX5_FPGA_STATUS_NONE = 0xFFFF,
+};
+
+struct mlx5_fpga_query {
+	enum mlx5_fpga_image admin_image;
+	enum mlx5_fpga_image oper_image;
+	enum mlx5_fpga_status status;
+};
+
+enum mlx5_fpga_qpc_field_select {
+	MLX5_FPGA_QPC_STATE = BIT(0),
+};
+
+struct mlx5_fpga_qp_counters {
+	u64 rx_ack_packets;
+	u64 rx_send_packets;
+	u64 tx_ack_packets;
+	u64 tx_send_packets;
+	u64 rx_total_drop;
+};
+
+int mlx5_fpga_caps(struct mlx5_core_dev *dev);
+int mlx5_fpga_query(struct mlx5_core_dev *dev, struct mlx5_fpga_query *query);
+int mlx5_fpga_ctrl_op(struct mlx5_core_dev *dev, u8 op);
+int mlx5_fpga_access_reg(struct mlx5_core_dev *dev, u8 size, u64 addr,
+			 void *buf, bool write);
+int mlx5_fpga_sbu_caps(struct mlx5_core_dev *dev, void *caps, int size);
+
+int mlx5_fpga_create_qp(struct mlx5_core_dev *dev, void *fpga_qpc,
+			u32 *fpga_qpn);
+int mlx5_fpga_modify_qp(struct mlx5_core_dev *dev, u32 fpga_qpn,
+			enum mlx5_fpga_qpc_field_select fields, void *fpga_qpc);
+int mlx5_fpga_query_qp(struct mlx5_core_dev *dev, u32 fpga_qpn, void *fpga_qpc);
+int mlx5_fpga_query_qp_counters(struct mlx5_core_dev *dev, u32 fpga_qpn,
+				bool clear, struct mlx5_fpga_qp_counters *data);
+int mlx5_fpga_destroy_qp(struct mlx5_core_dev *dev, u32 fpga_qpn);
+
+#endif /* __MLX5_FPGA_H__ */
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/fpga/conn.c b/drivers/net/ethernet/mellanox/mlx5/core/fpga/conn.c
new file mode 100644
index 000000000..d8d0b6bd5
--- /dev/null
+++ b/drivers/net/ethernet/mellanox/mlx5/core/fpga/conn.c
@@ -0,0 +1,1047 @@
+/*
+ * Copyright (c) 2017 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/addrconf.h>
+#include <linux/etherdevice.h>
+#include <linux/mlx5/vport.h>
+
+#include "mlx5_core.h"
+#include "lib/mlx5.h"
+#include "fpga/conn.h"
+
+#define MLX5_FPGA_PKEY 0xFFFF
+#define MLX5_FPGA_PKEY_INDEX 0 /* RoCE PKEY 0xFFFF is always at index 0 */
+#define MLX5_FPGA_RECV_SIZE 2048
+#define MLX5_FPGA_PORT_NUM 1
+#define MLX5_FPGA_CQ_BUDGET 64
+
+static int mlx5_fpga_conn_map_buf(struct mlx5_fpga_conn *conn,
+				  struct mlx5_fpga_dma_buf *buf)
+{
+	struct device *dma_device;
+	int err = 0;
+
+	if (unlikely(!buf->sg[0].data))
+		goto out;
+
+	dma_device = &conn->fdev->mdev->pdev->dev;
+	buf->sg[0].dma_addr = dma_map_single(dma_device, buf->sg[0].data,
+					     buf->sg[0].size, buf->dma_dir);
+	err = dma_mapping_error(dma_device, buf->sg[0].dma_addr);
+	if (unlikely(err)) {
+		mlx5_fpga_warn(conn->fdev, "DMA error on sg 0: %d\n", err);
+		err = -ENOMEM;
+		goto out;
+	}
+
+	if (!buf->sg[1].data)
+		goto out;
+
+	buf->sg[1].dma_addr = dma_map_single(dma_device, buf->sg[1].data,
+					     buf->sg[1].size, buf->dma_dir);
+	err = dma_mapping_error(dma_device, buf->sg[1].dma_addr);
+	if (unlikely(err)) {
+		mlx5_fpga_warn(conn->fdev, "DMA error on sg 1: %d\n", err);
+		dma_unmap_single(dma_device, buf->sg[0].dma_addr,
+				 buf->sg[0].size, buf->dma_dir);
+		err = -ENOMEM;
+	}
+
+out:
+	return err;
+}
+
+static void mlx5_fpga_conn_unmap_buf(struct mlx5_fpga_conn *conn,
+				     struct mlx5_fpga_dma_buf *buf)
+{
+	struct device *dma_device;
+
+	dma_device = &conn->fdev->mdev->pdev->dev;
+	if (buf->sg[1].data)
+		dma_unmap_single(dma_device, buf->sg[1].dma_addr,
+				 buf->sg[1].size, buf->dma_dir);
+
+	if (likely(buf->sg[0].data))
+		dma_unmap_single(dma_device, buf->sg[0].dma_addr,
+				 buf->sg[0].size, buf->dma_dir);
+}
+
+static int mlx5_fpga_conn_post_recv(struct mlx5_fpga_conn *conn,
+				    struct mlx5_fpga_dma_buf *buf)
+{
+	struct mlx5_wqe_data_seg *data;
+	unsigned int ix;
+	int err = 0;
+
+	err = mlx5_fpga_conn_map_buf(conn, buf);
+	if (unlikely(err))
+		goto out;
+
+	if (unlikely(conn->qp.rq.pc - conn->qp.rq.cc >= conn->qp.rq.size)) {
+		mlx5_fpga_conn_unmap_buf(conn, buf);
+		return -EBUSY;
+	}
+
+	ix = conn->qp.rq.pc & (conn->qp.rq.size - 1);
+	data = mlx5_wq_cyc_get_wqe(&conn->qp.wq.rq, ix);
+	data->byte_count = cpu_to_be32(buf->sg[0].size);
+	data->lkey = cpu_to_be32(conn->fdev->conn_res.mkey.key);
+	data->addr = cpu_to_be64(buf->sg[0].dma_addr);
+
+	conn->qp.rq.pc++;
+	conn->qp.rq.bufs[ix] = buf;
+
+	/* Make sure that descriptors are written before doorbell record. */
+	dma_wmb();
+	*conn->qp.wq.rq.db = cpu_to_be32(conn->qp.rq.pc & 0xffff);
+out:
+	return err;
+}
+
+static void mlx5_fpga_conn_notify_hw(struct mlx5_fpga_conn *conn, void *wqe)
+{
+	/* ensure wqe is visible to device before updating doorbell record */
+	dma_wmb();
+	*conn->qp.wq.sq.db = cpu_to_be32(conn->qp.sq.pc);
+	/* Make sure that doorbell record is visible before ringing */
+	wmb();
+	mlx5_write64(wqe, conn->fdev->conn_res.uar->map + MLX5_BF_OFFSET, NULL);
+}
+
+static void mlx5_fpga_conn_post_send(struct mlx5_fpga_conn *conn,
+				     struct mlx5_fpga_dma_buf *buf)
+{
+	struct mlx5_wqe_ctrl_seg *ctrl;
+	struct mlx5_wqe_data_seg *data;
+	unsigned int ix, sgi;
+	int size = 1;
+
+	ix = conn->qp.sq.pc & (conn->qp.sq.size - 1);
+
+	ctrl = mlx5_wq_cyc_get_wqe(&conn->qp.wq.sq, ix);
+	data = (void *)(ctrl + 1);
+
+	for (sgi = 0; sgi < ARRAY_SIZE(buf->sg); sgi++) {
+		if (!buf->sg[sgi].data)
+			break;
+		data->byte_count = cpu_to_be32(buf->sg[sgi].size);
+		data->lkey = cpu_to_be32(conn->fdev->conn_res.mkey.key);
+		data->addr = cpu_to_be64(buf->sg[sgi].dma_addr);
+		data++;
+		size++;
+	}
+
+	ctrl->imm = 0;
+	ctrl->fm_ce_se = MLX5_WQE_CTRL_CQ_UPDATE;
+	ctrl->opmod_idx_opcode = cpu_to_be32(((conn->qp.sq.pc & 0xffff) << 8) |
+					     MLX5_OPCODE_SEND);
+	ctrl->qpn_ds = cpu_to_be32(size | (conn->qp.mqp.qpn << 8));
+
+	conn->qp.sq.pc++;
+	conn->qp.sq.bufs[ix] = buf;
+	mlx5_fpga_conn_notify_hw(conn, ctrl);
+}
+
+int mlx5_fpga_conn_send(struct mlx5_fpga_conn *conn,
+			struct mlx5_fpga_dma_buf *buf)
+{
+	unsigned long flags;
+	int err;
+
+	if (!conn->qp.active)
+		return -ENOTCONN;
+
+	buf->dma_dir = DMA_TO_DEVICE;
+	err = mlx5_fpga_conn_map_buf(conn, buf);
+	if (err)
+		return err;
+
+	spin_lock_irqsave(&conn->qp.sq.lock, flags);
+
+	if (conn->qp.sq.pc - conn->qp.sq.cc >= conn->qp.sq.size) {
+		list_add_tail(&buf->list, &conn->qp.sq.backlog);
+		goto out_unlock;
+	}
+
+	mlx5_fpga_conn_post_send(conn, buf);
+
+out_unlock:
+	spin_unlock_irqrestore(&conn->qp.sq.lock, flags);
+	return err;
+}
+
+static int mlx5_fpga_conn_post_recv_buf(struct mlx5_fpga_conn *conn)
+{
+	struct mlx5_fpga_dma_buf *buf;
+	int err;
+
+	buf = kzalloc(sizeof(*buf) + MLX5_FPGA_RECV_SIZE, 0);
+	if (!buf)
+		return -ENOMEM;
+
+	buf->sg[0].data = (void *)(buf + 1);
+	buf->sg[0].size = MLX5_FPGA_RECV_SIZE;
+	buf->dma_dir = DMA_FROM_DEVICE;
+
+	err = mlx5_fpga_conn_post_recv(conn, buf);
+	if (err)
+		kfree(buf);
+
+	return err;
+}
+
+static int mlx5_fpga_conn_create_mkey(struct mlx5_core_dev *mdev, u32 pdn,
+				      struct mlx5_core_mkey *mkey)
+{
+	int inlen = MLX5_ST_SZ_BYTES(create_mkey_in);
+	void *mkc;
+	u32 *in;
+	int err;
+
+	in = kvzalloc(inlen, GFP_KERNEL);
+	if (!in)
+		return -ENOMEM;
+
+	mkc = MLX5_ADDR_OF(create_mkey_in, in, memory_key_mkey_entry);
+	MLX5_SET(mkc, mkc, access_mode_1_0, MLX5_MKC_ACCESS_MODE_PA);
+	MLX5_SET(mkc, mkc, lw, 1);
+	MLX5_SET(mkc, mkc, lr, 1);
+
+	MLX5_SET(mkc, mkc, pd, pdn);
+	MLX5_SET(mkc, mkc, length64, 1);
+	MLX5_SET(mkc, mkc, qpn, 0xffffff);
+
+	err = mlx5_core_create_mkey(mdev, mkey, in, inlen);
+
+	kvfree(in);
+	return err;
+}
+
+static void mlx5_fpga_conn_rq_cqe(struct mlx5_fpga_conn *conn,
+				  struct mlx5_cqe64 *cqe, u8 status)
+{
+	struct mlx5_fpga_dma_buf *buf;
+	int ix, err;
+
+	ix = be16_to_cpu(cqe->wqe_counter) & (conn->qp.rq.size - 1);
+	buf = conn->qp.rq.bufs[ix];
+	conn->qp.rq.bufs[ix] = NULL;
+	conn->qp.rq.cc++;
+
+	if (unlikely(status && (status != MLX5_CQE_SYNDROME_WR_FLUSH_ERR)))
+		mlx5_fpga_warn(conn->fdev, "RQ buf %p on FPGA QP %u completion status %d\n",
+			       buf, conn->fpga_qpn, status);
+	else
+		mlx5_fpga_dbg(conn->fdev, "RQ buf %p on FPGA QP %u completion status %d\n",
+			      buf, conn->fpga_qpn, status);
+
+	mlx5_fpga_conn_unmap_buf(conn, buf);
+
+	if (unlikely(status || !conn->qp.active)) {
+		conn->qp.active = false;
+		kfree(buf);
+		return;
+	}
+
+	buf->sg[0].size = be32_to_cpu(cqe->byte_cnt);
+	mlx5_fpga_dbg(conn->fdev, "Message with %u bytes received successfully\n",
+		      buf->sg[0].size);
+	conn->recv_cb(conn->cb_arg, buf);
+
+	buf->sg[0].size = MLX5_FPGA_RECV_SIZE;
+	err = mlx5_fpga_conn_post_recv(conn, buf);
+	if (unlikely(err)) {
+		mlx5_fpga_warn(conn->fdev,
+			       "Failed to re-post recv buf: %d\n", err);
+		kfree(buf);
+	}
+}
+
+static void mlx5_fpga_conn_sq_cqe(struct mlx5_fpga_conn *conn,
+				  struct mlx5_cqe64 *cqe, u8 status)
+{
+	struct mlx5_fpga_dma_buf *buf, *nextbuf;
+	unsigned long flags;
+	int ix;
+
+	spin_lock_irqsave(&conn->qp.sq.lock, flags);
+
+	ix = be16_to_cpu(cqe->wqe_counter) & (conn->qp.sq.size - 1);
+	buf = conn->qp.sq.bufs[ix];
+	conn->qp.sq.bufs[ix] = NULL;
+	conn->qp.sq.cc++;
+
+	/* Handle backlog still under the spinlock to ensure message post order */
+	if (unlikely(!list_empty(&conn->qp.sq.backlog))) {
+		if (likely(conn->qp.active)) {
+			nextbuf = list_first_entry(&conn->qp.sq.backlog,
+						   struct mlx5_fpga_dma_buf, list);
+			list_del(&nextbuf->list);
+			mlx5_fpga_conn_post_send(conn, nextbuf);
+		}
+	}
+
+	spin_unlock_irqrestore(&conn->qp.sq.lock, flags);
+
+	if (unlikely(status && (status != MLX5_CQE_SYNDROME_WR_FLUSH_ERR)))
+		mlx5_fpga_warn(conn->fdev, "SQ buf %p on FPGA QP %u completion status %d\n",
+			       buf, conn->fpga_qpn, status);
+	else
+		mlx5_fpga_dbg(conn->fdev, "SQ buf %p on FPGA QP %u completion status %d\n",
+			      buf, conn->fpga_qpn, status);
+
+	mlx5_fpga_conn_unmap_buf(conn, buf);
+
+	if (likely(buf->complete))
+		buf->complete(conn, conn->fdev, buf, status);
+
+	if (unlikely(status))
+		conn->qp.active = false;
+}
+
+static void mlx5_fpga_conn_handle_cqe(struct mlx5_fpga_conn *conn,
+				      struct mlx5_cqe64 *cqe)
+{
+	u8 opcode, status = 0;
+
+	opcode = cqe->op_own >> 4;
+
+	switch (opcode) {
+	case MLX5_CQE_REQ_ERR:
+		status = ((struct mlx5_err_cqe *)cqe)->syndrome;
+		/* Fall through */
+	case MLX5_CQE_REQ:
+		mlx5_fpga_conn_sq_cqe(conn, cqe, status);
+		break;
+
+	case MLX5_CQE_RESP_ERR:
+		status = ((struct mlx5_err_cqe *)cqe)->syndrome;
+		/* Fall through */
+	case MLX5_CQE_RESP_SEND:
+		mlx5_fpga_conn_rq_cqe(conn, cqe, status);
+		break;
+	default:
+		mlx5_fpga_warn(conn->fdev, "Unexpected cqe opcode %u\n",
+			       opcode);
+	}
+}
+
+static void mlx5_fpga_conn_arm_cq(struct mlx5_fpga_conn *conn)
+{
+	mlx5_cq_arm(&conn->cq.mcq, MLX5_CQ_DB_REQ_NOT,
+		    conn->fdev->conn_res.uar->map, conn->cq.wq.cc);
+}
+
+static void mlx5_fpga_conn_cq_event(struct mlx5_core_cq *mcq,
+				    enum mlx5_event event)
+{
+	struct mlx5_fpga_conn *conn;
+
+	conn = container_of(mcq, struct mlx5_fpga_conn, cq.mcq);
+	mlx5_fpga_warn(conn->fdev, "CQ event %u on CQ #%u\n", event, mcq->cqn);
+}
+
+static void mlx5_fpga_conn_event(struct mlx5_core_qp *mqp, int event)
+{
+	struct mlx5_fpga_conn *conn;
+
+	conn = container_of(mqp, struct mlx5_fpga_conn, qp.mqp);
+	mlx5_fpga_warn(conn->fdev, "QP event %u on QP #%u\n", event, mqp->qpn);
+}
+
+static inline void mlx5_fpga_conn_cqes(struct mlx5_fpga_conn *conn,
+				       unsigned int budget)
+{
+	struct mlx5_cqe64 *cqe;
+
+	while (budget) {
+		cqe = mlx5_cqwq_get_cqe(&conn->cq.wq);
+		if (!cqe)
+			break;
+
+		budget--;
+		mlx5_cqwq_pop(&conn->cq.wq);
+		mlx5_fpga_conn_handle_cqe(conn, cqe);
+		mlx5_cqwq_update_db_record(&conn->cq.wq);
+	}
+	if (!budget) {
+		tasklet_schedule(&conn->cq.tasklet);
+		return;
+	}
+
+	mlx5_fpga_dbg(conn->fdev, "Re-arming CQ with cc# %u\n", conn->cq.wq.cc);
+	/* ensure cq space is freed before enabling more cqes */
+	wmb();
+	mlx5_fpga_conn_arm_cq(conn);
+}
+
+static void mlx5_fpga_conn_cq_tasklet(unsigned long data)
+{
+	struct mlx5_fpga_conn *conn = (void *)data;
+
+	if (unlikely(!conn->qp.active))
+		return;
+	mlx5_fpga_conn_cqes(conn, MLX5_FPGA_CQ_BUDGET);
+}
+
+static void mlx5_fpga_conn_cq_complete(struct mlx5_core_cq *mcq)
+{
+	struct mlx5_fpga_conn *conn;
+
+	conn = container_of(mcq, struct mlx5_fpga_conn, cq.mcq);
+	if (unlikely(!conn->qp.active))
+		return;
+	mlx5_fpga_conn_cqes(conn, MLX5_FPGA_CQ_BUDGET);
+}
+
+static int mlx5_fpga_conn_create_cq(struct mlx5_fpga_conn *conn, int cq_size)
+{
+	struct mlx5_fpga_device *fdev = conn->fdev;
+	struct mlx5_core_dev *mdev = fdev->mdev;
+	u32 temp_cqc[MLX5_ST_SZ_DW(cqc)] = {0};
+	struct mlx5_wq_param wqp;
+	struct mlx5_cqe64 *cqe;
+	int inlen, err, eqn;
+	unsigned int irqn;
+	void *cqc, *in;
+	__be64 *pas;
+	u32 i;
+
+	cq_size = roundup_pow_of_two(cq_size);
+	MLX5_SET(cqc, temp_cqc, log_cq_size, ilog2(cq_size));
+
+	wqp.buf_numa_node = mdev->priv.numa_node;
+	wqp.db_numa_node  = mdev->priv.numa_node;
+
+	err = mlx5_cqwq_create(mdev, &wqp, temp_cqc, &conn->cq.wq,
+			       &conn->cq.wq_ctrl);
+	if (err)
+		return err;
+
+	for (i = 0; i < mlx5_cqwq_get_size(&conn->cq.wq); i++) {
+		cqe = mlx5_cqwq_get_wqe(&conn->cq.wq, i);
+		cqe->op_own = MLX5_CQE_INVALID << 4 | MLX5_CQE_OWNER_MASK;
+	}
+
+	inlen = MLX5_ST_SZ_BYTES(create_cq_in) +
+		sizeof(u64) * conn->cq.wq_ctrl.buf.npages;
+	in = kvzalloc(inlen, GFP_KERNEL);
+	if (!in) {
+		err = -ENOMEM;
+		goto err_cqwq;
+	}
+
+	err = mlx5_vector2eqn(mdev, smp_processor_id(), &eqn, &irqn);
+	if (err) {
+		kvfree(in);
+		goto err_cqwq;
+	}
+
+	cqc = MLX5_ADDR_OF(create_cq_in, in, cq_context);
+	MLX5_SET(cqc, cqc, log_cq_size, ilog2(cq_size));
+	MLX5_SET(cqc, cqc, c_eqn, eqn);
+	MLX5_SET(cqc, cqc, uar_page, fdev->conn_res.uar->index);
+	MLX5_SET(cqc, cqc, log_page_size, conn->cq.wq_ctrl.buf.page_shift -
+			   MLX5_ADAPTER_PAGE_SHIFT);
+	MLX5_SET64(cqc, cqc, dbr_addr, conn->cq.wq_ctrl.db.dma);
+
+	pas = (__be64 *)MLX5_ADDR_OF(create_cq_in, in, pas);
+	mlx5_fill_page_frag_array(&conn->cq.wq_ctrl.buf, pas);
+
+	err = mlx5_core_create_cq(mdev, &conn->cq.mcq, in, inlen);
+	kvfree(in);
+
+	if (err)
+		goto err_cqwq;
+
+	conn->cq.mcq.cqe_sz     = 64;
+	conn->cq.mcq.set_ci_db  = conn->cq.wq_ctrl.db.db;
+	conn->cq.mcq.arm_db     = conn->cq.wq_ctrl.db.db + 1;
+	*conn->cq.mcq.set_ci_db = 0;
+	*conn->cq.mcq.arm_db    = 0;
+	conn->cq.mcq.vector     = 0;
+	conn->cq.mcq.comp       = mlx5_fpga_conn_cq_complete;
+	conn->cq.mcq.event      = mlx5_fpga_conn_cq_event;
+	conn->cq.mcq.irqn       = irqn;
+	conn->cq.mcq.uar        = fdev->conn_res.uar;
+	tasklet_init(&conn->cq.tasklet, mlx5_fpga_conn_cq_tasklet,
+		     (unsigned long)conn);
+
+	mlx5_fpga_dbg(fdev, "Created CQ #0x%x\n", conn->cq.mcq.cqn);
+
+	goto out;
+
+err_cqwq:
+	mlx5_wq_destroy(&conn->cq.wq_ctrl);
+out:
+	return err;
+}
+
+static void mlx5_fpga_conn_destroy_cq(struct mlx5_fpga_conn *conn)
+{
+	tasklet_disable(&conn->cq.tasklet);
+	tasklet_kill(&conn->cq.tasklet);
+	mlx5_core_destroy_cq(conn->fdev->mdev, &conn->cq.mcq);
+	mlx5_wq_destroy(&conn->cq.wq_ctrl);
+}
+
+static int mlx5_fpga_conn_create_wq(struct mlx5_fpga_conn *conn, void *qpc)
+{
+	struct mlx5_fpga_device *fdev = conn->fdev;
+	struct mlx5_core_dev *mdev = fdev->mdev;
+	struct mlx5_wq_param wqp;
+
+	wqp.buf_numa_node = mdev->priv.numa_node;
+	wqp.db_numa_node  = mdev->priv.numa_node;
+
+	return mlx5_wq_qp_create(mdev, &wqp, qpc, &conn->qp.wq,
+				 &conn->qp.wq_ctrl);
+}
+
+static int mlx5_fpga_conn_create_qp(struct mlx5_fpga_conn *conn,
+				    unsigned int tx_size, unsigned int rx_size)
+{
+	struct mlx5_fpga_device *fdev = conn->fdev;
+	struct mlx5_core_dev *mdev = fdev->mdev;
+	u32 temp_qpc[MLX5_ST_SZ_DW(qpc)] = {0};
+	void *in = NULL, *qpc;
+	int err, inlen;
+
+	conn->qp.rq.pc = 0;
+	conn->qp.rq.cc = 0;
+	conn->qp.rq.size = roundup_pow_of_two(rx_size);
+	conn->qp.sq.pc = 0;
+	conn->qp.sq.cc = 0;
+	conn->qp.sq.size = roundup_pow_of_two(tx_size);
+
+	MLX5_SET(qpc, temp_qpc, log_rq_stride, ilog2(MLX5_SEND_WQE_DS) - 4);
+	MLX5_SET(qpc, temp_qpc, log_rq_size, ilog2(conn->qp.rq.size));
+	MLX5_SET(qpc, temp_qpc, log_sq_size, ilog2(conn->qp.sq.size));
+	err = mlx5_fpga_conn_create_wq(conn, temp_qpc);
+	if (err)
+		goto out;
+
+	conn->qp.rq.bufs = kvcalloc(conn->qp.rq.size,
+				    sizeof(conn->qp.rq.bufs[0]),
+				    GFP_KERNEL);
+	if (!conn->qp.rq.bufs) {
+		err = -ENOMEM;
+		goto err_wq;
+	}
+
+	conn->qp.sq.bufs = kvcalloc(conn->qp.sq.size,
+				    sizeof(conn->qp.sq.bufs[0]),
+				    GFP_KERNEL);
+	if (!conn->qp.sq.bufs) {
+		err = -ENOMEM;
+		goto err_rq_bufs;
+	}
+
+	inlen = MLX5_ST_SZ_BYTES(create_qp_in) +
+		MLX5_FLD_SZ_BYTES(create_qp_in, pas[0]) *
+		conn->qp.wq_ctrl.buf.npages;
+	in = kvzalloc(inlen, GFP_KERNEL);
+	if (!in) {
+		err = -ENOMEM;
+		goto err_sq_bufs;
+	}
+
+	qpc = MLX5_ADDR_OF(create_qp_in, in, qpc);
+	MLX5_SET(qpc, qpc, uar_page, fdev->conn_res.uar->index);
+	MLX5_SET(qpc, qpc, log_page_size,
+		 conn->qp.wq_ctrl.buf.page_shift - MLX5_ADAPTER_PAGE_SHIFT);
+	MLX5_SET(qpc, qpc, fre, 1);
+	MLX5_SET(qpc, qpc, rlky, 1);
+	MLX5_SET(qpc, qpc, st, MLX5_QP_ST_RC);
+	MLX5_SET(qpc, qpc, pm_state, MLX5_QP_PM_MIGRATED);
+	MLX5_SET(qpc, qpc, pd, fdev->conn_res.pdn);
+	MLX5_SET(qpc, qpc, log_rq_stride, ilog2(MLX5_SEND_WQE_DS) - 4);
+	MLX5_SET(qpc, qpc, log_rq_size, ilog2(conn->qp.rq.size));
+	MLX5_SET(qpc, qpc, rq_type, MLX5_NON_ZERO_RQ);
+	MLX5_SET(qpc, qpc, log_sq_size, ilog2(conn->qp.sq.size));
+	MLX5_SET(qpc, qpc, cqn_snd, conn->cq.mcq.cqn);
+	MLX5_SET(qpc, qpc, cqn_rcv, conn->cq.mcq.cqn);
+	MLX5_SET64(qpc, qpc, dbr_addr, conn->qp.wq_ctrl.db.dma);
+	if (MLX5_CAP_GEN(mdev, cqe_version) == 1)
+		MLX5_SET(qpc, qpc, user_index, 0xFFFFFF);
+
+	mlx5_fill_page_frag_array(&conn->qp.wq_ctrl.buf,
+				  (__be64 *)MLX5_ADDR_OF(create_qp_in, in, pas));
+
+	err = mlx5_core_create_qp(mdev, &conn->qp.mqp, in, inlen);
+	if (err)
+		goto err_sq_bufs;
+
+	conn->qp.mqp.event = mlx5_fpga_conn_event;
+	mlx5_fpga_dbg(fdev, "Created QP #0x%x\n", conn->qp.mqp.qpn);
+
+	goto out;
+
+err_sq_bufs:
+	kvfree(conn->qp.sq.bufs);
+err_rq_bufs:
+	kvfree(conn->qp.rq.bufs);
+err_wq:
+	mlx5_wq_destroy(&conn->qp.wq_ctrl);
+out:
+	kvfree(in);
+	return err;
+}
+
+static void mlx5_fpga_conn_free_recv_bufs(struct mlx5_fpga_conn *conn)
+{
+	int ix;
+
+	for (ix = 0; ix < conn->qp.rq.size; ix++) {
+		if (!conn->qp.rq.bufs[ix])
+			continue;
+		mlx5_fpga_conn_unmap_buf(conn, conn->qp.rq.bufs[ix]);
+		kfree(conn->qp.rq.bufs[ix]);
+		conn->qp.rq.bufs[ix] = NULL;
+	}
+}
+
+static void mlx5_fpga_conn_flush_send_bufs(struct mlx5_fpga_conn *conn)
+{
+	struct mlx5_fpga_dma_buf *buf, *temp;
+	int ix;
+
+	for (ix = 0; ix < conn->qp.sq.size; ix++) {
+		buf = conn->qp.sq.bufs[ix];
+		if (!buf)
+			continue;
+		conn->qp.sq.bufs[ix] = NULL;
+		mlx5_fpga_conn_unmap_buf(conn, buf);
+		if (!buf->complete)
+			continue;
+		buf->complete(conn, conn->fdev, buf, MLX5_CQE_SYNDROME_WR_FLUSH_ERR);
+	}
+	list_for_each_entry_safe(buf, temp, &conn->qp.sq.backlog, list) {
+		mlx5_fpga_conn_unmap_buf(conn, buf);
+		if (!buf->complete)
+			continue;
+		buf->complete(conn, conn->fdev, buf, MLX5_CQE_SYNDROME_WR_FLUSH_ERR);
+	}
+}
+
+static void mlx5_fpga_conn_destroy_qp(struct mlx5_fpga_conn *conn)
+{
+	mlx5_core_destroy_qp(conn->fdev->mdev, &conn->qp.mqp);
+	mlx5_fpga_conn_free_recv_bufs(conn);
+	mlx5_fpga_conn_flush_send_bufs(conn);
+	kvfree(conn->qp.sq.bufs);
+	kvfree(conn->qp.rq.bufs);
+	mlx5_wq_destroy(&conn->qp.wq_ctrl);
+}
+
+static inline int mlx5_fpga_conn_reset_qp(struct mlx5_fpga_conn *conn)
+{
+	struct mlx5_core_dev *mdev = conn->fdev->mdev;
+
+	mlx5_fpga_dbg(conn->fdev, "Modifying QP %u to RST\n", conn->qp.mqp.qpn);
+
+	return mlx5_core_qp_modify(mdev, MLX5_CMD_OP_2RST_QP, 0, NULL,
+				   &conn->qp.mqp);
+}
+
+static inline int mlx5_fpga_conn_init_qp(struct mlx5_fpga_conn *conn)
+{
+	struct mlx5_fpga_device *fdev = conn->fdev;
+	struct mlx5_core_dev *mdev = fdev->mdev;
+	u32 *qpc = NULL;
+	int err;
+
+	mlx5_fpga_dbg(conn->fdev, "Modifying QP %u to INIT\n", conn->qp.mqp.qpn);
+
+	qpc = kzalloc(MLX5_ST_SZ_BYTES(qpc), GFP_KERNEL);
+	if (!qpc) {
+		err = -ENOMEM;
+		goto out;
+	}
+
+	MLX5_SET(qpc, qpc, st, MLX5_QP_ST_RC);
+	MLX5_SET(qpc, qpc, pm_state, MLX5_QP_PM_MIGRATED);
+	MLX5_SET(qpc, qpc, primary_address_path.pkey_index, MLX5_FPGA_PKEY_INDEX);
+	MLX5_SET(qpc, qpc, primary_address_path.vhca_port_num, MLX5_FPGA_PORT_NUM);
+	MLX5_SET(qpc, qpc, pd, conn->fdev->conn_res.pdn);
+	MLX5_SET(qpc, qpc, cqn_snd, conn->cq.mcq.cqn);
+	MLX5_SET(qpc, qpc, cqn_rcv, conn->cq.mcq.cqn);
+	MLX5_SET64(qpc, qpc, dbr_addr, conn->qp.wq_ctrl.db.dma);
+
+	err = mlx5_core_qp_modify(mdev, MLX5_CMD_OP_RST2INIT_QP, 0, qpc,
+				  &conn->qp.mqp);
+	if (err) {
+		mlx5_fpga_warn(fdev, "qp_modify RST2INIT failed: %d\n", err);
+		goto out;
+	}
+
+out:
+	kfree(qpc);
+	return err;
+}
+
+static inline int mlx5_fpga_conn_rtr_qp(struct mlx5_fpga_conn *conn)
+{
+	struct mlx5_fpga_device *fdev = conn->fdev;
+	struct mlx5_core_dev *mdev = fdev->mdev;
+	u32 *qpc = NULL;
+	int err;
+
+	mlx5_fpga_dbg(conn->fdev, "QP RTR\n");
+
+	qpc = kzalloc(MLX5_ST_SZ_BYTES(qpc), GFP_KERNEL);
+	if (!qpc) {
+		err = -ENOMEM;
+		goto out;
+	}
+
+	MLX5_SET(qpc, qpc, mtu, MLX5_QPC_MTU_1K_BYTES);
+	MLX5_SET(qpc, qpc, log_msg_max, (u8)MLX5_CAP_GEN(mdev, log_max_msg));
+	MLX5_SET(qpc, qpc, remote_qpn, conn->fpga_qpn);
+	MLX5_SET(qpc, qpc, next_rcv_psn,
+		 MLX5_GET(fpga_qpc, conn->fpga_qpc, next_send_psn));
+	MLX5_SET(qpc, qpc, primary_address_path.pkey_index, MLX5_FPGA_PKEY_INDEX);
+	MLX5_SET(qpc, qpc, primary_address_path.vhca_port_num, MLX5_FPGA_PORT_NUM);
+	ether_addr_copy(MLX5_ADDR_OF(qpc, qpc, primary_address_path.rmac_47_32),
+			MLX5_ADDR_OF(fpga_qpc, conn->fpga_qpc, fpga_mac_47_32));
+	MLX5_SET(qpc, qpc, primary_address_path.udp_sport,
+		 MLX5_CAP_ROCE(mdev, r_roce_min_src_udp_port));
+	MLX5_SET(qpc, qpc, primary_address_path.src_addr_index,
+		 conn->qp.sgid_index);
+	MLX5_SET(qpc, qpc, primary_address_path.hop_limit, 0);
+	memcpy(MLX5_ADDR_OF(qpc, qpc, primary_address_path.rgid_rip),
+	       MLX5_ADDR_OF(fpga_qpc, conn->fpga_qpc, fpga_ip),
+	       MLX5_FLD_SZ_BYTES(qpc, primary_address_path.rgid_rip));
+
+	err = mlx5_core_qp_modify(mdev, MLX5_CMD_OP_INIT2RTR_QP, 0, qpc,
+				  &conn->qp.mqp);
+	if (err) {
+		mlx5_fpga_warn(fdev, "qp_modify RST2INIT failed: %d\n", err);
+		goto out;
+	}
+
+out:
+	kfree(qpc);
+	return err;
+}
+
+static inline int mlx5_fpga_conn_rts_qp(struct mlx5_fpga_conn *conn)
+{
+	struct mlx5_fpga_device *fdev = conn->fdev;
+	struct mlx5_core_dev *mdev = fdev->mdev;
+	u32 *qpc = NULL;
+	u32 opt_mask;
+	int err;
+
+	mlx5_fpga_dbg(conn->fdev, "QP RTS\n");
+
+	qpc = kzalloc(MLX5_ST_SZ_BYTES(qpc), GFP_KERNEL);
+	if (!qpc) {
+		err = -ENOMEM;
+		goto out;
+	}
+
+	MLX5_SET(qpc, qpc, log_ack_req_freq, 8);
+	MLX5_SET(qpc, qpc, min_rnr_nak, 0x12);
+	MLX5_SET(qpc, qpc, primary_address_path.ack_timeout, 0x12); /* ~1.07s */
+	MLX5_SET(qpc, qpc, next_send_psn,
+		 MLX5_GET(fpga_qpc, conn->fpga_qpc, next_rcv_psn));
+	MLX5_SET(qpc, qpc, retry_count, 7);
+	MLX5_SET(qpc, qpc, rnr_retry, 7); /* Infinite retry if RNR NACK */
+
+	opt_mask = MLX5_QP_OPTPAR_RNR_TIMEOUT;
+	err = mlx5_core_qp_modify(mdev, MLX5_CMD_OP_RTR2RTS_QP, opt_mask, qpc,
+				  &conn->qp.mqp);
+	if (err) {
+		mlx5_fpga_warn(fdev, "qp_modify RST2INIT failed: %d\n", err);
+		goto out;
+	}
+
+out:
+	kfree(qpc);
+	return err;
+}
+
+static int mlx5_fpga_conn_connect(struct mlx5_fpga_conn *conn)
+{
+	struct mlx5_fpga_device *fdev = conn->fdev;
+	int err;
+
+	MLX5_SET(fpga_qpc, conn->fpga_qpc, state, MLX5_FPGA_QPC_STATE_ACTIVE);
+	err = mlx5_fpga_modify_qp(conn->fdev->mdev, conn->fpga_qpn,
+				  MLX5_FPGA_QPC_STATE, &conn->fpga_qpc);
+	if (err) {
+		mlx5_fpga_err(fdev, "Failed to activate FPGA RC QP: %d\n", err);
+		goto out;
+	}
+
+	err = mlx5_fpga_conn_reset_qp(conn);
+	if (err) {
+		mlx5_fpga_err(fdev, "Failed to change QP state to reset\n");
+		goto err_fpga_qp;
+	}
+
+	err = mlx5_fpga_conn_init_qp(conn);
+	if (err) {
+		mlx5_fpga_err(fdev, "Failed to modify QP from RESET to INIT\n");
+		goto err_fpga_qp;
+	}
+	conn->qp.active = true;
+
+	while (!mlx5_fpga_conn_post_recv_buf(conn))
+		;
+
+	err = mlx5_fpga_conn_rtr_qp(conn);
+	if (err) {
+		mlx5_fpga_err(fdev, "Failed to change QP state from INIT to RTR\n");
+		goto err_recv_bufs;
+	}
+
+	err = mlx5_fpga_conn_rts_qp(conn);
+	if (err) {
+		mlx5_fpga_err(fdev, "Failed to change QP state from RTR to RTS\n");
+		goto err_recv_bufs;
+	}
+	goto out;
+
+err_recv_bufs:
+	mlx5_fpga_conn_free_recv_bufs(conn);
+err_fpga_qp:
+	MLX5_SET(fpga_qpc, conn->fpga_qpc, state, MLX5_FPGA_QPC_STATE_INIT);
+	if (mlx5_fpga_modify_qp(conn->fdev->mdev, conn->fpga_qpn,
+				MLX5_FPGA_QPC_STATE, &conn->fpga_qpc))
+		mlx5_fpga_err(fdev, "Failed to revert FPGA QP to INIT\n");
+out:
+	return err;
+}
+
+struct mlx5_fpga_conn *mlx5_fpga_conn_create(struct mlx5_fpga_device *fdev,
+					     struct mlx5_fpga_conn_attr *attr,
+					     enum mlx5_ifc_fpga_qp_type qp_type)
+{
+	struct mlx5_fpga_conn *ret, *conn;
+	u8 *remote_mac, *remote_ip;
+	int err;
+
+	if (!attr->recv_cb)
+		return ERR_PTR(-EINVAL);
+
+	conn = kzalloc(sizeof(*conn), GFP_KERNEL);
+	if (!conn)
+		return ERR_PTR(-ENOMEM);
+
+	conn->fdev = fdev;
+	INIT_LIST_HEAD(&conn->qp.sq.backlog);
+
+	spin_lock_init(&conn->qp.sq.lock);
+
+	conn->recv_cb = attr->recv_cb;
+	conn->cb_arg = attr->cb_arg;
+
+	remote_mac = MLX5_ADDR_OF(fpga_qpc, conn->fpga_qpc, remote_mac_47_32);
+	err = mlx5_query_nic_vport_mac_address(fdev->mdev, 0, remote_mac);
+	if (err) {
+		mlx5_fpga_err(fdev, "Failed to query local MAC: %d\n", err);
+		ret = ERR_PTR(err);
+		goto err;
+	}
+
+	/* Build Modified EUI-64 IPv6 address from the MAC address */
+	remote_ip = MLX5_ADDR_OF(fpga_qpc, conn->fpga_qpc, remote_ip);
+	remote_ip[0] = 0xfe;
+	remote_ip[1] = 0x80;
+	addrconf_addr_eui48(&remote_ip[8], remote_mac);
+
+	err = mlx5_core_reserved_gid_alloc(fdev->mdev, &conn->qp.sgid_index);
+	if (err) {
+		mlx5_fpga_err(fdev, "Failed to allocate SGID: %d\n", err);
+		ret = ERR_PTR(err);
+		goto err;
+	}
+
+	err = mlx5_core_roce_gid_set(fdev->mdev, conn->qp.sgid_index,
+				     MLX5_ROCE_VERSION_2,
+				     MLX5_ROCE_L3_TYPE_IPV6,
+				     remote_ip, remote_mac, true, 0,
+				     MLX5_FPGA_PORT_NUM);
+	if (err) {
+		mlx5_fpga_err(fdev, "Failed to set SGID: %d\n", err);
+		ret = ERR_PTR(err);
+		goto err_rsvd_gid;
+	}
+	mlx5_fpga_dbg(fdev, "Reserved SGID index %u\n", conn->qp.sgid_index);
+
+	/* Allow for one cqe per rx/tx wqe, plus one cqe for the next wqe,
+	 * created during processing of the cqe
+	 */
+	err = mlx5_fpga_conn_create_cq(conn,
+				       (attr->tx_size + attr->rx_size) * 2);
+	if (err) {
+		mlx5_fpga_err(fdev, "Failed to create CQ: %d\n", err);
+		ret = ERR_PTR(err);
+		goto err_gid;
+	}
+
+	mlx5_fpga_conn_arm_cq(conn);
+
+	err = mlx5_fpga_conn_create_qp(conn, attr->tx_size, attr->rx_size);
+	if (err) {
+		mlx5_fpga_err(fdev, "Failed to create QP: %d\n", err);
+		ret = ERR_PTR(err);
+		goto err_cq;
+	}
+
+	MLX5_SET(fpga_qpc, conn->fpga_qpc, state, MLX5_FPGA_QPC_STATE_INIT);
+	MLX5_SET(fpga_qpc, conn->fpga_qpc, qp_type, qp_type);
+	MLX5_SET(fpga_qpc, conn->fpga_qpc, st, MLX5_FPGA_QPC_ST_RC);
+	MLX5_SET(fpga_qpc, conn->fpga_qpc, ether_type, ETH_P_8021Q);
+	MLX5_SET(fpga_qpc, conn->fpga_qpc, vid, 0);
+	MLX5_SET(fpga_qpc, conn->fpga_qpc, next_rcv_psn, 1);
+	MLX5_SET(fpga_qpc, conn->fpga_qpc, next_send_psn, 0);
+	MLX5_SET(fpga_qpc, conn->fpga_qpc, pkey, MLX5_FPGA_PKEY);
+	MLX5_SET(fpga_qpc, conn->fpga_qpc, remote_qpn, conn->qp.mqp.qpn);
+	MLX5_SET(fpga_qpc, conn->fpga_qpc, rnr_retry, 7);
+	MLX5_SET(fpga_qpc, conn->fpga_qpc, retry_count, 7);
+
+	err = mlx5_fpga_create_qp(fdev->mdev, &conn->fpga_qpc,
+				  &conn->fpga_qpn);
+	if (err) {
+		mlx5_fpga_err(fdev, "Failed to create FPGA RC QP: %d\n", err);
+		ret = ERR_PTR(err);
+		goto err_qp;
+	}
+
+	err = mlx5_fpga_conn_connect(conn);
+	if (err) {
+		ret = ERR_PTR(err);
+		goto err_conn;
+	}
+
+	mlx5_fpga_dbg(fdev, "FPGA QPN is %u\n", conn->fpga_qpn);
+	ret = conn;
+	goto out;
+
+err_conn:
+	mlx5_fpga_destroy_qp(conn->fdev->mdev, conn->fpga_qpn);
+err_qp:
+	mlx5_fpga_conn_destroy_qp(conn);
+err_cq:
+	mlx5_fpga_conn_destroy_cq(conn);
+err_gid:
+	mlx5_core_roce_gid_set(fdev->mdev, conn->qp.sgid_index, 0, 0, NULL,
+			       NULL, false, 0, MLX5_FPGA_PORT_NUM);
+err_rsvd_gid:
+	mlx5_core_reserved_gid_free(fdev->mdev, conn->qp.sgid_index);
+err:
+	kfree(conn);
+out:
+	return ret;
+}
+
+void mlx5_fpga_conn_destroy(struct mlx5_fpga_conn *conn)
+{
+	struct mlx5_fpga_device *fdev = conn->fdev;
+	struct mlx5_core_dev *mdev = fdev->mdev;
+	int err = 0;
+
+	conn->qp.active = false;
+	tasklet_disable(&conn->cq.tasklet);
+	synchronize_irq(conn->cq.mcq.irqn);
+
+	mlx5_fpga_destroy_qp(conn->fdev->mdev, conn->fpga_qpn);
+	err = mlx5_core_qp_modify(mdev, MLX5_CMD_OP_2ERR_QP, 0, NULL,
+				  &conn->qp.mqp);
+	if (err)
+		mlx5_fpga_warn(fdev, "qp_modify 2ERR failed: %d\n", err);
+	mlx5_fpga_conn_destroy_qp(conn);
+	mlx5_fpga_conn_destroy_cq(conn);
+
+	mlx5_core_roce_gid_set(conn->fdev->mdev, conn->qp.sgid_index, 0, 0,
+			       NULL, NULL, false, 0, MLX5_FPGA_PORT_NUM);
+	mlx5_core_reserved_gid_free(conn->fdev->mdev, conn->qp.sgid_index);
+	kfree(conn);
+}
+
+int mlx5_fpga_conn_device_init(struct mlx5_fpga_device *fdev)
+{
+	int err;
+
+	err = mlx5_nic_vport_enable_roce(fdev->mdev);
+	if (err) {
+		mlx5_fpga_err(fdev, "Failed to enable RoCE: %d\n", err);
+		goto out;
+	}
+
+	fdev->conn_res.uar = mlx5_get_uars_page(fdev->mdev);
+	if (IS_ERR(fdev->conn_res.uar)) {
+		err = PTR_ERR(fdev->conn_res.uar);
+		mlx5_fpga_err(fdev, "get_uars_page failed, %d\n", err);
+		goto err_roce;
+	}
+	mlx5_fpga_dbg(fdev, "Allocated UAR index %u\n",
+		      fdev->conn_res.uar->index);
+
+	err = mlx5_core_alloc_pd(fdev->mdev, &fdev->conn_res.pdn);
+	if (err) {
+		mlx5_fpga_err(fdev, "alloc pd failed, %d\n", err);
+		goto err_uar;
+	}
+	mlx5_fpga_dbg(fdev, "Allocated PD %u\n", fdev->conn_res.pdn);
+
+	err = mlx5_fpga_conn_create_mkey(fdev->mdev, fdev->conn_res.pdn,
+					 &fdev->conn_res.mkey);
+	if (err) {
+		mlx5_fpga_err(fdev, "create mkey failed, %d\n", err);
+		goto err_dealloc_pd;
+	}
+	mlx5_fpga_dbg(fdev, "Created mkey 0x%x\n", fdev->conn_res.mkey.key);
+
+	return 0;
+
+err_dealloc_pd:
+	mlx5_core_dealloc_pd(fdev->mdev, fdev->conn_res.pdn);
+err_uar:
+	mlx5_put_uars_page(fdev->mdev, fdev->conn_res.uar);
+err_roce:
+	mlx5_nic_vport_disable_roce(fdev->mdev);
+out:
+	return err;
+}
+
+void mlx5_fpga_conn_device_cleanup(struct mlx5_fpga_device *fdev)
+{
+	mlx5_core_destroy_mkey(fdev->mdev, &fdev->conn_res.mkey);
+	mlx5_core_dealloc_pd(fdev->mdev, fdev->conn_res.pdn);
+	mlx5_put_uars_page(fdev->mdev, fdev->conn_res.uar);
+	mlx5_nic_vport_disable_roce(fdev->mdev);
+}
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/fpga/conn.h b/drivers/net/ethernet/mellanox/mlx5/core/fpga/conn.h
new file mode 100644
index 000000000..634ae10e2
--- /dev/null
+++ b/drivers/net/ethernet/mellanox/mlx5/core/fpga/conn.h
@@ -0,0 +1,96 @@
+/*
+ * Copyright (c) 2017 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 __MLX5_FPGA_CONN_H__
+#define __MLX5_FPGA_CONN_H__
+
+#include <linux/mlx5/cq.h>
+#include <linux/mlx5/qp.h>
+
+#include "fpga/core.h"
+#include "fpga/sdk.h"
+#include "wq.h"
+
+struct mlx5_fpga_conn {
+	struct mlx5_fpga_device *fdev;
+
+	void (*recv_cb)(void *cb_arg, struct mlx5_fpga_dma_buf *buf);
+	void *cb_arg;
+
+	/* FPGA QP */
+	u32 fpga_qpc[MLX5_ST_SZ_DW(fpga_qpc)];
+	u32 fpga_qpn;
+
+	/* CQ */
+	struct {
+		struct mlx5_cqwq wq;
+		struct mlx5_wq_ctrl wq_ctrl;
+		struct mlx5_core_cq mcq;
+		struct tasklet_struct tasklet;
+	} cq;
+
+	/* QP */
+	struct {
+		bool active;
+		int sgid_index;
+		struct mlx5_wq_qp wq;
+		struct mlx5_wq_ctrl wq_ctrl;
+		struct mlx5_core_qp mqp;
+		struct {
+			spinlock_t lock; /* Protects all SQ state */
+			unsigned int pc;
+			unsigned int cc;
+			unsigned int size;
+			struct mlx5_fpga_dma_buf **bufs;
+			struct list_head backlog;
+		} sq;
+		struct {
+			unsigned int pc;
+			unsigned int cc;
+			unsigned int size;
+			struct mlx5_fpga_dma_buf **bufs;
+		} rq;
+	} qp;
+};
+
+int mlx5_fpga_conn_device_init(struct mlx5_fpga_device *fdev);
+void mlx5_fpga_conn_device_cleanup(struct mlx5_fpga_device *fdev);
+struct mlx5_fpga_conn *
+mlx5_fpga_conn_create(struct mlx5_fpga_device *fdev,
+		      struct mlx5_fpga_conn_attr *attr,
+		      enum mlx5_ifc_fpga_qp_type qp_type);
+void mlx5_fpga_conn_destroy(struct mlx5_fpga_conn *conn);
+int mlx5_fpga_conn_send(struct mlx5_fpga_conn *conn,
+			struct mlx5_fpga_dma_buf *buf);
+
+#endif /* __MLX5_FPGA_CONN_H__ */
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/fpga/core.c b/drivers/net/ethernet/mellanox/mlx5/core/fpga/core.c
new file mode 100644
index 000000000..310f9e7d8
--- /dev/null
+++ b/drivers/net/ethernet/mellanox/mlx5/core/fpga/core.c
@@ -0,0 +1,327 @@
+/*
+ * Copyright (c) 2017, 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/module.h>
+#include <linux/etherdevice.h>
+#include <linux/mlx5/driver.h>
+
+#include "mlx5_core.h"
+#include "lib/mlx5.h"
+#include "fpga/core.h"
+#include "fpga/conn.h"
+
+static const char *const mlx5_fpga_error_strings[] = {
+	"Null Syndrome",
+	"Corrupted DDR",
+	"Flash Timeout",
+	"Internal Link Error",
+	"Watchdog HW Failure",
+	"I2C Failure",
+	"Image Changed",
+	"Temperature Critical",
+};
+
+static const char * const mlx5_fpga_qp_error_strings[] = {
+	"Null Syndrome",
+	"Retry Counter Expired",
+	"RNR Expired",
+};
+static struct mlx5_fpga_device *mlx5_fpga_device_alloc(void)
+{
+	struct mlx5_fpga_device *fdev = NULL;
+
+	fdev = kzalloc(sizeof(*fdev), GFP_KERNEL);
+	if (!fdev)
+		return NULL;
+
+	spin_lock_init(&fdev->state_lock);
+	fdev->state = MLX5_FPGA_STATUS_NONE;
+	return fdev;
+}
+
+static const char *mlx5_fpga_image_name(enum mlx5_fpga_image image)
+{
+	switch (image) {
+	case MLX5_FPGA_IMAGE_USER:
+		return "user";
+	case MLX5_FPGA_IMAGE_FACTORY:
+		return "factory";
+	default:
+		return "unknown";
+	}
+}
+
+static const char *mlx5_fpga_device_name(u32 device)
+{
+	switch (device) {
+	case MLX5_FPGA_DEVICE_KU040:
+		return "ku040";
+	case MLX5_FPGA_DEVICE_KU060:
+		return "ku060";
+	case MLX5_FPGA_DEVICE_KU060_2:
+		return "ku060_2";
+	case MLX5_FPGA_DEVICE_UNKNOWN:
+	default:
+		return "unknown";
+	}
+}
+
+static int mlx5_fpga_device_load_check(struct mlx5_fpga_device *fdev)
+{
+	struct mlx5_fpga_query query;
+	int err;
+
+	err = mlx5_fpga_query(fdev->mdev, &query);
+	if (err) {
+		mlx5_fpga_err(fdev, "Failed to query status: %d\n", err);
+		return err;
+	}
+
+	fdev->last_admin_image = query.admin_image;
+	fdev->last_oper_image = query.oper_image;
+
+	mlx5_fpga_dbg(fdev, "Status %u; Admin image %u; Oper image %u\n",
+		      query.status, query.admin_image, query.oper_image);
+
+	if (query.status != MLX5_FPGA_STATUS_SUCCESS) {
+		mlx5_fpga_err(fdev, "%s image failed to load; status %u\n",
+			      mlx5_fpga_image_name(fdev->last_oper_image),
+			      query.status);
+		return -EIO;
+	}
+
+	return 0;
+}
+
+static int mlx5_fpga_device_brb(struct mlx5_fpga_device *fdev)
+{
+	int err;
+	struct mlx5_core_dev *mdev = fdev->mdev;
+
+	err = mlx5_fpga_ctrl_op(mdev, MLX5_FPGA_CTRL_OPERATION_SANDBOX_BYPASS_ON);
+	if (err) {
+		mlx5_fpga_err(fdev, "Failed to set bypass on: %d\n", err);
+		return err;
+	}
+	err = mlx5_fpga_ctrl_op(mdev, MLX5_FPGA_CTRL_OPERATION_RESET_SANDBOX);
+	if (err) {
+		mlx5_fpga_err(fdev, "Failed to reset SBU: %d\n", err);
+		return err;
+	}
+	err = mlx5_fpga_ctrl_op(mdev, MLX5_FPGA_CTRL_OPERATION_SANDBOX_BYPASS_OFF);
+	if (err) {
+		mlx5_fpga_err(fdev, "Failed to set bypass off: %d\n", err);
+		return err;
+	}
+	return 0;
+}
+
+int mlx5_fpga_device_start(struct mlx5_core_dev *mdev)
+{
+	struct mlx5_fpga_device *fdev = mdev->fpga;
+	unsigned int max_num_qps;
+	unsigned long flags;
+	u32 fpga_device_id;
+	int err;
+
+	if (!fdev)
+		return 0;
+
+	err = mlx5_fpga_device_load_check(fdev);
+	if (err)
+		goto out;
+
+	err = mlx5_fpga_caps(fdev->mdev);
+	if (err)
+		goto out;
+
+	fpga_device_id = MLX5_CAP_FPGA(fdev->mdev, fpga_device);
+	mlx5_fpga_info(fdev, "%s:%u; %s image, version %u; SBU %06x:%04x version %d\n",
+		       mlx5_fpga_device_name(fpga_device_id),
+		       fpga_device_id,
+		       mlx5_fpga_image_name(fdev->last_oper_image),
+		       MLX5_CAP_FPGA(fdev->mdev, image_version),
+		       MLX5_CAP_FPGA(fdev->mdev, ieee_vendor_id),
+		       MLX5_CAP_FPGA(fdev->mdev, sandbox_product_id),
+		       MLX5_CAP_FPGA(fdev->mdev, sandbox_product_version));
+
+	max_num_qps = MLX5_CAP_FPGA(mdev, shell_caps.max_num_qps);
+	if (!max_num_qps) {
+		mlx5_fpga_err(fdev, "FPGA reports 0 QPs in SHELL_CAPS\n");
+		err = -ENOTSUPP;
+		goto out;
+	}
+
+	err = mlx5_core_reserve_gids(mdev, max_num_qps);
+	if (err)
+		goto out;
+
+	err = mlx5_fpga_conn_device_init(fdev);
+	if (err)
+		goto err_rsvd_gid;
+
+	if (fdev->last_oper_image == MLX5_FPGA_IMAGE_USER) {
+		err = mlx5_fpga_device_brb(fdev);
+		if (err)
+			goto err_conn_init;
+	}
+
+	goto out;
+
+err_conn_init:
+	mlx5_fpga_conn_device_cleanup(fdev);
+
+err_rsvd_gid:
+	mlx5_core_unreserve_gids(mdev, max_num_qps);
+out:
+	spin_lock_irqsave(&fdev->state_lock, flags);
+	fdev->state = err ? MLX5_FPGA_STATUS_FAILURE : MLX5_FPGA_STATUS_SUCCESS;
+	spin_unlock_irqrestore(&fdev->state_lock, flags);
+	return err;
+}
+
+int mlx5_fpga_init(struct mlx5_core_dev *mdev)
+{
+	struct mlx5_fpga_device *fdev = NULL;
+
+	if (!MLX5_CAP_GEN(mdev, fpga)) {
+		mlx5_core_dbg(mdev, "FPGA capability not present\n");
+		return 0;
+	}
+
+	mlx5_core_dbg(mdev, "Initializing FPGA\n");
+
+	fdev = mlx5_fpga_device_alloc();
+	if (!fdev)
+		return -ENOMEM;
+
+	fdev->mdev = mdev;
+	mdev->fpga = fdev;
+
+	return 0;
+}
+
+void mlx5_fpga_device_stop(struct mlx5_core_dev *mdev)
+{
+	struct mlx5_fpga_device *fdev = mdev->fpga;
+	unsigned int max_num_qps;
+	unsigned long flags;
+	int err;
+
+	if (!fdev)
+		return;
+
+	spin_lock_irqsave(&fdev->state_lock, flags);
+	if (fdev->state != MLX5_FPGA_STATUS_SUCCESS) {
+		spin_unlock_irqrestore(&fdev->state_lock, flags);
+		return;
+	}
+	fdev->state = MLX5_FPGA_STATUS_NONE;
+	spin_unlock_irqrestore(&fdev->state_lock, flags);
+
+	if (fdev->last_oper_image == MLX5_FPGA_IMAGE_USER) {
+		err = mlx5_fpga_ctrl_op(mdev, MLX5_FPGA_CTRL_OPERATION_SANDBOX_BYPASS_ON);
+		if (err)
+			mlx5_fpga_err(fdev, "Failed to re-set SBU bypass on: %d\n",
+				      err);
+	}
+
+	mlx5_fpga_conn_device_cleanup(fdev);
+	max_num_qps = MLX5_CAP_FPGA(mdev, shell_caps.max_num_qps);
+	mlx5_core_unreserve_gids(mdev, max_num_qps);
+}
+
+void mlx5_fpga_cleanup(struct mlx5_core_dev *mdev)
+{
+	struct mlx5_fpga_device *fdev = mdev->fpga;
+
+	mlx5_fpga_device_stop(mdev);
+	kfree(fdev);
+	mdev->fpga = NULL;
+}
+
+static const char *mlx5_fpga_syndrome_to_string(u8 syndrome)
+{
+	if (syndrome < ARRAY_SIZE(mlx5_fpga_error_strings))
+		return mlx5_fpga_error_strings[syndrome];
+	return "Unknown";
+}
+
+static const char *mlx5_fpga_qp_syndrome_to_string(u8 syndrome)
+{
+	if (syndrome < ARRAY_SIZE(mlx5_fpga_qp_error_strings))
+		return mlx5_fpga_qp_error_strings[syndrome];
+	return "Unknown";
+}
+
+void mlx5_fpga_event(struct mlx5_core_dev *mdev, u8 event, void *data)
+{
+	struct mlx5_fpga_device *fdev = mdev->fpga;
+	const char *event_name;
+	bool teardown = false;
+	unsigned long flags;
+	u8 syndrome;
+
+	switch (event) {
+	case MLX5_EVENT_TYPE_FPGA_ERROR:
+		syndrome = MLX5_GET(fpga_error_event, data, syndrome);
+		event_name = mlx5_fpga_syndrome_to_string(syndrome);
+		break;
+	case MLX5_EVENT_TYPE_FPGA_QP_ERROR:
+		syndrome = MLX5_GET(fpga_qp_error_event, data, syndrome);
+		event_name = mlx5_fpga_qp_syndrome_to_string(syndrome);
+		break;
+	default:
+		mlx5_fpga_warn_ratelimited(fdev, "Unexpected event %u\n",
+					   event);
+		return;
+	}
+
+	spin_lock_irqsave(&fdev->state_lock, flags);
+	switch (fdev->state) {
+	case MLX5_FPGA_STATUS_SUCCESS:
+		mlx5_fpga_warn(fdev, "Error %u: %s\n", syndrome, event_name);
+		teardown = true;
+		break;
+	default:
+		mlx5_fpga_warn_ratelimited(fdev, "Unexpected error event %u: %s\n",
+					   syndrome, event_name);
+	}
+	spin_unlock_irqrestore(&fdev->state_lock, flags);
+	/* We tear-down the card's interfaces and functionality because
+	 * the FPGA bump-on-the-wire is misbehaving and we lose ability
+	 * to communicate with the network. User may still be able to
+	 * recover by re-programming or debugging the FPGA
+	 */
+	if (teardown)
+		mlx5_trigger_health_work(fdev->mdev);
+}
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/fpga/core.h b/drivers/net/ethernet/mellanox/mlx5/core/fpga/core.h
new file mode 100644
index 000000000..3e2355c8d
--- /dev/null
+++ b/drivers/net/ethernet/mellanox/mlx5/core/fpga/core.h
@@ -0,0 +1,114 @@
+/*
+ * Copyright (c) 2017, Mellanox Technologies, Ltd.  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 __MLX5_FPGA_CORE_H__
+#define __MLX5_FPGA_CORE_H__
+
+#ifdef CONFIG_MLX5_FPGA
+
+#include "fpga/cmd.h"
+
+/* Represents an Innova device */
+struct mlx5_fpga_device {
+	struct mlx5_core_dev *mdev;
+	spinlock_t state_lock; /* Protects state transitions */
+	enum mlx5_fpga_status state;
+	enum mlx5_fpga_image last_admin_image;
+	enum mlx5_fpga_image last_oper_image;
+
+	/* QP Connection resources */
+	struct {
+		u32 pdn;
+		struct mlx5_core_mkey mkey;
+		struct mlx5_uars_page *uar;
+	} conn_res;
+
+	struct mlx5_fpga_ipsec *ipsec;
+	struct mlx5_fpga_tls *tls;
+};
+
+#define mlx5_fpga_dbg(__adev, format, ...) \
+	dev_dbg(&(__adev)->mdev->pdev->dev, "FPGA: %s:%d:(pid %d): " format, \
+		 __func__, __LINE__, current->pid, ##__VA_ARGS__)
+
+#define mlx5_fpga_err(__adev, format, ...) \
+	dev_err(&(__adev)->mdev->pdev->dev, "FPGA: %s:%d:(pid %d): " format, \
+		__func__, __LINE__, current->pid, ##__VA_ARGS__)
+
+#define mlx5_fpga_warn(__adev, format, ...) \
+	dev_warn(&(__adev)->mdev->pdev->dev, "FPGA: %s:%d:(pid %d): " format, \
+		__func__, __LINE__, current->pid, ##__VA_ARGS__)
+
+#define mlx5_fpga_warn_ratelimited(__adev, format, ...) \
+	dev_warn_ratelimited(&(__adev)->mdev->pdev->dev, "FPGA: %s:%d: " \
+		format, __func__, __LINE__, ##__VA_ARGS__)
+
+#define mlx5_fpga_notice(__adev, format, ...) \
+	dev_notice(&(__adev)->mdev->pdev->dev, "FPGA: " format, ##__VA_ARGS__)
+
+#define mlx5_fpga_info(__adev, format, ...) \
+	dev_info(&(__adev)->mdev->pdev->dev, "FPGA: " format, ##__VA_ARGS__)
+
+int mlx5_fpga_init(struct mlx5_core_dev *mdev);
+void mlx5_fpga_cleanup(struct mlx5_core_dev *mdev);
+int mlx5_fpga_device_start(struct mlx5_core_dev *mdev);
+void mlx5_fpga_device_stop(struct mlx5_core_dev *mdev);
+void mlx5_fpga_event(struct mlx5_core_dev *mdev, u8 event, void *data);
+
+#else
+
+static inline int mlx5_fpga_init(struct mlx5_core_dev *mdev)
+{
+	return 0;
+}
+
+static inline void mlx5_fpga_cleanup(struct mlx5_core_dev *mdev)
+{
+}
+
+static inline int mlx5_fpga_device_start(struct mlx5_core_dev *mdev)
+{
+	return 0;
+}
+
+static inline void mlx5_fpga_device_stop(struct mlx5_core_dev *mdev)
+{
+}
+
+static inline void mlx5_fpga_event(struct mlx5_core_dev *mdev, u8 event,
+				   void *data)
+{
+}
+
+#endif
+
+#endif /* __MLX5_FPGA_CORE_H__ */
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/fpga/ipsec.c b/drivers/net/ethernet/mellanox/mlx5/core/fpga/ipsec.c
new file mode 100644
index 000000000..715ccafc9
--- /dev/null
+++ b/drivers/net/ethernet/mellanox/mlx5/core/fpga/ipsec.c
@@ -0,0 +1,1533 @@
+/*
+ * Copyright (c) 2017 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/rhashtable.h>
+#include <linux/mlx5/driver.h>
+#include <linux/mlx5/fs_helpers.h>
+#include <linux/mlx5/fs.h>
+#include <linux/rbtree.h>
+
+#include "mlx5_core.h"
+#include "fs_cmd.h"
+#include "fpga/ipsec.h"
+#include "fpga/sdk.h"
+#include "fpga/core.h"
+
+enum mlx5_fpga_ipsec_cmd_status {
+	MLX5_FPGA_IPSEC_CMD_PENDING,
+	MLX5_FPGA_IPSEC_CMD_SEND_FAIL,
+	MLX5_FPGA_IPSEC_CMD_COMPLETE,
+};
+
+struct mlx5_fpga_ipsec_cmd_context {
+	struct mlx5_fpga_dma_buf buf;
+	enum mlx5_fpga_ipsec_cmd_status status;
+	struct mlx5_ifc_fpga_ipsec_cmd_resp resp;
+	int status_code;
+	struct completion complete;
+	struct mlx5_fpga_device *dev;
+	struct list_head list; /* Item in pending_cmds */
+	u8 command[0];
+};
+
+struct mlx5_fpga_esp_xfrm;
+
+struct mlx5_fpga_ipsec_sa_ctx {
+	struct rhash_head		hash;
+	struct mlx5_ifc_fpga_ipsec_sa	hw_sa;
+	struct mlx5_core_dev		*dev;
+	struct mlx5_fpga_esp_xfrm	*fpga_xfrm;
+};
+
+struct mlx5_fpga_esp_xfrm {
+	unsigned int			num_rules;
+	struct mlx5_fpga_ipsec_sa_ctx	*sa_ctx;
+	struct mutex			lock; /* xfrm lock */
+	struct mlx5_accel_esp_xfrm	accel_xfrm;
+};
+
+struct mlx5_fpga_ipsec_rule {
+	struct rb_node			node;
+	struct fs_fte			*fte;
+	struct mlx5_fpga_ipsec_sa_ctx	*ctx;
+};
+
+static const struct rhashtable_params rhash_sa = {
+	/* Keep out "cmd" field from the key as it's
+	 * value is not constant during the lifetime
+	 * of the key object.
+	 */
+	.key_len = FIELD_SIZEOF(struct mlx5_fpga_ipsec_sa_ctx, hw_sa) -
+		   FIELD_SIZEOF(struct mlx5_ifc_fpga_ipsec_sa_v1, cmd),
+	.key_offset = offsetof(struct mlx5_fpga_ipsec_sa_ctx, hw_sa) +
+		      FIELD_SIZEOF(struct mlx5_ifc_fpga_ipsec_sa_v1, cmd),
+	.head_offset = offsetof(struct mlx5_fpga_ipsec_sa_ctx, hash),
+	.automatic_shrinking = true,
+	.min_size = 1,
+};
+
+struct mlx5_fpga_ipsec {
+	struct mlx5_fpga_device *fdev;
+	struct list_head pending_cmds;
+	spinlock_t pending_cmds_lock; /* Protects pending_cmds */
+	u32 caps[MLX5_ST_SZ_DW(ipsec_extended_cap)];
+	struct mlx5_fpga_conn *conn;
+
+	struct notifier_block	fs_notifier_ingress_bypass;
+	struct notifier_block	fs_notifier_egress;
+
+	/* Map hardware SA           -->  SA context
+	 *     (mlx5_fpga_ipsec_sa)       (mlx5_fpga_ipsec_sa_ctx)
+	 * We will use this hash to avoid SAs duplication in fpga which
+	 * aren't allowed
+	 */
+	struct rhashtable sa_hash;	/* hw_sa -> mlx5_fpga_ipsec_sa_ctx */
+	struct mutex sa_hash_lock;
+
+	/* Tree holding all rules for this fpga device
+	 * Key for searching a rule (mlx5_fpga_ipsec_rule) is (ft, id)
+	 */
+	struct rb_root rules_rb;
+	struct mutex rules_rb_lock; /* rules lock */
+};
+
+static bool mlx5_fpga_is_ipsec_device(struct mlx5_core_dev *mdev)
+{
+	if (!mdev->fpga || !MLX5_CAP_GEN(mdev, fpga))
+		return false;
+
+	if (MLX5_CAP_FPGA(mdev, ieee_vendor_id) !=
+	    MLX5_FPGA_CAP_SANDBOX_VENDOR_ID_MLNX)
+		return false;
+
+	if (MLX5_CAP_FPGA(mdev, sandbox_product_id) !=
+	    MLX5_FPGA_CAP_SANDBOX_PRODUCT_ID_IPSEC)
+		return false;
+
+	return true;
+}
+
+static void mlx5_fpga_ipsec_send_complete(struct mlx5_fpga_conn *conn,
+					  struct mlx5_fpga_device *fdev,
+					  struct mlx5_fpga_dma_buf *buf,
+					  u8 status)
+{
+	struct mlx5_fpga_ipsec_cmd_context *context;
+
+	if (status) {
+		context = container_of(buf, struct mlx5_fpga_ipsec_cmd_context,
+				       buf);
+		mlx5_fpga_warn(fdev, "IPSec command send failed with status %u\n",
+			       status);
+		context->status = MLX5_FPGA_IPSEC_CMD_SEND_FAIL;
+		complete(&context->complete);
+	}
+}
+
+static inline
+int syndrome_to_errno(enum mlx5_ifc_fpga_ipsec_response_syndrome syndrome)
+{
+	switch (syndrome) {
+	case MLX5_FPGA_IPSEC_RESPONSE_SUCCESS:
+		return 0;
+	case MLX5_FPGA_IPSEC_RESPONSE_SADB_ISSUE:
+		return -EEXIST;
+	case MLX5_FPGA_IPSEC_RESPONSE_ILLEGAL_REQUEST:
+		return -EINVAL;
+	case MLX5_FPGA_IPSEC_RESPONSE_WRITE_RESPONSE_ISSUE:
+		return -EIO;
+	}
+	return -EIO;
+}
+
+static void mlx5_fpga_ipsec_recv(void *cb_arg, struct mlx5_fpga_dma_buf *buf)
+{
+	struct mlx5_ifc_fpga_ipsec_cmd_resp *resp = buf->sg[0].data;
+	struct mlx5_fpga_ipsec_cmd_context *context;
+	enum mlx5_ifc_fpga_ipsec_response_syndrome syndrome;
+	struct mlx5_fpga_device *fdev = cb_arg;
+	unsigned long flags;
+
+	if (buf->sg[0].size < sizeof(*resp)) {
+		mlx5_fpga_warn(fdev, "Short receive from FPGA IPSec: %u < %zu bytes\n",
+			       buf->sg[0].size, sizeof(*resp));
+		return;
+	}
+
+	mlx5_fpga_dbg(fdev, "mlx5_ipsec recv_cb syndrome %08x\n",
+		      ntohl(resp->syndrome));
+
+	spin_lock_irqsave(&fdev->ipsec->pending_cmds_lock, flags);
+	context = list_first_entry_or_null(&fdev->ipsec->pending_cmds,
+					   struct mlx5_fpga_ipsec_cmd_context,
+					   list);
+	if (context)
+		list_del(&context->list);
+	spin_unlock_irqrestore(&fdev->ipsec->pending_cmds_lock, flags);
+
+	if (!context) {
+		mlx5_fpga_warn(fdev, "Received IPSec offload response without pending command request\n");
+		return;
+	}
+	mlx5_fpga_dbg(fdev, "Handling response for %p\n", context);
+
+	syndrome = ntohl(resp->syndrome);
+	context->status_code = syndrome_to_errno(syndrome);
+	context->status = MLX5_FPGA_IPSEC_CMD_COMPLETE;
+	memcpy(&context->resp, resp, sizeof(*resp));
+
+	if (context->status_code)
+		mlx5_fpga_warn(fdev, "IPSec command failed with syndrome %08x\n",
+			       syndrome);
+
+	complete(&context->complete);
+}
+
+static void *mlx5_fpga_ipsec_cmd_exec(struct mlx5_core_dev *mdev,
+				      const void *cmd, int cmd_size)
+{
+	struct mlx5_fpga_ipsec_cmd_context *context;
+	struct mlx5_fpga_device *fdev = mdev->fpga;
+	unsigned long flags;
+	int res;
+
+	if (!fdev || !fdev->ipsec)
+		return ERR_PTR(-EOPNOTSUPP);
+
+	if (cmd_size & 3)
+		return ERR_PTR(-EINVAL);
+
+	context = kzalloc(sizeof(*context) + cmd_size, GFP_ATOMIC);
+	if (!context)
+		return ERR_PTR(-ENOMEM);
+
+	context->status = MLX5_FPGA_IPSEC_CMD_PENDING;
+	context->dev = fdev;
+	context->buf.complete = mlx5_fpga_ipsec_send_complete;
+	init_completion(&context->complete);
+	memcpy(&context->command, cmd, cmd_size);
+	context->buf.sg[0].size = cmd_size;
+	context->buf.sg[0].data = &context->command;
+
+	spin_lock_irqsave(&fdev->ipsec->pending_cmds_lock, flags);
+	res = mlx5_fpga_sbu_conn_sendmsg(fdev->ipsec->conn, &context->buf);
+	if (!res)
+		list_add_tail(&context->list, &fdev->ipsec->pending_cmds);
+	spin_unlock_irqrestore(&fdev->ipsec->pending_cmds_lock, flags);
+
+	if (res) {
+		mlx5_fpga_warn(fdev, "Failed to send IPSec command: %d\n", res);
+		kfree(context);
+		return ERR_PTR(res);
+	}
+
+	/* Context should be freed by the caller after completion. */
+	return context;
+}
+
+static int mlx5_fpga_ipsec_cmd_wait(void *ctx)
+{
+	struct mlx5_fpga_ipsec_cmd_context *context = ctx;
+	unsigned long timeout =
+		msecs_to_jiffies(MLX5_FPGA_CMD_TIMEOUT_MSEC);
+	int res;
+
+	res = wait_for_completion_timeout(&context->complete, timeout);
+	if (!res) {
+		mlx5_fpga_warn(context->dev, "Failure waiting for IPSec command response\n");
+		return -ETIMEDOUT;
+	}
+
+	if (context->status == MLX5_FPGA_IPSEC_CMD_COMPLETE)
+		res = context->status_code;
+	else
+		res = -EIO;
+
+	return res;
+}
+
+static inline bool is_v2_sadb_supported(struct mlx5_fpga_ipsec *fipsec)
+{
+	if (MLX5_GET(ipsec_extended_cap, fipsec->caps, v2_command))
+		return true;
+	return false;
+}
+
+static int mlx5_fpga_ipsec_update_hw_sa(struct mlx5_fpga_device *fdev,
+					struct mlx5_ifc_fpga_ipsec_sa *hw_sa,
+					int opcode)
+{
+	struct mlx5_core_dev *dev = fdev->mdev;
+	struct mlx5_ifc_fpga_ipsec_sa *sa;
+	struct mlx5_fpga_ipsec_cmd_context *cmd_context;
+	size_t sa_cmd_size;
+	int err;
+
+	hw_sa->ipsec_sa_v1.cmd = htonl(opcode);
+	if (is_v2_sadb_supported(fdev->ipsec))
+		sa_cmd_size = sizeof(*hw_sa);
+	else
+		sa_cmd_size = sizeof(hw_sa->ipsec_sa_v1);
+
+	cmd_context = (struct mlx5_fpga_ipsec_cmd_context *)
+			mlx5_fpga_ipsec_cmd_exec(dev, hw_sa, sa_cmd_size);
+	if (IS_ERR(cmd_context))
+		return PTR_ERR(cmd_context);
+
+	err = mlx5_fpga_ipsec_cmd_wait(cmd_context);
+	if (err)
+		goto out;
+
+	sa = (struct mlx5_ifc_fpga_ipsec_sa *)&cmd_context->command;
+	if (sa->ipsec_sa_v1.sw_sa_handle != cmd_context->resp.sw_sa_handle) {
+		mlx5_fpga_err(fdev, "mismatch SA handle. cmd 0x%08x vs resp 0x%08x\n",
+			      ntohl(sa->ipsec_sa_v1.sw_sa_handle),
+			      ntohl(cmd_context->resp.sw_sa_handle));
+		err = -EIO;
+	}
+
+out:
+	kfree(cmd_context);
+	return err;
+}
+
+u32 mlx5_fpga_ipsec_device_caps(struct mlx5_core_dev *mdev)
+{
+	struct mlx5_fpga_device *fdev = mdev->fpga;
+	u32 ret = 0;
+
+	if (mlx5_fpga_is_ipsec_device(mdev)) {
+		ret |= MLX5_ACCEL_IPSEC_CAP_DEVICE;
+		ret |= MLX5_ACCEL_IPSEC_CAP_REQUIRED_METADATA;
+	} else {
+		return ret;
+	}
+
+	if (!fdev->ipsec)
+		return ret;
+
+	if (MLX5_GET(ipsec_extended_cap, fdev->ipsec->caps, esp))
+		ret |= MLX5_ACCEL_IPSEC_CAP_ESP;
+
+	if (MLX5_GET(ipsec_extended_cap, fdev->ipsec->caps, ipv6))
+		ret |= MLX5_ACCEL_IPSEC_CAP_IPV6;
+
+	if (MLX5_GET(ipsec_extended_cap, fdev->ipsec->caps, lso))
+		ret |= MLX5_ACCEL_IPSEC_CAP_LSO;
+
+	if (MLX5_GET(ipsec_extended_cap, fdev->ipsec->caps, rx_no_trailer))
+		ret |= MLX5_ACCEL_IPSEC_CAP_RX_NO_TRAILER;
+
+	if (MLX5_GET(ipsec_extended_cap, fdev->ipsec->caps, esn)) {
+		ret |= MLX5_ACCEL_IPSEC_CAP_ESN;
+		ret |= MLX5_ACCEL_IPSEC_CAP_TX_IV_IS_ESN;
+	}
+
+	return ret;
+}
+
+unsigned int mlx5_fpga_ipsec_counters_count(struct mlx5_core_dev *mdev)
+{
+	struct mlx5_fpga_device *fdev = mdev->fpga;
+
+	if (!fdev || !fdev->ipsec)
+		return 0;
+
+	return MLX5_GET(ipsec_extended_cap, fdev->ipsec->caps,
+			number_of_ipsec_counters);
+}
+
+int mlx5_fpga_ipsec_counters_read(struct mlx5_core_dev *mdev, u64 *counters,
+				  unsigned int counters_count)
+{
+	struct mlx5_fpga_device *fdev = mdev->fpga;
+	unsigned int i;
+	__be32 *data;
+	u32 count;
+	u64 addr;
+	int ret;
+
+	if (!fdev || !fdev->ipsec)
+		return 0;
+
+	addr = (u64)MLX5_GET(ipsec_extended_cap, fdev->ipsec->caps,
+			     ipsec_counters_addr_low) +
+	       ((u64)MLX5_GET(ipsec_extended_cap, fdev->ipsec->caps,
+			     ipsec_counters_addr_high) << 32);
+
+	count = mlx5_fpga_ipsec_counters_count(mdev);
+
+	data = kzalloc(array3_size(sizeof(*data), count, 2), GFP_KERNEL);
+	if (!data) {
+		ret = -ENOMEM;
+		goto out;
+	}
+
+	ret = mlx5_fpga_mem_read(fdev, count * sizeof(u64), addr, data,
+				 MLX5_FPGA_ACCESS_TYPE_DONTCARE);
+	if (ret < 0) {
+		mlx5_fpga_err(fdev, "Failed to read IPSec counters from HW: %d\n",
+			      ret);
+		goto out;
+	}
+	ret = 0;
+
+	if (count > counters_count)
+		count = counters_count;
+
+	/* Each counter is low word, then high. But each word is big-endian */
+	for (i = 0; i < count; i++)
+		counters[i] = (u64)ntohl(data[i * 2]) |
+			      ((u64)ntohl(data[i * 2 + 1]) << 32);
+
+out:
+	kfree(data);
+	return ret;
+}
+
+static int mlx5_fpga_ipsec_set_caps(struct mlx5_core_dev *mdev, u32 flags)
+{
+	struct mlx5_fpga_ipsec_cmd_context *context;
+	struct mlx5_ifc_fpga_ipsec_cmd_cap cmd = {0};
+	int err;
+
+	cmd.cmd = htonl(MLX5_FPGA_IPSEC_CMD_OP_SET_CAP);
+	cmd.flags = htonl(flags);
+	context = mlx5_fpga_ipsec_cmd_exec(mdev, &cmd, sizeof(cmd));
+	if (IS_ERR(context))
+		return PTR_ERR(context);
+
+	err = mlx5_fpga_ipsec_cmd_wait(context);
+	if (err)
+		goto out;
+
+	if ((context->resp.flags & cmd.flags) != cmd.flags) {
+		mlx5_fpga_err(context->dev, "Failed to set capabilities. cmd 0x%08x vs resp 0x%08x\n",
+			      cmd.flags,
+			      context->resp.flags);
+		err = -EIO;
+	}
+
+out:
+	kfree(context);
+	return err;
+}
+
+static int mlx5_fpga_ipsec_enable_supported_caps(struct mlx5_core_dev *mdev)
+{
+	u32 dev_caps = mlx5_fpga_ipsec_device_caps(mdev);
+	u32 flags = 0;
+
+	if (dev_caps & MLX5_ACCEL_IPSEC_CAP_RX_NO_TRAILER)
+		flags |= MLX5_FPGA_IPSEC_CAP_NO_TRAILER;
+
+	return mlx5_fpga_ipsec_set_caps(mdev, flags);
+}
+
+static void
+mlx5_fpga_ipsec_build_hw_xfrm(struct mlx5_core_dev *mdev,
+			      const struct mlx5_accel_esp_xfrm_attrs *xfrm_attrs,
+			      struct mlx5_ifc_fpga_ipsec_sa *hw_sa)
+{
+	const struct aes_gcm_keymat *aes_gcm = &xfrm_attrs->keymat.aes_gcm;
+
+	/* key */
+	memcpy(&hw_sa->ipsec_sa_v1.key_enc, aes_gcm->aes_key,
+	       aes_gcm->key_len / 8);
+	/* Duplicate 128 bit key twice according to HW layout */
+	if (aes_gcm->key_len == 128)
+		memcpy(&hw_sa->ipsec_sa_v1.key_enc[16],
+		       aes_gcm->aes_key, aes_gcm->key_len / 8);
+
+	/* salt and seq_iv */
+	memcpy(&hw_sa->ipsec_sa_v1.gcm.salt_iv, &aes_gcm->seq_iv,
+	       sizeof(aes_gcm->seq_iv));
+	memcpy(&hw_sa->ipsec_sa_v1.gcm.salt, &aes_gcm->salt,
+	       sizeof(aes_gcm->salt));
+
+	/* esn */
+	if (xfrm_attrs->flags & MLX5_ACCEL_ESP_FLAGS_ESN_TRIGGERED) {
+		hw_sa->ipsec_sa_v1.flags |= MLX5_FPGA_IPSEC_SA_ESN_EN;
+		hw_sa->ipsec_sa_v1.flags |=
+				(xfrm_attrs->flags &
+				 MLX5_ACCEL_ESP_FLAGS_ESN_STATE_OVERLAP) ?
+					MLX5_FPGA_IPSEC_SA_ESN_OVERLAP : 0;
+		hw_sa->esn = htonl(xfrm_attrs->esn);
+	} else {
+		hw_sa->ipsec_sa_v1.flags &= ~MLX5_FPGA_IPSEC_SA_ESN_EN;
+		hw_sa->ipsec_sa_v1.flags &=
+				~(xfrm_attrs->flags &
+				  MLX5_ACCEL_ESP_FLAGS_ESN_STATE_OVERLAP) ?
+					MLX5_FPGA_IPSEC_SA_ESN_OVERLAP : 0;
+		hw_sa->esn = 0;
+	}
+
+	/* rx handle */
+	hw_sa->ipsec_sa_v1.sw_sa_handle = htonl(xfrm_attrs->sa_handle);
+
+	/* enc mode */
+	switch (aes_gcm->key_len) {
+	case 128:
+		hw_sa->ipsec_sa_v1.enc_mode =
+			MLX5_FPGA_IPSEC_SA_ENC_MODE_AES_GCM_128_AUTH_128;
+		break;
+	case 256:
+		hw_sa->ipsec_sa_v1.enc_mode =
+			MLX5_FPGA_IPSEC_SA_ENC_MODE_AES_GCM_256_AUTH_128;
+		break;
+	}
+
+	/* flags */
+	hw_sa->ipsec_sa_v1.flags |= MLX5_FPGA_IPSEC_SA_SA_VALID |
+			MLX5_FPGA_IPSEC_SA_SPI_EN |
+			MLX5_FPGA_IPSEC_SA_IP_ESP;
+
+	if (xfrm_attrs->action & MLX5_ACCEL_ESP_ACTION_ENCRYPT)
+		hw_sa->ipsec_sa_v1.flags |= MLX5_FPGA_IPSEC_SA_DIR_SX;
+	else
+		hw_sa->ipsec_sa_v1.flags &= ~MLX5_FPGA_IPSEC_SA_DIR_SX;
+}
+
+static void
+mlx5_fpga_ipsec_build_hw_sa(struct mlx5_core_dev *mdev,
+			    struct mlx5_accel_esp_xfrm_attrs *xfrm_attrs,
+			    const __be32 saddr[4],
+			    const __be32 daddr[4],
+			    const __be32 spi, bool is_ipv6,
+			    struct mlx5_ifc_fpga_ipsec_sa *hw_sa)
+{
+	mlx5_fpga_ipsec_build_hw_xfrm(mdev, xfrm_attrs, hw_sa);
+
+	/* IPs */
+	memcpy(hw_sa->ipsec_sa_v1.sip, saddr, sizeof(hw_sa->ipsec_sa_v1.sip));
+	memcpy(hw_sa->ipsec_sa_v1.dip, daddr, sizeof(hw_sa->ipsec_sa_v1.dip));
+
+	/* SPI */
+	hw_sa->ipsec_sa_v1.spi = spi;
+
+	/* flags */
+	if (is_ipv6)
+		hw_sa->ipsec_sa_v1.flags |= MLX5_FPGA_IPSEC_SA_IPV6;
+}
+
+static bool is_full_mask(const void *p, size_t len)
+{
+	WARN_ON(len % 4);
+
+	return !memchr_inv(p, 0xff, len);
+}
+
+static bool validate_fpga_full_mask(struct mlx5_core_dev *dev,
+				    const u32 *match_c,
+				    const u32 *match_v)
+{
+	const void *misc_params_c = MLX5_ADDR_OF(fte_match_param,
+						 match_c,
+						 misc_parameters);
+	const void *headers_c = MLX5_ADDR_OF(fte_match_param,
+					     match_c,
+					     outer_headers);
+	const void *headers_v = MLX5_ADDR_OF(fte_match_param,
+					     match_v,
+					     outer_headers);
+
+	if (mlx5_fs_is_outer_ipv4_flow(dev, headers_c, headers_v)) {
+		const void *s_ipv4_c = MLX5_ADDR_OF(fte_match_set_lyr_2_4,
+						    headers_c,
+						    src_ipv4_src_ipv6.ipv4_layout.ipv4);
+		const void *d_ipv4_c = MLX5_ADDR_OF(fte_match_set_lyr_2_4,
+						    headers_c,
+						    dst_ipv4_dst_ipv6.ipv4_layout.ipv4);
+
+		if (!is_full_mask(s_ipv4_c, MLX5_FLD_SZ_BYTES(ipv4_layout,
+							      ipv4)) ||
+		    !is_full_mask(d_ipv4_c, MLX5_FLD_SZ_BYTES(ipv4_layout,
+							      ipv4)))
+			return false;
+	} else {
+		const void *s_ipv6_c = MLX5_ADDR_OF(fte_match_set_lyr_2_4,
+						    headers_c,
+						    src_ipv4_src_ipv6.ipv6_layout.ipv6);
+		const void *d_ipv6_c = MLX5_ADDR_OF(fte_match_set_lyr_2_4,
+						    headers_c,
+						    dst_ipv4_dst_ipv6.ipv6_layout.ipv6);
+
+		if (!is_full_mask(s_ipv6_c, MLX5_FLD_SZ_BYTES(ipv6_layout,
+							      ipv6)) ||
+		    !is_full_mask(d_ipv6_c, MLX5_FLD_SZ_BYTES(ipv6_layout,
+							      ipv6)))
+			return false;
+	}
+
+	if (!is_full_mask(MLX5_ADDR_OF(fte_match_set_misc, misc_params_c,
+				       outer_esp_spi),
+			  MLX5_FLD_SZ_BYTES(fte_match_set_misc, outer_esp_spi)))
+		return false;
+
+	return true;
+}
+
+static bool mlx5_is_fpga_ipsec_rule(struct mlx5_core_dev *dev,
+				    u8 match_criteria_enable,
+				    const u32 *match_c,
+				    const u32 *match_v)
+{
+	u32 ipsec_dev_caps = mlx5_accel_ipsec_device_caps(dev);
+	bool ipv6_flow;
+
+	ipv6_flow = mlx5_fs_is_outer_ipv6_flow(dev, match_c, match_v);
+
+	if (!(match_criteria_enable & MLX5_MATCH_OUTER_HEADERS) ||
+	    mlx5_fs_is_outer_udp_flow(match_c, match_v) ||
+	    mlx5_fs_is_outer_tcp_flow(match_c, match_v) ||
+	    mlx5_fs_is_vxlan_flow(match_c) ||
+	    !(mlx5_fs_is_outer_ipv4_flow(dev, match_c, match_v) ||
+	      ipv6_flow))
+		return false;
+
+	if (!(ipsec_dev_caps & MLX5_ACCEL_IPSEC_CAP_DEVICE))
+		return false;
+
+	if (!(ipsec_dev_caps & MLX5_ACCEL_IPSEC_CAP_ESP) &&
+	    mlx5_fs_is_outer_ipsec_flow(match_c))
+		return false;
+
+	if (!(ipsec_dev_caps & MLX5_ACCEL_IPSEC_CAP_IPV6) &&
+	    ipv6_flow)
+		return false;
+
+	if (!validate_fpga_full_mask(dev, match_c, match_v))
+		return false;
+
+	return true;
+}
+
+static bool mlx5_is_fpga_egress_ipsec_rule(struct mlx5_core_dev *dev,
+					   u8 match_criteria_enable,
+					   const u32 *match_c,
+					   const u32 *match_v,
+					   struct mlx5_flow_act *flow_act)
+{
+	const void *outer_c = MLX5_ADDR_OF(fte_match_param, match_c,
+					   outer_headers);
+	bool is_dmac = MLX5_GET(fte_match_set_lyr_2_4, outer_c, dmac_47_16) ||
+			MLX5_GET(fte_match_set_lyr_2_4, outer_c, dmac_15_0);
+	bool is_smac = MLX5_GET(fte_match_set_lyr_2_4, outer_c, smac_47_16) ||
+			MLX5_GET(fte_match_set_lyr_2_4, outer_c, smac_15_0);
+	int ret;
+
+	ret = mlx5_is_fpga_ipsec_rule(dev, match_criteria_enable, match_c,
+				      match_v);
+	if (!ret)
+		return ret;
+
+	if (is_dmac || is_smac ||
+	    (match_criteria_enable &
+	     ~(MLX5_MATCH_OUTER_HEADERS | MLX5_MATCH_MISC_PARAMETERS)) ||
+	    (flow_act->action & ~(MLX5_FLOW_CONTEXT_ACTION_ENCRYPT | MLX5_FLOW_CONTEXT_ACTION_ALLOW)) ||
+	     flow_act->has_flow_tag)
+		return false;
+
+	return true;
+}
+
+void *mlx5_fpga_ipsec_create_sa_ctx(struct mlx5_core_dev *mdev,
+				    struct mlx5_accel_esp_xfrm *accel_xfrm,
+				    const __be32 saddr[4],
+				    const __be32 daddr[4],
+				    const __be32 spi, bool is_ipv6)
+{
+	struct mlx5_fpga_ipsec_sa_ctx *sa_ctx;
+	struct mlx5_fpga_esp_xfrm *fpga_xfrm =
+			container_of(accel_xfrm, typeof(*fpga_xfrm),
+				     accel_xfrm);
+	struct mlx5_fpga_device *fdev = mdev->fpga;
+	struct mlx5_fpga_ipsec *fipsec = fdev->ipsec;
+	int opcode, err;
+	void *context;
+
+	/* alloc SA */
+	sa_ctx = kzalloc(sizeof(*sa_ctx), GFP_KERNEL);
+	if (!sa_ctx)
+		return ERR_PTR(-ENOMEM);
+
+	sa_ctx->dev = mdev;
+
+	/* build candidate SA */
+	mlx5_fpga_ipsec_build_hw_sa(mdev, &accel_xfrm->attrs,
+				    saddr, daddr, spi, is_ipv6,
+				    &sa_ctx->hw_sa);
+
+	mutex_lock(&fpga_xfrm->lock);
+
+	if (fpga_xfrm->sa_ctx) {        /* multiple rules for same accel_xfrm */
+		/* all rules must be with same IPs and SPI */
+		if (memcmp(&sa_ctx->hw_sa, &fpga_xfrm->sa_ctx->hw_sa,
+			   sizeof(sa_ctx->hw_sa))) {
+			context = ERR_PTR(-EINVAL);
+			goto exists;
+		}
+
+		++fpga_xfrm->num_rules;
+		context = fpga_xfrm->sa_ctx;
+		goto exists;
+	}
+
+	/* This is unbounded fpga_xfrm, try to add to hash */
+	mutex_lock(&fipsec->sa_hash_lock);
+
+	err = rhashtable_lookup_insert_fast(&fipsec->sa_hash, &sa_ctx->hash,
+					    rhash_sa);
+	if (err) {
+		/* Can't bound different accel_xfrm to already existing sa_ctx.
+		 * This is because we can't support multiple ketmats for
+		 * same IPs and SPI
+		 */
+		context = ERR_PTR(-EEXIST);
+		goto unlock_hash;
+	}
+
+	/* Bound accel_xfrm to sa_ctx */
+	opcode = is_v2_sadb_supported(fdev->ipsec) ?
+			MLX5_FPGA_IPSEC_CMD_OP_ADD_SA_V2 :
+			MLX5_FPGA_IPSEC_CMD_OP_ADD_SA;
+	err = mlx5_fpga_ipsec_update_hw_sa(fdev, &sa_ctx->hw_sa, opcode);
+	sa_ctx->hw_sa.ipsec_sa_v1.cmd = 0;
+	if (err) {
+		context = ERR_PTR(err);
+		goto delete_hash;
+	}
+
+	mutex_unlock(&fipsec->sa_hash_lock);
+
+	++fpga_xfrm->num_rules;
+	fpga_xfrm->sa_ctx = sa_ctx;
+	sa_ctx->fpga_xfrm = fpga_xfrm;
+
+	mutex_unlock(&fpga_xfrm->lock);
+
+	return sa_ctx;
+
+delete_hash:
+	WARN_ON(rhashtable_remove_fast(&fipsec->sa_hash, &sa_ctx->hash,
+				       rhash_sa));
+unlock_hash:
+	mutex_unlock(&fipsec->sa_hash_lock);
+
+exists:
+	mutex_unlock(&fpga_xfrm->lock);
+	kfree(sa_ctx);
+	return context;
+}
+
+static void *
+mlx5_fpga_ipsec_fs_create_sa_ctx(struct mlx5_core_dev *mdev,
+				 struct fs_fte *fte,
+				 bool is_egress)
+{
+	struct mlx5_accel_esp_xfrm *accel_xfrm;
+	__be32 saddr[4], daddr[4], spi;
+	struct mlx5_flow_group *fg;
+	bool is_ipv6 = false;
+
+	fs_get_obj(fg, fte->node.parent);
+	/* validate */
+	if (is_egress &&
+	    !mlx5_is_fpga_egress_ipsec_rule(mdev,
+					    fg->mask.match_criteria_enable,
+					    fg->mask.match_criteria,
+					    fte->val,
+					    &fte->action))
+		return ERR_PTR(-EINVAL);
+	else if (!mlx5_is_fpga_ipsec_rule(mdev,
+					  fg->mask.match_criteria_enable,
+					  fg->mask.match_criteria,
+					  fte->val))
+		return ERR_PTR(-EINVAL);
+
+	/* get xfrm context */
+	accel_xfrm =
+		(struct mlx5_accel_esp_xfrm *)fte->action.esp_id;
+
+	/* IPs */
+	if (mlx5_fs_is_outer_ipv4_flow(mdev, fg->mask.match_criteria,
+				       fte->val)) {
+		memcpy(&saddr[3],
+		       MLX5_ADDR_OF(fte_match_set_lyr_2_4,
+				    fte->val,
+				    src_ipv4_src_ipv6.ipv4_layout.ipv4),
+				    sizeof(saddr[3]));
+		memcpy(&daddr[3],
+		       MLX5_ADDR_OF(fte_match_set_lyr_2_4,
+				    fte->val,
+				    dst_ipv4_dst_ipv6.ipv4_layout.ipv4),
+				    sizeof(daddr[3]));
+	} else {
+		memcpy(saddr,
+		       MLX5_ADDR_OF(fte_match_param,
+				    fte->val,
+				    outer_headers.src_ipv4_src_ipv6.ipv6_layout.ipv6),
+				    sizeof(saddr));
+		memcpy(daddr,
+		       MLX5_ADDR_OF(fte_match_param,
+				    fte->val,
+				    outer_headers.dst_ipv4_dst_ipv6.ipv6_layout.ipv6),
+				    sizeof(daddr));
+		is_ipv6 = true;
+	}
+
+	/* SPI */
+	spi = MLX5_GET_BE(typeof(spi),
+			  fte_match_param, fte->val,
+			  misc_parameters.outer_esp_spi);
+
+	/* create */
+	return mlx5_fpga_ipsec_create_sa_ctx(mdev, accel_xfrm,
+					     saddr, daddr,
+					     spi, is_ipv6);
+}
+
+static void
+mlx5_fpga_ipsec_release_sa_ctx(struct mlx5_fpga_ipsec_sa_ctx *sa_ctx)
+{
+	struct mlx5_fpga_device *fdev = sa_ctx->dev->fpga;
+	struct mlx5_fpga_ipsec *fipsec = fdev->ipsec;
+	int opcode = is_v2_sadb_supported(fdev->ipsec) ?
+			MLX5_FPGA_IPSEC_CMD_OP_DEL_SA_V2 :
+			MLX5_FPGA_IPSEC_CMD_OP_DEL_SA;
+	int err;
+
+	err = mlx5_fpga_ipsec_update_hw_sa(fdev, &sa_ctx->hw_sa, opcode);
+	sa_ctx->hw_sa.ipsec_sa_v1.cmd = 0;
+	if (err) {
+		WARN_ON(err);
+		return;
+	}
+
+	mutex_lock(&fipsec->sa_hash_lock);
+	WARN_ON(rhashtable_remove_fast(&fipsec->sa_hash, &sa_ctx->hash,
+				       rhash_sa));
+	mutex_unlock(&fipsec->sa_hash_lock);
+}
+
+void mlx5_fpga_ipsec_delete_sa_ctx(void *context)
+{
+	struct mlx5_fpga_esp_xfrm *fpga_xfrm =
+			((struct mlx5_fpga_ipsec_sa_ctx *)context)->fpga_xfrm;
+
+	mutex_lock(&fpga_xfrm->lock);
+	if (!--fpga_xfrm->num_rules) {
+		mlx5_fpga_ipsec_release_sa_ctx(fpga_xfrm->sa_ctx);
+		kfree(fpga_xfrm->sa_ctx);
+		fpga_xfrm->sa_ctx = NULL;
+	}
+	mutex_unlock(&fpga_xfrm->lock);
+}
+
+static inline struct mlx5_fpga_ipsec_rule *
+_rule_search(struct rb_root *root, struct fs_fte *fte)
+{
+	struct rb_node *node = root->rb_node;
+
+	while (node) {
+		struct mlx5_fpga_ipsec_rule *rule =
+				container_of(node, struct mlx5_fpga_ipsec_rule,
+					     node);
+
+		if (rule->fte < fte)
+			node = node->rb_left;
+		else if (rule->fte > fte)
+			node = node->rb_right;
+		else
+			return rule;
+	}
+	return NULL;
+}
+
+static struct mlx5_fpga_ipsec_rule *
+rule_search(struct mlx5_fpga_ipsec *ipsec_dev, struct fs_fte *fte)
+{
+	struct mlx5_fpga_ipsec_rule *rule;
+
+	mutex_lock(&ipsec_dev->rules_rb_lock);
+	rule = _rule_search(&ipsec_dev->rules_rb, fte);
+	mutex_unlock(&ipsec_dev->rules_rb_lock);
+
+	return rule;
+}
+
+static inline int _rule_insert(struct rb_root *root,
+			       struct mlx5_fpga_ipsec_rule *rule)
+{
+	struct rb_node **new = &root->rb_node, *parent = NULL;
+
+	/* Figure out where to put new node */
+	while (*new) {
+		struct mlx5_fpga_ipsec_rule *this =
+				container_of(*new, struct mlx5_fpga_ipsec_rule,
+					     node);
+
+		parent = *new;
+		if (rule->fte < this->fte)
+			new = &((*new)->rb_left);
+		else if (rule->fte > this->fte)
+			new = &((*new)->rb_right);
+		else
+			return -EEXIST;
+	}
+
+	/* Add new node and rebalance tree. */
+	rb_link_node(&rule->node, parent, new);
+	rb_insert_color(&rule->node, root);
+
+	return 0;
+}
+
+static int rule_insert(struct mlx5_fpga_ipsec *ipsec_dev,
+		       struct mlx5_fpga_ipsec_rule *rule)
+{
+	int ret;
+
+	mutex_lock(&ipsec_dev->rules_rb_lock);
+	ret = _rule_insert(&ipsec_dev->rules_rb, rule);
+	mutex_unlock(&ipsec_dev->rules_rb_lock);
+
+	return ret;
+}
+
+static inline void _rule_delete(struct mlx5_fpga_ipsec *ipsec_dev,
+				struct mlx5_fpga_ipsec_rule *rule)
+{
+	struct rb_root *root = &ipsec_dev->rules_rb;
+
+	mutex_lock(&ipsec_dev->rules_rb_lock);
+	rb_erase(&rule->node, root);
+	mutex_unlock(&ipsec_dev->rules_rb_lock);
+}
+
+static void rule_delete(struct mlx5_fpga_ipsec *ipsec_dev,
+			struct mlx5_fpga_ipsec_rule *rule)
+{
+	_rule_delete(ipsec_dev, rule);
+	kfree(rule);
+}
+
+struct mailbox_mod {
+	uintptr_t			saved_esp_id;
+	u32				saved_action;
+	u32				saved_outer_esp_spi_value;
+};
+
+static void restore_spec_mailbox(struct fs_fte *fte,
+				 struct mailbox_mod *mbox_mod)
+{
+	char *misc_params_v = MLX5_ADDR_OF(fte_match_param,
+					   fte->val,
+					   misc_parameters);
+
+	MLX5_SET(fte_match_set_misc, misc_params_v, outer_esp_spi,
+		 mbox_mod->saved_outer_esp_spi_value);
+	fte->action.action |= mbox_mod->saved_action;
+	fte->action.esp_id = (uintptr_t)mbox_mod->saved_esp_id;
+}
+
+static void modify_spec_mailbox(struct mlx5_core_dev *mdev,
+				struct fs_fte *fte,
+				struct mailbox_mod *mbox_mod)
+{
+	char *misc_params_v = MLX5_ADDR_OF(fte_match_param,
+					   fte->val,
+					   misc_parameters);
+
+	mbox_mod->saved_esp_id = fte->action.esp_id;
+	mbox_mod->saved_action = fte->action.action &
+			(MLX5_FLOW_CONTEXT_ACTION_ENCRYPT |
+			 MLX5_FLOW_CONTEXT_ACTION_DECRYPT);
+	mbox_mod->saved_outer_esp_spi_value =
+			MLX5_GET(fte_match_set_misc, misc_params_v,
+				 outer_esp_spi);
+
+	fte->action.esp_id = 0;
+	fte->action.action &= ~(MLX5_FLOW_CONTEXT_ACTION_ENCRYPT |
+				MLX5_FLOW_CONTEXT_ACTION_DECRYPT);
+	if (!MLX5_CAP_FLOWTABLE(mdev,
+				flow_table_properties_nic_receive.ft_field_support.outer_esp_spi))
+		MLX5_SET(fte_match_set_misc, misc_params_v, outer_esp_spi, 0);
+}
+
+static enum fs_flow_table_type egress_to_fs_ft(bool egress)
+{
+	return egress ? FS_FT_NIC_TX : FS_FT_NIC_RX;
+}
+
+static int fpga_ipsec_fs_create_flow_group(struct mlx5_core_dev *dev,
+					   struct mlx5_flow_table *ft,
+					   u32 *in,
+					   unsigned int *group_id,
+					   bool is_egress)
+{
+	int (*create_flow_group)(struct mlx5_core_dev *dev,
+				 struct mlx5_flow_table *ft, u32 *in,
+				 unsigned int *group_id) =
+		mlx5_fs_cmd_get_default(egress_to_fs_ft(is_egress))->create_flow_group;
+	char *misc_params_c = MLX5_ADDR_OF(create_flow_group_in, in,
+					   match_criteria.misc_parameters);
+	u32 saved_outer_esp_spi_mask;
+	u8 match_criteria_enable;
+	int ret;
+
+	if (MLX5_CAP_FLOWTABLE(dev,
+			       flow_table_properties_nic_receive.ft_field_support.outer_esp_spi))
+		return create_flow_group(dev, ft, in, group_id);
+
+	match_criteria_enable =
+		MLX5_GET(create_flow_group_in, in, match_criteria_enable);
+	saved_outer_esp_spi_mask =
+		MLX5_GET(fte_match_set_misc, misc_params_c, outer_esp_spi);
+	if (!match_criteria_enable || !saved_outer_esp_spi_mask)
+		return create_flow_group(dev, ft, in, group_id);
+
+	MLX5_SET(fte_match_set_misc, misc_params_c, outer_esp_spi, 0);
+
+	if (!(*misc_params_c) &&
+	    !memcmp(misc_params_c, misc_params_c + 1, MLX5_ST_SZ_BYTES(fte_match_set_misc) - 1))
+		MLX5_SET(create_flow_group_in, in, match_criteria_enable,
+			 match_criteria_enable & ~MLX5_MATCH_MISC_PARAMETERS);
+
+	ret = create_flow_group(dev, ft, in, group_id);
+
+	MLX5_SET(fte_match_set_misc, misc_params_c, outer_esp_spi, saved_outer_esp_spi_mask);
+	MLX5_SET(create_flow_group_in, in, match_criteria_enable, match_criteria_enable);
+
+	return ret;
+}
+
+static int fpga_ipsec_fs_create_fte(struct mlx5_core_dev *dev,
+				    struct mlx5_flow_table *ft,
+				    struct mlx5_flow_group *fg,
+				    struct fs_fte *fte,
+				    bool is_egress)
+{
+	int (*create_fte)(struct mlx5_core_dev *dev,
+			  struct mlx5_flow_table *ft,
+			  struct mlx5_flow_group *fg,
+			  struct fs_fte *fte) =
+		mlx5_fs_cmd_get_default(egress_to_fs_ft(is_egress))->create_fte;
+	struct mlx5_fpga_device *fdev = dev->fpga;
+	struct mlx5_fpga_ipsec *fipsec = fdev->ipsec;
+	struct mlx5_fpga_ipsec_rule *rule;
+	bool is_esp = fte->action.esp_id;
+	struct mailbox_mod mbox_mod;
+	int ret;
+
+	if (!is_esp ||
+	    !(fte->action.action &
+	      (MLX5_FLOW_CONTEXT_ACTION_ENCRYPT |
+	       MLX5_FLOW_CONTEXT_ACTION_DECRYPT)))
+		return create_fte(dev, ft, fg, fte);
+
+	rule = kzalloc(sizeof(*rule), GFP_KERNEL);
+	if (!rule)
+		return -ENOMEM;
+
+	rule->ctx = mlx5_fpga_ipsec_fs_create_sa_ctx(dev, fte, is_egress);
+	if (IS_ERR(rule->ctx)) {
+		int err = PTR_ERR(rule->ctx);
+		kfree(rule);
+		return err;
+	}
+
+	rule->fte = fte;
+	WARN_ON(rule_insert(fipsec, rule));
+
+	modify_spec_mailbox(dev, fte, &mbox_mod);
+	ret = create_fte(dev, ft, fg, fte);
+	restore_spec_mailbox(fte, &mbox_mod);
+	if (ret) {
+		_rule_delete(fipsec, rule);
+		mlx5_fpga_ipsec_delete_sa_ctx(rule->ctx);
+		kfree(rule);
+	}
+
+	return ret;
+}
+
+static int fpga_ipsec_fs_update_fte(struct mlx5_core_dev *dev,
+				    struct mlx5_flow_table *ft,
+				    unsigned int group_id,
+				    int modify_mask,
+				    struct fs_fte *fte,
+				    bool is_egress)
+{
+	int (*update_fte)(struct mlx5_core_dev *dev,
+			  struct mlx5_flow_table *ft,
+			  unsigned int group_id,
+			  int modify_mask,
+			  struct fs_fte *fte) =
+		mlx5_fs_cmd_get_default(egress_to_fs_ft(is_egress))->update_fte;
+	bool is_esp = fte->action.esp_id;
+	struct mailbox_mod mbox_mod;
+	int ret;
+
+	if (!is_esp ||
+	    !(fte->action.action &
+	      (MLX5_FLOW_CONTEXT_ACTION_ENCRYPT |
+	       MLX5_FLOW_CONTEXT_ACTION_DECRYPT)))
+		return update_fte(dev, ft, group_id, modify_mask, fte);
+
+	modify_spec_mailbox(dev, fte, &mbox_mod);
+	ret = update_fte(dev, ft, group_id, modify_mask, fte);
+	restore_spec_mailbox(fte, &mbox_mod);
+
+	return ret;
+}
+
+static int fpga_ipsec_fs_delete_fte(struct mlx5_core_dev *dev,
+				    struct mlx5_flow_table *ft,
+				    struct fs_fte *fte,
+				    bool is_egress)
+{
+	int (*delete_fte)(struct mlx5_core_dev *dev,
+			  struct mlx5_flow_table *ft,
+			  struct fs_fte *fte) =
+		mlx5_fs_cmd_get_default(egress_to_fs_ft(is_egress))->delete_fte;
+	struct mlx5_fpga_device *fdev = dev->fpga;
+	struct mlx5_fpga_ipsec *fipsec = fdev->ipsec;
+	struct mlx5_fpga_ipsec_rule *rule;
+	bool is_esp = fte->action.esp_id;
+	struct mailbox_mod mbox_mod;
+	int ret;
+
+	if (!is_esp ||
+	    !(fte->action.action &
+	      (MLX5_FLOW_CONTEXT_ACTION_ENCRYPT |
+	       MLX5_FLOW_CONTEXT_ACTION_DECRYPT)))
+		return delete_fte(dev, ft, fte);
+
+	rule = rule_search(fipsec, fte);
+	if (!rule)
+		return -ENOENT;
+
+	mlx5_fpga_ipsec_delete_sa_ctx(rule->ctx);
+	rule_delete(fipsec, rule);
+
+	modify_spec_mailbox(dev, fte, &mbox_mod);
+	ret = delete_fte(dev, ft, fte);
+	restore_spec_mailbox(fte, &mbox_mod);
+
+	return ret;
+}
+
+static int
+mlx5_fpga_ipsec_fs_create_flow_group_egress(struct mlx5_core_dev *dev,
+					    struct mlx5_flow_table *ft,
+					    u32 *in,
+					    unsigned int *group_id)
+{
+	return fpga_ipsec_fs_create_flow_group(dev, ft, in, group_id, true);
+}
+
+static int
+mlx5_fpga_ipsec_fs_create_fte_egress(struct mlx5_core_dev *dev,
+				     struct mlx5_flow_table *ft,
+				     struct mlx5_flow_group *fg,
+				     struct fs_fte *fte)
+{
+	return fpga_ipsec_fs_create_fte(dev, ft, fg, fte, true);
+}
+
+static int
+mlx5_fpga_ipsec_fs_update_fte_egress(struct mlx5_core_dev *dev,
+				     struct mlx5_flow_table *ft,
+				     unsigned int group_id,
+				     int modify_mask,
+				     struct fs_fte *fte)
+{
+	return fpga_ipsec_fs_update_fte(dev, ft, group_id, modify_mask, fte,
+					true);
+}
+
+static int
+mlx5_fpga_ipsec_fs_delete_fte_egress(struct mlx5_core_dev *dev,
+				     struct mlx5_flow_table *ft,
+				     struct fs_fte *fte)
+{
+	return fpga_ipsec_fs_delete_fte(dev, ft, fte, true);
+}
+
+static int
+mlx5_fpga_ipsec_fs_create_flow_group_ingress(struct mlx5_core_dev *dev,
+					     struct mlx5_flow_table *ft,
+					     u32 *in,
+					     unsigned int *group_id)
+{
+	return fpga_ipsec_fs_create_flow_group(dev, ft, in, group_id, false);
+}
+
+static int
+mlx5_fpga_ipsec_fs_create_fte_ingress(struct mlx5_core_dev *dev,
+				      struct mlx5_flow_table *ft,
+				      struct mlx5_flow_group *fg,
+				      struct fs_fte *fte)
+{
+	return fpga_ipsec_fs_create_fte(dev, ft, fg, fte, false);
+}
+
+static int
+mlx5_fpga_ipsec_fs_update_fte_ingress(struct mlx5_core_dev *dev,
+				      struct mlx5_flow_table *ft,
+				      unsigned int group_id,
+				      int modify_mask,
+				      struct fs_fte *fte)
+{
+	return fpga_ipsec_fs_update_fte(dev, ft, group_id, modify_mask, fte,
+					false);
+}
+
+static int
+mlx5_fpga_ipsec_fs_delete_fte_ingress(struct mlx5_core_dev *dev,
+				      struct mlx5_flow_table *ft,
+				      struct fs_fte *fte)
+{
+	return fpga_ipsec_fs_delete_fte(dev, ft, fte, false);
+}
+
+static struct mlx5_flow_cmds fpga_ipsec_ingress;
+static struct mlx5_flow_cmds fpga_ipsec_egress;
+
+const struct mlx5_flow_cmds *mlx5_fs_cmd_get_default_ipsec_fpga_cmds(enum fs_flow_table_type type)
+{
+	switch (type) {
+	case FS_FT_NIC_RX:
+		return &fpga_ipsec_ingress;
+	case FS_FT_NIC_TX:
+		return &fpga_ipsec_egress;
+	default:
+		WARN_ON(true);
+		return NULL;
+	}
+}
+
+int mlx5_fpga_ipsec_init(struct mlx5_core_dev *mdev)
+{
+	struct mlx5_fpga_conn_attr init_attr = {0};
+	struct mlx5_fpga_device *fdev = mdev->fpga;
+	struct mlx5_fpga_conn *conn;
+	int err;
+
+	if (!mlx5_fpga_is_ipsec_device(mdev))
+		return 0;
+
+	fdev->ipsec = kzalloc(sizeof(*fdev->ipsec), GFP_KERNEL);
+	if (!fdev->ipsec)
+		return -ENOMEM;
+
+	fdev->ipsec->fdev = fdev;
+
+	err = mlx5_fpga_get_sbu_caps(fdev, sizeof(fdev->ipsec->caps),
+				     fdev->ipsec->caps);
+	if (err) {
+		mlx5_fpga_err(fdev, "Failed to retrieve IPSec extended capabilities: %d\n",
+			      err);
+		goto error;
+	}
+
+	INIT_LIST_HEAD(&fdev->ipsec->pending_cmds);
+	spin_lock_init(&fdev->ipsec->pending_cmds_lock);
+
+	init_attr.rx_size = SBU_QP_QUEUE_SIZE;
+	init_attr.tx_size = SBU_QP_QUEUE_SIZE;
+	init_attr.recv_cb = mlx5_fpga_ipsec_recv;
+	init_attr.cb_arg = fdev;
+	conn = mlx5_fpga_sbu_conn_create(fdev, &init_attr);
+	if (IS_ERR(conn)) {
+		err = PTR_ERR(conn);
+		mlx5_fpga_err(fdev, "Error creating IPSec command connection %d\n",
+			      err);
+		goto error;
+	}
+	fdev->ipsec->conn = conn;
+
+	err = rhashtable_init(&fdev->ipsec->sa_hash, &rhash_sa);
+	if (err)
+		goto err_destroy_conn;
+	mutex_init(&fdev->ipsec->sa_hash_lock);
+
+	fdev->ipsec->rules_rb = RB_ROOT;
+	mutex_init(&fdev->ipsec->rules_rb_lock);
+
+	err = mlx5_fpga_ipsec_enable_supported_caps(mdev);
+	if (err) {
+		mlx5_fpga_err(fdev, "Failed to enable IPSec extended capabilities: %d\n",
+			      err);
+		goto err_destroy_hash;
+	}
+
+	return 0;
+
+err_destroy_hash:
+	rhashtable_destroy(&fdev->ipsec->sa_hash);
+
+err_destroy_conn:
+	mlx5_fpga_sbu_conn_destroy(conn);
+
+error:
+	kfree(fdev->ipsec);
+	fdev->ipsec = NULL;
+	return err;
+}
+
+static void destroy_rules_rb(struct rb_root *root)
+{
+	struct mlx5_fpga_ipsec_rule *r, *tmp;
+
+	rbtree_postorder_for_each_entry_safe(r, tmp, root, node) {
+		rb_erase(&r->node, root);
+		mlx5_fpga_ipsec_delete_sa_ctx(r->ctx);
+		kfree(r);
+	}
+}
+
+void mlx5_fpga_ipsec_cleanup(struct mlx5_core_dev *mdev)
+{
+	struct mlx5_fpga_device *fdev = mdev->fpga;
+
+	if (!mlx5_fpga_is_ipsec_device(mdev))
+		return;
+
+	destroy_rules_rb(&fdev->ipsec->rules_rb);
+	rhashtable_destroy(&fdev->ipsec->sa_hash);
+
+	mlx5_fpga_sbu_conn_destroy(fdev->ipsec->conn);
+	kfree(fdev->ipsec);
+	fdev->ipsec = NULL;
+}
+
+void mlx5_fpga_ipsec_build_fs_cmds(void)
+{
+	/* ingress */
+	fpga_ipsec_ingress.create_flow_table =
+		mlx5_fs_cmd_get_default(egress_to_fs_ft(false))->create_flow_table;
+	fpga_ipsec_ingress.destroy_flow_table =
+		mlx5_fs_cmd_get_default(egress_to_fs_ft(false))->destroy_flow_table;
+	fpga_ipsec_ingress.modify_flow_table =
+		mlx5_fs_cmd_get_default(egress_to_fs_ft(false))->modify_flow_table;
+	fpga_ipsec_ingress.create_flow_group =
+		mlx5_fpga_ipsec_fs_create_flow_group_ingress;
+	fpga_ipsec_ingress.destroy_flow_group =
+		 mlx5_fs_cmd_get_default(egress_to_fs_ft(false))->destroy_flow_group;
+	fpga_ipsec_ingress.create_fte =
+		mlx5_fpga_ipsec_fs_create_fte_ingress;
+	fpga_ipsec_ingress.update_fte =
+		mlx5_fpga_ipsec_fs_update_fte_ingress;
+	fpga_ipsec_ingress.delete_fte =
+		mlx5_fpga_ipsec_fs_delete_fte_ingress;
+	fpga_ipsec_ingress.update_root_ft =
+		mlx5_fs_cmd_get_default(egress_to_fs_ft(false))->update_root_ft;
+
+	/* egress */
+	fpga_ipsec_egress.create_flow_table =
+		mlx5_fs_cmd_get_default(egress_to_fs_ft(true))->create_flow_table;
+	fpga_ipsec_egress.destroy_flow_table =
+		mlx5_fs_cmd_get_default(egress_to_fs_ft(true))->destroy_flow_table;
+	fpga_ipsec_egress.modify_flow_table =
+		mlx5_fs_cmd_get_default(egress_to_fs_ft(true))->modify_flow_table;
+	fpga_ipsec_egress.create_flow_group =
+		mlx5_fpga_ipsec_fs_create_flow_group_egress;
+	fpga_ipsec_egress.destroy_flow_group =
+		mlx5_fs_cmd_get_default(egress_to_fs_ft(true))->destroy_flow_group;
+	fpga_ipsec_egress.create_fte =
+		mlx5_fpga_ipsec_fs_create_fte_egress;
+	fpga_ipsec_egress.update_fte =
+		mlx5_fpga_ipsec_fs_update_fte_egress;
+	fpga_ipsec_egress.delete_fte =
+		mlx5_fpga_ipsec_fs_delete_fte_egress;
+	fpga_ipsec_egress.update_root_ft =
+		mlx5_fs_cmd_get_default(egress_to_fs_ft(true))->update_root_ft;
+}
+
+static int
+mlx5_fpga_esp_validate_xfrm_attrs(struct mlx5_core_dev *mdev,
+				  const struct mlx5_accel_esp_xfrm_attrs *attrs)
+{
+	if (attrs->tfc_pad) {
+		mlx5_core_err(mdev, "Cannot offload xfrm states with tfc padding\n");
+		return -EOPNOTSUPP;
+	}
+
+	if (attrs->replay_type != MLX5_ACCEL_ESP_REPLAY_NONE) {
+		mlx5_core_err(mdev, "Cannot offload xfrm states with anti replay\n");
+		return -EOPNOTSUPP;
+	}
+
+	if (attrs->keymat_type != MLX5_ACCEL_ESP_KEYMAT_AES_GCM) {
+		mlx5_core_err(mdev, "Only aes gcm keymat is supported\n");
+		return -EOPNOTSUPP;
+	}
+
+	if (attrs->keymat.aes_gcm.iv_algo !=
+	    MLX5_ACCEL_ESP_AES_GCM_IV_ALGO_SEQ) {
+		mlx5_core_err(mdev, "Only iv sequence algo is supported\n");
+		return -EOPNOTSUPP;
+	}
+
+	if (attrs->keymat.aes_gcm.icv_len != 128) {
+		mlx5_core_err(mdev, "Cannot offload xfrm states with AEAD ICV length other than 128bit\n");
+		return -EOPNOTSUPP;
+	}
+
+	if (attrs->keymat.aes_gcm.key_len != 128 &&
+	    attrs->keymat.aes_gcm.key_len != 256) {
+		mlx5_core_err(mdev, "Cannot offload xfrm states with AEAD key length other than 128/256 bit\n");
+		return -EOPNOTSUPP;
+	}
+
+	if ((attrs->flags & MLX5_ACCEL_ESP_FLAGS_ESN_TRIGGERED) &&
+	    (!MLX5_GET(ipsec_extended_cap, mdev->fpga->ipsec->caps,
+		       v2_command))) {
+		mlx5_core_err(mdev, "Cannot offload xfrm states with AEAD key length other than 128/256 bit\n");
+		return -EOPNOTSUPP;
+	}
+
+	return 0;
+}
+
+struct mlx5_accel_esp_xfrm *
+mlx5_fpga_esp_create_xfrm(struct mlx5_core_dev *mdev,
+			  const struct mlx5_accel_esp_xfrm_attrs *attrs,
+			  u32 flags)
+{
+	struct mlx5_fpga_esp_xfrm *fpga_xfrm;
+
+	if (!(flags & MLX5_ACCEL_XFRM_FLAG_REQUIRE_METADATA)) {
+		mlx5_core_warn(mdev, "Tried to create an esp action without metadata\n");
+		return ERR_PTR(-EINVAL);
+	}
+
+	if (mlx5_fpga_esp_validate_xfrm_attrs(mdev, attrs)) {
+		mlx5_core_warn(mdev, "Tried to create an esp with unsupported attrs\n");
+		return ERR_PTR(-EOPNOTSUPP);
+	}
+
+	fpga_xfrm = kzalloc(sizeof(*fpga_xfrm), GFP_KERNEL);
+	if (!fpga_xfrm)
+		return ERR_PTR(-ENOMEM);
+
+	mutex_init(&fpga_xfrm->lock);
+	memcpy(&fpga_xfrm->accel_xfrm.attrs, attrs,
+	       sizeof(fpga_xfrm->accel_xfrm.attrs));
+
+	return &fpga_xfrm->accel_xfrm;
+}
+
+void mlx5_fpga_esp_destroy_xfrm(struct mlx5_accel_esp_xfrm *xfrm)
+{
+	struct mlx5_fpga_esp_xfrm *fpga_xfrm =
+			container_of(xfrm, struct mlx5_fpga_esp_xfrm,
+				     accel_xfrm);
+	/* assuming no sa_ctx are connected to this xfrm_ctx */
+	kfree(fpga_xfrm);
+}
+
+int mlx5_fpga_esp_modify_xfrm(struct mlx5_accel_esp_xfrm *xfrm,
+			      const struct mlx5_accel_esp_xfrm_attrs *attrs)
+{
+	struct mlx5_core_dev *mdev = xfrm->mdev;
+	struct mlx5_fpga_device *fdev = mdev->fpga;
+	struct mlx5_fpga_ipsec *fipsec = fdev->ipsec;
+	struct mlx5_fpga_esp_xfrm *fpga_xfrm;
+	struct mlx5_ifc_fpga_ipsec_sa org_hw_sa;
+
+	int err = 0;
+
+	if (!memcmp(&xfrm->attrs, attrs, sizeof(xfrm->attrs)))
+		return 0;
+
+	if (mlx5_fpga_esp_validate_xfrm_attrs(mdev, attrs)) {
+		mlx5_core_warn(mdev, "Tried to create an esp with unsupported attrs\n");
+		return -EOPNOTSUPP;
+	}
+
+	if (is_v2_sadb_supported(fipsec)) {
+		mlx5_core_warn(mdev, "Modify esp is not supported\n");
+		return -EOPNOTSUPP;
+	}
+
+	fpga_xfrm = container_of(xfrm, struct mlx5_fpga_esp_xfrm, accel_xfrm);
+
+	mutex_lock(&fpga_xfrm->lock);
+
+	if (!fpga_xfrm->sa_ctx)
+		/* Unbounded xfrm, chane only sw attrs */
+		goto change_sw_xfrm_attrs;
+
+	/* copy original hw sa */
+	memcpy(&org_hw_sa, &fpga_xfrm->sa_ctx->hw_sa, sizeof(org_hw_sa));
+	mutex_lock(&fipsec->sa_hash_lock);
+	/* remove original hw sa from hash */
+	WARN_ON(rhashtable_remove_fast(&fipsec->sa_hash,
+				       &fpga_xfrm->sa_ctx->hash, rhash_sa));
+	/* update hw_sa with new xfrm attrs*/
+	mlx5_fpga_ipsec_build_hw_xfrm(xfrm->mdev, attrs,
+				      &fpga_xfrm->sa_ctx->hw_sa);
+	/* try to insert new hw_sa to hash */
+	err = rhashtable_insert_fast(&fipsec->sa_hash,
+				     &fpga_xfrm->sa_ctx->hash, rhash_sa);
+	if (err)
+		goto rollback_sa;
+
+	/* modify device with new hw_sa */
+	err = mlx5_fpga_ipsec_update_hw_sa(fdev, &fpga_xfrm->sa_ctx->hw_sa,
+					   MLX5_FPGA_IPSEC_CMD_OP_MOD_SA_V2);
+	fpga_xfrm->sa_ctx->hw_sa.ipsec_sa_v1.cmd = 0;
+	if (err)
+		WARN_ON(rhashtable_remove_fast(&fipsec->sa_hash,
+					       &fpga_xfrm->sa_ctx->hash,
+					       rhash_sa));
+rollback_sa:
+	if (err) {
+		/* return original hw_sa to hash */
+		memcpy(&fpga_xfrm->sa_ctx->hw_sa, &org_hw_sa,
+		       sizeof(org_hw_sa));
+		WARN_ON(rhashtable_insert_fast(&fipsec->sa_hash,
+					       &fpga_xfrm->sa_ctx->hash,
+					       rhash_sa));
+	}
+	mutex_unlock(&fipsec->sa_hash_lock);
+
+change_sw_xfrm_attrs:
+	if (!err)
+		memcpy(&xfrm->attrs, attrs, sizeof(xfrm->attrs));
+	mutex_unlock(&fpga_xfrm->lock);
+	return err;
+}
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/fpga/ipsec.h b/drivers/net/ethernet/mellanox/mlx5/core/fpga/ipsec.h
new file mode 100644
index 000000000..2b5e63b0d
--- /dev/null
+++ b/drivers/net/ethernet/mellanox/mlx5/core/fpga/ipsec.h
@@ -0,0 +1,142 @@
+/*
+ * Copyright (c) 2017 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 __MLX5_FPGA_IPSEC_H__
+#define __MLX5_FPGA_IPSEC_H__
+
+#include "accel/ipsec.h"
+#include "fs_cmd.h"
+
+#ifdef CONFIG_MLX5_FPGA
+
+u32 mlx5_fpga_ipsec_device_caps(struct mlx5_core_dev *mdev);
+unsigned int mlx5_fpga_ipsec_counters_count(struct mlx5_core_dev *mdev);
+int mlx5_fpga_ipsec_counters_read(struct mlx5_core_dev *mdev, u64 *counters,
+				  unsigned int counters_count);
+
+void *mlx5_fpga_ipsec_create_sa_ctx(struct mlx5_core_dev *mdev,
+				    struct mlx5_accel_esp_xfrm *accel_xfrm,
+				    const __be32 saddr[4],
+				    const __be32 daddr[4],
+				    const __be32 spi, bool is_ipv6);
+void mlx5_fpga_ipsec_delete_sa_ctx(void *context);
+
+int mlx5_fpga_ipsec_init(struct mlx5_core_dev *mdev);
+void mlx5_fpga_ipsec_cleanup(struct mlx5_core_dev *mdev);
+void mlx5_fpga_ipsec_build_fs_cmds(void);
+
+struct mlx5_accel_esp_xfrm *
+mlx5_fpga_esp_create_xfrm(struct mlx5_core_dev *mdev,
+			  const struct mlx5_accel_esp_xfrm_attrs *attrs,
+			  u32 flags);
+void mlx5_fpga_esp_destroy_xfrm(struct mlx5_accel_esp_xfrm *xfrm);
+int mlx5_fpga_esp_modify_xfrm(struct mlx5_accel_esp_xfrm *xfrm,
+			      const struct mlx5_accel_esp_xfrm_attrs *attrs);
+
+const struct mlx5_flow_cmds *
+mlx5_fs_cmd_get_default_ipsec_fpga_cmds(enum fs_flow_table_type type);
+
+#else
+
+static inline u32 mlx5_fpga_ipsec_device_caps(struct mlx5_core_dev *mdev)
+{
+	return 0;
+}
+
+static inline unsigned int
+mlx5_fpga_ipsec_counters_count(struct mlx5_core_dev *mdev)
+{
+	return 0;
+}
+
+static inline int mlx5_fpga_ipsec_counters_read(struct mlx5_core_dev *mdev,
+						u64 *counters)
+{
+	return 0;
+}
+
+static inline void *
+mlx5_fpga_ipsec_create_sa_ctx(struct mlx5_core_dev *mdev,
+			      struct mlx5_accel_esp_xfrm *accel_xfrm,
+			      const __be32 saddr[4],
+			      const __be32 daddr[4],
+			      const __be32 spi, bool is_ipv6)
+{
+	return NULL;
+}
+
+static inline void mlx5_fpga_ipsec_delete_sa_ctx(void *context)
+{
+}
+
+static inline int mlx5_fpga_ipsec_init(struct mlx5_core_dev *mdev)
+{
+	return 0;
+}
+
+static inline void mlx5_fpga_ipsec_cleanup(struct mlx5_core_dev *mdev)
+{
+}
+
+static inline void mlx5_fpga_ipsec_build_fs_cmds(void)
+{
+}
+
+static inline struct mlx5_accel_esp_xfrm *
+mlx5_fpga_esp_create_xfrm(struct mlx5_core_dev *mdev,
+			  const struct mlx5_accel_esp_xfrm_attrs *attrs,
+			  u32 flags)
+{
+	return ERR_PTR(-EOPNOTSUPP);
+}
+
+static inline void mlx5_fpga_esp_destroy_xfrm(struct mlx5_accel_esp_xfrm *xfrm)
+{
+}
+
+static inline int
+mlx5_fpga_esp_modify_xfrm(struct mlx5_accel_esp_xfrm *xfrm,
+			  const struct mlx5_accel_esp_xfrm_attrs *attrs)
+{
+	return -EOPNOTSUPP;
+}
+
+static inline const struct mlx5_flow_cmds *
+mlx5_fs_cmd_get_default_ipsec_fpga_cmds(enum fs_flow_table_type type)
+{
+	return mlx5_fs_cmd_get_default(type);
+}
+
+#endif /* CONFIG_MLX5_FPGA */
+
+#endif	/* __MLX5_FPGA_SADB_H__ */
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/fpga/sdk.c b/drivers/net/ethernet/mellanox/mlx5/core/fpga/sdk.c
new file mode 100644
index 000000000..14962969c
--- /dev/null
+++ b/drivers/net/ethernet/mellanox/mlx5/core/fpga/sdk.c
@@ -0,0 +1,170 @@
+/*
+ * Copyright (c) 2017 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/mlx5/device.h>
+
+#include "fpga/core.h"
+#include "fpga/conn.h"
+#include "fpga/sdk.h"
+
+struct mlx5_fpga_conn *
+mlx5_fpga_sbu_conn_create(struct mlx5_fpga_device *fdev,
+			  struct mlx5_fpga_conn_attr *attr)
+{
+	return mlx5_fpga_conn_create(fdev, attr, MLX5_FPGA_QPC_QP_TYPE_SANDBOX_QP);
+}
+EXPORT_SYMBOL(mlx5_fpga_sbu_conn_create);
+
+void mlx5_fpga_sbu_conn_destroy(struct mlx5_fpga_conn *conn)
+{
+	mlx5_fpga_conn_destroy(conn);
+}
+EXPORT_SYMBOL(mlx5_fpga_sbu_conn_destroy);
+
+int mlx5_fpga_sbu_conn_sendmsg(struct mlx5_fpga_conn *conn,
+			       struct mlx5_fpga_dma_buf *buf)
+{
+	return mlx5_fpga_conn_send(conn, buf);
+}
+EXPORT_SYMBOL(mlx5_fpga_sbu_conn_sendmsg);
+
+static int mlx5_fpga_mem_read_i2c(struct mlx5_fpga_device *fdev, size_t size,
+				  u64 addr, u8 *buf)
+{
+	size_t max_size = MLX5_FPGA_ACCESS_REG_SIZE_MAX;
+	size_t bytes_done = 0;
+	u8 actual_size;
+	int err;
+
+	if (!size)
+		return -EINVAL;
+
+	if (!fdev->mdev)
+		return -ENOTCONN;
+
+	while (bytes_done < size) {
+		actual_size = min(max_size, (size - bytes_done));
+
+		err = mlx5_fpga_access_reg(fdev->mdev, actual_size,
+					   addr + bytes_done,
+					   buf + bytes_done, false);
+		if (err) {
+			mlx5_fpga_err(fdev, "Failed to read over I2C: %d\n",
+				      err);
+			break;
+		}
+
+		bytes_done += actual_size;
+	}
+
+	return err;
+}
+
+static int mlx5_fpga_mem_write_i2c(struct mlx5_fpga_device *fdev, size_t size,
+				   u64 addr, u8 *buf)
+{
+	size_t max_size = MLX5_FPGA_ACCESS_REG_SIZE_MAX;
+	size_t bytes_done = 0;
+	u8 actual_size;
+	int err;
+
+	if (!size)
+		return -EINVAL;
+
+	if (!fdev->mdev)
+		return -ENOTCONN;
+
+	while (bytes_done < size) {
+		actual_size = min(max_size, (size - bytes_done));
+
+		err = mlx5_fpga_access_reg(fdev->mdev, actual_size,
+					   addr + bytes_done,
+					   buf + bytes_done, true);
+		if (err) {
+			mlx5_fpga_err(fdev, "Failed to write FPGA crspace\n");
+			break;
+		}
+
+		bytes_done += actual_size;
+	}
+
+	return err;
+}
+
+int mlx5_fpga_mem_read(struct mlx5_fpga_device *fdev, size_t size, u64 addr,
+		       void *buf, enum mlx5_fpga_access_type access_type)
+{
+	int ret;
+
+	switch (access_type) {
+	case MLX5_FPGA_ACCESS_TYPE_I2C:
+		ret = mlx5_fpga_mem_read_i2c(fdev, size, addr, buf);
+		if (ret)
+			return ret;
+		break;
+	default:
+		mlx5_fpga_warn(fdev, "Unexpected read access_type %u\n",
+			       access_type);
+		return -EACCES;
+	}
+
+	return size;
+}
+EXPORT_SYMBOL(mlx5_fpga_mem_read);
+
+int mlx5_fpga_mem_write(struct mlx5_fpga_device *fdev, size_t size, u64 addr,
+			void *buf, enum mlx5_fpga_access_type access_type)
+{
+	int ret;
+
+	switch (access_type) {
+	case MLX5_FPGA_ACCESS_TYPE_I2C:
+		ret = mlx5_fpga_mem_write_i2c(fdev, size, addr, buf);
+		if (ret)
+			return ret;
+		break;
+	default:
+		mlx5_fpga_warn(fdev, "Unexpected write access_type %u\n",
+			       access_type);
+		return -EACCES;
+	}
+
+	return size;
+}
+EXPORT_SYMBOL(mlx5_fpga_mem_write);
+
+int mlx5_fpga_get_sbu_caps(struct mlx5_fpga_device *fdev, int size, void *buf)
+{
+	return mlx5_fpga_sbu_caps(fdev->mdev, buf, size);
+}
+EXPORT_SYMBOL(mlx5_fpga_get_sbu_caps);
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/fpga/sdk.h b/drivers/net/ethernet/mellanox/mlx5/core/fpga/sdk.h
new file mode 100644
index 000000000..656f96be6
--- /dev/null
+++ b/drivers/net/ethernet/mellanox/mlx5/core/fpga/sdk.h
@@ -0,0 +1,212 @@
+/*
+ * Copyright (c) 2017 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 MLX5_FPGA_SDK_H
+#define MLX5_FPGA_SDK_H
+
+#include <linux/types.h>
+#include <linux/dma-direction.h>
+
+/**
+ * DOC: Innova SDK
+ * This header defines the in-kernel API for Innova FPGA client drivers.
+ */
+#define SBU_QP_QUEUE_SIZE 8
+#define MLX5_FPGA_CMD_TIMEOUT_MSEC (60 * 1000)
+
+/**
+ * enum mlx5_fpga_access_type - Enumerated the different methods possible for
+ * accessing the device memory address space
+ */
+enum mlx5_fpga_access_type {
+	/** Use the slow CX-FPGA I2C bus */
+	MLX5_FPGA_ACCESS_TYPE_I2C = 0x0,
+	/** Use the fastest available method */
+	MLX5_FPGA_ACCESS_TYPE_DONTCARE = 0x0,
+};
+
+struct mlx5_fpga_conn;
+struct mlx5_fpga_device;
+
+/**
+ * struct mlx5_fpga_dma_entry - A scatter-gather DMA entry
+ */
+struct mlx5_fpga_dma_entry {
+	/** @data: Virtual address pointer to the data */
+	void *data;
+	/** @size: Size in bytes of the data */
+	unsigned int size;
+	/** @dma_addr: Private member. Physical DMA-mapped address of the data */
+	dma_addr_t dma_addr;
+};
+
+/**
+ * struct mlx5_fpga_dma_buf - A packet buffer
+ * May contain up to 2 scatter-gather data entries
+ */
+struct mlx5_fpga_dma_buf {
+	/** @dma_dir: DMA direction */
+	enum dma_data_direction dma_dir;
+	/** @sg: Scatter-gather entries pointing to the data in memory */
+	struct mlx5_fpga_dma_entry sg[2];
+	/** @list: Item in SQ backlog, for TX packets */
+	struct list_head list;
+	/**
+	 * @complete: Completion routine, for TX packets
+	 * @conn: FPGA Connection this packet was sent to
+	 * @fdev: FPGA device this packet was sent to
+	 * @buf: The packet buffer
+	 * @status: 0 if successful, or an error code otherwise
+	 */
+	void (*complete)(struct mlx5_fpga_conn *conn,
+			 struct mlx5_fpga_device *fdev,
+			 struct mlx5_fpga_dma_buf *buf, u8 status);
+};
+
+/**
+ * struct mlx5_fpga_conn_attr - FPGA connection attributes
+ * Describes the attributes of a connection
+ */
+struct mlx5_fpga_conn_attr {
+	/** @tx_size: Size of connection TX queue, in packets */
+	unsigned int tx_size;
+	/** @rx_size: Size of connection RX queue, in packets */
+	unsigned int rx_size;
+	/**
+	 * @recv_cb: Callback function which is called for received packets
+	 * @cb_arg: The value provided in mlx5_fpga_conn_attr.cb_arg
+	 * @buf: A buffer containing a received packet
+	 *
+	 * buf is guaranteed to only contain a single scatter-gather entry.
+	 * The size of the actual packet received is specified in buf.sg[0].size
+	 * When this callback returns, the packet buffer may be re-used for
+	 * subsequent receives.
+	 */
+	void (*recv_cb)(void *cb_arg, struct mlx5_fpga_dma_buf *buf);
+	void *cb_arg;
+};
+
+/**
+ * mlx5_fpga_sbu_conn_create() - Initialize a new FPGA SBU connection
+ * @fdev: The FPGA device
+ * @attr: Attributes of the new connection
+ *
+ * Sets up a new FPGA SBU connection with the specified attributes.
+ * The receive callback function may be called for incoming messages even
+ * before this function returns.
+ *
+ * The caller must eventually destroy the connection by calling
+ * mlx5_fpga_sbu_conn_destroy.
+ *
+ * Return: A new connection, or ERR_PTR() error value otherwise.
+ */
+struct mlx5_fpga_conn *
+mlx5_fpga_sbu_conn_create(struct mlx5_fpga_device *fdev,
+			  struct mlx5_fpga_conn_attr *attr);
+
+/**
+ * mlx5_fpga_sbu_conn_destroy() - Destroy an FPGA SBU connection
+ * @conn: The FPGA SBU connection to destroy
+ *
+ * Cleans up an FPGA SBU connection which was previously created with
+ * mlx5_fpga_sbu_conn_create.
+ */
+void mlx5_fpga_sbu_conn_destroy(struct mlx5_fpga_conn *conn);
+
+/**
+ * mlx5_fpga_sbu_conn_sendmsg() - Queue the transmission of a packet
+ * @fdev: An FPGA SBU connection
+ * @buf: The packet buffer
+ *
+ * Queues a packet for transmission over an FPGA SBU connection.
+ * The buffer should not be modified or freed until completion.
+ * Upon completion, the buf's complete() callback is invoked, indicating the
+ * success or error status of the transmission.
+ *
+ * Return: 0 if successful, or an error value otherwise.
+ */
+int mlx5_fpga_sbu_conn_sendmsg(struct mlx5_fpga_conn *conn,
+			       struct mlx5_fpga_dma_buf *buf);
+
+/**
+ * mlx5_fpga_mem_read() - Read from FPGA memory address space
+ * @fdev: The FPGA device
+ * @size: Size of chunk to read, in bytes
+ * @addr: Starting address to read from, in FPGA address space
+ * @buf: Buffer to read into
+ * @access_type: Method for reading
+ *
+ * Reads from the specified address into the specified buffer.
+ * The address may point to configuration space or to DDR.
+ * Large reads may be performed internally as several non-atomic operations.
+ * This function may sleep, so should not be called from atomic contexts.
+ *
+ * Return: 0 if successful, or an error value otherwise.
+ */
+int mlx5_fpga_mem_read(struct mlx5_fpga_device *fdev, size_t size, u64 addr,
+		       void *buf, enum mlx5_fpga_access_type access_type);
+
+/**
+ * mlx5_fpga_mem_write() - Write to FPGA memory address space
+ * @fdev: The FPGA device
+ * @size: Size of chunk to write, in bytes
+ * @addr: Starting address to write to, in FPGA address space
+ * @buf: Buffer which contains data to write
+ * @access_type: Method for writing
+ *
+ * Writes the specified buffer data to FPGA memory at the specified address.
+ * The address may point to configuration space or to DDR.
+ * Large writes may be performed internally as several non-atomic operations.
+ * This function may sleep, so should not be called from atomic contexts.
+ *
+ * Return: 0 if successful, or an error value otherwise.
+ */
+int mlx5_fpga_mem_write(struct mlx5_fpga_device *fdev, size_t size, u64 addr,
+			void *buf, enum mlx5_fpga_access_type access_type);
+
+/**
+ * mlx5_fpga_get_sbu_caps() - Read the SBU capabilities
+ * @fdev: The FPGA device
+ * @size: Size of the buffer to read into
+ * @buf: Buffer to read the capabilities into
+ *
+ * Reads the FPGA SBU capabilities into the specified buffer.
+ * The format of the capabilities buffer is SBU-dependent.
+ *
+ * Return: 0 if successful
+ *         -EINVAL if the buffer is not large enough to contain SBU caps
+ *         or any other error value otherwise.
+ */
+int mlx5_fpga_get_sbu_caps(struct mlx5_fpga_device *fdev, int size, void *buf);
+
+#endif /* MLX5_FPGA_SDK_H */
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/fpga/tls.c b/drivers/net/ethernet/mellanox/mlx5/core/fpga/tls.c
new file mode 100644
index 000000000..22a2ef111
--- /dev/null
+++ b/drivers/net/ethernet/mellanox/mlx5/core/fpga/tls.c
@@ -0,0 +1,622 @@
+/*
+ * 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 <linux/mlx5/device.h>
+#include "fpga/tls.h"
+#include "fpga/cmd.h"
+#include "fpga/sdk.h"
+#include "fpga/core.h"
+#include "accel/tls.h"
+
+struct mlx5_fpga_tls_command_context;
+
+typedef void (*mlx5_fpga_tls_command_complete)
+	(struct mlx5_fpga_conn *conn, struct mlx5_fpga_device *fdev,
+	 struct mlx5_fpga_tls_command_context *ctx,
+	 struct mlx5_fpga_dma_buf *resp);
+
+struct mlx5_fpga_tls_command_context {
+	struct list_head list;
+	/* There is no guarantee on the order between the TX completion
+	 * and the command response.
+	 * The TX completion is going to touch cmd->buf even in
+	 * the case of successful transmission.
+	 * So instead of requiring separate allocations for cmd
+	 * and cmd->buf we've decided to use a reference counter
+	 */
+	refcount_t ref;
+	struct mlx5_fpga_dma_buf buf;
+	mlx5_fpga_tls_command_complete complete;
+};
+
+static void
+mlx5_fpga_tls_put_command_ctx(struct mlx5_fpga_tls_command_context *ctx)
+{
+	if (refcount_dec_and_test(&ctx->ref))
+		kfree(ctx);
+}
+
+static void mlx5_fpga_tls_cmd_complete(struct mlx5_fpga_device *fdev,
+				       struct mlx5_fpga_dma_buf *resp)
+{
+	struct mlx5_fpga_conn *conn = fdev->tls->conn;
+	struct mlx5_fpga_tls_command_context *ctx;
+	struct mlx5_fpga_tls *tls = fdev->tls;
+	unsigned long flags;
+
+	spin_lock_irqsave(&tls->pending_cmds_lock, flags);
+	ctx = list_first_entry(&tls->pending_cmds,
+			       struct mlx5_fpga_tls_command_context, list);
+	list_del(&ctx->list);
+	spin_unlock_irqrestore(&tls->pending_cmds_lock, flags);
+	ctx->complete(conn, fdev, ctx, resp);
+}
+
+static void mlx5_fpga_cmd_send_complete(struct mlx5_fpga_conn *conn,
+					struct mlx5_fpga_device *fdev,
+					struct mlx5_fpga_dma_buf *buf,
+					u8 status)
+{
+	struct mlx5_fpga_tls_command_context *ctx =
+	    container_of(buf, struct mlx5_fpga_tls_command_context, buf);
+
+	mlx5_fpga_tls_put_command_ctx(ctx);
+
+	if (unlikely(status))
+		mlx5_fpga_tls_cmd_complete(fdev, NULL);
+}
+
+static void mlx5_fpga_tls_cmd_send(struct mlx5_fpga_device *fdev,
+				   struct mlx5_fpga_tls_command_context *cmd,
+				   mlx5_fpga_tls_command_complete complete)
+{
+	struct mlx5_fpga_tls *tls = fdev->tls;
+	unsigned long flags;
+	int ret;
+
+	refcount_set(&cmd->ref, 2);
+	cmd->complete = complete;
+	cmd->buf.complete = mlx5_fpga_cmd_send_complete;
+
+	spin_lock_irqsave(&tls->pending_cmds_lock, flags);
+	/* mlx5_fpga_sbu_conn_sendmsg is called under pending_cmds_lock
+	 * to make sure commands are inserted to the tls->pending_cmds list
+	 * and the command QP in the same order.
+	 */
+	ret = mlx5_fpga_sbu_conn_sendmsg(tls->conn, &cmd->buf);
+	if (likely(!ret))
+		list_add_tail(&cmd->list, &tls->pending_cmds);
+	else
+		complete(tls->conn, fdev, cmd, NULL);
+	spin_unlock_irqrestore(&tls->pending_cmds_lock, flags);
+}
+
+/* Start of context identifiers range (inclusive) */
+#define SWID_START	0
+/* End of context identifiers range (exclusive) */
+#define SWID_END	BIT(24)
+
+static int mlx5_fpga_tls_alloc_swid(struct idr *idr, spinlock_t *idr_spinlock,
+				    void *ptr)
+{
+	unsigned long flags;
+	int ret;
+
+	/* TLS metadata format is 1 byte for syndrome followed
+	 * by 3 bytes of swid (software ID)
+	 * swid must not exceed 3 bytes.
+	 * See tls_rxtx.c:insert_pet() for details
+	 */
+	BUILD_BUG_ON((SWID_END - 1) & 0xFF000000);
+
+	idr_preload(GFP_KERNEL);
+	spin_lock_irqsave(idr_spinlock, flags);
+	ret = idr_alloc(idr, ptr, SWID_START, SWID_END, GFP_ATOMIC);
+	spin_unlock_irqrestore(idr_spinlock, flags);
+	idr_preload_end();
+
+	return ret;
+}
+
+static void *mlx5_fpga_tls_release_swid(struct idr *idr,
+					spinlock_t *idr_spinlock, u32 swid)
+{
+	unsigned long flags;
+	void *ptr;
+
+	spin_lock_irqsave(idr_spinlock, flags);
+	ptr = idr_remove(idr, swid);
+	spin_unlock_irqrestore(idr_spinlock, flags);
+	return ptr;
+}
+
+static void mlx_tls_kfree_complete(struct mlx5_fpga_conn *conn,
+				   struct mlx5_fpga_device *fdev,
+				   struct mlx5_fpga_dma_buf *buf, u8 status)
+{
+	kfree(buf);
+}
+
+static void
+mlx5_fpga_tls_teardown_completion(struct mlx5_fpga_conn *conn,
+				  struct mlx5_fpga_device *fdev,
+				  struct mlx5_fpga_tls_command_context *cmd,
+				  struct mlx5_fpga_dma_buf *resp)
+{
+	if (resp) {
+		u32 syndrome = MLX5_GET(tls_resp, resp->sg[0].data, syndrome);
+
+		if (syndrome)
+			mlx5_fpga_err(fdev,
+				      "Teardown stream failed with syndrome = %d",
+				      syndrome);
+	}
+	mlx5_fpga_tls_put_command_ctx(cmd);
+}
+
+static void mlx5_fpga_tls_flow_to_cmd(void *flow, void *cmd)
+{
+	memcpy(MLX5_ADDR_OF(tls_cmd, cmd, src_port), flow,
+	       MLX5_BYTE_OFF(tls_flow, ipv6));
+
+	MLX5_SET(tls_cmd, cmd, ipv6, MLX5_GET(tls_flow, flow, ipv6));
+	MLX5_SET(tls_cmd, cmd, direction_sx,
+		 MLX5_GET(tls_flow, flow, direction_sx));
+}
+
+int mlx5_fpga_tls_resync_rx(struct mlx5_core_dev *mdev, u32 handle, u32 seq,
+			    u64 rcd_sn)
+{
+	struct mlx5_fpga_dma_buf *buf;
+	int size = sizeof(*buf) + MLX5_TLS_COMMAND_SIZE;
+	void *flow;
+	void *cmd;
+	int ret;
+
+	buf = kzalloc(size, GFP_ATOMIC);
+	if (!buf)
+		return -ENOMEM;
+
+	cmd = (buf + 1);
+
+	rcu_read_lock();
+	flow = idr_find(&mdev->fpga->tls->rx_idr, ntohl(handle));
+	if (unlikely(!flow)) {
+		rcu_read_unlock();
+		WARN_ONCE(1, "Received NULL pointer for handle\n");
+		kfree(buf);
+		return -EINVAL;
+	}
+	mlx5_fpga_tls_flow_to_cmd(flow, cmd);
+	rcu_read_unlock();
+
+	MLX5_SET(tls_cmd, cmd, swid, ntohl(handle));
+	MLX5_SET64(tls_cmd, cmd, tls_rcd_sn, be64_to_cpu(rcd_sn));
+	MLX5_SET(tls_cmd, cmd, tcp_sn, seq);
+	MLX5_SET(tls_cmd, cmd, command_type, CMD_RESYNC_RX);
+
+	buf->sg[0].data = cmd;
+	buf->sg[0].size = MLX5_TLS_COMMAND_SIZE;
+	buf->complete = mlx_tls_kfree_complete;
+
+	ret = mlx5_fpga_sbu_conn_sendmsg(mdev->fpga->tls->conn, buf);
+	if (ret < 0)
+		kfree(buf);
+
+	return ret;
+}
+
+static void mlx5_fpga_tls_send_teardown_cmd(struct mlx5_core_dev *mdev,
+					    void *flow, u32 swid, gfp_t flags)
+{
+	struct mlx5_fpga_tls_command_context *ctx;
+	struct mlx5_fpga_dma_buf *buf;
+	void *cmd;
+
+	ctx = kzalloc(sizeof(*ctx) + MLX5_TLS_COMMAND_SIZE, flags);
+	if (!ctx)
+		return;
+
+	buf = &ctx->buf;
+	cmd = (ctx + 1);
+	MLX5_SET(tls_cmd, cmd, command_type, CMD_TEARDOWN_STREAM);
+	MLX5_SET(tls_cmd, cmd, swid, swid);
+
+	mlx5_fpga_tls_flow_to_cmd(flow, cmd);
+	kfree(flow);
+
+	buf->sg[0].data = cmd;
+	buf->sg[0].size = MLX5_TLS_COMMAND_SIZE;
+
+	mlx5_fpga_tls_cmd_send(mdev->fpga, ctx,
+			       mlx5_fpga_tls_teardown_completion);
+}
+
+void mlx5_fpga_tls_del_flow(struct mlx5_core_dev *mdev, u32 swid,
+			    gfp_t flags, bool direction_sx)
+{
+	struct mlx5_fpga_tls *tls = mdev->fpga->tls;
+	void *flow;
+
+	if (direction_sx)
+		flow = mlx5_fpga_tls_release_swid(&tls->tx_idr,
+						  &tls->tx_idr_spinlock,
+						  swid);
+	else
+		flow = mlx5_fpga_tls_release_swid(&tls->rx_idr,
+						  &tls->rx_idr_spinlock,
+						  swid);
+
+	if (!flow) {
+		mlx5_fpga_err(mdev->fpga, "No flow information for swid %u\n",
+			      swid);
+		return;
+	}
+
+	synchronize_rcu(); /* before kfree(flow) */
+	mlx5_fpga_tls_send_teardown_cmd(mdev, flow, swid, flags);
+}
+
+enum mlx5_fpga_setup_stream_status {
+	MLX5_FPGA_CMD_PENDING,
+	MLX5_FPGA_CMD_SEND_FAILED,
+	MLX5_FPGA_CMD_RESPONSE_RECEIVED,
+	MLX5_FPGA_CMD_ABANDONED,
+};
+
+struct mlx5_setup_stream_context {
+	struct mlx5_fpga_tls_command_context cmd;
+	atomic_t status;
+	u32 syndrome;
+	struct completion comp;
+};
+
+static void
+mlx5_fpga_tls_setup_completion(struct mlx5_fpga_conn *conn,
+			       struct mlx5_fpga_device *fdev,
+			       struct mlx5_fpga_tls_command_context *cmd,
+			       struct mlx5_fpga_dma_buf *resp)
+{
+	struct mlx5_setup_stream_context *ctx =
+	    container_of(cmd, struct mlx5_setup_stream_context, cmd);
+	int status = MLX5_FPGA_CMD_SEND_FAILED;
+	void *tls_cmd = ctx + 1;
+
+	/* If we failed to send to command resp == NULL */
+	if (resp) {
+		ctx->syndrome = MLX5_GET(tls_resp, resp->sg[0].data, syndrome);
+		status = MLX5_FPGA_CMD_RESPONSE_RECEIVED;
+	}
+
+	status = atomic_xchg_release(&ctx->status, status);
+	if (likely(status != MLX5_FPGA_CMD_ABANDONED)) {
+		complete(&ctx->comp);
+		return;
+	}
+
+	mlx5_fpga_err(fdev, "Command was abandoned, syndrome = %u\n",
+		      ctx->syndrome);
+
+	if (!ctx->syndrome) {
+		/* The process was killed while waiting for the context to be
+		 * added, and the add completed successfully.
+		 * We need to destroy the HW context, and we can't can't reuse
+		 * the command context because we might not have received
+		 * the tx completion yet.
+		 */
+		mlx5_fpga_tls_del_flow(fdev->mdev,
+				       MLX5_GET(tls_cmd, tls_cmd, swid),
+				       GFP_ATOMIC,
+				       MLX5_GET(tls_cmd, tls_cmd,
+						direction_sx));
+	}
+
+	mlx5_fpga_tls_put_command_ctx(cmd);
+}
+
+static int mlx5_fpga_tls_setup_stream_cmd(struct mlx5_core_dev *mdev,
+					  struct mlx5_setup_stream_context *ctx)
+{
+	struct mlx5_fpga_dma_buf *buf;
+	void *cmd = ctx + 1;
+	int status, ret = 0;
+
+	buf = &ctx->cmd.buf;
+	buf->sg[0].data = cmd;
+	buf->sg[0].size = MLX5_TLS_COMMAND_SIZE;
+	MLX5_SET(tls_cmd, cmd, command_type, CMD_SETUP_STREAM);
+
+	init_completion(&ctx->comp);
+	atomic_set(&ctx->status, MLX5_FPGA_CMD_PENDING);
+	ctx->syndrome = -1;
+
+	mlx5_fpga_tls_cmd_send(mdev->fpga, &ctx->cmd,
+			       mlx5_fpga_tls_setup_completion);
+	wait_for_completion_killable(&ctx->comp);
+
+	status = atomic_xchg_acquire(&ctx->status, MLX5_FPGA_CMD_ABANDONED);
+	if (unlikely(status == MLX5_FPGA_CMD_PENDING))
+	/* ctx is going to be released in mlx5_fpga_tls_setup_completion */
+		return -EINTR;
+
+	if (unlikely(ctx->syndrome))
+		ret = -ENOMEM;
+
+	mlx5_fpga_tls_put_command_ctx(&ctx->cmd);
+	return ret;
+}
+
+static void mlx5_fpga_tls_hw_qp_recv_cb(void *cb_arg,
+					struct mlx5_fpga_dma_buf *buf)
+{
+	struct mlx5_fpga_device *fdev = (struct mlx5_fpga_device *)cb_arg;
+
+	mlx5_fpga_tls_cmd_complete(fdev, buf);
+}
+
+bool mlx5_fpga_is_tls_device(struct mlx5_core_dev *mdev)
+{
+	if (!mdev->fpga || !MLX5_CAP_GEN(mdev, fpga))
+		return false;
+
+	if (MLX5_CAP_FPGA(mdev, ieee_vendor_id) !=
+	    MLX5_FPGA_CAP_SANDBOX_VENDOR_ID_MLNX)
+		return false;
+
+	if (MLX5_CAP_FPGA(mdev, sandbox_product_id) !=
+	    MLX5_FPGA_CAP_SANDBOX_PRODUCT_ID_TLS)
+		return false;
+
+	if (MLX5_CAP_FPGA(mdev, sandbox_product_version) != 0)
+		return false;
+
+	return true;
+}
+
+static int mlx5_fpga_tls_get_caps(struct mlx5_fpga_device *fdev,
+				  u32 *p_caps)
+{
+	int err, cap_size = MLX5_ST_SZ_BYTES(tls_extended_cap);
+	u32 caps = 0;
+	void *buf;
+
+	buf = kzalloc(cap_size, GFP_KERNEL);
+	if (!buf)
+		return -ENOMEM;
+
+	err = mlx5_fpga_get_sbu_caps(fdev, cap_size, buf);
+	if (err)
+		goto out;
+
+	if (MLX5_GET(tls_extended_cap, buf, tx))
+		caps |= MLX5_ACCEL_TLS_TX;
+	if (MLX5_GET(tls_extended_cap, buf, rx))
+		caps |= MLX5_ACCEL_TLS_RX;
+	if (MLX5_GET(tls_extended_cap, buf, tls_v12))
+		caps |= MLX5_ACCEL_TLS_V12;
+	if (MLX5_GET(tls_extended_cap, buf, tls_v13))
+		caps |= MLX5_ACCEL_TLS_V13;
+	if (MLX5_GET(tls_extended_cap, buf, lro))
+		caps |= MLX5_ACCEL_TLS_LRO;
+	if (MLX5_GET(tls_extended_cap, buf, ipv6))
+		caps |= MLX5_ACCEL_TLS_IPV6;
+
+	if (MLX5_GET(tls_extended_cap, buf, aes_gcm_128))
+		caps |= MLX5_ACCEL_TLS_AES_GCM128;
+	if (MLX5_GET(tls_extended_cap, buf, aes_gcm_256))
+		caps |= MLX5_ACCEL_TLS_AES_GCM256;
+
+	*p_caps = caps;
+	err = 0;
+out:
+	kfree(buf);
+	return err;
+}
+
+int mlx5_fpga_tls_init(struct mlx5_core_dev *mdev)
+{
+	struct mlx5_fpga_device *fdev = mdev->fpga;
+	struct mlx5_fpga_conn_attr init_attr = {0};
+	struct mlx5_fpga_conn *conn;
+	struct mlx5_fpga_tls *tls;
+	int err = 0;
+
+	if (!mlx5_fpga_is_tls_device(mdev) || !fdev)
+		return 0;
+
+	tls = kzalloc(sizeof(*tls), GFP_KERNEL);
+	if (!tls)
+		return -ENOMEM;
+
+	err = mlx5_fpga_tls_get_caps(fdev, &tls->caps);
+	if (err)
+		goto error;
+
+	if (!(tls->caps & (MLX5_ACCEL_TLS_V12 | MLX5_ACCEL_TLS_AES_GCM128))) {
+		err = -ENOTSUPP;
+		goto error;
+	}
+
+	init_attr.rx_size = SBU_QP_QUEUE_SIZE;
+	init_attr.tx_size = SBU_QP_QUEUE_SIZE;
+	init_attr.recv_cb = mlx5_fpga_tls_hw_qp_recv_cb;
+	init_attr.cb_arg = fdev;
+	conn = mlx5_fpga_sbu_conn_create(fdev, &init_attr);
+	if (IS_ERR(conn)) {
+		err = PTR_ERR(conn);
+		mlx5_fpga_err(fdev, "Error creating TLS command connection %d\n",
+			      err);
+		goto error;
+	}
+
+	tls->conn = conn;
+	spin_lock_init(&tls->pending_cmds_lock);
+	INIT_LIST_HEAD(&tls->pending_cmds);
+
+	idr_init(&tls->tx_idr);
+	idr_init(&tls->rx_idr);
+	spin_lock_init(&tls->tx_idr_spinlock);
+	spin_lock_init(&tls->rx_idr_spinlock);
+	fdev->tls = tls;
+	return 0;
+
+error:
+	kfree(tls);
+	return err;
+}
+
+void mlx5_fpga_tls_cleanup(struct mlx5_core_dev *mdev)
+{
+	struct mlx5_fpga_device *fdev = mdev->fpga;
+
+	if (!fdev || !fdev->tls)
+		return;
+
+	mlx5_fpga_sbu_conn_destroy(fdev->tls->conn);
+	kfree(fdev->tls);
+	fdev->tls = NULL;
+}
+
+static void mlx5_fpga_tls_set_aes_gcm128_ctx(void *cmd,
+					     struct tls_crypto_info *info,
+					     __be64 *rcd_sn)
+{
+	struct tls12_crypto_info_aes_gcm_128 *crypto_info =
+	    (struct tls12_crypto_info_aes_gcm_128 *)info;
+
+	memcpy(MLX5_ADDR_OF(tls_cmd, cmd, tls_rcd_sn), crypto_info->rec_seq,
+	       TLS_CIPHER_AES_GCM_128_REC_SEQ_SIZE);
+
+	memcpy(MLX5_ADDR_OF(tls_cmd, cmd, tls_implicit_iv),
+	       crypto_info->salt, TLS_CIPHER_AES_GCM_128_SALT_SIZE);
+	memcpy(MLX5_ADDR_OF(tls_cmd, cmd, encryption_key),
+	       crypto_info->key, TLS_CIPHER_AES_GCM_128_KEY_SIZE);
+
+	/* in AES-GCM 128 we need to write the key twice */
+	memcpy(MLX5_ADDR_OF(tls_cmd, cmd, encryption_key) +
+		   TLS_CIPHER_AES_GCM_128_KEY_SIZE,
+	       crypto_info->key, TLS_CIPHER_AES_GCM_128_KEY_SIZE);
+
+	MLX5_SET(tls_cmd, cmd, alg, MLX5_TLS_ALG_AES_GCM_128);
+}
+
+static int mlx5_fpga_tls_set_key_material(void *cmd, u32 caps,
+					  struct tls_crypto_info *crypto_info)
+{
+	__be64 rcd_sn;
+
+	switch (crypto_info->cipher_type) {
+	case TLS_CIPHER_AES_GCM_128:
+		if (!(caps & MLX5_ACCEL_TLS_AES_GCM128))
+			return -EINVAL;
+		mlx5_fpga_tls_set_aes_gcm128_ctx(cmd, crypto_info, &rcd_sn);
+		break;
+	default:
+		return -EINVAL;
+	}
+
+	return 0;
+}
+
+static int _mlx5_fpga_tls_add_flow(struct mlx5_core_dev *mdev, void *flow,
+				   struct tls_crypto_info *crypto_info,
+				   u32 swid, u32 tcp_sn)
+{
+	u32 caps = mlx5_fpga_tls_device_caps(mdev);
+	struct mlx5_setup_stream_context *ctx;
+	int ret = -ENOMEM;
+	size_t cmd_size;
+	void *cmd;
+
+	cmd_size = MLX5_TLS_COMMAND_SIZE + sizeof(*ctx);
+	ctx = kzalloc(cmd_size, GFP_KERNEL);
+	if (!ctx)
+		goto out;
+
+	cmd = ctx + 1;
+	ret = mlx5_fpga_tls_set_key_material(cmd, caps, crypto_info);
+	if (ret)
+		goto free_ctx;
+
+	mlx5_fpga_tls_flow_to_cmd(flow, cmd);
+
+	MLX5_SET(tls_cmd, cmd, swid, swid);
+	MLX5_SET(tls_cmd, cmd, tcp_sn, tcp_sn);
+
+	return mlx5_fpga_tls_setup_stream_cmd(mdev, ctx);
+
+free_ctx:
+	kfree(ctx);
+out:
+	return ret;
+}
+
+int mlx5_fpga_tls_add_flow(struct mlx5_core_dev *mdev, void *flow,
+			   struct tls_crypto_info *crypto_info,
+			   u32 start_offload_tcp_sn, u32 *p_swid,
+			   bool direction_sx)
+{
+	struct mlx5_fpga_tls *tls = mdev->fpga->tls;
+	int ret = -ENOMEM;
+	u32 swid;
+
+	if (direction_sx)
+		ret = mlx5_fpga_tls_alloc_swid(&tls->tx_idr,
+					       &tls->tx_idr_spinlock, flow);
+	else
+		ret = mlx5_fpga_tls_alloc_swid(&tls->rx_idr,
+					       &tls->rx_idr_spinlock, flow);
+
+	if (ret < 0)
+		return ret;
+
+	swid = ret;
+	MLX5_SET(tls_flow, flow, direction_sx, direction_sx ? 1 : 0);
+
+	ret = _mlx5_fpga_tls_add_flow(mdev, flow, crypto_info, swid,
+				      start_offload_tcp_sn);
+	if (ret && ret != -EINTR)
+		goto free_swid;
+
+	*p_swid = swid;
+	return 0;
+free_swid:
+	if (direction_sx)
+		mlx5_fpga_tls_release_swid(&tls->tx_idr,
+					   &tls->tx_idr_spinlock, swid);
+	else
+		mlx5_fpga_tls_release_swid(&tls->rx_idr,
+					   &tls->rx_idr_spinlock, swid);
+
+	return ret;
+}
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/fpga/tls.h b/drivers/net/ethernet/mellanox/mlx5/core/fpga/tls.h
new file mode 100644
index 000000000..3b2e37bf7
--- /dev/null
+++ b/drivers/net/ethernet/mellanox/mlx5/core/fpga/tls.h
@@ -0,0 +1,74 @@
+/*
+ * 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.
+ *
+ */
+
+#ifndef __MLX5_FPGA_TLS_H__
+#define __MLX5_FPGA_TLS_H__
+
+#include <linux/mlx5/driver.h>
+
+#include <net/tls.h>
+#include "fpga/core.h"
+
+struct mlx5_fpga_tls {
+	struct list_head pending_cmds;
+	spinlock_t pending_cmds_lock; /* Protects pending_cmds */
+	u32 caps;
+	struct mlx5_fpga_conn *conn;
+
+	struct idr tx_idr;
+	struct idr rx_idr;
+	spinlock_t tx_idr_spinlock; /* protects the IDR */
+	spinlock_t rx_idr_spinlock; /* protects the IDR */
+};
+
+int mlx5_fpga_tls_add_flow(struct mlx5_core_dev *mdev, void *flow,
+			   struct tls_crypto_info *crypto_info,
+			   u32 start_offload_tcp_sn, u32 *p_swid,
+			   bool direction_sx);
+
+void mlx5_fpga_tls_del_flow(struct mlx5_core_dev *mdev, u32 swid,
+			    gfp_t flags, bool direction_sx);
+
+bool mlx5_fpga_is_tls_device(struct mlx5_core_dev *mdev);
+int mlx5_fpga_tls_init(struct mlx5_core_dev *mdev);
+void mlx5_fpga_tls_cleanup(struct mlx5_core_dev *mdev);
+
+static inline u32 mlx5_fpga_tls_device_caps(struct mlx5_core_dev *mdev)
+{
+	return mdev->fpga->tls->caps;
+}
+
+int mlx5_fpga_tls_resync_rx(struct mlx5_core_dev *mdev, u32 handle, u32 seq,
+			    u64 rcd_sn);
+
+#endif /* __MLX5_FPGA_TLS_H__ */
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/fs_cmd.c b/drivers/net/ethernet/mellanox/mlx5/core/fs_cmd.c
new file mode 100644
index 000000000..8e01f8180
--- /dev/null
+++ b/drivers/net/ethernet/mellanox/mlx5/core/fs_cmd.c
@@ -0,0 +1,768 @@
+/*
+ * Copyright (c) 2015, 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/mlx5/driver.h>
+#include <linux/mlx5/device.h>
+#include <linux/mlx5/mlx5_ifc.h>
+
+#include "fs_core.h"
+#include "fs_cmd.h"
+#include "mlx5_core.h"
+#include "eswitch.h"
+
+static int mlx5_cmd_stub_update_root_ft(struct mlx5_core_dev *dev,
+					struct mlx5_flow_table *ft,
+					u32 underlay_qpn,
+					bool disconnect)
+{
+	return 0;
+}
+
+static int mlx5_cmd_stub_create_flow_table(struct mlx5_core_dev *dev,
+					   u16 vport,
+					   enum fs_flow_table_op_mod op_mod,
+					   enum fs_flow_table_type type,
+					   unsigned int level,
+					   unsigned int log_size,
+					   struct mlx5_flow_table *next_ft,
+					   unsigned int *table_id, u32 flags)
+{
+	return 0;
+}
+
+static int mlx5_cmd_stub_destroy_flow_table(struct mlx5_core_dev *dev,
+					    struct mlx5_flow_table *ft)
+{
+	return 0;
+}
+
+static int mlx5_cmd_stub_modify_flow_table(struct mlx5_core_dev *dev,
+					   struct mlx5_flow_table *ft,
+					   struct mlx5_flow_table *next_ft)
+{
+	return 0;
+}
+
+static int mlx5_cmd_stub_create_flow_group(struct mlx5_core_dev *dev,
+					   struct mlx5_flow_table *ft,
+					   u32 *in,
+					   unsigned int *group_id)
+{
+	return 0;
+}
+
+static int mlx5_cmd_stub_destroy_flow_group(struct mlx5_core_dev *dev,
+					    struct mlx5_flow_table *ft,
+					    unsigned int group_id)
+{
+	return 0;
+}
+
+static int mlx5_cmd_stub_create_fte(struct mlx5_core_dev *dev,
+				    struct mlx5_flow_table *ft,
+				    struct mlx5_flow_group *group,
+				    struct fs_fte *fte)
+{
+	return 0;
+}
+
+static int mlx5_cmd_stub_update_fte(struct mlx5_core_dev *dev,
+				    struct mlx5_flow_table *ft,
+				    unsigned int group_id,
+				    int modify_mask,
+				    struct fs_fte *fte)
+{
+	return -EOPNOTSUPP;
+}
+
+static int mlx5_cmd_stub_delete_fte(struct mlx5_core_dev *dev,
+				    struct mlx5_flow_table *ft,
+				    struct fs_fte *fte)
+{
+	return 0;
+}
+
+static int mlx5_cmd_update_root_ft(struct mlx5_core_dev *dev,
+				   struct mlx5_flow_table *ft, u32 underlay_qpn,
+				   bool disconnect)
+{
+	u32 in[MLX5_ST_SZ_DW(set_flow_table_root_in)]   = {0};
+	u32 out[MLX5_ST_SZ_DW(set_flow_table_root_out)] = {0};
+
+	if ((MLX5_CAP_GEN(dev, port_type) == MLX5_CAP_PORT_TYPE_IB) &&
+	    underlay_qpn == 0)
+		return 0;
+
+	MLX5_SET(set_flow_table_root_in, in, opcode,
+		 MLX5_CMD_OP_SET_FLOW_TABLE_ROOT);
+	MLX5_SET(set_flow_table_root_in, in, table_type, ft->type);
+
+	if (disconnect) {
+		MLX5_SET(set_flow_table_root_in, in, op_mod, 1);
+		MLX5_SET(set_flow_table_root_in, in, table_id, 0);
+	} else {
+		MLX5_SET(set_flow_table_root_in, in, op_mod, 0);
+		MLX5_SET(set_flow_table_root_in, in, table_id, ft->id);
+	}
+
+	MLX5_SET(set_flow_table_root_in, in, underlay_qpn, underlay_qpn);
+	if (ft->vport) {
+		MLX5_SET(set_flow_table_root_in, in, vport_number, ft->vport);
+		MLX5_SET(set_flow_table_root_in, in, other_vport, 1);
+	}
+
+	return mlx5_cmd_exec(dev, in, sizeof(in), out, sizeof(out));
+}
+
+static int mlx5_cmd_create_flow_table(struct mlx5_core_dev *dev,
+				      u16 vport,
+				      enum fs_flow_table_op_mod op_mod,
+				      enum fs_flow_table_type type,
+				      unsigned int level,
+				      unsigned int log_size,
+				      struct mlx5_flow_table *next_ft,
+				      unsigned int *table_id, u32 flags)
+{
+	int en_encap_decap = !!(flags & MLX5_FLOW_TABLE_TUNNEL_EN);
+	u32 out[MLX5_ST_SZ_DW(create_flow_table_out)] = {0};
+	u32 in[MLX5_ST_SZ_DW(create_flow_table_in)]   = {0};
+	int err;
+
+	MLX5_SET(create_flow_table_in, in, opcode,
+		 MLX5_CMD_OP_CREATE_FLOW_TABLE);
+
+	MLX5_SET(create_flow_table_in, in, table_type, type);
+	MLX5_SET(create_flow_table_in, in, flow_table_context.level, level);
+	MLX5_SET(create_flow_table_in, in, flow_table_context.log_size, log_size);
+	if (vport) {
+		MLX5_SET(create_flow_table_in, in, vport_number, vport);
+		MLX5_SET(create_flow_table_in, in, other_vport, 1);
+	}
+
+	MLX5_SET(create_flow_table_in, in, flow_table_context.decap_en,
+		 en_encap_decap);
+	MLX5_SET(create_flow_table_in, in, flow_table_context.encap_en,
+		 en_encap_decap);
+
+	switch (op_mod) {
+	case FS_FT_OP_MOD_NORMAL:
+		if (next_ft) {
+			MLX5_SET(create_flow_table_in, in,
+				 flow_table_context.table_miss_action, 1);
+			MLX5_SET(create_flow_table_in, in,
+				 flow_table_context.table_miss_id, next_ft->id);
+		}
+		break;
+
+	case FS_FT_OP_MOD_LAG_DEMUX:
+		MLX5_SET(create_flow_table_in, in, op_mod, 0x1);
+		if (next_ft)
+			MLX5_SET(create_flow_table_in, in,
+				 flow_table_context.lag_master_next_table_id,
+				 next_ft->id);
+		break;
+	}
+
+	err = mlx5_cmd_exec(dev, in, sizeof(in), out, sizeof(out));
+	if (!err)
+		*table_id = MLX5_GET(create_flow_table_out, out,
+				     table_id);
+	return err;
+}
+
+static int mlx5_cmd_destroy_flow_table(struct mlx5_core_dev *dev,
+				       struct mlx5_flow_table *ft)
+{
+	u32 in[MLX5_ST_SZ_DW(destroy_flow_table_in)]   = {0};
+	u32 out[MLX5_ST_SZ_DW(destroy_flow_table_out)] = {0};
+
+	MLX5_SET(destroy_flow_table_in, in, opcode,
+		 MLX5_CMD_OP_DESTROY_FLOW_TABLE);
+	MLX5_SET(destroy_flow_table_in, in, table_type, ft->type);
+	MLX5_SET(destroy_flow_table_in, in, table_id, ft->id);
+	if (ft->vport) {
+		MLX5_SET(destroy_flow_table_in, in, vport_number, ft->vport);
+		MLX5_SET(destroy_flow_table_in, in, other_vport, 1);
+	}
+
+	return mlx5_cmd_exec(dev, in, sizeof(in), out, sizeof(out));
+}
+
+static int mlx5_cmd_modify_flow_table(struct mlx5_core_dev *dev,
+				      struct mlx5_flow_table *ft,
+				      struct mlx5_flow_table *next_ft)
+{
+	u32 in[MLX5_ST_SZ_DW(modify_flow_table_in)]   = {0};
+	u32 out[MLX5_ST_SZ_DW(modify_flow_table_out)] = {0};
+
+	MLX5_SET(modify_flow_table_in, in, opcode,
+		 MLX5_CMD_OP_MODIFY_FLOW_TABLE);
+	MLX5_SET(modify_flow_table_in, in, table_type, ft->type);
+	MLX5_SET(modify_flow_table_in, in, table_id, ft->id);
+
+	if (ft->op_mod == FS_FT_OP_MOD_LAG_DEMUX) {
+		MLX5_SET(modify_flow_table_in, in, modify_field_select,
+			 MLX5_MODIFY_FLOW_TABLE_LAG_NEXT_TABLE_ID);
+		if (next_ft) {
+			MLX5_SET(modify_flow_table_in, in,
+				 flow_table_context.lag_master_next_table_id, next_ft->id);
+		} else {
+			MLX5_SET(modify_flow_table_in, in,
+				 flow_table_context.lag_master_next_table_id, 0);
+		}
+	} else {
+		if (ft->vport) {
+			MLX5_SET(modify_flow_table_in, in, vport_number,
+				 ft->vport);
+			MLX5_SET(modify_flow_table_in, in, other_vport, 1);
+		}
+		MLX5_SET(modify_flow_table_in, in, modify_field_select,
+			 MLX5_MODIFY_FLOW_TABLE_MISS_TABLE_ID);
+		if (next_ft) {
+			MLX5_SET(modify_flow_table_in, in,
+				 flow_table_context.table_miss_action, 1);
+			MLX5_SET(modify_flow_table_in, in,
+				 flow_table_context.table_miss_id,
+				 next_ft->id);
+		} else {
+			MLX5_SET(modify_flow_table_in, in,
+				 flow_table_context.table_miss_action, 0);
+		}
+	}
+
+	return mlx5_cmd_exec(dev, in, sizeof(in), out, sizeof(out));
+}
+
+static int mlx5_cmd_create_flow_group(struct mlx5_core_dev *dev,
+				      struct mlx5_flow_table *ft,
+				      u32 *in,
+				      unsigned int *group_id)
+{
+	u32 out[MLX5_ST_SZ_DW(create_flow_group_out)] = {0};
+	int inlen = MLX5_ST_SZ_BYTES(create_flow_group_in);
+	int err;
+
+	MLX5_SET(create_flow_group_in, in, opcode,
+		 MLX5_CMD_OP_CREATE_FLOW_GROUP);
+	MLX5_SET(create_flow_group_in, in, table_type, ft->type);
+	MLX5_SET(create_flow_group_in, in, table_id, ft->id);
+	if (ft->vport) {
+		MLX5_SET(create_flow_group_in, in, vport_number, ft->vport);
+		MLX5_SET(create_flow_group_in, in, other_vport, 1);
+	}
+
+	err = mlx5_cmd_exec(dev, in, inlen, out, sizeof(out));
+	if (!err)
+		*group_id = MLX5_GET(create_flow_group_out, out,
+				     group_id);
+	return err;
+}
+
+static int mlx5_cmd_destroy_flow_group(struct mlx5_core_dev *dev,
+				       struct mlx5_flow_table *ft,
+				       unsigned int group_id)
+{
+	u32 out[MLX5_ST_SZ_DW(destroy_flow_group_out)] = {0};
+	u32 in[MLX5_ST_SZ_DW(destroy_flow_group_in)]   = {0};
+
+	MLX5_SET(destroy_flow_group_in, in, opcode,
+		 MLX5_CMD_OP_DESTROY_FLOW_GROUP);
+	MLX5_SET(destroy_flow_group_in, in, table_type, ft->type);
+	MLX5_SET(destroy_flow_group_in, in, table_id, ft->id);
+	MLX5_SET(destroy_flow_group_in, in, group_id, group_id);
+	if (ft->vport) {
+		MLX5_SET(destroy_flow_group_in, in, vport_number, ft->vport);
+		MLX5_SET(destroy_flow_group_in, in, other_vport, 1);
+	}
+
+	return mlx5_cmd_exec(dev, in, sizeof(in), out, sizeof(out));
+}
+
+static int mlx5_cmd_set_fte(struct mlx5_core_dev *dev,
+			    int opmod, int modify_mask,
+			    struct mlx5_flow_table *ft,
+			    unsigned group_id,
+			    struct fs_fte *fte)
+{
+	unsigned int inlen = MLX5_ST_SZ_BYTES(set_fte_in) +
+		fte->dests_size * MLX5_ST_SZ_BYTES(dest_format_struct);
+	u32 out[MLX5_ST_SZ_DW(set_fte_out)] = {0};
+	struct mlx5_flow_rule *dst;
+	void *in_flow_context, *vlan;
+	void *in_match_value;
+	void *in_dests;
+	u32 *in;
+	int err;
+
+	in = kvzalloc(inlen, GFP_KERNEL);
+	if (!in)
+		return -ENOMEM;
+
+	MLX5_SET(set_fte_in, in, opcode, MLX5_CMD_OP_SET_FLOW_TABLE_ENTRY);
+	MLX5_SET(set_fte_in, in, op_mod, opmod);
+	MLX5_SET(set_fte_in, in, modify_enable_mask, modify_mask);
+	MLX5_SET(set_fte_in, in, table_type, ft->type);
+	MLX5_SET(set_fte_in, in, table_id,   ft->id);
+	MLX5_SET(set_fte_in, in, flow_index, fte->index);
+	if (ft->vport) {
+		MLX5_SET(set_fte_in, in, vport_number, ft->vport);
+		MLX5_SET(set_fte_in, in, other_vport, 1);
+	}
+
+	in_flow_context = MLX5_ADDR_OF(set_fte_in, in, flow_context);
+	MLX5_SET(flow_context, in_flow_context, group_id, group_id);
+
+	MLX5_SET(flow_context, in_flow_context, flow_tag, fte->action.flow_tag);
+	MLX5_SET(flow_context, in_flow_context, action, fte->action.action);
+	MLX5_SET(flow_context, in_flow_context, encap_id, fte->action.encap_id);
+	MLX5_SET(flow_context, in_flow_context, modify_header_id,
+		 fte->action.modify_id);
+
+	vlan = MLX5_ADDR_OF(flow_context, in_flow_context, push_vlan);
+
+	MLX5_SET(vlan, vlan, ethtype, fte->action.vlan[0].ethtype);
+	MLX5_SET(vlan, vlan, vid, fte->action.vlan[0].vid);
+	MLX5_SET(vlan, vlan, prio, fte->action.vlan[0].prio);
+
+	vlan = MLX5_ADDR_OF(flow_context, in_flow_context, push_vlan_2);
+
+	MLX5_SET(vlan, vlan, ethtype, fte->action.vlan[1].ethtype);
+	MLX5_SET(vlan, vlan, vid, fte->action.vlan[1].vid);
+	MLX5_SET(vlan, vlan, prio, fte->action.vlan[1].prio);
+
+	in_match_value = MLX5_ADDR_OF(flow_context, in_flow_context,
+				      match_value);
+	memcpy(in_match_value, &fte->val, sizeof(fte->val));
+
+	in_dests = MLX5_ADDR_OF(flow_context, in_flow_context, destination);
+	if (fte->action.action & MLX5_FLOW_CONTEXT_ACTION_FWD_DEST) {
+		int list_size = 0;
+
+		list_for_each_entry(dst, &fte->node.children, node.list) {
+			unsigned int id, type = dst->dest_attr.type;
+
+			if (type == MLX5_FLOW_DESTINATION_TYPE_COUNTER)
+				continue;
+
+			switch (type) {
+			case MLX5_FLOW_DESTINATION_TYPE_FLOW_TABLE_NUM:
+				id = dst->dest_attr.ft_num;
+				type = MLX5_FLOW_DESTINATION_TYPE_FLOW_TABLE;
+				break;
+			case MLX5_FLOW_DESTINATION_TYPE_FLOW_TABLE:
+				id = dst->dest_attr.ft->id;
+				break;
+			case MLX5_FLOW_DESTINATION_TYPE_VPORT:
+				id = dst->dest_attr.vport.num;
+				MLX5_SET(dest_format_struct, in_dests,
+					 destination_eswitch_owner_vhca_id_valid,
+					 dst->dest_attr.vport.vhca_id_valid);
+				MLX5_SET(dest_format_struct, in_dests,
+					 destination_eswitch_owner_vhca_id,
+					 dst->dest_attr.vport.vhca_id);
+				break;
+			default:
+				id = dst->dest_attr.tir_num;
+			}
+
+			MLX5_SET(dest_format_struct, in_dests, destination_type,
+				 type);
+			MLX5_SET(dest_format_struct, in_dests, destination_id, id);
+			in_dests += MLX5_ST_SZ_BYTES(dest_format_struct);
+			list_size++;
+		}
+
+		MLX5_SET(flow_context, in_flow_context, destination_list_size,
+			 list_size);
+	}
+
+	if (fte->action.action & MLX5_FLOW_CONTEXT_ACTION_COUNT) {
+		int max_list_size = BIT(MLX5_CAP_FLOWTABLE_TYPE(dev,
+					log_max_flow_counter,
+					ft->type));
+		int list_size = 0;
+
+		list_for_each_entry(dst, &fte->node.children, node.list) {
+			if (dst->dest_attr.type !=
+			    MLX5_FLOW_DESTINATION_TYPE_COUNTER)
+				continue;
+
+			MLX5_SET(flow_counter_list, in_dests, flow_counter_id,
+				 dst->dest_attr.counter->id);
+			in_dests += MLX5_ST_SZ_BYTES(dest_format_struct);
+			list_size++;
+		}
+		if (list_size > max_list_size) {
+			err = -EINVAL;
+			goto err_out;
+		}
+
+		MLX5_SET(flow_context, in_flow_context, flow_counter_list_size,
+			 list_size);
+	}
+
+	err = mlx5_cmd_exec(dev, in, inlen, out, sizeof(out));
+err_out:
+	kvfree(in);
+	return err;
+}
+
+static int mlx5_cmd_create_fte(struct mlx5_core_dev *dev,
+			       struct mlx5_flow_table *ft,
+			       struct mlx5_flow_group *group,
+			       struct fs_fte *fte)
+{
+	unsigned int group_id = group->id;
+
+	return mlx5_cmd_set_fte(dev, 0, 0, ft, group_id, fte);
+}
+
+static int mlx5_cmd_update_fte(struct mlx5_core_dev *dev,
+			       struct mlx5_flow_table *ft,
+			       unsigned int group_id,
+			       int modify_mask,
+			       struct fs_fte *fte)
+{
+	int opmod;
+	int atomic_mod_cap = MLX5_CAP_FLOWTABLE(dev,
+						flow_table_properties_nic_receive.
+						flow_modify_en);
+	if (!atomic_mod_cap)
+		return -EOPNOTSUPP;
+	opmod = 1;
+
+	return	mlx5_cmd_set_fte(dev, opmod, modify_mask, ft, group_id, fte);
+}
+
+static int mlx5_cmd_delete_fte(struct mlx5_core_dev *dev,
+			       struct mlx5_flow_table *ft,
+			       struct fs_fte *fte)
+{
+	u32 out[MLX5_ST_SZ_DW(delete_fte_out)] = {0};
+	u32 in[MLX5_ST_SZ_DW(delete_fte_in)]   = {0};
+
+	MLX5_SET(delete_fte_in, in, opcode, MLX5_CMD_OP_DELETE_FLOW_TABLE_ENTRY);
+	MLX5_SET(delete_fte_in, in, table_type, ft->type);
+	MLX5_SET(delete_fte_in, in, table_id, ft->id);
+	MLX5_SET(delete_fte_in, in, flow_index, fte->index);
+	if (ft->vport) {
+		MLX5_SET(delete_fte_in, in, vport_number, ft->vport);
+		MLX5_SET(delete_fte_in, in, other_vport, 1);
+	}
+
+	return mlx5_cmd_exec(dev, in, sizeof(in), out, sizeof(out));
+}
+
+int mlx5_cmd_fc_alloc(struct mlx5_core_dev *dev, u32 *id)
+{
+	u32 in[MLX5_ST_SZ_DW(alloc_flow_counter_in)]   = {0};
+	u32 out[MLX5_ST_SZ_DW(alloc_flow_counter_out)] = {0};
+	int err;
+
+	MLX5_SET(alloc_flow_counter_in, in, opcode,
+		 MLX5_CMD_OP_ALLOC_FLOW_COUNTER);
+
+	err = mlx5_cmd_exec(dev, in, sizeof(in), out, sizeof(out));
+	if (!err)
+		*id = MLX5_GET(alloc_flow_counter_out, out, flow_counter_id);
+	return err;
+}
+
+int mlx5_cmd_fc_free(struct mlx5_core_dev *dev, u32 id)
+{
+	u32 in[MLX5_ST_SZ_DW(dealloc_flow_counter_in)]   = {0};
+	u32 out[MLX5_ST_SZ_DW(dealloc_flow_counter_out)] = {0};
+
+	MLX5_SET(dealloc_flow_counter_in, in, opcode,
+		 MLX5_CMD_OP_DEALLOC_FLOW_COUNTER);
+	MLX5_SET(dealloc_flow_counter_in, in, flow_counter_id, id);
+	return mlx5_cmd_exec(dev, in, sizeof(in), out, sizeof(out));
+}
+
+int mlx5_cmd_fc_query(struct mlx5_core_dev *dev, u32 id,
+		      u64 *packets, u64 *bytes)
+{
+	u32 out[MLX5_ST_SZ_BYTES(query_flow_counter_out) +
+		MLX5_ST_SZ_BYTES(traffic_counter)]   = {0};
+	u32 in[MLX5_ST_SZ_DW(query_flow_counter_in)] = {0};
+	void *stats;
+	int err = 0;
+
+	MLX5_SET(query_flow_counter_in, in, opcode,
+		 MLX5_CMD_OP_QUERY_FLOW_COUNTER);
+	MLX5_SET(query_flow_counter_in, in, op_mod, 0);
+	MLX5_SET(query_flow_counter_in, in, flow_counter_id, id);
+	err = mlx5_cmd_exec(dev, in, sizeof(in), out, sizeof(out));
+	if (err)
+		return err;
+
+	stats = MLX5_ADDR_OF(query_flow_counter_out, out, flow_statistics);
+	*packets = MLX5_GET64(traffic_counter, stats, packets);
+	*bytes = MLX5_GET64(traffic_counter, stats, octets);
+	return 0;
+}
+
+struct mlx5_cmd_fc_bulk {
+	u32 id;
+	int num;
+	int outlen;
+	u32 out[0];
+};
+
+struct mlx5_cmd_fc_bulk *
+mlx5_cmd_fc_bulk_alloc(struct mlx5_core_dev *dev, u32 id, int num)
+{
+	struct mlx5_cmd_fc_bulk *b;
+	int outlen =
+		MLX5_ST_SZ_BYTES(query_flow_counter_out) +
+		MLX5_ST_SZ_BYTES(traffic_counter) * num;
+
+	b = kzalloc(sizeof(*b) + outlen, GFP_KERNEL);
+	if (!b)
+		return NULL;
+
+	b->id = id;
+	b->num = num;
+	b->outlen = outlen;
+
+	return b;
+}
+
+void mlx5_cmd_fc_bulk_free(struct mlx5_cmd_fc_bulk *b)
+{
+	kfree(b);
+}
+
+int
+mlx5_cmd_fc_bulk_query(struct mlx5_core_dev *dev, struct mlx5_cmd_fc_bulk *b)
+{
+	u32 in[MLX5_ST_SZ_DW(query_flow_counter_in)] = {0};
+
+	MLX5_SET(query_flow_counter_in, in, opcode,
+		 MLX5_CMD_OP_QUERY_FLOW_COUNTER);
+	MLX5_SET(query_flow_counter_in, in, op_mod, 0);
+	MLX5_SET(query_flow_counter_in, in, flow_counter_id, b->id);
+	MLX5_SET(query_flow_counter_in, in, num_of_counters, b->num);
+	return mlx5_cmd_exec(dev, in, sizeof(in), b->out, b->outlen);
+}
+
+void mlx5_cmd_fc_bulk_get(struct mlx5_core_dev *dev,
+			  struct mlx5_cmd_fc_bulk *b, u32 id,
+			  u64 *packets, u64 *bytes)
+{
+	int index = id - b->id;
+	void *stats;
+
+	if (index < 0 || index >= b->num) {
+		mlx5_core_warn(dev, "Flow counter id (0x%x) out of range (0x%x..0x%x). Counter ignored.\n",
+			       id, b->id, b->id + b->num - 1);
+		return;
+	}
+
+	stats = MLX5_ADDR_OF(query_flow_counter_out, b->out,
+			     flow_statistics[index]);
+	*packets = MLX5_GET64(traffic_counter, stats, packets);
+	*bytes = MLX5_GET64(traffic_counter, stats, octets);
+}
+
+int mlx5_encap_alloc(struct mlx5_core_dev *dev,
+		     int header_type,
+		     size_t size,
+		     void *encap_header,
+		     u32 *encap_id)
+{
+	int max_encap_size = MLX5_CAP_ESW(dev, max_encap_header_size);
+	u32 out[MLX5_ST_SZ_DW(alloc_encap_header_out)];
+	void *encap_header_in;
+	void *header;
+	int inlen;
+	int err;
+	u32 *in;
+
+	if (size > max_encap_size) {
+		mlx5_core_warn(dev, "encap size %zd too big, max supported is %d\n",
+			       size, max_encap_size);
+		return -EINVAL;
+	}
+
+	in = kzalloc(MLX5_ST_SZ_BYTES(alloc_encap_header_in) + size,
+		     GFP_KERNEL);
+	if (!in)
+		return -ENOMEM;
+
+	encap_header_in = MLX5_ADDR_OF(alloc_encap_header_in, in, encap_header);
+	header = MLX5_ADDR_OF(encap_header_in, encap_header_in, encap_header);
+	inlen = header - (void *)in  + size;
+
+	memset(in, 0, inlen);
+	MLX5_SET(alloc_encap_header_in, in, opcode,
+		 MLX5_CMD_OP_ALLOC_ENCAP_HEADER);
+	MLX5_SET(encap_header_in, encap_header_in, encap_header_size, size);
+	MLX5_SET(encap_header_in, encap_header_in, header_type, header_type);
+	memcpy(header, encap_header, size);
+
+	memset(out, 0, sizeof(out));
+	err = mlx5_cmd_exec(dev, in, inlen, out, sizeof(out));
+
+	*encap_id = MLX5_GET(alloc_encap_header_out, out, encap_id);
+	kfree(in);
+	return err;
+}
+
+void mlx5_encap_dealloc(struct mlx5_core_dev *dev, u32 encap_id)
+{
+	u32 in[MLX5_ST_SZ_DW(dealloc_encap_header_in)];
+	u32 out[MLX5_ST_SZ_DW(dealloc_encap_header_out)];
+
+	memset(in, 0, sizeof(in));
+	MLX5_SET(dealloc_encap_header_in, in, opcode,
+		 MLX5_CMD_OP_DEALLOC_ENCAP_HEADER);
+	MLX5_SET(dealloc_encap_header_in, in, encap_id, encap_id);
+
+	mlx5_cmd_exec(dev, in, sizeof(in), out, sizeof(out));
+}
+
+int mlx5_modify_header_alloc(struct mlx5_core_dev *dev,
+			     u8 namespace, u8 num_actions,
+			     void *modify_actions, u32 *modify_header_id)
+{
+	u32 out[MLX5_ST_SZ_DW(alloc_modify_header_context_out)];
+	int max_actions, actions_size, inlen, err;
+	void *actions_in;
+	u8 table_type;
+	u32 *in;
+
+	switch (namespace) {
+	case MLX5_FLOW_NAMESPACE_FDB:
+		max_actions = MLX5_CAP_ESW_FLOWTABLE_FDB(dev, max_modify_header_actions);
+		table_type = FS_FT_FDB;
+		break;
+	case MLX5_FLOW_NAMESPACE_KERNEL:
+		max_actions = MLX5_CAP_FLOWTABLE_NIC_RX(dev, max_modify_header_actions);
+		table_type = FS_FT_NIC_RX;
+		break;
+	default:
+		return -EOPNOTSUPP;
+	}
+
+	if (num_actions > max_actions) {
+		mlx5_core_warn(dev, "too many modify header actions %d, max supported %d\n",
+			       num_actions, max_actions);
+		return -EOPNOTSUPP;
+	}
+
+	actions_size = MLX5_UN_SZ_BYTES(set_action_in_add_action_in_auto) * num_actions;
+	inlen = MLX5_ST_SZ_BYTES(alloc_modify_header_context_in) + actions_size;
+
+	in = kzalloc(inlen, GFP_KERNEL);
+	if (!in)
+		return -ENOMEM;
+
+	MLX5_SET(alloc_modify_header_context_in, in, opcode,
+		 MLX5_CMD_OP_ALLOC_MODIFY_HEADER_CONTEXT);
+	MLX5_SET(alloc_modify_header_context_in, in, table_type, table_type);
+	MLX5_SET(alloc_modify_header_context_in, in, num_of_actions, num_actions);
+
+	actions_in = MLX5_ADDR_OF(alloc_modify_header_context_in, in, actions);
+	memcpy(actions_in, modify_actions, actions_size);
+
+	memset(out, 0, sizeof(out));
+	err = mlx5_cmd_exec(dev, in, inlen, out, sizeof(out));
+
+	*modify_header_id = MLX5_GET(alloc_modify_header_context_out, out, modify_header_id);
+	kfree(in);
+	return err;
+}
+
+void mlx5_modify_header_dealloc(struct mlx5_core_dev *dev, u32 modify_header_id)
+{
+	u32 in[MLX5_ST_SZ_DW(dealloc_modify_header_context_in)];
+	u32 out[MLX5_ST_SZ_DW(dealloc_modify_header_context_out)];
+
+	memset(in, 0, sizeof(in));
+	MLX5_SET(dealloc_modify_header_context_in, in, opcode,
+		 MLX5_CMD_OP_DEALLOC_MODIFY_HEADER_CONTEXT);
+	MLX5_SET(dealloc_modify_header_context_in, in, modify_header_id,
+		 modify_header_id);
+
+	mlx5_cmd_exec(dev, in, sizeof(in), out, sizeof(out));
+}
+
+static const struct mlx5_flow_cmds mlx5_flow_cmds = {
+	.create_flow_table = mlx5_cmd_create_flow_table,
+	.destroy_flow_table = mlx5_cmd_destroy_flow_table,
+	.modify_flow_table = mlx5_cmd_modify_flow_table,
+	.create_flow_group = mlx5_cmd_create_flow_group,
+	.destroy_flow_group = mlx5_cmd_destroy_flow_group,
+	.create_fte = mlx5_cmd_create_fte,
+	.update_fte = mlx5_cmd_update_fte,
+	.delete_fte = mlx5_cmd_delete_fte,
+	.update_root_ft = mlx5_cmd_update_root_ft,
+};
+
+static const struct mlx5_flow_cmds mlx5_flow_cmd_stubs = {
+	.create_flow_table = mlx5_cmd_stub_create_flow_table,
+	.destroy_flow_table = mlx5_cmd_stub_destroy_flow_table,
+	.modify_flow_table = mlx5_cmd_stub_modify_flow_table,
+	.create_flow_group = mlx5_cmd_stub_create_flow_group,
+	.destroy_flow_group = mlx5_cmd_stub_destroy_flow_group,
+	.create_fte = mlx5_cmd_stub_create_fte,
+	.update_fte = mlx5_cmd_stub_update_fte,
+	.delete_fte = mlx5_cmd_stub_delete_fte,
+	.update_root_ft = mlx5_cmd_stub_update_root_ft,
+};
+
+static const struct mlx5_flow_cmds *mlx5_fs_cmd_get_fw_cmds(void)
+{
+	return &mlx5_flow_cmds;
+}
+
+static const struct mlx5_flow_cmds *mlx5_fs_cmd_get_stub_cmds(void)
+{
+	return &mlx5_flow_cmd_stubs;
+}
+
+const struct mlx5_flow_cmds *mlx5_fs_cmd_get_default(enum fs_flow_table_type type)
+{
+	switch (type) {
+	case FS_FT_NIC_RX:
+	case FS_FT_ESW_EGRESS_ACL:
+	case FS_FT_ESW_INGRESS_ACL:
+	case FS_FT_FDB:
+	case FS_FT_SNIFFER_RX:
+	case FS_FT_SNIFFER_TX:
+		return mlx5_fs_cmd_get_fw_cmds();
+	case FS_FT_NIC_TX:
+	default:
+		return mlx5_fs_cmd_get_stub_cmds();
+	}
+}
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/fs_cmd.h b/drivers/net/ethernet/mellanox/mlx5/core/fs_cmd.h
new file mode 100644
index 000000000..6228ba7bf
--- /dev/null
+++ b/drivers/net/ethernet/mellanox/mlx5/core/fs_cmd.h
@@ -0,0 +1,101 @@
+/*
+ * Copyright (c) 2015, 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 _MLX5_FS_CMD_
+#define _MLX5_FS_CMD_
+
+#include "fs_core.h"
+
+struct mlx5_flow_cmds {
+	int (*create_flow_table)(struct mlx5_core_dev *dev,
+				 u16 vport,
+				 enum fs_flow_table_op_mod op_mod,
+				 enum fs_flow_table_type type,
+				 unsigned int level, unsigned int log_size,
+				 struct mlx5_flow_table *next_ft,
+				 unsigned int *table_id, u32 flags);
+	int (*destroy_flow_table)(struct mlx5_core_dev *dev,
+				  struct mlx5_flow_table *ft);
+
+	int (*modify_flow_table)(struct mlx5_core_dev *dev,
+				 struct mlx5_flow_table *ft,
+				 struct mlx5_flow_table *next_ft);
+
+	int (*create_flow_group)(struct mlx5_core_dev *dev,
+				 struct mlx5_flow_table *ft,
+				 u32 *in,
+				 unsigned int *group_id);
+
+	int (*destroy_flow_group)(struct mlx5_core_dev *dev,
+				  struct mlx5_flow_table *ft,
+				  unsigned int group_id);
+
+	int (*create_fte)(struct mlx5_core_dev *dev,
+			  struct mlx5_flow_table *ft,
+			  struct mlx5_flow_group *fg,
+			  struct fs_fte *fte);
+
+	int (*update_fte)(struct mlx5_core_dev *dev,
+			  struct mlx5_flow_table *ft,
+			  unsigned int group_id,
+			  int modify_mask,
+			  struct fs_fte *fte);
+
+	int (*delete_fte)(struct mlx5_core_dev *dev,
+			  struct mlx5_flow_table *ft,
+			  struct fs_fte *fte);
+
+	int (*update_root_ft)(struct mlx5_core_dev *dev,
+			      struct mlx5_flow_table *ft,
+			      u32 underlay_qpn,
+			      bool disconnect);
+};
+
+int mlx5_cmd_fc_alloc(struct mlx5_core_dev *dev, u32 *id);
+int mlx5_cmd_fc_free(struct mlx5_core_dev *dev, u32 id);
+int mlx5_cmd_fc_query(struct mlx5_core_dev *dev, u32 id,
+		      u64 *packets, u64 *bytes);
+
+struct mlx5_cmd_fc_bulk;
+
+struct mlx5_cmd_fc_bulk *
+mlx5_cmd_fc_bulk_alloc(struct mlx5_core_dev *dev, u32 id, int num);
+void mlx5_cmd_fc_bulk_free(struct mlx5_cmd_fc_bulk *b);
+int
+mlx5_cmd_fc_bulk_query(struct mlx5_core_dev *dev, struct mlx5_cmd_fc_bulk *b);
+void mlx5_cmd_fc_bulk_get(struct mlx5_core_dev *dev,
+			  struct mlx5_cmd_fc_bulk *b, u32 id,
+			  u64 *packets, u64 *bytes);
+
+const struct mlx5_flow_cmds *mlx5_fs_cmd_get_default(enum fs_flow_table_type type);
+
+#endif
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/fs_core.c b/drivers/net/ethernet/mellanox/mlx5/core/fs_core.c
new file mode 100644
index 000000000..f0aa7f0e5
--- /dev/null
+++ b/drivers/net/ethernet/mellanox/mlx5/core/fs_core.c
@@ -0,0 +1,2720 @@
+/*
+ * Copyright (c) 2015, 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/mutex.h>
+#include <linux/mlx5/driver.h>
+#include <linux/mlx5/eswitch.h>
+
+#include "mlx5_core.h"
+#include "fs_core.h"
+#include "fs_cmd.h"
+#include "diag/fs_tracepoint.h"
+#include "accel/ipsec.h"
+#include "fpga/ipsec.h"
+
+#define INIT_TREE_NODE_ARRAY_SIZE(...)	(sizeof((struct init_tree_node[]){__VA_ARGS__}) /\
+					 sizeof(struct init_tree_node))
+
+#define ADD_PRIO(num_prios_val, min_level_val, num_levels_val, caps_val,\
+		 ...) {.type = FS_TYPE_PRIO,\
+	.min_ft_level = min_level_val,\
+	.num_levels = num_levels_val,\
+	.num_leaf_prios = num_prios_val,\
+	.caps = caps_val,\
+	.children = (struct init_tree_node[]) {__VA_ARGS__},\
+	.ar_size = INIT_TREE_NODE_ARRAY_SIZE(__VA_ARGS__) \
+}
+
+#define ADD_MULTIPLE_PRIO(num_prios_val, num_levels_val, ...)\
+	ADD_PRIO(num_prios_val, 0, num_levels_val, {},\
+		 __VA_ARGS__)\
+
+#define ADD_NS(...) {.type = FS_TYPE_NAMESPACE,\
+	.children = (struct init_tree_node[]) {__VA_ARGS__},\
+	.ar_size = INIT_TREE_NODE_ARRAY_SIZE(__VA_ARGS__) \
+}
+
+#define INIT_CAPS_ARRAY_SIZE(...) (sizeof((long[]){__VA_ARGS__}) /\
+				   sizeof(long))
+
+#define FS_CAP(cap) (__mlx5_bit_off(flow_table_nic_cap, cap))
+
+#define FS_REQUIRED_CAPS(...) {.arr_sz = INIT_CAPS_ARRAY_SIZE(__VA_ARGS__), \
+			       .caps = (long[]) {__VA_ARGS__} }
+
+#define FS_CHAINING_CAPS  FS_REQUIRED_CAPS(FS_CAP(flow_table_properties_nic_receive.flow_modify_en), \
+					   FS_CAP(flow_table_properties_nic_receive.modify_root), \
+					   FS_CAP(flow_table_properties_nic_receive.identified_miss_table_mode), \
+					   FS_CAP(flow_table_properties_nic_receive.flow_table_modify))
+
+#define LEFTOVERS_NUM_LEVELS 1
+#define LEFTOVERS_NUM_PRIOS 1
+
+#define BY_PASS_PRIO_NUM_LEVELS 1
+#define BY_PASS_MIN_LEVEL (ETHTOOL_MIN_LEVEL + MLX5_BY_PASS_NUM_PRIOS +\
+			   LEFTOVERS_NUM_PRIOS)
+
+#define ETHTOOL_PRIO_NUM_LEVELS 1
+#define ETHTOOL_NUM_PRIOS 11
+#define ETHTOOL_MIN_LEVEL (KERNEL_MIN_LEVEL + ETHTOOL_NUM_PRIOS)
+/* Vlan, mac, ttc, inner ttc, aRFS */
+#define KERNEL_NIC_PRIO_NUM_LEVELS 5
+#define KERNEL_NIC_NUM_PRIOS 1
+/* One more level for tc */
+#define KERNEL_MIN_LEVEL (KERNEL_NIC_PRIO_NUM_LEVELS + 1)
+
+#define KERNEL_NIC_TC_NUM_PRIOS  1
+#define KERNEL_NIC_TC_NUM_LEVELS 2
+
+#define ANCHOR_NUM_LEVELS 1
+#define ANCHOR_NUM_PRIOS 1
+#define ANCHOR_MIN_LEVEL (BY_PASS_MIN_LEVEL + 1)
+
+#define OFFLOADS_MAX_FT 1
+#define OFFLOADS_NUM_PRIOS 1
+#define OFFLOADS_MIN_LEVEL (ANCHOR_MIN_LEVEL + 1)
+
+#define LAG_PRIO_NUM_LEVELS 1
+#define LAG_NUM_PRIOS 1
+#define LAG_MIN_LEVEL (OFFLOADS_MIN_LEVEL + 1)
+
+struct node_caps {
+	size_t	arr_sz;
+	long	*caps;
+};
+
+static struct init_tree_node {
+	enum fs_node_type	type;
+	struct init_tree_node *children;
+	int ar_size;
+	struct node_caps caps;
+	int min_ft_level;
+	int num_leaf_prios;
+	int prio;
+	int num_levels;
+} root_fs = {
+	.type = FS_TYPE_NAMESPACE,
+	.ar_size = 7,
+	.children = (struct init_tree_node[]) {
+		ADD_PRIO(0, BY_PASS_MIN_LEVEL, 0,
+			 FS_CHAINING_CAPS,
+			 ADD_NS(ADD_MULTIPLE_PRIO(MLX5_BY_PASS_NUM_PRIOS,
+						  BY_PASS_PRIO_NUM_LEVELS))),
+		ADD_PRIO(0, LAG_MIN_LEVEL, 0,
+			 FS_CHAINING_CAPS,
+			 ADD_NS(ADD_MULTIPLE_PRIO(LAG_NUM_PRIOS,
+						  LAG_PRIO_NUM_LEVELS))),
+		ADD_PRIO(0, OFFLOADS_MIN_LEVEL, 0, {},
+			 ADD_NS(ADD_MULTIPLE_PRIO(OFFLOADS_NUM_PRIOS, OFFLOADS_MAX_FT))),
+		ADD_PRIO(0, ETHTOOL_MIN_LEVEL, 0,
+			 FS_CHAINING_CAPS,
+			 ADD_NS(ADD_MULTIPLE_PRIO(ETHTOOL_NUM_PRIOS,
+						  ETHTOOL_PRIO_NUM_LEVELS))),
+		ADD_PRIO(0, KERNEL_MIN_LEVEL, 0, {},
+			 ADD_NS(ADD_MULTIPLE_PRIO(KERNEL_NIC_TC_NUM_PRIOS, KERNEL_NIC_TC_NUM_LEVELS),
+				ADD_MULTIPLE_PRIO(KERNEL_NIC_NUM_PRIOS,
+						  KERNEL_NIC_PRIO_NUM_LEVELS))),
+		ADD_PRIO(0, BY_PASS_MIN_LEVEL, 0,
+			 FS_CHAINING_CAPS,
+			 ADD_NS(ADD_MULTIPLE_PRIO(LEFTOVERS_NUM_PRIOS, LEFTOVERS_NUM_LEVELS))),
+		ADD_PRIO(0, ANCHOR_MIN_LEVEL, 0, {},
+			 ADD_NS(ADD_MULTIPLE_PRIO(ANCHOR_NUM_PRIOS, ANCHOR_NUM_LEVELS))),
+	}
+};
+
+enum fs_i_lock_class {
+	FS_LOCK_GRANDPARENT,
+	FS_LOCK_PARENT,
+	FS_LOCK_CHILD
+};
+
+static const struct rhashtable_params rhash_fte = {
+	.key_len = FIELD_SIZEOF(struct fs_fte, val),
+	.key_offset = offsetof(struct fs_fte, val),
+	.head_offset = offsetof(struct fs_fte, hash),
+	.automatic_shrinking = true,
+	.min_size = 1,
+};
+
+static const struct rhashtable_params rhash_fg = {
+	.key_len = FIELD_SIZEOF(struct mlx5_flow_group, mask),
+	.key_offset = offsetof(struct mlx5_flow_group, mask),
+	.head_offset = offsetof(struct mlx5_flow_group, hash),
+	.automatic_shrinking = true,
+	.min_size = 1,
+
+};
+
+static void del_hw_flow_table(struct fs_node *node);
+static void del_hw_flow_group(struct fs_node *node);
+static void del_hw_fte(struct fs_node *node);
+static void del_sw_flow_table(struct fs_node *node);
+static void del_sw_flow_group(struct fs_node *node);
+static void del_sw_fte(struct fs_node *node);
+static void del_sw_prio(struct fs_node *node);
+static void del_sw_ns(struct fs_node *node);
+/* Delete rule (destination) is special case that 
+ * requires to lock the FTE for all the deletion process.
+ */
+static void del_sw_hw_rule(struct fs_node *node);
+static bool mlx5_flow_dests_cmp(struct mlx5_flow_destination *d1,
+				struct mlx5_flow_destination *d2);
+static void cleanup_root_ns(struct mlx5_flow_root_namespace *root_ns);
+static struct mlx5_flow_rule *
+find_flow_rule(struct fs_fte *fte,
+	       struct mlx5_flow_destination *dest);
+
+static void tree_init_node(struct fs_node *node,
+			   void (*del_hw_func)(struct fs_node *),
+			   void (*del_sw_func)(struct fs_node *))
+{
+	refcount_set(&node->refcount, 1);
+	INIT_LIST_HEAD(&node->list);
+	INIT_LIST_HEAD(&node->children);
+	init_rwsem(&node->lock);
+	node->del_hw_func = del_hw_func;
+	node->del_sw_func = del_sw_func;
+	node->active = false;
+}
+
+static void tree_add_node(struct fs_node *node, struct fs_node *parent)
+{
+	if (parent)
+		refcount_inc(&parent->refcount);
+	node->parent = parent;
+
+	/* Parent is the root */
+	if (!parent)
+		node->root = node;
+	else
+		node->root = parent->root;
+}
+
+static int tree_get_node(struct fs_node *node)
+{
+	return refcount_inc_not_zero(&node->refcount);
+}
+
+static void nested_down_read_ref_node(struct fs_node *node,
+				      enum fs_i_lock_class class)
+{
+	if (node) {
+		down_read_nested(&node->lock, class);
+		refcount_inc(&node->refcount);
+	}
+}
+
+static void nested_down_write_ref_node(struct fs_node *node,
+				       enum fs_i_lock_class class)
+{
+	if (node) {
+		down_write_nested(&node->lock, class);
+		refcount_inc(&node->refcount);
+	}
+}
+
+static void down_write_ref_node(struct fs_node *node)
+{
+	if (node) {
+		down_write(&node->lock);
+		refcount_inc(&node->refcount);
+	}
+}
+
+static void up_read_ref_node(struct fs_node *node)
+{
+	refcount_dec(&node->refcount);
+	up_read(&node->lock);
+}
+
+static void up_write_ref_node(struct fs_node *node)
+{
+	refcount_dec(&node->refcount);
+	up_write(&node->lock);
+}
+
+static void tree_put_node(struct fs_node *node)
+{
+	struct fs_node *parent_node = node->parent;
+
+	if (refcount_dec_and_test(&node->refcount)) {
+		if (node->del_hw_func)
+			node->del_hw_func(node);
+		if (parent_node) {
+			/* Only root namespace doesn't have parent and we just
+			 * need to free its node.
+			 */
+			down_write_ref_node(parent_node);
+			list_del_init(&node->list);
+			if (node->del_sw_func)
+				node->del_sw_func(node);
+			up_write_ref_node(parent_node);
+		} else {
+			kfree(node);
+		}
+		node = NULL;
+	}
+	if (!node && parent_node)
+		tree_put_node(parent_node);
+}
+
+static int tree_remove_node(struct fs_node *node)
+{
+	if (refcount_read(&node->refcount) > 1) {
+		refcount_dec(&node->refcount);
+		return -EEXIST;
+	}
+	tree_put_node(node);
+	return 0;
+}
+
+static struct fs_prio *find_prio(struct mlx5_flow_namespace *ns,
+				 unsigned int prio)
+{
+	struct fs_prio *iter_prio;
+
+	fs_for_each_prio(iter_prio, ns) {
+		if (iter_prio->prio == prio)
+			return iter_prio;
+	}
+
+	return NULL;
+}
+
+static bool check_valid_spec(const struct mlx5_flow_spec *spec)
+{
+	int i;
+
+	for (i = 0; i < MLX5_ST_SZ_DW_MATCH_PARAM; i++)
+		if (spec->match_value[i] & ~spec->match_criteria[i]) {
+			pr_warn("mlx5_core: match_value differs from match_criteria\n");
+			return false;
+		}
+
+	return true;
+}
+
+static struct mlx5_flow_root_namespace *find_root(struct fs_node *node)
+{
+	struct fs_node *root;
+	struct mlx5_flow_namespace *ns;
+
+	root = node->root;
+
+	if (WARN_ON(root->type != FS_TYPE_NAMESPACE)) {
+		pr_warn("mlx5: flow steering node is not in tree or garbaged\n");
+		return NULL;
+	}
+
+	ns = container_of(root, struct mlx5_flow_namespace, node);
+	return container_of(ns, struct mlx5_flow_root_namespace, ns);
+}
+
+static inline struct mlx5_flow_steering *get_steering(struct fs_node *node)
+{
+	struct mlx5_flow_root_namespace *root = find_root(node);
+
+	if (root)
+		return root->dev->priv.steering;
+	return NULL;
+}
+
+static inline struct mlx5_core_dev *get_dev(struct fs_node *node)
+{
+	struct mlx5_flow_root_namespace *root = find_root(node);
+
+	if (root)
+		return root->dev;
+	return NULL;
+}
+
+static void del_sw_ns(struct fs_node *node)
+{
+	kfree(node);
+}
+
+static void del_sw_prio(struct fs_node *node)
+{
+	kfree(node);
+}
+
+static void del_hw_flow_table(struct fs_node *node)
+{
+	struct mlx5_flow_root_namespace *root;
+	struct mlx5_flow_table *ft;
+	struct mlx5_core_dev *dev;
+	int err;
+
+	fs_get_obj(ft, node);
+	dev = get_dev(&ft->node);
+	root = find_root(&ft->node);
+
+	if (node->active) {
+		err = root->cmds->destroy_flow_table(dev, ft);
+		if (err)
+			mlx5_core_warn(dev, "flow steering can't destroy ft\n");
+	}
+}
+
+static void del_sw_flow_table(struct fs_node *node)
+{
+	struct mlx5_flow_table *ft;
+	struct fs_prio *prio;
+
+	fs_get_obj(ft, node);
+
+	rhltable_destroy(&ft->fgs_hash);
+	fs_get_obj(prio, ft->node.parent);
+	prio->num_ft--;
+	kfree(ft);
+}
+
+static void del_sw_hw_rule(struct fs_node *node)
+{
+	struct mlx5_flow_root_namespace *root;
+	struct mlx5_flow_rule *rule;
+	struct mlx5_flow_table *ft;
+	struct mlx5_flow_group *fg;
+	struct fs_fte *fte;
+	int modify_mask;
+	struct mlx5_core_dev *dev = get_dev(node);
+	int err;
+	bool update_fte = false;
+
+	fs_get_obj(rule, node);
+	fs_get_obj(fte, rule->node.parent);
+	fs_get_obj(fg, fte->node.parent);
+	fs_get_obj(ft, fg->node.parent);
+	trace_mlx5_fs_del_rule(rule);
+	if (rule->sw_action == MLX5_FLOW_CONTEXT_ACTION_FWD_NEXT_PRIO) {
+		mutex_lock(&rule->dest_attr.ft->lock);
+		list_del(&rule->next_ft);
+		mutex_unlock(&rule->dest_attr.ft->lock);
+	}
+
+	if (rule->dest_attr.type == MLX5_FLOW_DESTINATION_TYPE_COUNTER  &&
+	    --fte->dests_size) {
+		modify_mask = BIT(MLX5_SET_FTE_MODIFY_ENABLE_MASK_ACTION) |
+			      BIT(MLX5_SET_FTE_MODIFY_ENABLE_MASK_FLOW_COUNTERS);
+		fte->action.action &= ~MLX5_FLOW_CONTEXT_ACTION_COUNT;
+		update_fte = true;
+		goto out;
+	}
+
+	if ((fte->action.action & MLX5_FLOW_CONTEXT_ACTION_FWD_DEST) &&
+	    --fte->dests_size) {
+		modify_mask = BIT(MLX5_SET_FTE_MODIFY_ENABLE_MASK_DESTINATION_LIST);
+		update_fte = true;
+	}
+out:
+	root = find_root(&ft->node);
+	if (update_fte && fte->dests_size) {
+		err = root->cmds->update_fte(dev, ft, fg->id, modify_mask, fte);
+		if (err)
+			mlx5_core_warn(dev,
+				       "%s can't del rule fg id=%d fte_index=%d\n",
+				       __func__, fg->id, fte->index);
+	}
+	kfree(rule);
+}
+
+static void del_hw_fte(struct fs_node *node)
+{
+	struct mlx5_flow_root_namespace *root;
+	struct mlx5_flow_table *ft;
+	struct mlx5_flow_group *fg;
+	struct mlx5_core_dev *dev;
+	struct fs_fte *fte;
+	int err;
+
+	fs_get_obj(fte, node);
+	fs_get_obj(fg, fte->node.parent);
+	fs_get_obj(ft, fg->node.parent);
+
+	trace_mlx5_fs_del_fte(fte);
+	dev = get_dev(&ft->node);
+	root = find_root(&ft->node);
+	if (node->active) {
+		err = root->cmds->delete_fte(dev, ft, fte);
+		if (err)
+			mlx5_core_warn(dev,
+				       "flow steering can't delete fte in index %d of flow group id %d\n",
+				       fte->index, fg->id);
+		node->active = 0;
+	}
+}
+
+static void del_sw_fte(struct fs_node *node)
+{
+	struct mlx5_flow_steering *steering = get_steering(node);
+	struct mlx5_flow_group *fg;
+	struct fs_fte *fte;
+	int err;
+
+	fs_get_obj(fte, node);
+	fs_get_obj(fg, fte->node.parent);
+
+	err = rhashtable_remove_fast(&fg->ftes_hash,
+				     &fte->hash,
+				     rhash_fte);
+	WARN_ON(err);
+	ida_simple_remove(&fg->fte_allocator, fte->index - fg->start_index);
+	kmem_cache_free(steering->ftes_cache, fte);
+}
+
+static void del_hw_flow_group(struct fs_node *node)
+{
+	struct mlx5_flow_root_namespace *root;
+	struct mlx5_flow_group *fg;
+	struct mlx5_flow_table *ft;
+	struct mlx5_core_dev *dev;
+
+	fs_get_obj(fg, node);
+	fs_get_obj(ft, fg->node.parent);
+	dev = get_dev(&ft->node);
+	trace_mlx5_fs_del_fg(fg);
+
+	root = find_root(&ft->node);
+	if (fg->node.active && root->cmds->destroy_flow_group(dev, ft, fg->id))
+		mlx5_core_warn(dev, "flow steering can't destroy fg %d of ft %d\n",
+			       fg->id, ft->id);
+}
+
+static void del_sw_flow_group(struct fs_node *node)
+{
+	struct mlx5_flow_steering *steering = get_steering(node);
+	struct mlx5_flow_group *fg;
+	struct mlx5_flow_table *ft;
+	int err;
+
+	fs_get_obj(fg, node);
+	fs_get_obj(ft, fg->node.parent);
+
+	rhashtable_destroy(&fg->ftes_hash);
+	ida_destroy(&fg->fte_allocator);
+	if (ft->autogroup.active && fg->max_ftes == ft->autogroup.group_size)
+		ft->autogroup.num_groups--;
+	err = rhltable_remove(&ft->fgs_hash,
+			      &fg->hash,
+			      rhash_fg);
+	WARN_ON(err);
+	kmem_cache_free(steering->fgs_cache, fg);
+}
+
+static int insert_fte(struct mlx5_flow_group *fg, struct fs_fte *fte)
+{
+	int index;
+	int ret;
+
+	index = ida_simple_get(&fg->fte_allocator, 0, fg->max_ftes, GFP_KERNEL);
+	if (index < 0)
+		return index;
+
+	fte->index = index + fg->start_index;
+	ret = rhashtable_insert_fast(&fg->ftes_hash,
+				     &fte->hash,
+				     rhash_fte);
+	if (ret)
+		goto err_ida_remove;
+
+	tree_add_node(&fte->node, &fg->node);
+	list_add_tail(&fte->node.list, &fg->node.children);
+	return 0;
+
+err_ida_remove:
+	ida_simple_remove(&fg->fte_allocator, index);
+	return ret;
+}
+
+static struct fs_fte *alloc_fte(struct mlx5_flow_table *ft,
+				u32 *match_value,
+				struct mlx5_flow_act *flow_act)
+{
+	struct mlx5_flow_steering *steering = get_steering(&ft->node);
+	struct fs_fte *fte;
+
+	fte = kmem_cache_zalloc(steering->ftes_cache, GFP_KERNEL);
+	if (!fte)
+		return ERR_PTR(-ENOMEM);
+
+	memcpy(fte->val, match_value, sizeof(fte->val));
+	fte->node.type =  FS_TYPE_FLOW_ENTRY;
+	fte->action = *flow_act;
+
+	tree_init_node(&fte->node, del_hw_fte, del_sw_fte);
+
+	return fte;
+}
+
+static void dealloc_flow_group(struct mlx5_flow_steering *steering,
+			       struct mlx5_flow_group *fg)
+{
+	rhashtable_destroy(&fg->ftes_hash);
+	kmem_cache_free(steering->fgs_cache, fg);
+}
+
+static struct mlx5_flow_group *alloc_flow_group(struct mlx5_flow_steering *steering,
+						u8 match_criteria_enable,
+						void *match_criteria,
+						int start_index,
+						int end_index)
+{
+	struct mlx5_flow_group *fg;
+	int ret;
+
+	fg = kmem_cache_zalloc(steering->fgs_cache, GFP_KERNEL);
+	if (!fg)
+		return ERR_PTR(-ENOMEM);
+
+	ret = rhashtable_init(&fg->ftes_hash, &rhash_fte);
+	if (ret) {
+		kmem_cache_free(steering->fgs_cache, fg);
+		return ERR_PTR(ret);
+}
+	ida_init(&fg->fte_allocator);
+	fg->mask.match_criteria_enable = match_criteria_enable;
+	memcpy(&fg->mask.match_criteria, match_criteria,
+	       sizeof(fg->mask.match_criteria));
+	fg->node.type =  FS_TYPE_FLOW_GROUP;
+	fg->start_index = start_index;
+	fg->max_ftes = end_index - start_index + 1;
+
+	return fg;
+}
+
+static struct mlx5_flow_group *alloc_insert_flow_group(struct mlx5_flow_table *ft,
+						       u8 match_criteria_enable,
+						       void *match_criteria,
+						       int start_index,
+						       int end_index,
+						       struct list_head *prev)
+{
+	struct mlx5_flow_steering *steering = get_steering(&ft->node);
+	struct mlx5_flow_group *fg;
+	int ret;
+
+	fg = alloc_flow_group(steering, match_criteria_enable, match_criteria,
+			      start_index, end_index);
+	if (IS_ERR(fg))
+		return fg;
+
+	/* initialize refcnt, add to parent list */
+	ret = rhltable_insert(&ft->fgs_hash,
+			      &fg->hash,
+			      rhash_fg);
+	if (ret) {
+		dealloc_flow_group(steering, fg);
+		return ERR_PTR(ret);
+	}
+
+	tree_init_node(&fg->node, del_hw_flow_group, del_sw_flow_group);
+	tree_add_node(&fg->node, &ft->node);
+	/* Add node to group list */
+	list_add(&fg->node.list, prev);
+	atomic_inc(&ft->node.version);
+
+	return fg;
+}
+
+static struct mlx5_flow_table *alloc_flow_table(int level, u16 vport, int max_fte,
+						enum fs_flow_table_type table_type,
+						enum fs_flow_table_op_mod op_mod,
+						u32 flags)
+{
+	struct mlx5_flow_table *ft;
+	int ret;
+
+	ft  = kzalloc(sizeof(*ft), GFP_KERNEL);
+	if (!ft)
+		return ERR_PTR(-ENOMEM);
+
+	ret = rhltable_init(&ft->fgs_hash, &rhash_fg);
+	if (ret) {
+		kfree(ft);
+		return ERR_PTR(ret);
+	}
+
+	ft->level = level;
+	ft->node.type = FS_TYPE_FLOW_TABLE;
+	ft->op_mod = op_mod;
+	ft->type = table_type;
+	ft->vport = vport;
+	ft->max_fte = max_fte;
+	ft->flags = flags;
+	INIT_LIST_HEAD(&ft->fwd_rules);
+	mutex_init(&ft->lock);
+
+	return ft;
+}
+
+/* If reverse is false, then we search for the first flow table in the
+ * root sub-tree from start(closest from right), else we search for the
+ * last flow table in the root sub-tree till start(closest from left).
+ */
+static struct mlx5_flow_table *find_closest_ft_recursive(struct fs_node  *root,
+							 struct list_head *start,
+							 bool reverse)
+{
+#define list_advance_entry(pos, reverse)		\
+	((reverse) ? list_prev_entry(pos, list) : list_next_entry(pos, list))
+
+#define list_for_each_advance_continue(pos, head, reverse)	\
+	for (pos = list_advance_entry(pos, reverse);		\
+	     &pos->list != (head);				\
+	     pos = list_advance_entry(pos, reverse))
+
+	struct fs_node *iter = list_entry(start, struct fs_node, list);
+	struct mlx5_flow_table *ft = NULL;
+
+	if (!root)
+		return NULL;
+
+	list_for_each_advance_continue(iter, &root->children, reverse) {
+		if (iter->type == FS_TYPE_FLOW_TABLE) {
+			fs_get_obj(ft, iter);
+			return ft;
+		}
+		ft = find_closest_ft_recursive(iter, &iter->children, reverse);
+		if (ft)
+			return ft;
+	}
+
+	return ft;
+}
+
+/* If reverse if false then return the first flow table in next priority of
+ * prio in the tree, else return the last flow table in the previous priority
+ * of prio in the tree.
+ */
+static struct mlx5_flow_table *find_closest_ft(struct fs_prio *prio, bool reverse)
+{
+	struct mlx5_flow_table *ft = NULL;
+	struct fs_node *curr_node;
+	struct fs_node *parent;
+
+	parent = prio->node.parent;
+	curr_node = &prio->node;
+	while (!ft && parent) {
+		ft = find_closest_ft_recursive(parent, &curr_node->list, reverse);
+		curr_node = parent;
+		parent = curr_node->parent;
+	}
+	return ft;
+}
+
+/* Assuming all the tree is locked by mutex chain lock */
+static struct mlx5_flow_table *find_next_chained_ft(struct fs_prio *prio)
+{
+	return find_closest_ft(prio, false);
+}
+
+/* Assuming all the tree is locked by mutex chain lock */
+static struct mlx5_flow_table *find_prev_chained_ft(struct fs_prio *prio)
+{
+	return find_closest_ft(prio, true);
+}
+
+static int connect_fts_in_prio(struct mlx5_core_dev *dev,
+			       struct fs_prio *prio,
+			       struct mlx5_flow_table *ft)
+{
+	struct mlx5_flow_root_namespace *root = find_root(&prio->node);
+	struct mlx5_flow_table *iter;
+	int i = 0;
+	int err;
+
+	fs_for_each_ft(iter, prio) {
+		i++;
+		err = root->cmds->modify_flow_table(dev, iter, ft);
+		if (err) {
+			mlx5_core_warn(dev, "Failed to modify flow table %d\n",
+				       iter->id);
+			/* The driver is out of sync with the FW */
+			if (i > 1)
+				WARN_ON(true);
+			return err;
+		}
+	}
+	return 0;
+}
+
+/* Connect flow tables from previous priority of prio to ft */
+static int connect_prev_fts(struct mlx5_core_dev *dev,
+			    struct mlx5_flow_table *ft,
+			    struct fs_prio *prio)
+{
+	struct mlx5_flow_table *prev_ft;
+
+	prev_ft = find_prev_chained_ft(prio);
+	if (prev_ft) {
+		struct fs_prio *prev_prio;
+
+		fs_get_obj(prev_prio, prev_ft->node.parent);
+		return connect_fts_in_prio(dev, prev_prio, ft);
+	}
+	return 0;
+}
+
+static int update_root_ft_create(struct mlx5_flow_table *ft, struct fs_prio
+				 *prio)
+{
+	struct mlx5_flow_root_namespace *root = find_root(&prio->node);
+	struct mlx5_ft_underlay_qp *uqp;
+	int min_level = INT_MAX;
+	int err;
+	u32 qpn;
+
+	if (root->root_ft)
+		min_level = root->root_ft->level;
+
+	if (ft->level >= min_level)
+		return 0;
+
+	if (list_empty(&root->underlay_qpns)) {
+		/* Don't set any QPN (zero) in case QPN list is empty */
+		qpn = 0;
+		err = root->cmds->update_root_ft(root->dev, ft, qpn, false);
+	} else {
+		list_for_each_entry(uqp, &root->underlay_qpns, list) {
+			qpn = uqp->qpn;
+			err = root->cmds->update_root_ft(root->dev, ft,
+							 qpn, false);
+			if (err)
+				break;
+		}
+	}
+
+	if (err)
+		mlx5_core_warn(root->dev,
+			       "Update root flow table of id(%u) qpn(%d) failed\n",
+			       ft->id, qpn);
+	else
+		root->root_ft = ft;
+
+	return err;
+}
+
+static int _mlx5_modify_rule_destination(struct mlx5_flow_rule *rule,
+					 struct mlx5_flow_destination *dest)
+{
+	struct mlx5_flow_root_namespace *root;
+	struct mlx5_flow_table *ft;
+	struct mlx5_flow_group *fg;
+	struct fs_fte *fte;
+	int modify_mask = BIT(MLX5_SET_FTE_MODIFY_ENABLE_MASK_DESTINATION_LIST);
+	int err = 0;
+
+	fs_get_obj(fte, rule->node.parent);
+	if (!(fte->action.action & MLX5_FLOW_CONTEXT_ACTION_FWD_DEST))
+		return -EINVAL;
+	down_write_ref_node(&fte->node);
+	fs_get_obj(fg, fte->node.parent);
+	fs_get_obj(ft, fg->node.parent);
+
+	memcpy(&rule->dest_attr, dest, sizeof(*dest));
+	root = find_root(&ft->node);
+	err = root->cmds->update_fte(get_dev(&ft->node), ft, fg->id,
+				     modify_mask, fte);
+	up_write_ref_node(&fte->node);
+
+	return err;
+}
+
+int mlx5_modify_rule_destination(struct mlx5_flow_handle *handle,
+				 struct mlx5_flow_destination *new_dest,
+				 struct mlx5_flow_destination *old_dest)
+{
+	int i;
+
+	if (!old_dest) {
+		if (handle->num_rules != 1)
+			return -EINVAL;
+		return _mlx5_modify_rule_destination(handle->rule[0],
+						     new_dest);
+	}
+
+	for (i = 0; i < handle->num_rules; i++) {
+		if (mlx5_flow_dests_cmp(new_dest, &handle->rule[i]->dest_attr))
+			return _mlx5_modify_rule_destination(handle->rule[i],
+							     new_dest);
+	}
+
+	return -EINVAL;
+}
+
+/* Modify/set FWD rules that point on old_next_ft to point on new_next_ft  */
+static int connect_fwd_rules(struct mlx5_core_dev *dev,
+			     struct mlx5_flow_table *new_next_ft,
+			     struct mlx5_flow_table *old_next_ft)
+{
+	struct mlx5_flow_destination dest = {};
+	struct mlx5_flow_rule *iter;
+	int err = 0;
+
+	/* new_next_ft and old_next_ft could be NULL only
+	 * when we create/destroy the anchor flow table.
+	 */
+	if (!new_next_ft || !old_next_ft)
+		return 0;
+
+	dest.type = MLX5_FLOW_DESTINATION_TYPE_FLOW_TABLE;
+	dest.ft = new_next_ft;
+
+	mutex_lock(&old_next_ft->lock);
+	list_splice_init(&old_next_ft->fwd_rules, &new_next_ft->fwd_rules);
+	mutex_unlock(&old_next_ft->lock);
+	list_for_each_entry(iter, &new_next_ft->fwd_rules, next_ft) {
+		err = _mlx5_modify_rule_destination(iter, &dest);
+		if (err)
+			pr_err("mlx5_core: failed to modify rule to point on flow table %d\n",
+			       new_next_ft->id);
+	}
+	return 0;
+}
+
+static int connect_flow_table(struct mlx5_core_dev *dev, struct mlx5_flow_table *ft,
+			      struct fs_prio *prio)
+{
+	struct mlx5_flow_table *next_ft, *first_ft;
+	int err = 0;
+
+	/* Connect_prev_fts and update_root_ft_create are mutually exclusive */
+
+	first_ft = list_first_entry_or_null(&prio->node.children,
+					    struct mlx5_flow_table, node.list);
+	if (!first_ft || first_ft->level > ft->level) {
+		err = connect_prev_fts(dev, ft, prio);
+		if (err)
+			return err;
+
+		next_ft = first_ft ? first_ft : find_next_chained_ft(prio);
+		err = connect_fwd_rules(dev, ft, next_ft);
+		if (err)
+			return err;
+	}
+
+	if (MLX5_CAP_FLOWTABLE(dev,
+			       flow_table_properties_nic_receive.modify_root))
+		err = update_root_ft_create(ft, prio);
+	return err;
+}
+
+static void list_add_flow_table(struct mlx5_flow_table *ft,
+				struct fs_prio *prio)
+{
+	struct list_head *prev = &prio->node.children;
+	struct mlx5_flow_table *iter;
+
+	fs_for_each_ft(iter, prio) {
+		if (iter->level > ft->level)
+			break;
+		prev = &iter->node.list;
+	}
+	list_add(&ft->node.list, prev);
+}
+
+static struct mlx5_flow_table *__mlx5_create_flow_table(struct mlx5_flow_namespace *ns,
+							struct mlx5_flow_table_attr *ft_attr,
+							enum fs_flow_table_op_mod op_mod,
+							u16 vport)
+{
+	struct mlx5_flow_root_namespace *root = find_root(&ns->node);
+	struct mlx5_flow_table *next_ft = NULL;
+	struct fs_prio *fs_prio = NULL;
+	struct mlx5_flow_table *ft;
+	int log_table_sz;
+	int err;
+
+	if (!root) {
+		pr_err("mlx5: flow steering failed to find root of namespace\n");
+		return ERR_PTR(-ENODEV);
+	}
+
+	mutex_lock(&root->chain_lock);
+	fs_prio = find_prio(ns, ft_attr->prio);
+	if (!fs_prio) {
+		err = -EINVAL;
+		goto unlock_root;
+	}
+	if (ft_attr->level >= fs_prio->num_levels) {
+		err = -ENOSPC;
+		goto unlock_root;
+	}
+	/* The level is related to the
+	 * priority level range.
+	 */
+	ft_attr->level += fs_prio->start_level;
+	ft = alloc_flow_table(ft_attr->level,
+			      vport,
+			      ft_attr->max_fte ? roundup_pow_of_two(ft_attr->max_fte) : 0,
+			      root->table_type,
+			      op_mod, ft_attr->flags);
+	if (IS_ERR(ft)) {
+		err = PTR_ERR(ft);
+		goto unlock_root;
+	}
+
+	tree_init_node(&ft->node, del_hw_flow_table, del_sw_flow_table);
+	log_table_sz = ft->max_fte ? ilog2(ft->max_fte) : 0;
+	next_ft = find_next_chained_ft(fs_prio);
+	err = root->cmds->create_flow_table(root->dev, ft->vport, ft->op_mod,
+					    ft->type, ft->level, log_table_sz,
+					    next_ft, &ft->id, ft->flags);
+	if (err)
+		goto free_ft;
+
+	err = connect_flow_table(root->dev, ft, fs_prio);
+	if (err)
+		goto destroy_ft;
+	ft->node.active = true;
+	down_write_ref_node(&fs_prio->node);
+	tree_add_node(&ft->node, &fs_prio->node);
+	list_add_flow_table(ft, fs_prio);
+	fs_prio->num_ft++;
+	up_write_ref_node(&fs_prio->node);
+	mutex_unlock(&root->chain_lock);
+	return ft;
+destroy_ft:
+	root->cmds->destroy_flow_table(root->dev, ft);
+free_ft:
+	rhltable_destroy(&ft->fgs_hash);
+	kfree(ft);
+unlock_root:
+	mutex_unlock(&root->chain_lock);
+	return ERR_PTR(err);
+}
+
+struct mlx5_flow_table *mlx5_create_flow_table(struct mlx5_flow_namespace *ns,
+					       struct mlx5_flow_table_attr *ft_attr)
+{
+	return __mlx5_create_flow_table(ns, ft_attr, FS_FT_OP_MOD_NORMAL, 0);
+}
+
+struct mlx5_flow_table *mlx5_create_vport_flow_table(struct mlx5_flow_namespace *ns,
+						     int prio, int max_fte,
+						     u32 level, u16 vport)
+{
+	struct mlx5_flow_table_attr ft_attr = {};
+
+	ft_attr.max_fte = max_fte;
+	ft_attr.level   = level;
+	ft_attr.prio    = prio;
+
+	return __mlx5_create_flow_table(ns, &ft_attr, FS_FT_OP_MOD_NORMAL, vport);
+}
+
+struct mlx5_flow_table*
+mlx5_create_lag_demux_flow_table(struct mlx5_flow_namespace *ns,
+				 int prio, u32 level)
+{
+	struct mlx5_flow_table_attr ft_attr = {};
+
+	ft_attr.level = level;
+	ft_attr.prio  = prio;
+	return __mlx5_create_flow_table(ns, &ft_attr, FS_FT_OP_MOD_LAG_DEMUX, 0);
+}
+EXPORT_SYMBOL(mlx5_create_lag_demux_flow_table);
+
+struct mlx5_flow_table*
+mlx5_create_auto_grouped_flow_table(struct mlx5_flow_namespace *ns,
+				    int prio,
+				    int num_flow_table_entries,
+				    int max_num_groups,
+				    u32 level,
+				    u32 flags)
+{
+	struct mlx5_flow_table_attr ft_attr = {};
+	struct mlx5_flow_table *ft;
+
+	if (max_num_groups > num_flow_table_entries)
+		return ERR_PTR(-EINVAL);
+
+	ft_attr.max_fte = num_flow_table_entries;
+	ft_attr.prio    = prio;
+	ft_attr.level   = level;
+	ft_attr.flags   = flags;
+
+	ft = mlx5_create_flow_table(ns, &ft_attr);
+	if (IS_ERR(ft))
+		return ft;
+
+	ft->autogroup.active = true;
+	ft->autogroup.required_groups = max_num_groups;
+	/* We save place for flow groups in addition to max types */
+	ft->autogroup.group_size = ft->max_fte / (max_num_groups + 1);
+
+	return ft;
+}
+EXPORT_SYMBOL(mlx5_create_auto_grouped_flow_table);
+
+struct mlx5_flow_group *mlx5_create_flow_group(struct mlx5_flow_table *ft,
+					       u32 *fg_in)
+{
+	struct mlx5_flow_root_namespace *root = find_root(&ft->node);
+	void *match_criteria = MLX5_ADDR_OF(create_flow_group_in,
+					    fg_in, match_criteria);
+	u8 match_criteria_enable = MLX5_GET(create_flow_group_in,
+					    fg_in,
+					    match_criteria_enable);
+	int start_index = MLX5_GET(create_flow_group_in, fg_in,
+				   start_flow_index);
+	int end_index = MLX5_GET(create_flow_group_in, fg_in,
+				 end_flow_index);
+	struct mlx5_core_dev *dev = get_dev(&ft->node);
+	struct mlx5_flow_group *fg;
+	int err;
+
+	if (ft->autogroup.active)
+		return ERR_PTR(-EPERM);
+
+	down_write_ref_node(&ft->node);
+	fg = alloc_insert_flow_group(ft, match_criteria_enable, match_criteria,
+				     start_index, end_index,
+				     ft->node.children.prev);
+	up_write_ref_node(&ft->node);
+	if (IS_ERR(fg))
+		return fg;
+
+	err = root->cmds->create_flow_group(dev, ft, fg_in, &fg->id);
+	if (err) {
+		tree_put_node(&fg->node);
+		return ERR_PTR(err);
+	}
+	trace_mlx5_fs_add_fg(fg);
+	fg->node.active = true;
+
+	return fg;
+}
+
+static struct mlx5_flow_rule *alloc_rule(struct mlx5_flow_destination *dest)
+{
+	struct mlx5_flow_rule *rule;
+
+	rule = kzalloc(sizeof(*rule), GFP_KERNEL);
+	if (!rule)
+		return NULL;
+
+	INIT_LIST_HEAD(&rule->next_ft);
+	rule->node.type = FS_TYPE_FLOW_DEST;
+	if (dest)
+		memcpy(&rule->dest_attr, dest, sizeof(*dest));
+
+	return rule;
+}
+
+static struct mlx5_flow_handle *alloc_handle(int num_rules)
+{
+	struct mlx5_flow_handle *handle;
+
+	handle = kzalloc(struct_size(handle, rule, num_rules), GFP_KERNEL);
+	if (!handle)
+		return NULL;
+
+	handle->num_rules = num_rules;
+
+	return handle;
+}
+
+static void destroy_flow_handle(struct fs_fte *fte,
+				struct mlx5_flow_handle *handle,
+				struct mlx5_flow_destination *dest,
+				int i)
+{
+	for (; --i >= 0;) {
+		if (refcount_dec_and_test(&handle->rule[i]->node.refcount)) {
+			fte->dests_size--;
+			list_del(&handle->rule[i]->node.list);
+			kfree(handle->rule[i]);
+		}
+	}
+	kfree(handle);
+}
+
+static struct mlx5_flow_handle *
+create_flow_handle(struct fs_fte *fte,
+		   struct mlx5_flow_destination *dest,
+		   int dest_num,
+		   int *modify_mask,
+		   bool *new_rule)
+{
+	struct mlx5_flow_handle *handle;
+	struct mlx5_flow_rule *rule = NULL;
+	static int count = BIT(MLX5_SET_FTE_MODIFY_ENABLE_MASK_FLOW_COUNTERS);
+	static int dst = BIT(MLX5_SET_FTE_MODIFY_ENABLE_MASK_DESTINATION_LIST);
+	int type;
+	int i = 0;
+
+	handle = alloc_handle((dest_num) ? dest_num : 1);
+	if (!handle)
+		return ERR_PTR(-ENOMEM);
+
+	do {
+		if (dest) {
+			rule = find_flow_rule(fte, dest + i);
+			if (rule) {
+				refcount_inc(&rule->node.refcount);
+				goto rule_found;
+			}
+		}
+
+		*new_rule = true;
+		rule = alloc_rule(dest + i);
+		if (!rule)
+			goto free_rules;
+
+		/* Add dest to dests list- we need flow tables to be in the
+		 * end of the list for forward to next prio rules.
+		 */
+		tree_init_node(&rule->node, NULL, del_sw_hw_rule);
+		if (dest &&
+		    dest[i].type != MLX5_FLOW_DESTINATION_TYPE_FLOW_TABLE)
+			list_add(&rule->node.list, &fte->node.children);
+		else
+			list_add_tail(&rule->node.list, &fte->node.children);
+		if (dest) {
+			fte->dests_size++;
+
+			type = dest[i].type ==
+				MLX5_FLOW_DESTINATION_TYPE_COUNTER;
+			*modify_mask |= type ? count : dst;
+		}
+rule_found:
+		handle->rule[i] = rule;
+	} while (++i < dest_num);
+
+	return handle;
+
+free_rules:
+	destroy_flow_handle(fte, handle, dest, i);
+	return ERR_PTR(-ENOMEM);
+}
+
+/* fte should not be deleted while calling this function */
+static struct mlx5_flow_handle *
+add_rule_fte(struct fs_fte *fte,
+	     struct mlx5_flow_group *fg,
+	     struct mlx5_flow_destination *dest,
+	     int dest_num,
+	     bool update_action)
+{
+	struct mlx5_flow_root_namespace *root;
+	struct mlx5_flow_handle *handle;
+	struct mlx5_flow_table *ft;
+	int modify_mask = 0;
+	int err;
+	bool new_rule = false;
+
+	handle = create_flow_handle(fte, dest, dest_num, &modify_mask,
+				    &new_rule);
+	if (IS_ERR(handle) || !new_rule)
+		goto out;
+
+	if (update_action)
+		modify_mask |= BIT(MLX5_SET_FTE_MODIFY_ENABLE_MASK_ACTION);
+
+	fs_get_obj(ft, fg->node.parent);
+	root = find_root(&fg->node);
+	if (!(fte->status & FS_FTE_STATUS_EXISTING))
+		err = root->cmds->create_fte(get_dev(&ft->node),
+					     ft, fg, fte);
+	else
+		err = root->cmds->update_fte(get_dev(&ft->node), ft, fg->id,
+						     modify_mask, fte);
+	if (err)
+		goto free_handle;
+
+	fte->node.active = true;
+	fte->status |= FS_FTE_STATUS_EXISTING;
+	atomic_inc(&fte->node.version);
+
+out:
+	return handle;
+
+free_handle:
+	destroy_flow_handle(fte, handle, dest, handle->num_rules);
+	return ERR_PTR(err);
+}
+
+static struct mlx5_flow_group *alloc_auto_flow_group(struct mlx5_flow_table  *ft,
+						     struct mlx5_flow_spec *spec)
+{
+	struct list_head *prev = &ft->node.children;
+	struct mlx5_flow_group *fg;
+	unsigned int candidate_index = 0;
+	unsigned int group_size = 0;
+
+	if (!ft->autogroup.active)
+		return ERR_PTR(-ENOENT);
+
+	if (ft->autogroup.num_groups < ft->autogroup.required_groups)
+		group_size = ft->autogroup.group_size;
+
+	/*  ft->max_fte == ft->autogroup.max_types */
+	if (group_size == 0)
+		group_size = 1;
+
+	/* sorted by start_index */
+	fs_for_each_fg(fg, ft) {
+		if (candidate_index + group_size > fg->start_index)
+			candidate_index = fg->start_index + fg->max_ftes;
+		else
+			break;
+		prev = &fg->node.list;
+	}
+
+	if (candidate_index + group_size > ft->max_fte)
+		return ERR_PTR(-ENOSPC);
+
+	fg = alloc_insert_flow_group(ft,
+				     spec->match_criteria_enable,
+				     spec->match_criteria,
+				     candidate_index,
+				     candidate_index + group_size - 1,
+				     prev);
+	if (IS_ERR(fg))
+		goto out;
+
+	if (group_size == ft->autogroup.group_size)
+		ft->autogroup.num_groups++;
+
+out:
+	return fg;
+}
+
+static int create_auto_flow_group(struct mlx5_flow_table *ft,
+				  struct mlx5_flow_group *fg)
+{
+	struct mlx5_flow_root_namespace *root = find_root(&ft->node);
+	struct mlx5_core_dev *dev = get_dev(&ft->node);
+	int inlen = MLX5_ST_SZ_BYTES(create_flow_group_in);
+	void *match_criteria_addr;
+	u8 src_esw_owner_mask_on;
+	void *misc;
+	int err;
+	u32 *in;
+
+	in = kvzalloc(inlen, GFP_KERNEL);
+	if (!in)
+		return -ENOMEM;
+
+	MLX5_SET(create_flow_group_in, in, match_criteria_enable,
+		 fg->mask.match_criteria_enable);
+	MLX5_SET(create_flow_group_in, in, start_flow_index, fg->start_index);
+	MLX5_SET(create_flow_group_in, in, end_flow_index,   fg->start_index +
+		 fg->max_ftes - 1);
+
+	misc = MLX5_ADDR_OF(fte_match_param, fg->mask.match_criteria,
+			    misc_parameters);
+	src_esw_owner_mask_on = !!MLX5_GET(fte_match_set_misc, misc,
+					 source_eswitch_owner_vhca_id);
+	MLX5_SET(create_flow_group_in, in,
+		 source_eswitch_owner_vhca_id_valid, src_esw_owner_mask_on);
+
+	match_criteria_addr = MLX5_ADDR_OF(create_flow_group_in,
+					   in, match_criteria);
+	memcpy(match_criteria_addr, fg->mask.match_criteria,
+	       sizeof(fg->mask.match_criteria));
+
+	err = root->cmds->create_flow_group(dev, ft, in, &fg->id);
+	if (!err) {
+		fg->node.active = true;
+		trace_mlx5_fs_add_fg(fg);
+	}
+
+	kvfree(in);
+	return err;
+}
+
+static bool mlx5_flow_dests_cmp(struct mlx5_flow_destination *d1,
+				struct mlx5_flow_destination *d2)
+{
+	if (d1->type == d2->type) {
+		if ((d1->type == MLX5_FLOW_DESTINATION_TYPE_VPORT &&
+		     d1->vport.num == d2->vport.num) ||
+		    (d1->type == MLX5_FLOW_DESTINATION_TYPE_FLOW_TABLE &&
+		     d1->ft == d2->ft) ||
+		    (d1->type == MLX5_FLOW_DESTINATION_TYPE_TIR &&
+		     d1->tir_num == d2->tir_num) ||
+		    (d1->type == MLX5_FLOW_DESTINATION_TYPE_FLOW_TABLE_NUM &&
+		     d1->ft_num == d2->ft_num))
+			return true;
+	}
+
+	return false;
+}
+
+static struct mlx5_flow_rule *find_flow_rule(struct fs_fte *fte,
+					     struct mlx5_flow_destination *dest)
+{
+	struct mlx5_flow_rule *rule;
+
+	list_for_each_entry(rule, &fte->node.children, node.list) {
+		if (mlx5_flow_dests_cmp(&rule->dest_attr, dest))
+			return rule;
+	}
+	return NULL;
+}
+
+static bool check_conflicting_actions(u32 action1, u32 action2)
+{
+	u32 xored_actions = action1 ^ action2;
+
+	/* if one rule only wants to count, it's ok */
+	if (action1 == MLX5_FLOW_CONTEXT_ACTION_COUNT ||
+	    action2 == MLX5_FLOW_CONTEXT_ACTION_COUNT)
+		return false;
+
+	if (xored_actions & (MLX5_FLOW_CONTEXT_ACTION_DROP  |
+			     MLX5_FLOW_CONTEXT_ACTION_ENCAP |
+			     MLX5_FLOW_CONTEXT_ACTION_DECAP |
+			     MLX5_FLOW_CONTEXT_ACTION_MOD_HDR  |
+			     MLX5_FLOW_CONTEXT_ACTION_VLAN_POP |
+			     MLX5_FLOW_CONTEXT_ACTION_VLAN_PUSH |
+			     MLX5_FLOW_CONTEXT_ACTION_VLAN_POP_2 |
+			     MLX5_FLOW_CONTEXT_ACTION_VLAN_PUSH_2))
+		return true;
+
+	return false;
+}
+
+static int check_conflicting_ftes(struct fs_fte *fte, const struct mlx5_flow_act *flow_act)
+{
+	if (check_conflicting_actions(flow_act->action, fte->action.action)) {
+		mlx5_core_warn(get_dev(&fte->node),
+			       "Found two FTEs with conflicting actions\n");
+		return -EEXIST;
+	}
+
+	if (flow_act->has_flow_tag &&
+	    fte->action.flow_tag != flow_act->flow_tag) {
+		mlx5_core_warn(get_dev(&fte->node),
+			       "FTE flow tag %u already exists with different flow tag %u\n",
+			       fte->action.flow_tag,
+			       flow_act->flow_tag);
+		return -EEXIST;
+	}
+
+	return 0;
+}
+
+static struct mlx5_flow_handle *add_rule_fg(struct mlx5_flow_group *fg,
+					    u32 *match_value,
+					    struct mlx5_flow_act *flow_act,
+					    struct mlx5_flow_destination *dest,
+					    int dest_num,
+					    struct fs_fte *fte)
+{
+	struct mlx5_flow_handle *handle;
+	int old_action;
+	int i;
+	int ret;
+
+	ret = check_conflicting_ftes(fte, flow_act);
+	if (ret)
+		return ERR_PTR(ret);
+
+	old_action = fte->action.action;
+	fte->action.action |= flow_act->action;
+	handle = add_rule_fte(fte, fg, dest, dest_num,
+			      old_action != flow_act->action);
+	if (IS_ERR(handle)) {
+		fte->action.action = old_action;
+		return handle;
+	}
+	trace_mlx5_fs_set_fte(fte, false);
+
+	for (i = 0; i < handle->num_rules; i++) {
+		if (refcount_read(&handle->rule[i]->node.refcount) == 1) {
+			tree_add_node(&handle->rule[i]->node, &fte->node);
+			trace_mlx5_fs_add_rule(handle->rule[i]);
+		}
+	}
+	return handle;
+}
+
+struct mlx5_fc *mlx5_flow_rule_counter(struct mlx5_flow_handle *handle)
+{
+	struct mlx5_flow_rule *dst;
+	struct fs_fte *fte;
+
+	fs_get_obj(fte, handle->rule[0]->node.parent);
+
+	fs_for_each_dst(dst, fte) {
+		if (dst->dest_attr.type == MLX5_FLOW_DESTINATION_TYPE_COUNTER)
+			return dst->dest_attr.counter;
+	}
+
+	return NULL;
+}
+
+static bool counter_is_valid(struct mlx5_fc *counter, u32 action)
+{
+	if (!(action & MLX5_FLOW_CONTEXT_ACTION_COUNT))
+		return !counter;
+
+	if (!counter)
+		return false;
+
+	return (action & (MLX5_FLOW_CONTEXT_ACTION_DROP |
+			  MLX5_FLOW_CONTEXT_ACTION_FWD_DEST));
+}
+
+static bool dest_is_valid(struct mlx5_flow_destination *dest,
+			  u32 action,
+			  struct mlx5_flow_table *ft)
+{
+	if (dest && (dest->type == MLX5_FLOW_DESTINATION_TYPE_COUNTER))
+		return counter_is_valid(dest->counter, action);
+
+	if (!(action & MLX5_FLOW_CONTEXT_ACTION_FWD_DEST))
+		return true;
+
+	if (!dest || ((dest->type ==
+	    MLX5_FLOW_DESTINATION_TYPE_FLOW_TABLE) &&
+	    (dest->ft->level <= ft->level)))
+		return false;
+	return true;
+}
+
+struct match_list {
+	struct list_head	list;
+	struct mlx5_flow_group *g;
+};
+
+struct match_list_head {
+	struct list_head  list;
+	struct match_list first;
+};
+
+static void free_match_list(struct match_list_head *head)
+{
+	if (!list_empty(&head->list)) {
+		struct match_list *iter, *match_tmp;
+
+		list_del(&head->first.list);
+		tree_put_node(&head->first.g->node);
+		list_for_each_entry_safe(iter, match_tmp, &head->list,
+					 list) {
+			tree_put_node(&iter->g->node);
+			list_del(&iter->list);
+			kfree(iter);
+		}
+	}
+}
+
+static int build_match_list(struct match_list_head *match_head,
+			    struct mlx5_flow_table *ft,
+			    struct mlx5_flow_spec *spec)
+{
+	struct rhlist_head *tmp, *list;
+	struct mlx5_flow_group *g;
+	int err = 0;
+
+	rcu_read_lock();
+	INIT_LIST_HEAD(&match_head->list);
+	/* Collect all fgs which has a matching match_criteria */
+	list = rhltable_lookup(&ft->fgs_hash, spec, rhash_fg);
+	/* RCU is atomic, we can't execute FW commands here */
+	rhl_for_each_entry_rcu(g, tmp, list, hash) {
+		struct match_list *curr_match;
+
+		if (likely(list_empty(&match_head->list))) {
+			if (!tree_get_node(&g->node))
+				continue;
+			match_head->first.g = g;
+			list_add_tail(&match_head->first.list,
+				      &match_head->list);
+			continue;
+		}
+
+		curr_match = kmalloc(sizeof(*curr_match), GFP_ATOMIC);
+		if (!curr_match) {
+			rcu_read_unlock();
+			free_match_list(match_head);
+			return -ENOMEM;
+		}
+		if (!tree_get_node(&g->node)) {
+			kfree(curr_match);
+			continue;
+		}
+		curr_match->g = g;
+		list_add_tail(&curr_match->list, &match_head->list);
+	}
+	rcu_read_unlock();
+	return err;
+}
+
+static u64 matched_fgs_get_version(struct list_head *match_head)
+{
+	struct match_list *iter;
+	u64 version = 0;
+
+	list_for_each_entry(iter, match_head, list)
+		version += (u64)atomic_read(&iter->g->node.version);
+	return version;
+}
+
+static struct fs_fte *
+lookup_fte_locked(struct mlx5_flow_group *g,
+		  u32 *match_value,
+		  bool take_write)
+{
+	struct fs_fte *fte_tmp;
+
+	if (take_write)
+		nested_down_write_ref_node(&g->node, FS_LOCK_PARENT);
+	else
+		nested_down_read_ref_node(&g->node, FS_LOCK_PARENT);
+	fte_tmp = rhashtable_lookup_fast(&g->ftes_hash, match_value,
+					 rhash_fte);
+	if (!fte_tmp || !tree_get_node(&fte_tmp->node)) {
+		fte_tmp = NULL;
+		goto out;
+	}
+	if (!fte_tmp->node.active) {
+		tree_put_node(&fte_tmp->node);
+		fte_tmp = NULL;
+		goto out;
+	}
+
+	nested_down_write_ref_node(&fte_tmp->node, FS_LOCK_CHILD);
+out:
+	if (take_write)
+		up_write_ref_node(&g->node);
+	else
+		up_read_ref_node(&g->node);
+	return fte_tmp;
+}
+
+static struct mlx5_flow_handle *
+try_add_to_existing_fg(struct mlx5_flow_table *ft,
+		       struct list_head *match_head,
+		       struct mlx5_flow_spec *spec,
+		       struct mlx5_flow_act *flow_act,
+		       struct mlx5_flow_destination *dest,
+		       int dest_num,
+		       int ft_version)
+{
+	struct mlx5_flow_steering *steering = get_steering(&ft->node);
+	struct mlx5_flow_group *g;
+	struct mlx5_flow_handle *rule;
+	struct match_list *iter;
+	bool take_write = false;
+	struct fs_fte *fte;
+	u64  version;
+	int err;
+
+	fte = alloc_fte(ft, spec->match_value, flow_act);
+	if (IS_ERR(fte))
+		return  ERR_PTR(-ENOMEM);
+
+search_again_locked:
+	version = matched_fgs_get_version(match_head);
+	/* Try to find a fg that already contains a matching fte */
+	list_for_each_entry(iter, match_head, list) {
+		struct fs_fte *fte_tmp;
+
+		g = iter->g;
+		fte_tmp = lookup_fte_locked(g, spec->match_value, take_write);
+		if (!fte_tmp)
+			continue;
+		rule = add_rule_fg(g, spec->match_value,
+				   flow_act, dest, dest_num, fte_tmp);
+		up_write_ref_node(&fte_tmp->node);
+		tree_put_node(&fte_tmp->node);
+		kmem_cache_free(steering->ftes_cache, fte);
+		return rule;
+	}
+
+	/* Check the ft version, for case that new flow group
+	 * was added while the fgs weren't locked
+	 */
+	if (atomic_read(&ft->node.version) != ft_version) {
+		rule = ERR_PTR(-EAGAIN);
+		goto out;
+	}
+
+	/* Check the fgs version, for case the new FTE with the
+	 * same values was added while the fgs weren't locked
+	 */
+	if (version != matched_fgs_get_version(match_head)) {
+		take_write = true;
+		goto search_again_locked;
+	}
+
+	list_for_each_entry(iter, match_head, list) {
+		g = iter->g;
+
+		if (!g->node.active)
+			continue;
+
+		nested_down_write_ref_node(&g->node, FS_LOCK_PARENT);
+
+		err = insert_fte(g, fte);
+		if (err) {
+			up_write_ref_node(&g->node);
+			if (err == -ENOSPC)
+				continue;
+			kmem_cache_free(steering->ftes_cache, fte);
+			return ERR_PTR(err);
+		}
+
+		nested_down_write_ref_node(&fte->node, FS_LOCK_CHILD);
+		up_write_ref_node(&g->node);
+		rule = add_rule_fg(g, spec->match_value,
+				   flow_act, dest, dest_num, fte);
+		up_write_ref_node(&fte->node);
+		tree_put_node(&fte->node);
+		return rule;
+	}
+	rule = ERR_PTR(-ENOENT);
+out:
+	kmem_cache_free(steering->ftes_cache, fte);
+	return rule;
+}
+
+static struct mlx5_flow_handle *
+_mlx5_add_flow_rules(struct mlx5_flow_table *ft,
+		     struct mlx5_flow_spec *spec,
+		     struct mlx5_flow_act *flow_act,
+		     struct mlx5_flow_destination *dest,
+		     int dest_num)
+
+{
+	struct mlx5_flow_steering *steering = get_steering(&ft->node);
+	struct mlx5_flow_group *g;
+	struct mlx5_flow_handle *rule;
+	struct match_list_head match_head;
+	bool take_write = false;
+	struct fs_fte *fte;
+	int version;
+	int err;
+	int i;
+
+	if (!check_valid_spec(spec))
+		return ERR_PTR(-EINVAL);
+
+	for (i = 0; i < dest_num; i++) {
+		if (!dest_is_valid(&dest[i], flow_act->action, ft))
+			return ERR_PTR(-EINVAL);
+	}
+	nested_down_read_ref_node(&ft->node, FS_LOCK_GRANDPARENT);
+search_again_locked:
+	version = atomic_read(&ft->node.version);
+
+	/* Collect all fgs which has a matching match_criteria */
+	err = build_match_list(&match_head, ft, spec);
+	if (err) {
+		if (take_write)
+			up_write_ref_node(&ft->node);
+		else
+			up_read_ref_node(&ft->node);
+		return ERR_PTR(err);
+	}
+
+	if (!take_write)
+		up_read_ref_node(&ft->node);
+
+	rule = try_add_to_existing_fg(ft, &match_head.list, spec, flow_act, dest,
+				      dest_num, version);
+	free_match_list(&match_head);
+	if (!IS_ERR(rule) ||
+	    (PTR_ERR(rule) != -ENOENT && PTR_ERR(rule) != -EAGAIN)) {
+		if (take_write)
+			up_write_ref_node(&ft->node);
+		return rule;
+	}
+
+	if (!take_write) {
+		nested_down_write_ref_node(&ft->node, FS_LOCK_GRANDPARENT);
+		take_write = true;
+	}
+
+	if (PTR_ERR(rule) == -EAGAIN ||
+	    version != atomic_read(&ft->node.version))
+		goto search_again_locked;
+
+	g = alloc_auto_flow_group(ft, spec);
+	if (IS_ERR(g)) {
+		rule = ERR_CAST(g);
+		up_write_ref_node(&ft->node);
+		return rule;
+	}
+
+	nested_down_write_ref_node(&g->node, FS_LOCK_PARENT);
+	up_write_ref_node(&ft->node);
+
+	err = create_auto_flow_group(ft, g);
+	if (err)
+		goto err_release_fg;
+
+	fte = alloc_fte(ft, spec->match_value, flow_act);
+	if (IS_ERR(fte)) {
+		err = PTR_ERR(fte);
+		goto err_release_fg;
+	}
+
+	err = insert_fte(g, fte);
+	if (err) {
+		kmem_cache_free(steering->ftes_cache, fte);
+		goto err_release_fg;
+	}
+
+	nested_down_write_ref_node(&fte->node, FS_LOCK_CHILD);
+	up_write_ref_node(&g->node);
+	rule = add_rule_fg(g, spec->match_value, flow_act, dest,
+			   dest_num, fte);
+	up_write_ref_node(&fte->node);
+	tree_put_node(&fte->node);
+	tree_put_node(&g->node);
+	return rule;
+
+err_release_fg:
+	up_write_ref_node(&g->node);
+	tree_put_node(&g->node);
+	return ERR_PTR(err);
+}
+
+static bool fwd_next_prio_supported(struct mlx5_flow_table *ft)
+{
+	return ((ft->type == FS_FT_NIC_RX) &&
+		(MLX5_CAP_FLOWTABLE(get_dev(&ft->node), nic_rx_multi_path_tirs)));
+}
+
+struct mlx5_flow_handle *
+mlx5_add_flow_rules(struct mlx5_flow_table *ft,
+		    struct mlx5_flow_spec *spec,
+		    struct mlx5_flow_act *flow_act,
+		    struct mlx5_flow_destination *dest,
+		    int num_dest)
+{
+	struct mlx5_flow_root_namespace *root = find_root(&ft->node);
+	struct mlx5_flow_destination gen_dest = {};
+	struct mlx5_flow_table *next_ft = NULL;
+	struct mlx5_flow_handle *handle = NULL;
+	u32 sw_action = flow_act->action;
+	struct fs_prio *prio;
+
+	fs_get_obj(prio, ft->node.parent);
+	if (flow_act->action == MLX5_FLOW_CONTEXT_ACTION_FWD_NEXT_PRIO) {
+		if (!fwd_next_prio_supported(ft))
+			return ERR_PTR(-EOPNOTSUPP);
+		if (num_dest)
+			return ERR_PTR(-EINVAL);
+		mutex_lock(&root->chain_lock);
+		next_ft = find_next_chained_ft(prio);
+		if (next_ft) {
+			gen_dest.type = MLX5_FLOW_DESTINATION_TYPE_FLOW_TABLE;
+			gen_dest.ft = next_ft;
+			dest = &gen_dest;
+			num_dest = 1;
+			flow_act->action = MLX5_FLOW_CONTEXT_ACTION_FWD_DEST;
+		} else {
+			mutex_unlock(&root->chain_lock);
+			return ERR_PTR(-EOPNOTSUPP);
+		}
+	}
+
+	handle = _mlx5_add_flow_rules(ft, spec, flow_act, dest, num_dest);
+
+	if (sw_action == MLX5_FLOW_CONTEXT_ACTION_FWD_NEXT_PRIO) {
+		if (!IS_ERR_OR_NULL(handle) &&
+		    (list_empty(&handle->rule[0]->next_ft))) {
+			mutex_lock(&next_ft->lock);
+			list_add(&handle->rule[0]->next_ft,
+				 &next_ft->fwd_rules);
+			mutex_unlock(&next_ft->lock);
+			handle->rule[0]->sw_action = MLX5_FLOW_CONTEXT_ACTION_FWD_NEXT_PRIO;
+		}
+		mutex_unlock(&root->chain_lock);
+	}
+	return handle;
+}
+EXPORT_SYMBOL(mlx5_add_flow_rules);
+
+void mlx5_del_flow_rules(struct mlx5_flow_handle *handle)
+{
+	int i;
+
+	for (i = handle->num_rules - 1; i >= 0; i--)
+		tree_remove_node(&handle->rule[i]->node);
+	kfree(handle);
+}
+EXPORT_SYMBOL(mlx5_del_flow_rules);
+
+/* Assuming prio->node.children(flow tables) is sorted by level */
+static struct mlx5_flow_table *find_next_ft(struct mlx5_flow_table *ft)
+{
+	struct fs_prio *prio;
+
+	fs_get_obj(prio, ft->node.parent);
+
+	if (!list_is_last(&ft->node.list, &prio->node.children))
+		return list_next_entry(ft, node.list);
+	return find_next_chained_ft(prio);
+}
+
+static int update_root_ft_destroy(struct mlx5_flow_table *ft)
+{
+	struct mlx5_flow_root_namespace *root = find_root(&ft->node);
+	struct mlx5_ft_underlay_qp *uqp;
+	struct mlx5_flow_table *new_root_ft = NULL;
+	int err = 0;
+	u32 qpn;
+
+	if (root->root_ft != ft)
+		return 0;
+
+	new_root_ft = find_next_ft(ft);
+	if (!new_root_ft) {
+		root->root_ft = NULL;
+		return 0;
+	}
+
+	if (list_empty(&root->underlay_qpns)) {
+		/* Don't set any QPN (zero) in case QPN list is empty */
+		qpn = 0;
+		err = root->cmds->update_root_ft(root->dev, new_root_ft,
+						 qpn, false);
+	} else {
+		list_for_each_entry(uqp, &root->underlay_qpns, list) {
+			qpn = uqp->qpn;
+			err = root->cmds->update_root_ft(root->dev,
+							 new_root_ft, qpn,
+							 false);
+			if (err)
+				break;
+		}
+	}
+
+	if (err)
+		mlx5_core_warn(root->dev,
+			       "Update root flow table of id(%u) qpn(%d) failed\n",
+			       ft->id, qpn);
+	else
+		root->root_ft = new_root_ft;
+
+	return 0;
+}
+
+/* Connect flow table from previous priority to
+ * the next flow table.
+ */
+static int disconnect_flow_table(struct mlx5_flow_table *ft)
+{
+	struct mlx5_core_dev *dev = get_dev(&ft->node);
+	struct mlx5_flow_table *next_ft;
+	struct fs_prio *prio;
+	int err = 0;
+
+	err = update_root_ft_destroy(ft);
+	if (err)
+		return err;
+
+	fs_get_obj(prio, ft->node.parent);
+	if  (!(list_first_entry(&prio->node.children,
+				struct mlx5_flow_table,
+				node.list) == ft))
+		return 0;
+
+	next_ft = find_next_ft(ft);
+	err = connect_fwd_rules(dev, next_ft, ft);
+	if (err)
+		return err;
+
+	err = connect_prev_fts(dev, next_ft, prio);
+	if (err)
+		mlx5_core_warn(dev, "Failed to disconnect flow table %d\n",
+			       ft->id);
+	return err;
+}
+
+int mlx5_destroy_flow_table(struct mlx5_flow_table *ft)
+{
+	struct mlx5_flow_root_namespace *root = find_root(&ft->node);
+	int err = 0;
+
+	mutex_lock(&root->chain_lock);
+	err = disconnect_flow_table(ft);
+	if (err) {
+		mutex_unlock(&root->chain_lock);
+		return err;
+	}
+	if (tree_remove_node(&ft->node))
+		mlx5_core_warn(get_dev(&ft->node), "Flow table %d wasn't destroyed, refcount > 1\n",
+			       ft->id);
+	mutex_unlock(&root->chain_lock);
+
+	return err;
+}
+EXPORT_SYMBOL(mlx5_destroy_flow_table);
+
+void mlx5_destroy_flow_group(struct mlx5_flow_group *fg)
+{
+	if (tree_remove_node(&fg->node))
+		mlx5_core_warn(get_dev(&fg->node), "Flow group %d wasn't destroyed, refcount > 1\n",
+			       fg->id);
+}
+
+struct mlx5_flow_namespace *mlx5_get_flow_namespace(struct mlx5_core_dev *dev,
+						    enum mlx5_flow_namespace_type type)
+{
+	struct mlx5_flow_steering *steering = dev->priv.steering;
+	struct mlx5_flow_root_namespace *root_ns;
+	int prio;
+	struct fs_prio *fs_prio;
+	struct mlx5_flow_namespace *ns;
+
+	if (!steering)
+		return NULL;
+
+	switch (type) {
+	case MLX5_FLOW_NAMESPACE_BYPASS:
+	case MLX5_FLOW_NAMESPACE_LAG:
+	case MLX5_FLOW_NAMESPACE_OFFLOADS:
+	case MLX5_FLOW_NAMESPACE_ETHTOOL:
+	case MLX5_FLOW_NAMESPACE_KERNEL:
+	case MLX5_FLOW_NAMESPACE_LEFTOVERS:
+	case MLX5_FLOW_NAMESPACE_ANCHOR:
+		prio = type;
+		break;
+	case MLX5_FLOW_NAMESPACE_FDB:
+		if (steering->fdb_root_ns)
+			return &steering->fdb_root_ns->ns;
+		else
+			return NULL;
+	case MLX5_FLOW_NAMESPACE_SNIFFER_RX:
+		if (steering->sniffer_rx_root_ns)
+			return &steering->sniffer_rx_root_ns->ns;
+		else
+			return NULL;
+	case MLX5_FLOW_NAMESPACE_SNIFFER_TX:
+		if (steering->sniffer_tx_root_ns)
+			return &steering->sniffer_tx_root_ns->ns;
+		else
+			return NULL;
+	case MLX5_FLOW_NAMESPACE_EGRESS:
+		if (steering->egress_root_ns)
+			return &steering->egress_root_ns->ns;
+		else
+			return NULL;
+	default:
+		return NULL;
+	}
+
+	root_ns = steering->root_ns;
+	if (!root_ns)
+		return NULL;
+
+	fs_prio = find_prio(&root_ns->ns, prio);
+	if (!fs_prio)
+		return NULL;
+
+	ns = list_first_entry(&fs_prio->node.children,
+			      typeof(*ns),
+			      node.list);
+
+	return ns;
+}
+EXPORT_SYMBOL(mlx5_get_flow_namespace);
+
+struct mlx5_flow_namespace *mlx5_get_flow_vport_acl_namespace(struct mlx5_core_dev *dev,
+							      enum mlx5_flow_namespace_type type,
+							      int vport)
+{
+	struct mlx5_flow_steering *steering = dev->priv.steering;
+
+	if (!steering || vport >= MLX5_TOTAL_VPORTS(dev))
+		return NULL;
+
+	switch (type) {
+	case MLX5_FLOW_NAMESPACE_ESW_EGRESS:
+		if (steering->esw_egress_root_ns &&
+		    steering->esw_egress_root_ns[vport])
+			return &steering->esw_egress_root_ns[vport]->ns;
+		else
+			return NULL;
+	case MLX5_FLOW_NAMESPACE_ESW_INGRESS:
+		if (steering->esw_ingress_root_ns &&
+		    steering->esw_ingress_root_ns[vport])
+			return &steering->esw_ingress_root_ns[vport]->ns;
+		else
+			return NULL;
+	default:
+		return NULL;
+	}
+}
+
+static struct fs_prio *fs_create_prio(struct mlx5_flow_namespace *ns,
+				      unsigned int prio, int num_levels)
+{
+	struct fs_prio *fs_prio;
+
+	fs_prio = kzalloc(sizeof(*fs_prio), GFP_KERNEL);
+	if (!fs_prio)
+		return ERR_PTR(-ENOMEM);
+
+	fs_prio->node.type = FS_TYPE_PRIO;
+	tree_init_node(&fs_prio->node, NULL, del_sw_prio);
+	tree_add_node(&fs_prio->node, &ns->node);
+	fs_prio->num_levels = num_levels;
+	fs_prio->prio = prio;
+	list_add_tail(&fs_prio->node.list, &ns->node.children);
+
+	return fs_prio;
+}
+
+static struct mlx5_flow_namespace *fs_init_namespace(struct mlx5_flow_namespace
+						     *ns)
+{
+	ns->node.type = FS_TYPE_NAMESPACE;
+
+	return ns;
+}
+
+static struct mlx5_flow_namespace *fs_create_namespace(struct fs_prio *prio)
+{
+	struct mlx5_flow_namespace	*ns;
+
+	ns = kzalloc(sizeof(*ns), GFP_KERNEL);
+	if (!ns)
+		return ERR_PTR(-ENOMEM);
+
+	fs_init_namespace(ns);
+	tree_init_node(&ns->node, NULL, del_sw_ns);
+	tree_add_node(&ns->node, &prio->node);
+	list_add_tail(&ns->node.list, &prio->node.children);
+
+	return ns;
+}
+
+static int create_leaf_prios(struct mlx5_flow_namespace *ns, int prio,
+			     struct init_tree_node *prio_metadata)
+{
+	struct fs_prio *fs_prio;
+	int i;
+
+	for (i = 0; i < prio_metadata->num_leaf_prios; i++) {
+		fs_prio = fs_create_prio(ns, prio++, prio_metadata->num_levels);
+		if (IS_ERR(fs_prio))
+			return PTR_ERR(fs_prio);
+	}
+	return 0;
+}
+
+#define FLOW_TABLE_BIT_SZ 1
+#define GET_FLOW_TABLE_CAP(dev, offset) \
+	((be32_to_cpu(*((__be32 *)(dev->caps.hca_cur[MLX5_CAP_FLOW_TABLE]) +	\
+			offset / 32)) >>					\
+	  (32 - FLOW_TABLE_BIT_SZ - (offset & 0x1f))) & FLOW_TABLE_BIT_SZ)
+static bool has_required_caps(struct mlx5_core_dev *dev, struct node_caps *caps)
+{
+	int i;
+
+	for (i = 0; i < caps->arr_sz; i++) {
+		if (!GET_FLOW_TABLE_CAP(dev, caps->caps[i]))
+			return false;
+	}
+	return true;
+}
+
+static int init_root_tree_recursive(struct mlx5_flow_steering *steering,
+				    struct init_tree_node *init_node,
+				    struct fs_node *fs_parent_node,
+				    struct init_tree_node *init_parent_node,
+				    int prio)
+{
+	int max_ft_level = MLX5_CAP_FLOWTABLE(steering->dev,
+					      flow_table_properties_nic_receive.
+					      max_ft_level);
+	struct mlx5_flow_namespace *fs_ns;
+	struct fs_prio *fs_prio;
+	struct fs_node *base;
+	int i;
+	int err;
+
+	if (init_node->type == FS_TYPE_PRIO) {
+		if ((init_node->min_ft_level > max_ft_level) ||
+		    !has_required_caps(steering->dev, &init_node->caps))
+			return 0;
+
+		fs_get_obj(fs_ns, fs_parent_node);
+		if (init_node->num_leaf_prios)
+			return create_leaf_prios(fs_ns, prio, init_node);
+		fs_prio = fs_create_prio(fs_ns, prio, init_node->num_levels);
+		if (IS_ERR(fs_prio))
+			return PTR_ERR(fs_prio);
+		base = &fs_prio->node;
+	} else if (init_node->type == FS_TYPE_NAMESPACE) {
+		fs_get_obj(fs_prio, fs_parent_node);
+		fs_ns = fs_create_namespace(fs_prio);
+		if (IS_ERR(fs_ns))
+			return PTR_ERR(fs_ns);
+		base = &fs_ns->node;
+	} else {
+		return -EINVAL;
+	}
+	prio = 0;
+	for (i = 0; i < init_node->ar_size; i++) {
+		err = init_root_tree_recursive(steering, &init_node->children[i],
+					       base, init_node, prio);
+		if (err)
+			return err;
+		if (init_node->children[i].type == FS_TYPE_PRIO &&
+		    init_node->children[i].num_leaf_prios) {
+			prio += init_node->children[i].num_leaf_prios;
+		}
+	}
+
+	return 0;
+}
+
+static int init_root_tree(struct mlx5_flow_steering *steering,
+			  struct init_tree_node *init_node,
+			  struct fs_node *fs_parent_node)
+{
+	int i;
+	struct mlx5_flow_namespace *fs_ns;
+	int err;
+
+	fs_get_obj(fs_ns, fs_parent_node);
+	for (i = 0; i < init_node->ar_size; i++) {
+		err = init_root_tree_recursive(steering, &init_node->children[i],
+					       &fs_ns->node,
+					       init_node, i);
+		if (err)
+			return err;
+	}
+	return 0;
+}
+
+static struct mlx5_flow_root_namespace
+*create_root_ns(struct mlx5_flow_steering *steering,
+		enum fs_flow_table_type table_type)
+{
+	const struct mlx5_flow_cmds *cmds = mlx5_fs_cmd_get_default(table_type);
+	struct mlx5_flow_root_namespace *root_ns;
+	struct mlx5_flow_namespace *ns;
+
+	if (mlx5_accel_ipsec_device_caps(steering->dev) & MLX5_ACCEL_IPSEC_CAP_DEVICE &&
+	    (table_type == FS_FT_NIC_RX || table_type == FS_FT_NIC_TX))
+		cmds = mlx5_fs_cmd_get_default_ipsec_fpga_cmds(table_type);
+
+	/* Create the root namespace */
+	root_ns = kzalloc(sizeof(*root_ns), GFP_KERNEL);
+	if (!root_ns)
+		return NULL;
+
+	root_ns->dev = steering->dev;
+	root_ns->table_type = table_type;
+	root_ns->cmds = cmds;
+
+	INIT_LIST_HEAD(&root_ns->underlay_qpns);
+
+	ns = &root_ns->ns;
+	fs_init_namespace(ns);
+	mutex_init(&root_ns->chain_lock);
+	tree_init_node(&ns->node, NULL, NULL);
+	tree_add_node(&ns->node, NULL);
+
+	return root_ns;
+}
+
+static void set_prio_attrs_in_prio(struct fs_prio *prio, int acc_level);
+
+static int set_prio_attrs_in_ns(struct mlx5_flow_namespace *ns, int acc_level)
+{
+	struct fs_prio *prio;
+
+	fs_for_each_prio(prio, ns) {
+		 /* This updates prio start_level and num_levels */
+		set_prio_attrs_in_prio(prio, acc_level);
+		acc_level += prio->num_levels;
+	}
+	return acc_level;
+}
+
+static void set_prio_attrs_in_prio(struct fs_prio *prio, int acc_level)
+{
+	struct mlx5_flow_namespace *ns;
+	int acc_level_ns = acc_level;
+
+	prio->start_level = acc_level;
+	fs_for_each_ns(ns, prio)
+		/* This updates start_level and num_levels of ns's priority descendants */
+		acc_level_ns = set_prio_attrs_in_ns(ns, acc_level);
+	if (!prio->num_levels)
+		prio->num_levels = acc_level_ns - prio->start_level;
+	WARN_ON(prio->num_levels < acc_level_ns - prio->start_level);
+}
+
+static void set_prio_attrs(struct mlx5_flow_root_namespace *root_ns)
+{
+	struct mlx5_flow_namespace *ns = &root_ns->ns;
+	struct fs_prio *prio;
+	int start_level = 0;
+
+	fs_for_each_prio(prio, ns) {
+		set_prio_attrs_in_prio(prio, start_level);
+		start_level += prio->num_levels;
+	}
+}
+
+#define ANCHOR_PRIO 0
+#define ANCHOR_SIZE 1
+#define ANCHOR_LEVEL 0
+static int create_anchor_flow_table(struct mlx5_flow_steering *steering)
+{
+	struct mlx5_flow_namespace *ns = NULL;
+	struct mlx5_flow_table_attr ft_attr = {};
+	struct mlx5_flow_table *ft;
+
+	ns = mlx5_get_flow_namespace(steering->dev, MLX5_FLOW_NAMESPACE_ANCHOR);
+	if (WARN_ON(!ns))
+		return -EINVAL;
+
+	ft_attr.max_fte = ANCHOR_SIZE;
+	ft_attr.level   = ANCHOR_LEVEL;
+	ft_attr.prio    = ANCHOR_PRIO;
+
+	ft = mlx5_create_flow_table(ns, &ft_attr);
+	if (IS_ERR(ft)) {
+		mlx5_core_err(steering->dev, "Failed to create last anchor flow table");
+		return PTR_ERR(ft);
+	}
+	return 0;
+}
+
+static int init_root_ns(struct mlx5_flow_steering *steering)
+{
+	int err;
+
+	steering->root_ns = create_root_ns(steering, FS_FT_NIC_RX);
+	if (!steering->root_ns)
+		return -ENOMEM;
+
+	err = init_root_tree(steering, &root_fs, &steering->root_ns->ns.node);
+	if (err)
+		goto out_err;
+
+	set_prio_attrs(steering->root_ns);
+	err = create_anchor_flow_table(steering);
+	if (err)
+		goto out_err;
+
+	return 0;
+
+out_err:
+	cleanup_root_ns(steering->root_ns);
+	steering->root_ns = NULL;
+	return err;
+}
+
+static void clean_tree(struct fs_node *node)
+{
+	if (node) {
+		struct fs_node *iter;
+		struct fs_node *temp;
+
+		tree_get_node(node);
+		list_for_each_entry_safe(iter, temp, &node->children, list)
+			clean_tree(iter);
+		tree_put_node(node);
+		tree_remove_node(node);
+	}
+}
+
+static void cleanup_root_ns(struct mlx5_flow_root_namespace *root_ns)
+{
+	if (!root_ns)
+		return;
+
+	clean_tree(&root_ns->ns.node);
+}
+
+static void cleanup_egress_acls_root_ns(struct mlx5_core_dev *dev)
+{
+	struct mlx5_flow_steering *steering = dev->priv.steering;
+	int i;
+
+	if (!steering->esw_egress_root_ns)
+		return;
+
+	for (i = 0; i < MLX5_TOTAL_VPORTS(dev); i++)
+		cleanup_root_ns(steering->esw_egress_root_ns[i]);
+
+	kfree(steering->esw_egress_root_ns);
+	steering->esw_egress_root_ns = NULL;
+}
+
+static void cleanup_ingress_acls_root_ns(struct mlx5_core_dev *dev)
+{
+	struct mlx5_flow_steering *steering = dev->priv.steering;
+	int i;
+
+	if (!steering->esw_ingress_root_ns)
+		return;
+
+	for (i = 0; i < MLX5_TOTAL_VPORTS(dev); i++)
+		cleanup_root_ns(steering->esw_ingress_root_ns[i]);
+
+	kfree(steering->esw_ingress_root_ns);
+	steering->esw_ingress_root_ns = NULL;
+}
+
+void mlx5_cleanup_fs(struct mlx5_core_dev *dev)
+{
+	struct mlx5_flow_steering *steering = dev->priv.steering;
+
+	cleanup_root_ns(steering->root_ns);
+	cleanup_egress_acls_root_ns(dev);
+	cleanup_ingress_acls_root_ns(dev);
+	cleanup_root_ns(steering->fdb_root_ns);
+	cleanup_root_ns(steering->sniffer_rx_root_ns);
+	cleanup_root_ns(steering->sniffer_tx_root_ns);
+	cleanup_root_ns(steering->egress_root_ns);
+	mlx5_cleanup_fc_stats(dev);
+	kmem_cache_destroy(steering->ftes_cache);
+	kmem_cache_destroy(steering->fgs_cache);
+	kfree(steering);
+}
+
+static int init_sniffer_tx_root_ns(struct mlx5_flow_steering *steering)
+{
+	struct fs_prio *prio;
+
+	steering->sniffer_tx_root_ns = create_root_ns(steering, FS_FT_SNIFFER_TX);
+	if (!steering->sniffer_tx_root_ns)
+		return -ENOMEM;
+
+	/* Create single prio */
+	prio = fs_create_prio(&steering->sniffer_tx_root_ns->ns, 0, 1);
+	if (IS_ERR(prio)) {
+		cleanup_root_ns(steering->sniffer_tx_root_ns);
+		return PTR_ERR(prio);
+	}
+	return 0;
+}
+
+static int init_sniffer_rx_root_ns(struct mlx5_flow_steering *steering)
+{
+	struct fs_prio *prio;
+
+	steering->sniffer_rx_root_ns = create_root_ns(steering, FS_FT_SNIFFER_RX);
+	if (!steering->sniffer_rx_root_ns)
+		return -ENOMEM;
+
+	/* Create single prio */
+	prio = fs_create_prio(&steering->sniffer_rx_root_ns->ns, 0, 1);
+	if (IS_ERR(prio)) {
+		cleanup_root_ns(steering->sniffer_rx_root_ns);
+		return PTR_ERR(prio);
+	}
+	return 0;
+}
+
+static int init_fdb_root_ns(struct mlx5_flow_steering *steering)
+{
+	struct fs_prio *prio;
+
+	steering->fdb_root_ns = create_root_ns(steering, FS_FT_FDB);
+	if (!steering->fdb_root_ns)
+		return -ENOMEM;
+
+	prio = fs_create_prio(&steering->fdb_root_ns->ns, 0, 2);
+	if (IS_ERR(prio))
+		goto out_err;
+
+	prio = fs_create_prio(&steering->fdb_root_ns->ns, 1, 1);
+	if (IS_ERR(prio))
+		goto out_err;
+
+	set_prio_attrs(steering->fdb_root_ns);
+	return 0;
+
+out_err:
+	cleanup_root_ns(steering->fdb_root_ns);
+	steering->fdb_root_ns = NULL;
+	return PTR_ERR(prio);
+}
+
+static int init_egress_acl_root_ns(struct mlx5_flow_steering *steering, int vport)
+{
+	struct fs_prio *prio;
+
+	steering->esw_egress_root_ns[vport] = create_root_ns(steering, FS_FT_ESW_EGRESS_ACL);
+	if (!steering->esw_egress_root_ns[vport])
+		return -ENOMEM;
+
+	/* create 1 prio*/
+	prio = fs_create_prio(&steering->esw_egress_root_ns[vport]->ns, 0, 1);
+	return PTR_ERR_OR_ZERO(prio);
+}
+
+static int init_ingress_acl_root_ns(struct mlx5_flow_steering *steering, int vport)
+{
+	struct fs_prio *prio;
+
+	steering->esw_ingress_root_ns[vport] = create_root_ns(steering, FS_FT_ESW_INGRESS_ACL);
+	if (!steering->esw_ingress_root_ns[vport])
+		return -ENOMEM;
+
+	/* create 1 prio*/
+	prio = fs_create_prio(&steering->esw_ingress_root_ns[vport]->ns, 0, 1);
+	return PTR_ERR_OR_ZERO(prio);
+}
+
+static int init_egress_acls_root_ns(struct mlx5_core_dev *dev)
+{
+	struct mlx5_flow_steering *steering = dev->priv.steering;
+	int err;
+	int i;
+
+	steering->esw_egress_root_ns = kcalloc(MLX5_TOTAL_VPORTS(dev),
+					       sizeof(*steering->esw_egress_root_ns),
+					       GFP_KERNEL);
+	if (!steering->esw_egress_root_ns)
+		return -ENOMEM;
+
+	for (i = 0; i < MLX5_TOTAL_VPORTS(dev); i++) {
+		err = init_egress_acl_root_ns(steering, i);
+		if (err)
+			goto cleanup_root_ns;
+	}
+
+	return 0;
+
+cleanup_root_ns:
+	for (i--; i >= 0; i--)
+		cleanup_root_ns(steering->esw_egress_root_ns[i]);
+	kfree(steering->esw_egress_root_ns);
+	steering->esw_egress_root_ns = NULL;
+	return err;
+}
+
+static int init_ingress_acls_root_ns(struct mlx5_core_dev *dev)
+{
+	struct mlx5_flow_steering *steering = dev->priv.steering;
+	int err;
+	int i;
+
+	steering->esw_ingress_root_ns = kcalloc(MLX5_TOTAL_VPORTS(dev),
+						sizeof(*steering->esw_ingress_root_ns),
+						GFP_KERNEL);
+	if (!steering->esw_ingress_root_ns)
+		return -ENOMEM;
+
+	for (i = 0; i < MLX5_TOTAL_VPORTS(dev); i++) {
+		err = init_ingress_acl_root_ns(steering, i);
+		if (err)
+			goto cleanup_root_ns;
+	}
+
+	return 0;
+
+cleanup_root_ns:
+	for (i--; i >= 0; i--)
+		cleanup_root_ns(steering->esw_ingress_root_ns[i]);
+	kfree(steering->esw_ingress_root_ns);
+	steering->esw_ingress_root_ns = NULL;
+	return err;
+}
+
+static int init_egress_root_ns(struct mlx5_flow_steering *steering)
+{
+	struct fs_prio *prio;
+
+	steering->egress_root_ns = create_root_ns(steering,
+						  FS_FT_NIC_TX);
+	if (!steering->egress_root_ns)
+		return -ENOMEM;
+
+	/* create 1 prio*/
+	prio = fs_create_prio(&steering->egress_root_ns->ns, 0, 1);
+	return PTR_ERR_OR_ZERO(prio);
+}
+
+int mlx5_init_fs(struct mlx5_core_dev *dev)
+{
+	struct mlx5_flow_steering *steering;
+	int err = 0;
+
+	err = mlx5_init_fc_stats(dev);
+	if (err)
+		return err;
+
+	steering = kzalloc(sizeof(*steering), GFP_KERNEL);
+	if (!steering)
+		return -ENOMEM;
+	steering->dev = dev;
+	dev->priv.steering = steering;
+
+	steering->fgs_cache = kmem_cache_create("mlx5_fs_fgs",
+						sizeof(struct mlx5_flow_group), 0,
+						0, NULL);
+	steering->ftes_cache = kmem_cache_create("mlx5_fs_ftes", sizeof(struct fs_fte), 0,
+						 0, NULL);
+	if (!steering->ftes_cache || !steering->fgs_cache) {
+		err = -ENOMEM;
+		goto err;
+	}
+
+	if ((((MLX5_CAP_GEN(dev, port_type) == MLX5_CAP_PORT_TYPE_ETH) &&
+	      (MLX5_CAP_GEN(dev, nic_flow_table))) ||
+	     ((MLX5_CAP_GEN(dev, port_type) == MLX5_CAP_PORT_TYPE_IB) &&
+	      MLX5_CAP_GEN(dev, ipoib_enhanced_offloads))) &&
+	    MLX5_CAP_FLOWTABLE_NIC_RX(dev, ft_support)) {
+		err = init_root_ns(steering);
+		if (err)
+			goto err;
+	}
+
+	if (MLX5_ESWITCH_MANAGER(dev)) {
+		if (MLX5_CAP_ESW_FLOWTABLE_FDB(dev, ft_support)) {
+			err = init_fdb_root_ns(steering);
+			if (err)
+				goto err;
+		}
+		if (MLX5_CAP_ESW_EGRESS_ACL(dev, ft_support)) {
+			err = init_egress_acls_root_ns(dev);
+			if (err)
+				goto err;
+		}
+		if (MLX5_CAP_ESW_INGRESS_ACL(dev, ft_support)) {
+			err = init_ingress_acls_root_ns(dev);
+			if (err)
+				goto err;
+		}
+	}
+
+	if (MLX5_CAP_FLOWTABLE_SNIFFER_RX(dev, ft_support)) {
+		err = init_sniffer_rx_root_ns(steering);
+		if (err)
+			goto err;
+	}
+
+	if (MLX5_CAP_FLOWTABLE_SNIFFER_TX(dev, ft_support)) {
+		err = init_sniffer_tx_root_ns(steering);
+		if (err)
+			goto err;
+	}
+
+	if (MLX5_IPSEC_DEV(dev)) {
+		err = init_egress_root_ns(steering);
+		if (err)
+			goto err;
+	}
+
+	return 0;
+err:
+	mlx5_cleanup_fs(dev);
+	return err;
+}
+
+int mlx5_fs_add_rx_underlay_qpn(struct mlx5_core_dev *dev, u32 underlay_qpn)
+{
+	struct mlx5_flow_root_namespace *root = dev->priv.steering->root_ns;
+	struct mlx5_ft_underlay_qp *new_uqp;
+	int err = 0;
+
+	new_uqp = kzalloc(sizeof(*new_uqp), GFP_KERNEL);
+	if (!new_uqp)
+		return -ENOMEM;
+
+	mutex_lock(&root->chain_lock);
+
+	if (!root->root_ft) {
+		err = -EINVAL;
+		goto update_ft_fail;
+	}
+
+	err = root->cmds->update_root_ft(dev, root->root_ft, underlay_qpn,
+					 false);
+	if (err) {
+		mlx5_core_warn(dev, "Failed adding underlay QPN (%u) to root FT err(%d)\n",
+			       underlay_qpn, err);
+		goto update_ft_fail;
+	}
+
+	new_uqp->qpn = underlay_qpn;
+	list_add_tail(&new_uqp->list, &root->underlay_qpns);
+
+	mutex_unlock(&root->chain_lock);
+
+	return 0;
+
+update_ft_fail:
+	mutex_unlock(&root->chain_lock);
+	kfree(new_uqp);
+	return err;
+}
+EXPORT_SYMBOL(mlx5_fs_add_rx_underlay_qpn);
+
+int mlx5_fs_remove_rx_underlay_qpn(struct mlx5_core_dev *dev, u32 underlay_qpn)
+{
+	struct mlx5_flow_root_namespace *root = dev->priv.steering->root_ns;
+	struct mlx5_ft_underlay_qp *uqp;
+	bool found = false;
+	int err = 0;
+
+	mutex_lock(&root->chain_lock);
+	list_for_each_entry(uqp, &root->underlay_qpns, list) {
+		if (uqp->qpn == underlay_qpn) {
+			found = true;
+			break;
+		}
+	}
+
+	if (!found) {
+		mlx5_core_warn(dev, "Failed finding underlay qp (%u) in qpn list\n",
+			       underlay_qpn);
+		err = -EINVAL;
+		goto out;
+	}
+
+	err = root->cmds->update_root_ft(dev, root->root_ft, underlay_qpn,
+					 true);
+	if (err)
+		mlx5_core_warn(dev, "Failed removing underlay QPN (%u) from root FT err(%d)\n",
+			       underlay_qpn, err);
+
+	list_del(&uqp->list);
+	mutex_unlock(&root->chain_lock);
+	kfree(uqp);
+
+	return 0;
+
+out:
+	mutex_unlock(&root->chain_lock);
+	return err;
+}
+EXPORT_SYMBOL(mlx5_fs_remove_rx_underlay_qpn);
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/fs_core.h b/drivers/net/ethernet/mellanox/mlx5/core/fs_core.h
new file mode 100644
index 000000000..ba62fbce2
--- /dev/null
+++ b/drivers/net/ethernet/mellanox/mlx5/core/fs_core.h
@@ -0,0 +1,286 @@
+/*
+ * Copyright (c) 2015, 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 _MLX5_FS_CORE_
+#define _MLX5_FS_CORE_
+
+#include <linux/refcount.h>
+#include <linux/mlx5/fs.h>
+#include <linux/rhashtable.h>
+
+enum fs_node_type {
+	FS_TYPE_NAMESPACE,
+	FS_TYPE_PRIO,
+	FS_TYPE_FLOW_TABLE,
+	FS_TYPE_FLOW_GROUP,
+	FS_TYPE_FLOW_ENTRY,
+	FS_TYPE_FLOW_DEST
+};
+
+enum fs_flow_table_type {
+	FS_FT_NIC_RX          = 0x0,
+	FS_FT_NIC_TX          = 0x1,
+	FS_FT_ESW_EGRESS_ACL  = 0x2,
+	FS_FT_ESW_INGRESS_ACL = 0x3,
+	FS_FT_FDB             = 0X4,
+	FS_FT_SNIFFER_RX	= 0X5,
+	FS_FT_SNIFFER_TX	= 0X6,
+	FS_FT_MAX_TYPE = FS_FT_SNIFFER_TX,
+};
+
+enum fs_flow_table_op_mod {
+	FS_FT_OP_MOD_NORMAL,
+	FS_FT_OP_MOD_LAG_DEMUX,
+};
+
+enum fs_fte_status {
+	FS_FTE_STATUS_EXISTING = 1UL << 0,
+};
+
+struct mlx5_flow_steering {
+	struct mlx5_core_dev *dev;
+	struct kmem_cache               *fgs_cache;
+	struct kmem_cache               *ftes_cache;
+	struct mlx5_flow_root_namespace *root_ns;
+	struct mlx5_flow_root_namespace *fdb_root_ns;
+	struct mlx5_flow_root_namespace **esw_egress_root_ns;
+	struct mlx5_flow_root_namespace **esw_ingress_root_ns;
+	struct mlx5_flow_root_namespace	*sniffer_tx_root_ns;
+	struct mlx5_flow_root_namespace	*sniffer_rx_root_ns;
+	struct mlx5_flow_root_namespace	*egress_root_ns;
+};
+
+struct fs_node {
+	struct list_head	list;
+	struct list_head	children;
+	enum fs_node_type	type;
+	struct fs_node		*parent;
+	struct fs_node		*root;
+	/* lock the node for writing and traversing */
+	struct rw_semaphore	lock;
+	refcount_t		refcount;
+	bool			active;
+	void			(*del_hw_func)(struct fs_node *);
+	void			(*del_sw_func)(struct fs_node *);
+	atomic_t		version;
+};
+
+struct mlx5_flow_rule {
+	struct fs_node				node;
+	struct mlx5_flow_destination		dest_attr;
+	/* next_ft should be accessed under chain_lock and only of
+	 * destination type is FWD_NEXT_fT.
+	 */
+	struct list_head			next_ft;
+	u32					sw_action;
+};
+
+struct mlx5_flow_handle {
+	int num_rules;
+	struct mlx5_flow_rule *rule[];
+};
+
+/* Type of children is mlx5_flow_group */
+struct mlx5_flow_table {
+	struct fs_node			node;
+	u32				id;
+	u16				vport;
+	unsigned int			max_fte;
+	unsigned int			level;
+	enum fs_flow_table_type		type;
+	enum fs_flow_table_op_mod	op_mod;
+	struct {
+		bool			active;
+		unsigned int		required_groups;
+		unsigned int		group_size;
+		unsigned int		num_groups;
+	} autogroup;
+	/* Protect fwd_rules */
+	struct mutex			lock;
+	/* FWD rules that point on this flow table */
+	struct list_head		fwd_rules;
+	u32				flags;
+	struct rhltable			fgs_hash;
+};
+
+struct mlx5_fc_cache {
+	u64 packets;
+	u64 bytes;
+	u64 lastuse;
+};
+
+struct mlx5_fc {
+	struct rb_node node;
+	struct list_head list;
+
+	/* last{packets,bytes} members are used when calculating the delta since
+	 * last reading
+	 */
+	u64 lastpackets;
+	u64 lastbytes;
+
+	u32 id;
+	bool deleted;
+	bool aging;
+
+	struct mlx5_fc_cache cache ____cacheline_aligned_in_smp;
+};
+
+struct mlx5_ft_underlay_qp {
+	struct list_head list;
+	u32 qpn;
+};
+
+#define MLX5_FTE_MATCH_PARAM_RESERVED	reserved_at_800
+/* Calculate the fte_match_param length and without the reserved length.
+ * Make sure the reserved field is the last.
+ */
+#define MLX5_ST_SZ_DW_MATCH_PARAM					    \
+	((MLX5_BYTE_OFF(fte_match_param, MLX5_FTE_MATCH_PARAM_RESERVED) / sizeof(u32)) + \
+	 BUILD_BUG_ON_ZERO(MLX5_ST_SZ_BYTES(fte_match_param) !=		     \
+			   MLX5_FLD_SZ_BYTES(fte_match_param,		     \
+					     MLX5_FTE_MATCH_PARAM_RESERVED) +\
+			   MLX5_BYTE_OFF(fte_match_param,		     \
+					 MLX5_FTE_MATCH_PARAM_RESERVED)))
+
+/* Type of children is mlx5_flow_rule */
+struct fs_fte {
+	struct fs_node			node;
+	u32				val[MLX5_ST_SZ_DW_MATCH_PARAM];
+	u32				dests_size;
+	u32				index;
+	struct mlx5_flow_act		action;
+	enum fs_fte_status		status;
+	struct mlx5_fc			*counter;
+	struct rhash_head		hash;
+};
+
+/* Type of children is mlx5_flow_table/namespace */
+struct fs_prio {
+	struct fs_node			node;
+	unsigned int			num_levels;
+	unsigned int			start_level;
+	unsigned int			prio;
+	unsigned int			num_ft;
+};
+
+/* Type of children is fs_prio */
+struct mlx5_flow_namespace {
+	/* parent == NULL => root ns */
+	struct	fs_node			node;
+};
+
+struct mlx5_flow_group_mask {
+	u8	match_criteria_enable;
+	u32	match_criteria[MLX5_ST_SZ_DW_MATCH_PARAM];
+};
+
+/* Type of children is fs_fte */
+struct mlx5_flow_group {
+	struct fs_node			node;
+	struct mlx5_flow_group_mask	mask;
+	u32				start_index;
+	u32				max_ftes;
+	struct ida			fte_allocator;
+	u32				id;
+	struct rhashtable		ftes_hash;
+	struct rhlist_head		hash;
+};
+
+struct mlx5_flow_root_namespace {
+	struct mlx5_flow_namespace	ns;
+	enum   fs_flow_table_type	table_type;
+	struct mlx5_core_dev		*dev;
+	struct mlx5_flow_table		*root_ft;
+	/* Should be held when chaining flow tables */
+	struct mutex			chain_lock;
+	struct list_head		underlay_qpns;
+	const struct mlx5_flow_cmds	*cmds;
+};
+
+int mlx5_init_fc_stats(struct mlx5_core_dev *dev);
+void mlx5_cleanup_fc_stats(struct mlx5_core_dev *dev);
+void mlx5_fc_queue_stats_work(struct mlx5_core_dev *dev,
+			      struct delayed_work *dwork,
+			      unsigned long delay);
+void mlx5_fc_update_sampling_interval(struct mlx5_core_dev *dev,
+				      unsigned long interval);
+
+int mlx5_init_fs(struct mlx5_core_dev *dev);
+void mlx5_cleanup_fs(struct mlx5_core_dev *dev);
+
+#define fs_get_obj(v, _node)  {v = container_of((_node), typeof(*v), node); }
+
+#define fs_list_for_each_entry(pos, root)		\
+	list_for_each_entry(pos, root, node.list)
+
+#define fs_list_for_each_entry_safe(pos, tmp, root)		\
+	list_for_each_entry_safe(pos, tmp, root, node.list)
+
+#define fs_for_each_ns_or_ft_reverse(pos, prio)				\
+	list_for_each_entry_reverse(pos, &(prio)->node.children, list)
+
+#define fs_for_each_ns_or_ft(pos, prio)					\
+	list_for_each_entry(pos, (&(prio)->node.children), list)
+
+#define fs_for_each_prio(pos, ns)			\
+	fs_list_for_each_entry(pos, &(ns)->node.children)
+
+#define fs_for_each_ns(pos, prio)			\
+	fs_list_for_each_entry(pos, &(prio)->node.children)
+
+#define fs_for_each_ft(pos, prio)			\
+	fs_list_for_each_entry(pos, &(prio)->node.children)
+
+#define fs_for_each_ft_safe(pos, tmp, prio)			\
+	fs_list_for_each_entry_safe(pos, tmp, &(prio)->node.children)
+
+#define fs_for_each_fg(pos, ft)			\
+	fs_list_for_each_entry(pos, &(ft)->node.children)
+
+#define fs_for_each_fte(pos, fg)			\
+	fs_list_for_each_entry(pos, &(fg)->node.children)
+
+#define fs_for_each_dst(pos, fte)			\
+	fs_list_for_each_entry(pos, &(fte)->node.children)
+
+#define MLX5_CAP_FLOWTABLE_TYPE(mdev, cap, type) (		\
+	(type == FS_FT_NIC_RX) ? MLX5_CAP_FLOWTABLE_NIC_RX(mdev, cap) :		\
+	(type == FS_FT_ESW_EGRESS_ACL) ? MLX5_CAP_ESW_EGRESS_ACL(mdev, cap) :		\
+	(type == FS_FT_ESW_INGRESS_ACL) ? MLX5_CAP_ESW_INGRESS_ACL(mdev, cap) :		\
+	(type == FS_FT_FDB) ? MLX5_CAP_ESW_FLOWTABLE_FDB(mdev, cap) :		\
+	(type == FS_FT_SNIFFER_RX) ? MLX5_CAP_FLOWTABLE_SNIFFER_RX(mdev, cap) :		\
+	(type == FS_FT_SNIFFER_TX) ? MLX5_CAP_FLOWTABLE_SNIFFER_TX(mdev, cap) :		\
+	(BUILD_BUG_ON_ZERO(FS_FT_SNIFFER_TX != FS_FT_MAX_TYPE))\
+	)
+
+#endif
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/fs_counters.c b/drivers/net/ethernet/mellanox/mlx5/core/fs_counters.c
new file mode 100644
index 000000000..808ddd732
--- /dev/null
+++ b/drivers/net/ethernet/mellanox/mlx5/core/fs_counters.c
@@ -0,0 +1,360 @@
+/*
+ * Copyright (c) 2016, 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/mlx5/driver.h>
+#include <linux/mlx5/fs.h>
+#include <linux/rbtree.h>
+#include "mlx5_core.h"
+#include "fs_core.h"
+#include "fs_cmd.h"
+
+#define MLX5_FC_STATS_PERIOD msecs_to_jiffies(1000)
+/* Max number of counters to query in bulk read is 32K */
+#define MLX5_SW_MAX_COUNTERS_BULK BIT(15)
+
+/* locking scheme:
+ *
+ * It is the responsibility of the user to prevent concurrent calls or bad
+ * ordering to mlx5_fc_create(), mlx5_fc_destroy() and accessing a reference
+ * to struct mlx5_fc.
+ * e.g en_tc.c is protected by RTNL lock of its caller, and will never call a
+ * dump (access to struct mlx5_fc) after a counter is destroyed.
+ *
+ * access to counter list:
+ * - create (user context)
+ *   - mlx5_fc_create() only adds to an addlist to be used by
+ *     mlx5_fc_stats_query_work(). addlist is protected by a spinlock.
+ *   - spawn thread to do the actual destroy
+ *
+ * - destroy (user context)
+ *   - mark a counter as deleted
+ *   - spawn thread to do the actual del
+ *
+ * - dump (user context)
+ *   user should not call dump after destroy
+ *
+ * - query (single thread workqueue context)
+ *   destroy/dump - no conflict (see destroy)
+ *   query/dump - packets and bytes might be inconsistent (since update is not
+ *                atomic)
+ *   query/create - no conflict (see create)
+ *   since every create/destroy spawn the work, only after necessary time has
+ *   elapsed, the thread will actually query the hardware.
+ */
+
+static void mlx5_fc_stats_insert(struct rb_root *root, struct mlx5_fc *counter)
+{
+	struct rb_node **new = &root->rb_node;
+	struct rb_node *parent = NULL;
+
+	while (*new) {
+		struct mlx5_fc *this = rb_entry(*new, struct mlx5_fc, node);
+		int result = counter->id - this->id;
+
+		parent = *new;
+		if (result < 0)
+			new = &((*new)->rb_left);
+		else
+			new = &((*new)->rb_right);
+	}
+
+	/* Add new node and rebalance tree. */
+	rb_link_node(&counter->node, parent, new);
+	rb_insert_color(&counter->node, root);
+}
+
+/* The function returns the last node that was queried so the caller
+ * function can continue calling it till all counters are queried.
+ */
+static struct rb_node *mlx5_fc_stats_query(struct mlx5_core_dev *dev,
+					   struct mlx5_fc *first,
+					   u32 last_id)
+{
+	struct mlx5_cmd_fc_bulk *b;
+	struct rb_node *node = NULL;
+	u32 afirst_id;
+	int num;
+	int err;
+
+	int max_bulk = min_t(int, MLX5_SW_MAX_COUNTERS_BULK,
+			     (1 << MLX5_CAP_GEN(dev, log_max_flow_counter_bulk)));
+
+	/* first id must be aligned to 4 when using bulk query */
+	afirst_id = first->id & ~0x3;
+
+	/* number of counters to query inc. the last counter */
+	num = ALIGN(last_id - afirst_id + 1, 4);
+	if (num > max_bulk) {
+		num = max_bulk;
+		last_id = afirst_id + num - 1;
+	}
+
+	b = mlx5_cmd_fc_bulk_alloc(dev, afirst_id, num);
+	if (!b) {
+		mlx5_core_err(dev, "Error allocating resources for bulk query\n");
+		return NULL;
+	}
+
+	err = mlx5_cmd_fc_bulk_query(dev, b);
+	if (err) {
+		mlx5_core_err(dev, "Error doing bulk query: %d\n", err);
+		goto out;
+	}
+
+	for (node = &first->node; node; node = rb_next(node)) {
+		struct mlx5_fc *counter = rb_entry(node, struct mlx5_fc, node);
+		struct mlx5_fc_cache *c = &counter->cache;
+		u64 packets;
+		u64 bytes;
+
+		if (counter->id > last_id)
+			break;
+
+		mlx5_cmd_fc_bulk_get(dev, b,
+				     counter->id, &packets, &bytes);
+
+		if (c->packets == packets)
+			continue;
+
+		c->packets = packets;
+		c->bytes = bytes;
+		c->lastuse = jiffies;
+	}
+
+out:
+	mlx5_cmd_fc_bulk_free(b);
+
+	return node;
+}
+
+static void mlx5_fc_stats_work(struct work_struct *work)
+{
+	struct mlx5_core_dev *dev = container_of(work, struct mlx5_core_dev,
+						 priv.fc_stats.work.work);
+	struct mlx5_fc_stats *fc_stats = &dev->priv.fc_stats;
+	unsigned long now = jiffies;
+	struct mlx5_fc *counter = NULL;
+	struct mlx5_fc *last = NULL;
+	struct rb_node *node;
+	LIST_HEAD(tmplist);
+
+	spin_lock(&fc_stats->addlist_lock);
+
+	list_splice_tail_init(&fc_stats->addlist, &tmplist);
+
+	if (!list_empty(&tmplist) || !RB_EMPTY_ROOT(&fc_stats->counters))
+		queue_delayed_work(fc_stats->wq, &fc_stats->work,
+				   fc_stats->sampling_interval);
+
+	spin_unlock(&fc_stats->addlist_lock);
+
+	list_for_each_entry(counter, &tmplist, list)
+		mlx5_fc_stats_insert(&fc_stats->counters, counter);
+
+	node = rb_first(&fc_stats->counters);
+	while (node) {
+		counter = rb_entry(node, struct mlx5_fc, node);
+
+		node = rb_next(node);
+
+		if (counter->deleted) {
+			rb_erase(&counter->node, &fc_stats->counters);
+
+			mlx5_cmd_fc_free(dev, counter->id);
+
+			kfree(counter);
+			continue;
+		}
+
+		last = counter;
+	}
+
+	if (time_before(now, fc_stats->next_query) || !last)
+		return;
+
+	node = rb_first(&fc_stats->counters);
+	while (node) {
+		counter = rb_entry(node, struct mlx5_fc, node);
+
+		node = mlx5_fc_stats_query(dev, counter, last->id);
+	}
+
+	fc_stats->next_query = now + fc_stats->sampling_interval;
+}
+
+struct mlx5_fc *mlx5_fc_create(struct mlx5_core_dev *dev, bool aging)
+{
+	struct mlx5_fc_stats *fc_stats = &dev->priv.fc_stats;
+	struct mlx5_fc *counter;
+	int err;
+
+	counter = kzalloc(sizeof(*counter), GFP_KERNEL);
+	if (!counter)
+		return ERR_PTR(-ENOMEM);
+
+	err = mlx5_cmd_fc_alloc(dev, &counter->id);
+	if (err)
+		goto err_out;
+
+	if (aging) {
+		counter->cache.lastuse = jiffies;
+		counter->aging = true;
+
+		spin_lock(&fc_stats->addlist_lock);
+		list_add(&counter->list, &fc_stats->addlist);
+		spin_unlock(&fc_stats->addlist_lock);
+
+		mod_delayed_work(fc_stats->wq, &fc_stats->work, 0);
+	}
+
+	return counter;
+
+err_out:
+	kfree(counter);
+
+	return ERR_PTR(err);
+}
+EXPORT_SYMBOL(mlx5_fc_create);
+
+void mlx5_fc_destroy(struct mlx5_core_dev *dev, struct mlx5_fc *counter)
+{
+	struct mlx5_fc_stats *fc_stats = &dev->priv.fc_stats;
+
+	if (!counter)
+		return;
+
+	if (counter->aging) {
+		counter->deleted = true;
+		mod_delayed_work(fc_stats->wq, &fc_stats->work, 0);
+		return;
+	}
+
+	mlx5_cmd_fc_free(dev, counter->id);
+	kfree(counter);
+}
+EXPORT_SYMBOL(mlx5_fc_destroy);
+
+int mlx5_init_fc_stats(struct mlx5_core_dev *dev)
+{
+	struct mlx5_fc_stats *fc_stats = &dev->priv.fc_stats;
+
+	fc_stats->counters = RB_ROOT;
+	INIT_LIST_HEAD(&fc_stats->addlist);
+	spin_lock_init(&fc_stats->addlist_lock);
+
+	fc_stats->wq = create_singlethread_workqueue("mlx5_fc");
+	if (!fc_stats->wq)
+		return -ENOMEM;
+
+	fc_stats->sampling_interval = MLX5_FC_STATS_PERIOD;
+	INIT_DELAYED_WORK(&fc_stats->work, mlx5_fc_stats_work);
+
+	return 0;
+}
+
+void mlx5_cleanup_fc_stats(struct mlx5_core_dev *dev)
+{
+	struct mlx5_fc_stats *fc_stats = &dev->priv.fc_stats;
+	struct mlx5_fc *counter;
+	struct mlx5_fc *tmp;
+	struct rb_node *node;
+
+	cancel_delayed_work_sync(&dev->priv.fc_stats.work);
+	destroy_workqueue(dev->priv.fc_stats.wq);
+	dev->priv.fc_stats.wq = NULL;
+
+	list_for_each_entry_safe(counter, tmp, &fc_stats->addlist, list) {
+		list_del(&counter->list);
+
+		mlx5_cmd_fc_free(dev, counter->id);
+
+		kfree(counter);
+	}
+
+	node = rb_first(&fc_stats->counters);
+	while (node) {
+		counter = rb_entry(node, struct mlx5_fc, node);
+
+		node = rb_next(node);
+
+		rb_erase(&counter->node, &fc_stats->counters);
+
+		mlx5_cmd_fc_free(dev, counter->id);
+
+		kfree(counter);
+	}
+}
+
+int mlx5_fc_query(struct mlx5_core_dev *dev, struct mlx5_fc *counter,
+		  u64 *packets, u64 *bytes)
+{
+	return mlx5_cmd_fc_query(dev, counter->id, packets, bytes);
+}
+EXPORT_SYMBOL(mlx5_fc_query);
+
+u64 mlx5_fc_query_lastuse(struct mlx5_fc *counter)
+{
+	return counter->cache.lastuse;
+}
+
+void mlx5_fc_query_cached(struct mlx5_fc *counter,
+			  u64 *bytes, u64 *packets, u64 *lastuse)
+{
+	struct mlx5_fc_cache c;
+
+	c = counter->cache;
+
+	*bytes = c.bytes - counter->lastbytes;
+	*packets = c.packets - counter->lastpackets;
+	*lastuse = c.lastuse;
+
+	counter->lastbytes = c.bytes;
+	counter->lastpackets = c.packets;
+}
+
+void mlx5_fc_queue_stats_work(struct mlx5_core_dev *dev,
+			      struct delayed_work *dwork,
+			      unsigned long delay)
+{
+	struct mlx5_fc_stats *fc_stats = &dev->priv.fc_stats;
+
+	queue_delayed_work(fc_stats->wq, dwork, delay);
+}
+
+void mlx5_fc_update_sampling_interval(struct mlx5_core_dev *dev,
+				      unsigned long interval)
+{
+	struct mlx5_fc_stats *fc_stats = &dev->priv.fc_stats;
+
+	fc_stats->sampling_interval = min_t(unsigned long, interval,
+					    fc_stats->sampling_interval);
+}
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/fw.c b/drivers/net/ethernet/mellanox/mlx5/core/fw.c
new file mode 100644
index 000000000..41ad24f0d
--- /dev/null
+++ b/drivers/net/ethernet/mellanox/mlx5/core/fw.c
@@ -0,0 +1,527 @@
+/*
+ * Copyright (c) 2013-2015, 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/mlx5/driver.h>
+#include <linux/mlx5/cmd.h>
+#include <linux/mlx5/eswitch.h>
+#include <linux/module.h>
+#include "mlx5_core.h"
+#include "../../mlxfw/mlxfw.h"
+
+static int mlx5_cmd_query_adapter(struct mlx5_core_dev *dev, u32 *out,
+				  int outlen)
+{
+	u32 in[MLX5_ST_SZ_DW(query_adapter_in)] = {0};
+
+	MLX5_SET(query_adapter_in, in, opcode, MLX5_CMD_OP_QUERY_ADAPTER);
+	return mlx5_cmd_exec(dev, in, sizeof(in), out, outlen);
+}
+
+int mlx5_query_board_id(struct mlx5_core_dev *dev)
+{
+	u32 *out;
+	int outlen = MLX5_ST_SZ_BYTES(query_adapter_out);
+	int err;
+
+	out = kzalloc(outlen, GFP_KERNEL);
+	if (!out)
+		return -ENOMEM;
+
+	err = mlx5_cmd_query_adapter(dev, out, outlen);
+	if (err)
+		goto out;
+
+	memcpy(dev->board_id,
+	       MLX5_ADDR_OF(query_adapter_out, out,
+			    query_adapter_struct.vsd_contd_psid),
+	       MLX5_FLD_SZ_BYTES(query_adapter_out,
+				 query_adapter_struct.vsd_contd_psid));
+
+out:
+	kfree(out);
+	return err;
+}
+
+int mlx5_core_query_vendor_id(struct mlx5_core_dev *mdev, u32 *vendor_id)
+{
+	u32 *out;
+	int outlen = MLX5_ST_SZ_BYTES(query_adapter_out);
+	int err;
+
+	out = kzalloc(outlen, GFP_KERNEL);
+	if (!out)
+		return -ENOMEM;
+
+	err = mlx5_cmd_query_adapter(mdev, out, outlen);
+	if (err)
+		goto out;
+
+	*vendor_id = MLX5_GET(query_adapter_out, out,
+			      query_adapter_struct.ieee_vendor_id);
+out:
+	kfree(out);
+	return err;
+}
+EXPORT_SYMBOL(mlx5_core_query_vendor_id);
+
+static int mlx5_get_pcam_reg(struct mlx5_core_dev *dev)
+{
+	return mlx5_query_pcam_reg(dev, dev->caps.pcam,
+				   MLX5_PCAM_FEATURE_ENHANCED_FEATURES,
+				   MLX5_PCAM_REGS_5000_TO_507F);
+}
+
+static int mlx5_get_mcam_reg(struct mlx5_core_dev *dev)
+{
+	return mlx5_query_mcam_reg(dev, dev->caps.mcam,
+				   MLX5_MCAM_FEATURE_ENHANCED_FEATURES,
+				   MLX5_MCAM_REGS_FIRST_128);
+}
+
+static int mlx5_get_qcam_reg(struct mlx5_core_dev *dev)
+{
+	return mlx5_query_qcam_reg(dev, dev->caps.qcam,
+				   MLX5_QCAM_FEATURE_ENHANCED_FEATURES,
+				   MLX5_QCAM_REGS_FIRST_128);
+}
+
+int mlx5_query_hca_caps(struct mlx5_core_dev *dev)
+{
+	int err;
+
+	err = mlx5_core_get_caps(dev, MLX5_CAP_GENERAL);
+	if (err)
+		return err;
+
+	if (MLX5_CAP_GEN(dev, eth_net_offloads)) {
+		err = mlx5_core_get_caps(dev, MLX5_CAP_ETHERNET_OFFLOADS);
+		if (err)
+			return err;
+	}
+
+	if (MLX5_CAP_GEN(dev, ipoib_enhanced_offloads)) {
+		err = mlx5_core_get_caps(dev, MLX5_CAP_IPOIB_ENHANCED_OFFLOADS);
+		if (err)
+			return err;
+	}
+
+	if (MLX5_CAP_GEN(dev, pg)) {
+		err = mlx5_core_get_caps(dev, MLX5_CAP_ODP);
+		if (err)
+			return err;
+	}
+
+	if (MLX5_CAP_GEN(dev, atomic)) {
+		err = mlx5_core_get_caps(dev, MLX5_CAP_ATOMIC);
+		if (err)
+			return err;
+	}
+
+	if (MLX5_CAP_GEN(dev, roce)) {
+		err = mlx5_core_get_caps(dev, MLX5_CAP_ROCE);
+		if (err)
+			return err;
+	}
+
+	if (MLX5_CAP_GEN(dev, nic_flow_table) ||
+	    MLX5_CAP_GEN(dev, ipoib_enhanced_offloads)) {
+		err = mlx5_core_get_caps(dev, MLX5_CAP_FLOW_TABLE);
+		if (err)
+			return err;
+	}
+
+	if (MLX5_CAP_GEN(dev, vport_group_manager) &&
+	    MLX5_ESWITCH_MANAGER(dev)) {
+		err = mlx5_core_get_caps(dev, MLX5_CAP_ESWITCH_FLOW_TABLE);
+		if (err)
+			return err;
+	}
+
+	if (MLX5_ESWITCH_MANAGER(dev)) {
+		err = mlx5_core_get_caps(dev, MLX5_CAP_ESWITCH);
+		if (err)
+			return err;
+	}
+
+	if (MLX5_CAP_GEN(dev, vector_calc)) {
+		err = mlx5_core_get_caps(dev, MLX5_CAP_VECTOR_CALC);
+		if (err)
+			return err;
+	}
+
+	if (MLX5_CAP_GEN(dev, qos)) {
+		err = mlx5_core_get_caps(dev, MLX5_CAP_QOS);
+		if (err)
+			return err;
+	}
+
+	if (MLX5_CAP_GEN(dev, debug))
+		mlx5_core_get_caps(dev, MLX5_CAP_DEBUG);
+
+	if (MLX5_CAP_GEN(dev, pcam_reg))
+		mlx5_get_pcam_reg(dev);
+
+	if (MLX5_CAP_GEN(dev, mcam_reg))
+		mlx5_get_mcam_reg(dev);
+
+	if (MLX5_CAP_GEN(dev, qcam_reg))
+		mlx5_get_qcam_reg(dev);
+
+	if (MLX5_CAP_GEN(dev, device_memory)) {
+		err = mlx5_core_get_caps(dev, MLX5_CAP_DEV_MEM);
+		if (err)
+			return err;
+	}
+
+	return 0;
+}
+
+int mlx5_cmd_init_hca(struct mlx5_core_dev *dev, uint32_t *sw_owner_id)
+{
+	u32 out[MLX5_ST_SZ_DW(init_hca_out)] = {0};
+	u32 in[MLX5_ST_SZ_DW(init_hca_in)]   = {0};
+	int i;
+
+	MLX5_SET(init_hca_in, in, opcode, MLX5_CMD_OP_INIT_HCA);
+
+	if (MLX5_CAP_GEN(dev, sw_owner_id)) {
+		for (i = 0; i < 4; i++)
+			MLX5_ARRAY_SET(init_hca_in, in, sw_owner_id, i,
+				       sw_owner_id[i]);
+	}
+
+	return mlx5_cmd_exec(dev, in, sizeof(in), out, sizeof(out));
+}
+
+int mlx5_cmd_teardown_hca(struct mlx5_core_dev *dev)
+{
+	u32 out[MLX5_ST_SZ_DW(teardown_hca_out)] = {0};
+	u32 in[MLX5_ST_SZ_DW(teardown_hca_in)]   = {0};
+
+	MLX5_SET(teardown_hca_in, in, opcode, MLX5_CMD_OP_TEARDOWN_HCA);
+	return mlx5_cmd_exec(dev, in, sizeof(in), out, sizeof(out));
+}
+
+int mlx5_cmd_force_teardown_hca(struct mlx5_core_dev *dev)
+{
+	u32 out[MLX5_ST_SZ_DW(teardown_hca_out)] = {0};
+	u32 in[MLX5_ST_SZ_DW(teardown_hca_in)] = {0};
+	int force_state;
+	int ret;
+
+	if (!MLX5_CAP_GEN(dev, force_teardown)) {
+		mlx5_core_dbg(dev, "force teardown is not supported in the firmware\n");
+		return -EOPNOTSUPP;
+	}
+
+	MLX5_SET(teardown_hca_in, in, opcode, MLX5_CMD_OP_TEARDOWN_HCA);
+	MLX5_SET(teardown_hca_in, in, profile, MLX5_TEARDOWN_HCA_IN_PROFILE_FORCE_CLOSE);
+
+	ret = mlx5_cmd_exec_polling(dev, in, sizeof(in), out, sizeof(out));
+	if (ret)
+		return ret;
+
+	force_state = MLX5_GET(teardown_hca_out, out, force_state);
+	if (force_state == MLX5_TEARDOWN_HCA_OUT_FORCE_STATE_FAIL) {
+		mlx5_core_warn(dev, "teardown with force mode failed, doing normal teardown\n");
+		return -EIO;
+	}
+
+	return 0;
+}
+
+enum mlxsw_reg_mcc_instruction {
+	MLX5_REG_MCC_INSTRUCTION_LOCK_UPDATE_HANDLE = 0x01,
+	MLX5_REG_MCC_INSTRUCTION_RELEASE_UPDATE_HANDLE = 0x02,
+	MLX5_REG_MCC_INSTRUCTION_UPDATE_COMPONENT = 0x03,
+	MLX5_REG_MCC_INSTRUCTION_VERIFY_COMPONENT = 0x04,
+	MLX5_REG_MCC_INSTRUCTION_ACTIVATE = 0x06,
+	MLX5_REG_MCC_INSTRUCTION_CANCEL = 0x08,
+};
+
+static int mlx5_reg_mcc_set(struct mlx5_core_dev *dev,
+			    enum mlxsw_reg_mcc_instruction instr,
+			    u16 component_index, u32 update_handle,
+			    u32 component_size)
+{
+	u32 out[MLX5_ST_SZ_DW(mcc_reg)];
+	u32 in[MLX5_ST_SZ_DW(mcc_reg)];
+
+	memset(in, 0, sizeof(in));
+
+	MLX5_SET(mcc_reg, in, instruction, instr);
+	MLX5_SET(mcc_reg, in, component_index, component_index);
+	MLX5_SET(mcc_reg, in, update_handle, update_handle);
+	MLX5_SET(mcc_reg, in, component_size, component_size);
+
+	return mlx5_core_access_reg(dev, in, sizeof(in), out,
+				    sizeof(out), MLX5_REG_MCC, 0, 1);
+}
+
+static int mlx5_reg_mcc_query(struct mlx5_core_dev *dev,
+			      u32 *update_handle, u8 *error_code,
+			      u8 *control_state)
+{
+	u32 out[MLX5_ST_SZ_DW(mcc_reg)];
+	u32 in[MLX5_ST_SZ_DW(mcc_reg)];
+	int err;
+
+	memset(in, 0, sizeof(in));
+	memset(out, 0, sizeof(out));
+	MLX5_SET(mcc_reg, in, update_handle, *update_handle);
+
+	err = mlx5_core_access_reg(dev, in, sizeof(in), out,
+				   sizeof(out), MLX5_REG_MCC, 0, 0);
+	if (err)
+		goto out;
+
+	*update_handle = MLX5_GET(mcc_reg, out, update_handle);
+	*error_code = MLX5_GET(mcc_reg, out, error_code);
+	*control_state = MLX5_GET(mcc_reg, out, control_state);
+
+out:
+	return err;
+}
+
+static int mlx5_reg_mcda_set(struct mlx5_core_dev *dev,
+			     u32 update_handle,
+			     u32 offset, u16 size,
+			     u8 *data)
+{
+	int err, in_size = MLX5_ST_SZ_BYTES(mcda_reg) + size;
+	u32 out[MLX5_ST_SZ_DW(mcda_reg)];
+	int i, j, dw_size = size >> 2;
+	__be32 data_element;
+	u32 *in;
+
+	in = kzalloc(in_size, GFP_KERNEL);
+	if (!in)
+		return -ENOMEM;
+
+	MLX5_SET(mcda_reg, in, update_handle, update_handle);
+	MLX5_SET(mcda_reg, in, offset, offset);
+	MLX5_SET(mcda_reg, in, size, size);
+
+	for (i = 0; i < dw_size; i++) {
+		j = i * 4;
+		data_element = htonl(*(u32 *)&data[j]);
+		memcpy(MLX5_ADDR_OF(mcda_reg, in, data) + j, &data_element, 4);
+	}
+
+	err = mlx5_core_access_reg(dev, in, in_size, out,
+				   sizeof(out), MLX5_REG_MCDA, 0, 1);
+	kfree(in);
+	return err;
+}
+
+static int mlx5_reg_mcqi_query(struct mlx5_core_dev *dev,
+			       u16 component_index,
+			       u32 *max_component_size,
+			       u8 *log_mcda_word_size,
+			       u16 *mcda_max_write_size)
+{
+	u32 out[MLX5_ST_SZ_DW(mcqi_reg) + MLX5_ST_SZ_DW(mcqi_cap)];
+	int offset = MLX5_ST_SZ_DW(mcqi_reg);
+	u32 in[MLX5_ST_SZ_DW(mcqi_reg)];
+	int err;
+
+	memset(in, 0, sizeof(in));
+	memset(out, 0, sizeof(out));
+
+	MLX5_SET(mcqi_reg, in, component_index, component_index);
+	MLX5_SET(mcqi_reg, in, data_size, MLX5_ST_SZ_BYTES(mcqi_cap));
+
+	err = mlx5_core_access_reg(dev, in, sizeof(in), out,
+				   sizeof(out), MLX5_REG_MCQI, 0, 0);
+	if (err)
+		goto out;
+
+	*max_component_size = MLX5_GET(mcqi_cap, out + offset, max_component_size);
+	*log_mcda_word_size = MLX5_GET(mcqi_cap, out + offset, log_mcda_word_size);
+	*mcda_max_write_size = MLX5_GET(mcqi_cap, out + offset, mcda_max_write_size);
+
+out:
+	return err;
+}
+
+struct mlx5_mlxfw_dev {
+	struct mlxfw_dev mlxfw_dev;
+	struct mlx5_core_dev *mlx5_core_dev;
+};
+
+static int mlx5_component_query(struct mlxfw_dev *mlxfw_dev,
+				u16 component_index, u32 *p_max_size,
+				u8 *p_align_bits, u16 *p_max_write_size)
+{
+	struct mlx5_mlxfw_dev *mlx5_mlxfw_dev =
+		container_of(mlxfw_dev, struct mlx5_mlxfw_dev, mlxfw_dev);
+	struct mlx5_core_dev *dev = mlx5_mlxfw_dev->mlx5_core_dev;
+
+	return mlx5_reg_mcqi_query(dev, component_index, p_max_size,
+				   p_align_bits, p_max_write_size);
+}
+
+static int mlx5_fsm_lock(struct mlxfw_dev *mlxfw_dev, u32 *fwhandle)
+{
+	struct mlx5_mlxfw_dev *mlx5_mlxfw_dev =
+		container_of(mlxfw_dev, struct mlx5_mlxfw_dev, mlxfw_dev);
+	struct mlx5_core_dev *dev = mlx5_mlxfw_dev->mlx5_core_dev;
+	u8 control_state, error_code;
+	int err;
+
+	*fwhandle = 0;
+	err = mlx5_reg_mcc_query(dev, fwhandle, &error_code, &control_state);
+	if (err)
+		return err;
+
+	if (control_state != MLXFW_FSM_STATE_IDLE)
+		return -EBUSY;
+
+	return mlx5_reg_mcc_set(dev, MLX5_REG_MCC_INSTRUCTION_LOCK_UPDATE_HANDLE,
+				0, *fwhandle, 0);
+}
+
+static int mlx5_fsm_component_update(struct mlxfw_dev *mlxfw_dev, u32 fwhandle,
+				     u16 component_index, u32 component_size)
+{
+	struct mlx5_mlxfw_dev *mlx5_mlxfw_dev =
+		container_of(mlxfw_dev, struct mlx5_mlxfw_dev, mlxfw_dev);
+	struct mlx5_core_dev *dev = mlx5_mlxfw_dev->mlx5_core_dev;
+
+	return mlx5_reg_mcc_set(dev, MLX5_REG_MCC_INSTRUCTION_UPDATE_COMPONENT,
+				component_index, fwhandle, component_size);
+}
+
+static int mlx5_fsm_block_download(struct mlxfw_dev *mlxfw_dev, u32 fwhandle,
+				   u8 *data, u16 size, u32 offset)
+{
+	struct mlx5_mlxfw_dev *mlx5_mlxfw_dev =
+		container_of(mlxfw_dev, struct mlx5_mlxfw_dev, mlxfw_dev);
+	struct mlx5_core_dev *dev = mlx5_mlxfw_dev->mlx5_core_dev;
+
+	return mlx5_reg_mcda_set(dev, fwhandle, offset, size, data);
+}
+
+static int mlx5_fsm_component_verify(struct mlxfw_dev *mlxfw_dev, u32 fwhandle,
+				     u16 component_index)
+{
+	struct mlx5_mlxfw_dev *mlx5_mlxfw_dev =
+		container_of(mlxfw_dev, struct mlx5_mlxfw_dev, mlxfw_dev);
+	struct mlx5_core_dev *dev = mlx5_mlxfw_dev->mlx5_core_dev;
+
+	return mlx5_reg_mcc_set(dev, MLX5_REG_MCC_INSTRUCTION_VERIFY_COMPONENT,
+				component_index, fwhandle, 0);
+}
+
+static int mlx5_fsm_activate(struct mlxfw_dev *mlxfw_dev, u32 fwhandle)
+{
+	struct mlx5_mlxfw_dev *mlx5_mlxfw_dev =
+		container_of(mlxfw_dev, struct mlx5_mlxfw_dev, mlxfw_dev);
+	struct mlx5_core_dev *dev = mlx5_mlxfw_dev->mlx5_core_dev;
+
+	return mlx5_reg_mcc_set(dev, MLX5_REG_MCC_INSTRUCTION_ACTIVATE,	0,
+				fwhandle, 0);
+}
+
+static int mlx5_fsm_query_state(struct mlxfw_dev *mlxfw_dev, u32 fwhandle,
+				enum mlxfw_fsm_state *fsm_state,
+				enum mlxfw_fsm_state_err *fsm_state_err)
+{
+	struct mlx5_mlxfw_dev *mlx5_mlxfw_dev =
+		container_of(mlxfw_dev, struct mlx5_mlxfw_dev, mlxfw_dev);
+	struct mlx5_core_dev *dev = mlx5_mlxfw_dev->mlx5_core_dev;
+	u8 control_state, error_code;
+	int err;
+
+	err = mlx5_reg_mcc_query(dev, &fwhandle, &error_code, &control_state);
+	if (err)
+		return err;
+
+	*fsm_state = control_state;
+	*fsm_state_err = min_t(enum mlxfw_fsm_state_err, error_code,
+			       MLXFW_FSM_STATE_ERR_MAX);
+	return 0;
+}
+
+static void mlx5_fsm_cancel(struct mlxfw_dev *mlxfw_dev, u32 fwhandle)
+{
+	struct mlx5_mlxfw_dev *mlx5_mlxfw_dev =
+		container_of(mlxfw_dev, struct mlx5_mlxfw_dev, mlxfw_dev);
+	struct mlx5_core_dev *dev = mlx5_mlxfw_dev->mlx5_core_dev;
+
+	mlx5_reg_mcc_set(dev, MLX5_REG_MCC_INSTRUCTION_CANCEL, 0, fwhandle, 0);
+}
+
+static void mlx5_fsm_release(struct mlxfw_dev *mlxfw_dev, u32 fwhandle)
+{
+	struct mlx5_mlxfw_dev *mlx5_mlxfw_dev =
+		container_of(mlxfw_dev, struct mlx5_mlxfw_dev, mlxfw_dev);
+	struct mlx5_core_dev *dev = mlx5_mlxfw_dev->mlx5_core_dev;
+
+	mlx5_reg_mcc_set(dev, MLX5_REG_MCC_INSTRUCTION_RELEASE_UPDATE_HANDLE, 0,
+			 fwhandle, 0);
+}
+
+static const struct mlxfw_dev_ops mlx5_mlxfw_dev_ops = {
+	.component_query	= mlx5_component_query,
+	.fsm_lock		= mlx5_fsm_lock,
+	.fsm_component_update	= mlx5_fsm_component_update,
+	.fsm_block_download	= mlx5_fsm_block_download,
+	.fsm_component_verify	= mlx5_fsm_component_verify,
+	.fsm_activate		= mlx5_fsm_activate,
+	.fsm_query_state	= mlx5_fsm_query_state,
+	.fsm_cancel		= mlx5_fsm_cancel,
+	.fsm_release		= mlx5_fsm_release
+};
+
+int mlx5_firmware_flash(struct mlx5_core_dev *dev,
+			const struct firmware *firmware)
+{
+	struct mlx5_mlxfw_dev mlx5_mlxfw_dev = {
+		.mlxfw_dev = {
+			.ops = &mlx5_mlxfw_dev_ops,
+			.psid = dev->board_id,
+			.psid_size = strlen(dev->board_id),
+		},
+		.mlx5_core_dev = dev
+	};
+
+	if (!MLX5_CAP_GEN(dev, mcam_reg)  ||
+	    !MLX5_CAP_MCAM_REG(dev, mcqi) ||
+	    !MLX5_CAP_MCAM_REG(dev, mcc)  ||
+	    !MLX5_CAP_MCAM_REG(dev, mcda)) {
+		pr_info("%s flashing isn't supported by the running FW\n", __func__);
+		return -EOPNOTSUPP;
+	}
+
+	return mlxfw_firmware_flash(&mlx5_mlxfw_dev.mlxfw_dev, firmware);
+}
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/health.c b/drivers/net/ethernet/mellanox/mlx5/core/health.c
new file mode 100644
index 000000000..9f39aeca8
--- /dev/null
+++ b/drivers/net/ethernet/mellanox/mlx5/core/health.c
@@ -0,0 +1,401 @@
+/*
+ * Copyright (c) 2013-2015, 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/module.h>
+#include <linux/random.h>
+#include <linux/vmalloc.h>
+#include <linux/hardirq.h>
+#include <linux/mlx5/driver.h>
+#include <linux/mlx5/cmd.h>
+#include "mlx5_core.h"
+
+enum {
+	MLX5_HEALTH_POLL_INTERVAL	= 2 * HZ,
+	MAX_MISSES			= 3,
+};
+
+enum {
+	MLX5_HEALTH_SYNDR_FW_ERR		= 0x1,
+	MLX5_HEALTH_SYNDR_IRISC_ERR		= 0x7,
+	MLX5_HEALTH_SYNDR_HW_UNRECOVERABLE_ERR	= 0x8,
+	MLX5_HEALTH_SYNDR_CRC_ERR		= 0x9,
+	MLX5_HEALTH_SYNDR_FETCH_PCI_ERR		= 0xa,
+	MLX5_HEALTH_SYNDR_HW_FTL_ERR		= 0xb,
+	MLX5_HEALTH_SYNDR_ASYNC_EQ_OVERRUN_ERR	= 0xc,
+	MLX5_HEALTH_SYNDR_EQ_ERR		= 0xd,
+	MLX5_HEALTH_SYNDR_EQ_INV		= 0xe,
+	MLX5_HEALTH_SYNDR_FFSER_ERR		= 0xf,
+	MLX5_HEALTH_SYNDR_HIGH_TEMP		= 0x10
+};
+
+enum {
+	MLX5_NIC_IFC_FULL		= 0,
+	MLX5_NIC_IFC_DISABLED		= 1,
+	MLX5_NIC_IFC_NO_DRAM_NIC	= 2,
+	MLX5_NIC_IFC_INVALID		= 3
+};
+
+enum {
+	MLX5_DROP_NEW_HEALTH_WORK,
+	MLX5_DROP_NEW_RECOVERY_WORK,
+};
+
+static u8 get_nic_state(struct mlx5_core_dev *dev)
+{
+	return (ioread32be(&dev->iseg->cmdq_addr_l_sz) >> 8) & 3;
+}
+
+static void trigger_cmd_completions(struct mlx5_core_dev *dev)
+{
+	unsigned long flags;
+	u64 vector;
+
+	/* wait for pending handlers to complete */
+	synchronize_irq(pci_irq_vector(dev->pdev, MLX5_EQ_VEC_CMD));
+	spin_lock_irqsave(&dev->cmd.alloc_lock, flags);
+	vector = ~dev->cmd.bitmask & ((1ul << (1 << dev->cmd.log_sz)) - 1);
+	if (!vector)
+		goto no_trig;
+
+	vector |= MLX5_TRIGGERED_CMD_COMP;
+	spin_unlock_irqrestore(&dev->cmd.alloc_lock, flags);
+
+	mlx5_core_dbg(dev, "vector 0x%llx\n", vector);
+	mlx5_cmd_comp_handler(dev, vector, true);
+	return;
+
+no_trig:
+	spin_unlock_irqrestore(&dev->cmd.alloc_lock, flags);
+}
+
+static int in_fatal(struct mlx5_core_dev *dev)
+{
+	struct mlx5_core_health *health = &dev->priv.health;
+	struct health_buffer __iomem *h = health->health;
+
+	if (get_nic_state(dev) == MLX5_NIC_IFC_DISABLED)
+		return 1;
+
+	if (ioread32be(&h->fw_ver) == 0xffffffff)
+		return 1;
+
+	return 0;
+}
+
+void mlx5_enter_error_state(struct mlx5_core_dev *dev, bool force)
+{
+	mutex_lock(&dev->intf_state_mutex);
+	if (dev->state == MLX5_DEVICE_STATE_INTERNAL_ERROR)
+		goto unlock;
+
+	mlx5_core_err(dev, "start\n");
+	if (pci_channel_offline(dev->pdev) || in_fatal(dev) || force) {
+		dev->state = MLX5_DEVICE_STATE_INTERNAL_ERROR;
+		trigger_cmd_completions(dev);
+	}
+
+	mlx5_core_event(dev, MLX5_DEV_EVENT_SYS_ERROR, 1);
+	mlx5_core_err(dev, "end\n");
+
+unlock:
+	mutex_unlock(&dev->intf_state_mutex);
+}
+
+static void mlx5_handle_bad_state(struct mlx5_core_dev *dev)
+{
+	u8 nic_interface = get_nic_state(dev);
+
+	switch (nic_interface) {
+	case MLX5_NIC_IFC_FULL:
+		mlx5_core_warn(dev, "Expected to see disabled NIC but it is full driver\n");
+		break;
+
+	case MLX5_NIC_IFC_DISABLED:
+		mlx5_core_warn(dev, "starting teardown\n");
+		break;
+
+	case MLX5_NIC_IFC_NO_DRAM_NIC:
+		mlx5_core_warn(dev, "Expected to see disabled NIC but it is no dram nic\n");
+		break;
+	default:
+		mlx5_core_warn(dev, "Expected to see disabled NIC but it is has invalid value %d\n",
+			       nic_interface);
+	}
+
+	mlx5_disable_device(dev);
+}
+
+static void health_recover(struct work_struct *work)
+{
+	struct mlx5_core_health *health;
+	struct delayed_work *dwork;
+	struct mlx5_core_dev *dev;
+	struct mlx5_priv *priv;
+	u8 nic_state;
+
+	dwork = container_of(work, struct delayed_work, work);
+	health = container_of(dwork, struct mlx5_core_health, recover_work);
+	priv = container_of(health, struct mlx5_priv, health);
+	dev = container_of(priv, struct mlx5_core_dev, priv);
+
+	nic_state = get_nic_state(dev);
+	if (nic_state == MLX5_NIC_IFC_INVALID) {
+		dev_err(&dev->pdev->dev, "health recovery flow aborted since the nic state is invalid\n");
+		return;
+	}
+
+	dev_err(&dev->pdev->dev, "starting health recovery flow\n");
+	mlx5_recover_device(dev);
+}
+
+/* How much time to wait until health resetting the driver (in msecs) */
+#define MLX5_RECOVERY_DELAY_MSECS 60000
+static void health_care(struct work_struct *work)
+{
+	unsigned long recover_delay = msecs_to_jiffies(MLX5_RECOVERY_DELAY_MSECS);
+	struct mlx5_core_health *health;
+	struct mlx5_core_dev *dev;
+	struct mlx5_priv *priv;
+	unsigned long flags;
+
+	health = container_of(work, struct mlx5_core_health, work);
+	priv = container_of(health, struct mlx5_priv, health);
+	dev = container_of(priv, struct mlx5_core_dev, priv);
+	mlx5_core_warn(dev, "handling bad device here\n");
+	mlx5_handle_bad_state(dev);
+
+	spin_lock_irqsave(&health->wq_lock, flags);
+	if (!test_bit(MLX5_DROP_NEW_RECOVERY_WORK, &health->flags))
+		schedule_delayed_work(&health->recover_work, recover_delay);
+	else
+		dev_err(&dev->pdev->dev,
+			"new health works are not permitted at this stage\n");
+	spin_unlock_irqrestore(&health->wq_lock, flags);
+}
+
+static const char *hsynd_str(u8 synd)
+{
+	switch (synd) {
+	case MLX5_HEALTH_SYNDR_FW_ERR:
+		return "firmware internal error";
+	case MLX5_HEALTH_SYNDR_IRISC_ERR:
+		return "irisc not responding";
+	case MLX5_HEALTH_SYNDR_HW_UNRECOVERABLE_ERR:
+		return "unrecoverable hardware error";
+	case MLX5_HEALTH_SYNDR_CRC_ERR:
+		return "firmware CRC error";
+	case MLX5_HEALTH_SYNDR_FETCH_PCI_ERR:
+		return "ICM fetch PCI error";
+	case MLX5_HEALTH_SYNDR_HW_FTL_ERR:
+		return "HW fatal error\n";
+	case MLX5_HEALTH_SYNDR_ASYNC_EQ_OVERRUN_ERR:
+		return "async EQ buffer overrun";
+	case MLX5_HEALTH_SYNDR_EQ_ERR:
+		return "EQ error";
+	case MLX5_HEALTH_SYNDR_EQ_INV:
+		return "Invalid EQ referenced";
+	case MLX5_HEALTH_SYNDR_FFSER_ERR:
+		return "FFSER error";
+	case MLX5_HEALTH_SYNDR_HIGH_TEMP:
+		return "High temperature";
+	default:
+		return "unrecognized error";
+	}
+}
+
+static void print_health_info(struct mlx5_core_dev *dev)
+{
+	struct mlx5_core_health *health = &dev->priv.health;
+	struct health_buffer __iomem *h = health->health;
+	char fw_str[18];
+	u32 fw;
+	int i;
+
+	/* If the syndrome is 0, the device is OK and no need to print buffer */
+	if (!ioread8(&h->synd))
+		return;
+
+	for (i = 0; i < ARRAY_SIZE(h->assert_var); i++)
+		dev_err(&dev->pdev->dev, "assert_var[%d] 0x%08x\n", i, ioread32be(h->assert_var + i));
+
+	dev_err(&dev->pdev->dev, "assert_exit_ptr 0x%08x\n", ioread32be(&h->assert_exit_ptr));
+	dev_err(&dev->pdev->dev, "assert_callra 0x%08x\n", ioread32be(&h->assert_callra));
+	sprintf(fw_str, "%d.%d.%d", fw_rev_maj(dev), fw_rev_min(dev), fw_rev_sub(dev));
+	dev_err(&dev->pdev->dev, "fw_ver %s\n", fw_str);
+	dev_err(&dev->pdev->dev, "hw_id 0x%08x\n", ioread32be(&h->hw_id));
+	dev_err(&dev->pdev->dev, "irisc_index %d\n", ioread8(&h->irisc_index));
+	dev_err(&dev->pdev->dev, "synd 0x%x: %s\n", ioread8(&h->synd), hsynd_str(ioread8(&h->synd)));
+	dev_err(&dev->pdev->dev, "ext_synd 0x%04x\n", ioread16be(&h->ext_synd));
+	fw = ioread32be(&h->fw_ver);
+	dev_err(&dev->pdev->dev, "raw fw_ver 0x%08x\n", fw);
+}
+
+static unsigned long get_next_poll_jiffies(void)
+{
+	unsigned long next;
+
+	get_random_bytes(&next, sizeof(next));
+	next %= HZ;
+	next += jiffies + MLX5_HEALTH_POLL_INTERVAL;
+
+	return next;
+}
+
+void mlx5_trigger_health_work(struct mlx5_core_dev *dev)
+{
+	struct mlx5_core_health *health = &dev->priv.health;
+	unsigned long flags;
+
+	spin_lock_irqsave(&health->wq_lock, flags);
+	if (!test_bit(MLX5_DROP_NEW_HEALTH_WORK, &health->flags))
+		queue_work(health->wq, &health->work);
+	else
+		dev_err(&dev->pdev->dev,
+			"new health works are not permitted at this stage\n");
+	spin_unlock_irqrestore(&health->wq_lock, flags);
+}
+
+static void poll_health(struct timer_list *t)
+{
+	struct mlx5_core_dev *dev = from_timer(dev, t, priv.health.timer);
+	struct mlx5_core_health *health = &dev->priv.health;
+	u32 count;
+
+	if (dev->state == MLX5_DEVICE_STATE_INTERNAL_ERROR)
+		goto out;
+
+	count = ioread32be(health->health_counter);
+	if (count == health->prev)
+		++health->miss_counter;
+	else
+		health->miss_counter = 0;
+
+	health->prev = count;
+	if (health->miss_counter == MAX_MISSES) {
+		dev_err(&dev->pdev->dev, "device's health compromised - reached miss count\n");
+		print_health_info(dev);
+	}
+
+	if (in_fatal(dev) && !health->sick) {
+		health->sick = true;
+		print_health_info(dev);
+		mlx5_trigger_health_work(dev);
+	}
+
+out:
+	mod_timer(&health->timer, get_next_poll_jiffies());
+}
+
+void mlx5_start_health_poll(struct mlx5_core_dev *dev)
+{
+	struct mlx5_core_health *health = &dev->priv.health;
+
+	timer_setup(&health->timer, poll_health, 0);
+	health->sick = 0;
+	clear_bit(MLX5_DROP_NEW_HEALTH_WORK, &health->flags);
+	clear_bit(MLX5_DROP_NEW_RECOVERY_WORK, &health->flags);
+	health->health = &dev->iseg->health;
+	health->health_counter = &dev->iseg->health_counter;
+
+	health->timer.expires = round_jiffies(jiffies + MLX5_HEALTH_POLL_INTERVAL);
+	add_timer(&health->timer);
+}
+
+void mlx5_stop_health_poll(struct mlx5_core_dev *dev, bool disable_health)
+{
+	struct mlx5_core_health *health = &dev->priv.health;
+	unsigned long flags;
+
+	if (disable_health) {
+		spin_lock_irqsave(&health->wq_lock, flags);
+		set_bit(MLX5_DROP_NEW_HEALTH_WORK, &health->flags);
+		set_bit(MLX5_DROP_NEW_RECOVERY_WORK, &health->flags);
+		spin_unlock_irqrestore(&health->wq_lock, flags);
+	}
+
+	del_timer_sync(&health->timer);
+}
+
+void mlx5_drain_health_wq(struct mlx5_core_dev *dev)
+{
+	struct mlx5_core_health *health = &dev->priv.health;
+	unsigned long flags;
+
+	spin_lock_irqsave(&health->wq_lock, flags);
+	set_bit(MLX5_DROP_NEW_HEALTH_WORK, &health->flags);
+	set_bit(MLX5_DROP_NEW_RECOVERY_WORK, &health->flags);
+	spin_unlock_irqrestore(&health->wq_lock, flags);
+	cancel_delayed_work_sync(&health->recover_work);
+	cancel_work_sync(&health->work);
+}
+
+void mlx5_drain_health_recovery(struct mlx5_core_dev *dev)
+{
+	struct mlx5_core_health *health = &dev->priv.health;
+	unsigned long flags;
+
+	spin_lock_irqsave(&health->wq_lock, flags);
+	set_bit(MLX5_DROP_NEW_RECOVERY_WORK, &health->flags);
+	spin_unlock_irqrestore(&health->wq_lock, flags);
+	cancel_delayed_work_sync(&dev->priv.health.recover_work);
+}
+
+void mlx5_health_cleanup(struct mlx5_core_dev *dev)
+{
+	struct mlx5_core_health *health = &dev->priv.health;
+
+	destroy_workqueue(health->wq);
+}
+
+int mlx5_health_init(struct mlx5_core_dev *dev)
+{
+	struct mlx5_core_health *health;
+	char *name;
+
+	health = &dev->priv.health;
+	name = kmalloc(64, GFP_KERNEL);
+	if (!name)
+		return -ENOMEM;
+
+	strcpy(name, "mlx5_health");
+	strcat(name, dev_name(&dev->pdev->dev));
+	health->wq = create_singlethread_workqueue(name);
+	kfree(name);
+	if (!health->wq)
+		return -ENOMEM;
+	spin_lock_init(&health->wq_lock);
+	INIT_WORK(&health->work, health_care);
+	INIT_DELAYED_WORK(&health->recover_work, health_recover);
+
+	return 0;
+}
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/ipoib/Makefile b/drivers/net/ethernet/mellanox/mlx5/core/ipoib/Makefile
new file mode 100644
index 000000000..d8e17110f
--- /dev/null
+++ b/drivers/net/ethernet/mellanox/mlx5/core/ipoib/Makefile
@@ -0,0 +1 @@
+subdir-ccflags-y += -I$(src)/..
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/ipoib/ethtool.c b/drivers/net/ethernet/mellanox/mlx5/core/ipoib/ethtool.c
new file mode 100644
index 000000000..90cb50fe1
--- /dev/null
+++ b/drivers/net/ethernet/mellanox/mlx5/core/ipoib/ethtool.c
@@ -0,0 +1,258 @@
+/*
+ * Copyright (c) 2017, 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 "en.h"
+#include "ipoib.h"
+
+static void mlx5i_get_drvinfo(struct net_device *dev,
+			      struct ethtool_drvinfo *drvinfo)
+{
+	struct mlx5e_priv *priv = mlx5i_epriv(dev);
+
+	mlx5e_ethtool_get_drvinfo(priv, drvinfo);
+	strlcpy(drvinfo->driver, DRIVER_NAME "[ib_ipoib]",
+		sizeof(drvinfo->driver));
+}
+
+static void mlx5i_get_strings(struct net_device *dev, u32 stringset, u8 *data)
+{
+	struct mlx5e_priv *priv  = mlx5i_epriv(dev);
+
+	mlx5e_ethtool_get_strings(priv, stringset, data);
+}
+
+static int mlx5i_get_sset_count(struct net_device *dev, int sset)
+{
+	struct mlx5e_priv *priv = mlx5i_epriv(dev);
+
+	return mlx5e_ethtool_get_sset_count(priv, sset);
+}
+
+static void mlx5i_get_ethtool_stats(struct net_device *dev,
+				    struct ethtool_stats *stats,
+				    u64 *data)
+{
+	struct mlx5e_priv *priv = mlx5i_epriv(dev);
+
+	mlx5e_ethtool_get_ethtool_stats(priv, stats, data);
+}
+
+static int mlx5i_set_ringparam(struct net_device *dev,
+			       struct ethtool_ringparam *param)
+{
+	struct mlx5e_priv *priv = mlx5i_epriv(dev);
+
+	return mlx5e_ethtool_set_ringparam(priv, param);
+}
+
+static void mlx5i_get_ringparam(struct net_device *dev,
+				struct ethtool_ringparam *param)
+{
+	struct mlx5e_priv *priv = mlx5i_epriv(dev);
+
+	mlx5e_ethtool_get_ringparam(priv, param);
+}
+
+static int mlx5i_set_channels(struct net_device *dev,
+			      struct ethtool_channels *ch)
+{
+	struct mlx5e_priv *priv = mlx5i_epriv(dev);
+
+	return mlx5e_ethtool_set_channels(priv, ch);
+}
+
+static void mlx5i_get_channels(struct net_device *dev,
+			       struct ethtool_channels *ch)
+{
+	struct mlx5e_priv *priv = mlx5i_epriv(dev);
+
+	mlx5e_ethtool_get_channels(priv, ch);
+}
+
+static int mlx5i_set_coalesce(struct net_device *netdev,
+			      struct ethtool_coalesce *coal)
+{
+	struct mlx5e_priv *priv = mlx5i_epriv(netdev);
+
+	return mlx5e_ethtool_set_coalesce(priv, coal);
+}
+
+static int mlx5i_get_coalesce(struct net_device *netdev,
+			      struct ethtool_coalesce *coal)
+{
+	struct mlx5e_priv *priv = mlx5i_epriv(netdev);
+
+	return mlx5e_ethtool_get_coalesce(priv, coal);
+}
+
+static int mlx5i_get_ts_info(struct net_device *netdev,
+			     struct ethtool_ts_info *info)
+{
+	struct mlx5e_priv *priv = mlx5i_epriv(netdev);
+
+	return mlx5e_ethtool_get_ts_info(priv, info);
+}
+
+static int mlx5i_flash_device(struct net_device *netdev,
+			      struct ethtool_flash *flash)
+{
+	struct mlx5e_priv *priv = mlx5i_epriv(netdev);
+
+	return mlx5e_ethtool_flash_device(priv, flash);
+}
+
+enum mlx5_ptys_width {
+	MLX5_PTYS_WIDTH_1X	= 1 << 0,
+	MLX5_PTYS_WIDTH_2X	= 1 << 1,
+	MLX5_PTYS_WIDTH_4X	= 1 << 2,
+	MLX5_PTYS_WIDTH_8X	= 1 << 3,
+	MLX5_PTYS_WIDTH_12X	= 1 << 4,
+};
+
+static inline int mlx5_ptys_width_enum_to_int(enum mlx5_ptys_width width)
+{
+	switch (width) {
+	case MLX5_PTYS_WIDTH_1X:  return  1;
+	case MLX5_PTYS_WIDTH_2X:  return  2;
+	case MLX5_PTYS_WIDTH_4X:  return  4;
+	case MLX5_PTYS_WIDTH_8X:  return  8;
+	case MLX5_PTYS_WIDTH_12X: return 12;
+	default:		  return -1;
+	}
+}
+
+enum mlx5_ptys_rate {
+	MLX5_PTYS_RATE_SDR	= 1 << 0,
+	MLX5_PTYS_RATE_DDR	= 1 << 1,
+	MLX5_PTYS_RATE_QDR	= 1 << 2,
+	MLX5_PTYS_RATE_FDR10	= 1 << 3,
+	MLX5_PTYS_RATE_FDR	= 1 << 4,
+	MLX5_PTYS_RATE_EDR	= 1 << 5,
+	MLX5_PTYS_RATE_HDR	= 1 << 6,
+};
+
+static inline int mlx5_ptys_rate_enum_to_int(enum mlx5_ptys_rate rate)
+{
+	switch (rate) {
+	case MLX5_PTYS_RATE_SDR:   return 2500;
+	case MLX5_PTYS_RATE_DDR:   return 5000;
+	case MLX5_PTYS_RATE_QDR:
+	case MLX5_PTYS_RATE_FDR10: return 10000;
+	case MLX5_PTYS_RATE_FDR:   return 14000;
+	case MLX5_PTYS_RATE_EDR:   return 25000;
+	case MLX5_PTYS_RATE_HDR:   return 50000;
+	default:		   return -1;
+	}
+}
+
+static int mlx5i_get_port_settings(struct net_device *netdev,
+				   u16 *ib_link_width_oper, u16 *ib_proto_oper)
+{
+	struct mlx5e_priv *priv    = mlx5i_epriv(netdev);
+	struct mlx5_core_dev *mdev = priv->mdev;
+	u32 out[MLX5_ST_SZ_DW(ptys_reg)] = {0};
+	int ret;
+
+	ret = mlx5_query_port_ptys(mdev, out, sizeof(out), MLX5_PTYS_IB, 1);
+	if (ret)
+		return ret;
+
+	*ib_link_width_oper = MLX5_GET(ptys_reg, out, ib_link_width_oper);
+	*ib_proto_oper      = MLX5_GET(ptys_reg, out, ib_proto_oper);
+
+	return 0;
+}
+
+static int mlx5i_get_speed_settings(u16 ib_link_width_oper, u16 ib_proto_oper)
+{
+	int rate, width;
+
+	rate = mlx5_ptys_rate_enum_to_int(ib_proto_oper);
+	if (rate < 0)
+		return -EINVAL;
+	width = mlx5_ptys_width_enum_to_int(ib_link_width_oper);
+	if (width < 0)
+		return -EINVAL;
+
+	return rate * width;
+}
+
+static int mlx5i_get_link_ksettings(struct net_device *netdev,
+				    struct ethtool_link_ksettings *link_ksettings)
+{
+	u16 ib_link_width_oper;
+	u16 ib_proto_oper;
+	int speed, ret;
+
+	ret = mlx5i_get_port_settings(netdev, &ib_link_width_oper, &ib_proto_oper);
+	if (ret)
+		return ret;
+
+	ethtool_link_ksettings_zero_link_mode(link_ksettings, supported);
+	ethtool_link_ksettings_zero_link_mode(link_ksettings, advertising);
+
+	speed = mlx5i_get_speed_settings(ib_link_width_oper, ib_proto_oper);
+	if (speed < 0)
+		return -EINVAL;
+
+	link_ksettings->base.duplex = DUPLEX_FULL;
+	link_ksettings->base.port = PORT_OTHER;
+
+	link_ksettings->base.autoneg = AUTONEG_DISABLE;
+
+	link_ksettings->base.speed = speed;
+
+	return 0;
+}
+
+const struct ethtool_ops mlx5i_ethtool_ops = {
+	.get_drvinfo        = mlx5i_get_drvinfo,
+	.get_strings        = mlx5i_get_strings,
+	.get_sset_count     = mlx5i_get_sset_count,
+	.get_ethtool_stats  = mlx5i_get_ethtool_stats,
+	.get_ringparam      = mlx5i_get_ringparam,
+	.set_ringparam      = mlx5i_set_ringparam,
+	.flash_device       = mlx5i_flash_device,
+	.get_channels       = mlx5i_get_channels,
+	.set_channels       = mlx5i_set_channels,
+	.get_coalesce       = mlx5i_get_coalesce,
+	.set_coalesce       = mlx5i_set_coalesce,
+	.get_ts_info        = mlx5i_get_ts_info,
+	.get_link_ksettings = mlx5i_get_link_ksettings,
+	.get_link           = ethtool_op_get_link,
+};
+
+const struct ethtool_ops mlx5i_pkey_ethtool_ops = {
+	.get_drvinfo        = mlx5i_get_drvinfo,
+	.get_link           = ethtool_op_get_link,
+	.get_ts_info        = mlx5i_get_ts_info,
+};
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/ipoib/ipoib.c b/drivers/net/ethernet/mellanox/mlx5/core/ipoib/ipoib.c
new file mode 100644
index 000000000..db6aafcce
--- /dev/null
+++ b/drivers/net/ethernet/mellanox/mlx5/core/ipoib/ipoib.c
@@ -0,0 +1,697 @@
+/*
+ * Copyright (c) 2017, 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 <rdma/ib_verbs.h>
+#include <linux/mlx5/fs.h>
+#include "en.h"
+#include "ipoib.h"
+
+#define IB_DEFAULT_Q_KEY   0xb1b
+#define MLX5I_PARAMS_DEFAULT_LOG_RQ_SIZE 9
+
+static int mlx5i_open(struct net_device *netdev);
+static int mlx5i_close(struct net_device *netdev);
+static int mlx5i_change_mtu(struct net_device *netdev, int new_mtu);
+
+static const struct net_device_ops mlx5i_netdev_ops = {
+	.ndo_open                = mlx5i_open,
+	.ndo_stop                = mlx5i_close,
+	.ndo_init                = mlx5i_dev_init,
+	.ndo_uninit              = mlx5i_dev_cleanup,
+	.ndo_change_mtu          = mlx5i_change_mtu,
+	.ndo_do_ioctl            = mlx5i_ioctl,
+};
+
+/* IPoIB mlx5 netdev profile */
+static void mlx5i_build_nic_params(struct mlx5_core_dev *mdev,
+				   struct mlx5e_params *params)
+{
+	/* Override RQ params as IPoIB supports only LINKED LIST RQ for now */
+	MLX5E_SET_PFLAG(params, MLX5E_PFLAG_RX_STRIDING_RQ, false);
+	mlx5e_set_rq_type(mdev, params);
+	mlx5e_init_rq_type_params(mdev, params);
+
+	/* RQ size in ipoib by default is 512 */
+	params->log_rq_mtu_frames = is_kdump_kernel() ?
+		MLX5E_PARAMS_MINIMUM_LOG_RQ_SIZE :
+		MLX5I_PARAMS_DEFAULT_LOG_RQ_SIZE;
+
+	params->lro_en = false;
+	params->hard_mtu = MLX5_IB_GRH_BYTES + MLX5_IPOIB_HARD_LEN;
+}
+
+/* Called directly after IPoIB netdevice was created to initialize SW structs */
+void mlx5i_init(struct mlx5_core_dev *mdev,
+		struct net_device *netdev,
+		const struct mlx5e_profile *profile,
+		void *ppriv)
+{
+	struct mlx5e_priv *priv  = mlx5i_epriv(netdev);
+	u16 max_mtu;
+
+	/* priv init */
+	priv->mdev        = mdev;
+	priv->netdev      = netdev;
+	priv->profile     = profile;
+	priv->ppriv       = ppriv;
+	mutex_init(&priv->state_lock);
+
+	mlx5_query_port_max_mtu(mdev, &max_mtu, 1);
+	netdev->mtu = max_mtu;
+
+	mlx5e_build_nic_params(mdev, &priv->channels.params,
+			       profile->max_nch(mdev), netdev->mtu);
+	mlx5i_build_nic_params(mdev, &priv->channels.params);
+
+	mlx5e_timestamp_init(priv);
+
+	/* netdev init */
+	netdev->hw_features    |= NETIF_F_SG;
+	netdev->hw_features    |= NETIF_F_IP_CSUM;
+	netdev->hw_features    |= NETIF_F_IPV6_CSUM;
+	netdev->hw_features    |= NETIF_F_GRO;
+	netdev->hw_features    |= NETIF_F_TSO;
+	netdev->hw_features    |= NETIF_F_TSO6;
+	netdev->hw_features    |= NETIF_F_RXCSUM;
+	netdev->hw_features    |= NETIF_F_RXHASH;
+
+	netdev->netdev_ops = &mlx5i_netdev_ops;
+	netdev->ethtool_ops = &mlx5i_ethtool_ops;
+}
+
+/* Called directly before IPoIB netdevice is destroyed to cleanup SW structs */
+static void mlx5i_cleanup(struct mlx5e_priv *priv)
+{
+	/* Do nothing .. */
+}
+
+int mlx5i_init_underlay_qp(struct mlx5e_priv *priv)
+{
+	struct mlx5_core_dev *mdev = priv->mdev;
+	struct mlx5i_priv *ipriv = priv->ppriv;
+	struct mlx5_core_qp *qp = &ipriv->qp;
+	struct mlx5_qp_context *context;
+	int ret;
+
+	/* QP states */
+	context = kzalloc(sizeof(*context), GFP_KERNEL);
+	if (!context)
+		return -ENOMEM;
+
+	context->flags = cpu_to_be32(MLX5_QP_PM_MIGRATED << 11);
+	context->pri_path.port = 1;
+	context->pri_path.pkey_index = cpu_to_be16(ipriv->pkey_index);
+	context->qkey = cpu_to_be32(IB_DEFAULT_Q_KEY);
+
+	ret = mlx5_core_qp_modify(mdev, MLX5_CMD_OP_RST2INIT_QP, 0, context, qp);
+	if (ret) {
+		mlx5_core_err(mdev, "Failed to modify qp RST2INIT, err: %d\n", ret);
+		goto err_qp_modify_to_err;
+	}
+	memset(context, 0, sizeof(*context));
+
+	ret = mlx5_core_qp_modify(mdev, MLX5_CMD_OP_INIT2RTR_QP, 0, context, qp);
+	if (ret) {
+		mlx5_core_err(mdev, "Failed to modify qp INIT2RTR, err: %d\n", ret);
+		goto err_qp_modify_to_err;
+	}
+
+	ret = mlx5_core_qp_modify(mdev, MLX5_CMD_OP_RTR2RTS_QP, 0, context, qp);
+	if (ret) {
+		mlx5_core_err(mdev, "Failed to modify qp RTR2RTS, err: %d\n", ret);
+		goto err_qp_modify_to_err;
+	}
+
+	kfree(context);
+	return 0;
+
+err_qp_modify_to_err:
+	mlx5_core_qp_modify(mdev, MLX5_CMD_OP_2ERR_QP, 0, &context, qp);
+	kfree(context);
+	return ret;
+}
+
+void mlx5i_uninit_underlay_qp(struct mlx5e_priv *priv)
+{
+	struct mlx5i_priv *ipriv = priv->ppriv;
+	struct mlx5_core_dev *mdev = priv->mdev;
+	struct mlx5_qp_context context;
+	int err;
+
+	err = mlx5_core_qp_modify(mdev, MLX5_CMD_OP_2RST_QP, 0, &context,
+				  &ipriv->qp);
+	if (err)
+		mlx5_core_err(mdev, "Failed to modify qp 2RST, err: %d\n", err);
+}
+
+#define MLX5_QP_ENHANCED_ULP_STATELESS_MODE 2
+
+int mlx5i_create_underlay_qp(struct mlx5_core_dev *mdev, struct mlx5_core_qp *qp)
+{
+	u32 *in = NULL;
+	void *addr_path;
+	int ret = 0;
+	int inlen;
+	void *qpc;
+
+	inlen = MLX5_ST_SZ_BYTES(create_qp_in);
+	in = kvzalloc(inlen, GFP_KERNEL);
+	if (!in)
+		return -ENOMEM;
+
+	qpc = MLX5_ADDR_OF(create_qp_in, in, qpc);
+	MLX5_SET(qpc, qpc, st, MLX5_QP_ST_UD);
+	MLX5_SET(qpc, qpc, pm_state, MLX5_QP_PM_MIGRATED);
+	MLX5_SET(qpc, qpc, ulp_stateless_offload_mode,
+		 MLX5_QP_ENHANCED_ULP_STATELESS_MODE);
+
+	addr_path = MLX5_ADDR_OF(qpc, qpc, primary_address_path);
+	MLX5_SET(ads, addr_path, vhca_port_num, 1);
+	MLX5_SET(ads, addr_path, grh, 1);
+
+	ret = mlx5_core_create_qp(mdev, qp, in, inlen);
+	if (ret) {
+		mlx5_core_err(mdev, "Failed creating IPoIB QP err : %d\n", ret);
+		goto out;
+	}
+
+out:
+	kvfree(in);
+	return ret;
+}
+
+void mlx5i_destroy_underlay_qp(struct mlx5_core_dev *mdev, struct mlx5_core_qp *qp)
+{
+	mlx5_core_destroy_qp(mdev, qp);
+}
+
+static int mlx5i_init_tx(struct mlx5e_priv *priv)
+{
+	struct mlx5i_priv *ipriv = priv->ppriv;
+	int err;
+
+	err = mlx5i_create_underlay_qp(priv->mdev, &ipriv->qp);
+	if (err) {
+		mlx5_core_warn(priv->mdev, "create underlay QP failed, %d\n", err);
+		return err;
+	}
+
+	err = mlx5e_create_tis(priv->mdev, 0 /* tc */, ipriv->qp.qpn, &priv->tisn[0]);
+	if (err) {
+		mlx5_core_warn(priv->mdev, "create tis failed, %d\n", err);
+		goto err_destroy_underlay_qp;
+	}
+
+	return 0;
+
+err_destroy_underlay_qp:
+	mlx5i_destroy_underlay_qp(priv->mdev, &ipriv->qp);
+	return err;
+}
+
+static void mlx5i_cleanup_tx(struct mlx5e_priv *priv)
+{
+	struct mlx5i_priv *ipriv = priv->ppriv;
+
+	mlx5e_destroy_tis(priv->mdev, priv->tisn[0]);
+	mlx5i_destroy_underlay_qp(priv->mdev, &ipriv->qp);
+}
+
+static int mlx5i_create_flow_steering(struct mlx5e_priv *priv)
+{
+	struct ttc_params ttc_params = {};
+	int tt, err;
+
+	priv->fs.ns = mlx5_get_flow_namespace(priv->mdev,
+					       MLX5_FLOW_NAMESPACE_KERNEL);
+
+	if (!priv->fs.ns)
+		return -EINVAL;
+
+	err = mlx5e_arfs_create_tables(priv);
+	if (err) {
+		netdev_err(priv->netdev, "Failed to create arfs tables, err=%d\n",
+			   err);
+		priv->netdev->hw_features &= ~NETIF_F_NTUPLE;
+	}
+
+	mlx5e_set_ttc_basic_params(priv, &ttc_params);
+	mlx5e_set_inner_ttc_ft_params(&ttc_params);
+	for (tt = 0; tt < MLX5E_NUM_INDIR_TIRS; tt++)
+		ttc_params.indir_tirn[tt] = priv->inner_indir_tir[tt].tirn;
+
+	err = mlx5e_create_inner_ttc_table(priv, &ttc_params, &priv->fs.inner_ttc);
+	if (err) {
+		netdev_err(priv->netdev, "Failed to create inner ttc table, err=%d\n",
+			   err);
+		goto err_destroy_arfs_tables;
+	}
+
+	mlx5e_set_ttc_ft_params(&ttc_params);
+	for (tt = 0; tt < MLX5E_NUM_INDIR_TIRS; tt++)
+		ttc_params.indir_tirn[tt] = priv->indir_tir[tt].tirn;
+
+	err = mlx5e_create_ttc_table(priv, &ttc_params, &priv->fs.ttc);
+	if (err) {
+		netdev_err(priv->netdev, "Failed to create ttc table, err=%d\n",
+			   err);
+		goto err_destroy_inner_ttc_table;
+	}
+
+	return 0;
+
+err_destroy_inner_ttc_table:
+	mlx5e_destroy_inner_ttc_table(priv, &priv->fs.inner_ttc);
+err_destroy_arfs_tables:
+	mlx5e_arfs_destroy_tables(priv);
+
+	return err;
+}
+
+static void mlx5i_destroy_flow_steering(struct mlx5e_priv *priv)
+{
+	mlx5e_destroy_ttc_table(priv, &priv->fs.ttc);
+	mlx5e_destroy_inner_ttc_table(priv, &priv->fs.inner_ttc);
+	mlx5e_arfs_destroy_tables(priv);
+}
+
+static int mlx5i_init_rx(struct mlx5e_priv *priv)
+{
+	int err;
+
+	err = mlx5e_create_indirect_rqt(priv);
+	if (err)
+		return err;
+
+	err = mlx5e_create_direct_rqts(priv);
+	if (err)
+		goto err_destroy_indirect_rqts;
+
+	err = mlx5e_create_indirect_tirs(priv);
+	if (err)
+		goto err_destroy_direct_rqts;
+
+	err = mlx5e_create_direct_tirs(priv);
+	if (err)
+		goto err_destroy_indirect_tirs;
+
+	err = mlx5i_create_flow_steering(priv);
+	if (err)
+		goto err_destroy_direct_tirs;
+
+	return 0;
+
+err_destroy_direct_tirs:
+	mlx5e_destroy_direct_tirs(priv);
+err_destroy_indirect_tirs:
+	mlx5e_destroy_indirect_tirs(priv);
+err_destroy_direct_rqts:
+	mlx5e_destroy_direct_rqts(priv);
+err_destroy_indirect_rqts:
+	mlx5e_destroy_rqt(priv, &priv->indir_rqt);
+	return err;
+}
+
+static void mlx5i_cleanup_rx(struct mlx5e_priv *priv)
+{
+	mlx5i_destroy_flow_steering(priv);
+	mlx5e_destroy_direct_tirs(priv);
+	mlx5e_destroy_indirect_tirs(priv);
+	mlx5e_destroy_direct_rqts(priv);
+	mlx5e_destroy_rqt(priv, &priv->indir_rqt);
+}
+
+static const struct mlx5e_profile mlx5i_nic_profile = {
+	.init		   = mlx5i_init,
+	.cleanup	   = mlx5i_cleanup,
+	.init_tx	   = mlx5i_init_tx,
+	.cleanup_tx	   = mlx5i_cleanup_tx,
+	.init_rx	   = mlx5i_init_rx,
+	.cleanup_rx	   = mlx5i_cleanup_rx,
+	.enable		   = NULL, /* mlx5i_enable */
+	.disable	   = NULL, /* mlx5i_disable */
+	.update_stats	   = NULL, /* mlx5i_update_stats */
+	.max_nch	   = mlx5e_get_max_num_channels,
+	.update_carrier    = NULL, /* no HW update in IB link */
+	.rx_handlers.handle_rx_cqe       = mlx5i_handle_rx_cqe,
+	.rx_handlers.handle_rx_cqe_mpwqe = NULL, /* Not supported */
+	.max_tc		   = MLX5I_MAX_NUM_TC,
+};
+
+/* mlx5i netdev NDos */
+
+static int mlx5i_change_mtu(struct net_device *netdev, int new_mtu)
+{
+	struct mlx5e_priv *priv = mlx5i_epriv(netdev);
+	struct mlx5e_channels new_channels = {};
+	struct mlx5e_params *params;
+	int err = 0;
+
+	mutex_lock(&priv->state_lock);
+
+	params = &priv->channels.params;
+
+	if (!test_bit(MLX5E_STATE_OPENED, &priv->state)) {
+		params->sw_mtu = new_mtu;
+		netdev->mtu = params->sw_mtu;
+		goto out;
+	}
+
+	new_channels.params = *params;
+	new_channels.params.sw_mtu = new_mtu;
+	err = mlx5e_open_channels(priv, &new_channels);
+	if (err)
+		goto out;
+
+	mlx5e_switch_priv_channels(priv, &new_channels, NULL);
+	netdev->mtu = new_channels.params.sw_mtu;
+
+out:
+	mutex_unlock(&priv->state_lock);
+	return err;
+}
+
+int mlx5i_dev_init(struct net_device *dev)
+{
+	struct mlx5e_priv    *priv   = mlx5i_epriv(dev);
+	struct mlx5i_priv    *ipriv  = priv->ppriv;
+
+	/* Set dev address using underlay QP */
+	dev->dev_addr[1] = (ipriv->qp.qpn >> 16) & 0xff;
+	dev->dev_addr[2] = (ipriv->qp.qpn >>  8) & 0xff;
+	dev->dev_addr[3] = (ipriv->qp.qpn) & 0xff;
+
+	/* Add QPN to net-device mapping to HT */
+	mlx5i_pkey_add_qpn(dev ,ipriv->qp.qpn);
+
+	return 0;
+}
+
+int mlx5i_ioctl(struct net_device *dev, struct ifreq *ifr, int cmd)
+{
+	struct mlx5e_priv *priv = mlx5i_epriv(dev);
+
+	switch (cmd) {
+	case SIOCSHWTSTAMP:
+		return mlx5e_hwstamp_set(priv, ifr);
+	case SIOCGHWTSTAMP:
+		return mlx5e_hwstamp_get(priv, ifr);
+	default:
+		return -EOPNOTSUPP;
+	}
+}
+
+void mlx5i_dev_cleanup(struct net_device *dev)
+{
+	struct mlx5e_priv    *priv   = mlx5i_epriv(dev);
+	struct mlx5i_priv    *ipriv = priv->ppriv;
+
+	mlx5i_uninit_underlay_qp(priv);
+
+	/* Delete QPN to net-device mapping from HT */
+	mlx5i_pkey_del_qpn(dev, ipriv->qp.qpn);
+}
+
+static int mlx5i_open(struct net_device *netdev)
+{
+	struct mlx5e_priv *epriv = mlx5i_epriv(netdev);
+	struct mlx5i_priv *ipriv = epriv->ppriv;
+	struct mlx5_core_dev *mdev = epriv->mdev;
+	int err;
+
+	mutex_lock(&epriv->state_lock);
+
+	set_bit(MLX5E_STATE_OPENED, &epriv->state);
+
+	err = mlx5i_init_underlay_qp(epriv);
+	if (err) {
+		mlx5_core_warn(mdev, "prepare underlay qp state failed, %d\n", err);
+		goto err_clear_state_opened_flag;
+	}
+
+	err = mlx5_fs_add_rx_underlay_qpn(mdev, ipriv->qp.qpn);
+	if (err) {
+		mlx5_core_warn(mdev, "attach underlay qp to ft failed, %d\n", err);
+		goto err_reset_qp;
+	}
+
+	err = mlx5e_open_channels(epriv, &epriv->channels);
+	if (err)
+		goto err_remove_fs_underlay_qp;
+
+	mlx5e_refresh_tirs(epriv, false);
+	mlx5e_activate_priv_channels(epriv);
+
+	mutex_unlock(&epriv->state_lock);
+	return 0;
+
+err_remove_fs_underlay_qp:
+	mlx5_fs_remove_rx_underlay_qpn(mdev, ipriv->qp.qpn);
+err_reset_qp:
+	mlx5i_uninit_underlay_qp(epriv);
+err_clear_state_opened_flag:
+	clear_bit(MLX5E_STATE_OPENED, &epriv->state);
+	mutex_unlock(&epriv->state_lock);
+	return err;
+}
+
+static int mlx5i_close(struct net_device *netdev)
+{
+	struct mlx5e_priv *epriv = mlx5i_epriv(netdev);
+	struct mlx5i_priv *ipriv = epriv->ppriv;
+	struct mlx5_core_dev *mdev = epriv->mdev;
+
+	/* May already be CLOSED in case a previous configuration operation
+	 * (e.g RX/TX queue size change) that involves close&open failed.
+	 */
+	mutex_lock(&epriv->state_lock);
+
+	if (!test_bit(MLX5E_STATE_OPENED, &epriv->state))
+		goto unlock;
+
+	clear_bit(MLX5E_STATE_OPENED, &epriv->state);
+
+	netif_carrier_off(epriv->netdev);
+	mlx5_fs_remove_rx_underlay_qpn(mdev, ipriv->qp.qpn);
+	mlx5e_deactivate_priv_channels(epriv);
+	mlx5e_close_channels(&epriv->channels);
+	mlx5i_uninit_underlay_qp(epriv);
+unlock:
+	mutex_unlock(&epriv->state_lock);
+	return 0;
+}
+
+/* IPoIB RDMA netdev callbacks */
+static int mlx5i_attach_mcast(struct net_device *netdev, struct ib_device *hca,
+			      union ib_gid *gid, u16 lid, int set_qkey,
+			      u32 qkey)
+{
+	struct mlx5e_priv    *epriv = mlx5i_epriv(netdev);
+	struct mlx5_core_dev *mdev  = epriv->mdev;
+	struct mlx5i_priv    *ipriv = epriv->ppriv;
+	int err;
+
+	mlx5_core_dbg(mdev, "attaching QPN 0x%x, MGID %pI6\n", ipriv->qp.qpn, gid->raw);
+	err = mlx5_core_attach_mcg(mdev, gid, ipriv->qp.qpn);
+	if (err)
+		mlx5_core_warn(mdev, "failed attaching QPN 0x%x, MGID %pI6\n",
+			       ipriv->qp.qpn, gid->raw);
+
+	if (set_qkey) {
+		mlx5_core_dbg(mdev, "%s setting qkey 0x%x\n",
+			      netdev->name, qkey);
+		ipriv->qkey = qkey;
+	}
+
+	return err;
+}
+
+static int mlx5i_detach_mcast(struct net_device *netdev, struct ib_device *hca,
+			      union ib_gid *gid, u16 lid)
+{
+	struct mlx5e_priv    *epriv = mlx5i_epriv(netdev);
+	struct mlx5_core_dev *mdev  = epriv->mdev;
+	struct mlx5i_priv    *ipriv = epriv->ppriv;
+	int err;
+
+	mlx5_core_dbg(mdev, "detaching QPN 0x%x, MGID %pI6\n", ipriv->qp.qpn, gid->raw);
+
+	err = mlx5_core_detach_mcg(mdev, gid, ipriv->qp.qpn);
+	if (err)
+		mlx5_core_dbg(mdev, "failed detaching QPN 0x%x, MGID %pI6\n",
+			      ipriv->qp.qpn, gid->raw);
+
+	return err;
+}
+
+static int mlx5i_xmit(struct net_device *dev, struct sk_buff *skb,
+		      struct ib_ah *address, u32 dqpn)
+{
+	struct mlx5e_priv *epriv = mlx5i_epriv(dev);
+	struct mlx5e_txqsq *sq   = epriv->txq2sq[skb_get_queue_mapping(skb)];
+	struct mlx5_ib_ah *mah   = to_mah(address);
+	struct mlx5i_priv *ipriv = epriv->ppriv;
+
+	return mlx5i_sq_xmit(sq, skb, &mah->av, dqpn, ipriv->qkey);
+}
+
+static void mlx5i_set_pkey_index(struct net_device *netdev, int id)
+{
+	struct mlx5i_priv *ipriv = netdev_priv(netdev);
+
+	ipriv->pkey_index = (u16)id;
+}
+
+static int mlx5i_check_required_hca_cap(struct mlx5_core_dev *mdev)
+{
+	if (MLX5_CAP_GEN(mdev, port_type) != MLX5_CAP_PORT_TYPE_IB)
+		return -EOPNOTSUPP;
+
+	if (!MLX5_CAP_GEN(mdev, ipoib_enhanced_offloads)) {
+		mlx5_core_warn(mdev, "IPoIB enhanced offloads are not supported\n");
+		return -EOPNOTSUPP;
+	}
+
+	return 0;
+}
+
+static void mlx5_rdma_netdev_free(struct net_device *netdev)
+{
+	struct mlx5e_priv *priv = mlx5i_epriv(netdev);
+	struct mlx5i_priv *ipriv = priv->ppriv;
+	const struct mlx5e_profile *profile = priv->profile;
+
+	mlx5e_detach_netdev(priv);
+	profile->cleanup(priv);
+	destroy_workqueue(priv->wq);
+
+	if (!ipriv->sub_interface) {
+		mlx5i_pkey_qpn_ht_cleanup(netdev);
+		mlx5e_destroy_mdev_resources(priv->mdev);
+	}
+}
+
+struct net_device *mlx5_rdma_netdev_alloc(struct mlx5_core_dev *mdev,
+					  struct ib_device *ibdev,
+					  const char *name,
+					  void (*setup)(struct net_device *))
+{
+	const struct mlx5e_profile *profile;
+	struct net_device *netdev;
+	struct mlx5i_priv *ipriv;
+	struct mlx5e_priv *epriv;
+	struct rdma_netdev *rn;
+	bool sub_interface;
+	int nch;
+	int err;
+
+	if (mlx5i_check_required_hca_cap(mdev)) {
+		mlx5_core_warn(mdev, "Accelerated mode is not supported\n");
+		return ERR_PTR(-EOPNOTSUPP);
+	}
+
+	/* TODO: Need to find a better way to check if child device*/
+	sub_interface = (mdev->mlx5e_res.pdn != 0);
+
+	if (sub_interface)
+		profile = mlx5i_pkey_get_profile();
+	else
+		profile = &mlx5i_nic_profile;
+
+	nch = profile->max_nch(mdev);
+
+	netdev = alloc_netdev_mqs(sizeof(struct mlx5i_priv) + sizeof(struct mlx5e_priv),
+				  name, NET_NAME_UNKNOWN,
+				  setup,
+				  nch * MLX5E_MAX_NUM_TC,
+				  nch);
+	if (!netdev) {
+		mlx5_core_warn(mdev, "alloc_netdev_mqs failed\n");
+		return NULL;
+	}
+
+	ipriv = netdev_priv(netdev);
+	epriv = mlx5i_epriv(netdev);
+
+	epriv->wq = create_singlethread_workqueue("mlx5i");
+	if (!epriv->wq)
+		goto err_free_netdev;
+
+	ipriv->sub_interface = sub_interface;
+	if (!ipriv->sub_interface) {
+		err = mlx5i_pkey_qpn_ht_init(netdev);
+		if (err) {
+			mlx5_core_warn(mdev, "allocate qpn_to_netdev ht failed\n");
+			goto destroy_wq;
+		}
+
+		/* This should only be called once per mdev */
+		err = mlx5e_create_mdev_resources(mdev);
+		if (err)
+			goto destroy_ht;
+	}
+
+	profile->init(mdev, netdev, profile, ipriv);
+
+	err = mlx5e_attach_netdev(epriv);
+	if (err)
+		goto detach;
+	netif_carrier_off(netdev);
+
+	/* set rdma_netdev func pointers */
+	rn = &ipriv->rn;
+	rn->hca  = ibdev;
+	rn->send = mlx5i_xmit;
+	rn->attach_mcast = mlx5i_attach_mcast;
+	rn->detach_mcast = mlx5i_detach_mcast;
+	rn->set_id = mlx5i_set_pkey_index;
+
+	netdev->priv_destructor = mlx5_rdma_netdev_free;
+	netdev->needs_free_netdev = 1;
+
+	return netdev;
+
+detach:
+	profile->cleanup(epriv);
+	if (ipriv->sub_interface)
+		return NULL;
+	mlx5e_destroy_mdev_resources(mdev);
+destroy_ht:
+	mlx5i_pkey_qpn_ht_cleanup(netdev);
+destroy_wq:
+	destroy_workqueue(epriv->wq);
+err_free_netdev:
+	free_netdev(netdev);
+
+	return NULL;
+}
+EXPORT_SYMBOL(mlx5_rdma_netdev_alloc);
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/ipoib/ipoib.h b/drivers/net/ethernet/mellanox/mlx5/core/ipoib/ipoib.h
new file mode 100644
index 000000000..0982c579e
--- /dev/null
+++ b/drivers/net/ethernet/mellanox/mlx5/core/ipoib/ipoib.h
@@ -0,0 +1,125 @@
+/*
+ * Copyright (c) 2017, 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 __MLX5E_IPOB_H__
+#define __MLX5E_IPOB_H__
+
+#ifdef CONFIG_MLX5_CORE_IPOIB
+
+#include <linux/mlx5/fs.h>
+#include "en.h"
+
+#define MLX5I_MAX_NUM_TC 1
+
+extern const struct ethtool_ops mlx5i_ethtool_ops;
+extern const struct ethtool_ops mlx5i_pkey_ethtool_ops;
+
+#define MLX5_IB_GRH_BYTES       40
+#define MLX5_IPOIB_ENCAP_LEN    4
+#define MLX5_IPOIB_PSEUDO_LEN   20
+#define MLX5_IPOIB_HARD_LEN     (MLX5_IPOIB_PSEUDO_LEN + MLX5_IPOIB_ENCAP_LEN)
+
+/* ipoib rdma netdev's private data structure */
+struct mlx5i_priv {
+	struct rdma_netdev rn; /* keep this first */
+	struct mlx5_core_qp qp;
+	bool   sub_interface;
+	u32    qkey;
+	u16    pkey_index;
+	struct mlx5i_pkey_qpn_ht *qpn_htbl;
+	char  *mlx5e_priv[0];
+};
+
+/* Underlay QP create/destroy functions */
+int mlx5i_create_underlay_qp(struct mlx5_core_dev *mdev, struct mlx5_core_qp *qp);
+void mlx5i_destroy_underlay_qp(struct mlx5_core_dev *mdev, struct mlx5_core_qp *qp);
+
+/* Underlay QP state modification init/uninit functions */
+int mlx5i_init_underlay_qp(struct mlx5e_priv *priv);
+void mlx5i_uninit_underlay_qp(struct mlx5e_priv *priv);
+
+/* Allocate/Free underlay QPN to net-device hash table */
+int mlx5i_pkey_qpn_ht_init(struct net_device *netdev);
+void mlx5i_pkey_qpn_ht_cleanup(struct net_device *netdev);
+
+/* Add/Remove an underlay QPN to net-device mapping to/from the hash table */
+int mlx5i_pkey_add_qpn(struct net_device *netdev, u32 qpn);
+int mlx5i_pkey_del_qpn(struct net_device *netdev, u32 qpn);
+
+/* Get the net-device corresponding to the given underlay QPN */
+struct net_device *mlx5i_pkey_get_netdev(struct net_device *netdev, u32 qpn);
+
+/* Shared ndo functions */
+int mlx5i_dev_init(struct net_device *dev);
+void mlx5i_dev_cleanup(struct net_device *dev);
+int mlx5i_ioctl(struct net_device *dev, struct ifreq *ifr, int cmd);
+
+/* Parent profile functions */
+void mlx5i_init(struct mlx5_core_dev *mdev,
+		struct net_device *netdev,
+		const struct mlx5e_profile *profile,
+		void *ppriv);
+
+/* Get child interface nic profile */
+const struct mlx5e_profile *mlx5i_pkey_get_profile(void);
+
+/* Extract mlx5e_priv from IPoIB netdev */
+#define mlx5i_epriv(netdev) ((void *)(((struct mlx5i_priv *)netdev_priv(netdev))->mlx5e_priv))
+
+struct mlx5_wqe_eth_pad {
+	u8 rsvd0[16];
+};
+
+struct mlx5i_tx_wqe {
+	struct mlx5_wqe_ctrl_seg     ctrl;
+	struct mlx5_wqe_datagram_seg datagram;
+	struct mlx5_wqe_eth_pad      pad;
+	struct mlx5_wqe_eth_seg      eth;
+	struct mlx5_wqe_data_seg     data[0];
+};
+
+static inline void mlx5i_sq_fetch_wqe(struct mlx5e_txqsq *sq,
+				      struct mlx5i_tx_wqe **wqe,
+				      u16 pi)
+{
+	struct mlx5_wq_cyc *wq = &sq->wq;
+
+	*wqe = mlx5_wq_cyc_get_wqe(wq, pi);
+	memset(*wqe, 0, sizeof(**wqe));
+}
+
+netdev_tx_t mlx5i_sq_xmit(struct mlx5e_txqsq *sq, struct sk_buff *skb,
+			  struct mlx5_av *av, u32 dqpn, u32 dqkey);
+void mlx5i_handle_rx_cqe(struct mlx5e_rq *rq, struct mlx5_cqe64 *cqe);
+
+#endif /* CONFIG_MLX5_CORE_IPOIB */
+#endif /* __MLX5E_IPOB_H__ */
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/ipoib/ipoib_vlan.c b/drivers/net/ethernet/mellanox/mlx5/core/ipoib/ipoib_vlan.c
new file mode 100644
index 000000000..54a188f41
--- /dev/null
+++ b/drivers/net/ethernet/mellanox/mlx5/core/ipoib/ipoib_vlan.c
@@ -0,0 +1,357 @@
+/*
+ * Copyright (c) 2017, 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/hash.h>
+#include "ipoib.h"
+
+#define MLX5I_MAX_LOG_PKEY_SUP 7
+
+struct qpn_to_netdev {
+	struct net_device *netdev;
+	struct hlist_node hlist;
+	u32 underlay_qpn;
+};
+
+struct mlx5i_pkey_qpn_ht {
+	struct hlist_head buckets[1 << MLX5I_MAX_LOG_PKEY_SUP];
+	spinlock_t ht_lock; /* Synchronise with NAPI */
+};
+
+int mlx5i_pkey_qpn_ht_init(struct net_device *netdev)
+{
+	struct mlx5i_priv *ipriv = netdev_priv(netdev);
+	struct mlx5i_pkey_qpn_ht *qpn_htbl;
+
+	qpn_htbl = kzalloc(sizeof(*qpn_htbl), GFP_KERNEL);
+	if (!qpn_htbl)
+		return -ENOMEM;
+
+	ipriv->qpn_htbl = qpn_htbl;
+	spin_lock_init(&qpn_htbl->ht_lock);
+
+	return 0;
+}
+
+void mlx5i_pkey_qpn_ht_cleanup(struct net_device *netdev)
+{
+	struct mlx5i_priv *ipriv = netdev_priv(netdev);
+
+	kfree(ipriv->qpn_htbl);
+}
+
+static struct qpn_to_netdev *mlx5i_find_qpn_to_netdev_node(struct hlist_head *buckets,
+							   u32 qpn)
+{
+	struct hlist_head *h = &buckets[hash_32(qpn, MLX5I_MAX_LOG_PKEY_SUP)];
+	struct qpn_to_netdev *node;
+
+	hlist_for_each_entry(node, h, hlist) {
+		if (node->underlay_qpn == qpn)
+			return node;
+	}
+
+	return NULL;
+}
+
+int mlx5i_pkey_add_qpn(struct net_device *netdev, u32 qpn)
+{
+	struct mlx5i_priv *ipriv = netdev_priv(netdev);
+	struct mlx5i_pkey_qpn_ht *ht = ipriv->qpn_htbl;
+	u8 key = hash_32(qpn, MLX5I_MAX_LOG_PKEY_SUP);
+	struct qpn_to_netdev *new_node;
+
+	new_node = kzalloc(sizeof(*new_node), GFP_KERNEL);
+	if (!new_node)
+		return -ENOMEM;
+
+	new_node->netdev = netdev;
+	new_node->underlay_qpn = qpn;
+	spin_lock_bh(&ht->ht_lock);
+	hlist_add_head(&new_node->hlist, &ht->buckets[key]);
+	spin_unlock_bh(&ht->ht_lock);
+
+	return 0;
+}
+
+int mlx5i_pkey_del_qpn(struct net_device *netdev, u32 qpn)
+{
+	struct mlx5e_priv *epriv = mlx5i_epriv(netdev);
+	struct mlx5i_priv *ipriv = epriv->ppriv;
+	struct mlx5i_pkey_qpn_ht *ht = ipriv->qpn_htbl;
+	struct qpn_to_netdev *node;
+
+	node = mlx5i_find_qpn_to_netdev_node(ht->buckets, qpn);
+	if (!node) {
+		mlx5_core_warn(epriv->mdev, "QPN to netdev delete from HT failed\n");
+		return -EINVAL;
+	}
+
+	spin_lock_bh(&ht->ht_lock);
+	hlist_del_init(&node->hlist);
+	spin_unlock_bh(&ht->ht_lock);
+	kfree(node);
+
+	return 0;
+}
+
+struct net_device *mlx5i_pkey_get_netdev(struct net_device *netdev, u32 qpn)
+{
+	struct mlx5i_priv *ipriv = netdev_priv(netdev);
+	struct qpn_to_netdev *node;
+
+	node = mlx5i_find_qpn_to_netdev_node(ipriv->qpn_htbl->buckets, qpn);
+	if (!node)
+		return NULL;
+
+	return node->netdev;
+}
+
+static int mlx5i_pkey_open(struct net_device *netdev);
+static int mlx5i_pkey_close(struct net_device *netdev);
+static int mlx5i_pkey_dev_init(struct net_device *dev);
+static void mlx5i_pkey_dev_cleanup(struct net_device *netdev);
+static int mlx5i_pkey_change_mtu(struct net_device *netdev, int new_mtu);
+static int mlx5i_pkey_ioctl(struct net_device *dev, struct ifreq *ifr, int cmd);
+
+static const struct net_device_ops mlx5i_pkey_netdev_ops = {
+	.ndo_open                = mlx5i_pkey_open,
+	.ndo_stop                = mlx5i_pkey_close,
+	.ndo_init                = mlx5i_pkey_dev_init,
+	.ndo_uninit              = mlx5i_pkey_dev_cleanup,
+	.ndo_change_mtu          = mlx5i_pkey_change_mtu,
+	.ndo_do_ioctl            = mlx5i_pkey_ioctl,
+};
+
+/* Child NDOs */
+static int mlx5i_pkey_dev_init(struct net_device *dev)
+{
+	struct mlx5e_priv *priv = mlx5i_epriv(dev);
+	struct mlx5i_priv *ipriv, *parent_ipriv;
+	struct net_device *parent_dev;
+	int parent_ifindex;
+
+	ipriv = priv->ppriv;
+
+	/* Get QPN to netdevice hash table from parent */
+	parent_ifindex = dev->netdev_ops->ndo_get_iflink(dev);
+	parent_dev = dev_get_by_index(dev_net(dev), parent_ifindex);
+	if (!parent_dev) {
+		mlx5_core_warn(priv->mdev, "failed to get parent device\n");
+		return -EINVAL;
+	}
+
+	parent_ipriv = netdev_priv(parent_dev);
+	ipriv->qpn_htbl = parent_ipriv->qpn_htbl;
+	dev_put(parent_dev);
+
+	return mlx5i_dev_init(dev);
+}
+
+static int mlx5i_pkey_ioctl(struct net_device *dev, struct ifreq *ifr, int cmd)
+{
+	return mlx5i_ioctl(dev, ifr, cmd);
+}
+
+static void mlx5i_pkey_dev_cleanup(struct net_device *netdev)
+{
+	return mlx5i_dev_cleanup(netdev);
+}
+
+static int mlx5i_pkey_open(struct net_device *netdev)
+{
+	struct mlx5e_priv *epriv = mlx5i_epriv(netdev);
+	struct mlx5i_priv *ipriv = epriv->ppriv;
+	struct mlx5_core_dev *mdev = epriv->mdev;
+	int err;
+
+	mutex_lock(&epriv->state_lock);
+
+	set_bit(MLX5E_STATE_OPENED, &epriv->state);
+
+	err = mlx5i_init_underlay_qp(epriv);
+	if (err) {
+		mlx5_core_warn(mdev, "prepare child underlay qp state failed, %d\n", err);
+		goto err_release_lock;
+	}
+
+	err = mlx5_fs_add_rx_underlay_qpn(mdev, ipriv->qp.qpn);
+	if (err) {
+		mlx5_core_warn(mdev, "attach child underlay qp to ft failed, %d\n", err);
+		goto err_unint_underlay_qp;
+	}
+
+	err = mlx5e_create_tis(mdev, 0 /* tc */, ipriv->qp.qpn, &epriv->tisn[0]);
+	if (err) {
+		mlx5_core_warn(mdev, "create child tis failed, %d\n", err);
+		goto err_remove_rx_uderlay_qp;
+	}
+
+	err = mlx5e_open_channels(epriv, &epriv->channels);
+	if (err) {
+		mlx5_core_warn(mdev, "opening child channels failed, %d\n", err);
+		goto err_clear_state_opened_flag;
+	}
+	mlx5e_refresh_tirs(epriv, false);
+	mlx5e_activate_priv_channels(epriv);
+	mutex_unlock(&epriv->state_lock);
+
+	return 0;
+
+err_clear_state_opened_flag:
+	mlx5e_destroy_tis(mdev, epriv->tisn[0]);
+err_remove_rx_uderlay_qp:
+	mlx5_fs_remove_rx_underlay_qpn(mdev, ipriv->qp.qpn);
+err_unint_underlay_qp:
+	mlx5i_uninit_underlay_qp(epriv);
+err_release_lock:
+	clear_bit(MLX5E_STATE_OPENED, &epriv->state);
+	mutex_unlock(&epriv->state_lock);
+	return err;
+}
+
+static int mlx5i_pkey_close(struct net_device *netdev)
+{
+	struct mlx5e_priv *priv = mlx5i_epriv(netdev);
+	struct mlx5i_priv *ipriv = priv->ppriv;
+	struct mlx5_core_dev *mdev = priv->mdev;
+
+	mutex_lock(&priv->state_lock);
+
+	if (!test_bit(MLX5E_STATE_OPENED, &priv->state))
+		goto unlock;
+
+	clear_bit(MLX5E_STATE_OPENED, &priv->state);
+
+	netif_carrier_off(priv->netdev);
+	mlx5_fs_remove_rx_underlay_qpn(mdev, ipriv->qp.qpn);
+	mlx5i_uninit_underlay_qp(priv);
+	mlx5e_deactivate_priv_channels(priv);
+	mlx5e_close_channels(&priv->channels);
+	mlx5e_destroy_tis(mdev, priv->tisn[0]);
+unlock:
+	mutex_unlock(&priv->state_lock);
+	return 0;
+}
+
+static int mlx5i_pkey_change_mtu(struct net_device *netdev, int new_mtu)
+{
+	struct mlx5e_priv *priv = mlx5i_epriv(netdev);
+
+	mutex_lock(&priv->state_lock);
+	netdev->mtu = new_mtu;
+	mutex_unlock(&priv->state_lock);
+
+	return 0;
+}
+
+/* Called directly after IPoIB netdevice was created to initialize SW structs */
+static void mlx5i_pkey_init(struct mlx5_core_dev *mdev,
+			     struct net_device *netdev,
+			     const struct mlx5e_profile *profile,
+			     void *ppriv)
+{
+	struct mlx5e_priv *priv  = mlx5i_epriv(netdev);
+
+	mlx5i_init(mdev, netdev, profile, ppriv);
+
+	/* Override parent ndo */
+	netdev->netdev_ops = &mlx5i_pkey_netdev_ops;
+
+	/* Set child limited ethtool support */
+	netdev->ethtool_ops = &mlx5i_pkey_ethtool_ops;
+
+	/* Use dummy rqs */
+	priv->channels.params.log_rq_mtu_frames = MLX5E_PARAMS_MINIMUM_LOG_RQ_SIZE;
+}
+
+/* Called directly before IPoIB netdevice is destroyed to cleanup SW structs */
+static void mlx5i_pkey_cleanup(struct mlx5e_priv *priv)
+{
+	/* Do nothing .. */
+}
+
+static int mlx5i_pkey_init_tx(struct mlx5e_priv *priv)
+{
+	struct mlx5i_priv *ipriv = priv->ppriv;
+	int err;
+
+	err = mlx5i_create_underlay_qp(priv->mdev, &ipriv->qp);
+	if (err) {
+		mlx5_core_warn(priv->mdev, "create child underlay QP failed, %d\n", err);
+		return err;
+	}
+
+	return 0;
+}
+
+static void mlx5i_pkey_cleanup_tx(struct mlx5e_priv *priv)
+{
+	struct mlx5i_priv *ipriv = priv->ppriv;
+
+	mlx5i_destroy_underlay_qp(priv->mdev, &ipriv->qp);
+}
+
+static int mlx5i_pkey_init_rx(struct mlx5e_priv *priv)
+{
+	/* Since the rx resources are shared between child and parent, the
+	 * parent interface is taking care of rx resource allocation and init
+	 */
+	return 0;
+}
+
+static void mlx5i_pkey_cleanup_rx(struct mlx5e_priv *priv)
+{
+	/* Since the rx resources are shared between child and parent, the
+	 * parent interface is taking care of rx resource free and de-init
+	 */
+}
+
+static const struct mlx5e_profile mlx5i_pkey_nic_profile = {
+	.init		   = mlx5i_pkey_init,
+	.cleanup	   = mlx5i_pkey_cleanup,
+	.init_tx	   = mlx5i_pkey_init_tx,
+	.cleanup_tx	   = mlx5i_pkey_cleanup_tx,
+	.init_rx	   = mlx5i_pkey_init_rx,
+	.cleanup_rx	   = mlx5i_pkey_cleanup_rx,
+	.enable		   = NULL,
+	.disable	   = NULL,
+	.update_stats	   = NULL,
+	.max_nch	   = mlx5e_get_max_num_channels,
+	.rx_handlers.handle_rx_cqe       = mlx5i_handle_rx_cqe,
+	.rx_handlers.handle_rx_cqe_mpwqe = NULL, /* Not supported */
+	.max_tc		   = MLX5I_MAX_NUM_TC,
+};
+
+const struct mlx5e_profile *mlx5i_pkey_get_profile(void)
+{
+	return &mlx5i_pkey_nic_profile;
+}
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/lag.c b/drivers/net/ethernet/mellanox/mlx5/core/lag.c
new file mode 100644
index 000000000..582b2f180
--- /dev/null
+++ b/drivers/net/ethernet/mellanox/mlx5/core/lag.c
@@ -0,0 +1,691 @@
+/*
+ * Copyright (c) 2016, 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/netdevice.h>
+#include <linux/mlx5/driver.h>
+#include <linux/mlx5/vport.h>
+#include "mlx5_core.h"
+
+enum {
+	MLX5_LAG_FLAG_BONDED = 1 << 0,
+};
+
+struct lag_func {
+	struct mlx5_core_dev *dev;
+	struct net_device    *netdev;
+};
+
+/* Used for collection of netdev event info. */
+struct lag_tracker {
+	enum   netdev_lag_tx_type           tx_type;
+	struct netdev_lag_lower_state_info  netdev_state[MLX5_MAX_PORTS];
+	bool is_bonded;
+};
+
+/* LAG data of a ConnectX card.
+ * It serves both its phys functions.
+ */
+struct mlx5_lag {
+	u8                        flags;
+	u8                        v2p_map[MLX5_MAX_PORTS];
+	struct lag_func           pf[MLX5_MAX_PORTS];
+	struct lag_tracker        tracker;
+	struct delayed_work       bond_work;
+	struct notifier_block     nb;
+
+	/* Admin state. Allow lag only if allowed is true
+	 * even if network conditions for lag were met
+	 */
+	bool                      allowed;
+};
+
+/* General purpose, use for short periods of time.
+ * Beware of lock dependencies (preferably, no locks should be acquired
+ * under it).
+ */
+static DEFINE_MUTEX(lag_mutex);
+
+static int mlx5_cmd_create_lag(struct mlx5_core_dev *dev, u8 remap_port1,
+			       u8 remap_port2)
+{
+	u32   in[MLX5_ST_SZ_DW(create_lag_in)]   = {0};
+	u32   out[MLX5_ST_SZ_DW(create_lag_out)] = {0};
+	void *lag_ctx = MLX5_ADDR_OF(create_lag_in, in, ctx);
+
+	MLX5_SET(create_lag_in, in, opcode, MLX5_CMD_OP_CREATE_LAG);
+
+	MLX5_SET(lagc, lag_ctx, tx_remap_affinity_1, remap_port1);
+	MLX5_SET(lagc, lag_ctx, tx_remap_affinity_2, remap_port2);
+
+	return mlx5_cmd_exec(dev, in, sizeof(in), out, sizeof(out));
+}
+
+static int mlx5_cmd_modify_lag(struct mlx5_core_dev *dev, u8 remap_port1,
+			       u8 remap_port2)
+{
+	u32   in[MLX5_ST_SZ_DW(modify_lag_in)]   = {0};
+	u32   out[MLX5_ST_SZ_DW(modify_lag_out)] = {0};
+	void *lag_ctx = MLX5_ADDR_OF(modify_lag_in, in, ctx);
+
+	MLX5_SET(modify_lag_in, in, opcode, MLX5_CMD_OP_MODIFY_LAG);
+	MLX5_SET(modify_lag_in, in, field_select, 0x1);
+
+	MLX5_SET(lagc, lag_ctx, tx_remap_affinity_1, remap_port1);
+	MLX5_SET(lagc, lag_ctx, tx_remap_affinity_2, remap_port2);
+
+	return mlx5_cmd_exec(dev, in, sizeof(in), out, sizeof(out));
+}
+
+static int mlx5_cmd_destroy_lag(struct mlx5_core_dev *dev)
+{
+	u32  in[MLX5_ST_SZ_DW(destroy_lag_in)]  = {0};
+	u32 out[MLX5_ST_SZ_DW(destroy_lag_out)] = {0};
+
+	MLX5_SET(destroy_lag_in, in, opcode, MLX5_CMD_OP_DESTROY_LAG);
+
+	return mlx5_cmd_exec(dev, in, sizeof(in), out, sizeof(out));
+}
+
+int mlx5_cmd_create_vport_lag(struct mlx5_core_dev *dev)
+{
+	u32  in[MLX5_ST_SZ_DW(create_vport_lag_in)]  = {0};
+	u32 out[MLX5_ST_SZ_DW(create_vport_lag_out)] = {0};
+
+	MLX5_SET(create_vport_lag_in, in, opcode, MLX5_CMD_OP_CREATE_VPORT_LAG);
+
+	return mlx5_cmd_exec(dev, in, sizeof(in), out, sizeof(out));
+}
+EXPORT_SYMBOL(mlx5_cmd_create_vport_lag);
+
+int mlx5_cmd_destroy_vport_lag(struct mlx5_core_dev *dev)
+{
+	u32  in[MLX5_ST_SZ_DW(destroy_vport_lag_in)]  = {0};
+	u32 out[MLX5_ST_SZ_DW(destroy_vport_lag_out)] = {0};
+
+	MLX5_SET(destroy_vport_lag_in, in, opcode, MLX5_CMD_OP_DESTROY_VPORT_LAG);
+
+	return mlx5_cmd_exec(dev, in, sizeof(in), out, sizeof(out));
+}
+EXPORT_SYMBOL(mlx5_cmd_destroy_vport_lag);
+
+static int mlx5_cmd_query_cong_counter(struct mlx5_core_dev *dev,
+				       bool reset, void *out, int out_size)
+{
+	u32 in[MLX5_ST_SZ_DW(query_cong_statistics_in)] = { };
+
+	MLX5_SET(query_cong_statistics_in, in, opcode,
+		 MLX5_CMD_OP_QUERY_CONG_STATISTICS);
+	MLX5_SET(query_cong_statistics_in, in, clear, reset);
+	return mlx5_cmd_exec(dev, in, sizeof(in), out, out_size);
+}
+
+static struct mlx5_lag *mlx5_lag_dev_get(struct mlx5_core_dev *dev)
+{
+	return dev->priv.lag;
+}
+
+static int mlx5_lag_dev_get_netdev_idx(struct mlx5_lag *ldev,
+				       struct net_device *ndev)
+{
+	int i;
+
+	for (i = 0; i < MLX5_MAX_PORTS; i++)
+		if (ldev->pf[i].netdev == ndev)
+			return i;
+
+	return -1;
+}
+
+static bool mlx5_lag_is_bonded(struct mlx5_lag *ldev)
+{
+	return !!(ldev->flags & MLX5_LAG_FLAG_BONDED);
+}
+
+static void mlx5_infer_tx_affinity_mapping(struct lag_tracker *tracker,
+					   u8 *port1, u8 *port2)
+{
+	*port1 = 1;
+	*port2 = 2;
+	if (!tracker->netdev_state[0].tx_enabled ||
+	    !tracker->netdev_state[0].link_up) {
+		*port1 = 2;
+		return;
+	}
+
+	if (!tracker->netdev_state[1].tx_enabled ||
+	    !tracker->netdev_state[1].link_up)
+		*port2 = 1;
+}
+
+static void mlx5_activate_lag(struct mlx5_lag *ldev,
+			      struct lag_tracker *tracker)
+{
+	struct mlx5_core_dev *dev0 = ldev->pf[0].dev;
+	int err;
+
+	ldev->flags |= MLX5_LAG_FLAG_BONDED;
+
+	mlx5_infer_tx_affinity_mapping(tracker, &ldev->v2p_map[0],
+				       &ldev->v2p_map[1]);
+
+	err = mlx5_cmd_create_lag(dev0, ldev->v2p_map[0], ldev->v2p_map[1]);
+	if (err)
+		mlx5_core_err(dev0,
+			      "Failed to create LAG (%d)\n",
+			      err);
+}
+
+static void mlx5_deactivate_lag(struct mlx5_lag *ldev)
+{
+	struct mlx5_core_dev *dev0 = ldev->pf[0].dev;
+	int err;
+
+	ldev->flags &= ~MLX5_LAG_FLAG_BONDED;
+
+	err = mlx5_cmd_destroy_lag(dev0);
+	if (err)
+		mlx5_core_err(dev0,
+			      "Failed to destroy LAG (%d)\n",
+			      err);
+}
+
+static void mlx5_do_bond(struct mlx5_lag *ldev)
+{
+	struct mlx5_core_dev *dev0 = ldev->pf[0].dev;
+	struct mlx5_core_dev *dev1 = ldev->pf[1].dev;
+	struct lag_tracker tracker;
+	u8 v2p_port1, v2p_port2;
+	int i, err;
+	bool do_bond;
+
+	if (!dev0 || !dev1)
+		return;
+
+	mutex_lock(&lag_mutex);
+	tracker = ldev->tracker;
+	mutex_unlock(&lag_mutex);
+
+	do_bond = tracker.is_bonded && ldev->allowed;
+
+	if (do_bond && !mlx5_lag_is_bonded(ldev)) {
+		for (i = 0; i < MLX5_MAX_PORTS; i++)
+			mlx5_remove_dev_by_protocol(ldev->pf[i].dev,
+						    MLX5_INTERFACE_PROTOCOL_IB);
+
+		mlx5_activate_lag(ldev, &tracker);
+
+		mlx5_add_dev_by_protocol(dev0, MLX5_INTERFACE_PROTOCOL_IB);
+		mlx5_nic_vport_enable_roce(dev1);
+	} else if (do_bond && mlx5_lag_is_bonded(ldev)) {
+		mlx5_infer_tx_affinity_mapping(&tracker, &v2p_port1,
+					       &v2p_port2);
+
+		if ((v2p_port1 != ldev->v2p_map[0]) ||
+		    (v2p_port2 != ldev->v2p_map[1])) {
+			ldev->v2p_map[0] = v2p_port1;
+			ldev->v2p_map[1] = v2p_port2;
+
+			err = mlx5_cmd_modify_lag(dev0, v2p_port1, v2p_port2);
+			if (err)
+				mlx5_core_err(dev0,
+					      "Failed to modify LAG (%d)\n",
+					      err);
+		}
+	} else if (!do_bond && mlx5_lag_is_bonded(ldev)) {
+		mlx5_remove_dev_by_protocol(dev0, MLX5_INTERFACE_PROTOCOL_IB);
+		mlx5_nic_vport_disable_roce(dev1);
+
+		mlx5_deactivate_lag(ldev);
+
+		for (i = 0; i < MLX5_MAX_PORTS; i++)
+			if (ldev->pf[i].dev)
+				mlx5_add_dev_by_protocol(ldev->pf[i].dev,
+							 MLX5_INTERFACE_PROTOCOL_IB);
+	}
+}
+
+static void mlx5_queue_bond_work(struct mlx5_lag *ldev, unsigned long delay)
+{
+	schedule_delayed_work(&ldev->bond_work, delay);
+}
+
+static void mlx5_do_bond_work(struct work_struct *work)
+{
+	struct delayed_work *delayed_work = to_delayed_work(work);
+	struct mlx5_lag *ldev = container_of(delayed_work, struct mlx5_lag,
+					     bond_work);
+	int status;
+
+	status = mlx5_dev_list_trylock();
+	if (!status) {
+		/* 1 sec delay. */
+		mlx5_queue_bond_work(ldev, HZ);
+		return;
+	}
+
+	mlx5_do_bond(ldev);
+	mlx5_dev_list_unlock();
+}
+
+static int mlx5_handle_changeupper_event(struct mlx5_lag *ldev,
+					 struct lag_tracker *tracker,
+					 struct net_device *ndev,
+					 struct netdev_notifier_changeupper_info *info)
+{
+	struct net_device *upper = info->upper_dev, *ndev_tmp;
+	struct netdev_lag_upper_info *lag_upper_info = NULL;
+	bool is_bonded;
+	int bond_status = 0;
+	int num_slaves = 0;
+	int idx;
+
+	if (!netif_is_lag_master(upper))
+		return 0;
+
+	if (info->linking)
+		lag_upper_info = info->upper_info;
+
+	/* The event may still be of interest if the slave does not belong to
+	 * us, but is enslaved to a master which has one or more of our netdevs
+	 * as slaves (e.g., if a new slave is added to a master that bonds two
+	 * of our netdevs, we should unbond).
+	 */
+	rcu_read_lock();
+	for_each_netdev_in_bond_rcu(upper, ndev_tmp) {
+		idx = mlx5_lag_dev_get_netdev_idx(ldev, ndev_tmp);
+		if (idx > -1)
+			bond_status |= (1 << idx);
+
+		num_slaves++;
+	}
+	rcu_read_unlock();
+
+	/* None of this lagdev's netdevs are slaves of this master. */
+	if (!(bond_status & 0x3))
+		return 0;
+
+	if (lag_upper_info)
+		tracker->tx_type = lag_upper_info->tx_type;
+
+	/* Determine bonding status:
+	 * A device is considered bonded if both its physical ports are slaves
+	 * of the same lag master, and only them.
+	 * Lag mode must be activebackup or hash.
+	 */
+	is_bonded = (num_slaves == MLX5_MAX_PORTS) &&
+		    (bond_status == 0x3) &&
+		    ((tracker->tx_type == NETDEV_LAG_TX_TYPE_ACTIVEBACKUP) ||
+		     (tracker->tx_type == NETDEV_LAG_TX_TYPE_HASH));
+
+	if (tracker->is_bonded != is_bonded) {
+		tracker->is_bonded = is_bonded;
+		return 1;
+	}
+
+	return 0;
+}
+
+static int mlx5_handle_changelowerstate_event(struct mlx5_lag *ldev,
+					      struct lag_tracker *tracker,
+					      struct net_device *ndev,
+					      struct netdev_notifier_changelowerstate_info *info)
+{
+	struct netdev_lag_lower_state_info *lag_lower_info;
+	int idx;
+
+	if (!netif_is_lag_port(ndev))
+		return 0;
+
+	idx = mlx5_lag_dev_get_netdev_idx(ldev, ndev);
+	if (idx == -1)
+		return 0;
+
+	/* This information is used to determine virtual to physical
+	 * port mapping.
+	 */
+	lag_lower_info = info->lower_state_info;
+	if (!lag_lower_info)
+		return 0;
+
+	tracker->netdev_state[idx] = *lag_lower_info;
+
+	return 1;
+}
+
+static int mlx5_lag_netdev_event(struct notifier_block *this,
+				 unsigned long event, void *ptr)
+{
+	struct net_device *ndev = netdev_notifier_info_to_dev(ptr);
+	struct lag_tracker tracker;
+	struct mlx5_lag *ldev;
+	int changed = 0;
+
+	if (!net_eq(dev_net(ndev), &init_net))
+		return NOTIFY_DONE;
+
+	if ((event != NETDEV_CHANGEUPPER) && (event != NETDEV_CHANGELOWERSTATE))
+		return NOTIFY_DONE;
+
+	ldev    = container_of(this, struct mlx5_lag, nb);
+	tracker = ldev->tracker;
+
+	switch (event) {
+	case NETDEV_CHANGEUPPER:
+		changed = mlx5_handle_changeupper_event(ldev, &tracker, ndev,
+							ptr);
+		break;
+	case NETDEV_CHANGELOWERSTATE:
+		changed = mlx5_handle_changelowerstate_event(ldev, &tracker,
+							     ndev, ptr);
+		break;
+	}
+
+	mutex_lock(&lag_mutex);
+	ldev->tracker = tracker;
+	mutex_unlock(&lag_mutex);
+
+	if (changed)
+		mlx5_queue_bond_work(ldev, 0);
+
+	return NOTIFY_DONE;
+}
+
+static bool mlx5_lag_check_prereq(struct mlx5_lag *ldev)
+{
+	if ((ldev->pf[0].dev && mlx5_sriov_is_enabled(ldev->pf[0].dev)) ||
+	    (ldev->pf[1].dev && mlx5_sriov_is_enabled(ldev->pf[1].dev)))
+		return false;
+	else
+		return true;
+}
+
+static struct mlx5_lag *mlx5_lag_dev_alloc(void)
+{
+	struct mlx5_lag *ldev;
+
+	ldev = kzalloc(sizeof(*ldev), GFP_KERNEL);
+	if (!ldev)
+		return NULL;
+
+	INIT_DELAYED_WORK(&ldev->bond_work, mlx5_do_bond_work);
+	ldev->allowed = mlx5_lag_check_prereq(ldev);
+
+	return ldev;
+}
+
+static void mlx5_lag_dev_free(struct mlx5_lag *ldev)
+{
+	kfree(ldev);
+}
+
+static void mlx5_lag_dev_add_pf(struct mlx5_lag *ldev,
+				struct mlx5_core_dev *dev,
+				struct net_device *netdev)
+{
+	unsigned int fn = PCI_FUNC(dev->pdev->devfn);
+
+	if (fn >= MLX5_MAX_PORTS)
+		return;
+
+	mutex_lock(&lag_mutex);
+	ldev->pf[fn].dev    = dev;
+	ldev->pf[fn].netdev = netdev;
+	ldev->tracker.netdev_state[fn].link_up = 0;
+	ldev->tracker.netdev_state[fn].tx_enabled = 0;
+
+	ldev->allowed = mlx5_lag_check_prereq(ldev);
+	dev->priv.lag = ldev;
+
+	mutex_unlock(&lag_mutex);
+}
+
+static void mlx5_lag_dev_remove_pf(struct mlx5_lag *ldev,
+				   struct mlx5_core_dev *dev)
+{
+	int i;
+
+	for (i = 0; i < MLX5_MAX_PORTS; i++)
+		if (ldev->pf[i].dev == dev)
+			break;
+
+	if (i == MLX5_MAX_PORTS)
+		return;
+
+	mutex_lock(&lag_mutex);
+	memset(&ldev->pf[i], 0, sizeof(*ldev->pf));
+
+	dev->priv.lag = NULL;
+	ldev->allowed = mlx5_lag_check_prereq(ldev);
+	mutex_unlock(&lag_mutex);
+}
+
+/* Must be called with intf_mutex held */
+void mlx5_lag_add(struct mlx5_core_dev *dev, struct net_device *netdev)
+{
+	struct mlx5_lag *ldev = NULL;
+	struct mlx5_core_dev *tmp_dev;
+
+	if (!MLX5_CAP_GEN(dev, vport_group_manager) ||
+	    !MLX5_CAP_GEN(dev, lag_master) ||
+	    (MLX5_CAP_GEN(dev, num_lag_ports) != MLX5_MAX_PORTS))
+		return;
+
+	tmp_dev = mlx5_get_next_phys_dev(dev);
+	if (tmp_dev)
+		ldev = tmp_dev->priv.lag;
+
+	if (!ldev) {
+		ldev = mlx5_lag_dev_alloc();
+		if (!ldev) {
+			mlx5_core_err(dev, "Failed to alloc lag dev\n");
+			return;
+		}
+	}
+
+	mlx5_lag_dev_add_pf(ldev, dev, netdev);
+
+	if (!ldev->nb.notifier_call) {
+		ldev->nb.notifier_call = mlx5_lag_netdev_event;
+		if (register_netdevice_notifier(&ldev->nb)) {
+			ldev->nb.notifier_call = NULL;
+			mlx5_core_err(dev, "Failed to register LAG netdev notifier\n");
+		}
+	}
+}
+
+/* Must be called with intf_mutex held */
+void mlx5_lag_remove(struct mlx5_core_dev *dev)
+{
+	struct mlx5_lag *ldev;
+	int i;
+
+	ldev = mlx5_lag_dev_get(dev);
+	if (!ldev)
+		return;
+
+	if (mlx5_lag_is_bonded(ldev))
+		mlx5_deactivate_lag(ldev);
+
+	mlx5_lag_dev_remove_pf(ldev, dev);
+
+	for (i = 0; i < MLX5_MAX_PORTS; i++)
+		if (ldev->pf[i].dev)
+			break;
+
+	if (i == MLX5_MAX_PORTS) {
+		if (ldev->nb.notifier_call)
+			unregister_netdevice_notifier(&ldev->nb);
+		cancel_delayed_work_sync(&ldev->bond_work);
+		mlx5_lag_dev_free(ldev);
+	}
+}
+
+bool mlx5_lag_is_active(struct mlx5_core_dev *dev)
+{
+	struct mlx5_lag *ldev;
+	bool res;
+
+	mutex_lock(&lag_mutex);
+	ldev = mlx5_lag_dev_get(dev);
+	res  = ldev && mlx5_lag_is_bonded(ldev);
+	mutex_unlock(&lag_mutex);
+
+	return res;
+}
+EXPORT_SYMBOL(mlx5_lag_is_active);
+
+static int mlx5_lag_set_state(struct mlx5_core_dev *dev, bool allow)
+{
+	struct mlx5_lag *ldev;
+	int ret = 0;
+	bool lag_active;
+
+	mlx5_dev_list_lock();
+
+	ldev = mlx5_lag_dev_get(dev);
+	if (!ldev) {
+		ret = -ENODEV;
+		goto unlock;
+	}
+	lag_active = mlx5_lag_is_bonded(ldev);
+	if (!mlx5_lag_check_prereq(ldev) && allow) {
+		ret = -EINVAL;
+		goto unlock;
+	}
+	if (ldev->allowed == allow)
+		goto unlock;
+	ldev->allowed = allow;
+	if ((lag_active && !allow) || allow)
+		mlx5_do_bond(ldev);
+unlock:
+	mlx5_dev_list_unlock();
+	return ret;
+}
+
+int mlx5_lag_forbid(struct mlx5_core_dev *dev)
+{
+	return mlx5_lag_set_state(dev, false);
+}
+
+int mlx5_lag_allow(struct mlx5_core_dev *dev)
+{
+	return mlx5_lag_set_state(dev, true);
+}
+
+struct net_device *mlx5_lag_get_roce_netdev(struct mlx5_core_dev *dev)
+{
+	struct net_device *ndev = NULL;
+	struct mlx5_lag *ldev;
+
+	mutex_lock(&lag_mutex);
+	ldev = mlx5_lag_dev_get(dev);
+
+	if (!(ldev && mlx5_lag_is_bonded(ldev)))
+		goto unlock;
+
+	if (ldev->tracker.tx_type == NETDEV_LAG_TX_TYPE_ACTIVEBACKUP) {
+		ndev = ldev->tracker.netdev_state[0].tx_enabled ?
+		       ldev->pf[0].netdev : ldev->pf[1].netdev;
+	} else {
+		ndev = ldev->pf[0].netdev;
+	}
+	if (ndev)
+		dev_hold(ndev);
+
+unlock:
+	mutex_unlock(&lag_mutex);
+
+	return ndev;
+}
+EXPORT_SYMBOL(mlx5_lag_get_roce_netdev);
+
+bool mlx5_lag_intf_add(struct mlx5_interface *intf, struct mlx5_priv *priv)
+{
+	struct mlx5_core_dev *dev = container_of(priv, struct mlx5_core_dev,
+						 priv);
+	struct mlx5_lag *ldev;
+
+	if (intf->protocol != MLX5_INTERFACE_PROTOCOL_IB)
+		return true;
+
+	ldev = mlx5_lag_dev_get(dev);
+	if (!ldev || !mlx5_lag_is_bonded(ldev) || ldev->pf[0].dev == dev)
+		return true;
+
+	/* If bonded, we do not add an IB device for PF1. */
+	return false;
+}
+
+int mlx5_lag_query_cong_counters(struct mlx5_core_dev *dev,
+				 u64 *values,
+				 int num_counters,
+				 size_t *offsets)
+{
+	int outlen = MLX5_ST_SZ_BYTES(query_cong_statistics_out);
+	struct mlx5_core_dev *mdev[MLX5_MAX_PORTS];
+	struct mlx5_lag *ldev;
+	int num_ports;
+	int ret, i, j;
+	void *out;
+
+	out = kvzalloc(outlen, GFP_KERNEL);
+	if (!out)
+		return -ENOMEM;
+
+	memset(values, 0, sizeof(*values) * num_counters);
+
+	mutex_lock(&lag_mutex);
+	ldev = mlx5_lag_dev_get(dev);
+	if (ldev && mlx5_lag_is_bonded(ldev)) {
+		num_ports = MLX5_MAX_PORTS;
+		mdev[0] = ldev->pf[0].dev;
+		mdev[1] = ldev->pf[1].dev;
+	} else {
+		num_ports = 1;
+		mdev[0] = dev;
+	}
+
+	for (i = 0; i < num_ports; ++i) {
+		ret = mlx5_cmd_query_cong_counter(mdev[i], false, out, outlen);
+		if (ret)
+			goto unlock;
+
+		for (j = 0; j < num_counters; ++j)
+			values[j] += be64_to_cpup((__be64 *)(out + offsets[j]));
+	}
+
+unlock:
+	mutex_unlock(&lag_mutex);
+	kvfree(out);
+	return ret;
+}
+EXPORT_SYMBOL(mlx5_lag_query_cong_counters);
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/lib/Makefile b/drivers/net/ethernet/mellanox/mlx5/core/lib/Makefile
new file mode 100644
index 000000000..d8e17110f
--- /dev/null
+++ b/drivers/net/ethernet/mellanox/mlx5/core/lib/Makefile
@@ -0,0 +1 @@
+subdir-ccflags-y += -I$(src)/..
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/lib/clock.c b/drivers/net/ethernet/mellanox/mlx5/core/lib/clock.c
new file mode 100644
index 000000000..0fd62510f
--- /dev/null
+++ b/drivers/net/ethernet/mellanox/mlx5/core/lib/clock.c
@@ -0,0 +1,616 @@
+/*
+ * Copyright (c) 2015, 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/clocksource.h>
+#include <linux/highmem.h>
+#include <rdma/mlx5-abi.h>
+#include "en.h"
+#include "clock.h"
+
+enum {
+	MLX5_CYCLES_SHIFT	= 23
+};
+
+enum {
+	MLX5_PIN_MODE_IN		= 0x0,
+	MLX5_PIN_MODE_OUT		= 0x1,
+};
+
+enum {
+	MLX5_OUT_PATTERN_PULSE		= 0x0,
+	MLX5_OUT_PATTERN_PERIODIC	= 0x1,
+};
+
+enum {
+	MLX5_EVENT_MODE_DISABLE	= 0x0,
+	MLX5_EVENT_MODE_REPETETIVE	= 0x1,
+	MLX5_EVENT_MODE_ONCE_TILL_ARM	= 0x2,
+};
+
+enum {
+	MLX5_MTPPS_FS_ENABLE			= BIT(0x0),
+	MLX5_MTPPS_FS_PATTERN			= BIT(0x2),
+	MLX5_MTPPS_FS_PIN_MODE			= BIT(0x3),
+	MLX5_MTPPS_FS_TIME_STAMP		= BIT(0x4),
+	MLX5_MTPPS_FS_OUT_PULSE_DURATION	= BIT(0x5),
+	MLX5_MTPPS_FS_ENH_OUT_PER_ADJ		= BIT(0x7),
+};
+
+static u64 read_internal_timer(const struct cyclecounter *cc)
+{
+	struct mlx5_clock *clock = container_of(cc, struct mlx5_clock, cycles);
+	struct mlx5_core_dev *mdev = container_of(clock, struct mlx5_core_dev,
+						  clock);
+
+	return mlx5_read_internal_timer(mdev) & cc->mask;
+}
+
+static void mlx5_update_clock_info_page(struct mlx5_core_dev *mdev)
+{
+	struct mlx5_ib_clock_info *clock_info = mdev->clock_info;
+	struct mlx5_clock *clock = &mdev->clock;
+	u32 sign;
+
+	if (!clock_info)
+		return;
+
+	sign = smp_load_acquire(&clock_info->sign);
+	smp_store_mb(clock_info->sign,
+		     sign | MLX5_IB_CLOCK_INFO_KERNEL_UPDATING);
+
+	clock_info->cycles = clock->tc.cycle_last;
+	clock_info->mult   = clock->cycles.mult;
+	clock_info->nsec   = clock->tc.nsec;
+	clock_info->frac   = clock->tc.frac;
+
+	smp_store_release(&clock_info->sign,
+			  sign + MLX5_IB_CLOCK_INFO_KERNEL_UPDATING * 2);
+}
+
+static void mlx5_pps_out(struct work_struct *work)
+{
+	struct mlx5_pps *pps_info = container_of(work, struct mlx5_pps,
+						 out_work);
+	struct mlx5_clock *clock = container_of(pps_info, struct mlx5_clock,
+						pps_info);
+	struct mlx5_core_dev *mdev = container_of(clock, struct mlx5_core_dev,
+						  clock);
+	u32 in[MLX5_ST_SZ_DW(mtpps_reg)] = {0};
+	unsigned long flags;
+	int i;
+
+	for (i = 0; i < clock->ptp_info.n_pins; i++) {
+		u64 tstart;
+
+		write_lock_irqsave(&clock->lock, flags);
+		tstart = clock->pps_info.start[i];
+		clock->pps_info.start[i] = 0;
+		write_unlock_irqrestore(&clock->lock, flags);
+		if (!tstart)
+			continue;
+
+		MLX5_SET(mtpps_reg, in, pin, i);
+		MLX5_SET64(mtpps_reg, in, time_stamp, tstart);
+		MLX5_SET(mtpps_reg, in, field_select, MLX5_MTPPS_FS_TIME_STAMP);
+		mlx5_set_mtpps(mdev, in, sizeof(in));
+	}
+}
+
+static void mlx5_timestamp_overflow(struct work_struct *work)
+{
+	struct delayed_work *dwork = to_delayed_work(work);
+	struct mlx5_clock *clock = container_of(dwork, struct mlx5_clock,
+						overflow_work);
+	unsigned long flags;
+
+	write_lock_irqsave(&clock->lock, flags);
+	timecounter_read(&clock->tc);
+	mlx5_update_clock_info_page(clock->mdev);
+	write_unlock_irqrestore(&clock->lock, flags);
+	schedule_delayed_work(&clock->overflow_work, clock->overflow_period);
+}
+
+static int mlx5_ptp_settime(struct ptp_clock_info *ptp,
+			    const struct timespec64 *ts)
+{
+	struct mlx5_clock *clock = container_of(ptp, struct mlx5_clock,
+						 ptp_info);
+	u64 ns = timespec64_to_ns(ts);
+	unsigned long flags;
+
+	write_lock_irqsave(&clock->lock, flags);
+	timecounter_init(&clock->tc, &clock->cycles, ns);
+	mlx5_update_clock_info_page(clock->mdev);
+	write_unlock_irqrestore(&clock->lock, flags);
+
+	return 0;
+}
+
+static int mlx5_ptp_gettime(struct ptp_clock_info *ptp, struct timespec64 *ts)
+{
+	struct mlx5_clock *clock = container_of(ptp, struct mlx5_clock,
+						ptp_info);
+	u64 ns;
+	unsigned long flags;
+
+	write_lock_irqsave(&clock->lock, flags);
+	ns = timecounter_read(&clock->tc);
+	write_unlock_irqrestore(&clock->lock, flags);
+
+	*ts = ns_to_timespec64(ns);
+
+	return 0;
+}
+
+static int mlx5_ptp_adjtime(struct ptp_clock_info *ptp, s64 delta)
+{
+	struct mlx5_clock *clock = container_of(ptp, struct mlx5_clock,
+						ptp_info);
+	unsigned long flags;
+
+	write_lock_irqsave(&clock->lock, flags);
+	timecounter_adjtime(&clock->tc, delta);
+	mlx5_update_clock_info_page(clock->mdev);
+	write_unlock_irqrestore(&clock->lock, flags);
+
+	return 0;
+}
+
+static int mlx5_ptp_adjfreq(struct ptp_clock_info *ptp, s32 delta)
+{
+	u64 adj;
+	u32 diff;
+	unsigned long flags;
+	int neg_adj = 0;
+	struct mlx5_clock *clock = container_of(ptp, struct mlx5_clock,
+						ptp_info);
+
+	if (delta < 0) {
+		neg_adj = 1;
+		delta = -delta;
+	}
+
+	adj = clock->nominal_c_mult;
+	adj *= delta;
+	diff = div_u64(adj, 1000000000ULL);
+
+	write_lock_irqsave(&clock->lock, flags);
+	timecounter_read(&clock->tc);
+	clock->cycles.mult = neg_adj ? clock->nominal_c_mult - diff :
+				       clock->nominal_c_mult + diff;
+	mlx5_update_clock_info_page(clock->mdev);
+	write_unlock_irqrestore(&clock->lock, flags);
+
+	return 0;
+}
+
+static int mlx5_extts_configure(struct ptp_clock_info *ptp,
+				struct ptp_clock_request *rq,
+				int on)
+{
+	struct mlx5_clock *clock =
+			container_of(ptp, struct mlx5_clock, ptp_info);
+	struct mlx5_core_dev *mdev =
+			container_of(clock, struct mlx5_core_dev, clock);
+	u32 in[MLX5_ST_SZ_DW(mtpps_reg)] = {0};
+	u32 field_select = 0;
+	u8 pin_mode = 0;
+	u8 pattern = 0;
+	int pin = -1;
+	int err = 0;
+
+	if (!MLX5_PPS_CAP(mdev))
+		return -EOPNOTSUPP;
+
+	if (rq->extts.index >= clock->ptp_info.n_pins)
+		return -EINVAL;
+
+	if (on) {
+		pin = ptp_find_pin(clock->ptp, PTP_PF_EXTTS, rq->extts.index);
+		if (pin < 0)
+			return -EBUSY;
+		pin_mode = MLX5_PIN_MODE_IN;
+		pattern = !!(rq->extts.flags & PTP_FALLING_EDGE);
+		field_select = MLX5_MTPPS_FS_PIN_MODE |
+			       MLX5_MTPPS_FS_PATTERN |
+			       MLX5_MTPPS_FS_ENABLE;
+	} else {
+		pin = rq->extts.index;
+		field_select = MLX5_MTPPS_FS_ENABLE;
+	}
+
+	MLX5_SET(mtpps_reg, in, pin, pin);
+	MLX5_SET(mtpps_reg, in, pin_mode, pin_mode);
+	MLX5_SET(mtpps_reg, in, pattern, pattern);
+	MLX5_SET(mtpps_reg, in, enable, on);
+	MLX5_SET(mtpps_reg, in, field_select, field_select);
+
+	err = mlx5_set_mtpps(mdev, in, sizeof(in));
+	if (err)
+		return err;
+
+	return mlx5_set_mtppse(mdev, pin, 0,
+			       MLX5_EVENT_MODE_REPETETIVE & on);
+}
+
+static int mlx5_perout_configure(struct ptp_clock_info *ptp,
+				 struct ptp_clock_request *rq,
+				 int on)
+{
+	struct mlx5_clock *clock =
+			container_of(ptp, struct mlx5_clock, ptp_info);
+	struct mlx5_core_dev *mdev =
+			container_of(clock, struct mlx5_core_dev, clock);
+	u32 in[MLX5_ST_SZ_DW(mtpps_reg)] = {0};
+	u64 nsec_now, nsec_delta, time_stamp = 0;
+	u64 cycles_now, cycles_delta;
+	struct timespec64 ts;
+	unsigned long flags;
+	u32 field_select = 0;
+	u8 pin_mode = 0;
+	u8 pattern = 0;
+	int pin = -1;
+	int err = 0;
+	s64 ns;
+
+	if (!MLX5_PPS_CAP(mdev))
+		return -EOPNOTSUPP;
+
+	if (rq->perout.index >= clock->ptp_info.n_pins)
+		return -EINVAL;
+
+	if (on) {
+		pin = ptp_find_pin(clock->ptp, PTP_PF_PEROUT,
+				   rq->perout.index);
+		if (pin < 0)
+			return -EBUSY;
+
+		pin_mode = MLX5_PIN_MODE_OUT;
+		pattern = MLX5_OUT_PATTERN_PERIODIC;
+		ts.tv_sec = rq->perout.period.sec;
+		ts.tv_nsec = rq->perout.period.nsec;
+		ns = timespec64_to_ns(&ts);
+
+		if ((ns >> 1) != 500000000LL)
+			return -EINVAL;
+
+		ts.tv_sec = rq->perout.start.sec;
+		ts.tv_nsec = rq->perout.start.nsec;
+		ns = timespec64_to_ns(&ts);
+		cycles_now = mlx5_read_internal_timer(mdev);
+		write_lock_irqsave(&clock->lock, flags);
+		nsec_now = timecounter_cyc2time(&clock->tc, cycles_now);
+		nsec_delta = ns - nsec_now;
+		cycles_delta = div64_u64(nsec_delta << clock->cycles.shift,
+					 clock->cycles.mult);
+		write_unlock_irqrestore(&clock->lock, flags);
+		time_stamp = cycles_now + cycles_delta;
+		field_select = MLX5_MTPPS_FS_PIN_MODE |
+			       MLX5_MTPPS_FS_PATTERN |
+			       MLX5_MTPPS_FS_ENABLE |
+			       MLX5_MTPPS_FS_TIME_STAMP;
+	} else {
+		pin = rq->perout.index;
+		field_select = MLX5_MTPPS_FS_ENABLE;
+	}
+
+	MLX5_SET(mtpps_reg, in, pin, pin);
+	MLX5_SET(mtpps_reg, in, pin_mode, pin_mode);
+	MLX5_SET(mtpps_reg, in, pattern, pattern);
+	MLX5_SET(mtpps_reg, in, enable, on);
+	MLX5_SET64(mtpps_reg, in, time_stamp, time_stamp);
+	MLX5_SET(mtpps_reg, in, field_select, field_select);
+
+	err = mlx5_set_mtpps(mdev, in, sizeof(in));
+	if (err)
+		return err;
+
+	return mlx5_set_mtppse(mdev, pin, 0,
+			       MLX5_EVENT_MODE_REPETETIVE & on);
+}
+
+static int mlx5_pps_configure(struct ptp_clock_info *ptp,
+			      struct ptp_clock_request *rq,
+			      int on)
+{
+	struct mlx5_clock *clock =
+			container_of(ptp, struct mlx5_clock, ptp_info);
+
+	clock->pps_info.enabled = !!on;
+	return 0;
+}
+
+static int mlx5_ptp_enable(struct ptp_clock_info *ptp,
+			   struct ptp_clock_request *rq,
+			   int on)
+{
+	switch (rq->type) {
+	case PTP_CLK_REQ_EXTTS:
+		return mlx5_extts_configure(ptp, rq, on);
+	case PTP_CLK_REQ_PEROUT:
+		return mlx5_perout_configure(ptp, rq, on);
+	case PTP_CLK_REQ_PPS:
+		return mlx5_pps_configure(ptp, rq, on);
+	default:
+		return -EOPNOTSUPP;
+	}
+	return 0;
+}
+
+enum {
+	MLX5_MTPPS_REG_CAP_PIN_X_MODE_SUPPORT_PPS_IN = BIT(0),
+	MLX5_MTPPS_REG_CAP_PIN_X_MODE_SUPPORT_PPS_OUT = BIT(1),
+};
+
+static int mlx5_ptp_verify(struct ptp_clock_info *ptp, unsigned int pin,
+			   enum ptp_pin_function func, unsigned int chan)
+{
+	struct mlx5_clock *clock = container_of(ptp, struct mlx5_clock,
+						ptp_info);
+
+	switch (func) {
+	case PTP_PF_NONE:
+		return 0;
+	case PTP_PF_EXTTS:
+		return !(clock->pps_info.pin_caps[pin] &
+			 MLX5_MTPPS_REG_CAP_PIN_X_MODE_SUPPORT_PPS_IN);
+	case PTP_PF_PEROUT:
+		return !(clock->pps_info.pin_caps[pin] &
+			 MLX5_MTPPS_REG_CAP_PIN_X_MODE_SUPPORT_PPS_OUT);
+	default:
+		return -EOPNOTSUPP;
+	}
+
+	return -EOPNOTSUPP;
+}
+
+static const struct ptp_clock_info mlx5_ptp_clock_info = {
+	.owner		= THIS_MODULE,
+	.name		= "mlx5_p2p",
+	.max_adj	= 100000000,
+	.n_alarm	= 0,
+	.n_ext_ts	= 0,
+	.n_per_out	= 0,
+	.n_pins		= 0,
+	.pps		= 0,
+	.adjfreq	= mlx5_ptp_adjfreq,
+	.adjtime	= mlx5_ptp_adjtime,
+	.gettime64	= mlx5_ptp_gettime,
+	.settime64	= mlx5_ptp_settime,
+	.enable		= NULL,
+	.verify		= NULL,
+};
+
+static int mlx5_init_pin_config(struct mlx5_clock *clock)
+{
+	int i;
+
+	clock->ptp_info.pin_config =
+			kcalloc(clock->ptp_info.n_pins,
+				sizeof(*clock->ptp_info.pin_config),
+				GFP_KERNEL);
+	if (!clock->ptp_info.pin_config)
+		return -ENOMEM;
+	clock->ptp_info.enable = mlx5_ptp_enable;
+	clock->ptp_info.verify = mlx5_ptp_verify;
+	clock->ptp_info.pps = 1;
+
+	for (i = 0; i < clock->ptp_info.n_pins; i++) {
+		snprintf(clock->ptp_info.pin_config[i].name,
+			 sizeof(clock->ptp_info.pin_config[i].name),
+			 "mlx5_pps%d", i);
+		clock->ptp_info.pin_config[i].index = i;
+		clock->ptp_info.pin_config[i].func = PTP_PF_NONE;
+		clock->ptp_info.pin_config[i].chan = i;
+	}
+
+	return 0;
+}
+
+static void mlx5_get_pps_caps(struct mlx5_core_dev *mdev)
+{
+	struct mlx5_clock *clock = &mdev->clock;
+	u32 out[MLX5_ST_SZ_DW(mtpps_reg)] = {0};
+
+	mlx5_query_mtpps(mdev, out, sizeof(out));
+
+	clock->ptp_info.n_pins = MLX5_GET(mtpps_reg, out,
+					  cap_number_of_pps_pins);
+	clock->ptp_info.n_ext_ts = MLX5_GET(mtpps_reg, out,
+					    cap_max_num_of_pps_in_pins);
+	clock->ptp_info.n_per_out = MLX5_GET(mtpps_reg, out,
+					     cap_max_num_of_pps_out_pins);
+
+	clock->pps_info.pin_caps[0] = MLX5_GET(mtpps_reg, out, cap_pin_0_mode);
+	clock->pps_info.pin_caps[1] = MLX5_GET(mtpps_reg, out, cap_pin_1_mode);
+	clock->pps_info.pin_caps[2] = MLX5_GET(mtpps_reg, out, cap_pin_2_mode);
+	clock->pps_info.pin_caps[3] = MLX5_GET(mtpps_reg, out, cap_pin_3_mode);
+	clock->pps_info.pin_caps[4] = MLX5_GET(mtpps_reg, out, cap_pin_4_mode);
+	clock->pps_info.pin_caps[5] = MLX5_GET(mtpps_reg, out, cap_pin_5_mode);
+	clock->pps_info.pin_caps[6] = MLX5_GET(mtpps_reg, out, cap_pin_6_mode);
+	clock->pps_info.pin_caps[7] = MLX5_GET(mtpps_reg, out, cap_pin_7_mode);
+}
+
+void mlx5_pps_event(struct mlx5_core_dev *mdev,
+		    struct mlx5_eqe *eqe)
+{
+	struct mlx5_clock *clock = &mdev->clock;
+	struct ptp_clock_event ptp_event;
+	struct timespec64 ts;
+	u64 nsec_now, nsec_delta;
+	u64 cycles_now, cycles_delta;
+	int pin = eqe->data.pps.pin;
+	s64 ns;
+	unsigned long flags;
+
+	switch (clock->ptp_info.pin_config[pin].func) {
+	case PTP_PF_EXTTS:
+		ptp_event.index = pin;
+		ptp_event.timestamp =
+			mlx5_timecounter_cyc2time(clock,
+						  be64_to_cpu(eqe->data.pps.time_stamp));
+		if (clock->pps_info.enabled) {
+			ptp_event.type = PTP_CLOCK_PPSUSR;
+			ptp_event.pps_times.ts_real =
+					ns_to_timespec64(ptp_event.timestamp);
+		} else {
+			ptp_event.type = PTP_CLOCK_EXTTS;
+		}
+		ptp_clock_event(clock->ptp, &ptp_event);
+		break;
+	case PTP_PF_PEROUT:
+		mlx5_ptp_gettime(&clock->ptp_info, &ts);
+		cycles_now = mlx5_read_internal_timer(mdev);
+		ts.tv_sec += 1;
+		ts.tv_nsec = 0;
+		ns = timespec64_to_ns(&ts);
+		write_lock_irqsave(&clock->lock, flags);
+		nsec_now = timecounter_cyc2time(&clock->tc, cycles_now);
+		nsec_delta = ns - nsec_now;
+		cycles_delta = div64_u64(nsec_delta << clock->cycles.shift,
+					 clock->cycles.mult);
+		clock->pps_info.start[pin] = cycles_now + cycles_delta;
+		schedule_work(&clock->pps_info.out_work);
+		write_unlock_irqrestore(&clock->lock, flags);
+		break;
+	default:
+		mlx5_core_err(mdev, " Unhandled event\n");
+	}
+}
+
+void mlx5_init_clock(struct mlx5_core_dev *mdev)
+{
+	struct mlx5_clock *clock = &mdev->clock;
+	u64 overflow_cycles;
+	u64 ns;
+	u64 frac = 0;
+	u32 dev_freq;
+
+	dev_freq = MLX5_CAP_GEN(mdev, device_frequency_khz);
+	if (!dev_freq) {
+		mlx5_core_warn(mdev, "invalid device_frequency_khz, aborting HW clock init\n");
+		return;
+	}
+	rwlock_init(&clock->lock);
+	clock->cycles.read = read_internal_timer;
+	clock->cycles.shift = MLX5_CYCLES_SHIFT;
+	clock->cycles.mult = clocksource_khz2mult(dev_freq,
+						  clock->cycles.shift);
+	clock->nominal_c_mult = clock->cycles.mult;
+	clock->cycles.mask = CLOCKSOURCE_MASK(41);
+	clock->mdev = mdev;
+
+	timecounter_init(&clock->tc, &clock->cycles,
+			 ktime_to_ns(ktime_get_real()));
+
+	/* Calculate period in seconds to call the overflow watchdog - to make
+	 * sure counter is checked at least twice every wrap around.
+	 * The period is calculated as the minimum between max HW cycles count
+	 * (The clock source mask) and max amount of cycles that can be
+	 * multiplied by clock multiplier where the result doesn't exceed
+	 * 64bits.
+	 */
+	overflow_cycles = div64_u64(~0ULL >> 1, clock->cycles.mult);
+	overflow_cycles = min(overflow_cycles, div_u64(clock->cycles.mask, 3));
+
+	ns = cyclecounter_cyc2ns(&clock->cycles, overflow_cycles,
+				 frac, &frac);
+	do_div(ns, NSEC_PER_SEC / HZ);
+	clock->overflow_period = ns;
+
+	mdev->clock_info_page = alloc_page(GFP_KERNEL);
+	if (mdev->clock_info_page) {
+		mdev->clock_info = kmap(mdev->clock_info_page);
+		if (!mdev->clock_info) {
+			__free_page(mdev->clock_info_page);
+			mlx5_core_warn(mdev, "failed to map clock page\n");
+		} else {
+			mdev->clock_info->sign   = 0;
+			mdev->clock_info->nsec   = clock->tc.nsec;
+			mdev->clock_info->cycles = clock->tc.cycle_last;
+			mdev->clock_info->mask   = clock->cycles.mask;
+			mdev->clock_info->mult   = clock->nominal_c_mult;
+			mdev->clock_info->shift  = clock->cycles.shift;
+			mdev->clock_info->frac   = clock->tc.frac;
+			mdev->clock_info->overflow_period =
+						clock->overflow_period;
+		}
+	}
+
+	INIT_WORK(&clock->pps_info.out_work, mlx5_pps_out);
+	INIT_DELAYED_WORK(&clock->overflow_work, mlx5_timestamp_overflow);
+	if (clock->overflow_period)
+		schedule_delayed_work(&clock->overflow_work, 0);
+	else
+		mlx5_core_warn(mdev, "invalid overflow period, overflow_work is not scheduled\n");
+
+	/* Configure the PHC */
+	clock->ptp_info = mlx5_ptp_clock_info;
+
+	/* Initialize 1PPS data structures */
+	if (MLX5_PPS_CAP(mdev))
+		mlx5_get_pps_caps(mdev);
+	if (clock->ptp_info.n_pins)
+		mlx5_init_pin_config(clock);
+
+	clock->ptp = ptp_clock_register(&clock->ptp_info,
+					&mdev->pdev->dev);
+	if (IS_ERR(clock->ptp)) {
+		mlx5_core_warn(mdev, "ptp_clock_register failed %ld\n",
+			       PTR_ERR(clock->ptp));
+		clock->ptp = NULL;
+	}
+}
+
+void mlx5_cleanup_clock(struct mlx5_core_dev *mdev)
+{
+	struct mlx5_clock *clock = &mdev->clock;
+
+	if (!MLX5_CAP_GEN(mdev, device_frequency_khz))
+		return;
+
+	if (clock->ptp) {
+		ptp_clock_unregister(clock->ptp);
+		clock->ptp = NULL;
+	}
+
+	cancel_work_sync(&clock->pps_info.out_work);
+	cancel_delayed_work_sync(&clock->overflow_work);
+
+	if (mdev->clock_info) {
+		kunmap(mdev->clock_info_page);
+		__free_page(mdev->clock_info_page);
+		mdev->clock_info = NULL;
+	}
+
+	kfree(clock->ptp_info.pin_config);
+}
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/lib/clock.h b/drivers/net/ethernet/mellanox/mlx5/core/lib/clock.h
new file mode 100644
index 000000000..02e2e4575
--- /dev/null
+++ b/drivers/net/ethernet/mellanox/mlx5/core/lib/clock.h
@@ -0,0 +1,75 @@
+/*
+ * Copyright (c) 2017, Mellanox Technologies, Ltd.  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 __LIB_CLOCK_H__
+#define __LIB_CLOCK_H__
+
+#if IS_ENABLED(CONFIG_PTP_1588_CLOCK)
+void mlx5_init_clock(struct mlx5_core_dev *mdev);
+void mlx5_cleanup_clock(struct mlx5_core_dev *mdev);
+void mlx5_pps_event(struct mlx5_core_dev *dev, struct mlx5_eqe *eqe);
+
+static inline int mlx5_clock_get_ptp_index(struct mlx5_core_dev *mdev)
+{
+	return mdev->clock.ptp ? ptp_clock_index(mdev->clock.ptp) : -1;
+}
+
+static inline ktime_t mlx5_timecounter_cyc2time(struct mlx5_clock *clock,
+						u64 timestamp)
+{
+	u64 nsec;
+
+	read_lock(&clock->lock);
+	nsec = timecounter_cyc2time(&clock->tc, timestamp);
+	read_unlock(&clock->lock);
+
+	return ns_to_ktime(nsec);
+}
+
+#else
+static inline void mlx5_init_clock(struct mlx5_core_dev *mdev) {}
+static inline void mlx5_cleanup_clock(struct mlx5_core_dev *mdev) {}
+static inline void mlx5_pps_event(struct mlx5_core_dev *dev, struct mlx5_eqe *eqe) {}
+
+static inline int mlx5_clock_get_ptp_index(struct mlx5_core_dev *mdev)
+{
+	return -1;
+}
+
+static inline ktime_t mlx5_timecounter_cyc2time(struct mlx5_clock *clock,
+						u64 timestamp)
+{
+	return 0;
+}
+#endif
+
+#endif
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/lib/gid.c b/drivers/net/ethernet/mellanox/mlx5/core/lib/gid.c
new file mode 100644
index 000000000..7722a3f9b
--- /dev/null
+++ b/drivers/net/ethernet/mellanox/mlx5/core/lib/gid.c
@@ -0,0 +1,158 @@
+/*
+ * Copyright (c) 2017, 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/mlx5/driver.h>
+#include <linux/etherdevice.h>
+#include <linux/idr.h>
+#include "mlx5_core.h"
+#include "lib/mlx5.h"
+
+void mlx5_init_reserved_gids(struct mlx5_core_dev *dev)
+{
+	unsigned int tblsz = MLX5_CAP_ROCE(dev, roce_address_table_size);
+
+	ida_init(&dev->roce.reserved_gids.ida);
+	dev->roce.reserved_gids.start = tblsz;
+	dev->roce.reserved_gids.count = 0;
+}
+
+void mlx5_cleanup_reserved_gids(struct mlx5_core_dev *dev)
+{
+	WARN_ON(!ida_is_empty(&dev->roce.reserved_gids.ida));
+	dev->roce.reserved_gids.start = 0;
+	dev->roce.reserved_gids.count = 0;
+	ida_destroy(&dev->roce.reserved_gids.ida);
+}
+
+int mlx5_core_reserve_gids(struct mlx5_core_dev *dev, unsigned int count)
+{
+	if (test_bit(MLX5_INTERFACE_STATE_UP, &dev->intf_state)) {
+		mlx5_core_err(dev, "Cannot reserve GIDs when interfaces are up\n");
+		return -EPERM;
+	}
+	if (dev->roce.reserved_gids.start < count) {
+		mlx5_core_warn(dev, "GID table exhausted attempting to reserve %d more GIDs\n",
+			       count);
+		return -ENOMEM;
+	}
+	if (dev->roce.reserved_gids.count + count > MLX5_MAX_RESERVED_GIDS) {
+		mlx5_core_warn(dev, "Unable to reserve %d more GIDs\n", count);
+		return -ENOMEM;
+	}
+
+	dev->roce.reserved_gids.start -= count;
+	dev->roce.reserved_gids.count += count;
+	mlx5_core_dbg(dev, "Reserved %u GIDs starting at %u\n",
+		      dev->roce.reserved_gids.count,
+		      dev->roce.reserved_gids.start);
+	return 0;
+}
+
+void mlx5_core_unreserve_gids(struct mlx5_core_dev *dev, unsigned int count)
+{
+	WARN(test_bit(MLX5_INTERFACE_STATE_UP, &dev->intf_state), "Unreserving GIDs when interfaces are up");
+	WARN(count > dev->roce.reserved_gids.count, "Unreserving %u GIDs when only %u reserved",
+	     count, dev->roce.reserved_gids.count);
+
+	dev->roce.reserved_gids.start += count;
+	dev->roce.reserved_gids.count -= count;
+	mlx5_core_dbg(dev, "%u GIDs starting at %u left reserved\n",
+		      dev->roce.reserved_gids.count,
+		      dev->roce.reserved_gids.start);
+}
+
+int mlx5_core_reserved_gid_alloc(struct mlx5_core_dev *dev, int *gid_index)
+{
+	int end = dev->roce.reserved_gids.start +
+		  dev->roce.reserved_gids.count;
+	int index = 0;
+
+	index = ida_simple_get(&dev->roce.reserved_gids.ida,
+			       dev->roce.reserved_gids.start, end,
+			       GFP_KERNEL);
+	if (index < 0)
+		return index;
+
+	mlx5_core_dbg(dev, "Allocating reserved GID %u\n", index);
+	*gid_index = index;
+	return 0;
+}
+
+void mlx5_core_reserved_gid_free(struct mlx5_core_dev *dev, int gid_index)
+{
+	mlx5_core_dbg(dev, "Freeing reserved GID %u\n", gid_index);
+	ida_simple_remove(&dev->roce.reserved_gids.ida, gid_index);
+}
+
+unsigned int mlx5_core_reserved_gids_count(struct mlx5_core_dev *dev)
+{
+	return dev->roce.reserved_gids.count;
+}
+EXPORT_SYMBOL_GPL(mlx5_core_reserved_gids_count);
+
+int mlx5_core_roce_gid_set(struct mlx5_core_dev *dev, unsigned int index,
+			   u8 roce_version, u8 roce_l3_type, const u8 *gid,
+			   const u8 *mac, bool vlan, u16 vlan_id, u8 port_num)
+{
+#define MLX5_SET_RA(p, f, v) MLX5_SET(roce_addr_layout, p, f, v)
+	u32  in[MLX5_ST_SZ_DW(set_roce_address_in)] = {0};
+	u32 out[MLX5_ST_SZ_DW(set_roce_address_out)] = {0};
+	void *in_addr = MLX5_ADDR_OF(set_roce_address_in, in, roce_address);
+	char *addr_l3_addr = MLX5_ADDR_OF(roce_addr_layout, in_addr,
+					  source_l3_address);
+	void *addr_mac = MLX5_ADDR_OF(roce_addr_layout, in_addr,
+				      source_mac_47_32);
+	int gidsz = MLX5_FLD_SZ_BYTES(roce_addr_layout, source_l3_address);
+
+	if (MLX5_CAP_GEN(dev, port_type) != MLX5_CAP_PORT_TYPE_ETH)
+		return -EINVAL;
+
+	if (gid) {
+		if (vlan) {
+			MLX5_SET_RA(in_addr, vlan_valid, 1);
+			MLX5_SET_RA(in_addr, vlan_id, vlan_id);
+		}
+
+		ether_addr_copy(addr_mac, mac);
+		MLX5_SET_RA(in_addr, roce_version, roce_version);
+		MLX5_SET_RA(in_addr, roce_l3_type, roce_l3_type);
+		memcpy(addr_l3_addr, gid, gidsz);
+	}
+
+	if (MLX5_CAP_GEN(dev, num_vhca_ports) > 0)
+		MLX5_SET(set_roce_address_in, in, vhca_port_num, port_num);
+
+	MLX5_SET(set_roce_address_in, in, roce_address_index, index);
+	MLX5_SET(set_roce_address_in, in, opcode, MLX5_CMD_OP_SET_ROCE_ADDRESS);
+	return mlx5_cmd_exec(dev, in, sizeof(in), out, sizeof(out));
+}
+EXPORT_SYMBOL(mlx5_core_roce_gid_set);
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/lib/mlx5.h b/drivers/net/ethernet/mellanox/mlx5/core/lib/mlx5.h
new file mode 100644
index 000000000..7550b1cc8
--- /dev/null
+++ b/drivers/net/ethernet/mellanox/mlx5/core/lib/mlx5.h
@@ -0,0 +1,43 @@
+/*
+ * Copyright (c) 2017, Mellanox Technologies, Ltd.  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 __LIB_MLX5_H__
+#define __LIB_MLX5_H__
+
+void mlx5_init_reserved_gids(struct mlx5_core_dev *dev);
+void mlx5_cleanup_reserved_gids(struct mlx5_core_dev *dev);
+int  mlx5_core_reserve_gids(struct mlx5_core_dev *dev, unsigned int count);
+void mlx5_core_unreserve_gids(struct mlx5_core_dev *dev, unsigned int count);
+int  mlx5_core_reserved_gid_alloc(struct mlx5_core_dev *dev, int *gid_index);
+void mlx5_core_reserved_gid_free(struct mlx5_core_dev *dev, int gid_index);
+
+#endif
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/lib/mpfs.c b/drivers/net/ethernet/mellanox/mlx5/core/lib/mpfs.c
new file mode 100644
index 000000000..98359559c
--- /dev/null
+++ b/drivers/net/ethernet/mellanox/mlx5/core/lib/mpfs.c
@@ -0,0 +1,202 @@
+/*
+ * Copyright (c) 2017, 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/etherdevice.h>
+#include <linux/mlx5/driver.h>
+#include <linux/mlx5/mlx5_ifc.h>
+#include <linux/mlx5/eswitch.h>
+#include "mlx5_core.h"
+#include "lib/mpfs.h"
+
+/* HW L2 Table (MPFS) management */
+static int set_l2table_entry_cmd(struct mlx5_core_dev *dev, u32 index, u8 *mac)
+{
+	u32 in[MLX5_ST_SZ_DW(set_l2_table_entry_in)]   = {0};
+	u32 out[MLX5_ST_SZ_DW(set_l2_table_entry_out)] = {0};
+	u8 *in_mac_addr;
+
+	MLX5_SET(set_l2_table_entry_in, in, opcode, MLX5_CMD_OP_SET_L2_TABLE_ENTRY);
+	MLX5_SET(set_l2_table_entry_in, in, table_index, index);
+
+	in_mac_addr = MLX5_ADDR_OF(set_l2_table_entry_in, in, mac_address);
+	ether_addr_copy(&in_mac_addr[2], mac);
+
+	return mlx5_cmd_exec(dev, in, sizeof(in), out, sizeof(out));
+}
+
+static int del_l2table_entry_cmd(struct mlx5_core_dev *dev, u32 index)
+{
+	u32 in[MLX5_ST_SZ_DW(delete_l2_table_entry_in)]   = {0};
+	u32 out[MLX5_ST_SZ_DW(delete_l2_table_entry_out)] = {0};
+
+	MLX5_SET(delete_l2_table_entry_in, in, opcode, MLX5_CMD_OP_DELETE_L2_TABLE_ENTRY);
+	MLX5_SET(delete_l2_table_entry_in, in, table_index, index);
+	return mlx5_cmd_exec(dev, in, sizeof(in), out, sizeof(out));
+}
+
+/* UC L2 table hash node */
+struct l2table_node {
+	struct l2addr_node node;
+	u32                index; /* index in HW l2 table */
+};
+
+struct mlx5_mpfs {
+	struct hlist_head    hash[MLX5_L2_ADDR_HASH_SIZE];
+	struct mutex         lock; /* Synchronize l2 table access */
+	u32                  size;
+	unsigned long        *bitmap;
+};
+
+static int alloc_l2table_index(struct mlx5_mpfs *l2table, u32 *ix)
+{
+	int err = 0;
+
+	*ix = find_first_zero_bit(l2table->bitmap, l2table->size);
+	if (*ix >= l2table->size)
+		err = -ENOSPC;
+	else
+		__set_bit(*ix, l2table->bitmap);
+
+	return err;
+}
+
+static void free_l2table_index(struct mlx5_mpfs *l2table, u32 ix)
+{
+	__clear_bit(ix, l2table->bitmap);
+}
+
+int mlx5_mpfs_init(struct mlx5_core_dev *dev)
+{
+	int l2table_size = 1 << MLX5_CAP_GEN(dev, log_max_l2_table);
+	struct mlx5_mpfs *mpfs;
+
+	if (!MLX5_ESWITCH_MANAGER(dev))
+		return 0;
+
+	mpfs = kzalloc(sizeof(*mpfs), GFP_KERNEL);
+	if (!mpfs)
+		return -ENOMEM;
+
+	mutex_init(&mpfs->lock);
+	mpfs->size   = l2table_size;
+	mpfs->bitmap = kcalloc(BITS_TO_LONGS(l2table_size),
+			       sizeof(uintptr_t), GFP_KERNEL);
+	if (!mpfs->bitmap) {
+		kfree(mpfs);
+		return -ENOMEM;
+	}
+
+	dev->priv.mpfs = mpfs;
+	return 0;
+}
+
+void mlx5_mpfs_cleanup(struct mlx5_core_dev *dev)
+{
+	struct mlx5_mpfs *mpfs = dev->priv.mpfs;
+
+	if (!MLX5_ESWITCH_MANAGER(dev))
+		return;
+
+	WARN_ON(!hlist_empty(mpfs->hash));
+	kfree(mpfs->bitmap);
+	kfree(mpfs);
+}
+
+int mlx5_mpfs_add_mac(struct mlx5_core_dev *dev, u8 *mac)
+{
+	struct mlx5_mpfs *mpfs = dev->priv.mpfs;
+	struct l2table_node *l2addr;
+	u32 index;
+	int err;
+
+	if (!MLX5_ESWITCH_MANAGER(dev))
+		return 0;
+
+	mutex_lock(&mpfs->lock);
+
+	l2addr = l2addr_hash_find(mpfs->hash, mac, struct l2table_node);
+	if (l2addr) {
+		err = -EEXIST;
+		goto abort;
+	}
+
+	err = alloc_l2table_index(mpfs, &index);
+	if (err)
+		goto abort;
+
+	l2addr = l2addr_hash_add(mpfs->hash, mac, struct l2table_node, GFP_KERNEL);
+	if (!l2addr) {
+		free_l2table_index(mpfs, index);
+		err = -ENOMEM;
+		goto abort;
+	}
+
+	l2addr->index = index;
+	err = set_l2table_entry_cmd(dev, index, mac);
+	if (err) {
+		l2addr_hash_del(l2addr);
+		free_l2table_index(mpfs, index);
+	}
+
+	mlx5_core_dbg(dev, "MPFS mac added %pM, index (%d)\n", mac, index);
+abort:
+	mutex_unlock(&mpfs->lock);
+	return err;
+}
+
+int mlx5_mpfs_del_mac(struct mlx5_core_dev *dev, u8 *mac)
+{
+	struct mlx5_mpfs *mpfs = dev->priv.mpfs;
+	struct l2table_node *l2addr;
+	int err = 0;
+	u32 index;
+
+	if (!MLX5_ESWITCH_MANAGER(dev))
+		return 0;
+
+	mutex_lock(&mpfs->lock);
+
+	l2addr = l2addr_hash_find(mpfs->hash, mac, struct l2table_node);
+	if (!l2addr) {
+		err = -ENOENT;
+		goto unlock;
+	}
+
+	index = l2addr->index;
+	del_l2table_entry_cmd(dev, index);
+	l2addr_hash_del(l2addr);
+	free_l2table_index(mpfs, index);
+	mlx5_core_dbg(dev, "MPFS mac deleted %pM, index (%d)\n", mac, index);
+unlock:
+	mutex_unlock(&mpfs->lock);
+	return err;
+}
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/lib/mpfs.h b/drivers/net/ethernet/mellanox/mlx5/core/lib/mpfs.h
new file mode 100644
index 000000000..4a7b2c320
--- /dev/null
+++ b/drivers/net/ethernet/mellanox/mlx5/core/lib/mpfs.h
@@ -0,0 +1,95 @@
+/*
+ * Copyright (c) 2017, Mellanox Technologies, Ltd.  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 __MLX5_MPFS_H__
+#define __MLX5_MPFS_H__
+
+#include <linux/if_ether.h>
+#include <linux/mlx5/device.h>
+
+/* L2 -mac address based- hash helpers */
+#define MLX5_L2_ADDR_HASH_SIZE (BIT(BITS_PER_BYTE))
+#define MLX5_L2_ADDR_HASH(addr) (addr[5])
+
+struct l2addr_node {
+	struct hlist_node hlist;
+	u8                addr[ETH_ALEN];
+};
+
+#define for_each_l2hash_node(hn, tmp, hash, i) \
+	for (i = 0; i < MLX5_L2_ADDR_HASH_SIZE; i++) \
+		hlist_for_each_entry_safe(hn, tmp, &(hash)[i], hlist)
+
+#define l2addr_hash_find(hash, mac, type) ({                \
+	int ix = MLX5_L2_ADDR_HASH(mac);                    \
+	bool found = false;                                 \
+	type *ptr = NULL;                                   \
+							    \
+	hlist_for_each_entry(ptr, &(hash)[ix], node.hlist)  \
+		if (ether_addr_equal(ptr->node.addr, mac)) {\
+			found = true;                       \
+			break;                              \
+		}                                           \
+	if (!found)                                         \
+		ptr = NULL;                                 \
+	ptr;                                                \
+})
+
+#define l2addr_hash_add(hash, mac, type, gfp) ({            \
+	int ix = MLX5_L2_ADDR_HASH(mac);                    \
+	type *ptr = NULL;                                   \
+							    \
+	ptr = kzalloc(sizeof(type), gfp);                   \
+	if (ptr) {                                          \
+		ether_addr_copy(ptr->node.addr, mac);       \
+		hlist_add_head(&ptr->node.hlist, &(hash)[ix]);\
+	}                                                   \
+	ptr;                                                \
+})
+
+#define l2addr_hash_del(ptr) ({                             \
+	hlist_del(&(ptr)->node.hlist);                      \
+	kfree(ptr);                                         \
+})
+
+#ifdef CONFIG_MLX5_MPFS
+int  mlx5_mpfs_init(struct mlx5_core_dev *dev);
+void mlx5_mpfs_cleanup(struct mlx5_core_dev *dev);
+int  mlx5_mpfs_add_mac(struct mlx5_core_dev *dev, u8 *mac);
+int  mlx5_mpfs_del_mac(struct mlx5_core_dev *dev, u8 *mac);
+#else /* #ifndef CONFIG_MLX5_MPFS */
+static inline int  mlx5_mpfs_init(struct mlx5_core_dev *dev) { return 0; }
+static inline void mlx5_mpfs_cleanup(struct mlx5_core_dev *dev) {}
+static inline int  mlx5_mpfs_add_mac(struct mlx5_core_dev *dev, u8 *mac) { return 0; }
+static inline int  mlx5_mpfs_del_mac(struct mlx5_core_dev *dev, u8 *mac) { return 0; }
+#endif
+#endif
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/lib/vxlan.c b/drivers/net/ethernet/mellanox/mlx5/core/lib/vxlan.c
new file mode 100644
index 000000000..9a8fd7621
--- /dev/null
+++ b/drivers/net/ethernet/mellanox/mlx5/core/lib/vxlan.c
@@ -0,0 +1,230 @@
+/*
+ * Copyright (c) 2016, Mellanox Technologies, Ltd.  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/module.h>
+#include <linux/mlx5/driver.h>
+#include "mlx5_core.h"
+#include "vxlan.h"
+
+struct mlx5_vxlan {
+	struct mlx5_core_dev		*mdev;
+	spinlock_t			lock; /* protect vxlan table */
+	/* max_num_ports is usuallly 4, 16 buckets is more than enough */
+	DECLARE_HASHTABLE(htable, 4);
+	int				num_ports;
+	struct mutex                    sync_lock; /* sync add/del port HW operations */
+};
+
+struct mlx5_vxlan_port {
+	struct hlist_node hlist;
+	atomic_t refcount;
+	u16 udp_port;
+};
+
+static inline u8 mlx5_vxlan_max_udp_ports(struct mlx5_core_dev *mdev)
+{
+	return MLX5_CAP_ETH(mdev, max_vxlan_udp_ports) ?: 4;
+}
+
+static int mlx5_vxlan_core_add_port_cmd(struct mlx5_core_dev *mdev, u16 port)
+{
+	u32 in[MLX5_ST_SZ_DW(add_vxlan_udp_dport_in)]   = {0};
+	u32 out[MLX5_ST_SZ_DW(add_vxlan_udp_dport_out)] = {0};
+
+	MLX5_SET(add_vxlan_udp_dport_in, in, opcode,
+		 MLX5_CMD_OP_ADD_VXLAN_UDP_DPORT);
+	MLX5_SET(add_vxlan_udp_dport_in, in, vxlan_udp_port, port);
+	return mlx5_cmd_exec(mdev, in, sizeof(in), out, sizeof(out));
+}
+
+static int mlx5_vxlan_core_del_port_cmd(struct mlx5_core_dev *mdev, u16 port)
+{
+	u32 in[MLX5_ST_SZ_DW(delete_vxlan_udp_dport_in)]   = {0};
+	u32 out[MLX5_ST_SZ_DW(delete_vxlan_udp_dport_out)] = {0};
+
+	MLX5_SET(delete_vxlan_udp_dport_in, in, opcode,
+		 MLX5_CMD_OP_DELETE_VXLAN_UDP_DPORT);
+	MLX5_SET(delete_vxlan_udp_dport_in, in, vxlan_udp_port, port);
+	return mlx5_cmd_exec(mdev, in, sizeof(in), out, sizeof(out));
+}
+
+static struct mlx5_vxlan_port*
+mlx5_vxlan_lookup_port_locked(struct mlx5_vxlan *vxlan, u16 port)
+{
+	struct mlx5_vxlan_port *vxlanp;
+
+	hash_for_each_possible(vxlan->htable, vxlanp, hlist, port) {
+		if (vxlanp->udp_port == port)
+			return vxlanp;
+	}
+
+	return NULL;
+}
+
+struct mlx5_vxlan_port *mlx5_vxlan_lookup_port(struct mlx5_vxlan *vxlan, u16 port)
+{
+	struct mlx5_vxlan_port *vxlanp;
+
+	if (!mlx5_vxlan_allowed(vxlan))
+		return NULL;
+
+	spin_lock_bh(&vxlan->lock);
+	vxlanp = mlx5_vxlan_lookup_port_locked(vxlan, port);
+	spin_unlock_bh(&vxlan->lock);
+
+	return vxlanp;
+}
+
+int mlx5_vxlan_add_port(struct mlx5_vxlan *vxlan, u16 port)
+{
+	struct mlx5_vxlan_port *vxlanp;
+	int ret = -ENOSPC;
+
+	vxlanp = mlx5_vxlan_lookup_port(vxlan, port);
+	if (vxlanp) {
+		atomic_inc(&vxlanp->refcount);
+		return 0;
+	}
+
+	mutex_lock(&vxlan->sync_lock);
+	if (vxlan->num_ports >= mlx5_vxlan_max_udp_ports(vxlan->mdev)) {
+		mlx5_core_info(vxlan->mdev,
+			       "UDP port (%d) not offloaded, max number of UDP ports (%d) are already offloaded\n",
+			       port, mlx5_vxlan_max_udp_ports(vxlan->mdev));
+		ret = -ENOSPC;
+		goto unlock;
+	}
+
+	ret = mlx5_vxlan_core_add_port_cmd(vxlan->mdev, port);
+	if (ret)
+		goto unlock;
+
+	vxlanp = kzalloc(sizeof(*vxlanp), GFP_KERNEL);
+	if (!vxlanp) {
+		ret = -ENOMEM;
+		goto err_delete_port;
+	}
+
+	vxlanp->udp_port = port;
+	atomic_set(&vxlanp->refcount, 1);
+
+	spin_lock_bh(&vxlan->lock);
+	hash_add(vxlan->htable, &vxlanp->hlist, port);
+	spin_unlock_bh(&vxlan->lock);
+
+	vxlan->num_ports++;
+	mutex_unlock(&vxlan->sync_lock);
+	return 0;
+
+err_delete_port:
+	mlx5_vxlan_core_del_port_cmd(vxlan->mdev, port);
+
+unlock:
+	mutex_unlock(&vxlan->sync_lock);
+	return ret;
+}
+
+int mlx5_vxlan_del_port(struct mlx5_vxlan *vxlan, u16 port)
+{
+	struct mlx5_vxlan_port *vxlanp;
+	bool remove = false;
+	int ret = 0;
+
+	mutex_lock(&vxlan->sync_lock);
+
+	spin_lock_bh(&vxlan->lock);
+	vxlanp = mlx5_vxlan_lookup_port_locked(vxlan, port);
+	if (!vxlanp) {
+		ret = -ENOENT;
+		goto out_unlock;
+	}
+
+	if (atomic_dec_and_test(&vxlanp->refcount)) {
+		hash_del(&vxlanp->hlist);
+		remove = true;
+	}
+
+out_unlock:
+	spin_unlock_bh(&vxlan->lock);
+
+	if (remove) {
+		mlx5_vxlan_core_del_port_cmd(vxlan->mdev, port);
+		kfree(vxlanp);
+		vxlan->num_ports--;
+	}
+
+	mutex_unlock(&vxlan->sync_lock);
+
+	return ret;
+}
+
+struct mlx5_vxlan *mlx5_vxlan_create(struct mlx5_core_dev *mdev)
+{
+	struct mlx5_vxlan *vxlan;
+
+	if (!MLX5_CAP_ETH(mdev, tunnel_stateless_vxlan) || !mlx5_core_is_pf(mdev))
+		return ERR_PTR(-ENOTSUPP);
+
+	vxlan = kzalloc(sizeof(*vxlan), GFP_KERNEL);
+	if (!vxlan)
+		return ERR_PTR(-ENOMEM);
+
+	vxlan->mdev = mdev;
+	mutex_init(&vxlan->sync_lock);
+	spin_lock_init(&vxlan->lock);
+	hash_init(vxlan->htable);
+
+	/* Hardware adds 4789 by default */
+	mlx5_vxlan_add_port(vxlan, 4789);
+
+	return vxlan;
+}
+
+void mlx5_vxlan_destroy(struct mlx5_vxlan *vxlan)
+{
+	struct mlx5_vxlan_port *vxlanp;
+	struct hlist_node *tmp;
+	int bkt;
+
+	if (!mlx5_vxlan_allowed(vxlan))
+		return;
+
+	/* Lockless since we are the only hash table consumers*/
+	hash_for_each_safe(vxlan->htable, bkt, tmp, vxlanp, hlist) {
+		hash_del(&vxlanp->hlist);
+		mlx5_vxlan_core_del_port_cmd(vxlan->mdev, vxlanp->udp_port);
+		kfree(vxlanp);
+	}
+
+	kfree(vxlan);
+}
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/lib/vxlan.h b/drivers/net/ethernet/mellanox/mlx5/core/lib/vxlan.h
new file mode 100644
index 000000000..8fb0eb08f
--- /dev/null
+++ b/drivers/net/ethernet/mellanox/mlx5/core/lib/vxlan.h
@@ -0,0 +1,64 @@
+/*
+ * Copyright (c) 2016, Mellanox Technologies, Ltd.  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 __MLX5_VXLAN_H__
+#define __MLX5_VXLAN_H__
+
+#include <linux/mlx5/driver.h>
+
+struct mlx5_vxlan;
+struct mlx5_vxlan_port;
+
+static inline bool mlx5_vxlan_allowed(struct mlx5_vxlan *vxlan)
+{
+	/* not allowed reason is encoded in vxlan pointer as error,
+	 * on mlx5_vxlan_create
+	 */
+	return !IS_ERR_OR_NULL(vxlan);
+}
+
+#if IS_ENABLED(CONFIG_VXLAN)
+struct mlx5_vxlan *mlx5_vxlan_create(struct mlx5_core_dev *mdev);
+void mlx5_vxlan_destroy(struct mlx5_vxlan *vxlan);
+int mlx5_vxlan_add_port(struct mlx5_vxlan *vxlan, u16 port);
+int mlx5_vxlan_del_port(struct mlx5_vxlan *vxlan, u16 port);
+struct mlx5_vxlan_port *mlx5_vxlan_lookup_port(struct mlx5_vxlan *vxlan, u16 port);
+#else
+static inline struct mlx5_vxlan*
+mlx5_vxlan_create(struct mlx5_core_dev *mdev) { return ERR_PTR(-EOPNOTSUPP); }
+static inline void mlx5_vxlan_destroy(struct mlx5_vxlan *vxlan) { return; }
+static inline int mlx5_vxlan_add_port(struct mlx5_vxlan *vxlan, u16 port) { return -EOPNOTSUPP; }
+static inline int mlx5_vxlan_del_port(struct mlx5_vxlan *vxlan, u16 port) { return -EOPNOTSUPP; }
+static inline struct mx5_vxlan_port*
+mlx5_vxlan_lookup_port(struct mlx5_vxlan *vxlan, u16 port) { return NULL; }
+#endif
+
+#endif /* __MLX5_VXLAN_H__ */
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/mad.c b/drivers/net/ethernet/mellanox/mlx5/core/mad.c
new file mode 100644
index 000000000..3a3b0005f
--- /dev/null
+++ b/drivers/net/ethernet/mellanox/mlx5/core/mad.c
@@ -0,0 +1,75 @@
+/*
+ * Copyright (c) 2013-2015, 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/module.h>
+#include <linux/mlx5/driver.h>
+#include <linux/mlx5/cmd.h>
+#include "mlx5_core.h"
+
+int mlx5_core_mad_ifc(struct mlx5_core_dev *dev, const void *inb, void *outb,
+		      u16 opmod, u8 port)
+{
+	int outlen = MLX5_ST_SZ_BYTES(mad_ifc_out);
+	int inlen = MLX5_ST_SZ_BYTES(mad_ifc_in);
+	int err = -ENOMEM;
+	void *data;
+	void *resp;
+	u32 *out;
+	u32 *in;
+
+	in = kzalloc(inlen, GFP_KERNEL);
+	out = kzalloc(outlen, GFP_KERNEL);
+	if (!in || !out)
+		goto out;
+
+	MLX5_SET(mad_ifc_in, in, opcode, MLX5_CMD_OP_MAD_IFC);
+	MLX5_SET(mad_ifc_in, in, op_mod, opmod);
+	MLX5_SET(mad_ifc_in, in, port, port);
+
+	data = MLX5_ADDR_OF(mad_ifc_in, in, mad);
+	memcpy(data, inb, MLX5_FLD_SZ_BYTES(mad_ifc_in, mad));
+
+	err = mlx5_cmd_exec(dev, in, inlen, out, outlen);
+	if (err)
+		goto out;
+
+	resp = MLX5_ADDR_OF(mad_ifc_out, out, response_mad_packet);
+	memcpy(outb, resp,
+	       MLX5_FLD_SZ_BYTES(mad_ifc_out, response_mad_packet));
+
+out:
+	kfree(out);
+	kfree(in);
+	return err;
+}
+EXPORT_SYMBOL_GPL(mlx5_core_mad_ifc);
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/main.c b/drivers/net/ethernet/mellanox/mlx5/core/main.c
new file mode 100644
index 000000000..a2b25afa2
--- /dev/null
+++ b/drivers/net/ethernet/mellanox/mlx5/core/main.c
@@ -0,0 +1,1721 @@
+/*
+ * Copyright (c) 2013-2015, 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/highmem.h>
+#include <linux/module.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/interrupt.h>
+#include <linux/delay.h>
+#include <linux/mlx5/driver.h>
+#include <linux/mlx5/cq.h>
+#include <linux/mlx5/qp.h>
+#include <linux/mlx5/srq.h>
+#include <linux/debugfs.h>
+#include <linux/kmod.h>
+#include <linux/mlx5/mlx5_ifc.h>
+#include <linux/mlx5/vport.h>
+#ifdef CONFIG_RFS_ACCEL
+#include <linux/cpu_rmap.h>
+#endif
+#include <linux/version.h>
+#include <net/devlink.h>
+#include "mlx5_core.h"
+#include "fs_core.h"
+#include "lib/mpfs.h"
+#include "eswitch.h"
+#include "lib/mlx5.h"
+#include "fpga/core.h"
+#include "fpga/ipsec.h"
+#include "accel/ipsec.h"
+#include "accel/tls.h"
+#include "lib/clock.h"
+#include "lib/vxlan.h"
+#include "diag/fw_tracer.h"
+
+MODULE_AUTHOR("Eli Cohen <eli@mellanox.com>");
+MODULE_DESCRIPTION("Mellanox 5th generation network adapters (ConnectX series) core driver");
+MODULE_LICENSE("Dual BSD/GPL");
+MODULE_VERSION(DRIVER_VERSION);
+
+unsigned int mlx5_core_debug_mask;
+module_param_named(debug_mask, mlx5_core_debug_mask, uint, 0644);
+MODULE_PARM_DESC(debug_mask, "debug mask: 1 = dump cmd data, 2 = dump cmd exec time, 3 = both. Default=0");
+
+#define MLX5_DEFAULT_PROF	2
+static unsigned int prof_sel = MLX5_DEFAULT_PROF;
+module_param_named(prof_sel, prof_sel, uint, 0444);
+MODULE_PARM_DESC(prof_sel, "profile selector. Valid range 0 - 2");
+
+static u32 sw_owner_id[4];
+
+enum {
+	MLX5_ATOMIC_REQ_MODE_BE = 0x0,
+	MLX5_ATOMIC_REQ_MODE_HOST_ENDIANNESS = 0x1,
+};
+
+static struct mlx5_profile profile[] = {
+	[0] = {
+		.mask           = 0,
+	},
+	[1] = {
+		.mask		= MLX5_PROF_MASK_QP_SIZE,
+		.log_max_qp	= 12,
+	},
+	[2] = {
+		.mask		= MLX5_PROF_MASK_QP_SIZE |
+				  MLX5_PROF_MASK_MR_CACHE,
+		.log_max_qp	= 18,
+		.mr_cache[0]	= {
+			.size	= 500,
+			.limit	= 250
+		},
+		.mr_cache[1]	= {
+			.size	= 500,
+			.limit	= 250
+		},
+		.mr_cache[2]	= {
+			.size	= 500,
+			.limit	= 250
+		},
+		.mr_cache[3]	= {
+			.size	= 500,
+			.limit	= 250
+		},
+		.mr_cache[4]	= {
+			.size	= 500,
+			.limit	= 250
+		},
+		.mr_cache[5]	= {
+			.size	= 500,
+			.limit	= 250
+		},
+		.mr_cache[6]	= {
+			.size	= 500,
+			.limit	= 250
+		},
+		.mr_cache[7]	= {
+			.size	= 500,
+			.limit	= 250
+		},
+		.mr_cache[8]	= {
+			.size	= 500,
+			.limit	= 250
+		},
+		.mr_cache[9]	= {
+			.size	= 500,
+			.limit	= 250
+		},
+		.mr_cache[10]	= {
+			.size	= 500,
+			.limit	= 250
+		},
+		.mr_cache[11]	= {
+			.size	= 500,
+			.limit	= 250
+		},
+		.mr_cache[12]	= {
+			.size	= 64,
+			.limit	= 32
+		},
+		.mr_cache[13]	= {
+			.size	= 32,
+			.limit	= 16
+		},
+		.mr_cache[14]	= {
+			.size	= 16,
+			.limit	= 8
+		},
+		.mr_cache[15]	= {
+			.size	= 8,
+			.limit	= 4
+		},
+	},
+};
+
+#define FW_INIT_TIMEOUT_MILI		2000
+#define FW_INIT_WAIT_MS			2
+#define FW_PRE_INIT_TIMEOUT_MILI	10000
+
+static int wait_fw_init(struct mlx5_core_dev *dev, u32 max_wait_mili)
+{
+	unsigned long end = jiffies + msecs_to_jiffies(max_wait_mili);
+	int err = 0;
+
+	while (fw_initializing(dev)) {
+		if (time_after(jiffies, end)) {
+			err = -EBUSY;
+			break;
+		}
+		msleep(FW_INIT_WAIT_MS);
+	}
+
+	return err;
+}
+
+static void mlx5_set_driver_version(struct mlx5_core_dev *dev)
+{
+	int driver_ver_sz = MLX5_FLD_SZ_BYTES(set_driver_version_in,
+					      driver_version);
+	u8 in[MLX5_ST_SZ_BYTES(set_driver_version_in)] = {0};
+	u8 out[MLX5_ST_SZ_BYTES(set_driver_version_out)] = {0};
+	int remaining_size = driver_ver_sz;
+	char *string;
+
+	if (!MLX5_CAP_GEN(dev, driver_version))
+		return;
+
+	string = MLX5_ADDR_OF(set_driver_version_in, in, driver_version);
+
+	strncpy(string, "Linux", remaining_size);
+
+	remaining_size = max_t(int, 0, driver_ver_sz - strlen(string));
+	strncat(string, ",", remaining_size);
+
+	remaining_size = max_t(int, 0, driver_ver_sz - strlen(string));
+	strncat(string, DRIVER_NAME, remaining_size);
+
+	remaining_size = max_t(int, 0, driver_ver_sz - strlen(string));
+	strncat(string, ",", remaining_size);
+
+	remaining_size = max_t(int, 0, driver_ver_sz - strlen(string));
+
+	snprintf(string + strlen(string), remaining_size, "%u.%u.%u",
+		 (u8)((LINUX_VERSION_CODE >> 16) & 0xff), (u8)((LINUX_VERSION_CODE >> 8) & 0xff),
+		 (u16)(LINUX_VERSION_CODE & 0xffff));
+
+	/*Send the command*/
+	MLX5_SET(set_driver_version_in, in, opcode,
+		 MLX5_CMD_OP_SET_DRIVER_VERSION);
+
+	mlx5_cmd_exec(dev, in, sizeof(in), out, sizeof(out));
+}
+
+static int set_dma_caps(struct pci_dev *pdev)
+{
+	int err;
+
+	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");
+			return err;
+		}
+	}
+
+	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");
+			return err;
+		}
+	}
+
+	dma_set_max_seg_size(&pdev->dev, 2u * 1024 * 1024 * 1024);
+	return err;
+}
+
+static int mlx5_pci_enable_device(struct mlx5_core_dev *dev)
+{
+	struct pci_dev *pdev = dev->pdev;
+	int err = 0;
+
+	mutex_lock(&dev->pci_status_mutex);
+	if (dev->pci_status == MLX5_PCI_STATUS_DISABLED) {
+		err = pci_enable_device(pdev);
+		if (!err)
+			dev->pci_status = MLX5_PCI_STATUS_ENABLED;
+	}
+	mutex_unlock(&dev->pci_status_mutex);
+
+	return err;
+}
+
+static void mlx5_pci_disable_device(struct mlx5_core_dev *dev)
+{
+	struct pci_dev *pdev = dev->pdev;
+
+	mutex_lock(&dev->pci_status_mutex);
+	if (dev->pci_status == MLX5_PCI_STATUS_ENABLED) {
+		pci_disable_device(pdev);
+		dev->pci_status = MLX5_PCI_STATUS_DISABLED;
+	}
+	mutex_unlock(&dev->pci_status_mutex);
+}
+
+static int request_bar(struct pci_dev *pdev)
+{
+	int err = 0;
+
+	if (!(pci_resource_flags(pdev, 0) & IORESOURCE_MEM)) {
+		dev_err(&pdev->dev, "Missing registers BAR, aborting\n");
+		return -ENODEV;
+	}
+
+	err = pci_request_regions(pdev, DRIVER_NAME);
+	if (err)
+		dev_err(&pdev->dev, "Couldn't get PCI resources, aborting\n");
+
+	return err;
+}
+
+static void release_bar(struct pci_dev *pdev)
+{
+	pci_release_regions(pdev);
+}
+
+static int mlx5_alloc_irq_vectors(struct mlx5_core_dev *dev)
+{
+	struct mlx5_priv *priv = &dev->priv;
+	struct mlx5_eq_table *table = &priv->eq_table;
+	int num_eqs = MLX5_CAP_GEN(dev, max_num_eqs) ?
+		      MLX5_CAP_GEN(dev, max_num_eqs) :
+		      1 << MLX5_CAP_GEN(dev, log_max_eq);
+	int nvec;
+	int err;
+
+	nvec = MLX5_CAP_GEN(dev, num_ports) * num_online_cpus() +
+	       MLX5_EQ_VEC_COMP_BASE;
+	nvec = min_t(int, nvec, num_eqs);
+	if (nvec <= MLX5_EQ_VEC_COMP_BASE)
+		return -ENOMEM;
+
+	priv->irq_info = kcalloc(nvec, sizeof(*priv->irq_info), GFP_KERNEL);
+	if (!priv->irq_info)
+		return -ENOMEM;
+
+	nvec = pci_alloc_irq_vectors(dev->pdev,
+			MLX5_EQ_VEC_COMP_BASE + 1, nvec,
+			PCI_IRQ_MSIX);
+	if (nvec < 0) {
+		err = nvec;
+		goto err_free_irq_info;
+	}
+
+	table->num_comp_vectors = nvec - MLX5_EQ_VEC_COMP_BASE;
+
+	return 0;
+
+err_free_irq_info:
+	kfree(priv->irq_info);
+	return err;
+}
+
+static void mlx5_free_irq_vectors(struct mlx5_core_dev *dev)
+{
+	struct mlx5_priv *priv = &dev->priv;
+
+	pci_free_irq_vectors(dev->pdev);
+	kfree(priv->irq_info);
+}
+
+struct mlx5_reg_host_endianness {
+	u8	he;
+	u8      rsvd[15];
+};
+
+#define CAP_MASK(pos, size) ((u64)((1 << (size)) - 1) << (pos))
+
+enum {
+	MLX5_CAP_BITS_RW_MASK = CAP_MASK(MLX5_CAP_OFF_CMDIF_CSUM, 2) |
+				MLX5_DEV_CAP_FLAG_DCT,
+};
+
+static u16 to_fw_pkey_sz(struct mlx5_core_dev *dev, u32 size)
+{
+	switch (size) {
+	case 128:
+		return 0;
+	case 256:
+		return 1;
+	case 512:
+		return 2;
+	case 1024:
+		return 3;
+	case 2048:
+		return 4;
+	case 4096:
+		return 5;
+	default:
+		mlx5_core_warn(dev, "invalid pkey table size %d\n", size);
+		return 0;
+	}
+}
+
+static int mlx5_core_get_caps_mode(struct mlx5_core_dev *dev,
+				   enum mlx5_cap_type cap_type,
+				   enum mlx5_cap_mode cap_mode)
+{
+	u8 in[MLX5_ST_SZ_BYTES(query_hca_cap_in)];
+	int out_sz = MLX5_ST_SZ_BYTES(query_hca_cap_out);
+	void *out, *hca_caps;
+	u16 opmod = (cap_type << 1) | (cap_mode & 0x01);
+	int err;
+
+	memset(in, 0, sizeof(in));
+	out = kzalloc(out_sz, GFP_KERNEL);
+	if (!out)
+		return -ENOMEM;
+
+	MLX5_SET(query_hca_cap_in, in, opcode, MLX5_CMD_OP_QUERY_HCA_CAP);
+	MLX5_SET(query_hca_cap_in, in, op_mod, opmod);
+	err = mlx5_cmd_exec(dev, in, sizeof(in), out, out_sz);
+	if (err) {
+		mlx5_core_warn(dev,
+			       "QUERY_HCA_CAP : type(%x) opmode(%x) Failed(%d)\n",
+			       cap_type, cap_mode, err);
+		goto query_ex;
+	}
+
+	hca_caps =  MLX5_ADDR_OF(query_hca_cap_out, out, capability);
+
+	switch (cap_mode) {
+	case HCA_CAP_OPMOD_GET_MAX:
+		memcpy(dev->caps.hca_max[cap_type], hca_caps,
+		       MLX5_UN_SZ_BYTES(hca_cap_union));
+		break;
+	case HCA_CAP_OPMOD_GET_CUR:
+		memcpy(dev->caps.hca_cur[cap_type], hca_caps,
+		       MLX5_UN_SZ_BYTES(hca_cap_union));
+		break;
+	default:
+		mlx5_core_warn(dev,
+			       "Tried to query dev cap type(%x) with wrong opmode(%x)\n",
+			       cap_type, cap_mode);
+		err = -EINVAL;
+		break;
+	}
+query_ex:
+	kfree(out);
+	return err;
+}
+
+int mlx5_core_get_caps(struct mlx5_core_dev *dev, enum mlx5_cap_type cap_type)
+{
+	int ret;
+
+	ret = mlx5_core_get_caps_mode(dev, cap_type, HCA_CAP_OPMOD_GET_CUR);
+	if (ret)
+		return ret;
+	return mlx5_core_get_caps_mode(dev, cap_type, HCA_CAP_OPMOD_GET_MAX);
+}
+
+static int set_caps(struct mlx5_core_dev *dev, void *in, int in_sz, int opmod)
+{
+	u32 out[MLX5_ST_SZ_DW(set_hca_cap_out)] = {0};
+
+	MLX5_SET(set_hca_cap_in, in, opcode, MLX5_CMD_OP_SET_HCA_CAP);
+	MLX5_SET(set_hca_cap_in, in, op_mod, opmod << 1);
+	return mlx5_cmd_exec(dev, in, in_sz, out, sizeof(out));
+}
+
+static int handle_hca_cap_atomic(struct mlx5_core_dev *dev)
+{
+	void *set_ctx;
+	void *set_hca_cap;
+	int set_sz = MLX5_ST_SZ_BYTES(set_hca_cap_in);
+	int req_endianness;
+	int err;
+
+	if (MLX5_CAP_GEN(dev, atomic)) {
+		err = mlx5_core_get_caps(dev, MLX5_CAP_ATOMIC);
+		if (err)
+			return err;
+	} else {
+		return 0;
+	}
+
+	req_endianness =
+		MLX5_CAP_ATOMIC(dev,
+				supported_atomic_req_8B_endianness_mode_1);
+
+	if (req_endianness != MLX5_ATOMIC_REQ_MODE_HOST_ENDIANNESS)
+		return 0;
+
+	set_ctx = kzalloc(set_sz, GFP_KERNEL);
+	if (!set_ctx)
+		return -ENOMEM;
+
+	set_hca_cap = MLX5_ADDR_OF(set_hca_cap_in, set_ctx, capability);
+
+	/* Set requestor to host endianness */
+	MLX5_SET(atomic_caps, set_hca_cap, atomic_req_8B_endianness_mode,
+		 MLX5_ATOMIC_REQ_MODE_HOST_ENDIANNESS);
+
+	err = set_caps(dev, set_ctx, set_sz, MLX5_SET_HCA_CAP_OP_MOD_ATOMIC);
+
+	kfree(set_ctx);
+	return err;
+}
+
+static int handle_hca_cap(struct mlx5_core_dev *dev)
+{
+	void *set_ctx = NULL;
+	struct mlx5_profile *prof = dev->profile;
+	int err = -ENOMEM;
+	int set_sz = MLX5_ST_SZ_BYTES(set_hca_cap_in);
+	void *set_hca_cap;
+
+	set_ctx = kzalloc(set_sz, GFP_KERNEL);
+	if (!set_ctx)
+		goto query_ex;
+
+	err = mlx5_core_get_caps(dev, MLX5_CAP_GENERAL);
+	if (err)
+		goto query_ex;
+
+	set_hca_cap = MLX5_ADDR_OF(set_hca_cap_in, set_ctx,
+				   capability);
+	memcpy(set_hca_cap, dev->caps.hca_cur[MLX5_CAP_GENERAL],
+	       MLX5_ST_SZ_BYTES(cmd_hca_cap));
+
+	mlx5_core_dbg(dev, "Current Pkey table size %d Setting new size %d\n",
+		      mlx5_to_sw_pkey_sz(MLX5_CAP_GEN(dev, pkey_table_size)),
+		      128);
+	/* we limit the size of the pkey table to 128 entries for now */
+	MLX5_SET(cmd_hca_cap, set_hca_cap, pkey_table_size,
+		 to_fw_pkey_sz(dev, 128));
+
+	/* Check log_max_qp from HCA caps to set in current profile */
+	if (MLX5_CAP_GEN_MAX(dev, log_max_qp) < profile[prof_sel].log_max_qp) {
+		mlx5_core_warn(dev, "log_max_qp value in current profile is %d, changing it to HCA capability limit (%d)\n",
+			       profile[prof_sel].log_max_qp,
+			       MLX5_CAP_GEN_MAX(dev, log_max_qp));
+		profile[prof_sel].log_max_qp = MLX5_CAP_GEN_MAX(dev, log_max_qp);
+	}
+	if (prof->mask & MLX5_PROF_MASK_QP_SIZE)
+		MLX5_SET(cmd_hca_cap, set_hca_cap, log_max_qp,
+			 prof->log_max_qp);
+
+	/* disable cmdif checksum */
+	MLX5_SET(cmd_hca_cap, set_hca_cap, cmdif_checksum, 0);
+
+	/* Enable 4K UAR only when HCA supports it and page size is bigger
+	 * than 4K.
+	 */
+	if (MLX5_CAP_GEN_MAX(dev, uar_4k) && PAGE_SIZE > 4096)
+		MLX5_SET(cmd_hca_cap, set_hca_cap, uar_4k, 1);
+
+	MLX5_SET(cmd_hca_cap, set_hca_cap, log_uar_page_sz, PAGE_SHIFT - 12);
+
+	if (MLX5_CAP_GEN_MAX(dev, cache_line_128byte))
+		MLX5_SET(cmd_hca_cap,
+			 set_hca_cap,
+			 cache_line_128byte,
+			 cache_line_size() >= 128 ? 1 : 0);
+
+	if (MLX5_CAP_GEN_MAX(dev, dct))
+		MLX5_SET(cmd_hca_cap, set_hca_cap, dct, 1);
+
+	if (MLX5_CAP_GEN_MAX(dev, num_vhca_ports))
+		MLX5_SET(cmd_hca_cap,
+			 set_hca_cap,
+			 num_vhca_ports,
+			 MLX5_CAP_GEN_MAX(dev, num_vhca_ports));
+
+	err = set_caps(dev, set_ctx, set_sz,
+		       MLX5_SET_HCA_CAP_OP_MOD_GENERAL_DEVICE);
+
+query_ex:
+	kfree(set_ctx);
+	return err;
+}
+
+static int set_hca_ctrl(struct mlx5_core_dev *dev)
+{
+	struct mlx5_reg_host_endianness he_in;
+	struct mlx5_reg_host_endianness he_out;
+	int err;
+
+	if (!mlx5_core_is_pf(dev))
+		return 0;
+
+	memset(&he_in, 0, sizeof(he_in));
+	he_in.he = MLX5_SET_HOST_ENDIANNESS;
+	err = mlx5_core_access_reg(dev, &he_in,  sizeof(he_in),
+					&he_out, sizeof(he_out),
+					MLX5_REG_HOST_ENDIANNESS, 0, 1);
+	return err;
+}
+
+static int mlx5_core_set_hca_defaults(struct mlx5_core_dev *dev)
+{
+	int ret = 0;
+
+	/* Disable local_lb by default */
+	if (MLX5_CAP_GEN(dev, port_type) == MLX5_CAP_PORT_TYPE_ETH)
+		ret = mlx5_nic_vport_update_local_lb(dev, false);
+
+	return ret;
+}
+
+int mlx5_core_enable_hca(struct mlx5_core_dev *dev, u16 func_id)
+{
+	u32 out[MLX5_ST_SZ_DW(enable_hca_out)] = {0};
+	u32 in[MLX5_ST_SZ_DW(enable_hca_in)]   = {0};
+
+	MLX5_SET(enable_hca_in, in, opcode, MLX5_CMD_OP_ENABLE_HCA);
+	MLX5_SET(enable_hca_in, in, function_id, func_id);
+	return mlx5_cmd_exec(dev, &in, sizeof(in), &out, sizeof(out));
+}
+
+int mlx5_core_disable_hca(struct mlx5_core_dev *dev, u16 func_id)
+{
+	u32 out[MLX5_ST_SZ_DW(disable_hca_out)] = {0};
+	u32 in[MLX5_ST_SZ_DW(disable_hca_in)]   = {0};
+
+	MLX5_SET(disable_hca_in, in, opcode, MLX5_CMD_OP_DISABLE_HCA);
+	MLX5_SET(disable_hca_in, in, function_id, func_id);
+	return mlx5_cmd_exec(dev, in, sizeof(in), out, sizeof(out));
+}
+
+u64 mlx5_read_internal_timer(struct mlx5_core_dev *dev)
+{
+	u32 timer_h, timer_h1, timer_l;
+
+	timer_h = ioread32be(&dev->iseg->internal_timer_h);
+	timer_l = ioread32be(&dev->iseg->internal_timer_l);
+	timer_h1 = ioread32be(&dev->iseg->internal_timer_h);
+	if (timer_h != timer_h1) /* wrap around */
+		timer_l = ioread32be(&dev->iseg->internal_timer_l);
+
+	return (u64)timer_l | (u64)timer_h1 << 32;
+}
+
+static int mlx5_irq_set_affinity_hint(struct mlx5_core_dev *mdev, int i)
+{
+	struct mlx5_priv *priv  = &mdev->priv;
+	int vecidx = MLX5_EQ_VEC_COMP_BASE + i;
+	int irq = pci_irq_vector(mdev->pdev, vecidx);
+
+	if (!zalloc_cpumask_var(&priv->irq_info[vecidx].mask, GFP_KERNEL)) {
+		mlx5_core_warn(mdev, "zalloc_cpumask_var failed");
+		return -ENOMEM;
+	}
+
+	cpumask_set_cpu(cpumask_local_spread(i, priv->numa_node),
+			priv->irq_info[vecidx].mask);
+
+	if (IS_ENABLED(CONFIG_SMP) &&
+	    irq_set_affinity_hint(irq, priv->irq_info[vecidx].mask))
+		mlx5_core_warn(mdev, "irq_set_affinity_hint failed, irq 0x%.4x", irq);
+
+	return 0;
+}
+
+static void mlx5_irq_clear_affinity_hint(struct mlx5_core_dev *mdev, int i)
+{
+	int vecidx = MLX5_EQ_VEC_COMP_BASE + i;
+	struct mlx5_priv *priv  = &mdev->priv;
+	int irq = pci_irq_vector(mdev->pdev, vecidx);
+
+	irq_set_affinity_hint(irq, NULL);
+	free_cpumask_var(priv->irq_info[vecidx].mask);
+}
+
+static int mlx5_irq_set_affinity_hints(struct mlx5_core_dev *mdev)
+{
+	int err;
+	int i;
+
+	for (i = 0; i < mdev->priv.eq_table.num_comp_vectors; i++) {
+		err = mlx5_irq_set_affinity_hint(mdev, i);
+		if (err)
+			goto err_out;
+	}
+
+	return 0;
+
+err_out:
+	for (i--; i >= 0; i--)
+		mlx5_irq_clear_affinity_hint(mdev, i);
+
+	return err;
+}
+
+static void mlx5_irq_clear_affinity_hints(struct mlx5_core_dev *mdev)
+{
+	int i;
+
+	for (i = 0; i < mdev->priv.eq_table.num_comp_vectors; i++)
+		mlx5_irq_clear_affinity_hint(mdev, i);
+}
+
+int mlx5_vector2eqn(struct mlx5_core_dev *dev, int vector, int *eqn,
+		    unsigned int *irqn)
+{
+	struct mlx5_eq_table *table = &dev->priv.eq_table;
+	struct mlx5_eq *eq, *n;
+	int err = -ENOENT;
+
+	spin_lock(&table->lock);
+	list_for_each_entry_safe(eq, n, &table->comp_eqs_list, list) {
+		if (eq->index == vector) {
+			*eqn = eq->eqn;
+			*irqn = eq->irqn;
+			err = 0;
+			break;
+		}
+	}
+	spin_unlock(&table->lock);
+
+	return err;
+}
+EXPORT_SYMBOL(mlx5_vector2eqn);
+
+struct mlx5_eq *mlx5_eqn2eq(struct mlx5_core_dev *dev, int eqn)
+{
+	struct mlx5_eq_table *table = &dev->priv.eq_table;
+	struct mlx5_eq *eq;
+
+	spin_lock(&table->lock);
+	list_for_each_entry(eq, &table->comp_eqs_list, list)
+		if (eq->eqn == eqn) {
+			spin_unlock(&table->lock);
+			return eq;
+		}
+
+	spin_unlock(&table->lock);
+
+	return ERR_PTR(-ENOENT);
+}
+
+static void free_comp_eqs(struct mlx5_core_dev *dev)
+{
+	struct mlx5_eq_table *table = &dev->priv.eq_table;
+	struct mlx5_eq *eq, *n;
+
+#ifdef CONFIG_RFS_ACCEL
+	if (dev->rmap) {
+		free_irq_cpu_rmap(dev->rmap);
+		dev->rmap = NULL;
+	}
+#endif
+	spin_lock(&table->lock);
+	list_for_each_entry_safe(eq, n, &table->comp_eqs_list, list) {
+		list_del(&eq->list);
+		spin_unlock(&table->lock);
+		if (mlx5_destroy_unmap_eq(dev, eq))
+			mlx5_core_warn(dev, "failed to destroy EQ 0x%x\n",
+				       eq->eqn);
+		kfree(eq);
+		spin_lock(&table->lock);
+	}
+	spin_unlock(&table->lock);
+}
+
+static int alloc_comp_eqs(struct mlx5_core_dev *dev)
+{
+	struct mlx5_eq_table *table = &dev->priv.eq_table;
+	char name[MLX5_MAX_IRQ_NAME];
+	struct mlx5_eq *eq;
+	int ncomp_vec;
+	int nent;
+	int err;
+	int i;
+
+	INIT_LIST_HEAD(&table->comp_eqs_list);
+	ncomp_vec = table->num_comp_vectors;
+	nent = MLX5_COMP_EQ_SIZE;
+#ifdef CONFIG_RFS_ACCEL
+	dev->rmap = alloc_irq_cpu_rmap(ncomp_vec);
+	if (!dev->rmap)
+		return -ENOMEM;
+#endif
+	for (i = 0; i < ncomp_vec; i++) {
+		eq = kzalloc(sizeof(*eq), GFP_KERNEL);
+		if (!eq) {
+			err = -ENOMEM;
+			goto clean;
+		}
+
+#ifdef CONFIG_RFS_ACCEL
+		irq_cpu_rmap_add(dev->rmap, pci_irq_vector(dev->pdev,
+				 MLX5_EQ_VEC_COMP_BASE + i));
+#endif
+		snprintf(name, MLX5_MAX_IRQ_NAME, "mlx5_comp%d", i);
+		err = mlx5_create_map_eq(dev, eq,
+					 i + MLX5_EQ_VEC_COMP_BASE, nent, 0,
+					 name, MLX5_EQ_TYPE_COMP);
+		if (err) {
+			kfree(eq);
+			goto clean;
+		}
+		mlx5_core_dbg(dev, "allocated completion EQN %d\n", eq->eqn);
+		eq->index = i;
+		spin_lock(&table->lock);
+		list_add_tail(&eq->list, &table->comp_eqs_list);
+		spin_unlock(&table->lock);
+	}
+
+	return 0;
+
+clean:
+	free_comp_eqs(dev);
+	return err;
+}
+
+static int mlx5_core_set_issi(struct mlx5_core_dev *dev)
+{
+	u32 query_in[MLX5_ST_SZ_DW(query_issi_in)]   = {0};
+	u32 query_out[MLX5_ST_SZ_DW(query_issi_out)] = {0};
+	u32 sup_issi;
+	int err;
+
+	MLX5_SET(query_issi_in, query_in, opcode, MLX5_CMD_OP_QUERY_ISSI);
+	err = mlx5_cmd_exec(dev, query_in, sizeof(query_in),
+			    query_out, sizeof(query_out));
+	if (err) {
+		u32 syndrome;
+		u8 status;
+
+		mlx5_cmd_mbox_status(query_out, &status, &syndrome);
+		if (!status || syndrome == MLX5_DRIVER_SYND) {
+			mlx5_core_err(dev, "Failed to query ISSI err(%d) status(%d) synd(%d)\n",
+				      err, status, syndrome);
+			return err;
+		}
+
+		mlx5_core_warn(dev, "Query ISSI is not supported by FW, ISSI is 0\n");
+		dev->issi = 0;
+		return 0;
+	}
+
+	sup_issi = MLX5_GET(query_issi_out, query_out, supported_issi_dw0);
+
+	if (sup_issi & (1 << 1)) {
+		u32 set_in[MLX5_ST_SZ_DW(set_issi_in)]   = {0};
+		u32 set_out[MLX5_ST_SZ_DW(set_issi_out)] = {0};
+
+		MLX5_SET(set_issi_in, set_in, opcode, MLX5_CMD_OP_SET_ISSI);
+		MLX5_SET(set_issi_in, set_in, current_issi, 1);
+		err = mlx5_cmd_exec(dev, set_in, sizeof(set_in),
+				    set_out, sizeof(set_out));
+		if (err) {
+			mlx5_core_err(dev, "Failed to set ISSI to 1 err(%d)\n",
+				      err);
+			return err;
+		}
+
+		dev->issi = 1;
+
+		return 0;
+	} else if (sup_issi & (1 << 0) || !sup_issi) {
+		return 0;
+	}
+
+	return -EOPNOTSUPP;
+}
+
+static int mlx5_pci_init(struct mlx5_core_dev *dev, struct mlx5_priv *priv)
+{
+	struct pci_dev *pdev = dev->pdev;
+	int err = 0;
+
+	pci_set_drvdata(dev->pdev, dev);
+	strncpy(priv->name, dev_name(&pdev->dev), MLX5_MAX_NAME_LEN);
+	priv->name[MLX5_MAX_NAME_LEN - 1] = 0;
+
+	mutex_init(&priv->pgdir_mutex);
+	INIT_LIST_HEAD(&priv->pgdir_list);
+	spin_lock_init(&priv->mkey_lock);
+
+	mutex_init(&priv->alloc_mutex);
+
+	priv->numa_node = dev_to_node(&dev->pdev->dev);
+
+	if (mlx5_debugfs_root)
+		priv->dbg_root =
+			debugfs_create_dir(pci_name(pdev), mlx5_debugfs_root);
+
+	err = mlx5_pci_enable_device(dev);
+	if (err) {
+		dev_err(&pdev->dev, "Cannot enable PCI device, aborting\n");
+		goto err_dbg;
+	}
+
+	err = request_bar(pdev);
+	if (err) {
+		dev_err(&pdev->dev, "error requesting BARs, aborting\n");
+		goto err_disable;
+	}
+
+	pci_set_master(pdev);
+
+	err = set_dma_caps(pdev);
+	if (err) {
+		dev_err(&pdev->dev, "Failed setting DMA capabilities mask, aborting\n");
+		goto err_clr_master;
+	}
+
+	dev->iseg_base = pci_resource_start(dev->pdev, 0);
+	dev->iseg = ioremap(dev->iseg_base, sizeof(*dev->iseg));
+	if (!dev->iseg) {
+		err = -ENOMEM;
+		dev_err(&pdev->dev, "Failed mapping initialization segment, aborting\n");
+		goto err_clr_master;
+	}
+
+	return 0;
+
+err_clr_master:
+	pci_clear_master(dev->pdev);
+	release_bar(dev->pdev);
+err_disable:
+	mlx5_pci_disable_device(dev);
+
+err_dbg:
+	debugfs_remove(priv->dbg_root);
+	return err;
+}
+
+static void mlx5_pci_close(struct mlx5_core_dev *dev, struct mlx5_priv *priv)
+{
+	iounmap(dev->iseg);
+	pci_clear_master(dev->pdev);
+	release_bar(dev->pdev);
+	mlx5_pci_disable_device(dev);
+	debugfs_remove_recursive(priv->dbg_root);
+}
+
+static int mlx5_init_once(struct mlx5_core_dev *dev, struct mlx5_priv *priv)
+{
+	struct pci_dev *pdev = dev->pdev;
+	int err;
+
+	err = mlx5_query_board_id(dev);
+	if (err) {
+		dev_err(&pdev->dev, "query board id failed\n");
+		goto out;
+	}
+
+	err = mlx5_eq_init(dev);
+	if (err) {
+		dev_err(&pdev->dev, "failed to initialize eq\n");
+		goto out;
+	}
+
+	err = mlx5_cq_debugfs_init(dev);
+	if (err) {
+		dev_err(&pdev->dev, "failed to initialize cq debugfs\n");
+		goto err_eq_cleanup;
+	}
+
+	mlx5_init_qp_table(dev);
+
+	mlx5_init_srq_table(dev);
+
+	mlx5_init_mkey_table(dev);
+
+	mlx5_init_reserved_gids(dev);
+
+	mlx5_init_clock(dev);
+
+	dev->vxlan = mlx5_vxlan_create(dev);
+
+	err = mlx5_init_rl_table(dev);
+	if (err) {
+		dev_err(&pdev->dev, "Failed to init rate limiting\n");
+		goto err_tables_cleanup;
+	}
+
+	err = mlx5_mpfs_init(dev);
+	if (err) {
+		dev_err(&pdev->dev, "Failed to init l2 table %d\n", err);
+		goto err_rl_cleanup;
+	}
+
+	err = mlx5_eswitch_init(dev);
+	if (err) {
+		dev_err(&pdev->dev, "Failed to init eswitch %d\n", err);
+		goto err_mpfs_cleanup;
+	}
+
+	err = mlx5_sriov_init(dev);
+	if (err) {
+		dev_err(&pdev->dev, "Failed to init sriov %d\n", err);
+		goto err_eswitch_cleanup;
+	}
+
+	err = mlx5_fpga_init(dev);
+	if (err) {
+		dev_err(&pdev->dev, "Failed to init fpga device %d\n", err);
+		goto err_sriov_cleanup;
+	}
+
+	dev->tracer = mlx5_fw_tracer_create(dev);
+
+	return 0;
+
+err_sriov_cleanup:
+	mlx5_sriov_cleanup(dev);
+err_eswitch_cleanup:
+	mlx5_eswitch_cleanup(dev->priv.eswitch);
+err_mpfs_cleanup:
+	mlx5_mpfs_cleanup(dev);
+err_rl_cleanup:
+	mlx5_cleanup_rl_table(dev);
+err_tables_cleanup:
+	mlx5_vxlan_destroy(dev->vxlan);
+	mlx5_cleanup_mkey_table(dev);
+	mlx5_cleanup_srq_table(dev);
+	mlx5_cleanup_qp_table(dev);
+	mlx5_cq_debugfs_cleanup(dev);
+
+err_eq_cleanup:
+	mlx5_eq_cleanup(dev);
+
+out:
+	return err;
+}
+
+static void mlx5_cleanup_once(struct mlx5_core_dev *dev)
+{
+	mlx5_fw_tracer_destroy(dev->tracer);
+	mlx5_fpga_cleanup(dev);
+	mlx5_sriov_cleanup(dev);
+	mlx5_eswitch_cleanup(dev->priv.eswitch);
+	mlx5_mpfs_cleanup(dev);
+	mlx5_cleanup_rl_table(dev);
+	mlx5_vxlan_destroy(dev->vxlan);
+	mlx5_cleanup_clock(dev);
+	mlx5_cleanup_reserved_gids(dev);
+	mlx5_cleanup_mkey_table(dev);
+	mlx5_cleanup_srq_table(dev);
+	mlx5_cleanup_qp_table(dev);
+	mlx5_cq_debugfs_cleanup(dev);
+	mlx5_eq_cleanup(dev);
+}
+
+static int mlx5_load_one(struct mlx5_core_dev *dev, struct mlx5_priv *priv,
+			 bool boot)
+{
+	struct pci_dev *pdev = dev->pdev;
+	int err;
+
+	mutex_lock(&dev->intf_state_mutex);
+	if (test_bit(MLX5_INTERFACE_STATE_UP, &dev->intf_state)) {
+		dev_warn(&dev->pdev->dev, "%s: interface is up, NOP\n",
+			 __func__);
+		goto out;
+	}
+
+	dev_info(&pdev->dev, "firmware version: %d.%d.%d\n", fw_rev_maj(dev),
+		 fw_rev_min(dev), fw_rev_sub(dev));
+
+	/* Only PFs hold the relevant PCIe information for this query */
+	if (mlx5_core_is_pf(dev))
+		pcie_print_link_status(dev->pdev);
+
+	/* on load removing any previous indication of internal error, device is
+	 * up
+	 */
+	dev->state = MLX5_DEVICE_STATE_UP;
+
+	/* wait for firmware to accept initialization segments configurations
+	 */
+	err = wait_fw_init(dev, FW_PRE_INIT_TIMEOUT_MILI);
+	if (err) {
+		dev_err(&dev->pdev->dev, "Firmware over %d MS in pre-initializing state, aborting\n",
+			FW_PRE_INIT_TIMEOUT_MILI);
+		goto out_err;
+	}
+
+	err = mlx5_cmd_init(dev);
+	if (err) {
+		dev_err(&pdev->dev, "Failed initializing command interface, aborting\n");
+		goto out_err;
+	}
+
+	err = wait_fw_init(dev, FW_INIT_TIMEOUT_MILI);
+	if (err) {
+		dev_err(&dev->pdev->dev, "Firmware over %d MS in initializing state, aborting\n",
+			FW_INIT_TIMEOUT_MILI);
+		goto err_cmd_cleanup;
+	}
+
+	err = mlx5_core_enable_hca(dev, 0);
+	if (err) {
+		dev_err(&pdev->dev, "enable hca failed\n");
+		goto err_cmd_cleanup;
+	}
+
+	err = mlx5_core_set_issi(dev);
+	if (err) {
+		dev_err(&pdev->dev, "failed to set issi\n");
+		goto err_disable_hca;
+	}
+
+	err = mlx5_satisfy_startup_pages(dev, 1);
+	if (err) {
+		dev_err(&pdev->dev, "failed to allocate boot pages\n");
+		goto err_disable_hca;
+	}
+
+	err = set_hca_ctrl(dev);
+	if (err) {
+		dev_err(&pdev->dev, "set_hca_ctrl failed\n");
+		goto reclaim_boot_pages;
+	}
+
+	err = handle_hca_cap(dev);
+	if (err) {
+		dev_err(&pdev->dev, "handle_hca_cap failed\n");
+		goto reclaim_boot_pages;
+	}
+
+	err = handle_hca_cap_atomic(dev);
+	if (err) {
+		dev_err(&pdev->dev, "handle_hca_cap_atomic failed\n");
+		goto reclaim_boot_pages;
+	}
+
+	err = mlx5_satisfy_startup_pages(dev, 0);
+	if (err) {
+		dev_err(&pdev->dev, "failed to allocate init pages\n");
+		goto reclaim_boot_pages;
+	}
+
+	err = mlx5_pagealloc_start(dev);
+	if (err) {
+		dev_err(&pdev->dev, "mlx5_pagealloc_start failed\n");
+		goto reclaim_boot_pages;
+	}
+
+	err = mlx5_cmd_init_hca(dev, sw_owner_id);
+	if (err) {
+		dev_err(&pdev->dev, "init hca failed\n");
+		goto err_pagealloc_stop;
+	}
+
+	mlx5_set_driver_version(dev);
+
+	mlx5_start_health_poll(dev);
+
+	err = mlx5_query_hca_caps(dev);
+	if (err) {
+		dev_err(&pdev->dev, "query hca failed\n");
+		goto err_stop_poll;
+	}
+
+	if (boot) {
+		err = mlx5_init_once(dev, priv);
+		if (err) {
+			dev_err(&pdev->dev, "sw objs init failed\n");
+			goto err_stop_poll;
+		}
+	}
+
+	err = mlx5_alloc_irq_vectors(dev);
+	if (err) {
+		dev_err(&pdev->dev, "alloc irq vectors failed\n");
+		goto err_cleanup_once;
+	}
+
+	dev->priv.uar = mlx5_get_uars_page(dev);
+	if (IS_ERR(dev->priv.uar)) {
+		dev_err(&pdev->dev, "Failed allocating uar, aborting\n");
+		err = PTR_ERR(dev->priv.uar);
+		goto err_disable_msix;
+	}
+
+	err = mlx5_start_eqs(dev);
+	if (err) {
+		dev_err(&pdev->dev, "Failed to start pages and async EQs\n");
+		goto err_put_uars;
+	}
+
+	err = mlx5_fw_tracer_init(dev->tracer);
+	if (err) {
+		dev_err(&pdev->dev, "Failed to init FW tracer\n");
+		goto err_fw_tracer;
+	}
+
+	err = alloc_comp_eqs(dev);
+	if (err) {
+		dev_err(&pdev->dev, "Failed to alloc completion EQs\n");
+		goto err_comp_eqs;
+	}
+
+	err = mlx5_irq_set_affinity_hints(dev);
+	if (err) {
+		dev_err(&pdev->dev, "Failed to alloc affinity hint cpumask\n");
+		goto err_affinity_hints;
+	}
+
+	err = mlx5_fpga_device_start(dev);
+	if (err) {
+		dev_err(&pdev->dev, "fpga device start failed %d\n", err);
+		goto err_fpga_start;
+	}
+
+	err = mlx5_accel_ipsec_init(dev);
+	if (err) {
+		dev_err(&pdev->dev, "IPSec device start failed %d\n", err);
+		goto err_ipsec_start;
+	}
+
+	err = mlx5_accel_tls_init(dev);
+	if (err) {
+		dev_err(&pdev->dev, "TLS device start failed %d\n", err);
+		goto err_tls_start;
+	}
+
+	err = mlx5_init_fs(dev);
+	if (err) {
+		dev_err(&pdev->dev, "Failed to init flow steering\n");
+		goto err_fs;
+	}
+
+	err = mlx5_core_set_hca_defaults(dev);
+	if (err) {
+		dev_err(&pdev->dev, "Failed to set hca defaults\n");
+		goto err_fs;
+	}
+
+	err = mlx5_sriov_attach(dev);
+	if (err) {
+		dev_err(&pdev->dev, "sriov init failed %d\n", err);
+		goto err_sriov;
+	}
+
+	if (mlx5_device_registered(dev)) {
+		mlx5_attach_device(dev);
+	} else {
+		err = mlx5_register_device(dev);
+		if (err) {
+			dev_err(&pdev->dev, "mlx5_register_device failed %d\n", err);
+			goto err_reg_dev;
+		}
+	}
+
+	set_bit(MLX5_INTERFACE_STATE_UP, &dev->intf_state);
+out:
+	mutex_unlock(&dev->intf_state_mutex);
+
+	return 0;
+
+err_reg_dev:
+	mlx5_sriov_detach(dev);
+
+err_sriov:
+	mlx5_cleanup_fs(dev);
+
+err_fs:
+	mlx5_accel_tls_cleanup(dev);
+
+err_tls_start:
+	mlx5_accel_ipsec_cleanup(dev);
+
+err_ipsec_start:
+	mlx5_fpga_device_stop(dev);
+
+err_fpga_start:
+	mlx5_irq_clear_affinity_hints(dev);
+
+err_affinity_hints:
+	free_comp_eqs(dev);
+
+err_comp_eqs:
+	mlx5_fw_tracer_cleanup(dev->tracer);
+
+err_fw_tracer:
+	mlx5_stop_eqs(dev);
+
+err_put_uars:
+	mlx5_put_uars_page(dev, priv->uar);
+
+err_disable_msix:
+	mlx5_free_irq_vectors(dev);
+
+err_cleanup_once:
+	if (boot)
+		mlx5_cleanup_once(dev);
+
+err_stop_poll:
+	mlx5_stop_health_poll(dev, boot);
+	if (mlx5_cmd_teardown_hca(dev)) {
+		dev_err(&dev->pdev->dev, "tear_down_hca failed, skip cleanup\n");
+		goto out_err;
+	}
+
+err_pagealloc_stop:
+	mlx5_pagealloc_stop(dev);
+
+reclaim_boot_pages:
+	mlx5_reclaim_startup_pages(dev);
+
+err_disable_hca:
+	mlx5_core_disable_hca(dev, 0);
+
+err_cmd_cleanup:
+	mlx5_cmd_cleanup(dev);
+
+out_err:
+	dev->state = MLX5_DEVICE_STATE_INTERNAL_ERROR;
+	mutex_unlock(&dev->intf_state_mutex);
+
+	return err;
+}
+
+static int mlx5_unload_one(struct mlx5_core_dev *dev, struct mlx5_priv *priv,
+			   bool cleanup)
+{
+	int err = 0;
+
+	if (cleanup)
+		mlx5_drain_health_recovery(dev);
+
+	mutex_lock(&dev->intf_state_mutex);
+	if (!test_bit(MLX5_INTERFACE_STATE_UP, &dev->intf_state)) {
+		dev_warn(&dev->pdev->dev, "%s: interface is down, NOP\n",
+			 __func__);
+		if (cleanup)
+			mlx5_cleanup_once(dev);
+		goto out;
+	}
+
+	clear_bit(MLX5_INTERFACE_STATE_UP, &dev->intf_state);
+
+	if (mlx5_device_registered(dev))
+		mlx5_detach_device(dev);
+
+	mlx5_sriov_detach(dev);
+	mlx5_cleanup_fs(dev);
+	mlx5_accel_ipsec_cleanup(dev);
+	mlx5_accel_tls_cleanup(dev);
+	mlx5_fpga_device_stop(dev);
+	mlx5_irq_clear_affinity_hints(dev);
+	free_comp_eqs(dev);
+	mlx5_fw_tracer_cleanup(dev->tracer);
+	mlx5_stop_eqs(dev);
+	mlx5_put_uars_page(dev, priv->uar);
+	mlx5_free_irq_vectors(dev);
+	if (cleanup)
+		mlx5_cleanup_once(dev);
+	mlx5_stop_health_poll(dev, cleanup);
+	err = mlx5_cmd_teardown_hca(dev);
+	if (err) {
+		dev_err(&dev->pdev->dev, "tear_down_hca failed, skip cleanup\n");
+		goto out;
+	}
+	mlx5_pagealloc_stop(dev);
+	mlx5_reclaim_startup_pages(dev);
+	mlx5_core_disable_hca(dev, 0);
+	mlx5_cmd_cleanup(dev);
+
+out:
+	mutex_unlock(&dev->intf_state_mutex);
+	return err;
+}
+
+struct mlx5_core_event_handler {
+	void (*event)(struct mlx5_core_dev *dev,
+		      enum mlx5_dev_event event,
+		      void *data);
+};
+
+static const struct devlink_ops mlx5_devlink_ops = {
+#ifdef CONFIG_MLX5_ESWITCH
+	.eswitch_mode_set = mlx5_devlink_eswitch_mode_set,
+	.eswitch_mode_get = mlx5_devlink_eswitch_mode_get,
+	.eswitch_inline_mode_set = mlx5_devlink_eswitch_inline_mode_set,
+	.eswitch_inline_mode_get = mlx5_devlink_eswitch_inline_mode_get,
+	.eswitch_encap_mode_set = mlx5_devlink_eswitch_encap_mode_set,
+	.eswitch_encap_mode_get = mlx5_devlink_eswitch_encap_mode_get,
+#endif
+};
+
+#define MLX5_IB_MOD "mlx5_ib"
+static int init_one(struct pci_dev *pdev,
+		    const struct pci_device_id *id)
+{
+	struct mlx5_core_dev *dev;
+	struct devlink *devlink;
+	struct mlx5_priv *priv;
+	int err;
+
+	devlink = devlink_alloc(&mlx5_devlink_ops, sizeof(*dev));
+	if (!devlink) {
+		dev_err(&pdev->dev, "kzalloc failed\n");
+		return -ENOMEM;
+	}
+
+	dev = devlink_priv(devlink);
+	priv = &dev->priv;
+	priv->pci_dev_data = id->driver_data;
+
+	pci_set_drvdata(pdev, dev);
+
+	dev->pdev = pdev;
+	dev->event = mlx5_core_event;
+	dev->profile = &profile[prof_sel];
+
+	INIT_LIST_HEAD(&priv->ctx_list);
+	spin_lock_init(&priv->ctx_lock);
+	mutex_init(&dev->pci_status_mutex);
+	mutex_init(&dev->intf_state_mutex);
+
+	INIT_LIST_HEAD(&priv->waiting_events_list);
+	priv->is_accum_events = false;
+
+#ifdef CONFIG_INFINIBAND_ON_DEMAND_PAGING
+	err = init_srcu_struct(&priv->pfault_srcu);
+	if (err) {
+		dev_err(&pdev->dev, "init_srcu_struct failed with error code %d\n",
+			err);
+		goto clean_dev;
+	}
+#endif
+	mutex_init(&priv->bfregs.reg_head.lock);
+	mutex_init(&priv->bfregs.wc_head.lock);
+	INIT_LIST_HEAD(&priv->bfregs.reg_head.list);
+	INIT_LIST_HEAD(&priv->bfregs.wc_head.list);
+
+	err = mlx5_pci_init(dev, priv);
+	if (err) {
+		dev_err(&pdev->dev, "mlx5_pci_init failed with error code %d\n", err);
+		goto clean_srcu;
+	}
+
+	err = mlx5_health_init(dev);
+	if (err) {
+		dev_err(&pdev->dev, "mlx5_health_init failed with error code %d\n", err);
+		goto close_pci;
+	}
+
+	mlx5_pagealloc_init(dev);
+
+	err = mlx5_load_one(dev, priv, true);
+	if (err) {
+		dev_err(&pdev->dev, "mlx5_load_one failed with error code %d\n", err);
+		goto clean_health;
+	}
+
+	request_module_nowait(MLX5_IB_MOD);
+
+	err = devlink_register(devlink, &pdev->dev);
+	if (err)
+		goto clean_load;
+
+	pci_save_state(pdev);
+	return 0;
+
+clean_load:
+	mlx5_unload_one(dev, priv, true);
+clean_health:
+	mlx5_pagealloc_cleanup(dev);
+	mlx5_health_cleanup(dev);
+close_pci:
+	mlx5_pci_close(dev, priv);
+clean_srcu:
+#ifdef CONFIG_INFINIBAND_ON_DEMAND_PAGING
+	cleanup_srcu_struct(&priv->pfault_srcu);
+clean_dev:
+#endif
+	devlink_free(devlink);
+
+	return err;
+}
+
+static void remove_one(struct pci_dev *pdev)
+{
+	struct mlx5_core_dev *dev  = pci_get_drvdata(pdev);
+	struct devlink *devlink = priv_to_devlink(dev);
+	struct mlx5_priv *priv = &dev->priv;
+
+	devlink_unregister(devlink);
+	mlx5_unregister_device(dev);
+
+	if (mlx5_unload_one(dev, priv, true)) {
+		dev_err(&dev->pdev->dev, "mlx5_unload_one failed\n");
+		mlx5_health_cleanup(dev);
+		return;
+	}
+
+	mlx5_pagealloc_cleanup(dev);
+	mlx5_health_cleanup(dev);
+	mlx5_pci_close(dev, priv);
+#ifdef CONFIG_INFINIBAND_ON_DEMAND_PAGING
+	cleanup_srcu_struct(&priv->pfault_srcu);
+#endif
+	devlink_free(devlink);
+}
+
+static pci_ers_result_t mlx5_pci_err_detected(struct pci_dev *pdev,
+					      pci_channel_state_t state)
+{
+	struct mlx5_core_dev *dev = pci_get_drvdata(pdev);
+	struct mlx5_priv *priv = &dev->priv;
+
+	dev_info(&pdev->dev, "%s was called\n", __func__);
+
+	mlx5_enter_error_state(dev, false);
+	mlx5_unload_one(dev, priv, false);
+	/* In case of kernel call drain the health wq */
+	if (state) {
+		mlx5_drain_health_wq(dev);
+		mlx5_pci_disable_device(dev);
+	}
+
+	return state == pci_channel_io_perm_failure ?
+		PCI_ERS_RESULT_DISCONNECT : PCI_ERS_RESULT_NEED_RESET;
+}
+
+/* wait for the device to show vital signs by waiting
+ * for the health counter to start counting.
+ */
+static int wait_vital(struct pci_dev *pdev)
+{
+	struct mlx5_core_dev *dev = pci_get_drvdata(pdev);
+	struct mlx5_core_health *health = &dev->priv.health;
+	const int niter = 100;
+	u32 last_count = 0;
+	u32 count;
+	int i;
+
+	for (i = 0; i < niter; i++) {
+		count = ioread32be(health->health_counter);
+		if (count && count != 0xffffffff) {
+			if (last_count && last_count != count) {
+				dev_info(&pdev->dev, "Counter value 0x%x after %d iterations\n", count, i);
+				return 0;
+			}
+			last_count = count;
+		}
+		msleep(50);
+	}
+
+	return -ETIMEDOUT;
+}
+
+static pci_ers_result_t mlx5_pci_slot_reset(struct pci_dev *pdev)
+{
+	struct mlx5_core_dev *dev = pci_get_drvdata(pdev);
+	int err;
+
+	dev_info(&pdev->dev, "%s was called\n", __func__);
+
+	err = mlx5_pci_enable_device(dev);
+	if (err) {
+		dev_err(&pdev->dev, "%s: mlx5_pci_enable_device failed with error code: %d\n"
+			, __func__, err);
+		return PCI_ERS_RESULT_DISCONNECT;
+	}
+
+	pci_set_master(pdev);
+	pci_restore_state(pdev);
+	pci_save_state(pdev);
+
+	if (wait_vital(pdev)) {
+		dev_err(&pdev->dev, "%s: wait_vital timed out\n", __func__);
+		return PCI_ERS_RESULT_DISCONNECT;
+	}
+
+	return PCI_ERS_RESULT_RECOVERED;
+}
+
+static void mlx5_pci_resume(struct pci_dev *pdev)
+{
+	struct mlx5_core_dev *dev = pci_get_drvdata(pdev);
+	struct mlx5_priv *priv = &dev->priv;
+	int err;
+
+	dev_info(&pdev->dev, "%s was called\n", __func__);
+
+	err = mlx5_load_one(dev, priv, false);
+	if (err)
+		dev_err(&pdev->dev, "%s: mlx5_load_one failed with error code: %d\n"
+			, __func__, err);
+	else
+		dev_info(&pdev->dev, "%s: device recovered\n", __func__);
+}
+
+static const struct pci_error_handlers mlx5_err_handler = {
+	.error_detected = mlx5_pci_err_detected,
+	.slot_reset	= mlx5_pci_slot_reset,
+	.resume		= mlx5_pci_resume
+};
+
+static int mlx5_try_fast_unload(struct mlx5_core_dev *dev)
+{
+	int ret;
+
+	if (!MLX5_CAP_GEN(dev, force_teardown)) {
+		mlx5_core_dbg(dev, "force teardown is not supported in the firmware\n");
+		return -EOPNOTSUPP;
+	}
+
+	if (dev->state == MLX5_DEVICE_STATE_INTERNAL_ERROR) {
+		mlx5_core_dbg(dev, "Device in internal error state, giving up\n");
+		return -EAGAIN;
+	}
+
+	/* Panic tear down fw command will stop the PCI bus communication
+	 * with the HCA, so the health polll is no longer needed.
+	 */
+	mlx5_drain_health_wq(dev);
+	mlx5_stop_health_poll(dev, false);
+
+	ret = mlx5_cmd_force_teardown_hca(dev);
+	if (ret) {
+		mlx5_core_dbg(dev, "Firmware couldn't do fast unload error: %d\n", ret);
+		mlx5_start_health_poll(dev);
+		return ret;
+	}
+
+	mlx5_enter_error_state(dev, true);
+
+	/* Some platforms requiring freeing the IRQ's in the shutdown
+	 * flow. If they aren't freed they can't be allocated after
+	 * kexec. There is no need to cleanup the mlx5_core software
+	 * contexts.
+	 */
+	mlx5_irq_clear_affinity_hints(dev);
+	mlx5_core_eq_free_irqs(dev);
+
+	return 0;
+}
+
+static void shutdown(struct pci_dev *pdev)
+{
+	struct mlx5_core_dev *dev  = pci_get_drvdata(pdev);
+	struct mlx5_priv *priv = &dev->priv;
+	int err;
+
+	dev_info(&pdev->dev, "Shutdown was called\n");
+	err = mlx5_try_fast_unload(dev);
+	if (err)
+		mlx5_unload_one(dev, priv, false);
+	mlx5_pci_disable_device(dev);
+}
+
+static const struct pci_device_id mlx5_core_pci_table[] = {
+	{ PCI_VDEVICE(MELLANOX, PCI_DEVICE_ID_MELLANOX_CONNECTIB) },
+	{ PCI_VDEVICE(MELLANOX, 0x1012), MLX5_PCI_DEV_IS_VF},	/* Connect-IB VF */
+	{ PCI_VDEVICE(MELLANOX, PCI_DEVICE_ID_MELLANOX_CONNECTX4) },
+	{ PCI_VDEVICE(MELLANOX, 0x1014), MLX5_PCI_DEV_IS_VF},	/* ConnectX-4 VF */
+	{ PCI_VDEVICE(MELLANOX, PCI_DEVICE_ID_MELLANOX_CONNECTX4_LX) },
+	{ PCI_VDEVICE(MELLANOX, 0x1016), MLX5_PCI_DEV_IS_VF},	/* ConnectX-4LX VF */
+	{ PCI_VDEVICE(MELLANOX, 0x1017) },			/* ConnectX-5, PCIe 3.0 */
+	{ PCI_VDEVICE(MELLANOX, 0x1018), MLX5_PCI_DEV_IS_VF},	/* ConnectX-5 VF */
+	{ PCI_VDEVICE(MELLANOX, 0x1019) },			/* ConnectX-5 Ex */
+	{ PCI_VDEVICE(MELLANOX, 0x101a), MLX5_PCI_DEV_IS_VF},	/* ConnectX-5 Ex VF */
+	{ PCI_VDEVICE(MELLANOX, 0x101b) },			/* ConnectX-6 */
+	{ PCI_VDEVICE(MELLANOX, 0x101c), MLX5_PCI_DEV_IS_VF},	/* ConnectX-6 VF */
+	{ PCI_VDEVICE(MELLANOX, 0xa2d2) },			/* BlueField integrated ConnectX-5 network controller */
+	{ PCI_VDEVICE(MELLANOX, 0xa2d3), MLX5_PCI_DEV_IS_VF},	/* BlueField integrated ConnectX-5 network controller VF */
+	{ PCI_VDEVICE(MELLANOX, 0xa2d6) },			/* BlueField-2 integrated ConnectX-6 Dx network controller */
+	{ 0, }
+};
+
+MODULE_DEVICE_TABLE(pci, mlx5_core_pci_table);
+
+void mlx5_disable_device(struct mlx5_core_dev *dev)
+{
+	mlx5_pci_err_detected(dev->pdev, 0);
+}
+
+void mlx5_recover_device(struct mlx5_core_dev *dev)
+{
+	mlx5_pci_disable_device(dev);
+	if (mlx5_pci_slot_reset(dev->pdev) == PCI_ERS_RESULT_RECOVERED)
+		mlx5_pci_resume(dev->pdev);
+}
+
+static struct pci_driver mlx5_core_driver = {
+	.name           = DRIVER_NAME,
+	.id_table       = mlx5_core_pci_table,
+	.probe          = init_one,
+	.remove         = remove_one,
+	.shutdown	= shutdown,
+	.err_handler	= &mlx5_err_handler,
+	.sriov_configure   = mlx5_core_sriov_configure,
+};
+
+static void mlx5_core_verify_params(void)
+{
+	if (prof_sel >= ARRAY_SIZE(profile)) {
+		pr_warn("mlx5_core: WARNING: Invalid module parameter prof_sel %d, valid range 0-%zu, changing back to default(%d)\n",
+			prof_sel,
+			ARRAY_SIZE(profile) - 1,
+			MLX5_DEFAULT_PROF);
+		prof_sel = MLX5_DEFAULT_PROF;
+	}
+}
+
+static int __init init(void)
+{
+	int err;
+
+	get_random_bytes(&sw_owner_id, sizeof(sw_owner_id));
+
+	mlx5_core_verify_params();
+	mlx5_fpga_ipsec_build_fs_cmds();
+	mlx5_register_debugfs();
+
+	err = pci_register_driver(&mlx5_core_driver);
+	if (err)
+		goto err_debug;
+
+#ifdef CONFIG_MLX5_CORE_EN
+	mlx5e_init();
+#endif
+
+	return 0;
+
+err_debug:
+	mlx5_unregister_debugfs();
+	return err;
+}
+
+static void __exit cleanup(void)
+{
+#ifdef CONFIG_MLX5_CORE_EN
+	mlx5e_cleanup();
+#endif
+	pci_unregister_driver(&mlx5_core_driver);
+	mlx5_unregister_debugfs();
+}
+
+module_init(init);
+module_exit(cleanup);
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/mcg.c b/drivers/net/ethernet/mellanox/mlx5/core/mcg.c
new file mode 100644
index 000000000..ba2b09cc1
--- /dev/null
+++ b/drivers/net/ethernet/mellanox/mlx5/core/mcg.c
@@ -0,0 +1,66 @@
+/*
+ * Copyright (c) 2013-2015, 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/module.h>
+#include <linux/mlx5/driver.h>
+#include <linux/mlx5/cmd.h>
+#include <rdma/ib_verbs.h>
+#include "mlx5_core.h"
+
+int mlx5_core_attach_mcg(struct mlx5_core_dev *dev, union ib_gid *mgid, u32 qpn)
+{
+	u32 out[MLX5_ST_SZ_DW(attach_to_mcg_out)] = {0};
+	u32 in[MLX5_ST_SZ_DW(attach_to_mcg_in)]   = {0};
+	void *gid;
+
+	MLX5_SET(attach_to_mcg_in, in, opcode, MLX5_CMD_OP_ATTACH_TO_MCG);
+	MLX5_SET(attach_to_mcg_in, in, qpn, qpn);
+	gid = MLX5_ADDR_OF(attach_to_mcg_in, in, multicast_gid);
+	memcpy(gid, mgid, sizeof(*mgid));
+	return mlx5_cmd_exec(dev, in, sizeof(in), out, sizeof(out));
+}
+EXPORT_SYMBOL(mlx5_core_attach_mcg);
+
+int mlx5_core_detach_mcg(struct mlx5_core_dev *dev, union ib_gid *mgid, u32 qpn)
+{
+	u32 out[MLX5_ST_SZ_DW(detach_from_mcg_out)] = {0};
+	u32 in[MLX5_ST_SZ_DW(detach_from_mcg_in)]   = {0};
+	void *gid;
+
+	MLX5_SET(detach_from_mcg_in, in, opcode, MLX5_CMD_OP_DETACH_FROM_MCG);
+	MLX5_SET(detach_from_mcg_in, in, qpn, qpn);
+	gid = MLX5_ADDR_OF(detach_from_mcg_in, in, multicast_gid);
+	memcpy(gid, mgid, sizeof(*mgid));
+	return mlx5_cmd_exec(dev, in, sizeof(in), out, sizeof(out));
+}
+EXPORT_SYMBOL(mlx5_core_detach_mcg);
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/mlx5_core.h b/drivers/net/ethernet/mellanox/mlx5/core/mlx5_core.h
new file mode 100644
index 000000000..b4134fa0b
--- /dev/null
+++ b/drivers/net/ethernet/mellanox/mlx5/core/mlx5_core.h
@@ -0,0 +1,217 @@
+/*
+ * Copyright (c) 2013-2015, Mellanox Technologies, Ltd.  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 __MLX5_CORE_H__
+#define __MLX5_CORE_H__
+
+#include <linux/types.h>
+#include <linux/kernel.h>
+#include <linux/sched.h>
+#include <linux/if_link.h>
+#include <linux/firmware.h>
+#include <linux/mlx5/cq.h>
+
+#define DRIVER_NAME "mlx5_core"
+#define DRIVER_VERSION "5.0-0"
+
+extern uint mlx5_core_debug_mask;
+
+#define mlx5_core_dbg(__dev, format, ...)				\
+	dev_dbg(&(__dev)->pdev->dev, "%s:%d:(pid %d): " format,		\
+		 __func__, __LINE__, current->pid,			\
+		 ##__VA_ARGS__)
+
+#define mlx5_core_dbg_once(__dev, format, ...)				\
+	dev_dbg_once(&(__dev)->pdev->dev, "%s:%d:(pid %d): " format,	\
+		     __func__, __LINE__, current->pid,			\
+		     ##__VA_ARGS__)
+
+#define mlx5_core_dbg_mask(__dev, mask, format, ...)			\
+do {									\
+	if ((mask) & mlx5_core_debug_mask)				\
+		mlx5_core_dbg(__dev, format, ##__VA_ARGS__);		\
+} while (0)
+
+#define mlx5_core_err(__dev, format, ...)				\
+	dev_err(&(__dev)->pdev->dev, "%s:%d:(pid %d): " format,	\
+		__func__, __LINE__, current->pid,	\
+	       ##__VA_ARGS__)
+
+#define mlx5_core_err_rl(__dev, format, ...)				\
+	dev_err_ratelimited(&(__dev)->pdev->dev,			\
+			   "%s:%d:(pid %d): " format,			\
+			   __func__, __LINE__, current->pid,		\
+			   ##__VA_ARGS__)
+
+#define mlx5_core_warn(__dev, format, ...)				\
+	dev_warn(&(__dev)->pdev->dev, "%s:%d:(pid %d): " format,	\
+		 __func__, __LINE__, current->pid,			\
+		##__VA_ARGS__)
+
+#define mlx5_core_info(__dev, format, ...)				\
+	dev_info(&(__dev)->pdev->dev, format, ##__VA_ARGS__)
+
+enum {
+	MLX5_CMD_DATA, /* print command payload only */
+	MLX5_CMD_TIME, /* print command execution time */
+};
+
+enum {
+	MLX5_DRIVER_STATUS_ABORTED = 0xfe,
+	MLX5_DRIVER_SYND = 0xbadd00de,
+};
+
+int mlx5_query_hca_caps(struct mlx5_core_dev *dev);
+int mlx5_query_board_id(struct mlx5_core_dev *dev);
+int mlx5_cmd_init_hca(struct mlx5_core_dev *dev, uint32_t *sw_owner_id);
+int mlx5_cmd_teardown_hca(struct mlx5_core_dev *dev);
+int mlx5_cmd_force_teardown_hca(struct mlx5_core_dev *dev);
+void mlx5_core_event(struct mlx5_core_dev *dev, enum mlx5_dev_event event,
+		     unsigned long param);
+void mlx5_core_page_fault(struct mlx5_core_dev *dev,
+			  struct mlx5_pagefault *pfault);
+void mlx5_port_module_event(struct mlx5_core_dev *dev, struct mlx5_eqe *eqe);
+void mlx5_enter_error_state(struct mlx5_core_dev *dev, bool force);
+void mlx5_disable_device(struct mlx5_core_dev *dev);
+void mlx5_recover_device(struct mlx5_core_dev *dev);
+int mlx5_sriov_init(struct mlx5_core_dev *dev);
+void mlx5_sriov_cleanup(struct mlx5_core_dev *dev);
+int mlx5_sriov_attach(struct mlx5_core_dev *dev);
+void mlx5_sriov_detach(struct mlx5_core_dev *dev);
+int mlx5_core_sriov_configure(struct pci_dev *dev, int num_vfs);
+bool mlx5_sriov_is_enabled(struct mlx5_core_dev *dev);
+int mlx5_core_enable_hca(struct mlx5_core_dev *dev, u16 func_id);
+int mlx5_core_disable_hca(struct mlx5_core_dev *dev, u16 func_id);
+int mlx5_create_scheduling_element_cmd(struct mlx5_core_dev *dev, u8 hierarchy,
+				       void *context, u32 *element_id);
+int mlx5_modify_scheduling_element_cmd(struct mlx5_core_dev *dev, u8 hierarchy,
+				       void *context, u32 element_id,
+				       u32 modify_bitmask);
+int mlx5_destroy_scheduling_element_cmd(struct mlx5_core_dev *dev, u8 hierarchy,
+					u32 element_id);
+int mlx5_wait_for_vf_pages(struct mlx5_core_dev *dev);
+u64 mlx5_read_internal_timer(struct mlx5_core_dev *dev);
+
+int mlx5_eq_init(struct mlx5_core_dev *dev);
+void mlx5_eq_cleanup(struct mlx5_core_dev *dev);
+int mlx5_create_map_eq(struct mlx5_core_dev *dev, struct mlx5_eq *eq, u8 vecidx,
+		       int nent, u64 mask, const char *name,
+		       enum mlx5_eq_type type);
+int mlx5_destroy_unmap_eq(struct mlx5_core_dev *dev, struct mlx5_eq *eq);
+int mlx5_eq_add_cq(struct mlx5_eq *eq, struct mlx5_core_cq *cq);
+int mlx5_eq_del_cq(struct mlx5_eq *eq, struct mlx5_core_cq *cq);
+int mlx5_core_eq_query(struct mlx5_core_dev *dev, struct mlx5_eq *eq,
+		       u32 *out, int outlen);
+int mlx5_start_eqs(struct mlx5_core_dev *dev);
+void mlx5_stop_eqs(struct mlx5_core_dev *dev);
+/* This function should only be called after mlx5_cmd_force_teardown_hca */
+void mlx5_core_eq_free_irqs(struct mlx5_core_dev *dev);
+struct mlx5_eq *mlx5_eqn2eq(struct mlx5_core_dev *dev, int eqn);
+u32 mlx5_eq_poll_irq_disabled(struct mlx5_eq *eq);
+void mlx5_cq_tasklet_cb(unsigned long data);
+void mlx5_cmd_comp_handler(struct mlx5_core_dev *dev, u64 vec, bool forced);
+int mlx5_debug_eq_add(struct mlx5_core_dev *dev, struct mlx5_eq *eq);
+void mlx5_debug_eq_remove(struct mlx5_core_dev *dev, struct mlx5_eq *eq);
+int mlx5_eq_debugfs_init(struct mlx5_core_dev *dev);
+void mlx5_eq_debugfs_cleanup(struct mlx5_core_dev *dev);
+int mlx5_cq_debugfs_init(struct mlx5_core_dev *dev);
+void mlx5_cq_debugfs_cleanup(struct mlx5_core_dev *dev);
+
+int mlx5_query_pcam_reg(struct mlx5_core_dev *dev, u32 *pcam, u8 feature_group,
+			u8 access_reg_group);
+int mlx5_query_mcam_reg(struct mlx5_core_dev *dev, u32 *mcap, u8 feature_group,
+			u8 access_reg_group);
+int mlx5_query_qcam_reg(struct mlx5_core_dev *mdev, u32 *qcam,
+			u8 feature_group, u8 access_reg_group);
+
+void mlx5_lag_add(struct mlx5_core_dev *dev, struct net_device *netdev);
+void mlx5_lag_remove(struct mlx5_core_dev *dev);
+
+void mlx5_add_device(struct mlx5_interface *intf, struct mlx5_priv *priv);
+void mlx5_remove_device(struct mlx5_interface *intf, struct mlx5_priv *priv);
+void mlx5_attach_device(struct mlx5_core_dev *dev);
+void mlx5_detach_device(struct mlx5_core_dev *dev);
+bool mlx5_device_registered(struct mlx5_core_dev *dev);
+int mlx5_register_device(struct mlx5_core_dev *dev);
+void mlx5_unregister_device(struct mlx5_core_dev *dev);
+void mlx5_add_dev_by_protocol(struct mlx5_core_dev *dev, int protocol);
+void mlx5_remove_dev_by_protocol(struct mlx5_core_dev *dev, int protocol);
+struct mlx5_core_dev *mlx5_get_next_phys_dev(struct mlx5_core_dev *dev);
+void mlx5_dev_list_lock(void);
+void mlx5_dev_list_unlock(void);
+int mlx5_dev_list_trylock(void);
+int mlx5_encap_alloc(struct mlx5_core_dev *dev,
+		     int header_type,
+		     size_t size,
+		     void *encap_header,
+		     u32 *encap_id);
+void mlx5_encap_dealloc(struct mlx5_core_dev *dev, u32 encap_id);
+
+int mlx5_modify_header_alloc(struct mlx5_core_dev *dev,
+			     u8 namespace, u8 num_actions,
+			     void *modify_actions, u32 *modify_header_id);
+void mlx5_modify_header_dealloc(struct mlx5_core_dev *dev, u32 modify_header_id);
+
+bool mlx5_lag_intf_add(struct mlx5_interface *intf, struct mlx5_priv *priv);
+
+int mlx5_query_mtpps(struct mlx5_core_dev *dev, u32 *mtpps, u32 mtpps_size);
+int mlx5_set_mtpps(struct mlx5_core_dev *mdev, u32 *mtpps, u32 mtpps_size);
+int mlx5_query_mtppse(struct mlx5_core_dev *mdev, u8 pin, u8 *arm, u8 *mode);
+int mlx5_set_mtppse(struct mlx5_core_dev *mdev, u8 pin, u8 arm, u8 mode);
+
+#define MLX5_PPS_CAP(mdev) (MLX5_CAP_GEN((mdev), pps) &&		\
+			    MLX5_CAP_GEN((mdev), pps_modify) &&		\
+			    MLX5_CAP_MCAM_FEATURE((mdev), mtpps_fs) &&	\
+			    MLX5_CAP_MCAM_FEATURE((mdev), mtpps_enh_out_per_adj))
+
+int mlx5_firmware_flash(struct mlx5_core_dev *dev, const struct firmware *fw);
+
+void mlx5e_init(void);
+void mlx5e_cleanup(void);
+
+static inline int mlx5_lag_is_lacp_owner(struct mlx5_core_dev *dev)
+{
+	/* LACP owner conditions:
+	 * 1) Function is physical.
+	 * 2) LAG is supported by FW.
+	 * 3) LAG is managed by driver (currently the only option).
+	 */
+	return  MLX5_CAP_GEN(dev, vport_group_manager) &&
+		   (MLX5_CAP_GEN(dev, num_lag_ports) > 1) &&
+		    MLX5_CAP_GEN(dev, lag_master);
+}
+
+int mlx5_lag_allow(struct mlx5_core_dev *dev);
+int mlx5_lag_forbid(struct mlx5_core_dev *dev);
+
+void mlx5_reload_interface(struct mlx5_core_dev *mdev, int protocol);
+#endif /* __MLX5_CORE_H__ */
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/mr.c b/drivers/net/ethernet/mellanox/mlx5/core/mr.c
new file mode 100644
index 000000000..0670165af
--- /dev/null
+++ b/drivers/net/ethernet/mellanox/mlx5/core/mr.c
@@ -0,0 +1,193 @@
+/*
+ * Copyright (c) 2013-2015, 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/module.h>
+#include <linux/mlx5/driver.h>
+#include <linux/mlx5/cmd.h>
+#include "mlx5_core.h"
+
+void mlx5_init_mkey_table(struct mlx5_core_dev *dev)
+{
+	struct mlx5_mkey_table *table = &dev->priv.mkey_table;
+
+	memset(table, 0, sizeof(*table));
+	rwlock_init(&table->lock);
+	INIT_RADIX_TREE(&table->tree, GFP_ATOMIC);
+}
+
+void mlx5_cleanup_mkey_table(struct mlx5_core_dev *dev)
+{
+}
+
+int mlx5_core_create_mkey_cb(struct mlx5_core_dev *dev,
+			     struct mlx5_core_mkey *mkey,
+			     u32 *in, int inlen,
+			     u32 *out, int outlen,
+			     mlx5_cmd_cbk_t callback, void *context)
+{
+	struct mlx5_mkey_table *table = &dev->priv.mkey_table;
+	u32 lout[MLX5_ST_SZ_DW(create_mkey_out)] = {0};
+	u32 mkey_index;
+	void *mkc;
+	int err;
+	u8 key;
+
+	spin_lock_irq(&dev->priv.mkey_lock);
+	key = dev->priv.mkey_key++;
+	spin_unlock_irq(&dev->priv.mkey_lock);
+	mkc = MLX5_ADDR_OF(create_mkey_in, in, memory_key_mkey_entry);
+
+	MLX5_SET(create_mkey_in, in, opcode, MLX5_CMD_OP_CREATE_MKEY);
+	MLX5_SET(mkc, mkc, mkey_7_0, key);
+
+	if (callback)
+		return mlx5_cmd_exec_cb(dev, in, inlen, out, outlen,
+					callback, context);
+
+	err = mlx5_cmd_exec(dev, in, inlen, lout, sizeof(lout));
+	if (err)
+		return err;
+
+	mkey_index = MLX5_GET(create_mkey_out, lout, mkey_index);
+	mkey->iova = MLX5_GET64(mkc, mkc, start_addr);
+	mkey->size = MLX5_GET64(mkc, mkc, len);
+	mkey->key = mlx5_idx_to_mkey(mkey_index) | key;
+	mkey->pd = MLX5_GET(mkc, mkc, pd);
+
+	mlx5_core_dbg(dev, "out 0x%x, key 0x%x, mkey 0x%x\n",
+		      mkey_index, key, mkey->key);
+
+	/* connect to mkey tree */
+	write_lock_irq(&table->lock);
+	err = radix_tree_insert(&table->tree, mlx5_base_mkey(mkey->key), mkey);
+	write_unlock_irq(&table->lock);
+	if (err) {
+		mlx5_core_warn(dev, "failed radix tree insert of mkey 0x%x, %d\n",
+			       mlx5_base_mkey(mkey->key), err);
+		mlx5_core_destroy_mkey(dev, mkey);
+	}
+
+	return err;
+}
+EXPORT_SYMBOL(mlx5_core_create_mkey_cb);
+
+int mlx5_core_create_mkey(struct mlx5_core_dev *dev,
+			  struct mlx5_core_mkey *mkey,
+			  u32 *in, int inlen)
+{
+	return mlx5_core_create_mkey_cb(dev, mkey, in, inlen,
+					NULL, 0, NULL, NULL);
+}
+EXPORT_SYMBOL(mlx5_core_create_mkey);
+
+int mlx5_core_destroy_mkey(struct mlx5_core_dev *dev,
+			   struct mlx5_core_mkey *mkey)
+{
+	struct mlx5_mkey_table *table = &dev->priv.mkey_table;
+	u32 out[MLX5_ST_SZ_DW(destroy_mkey_out)] = {0};
+	u32 in[MLX5_ST_SZ_DW(destroy_mkey_in)]   = {0};
+	struct mlx5_core_mkey *deleted_mkey;
+	unsigned long flags;
+
+	write_lock_irqsave(&table->lock, flags);
+	deleted_mkey = radix_tree_delete(&table->tree, mlx5_base_mkey(mkey->key));
+	write_unlock_irqrestore(&table->lock, flags);
+	if (!deleted_mkey) {
+		mlx5_core_dbg(dev, "failed radix tree delete of mkey 0x%x\n",
+			      mlx5_base_mkey(mkey->key));
+		return -ENOENT;
+	}
+
+	MLX5_SET(destroy_mkey_in, in, opcode, MLX5_CMD_OP_DESTROY_MKEY);
+	MLX5_SET(destroy_mkey_in, in, mkey_index, mlx5_mkey_to_idx(mkey->key));
+	return mlx5_cmd_exec(dev, in, sizeof(in), out, sizeof(out));
+}
+EXPORT_SYMBOL(mlx5_core_destroy_mkey);
+
+int mlx5_core_query_mkey(struct mlx5_core_dev *dev, struct mlx5_core_mkey *mkey,
+			 u32 *out, int outlen)
+{
+	u32 in[MLX5_ST_SZ_DW(query_mkey_in)] = {0};
+
+	memset(out, 0, outlen);
+	MLX5_SET(query_mkey_in, in, opcode, MLX5_CMD_OP_QUERY_MKEY);
+	MLX5_SET(query_mkey_in, in, mkey_index, mlx5_mkey_to_idx(mkey->key));
+	return mlx5_cmd_exec(dev, in, sizeof(in), out, outlen);
+}
+EXPORT_SYMBOL(mlx5_core_query_mkey);
+
+static inline u32 mlx5_get_psv(u32 *out, int psv_index)
+{
+	switch (psv_index) {
+	case 1: return MLX5_GET(create_psv_out, out, psv1_index);
+	case 2: return MLX5_GET(create_psv_out, out, psv2_index);
+	case 3: return MLX5_GET(create_psv_out, out, psv3_index);
+	default: return MLX5_GET(create_psv_out, out, psv0_index);
+	}
+}
+
+int mlx5_core_create_psv(struct mlx5_core_dev *dev, u32 pdn,
+			 int npsvs, u32 *sig_index)
+{
+	u32 out[MLX5_ST_SZ_DW(create_psv_out)] = {0};
+	u32 in[MLX5_ST_SZ_DW(create_psv_in)]   = {0};
+	int i, err;
+
+	if (npsvs > MLX5_MAX_PSVS)
+		return -EINVAL;
+
+	MLX5_SET(create_psv_in, in, opcode, MLX5_CMD_OP_CREATE_PSV);
+	MLX5_SET(create_psv_in, in, pd, pdn);
+	MLX5_SET(create_psv_in, in, num_psv, npsvs);
+
+	err = mlx5_cmd_exec(dev, in, sizeof(in), out, sizeof(out));
+	if (err)
+		return err;
+
+	for (i = 0; i < npsvs; i++)
+		sig_index[i] = mlx5_get_psv(out, i);
+
+	return err;
+}
+EXPORT_SYMBOL(mlx5_core_create_psv);
+
+int mlx5_core_destroy_psv(struct mlx5_core_dev *dev, int psv_num)
+{
+	u32 out[MLX5_ST_SZ_DW(destroy_psv_out)] = {0};
+	u32 in[MLX5_ST_SZ_DW(destroy_psv_in)]   = {0};
+
+	MLX5_SET(destroy_psv_in, in, opcode, MLX5_CMD_OP_DESTROY_PSV);
+	MLX5_SET(destroy_psv_in, in, psvn, psv_num);
+	return mlx5_cmd_exec(dev, in, sizeof(in), out, sizeof(out));
+}
+EXPORT_SYMBOL(mlx5_core_destroy_psv);
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/pagealloc.c b/drivers/net/ethernet/mellanox/mlx5/core/pagealloc.c
new file mode 100644
index 000000000..9c3653e06
--- /dev/null
+++ b/drivers/net/ethernet/mellanox/mlx5/core/pagealloc.c
@@ -0,0 +1,596 @@
+/*
+ * Copyright (c) 2013-2015, 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/highmem.h>
+#include <linux/kernel.h>
+#include <linux/module.h>
+#include <linux/delay.h>
+#include <linux/mlx5/driver.h>
+#include <linux/mlx5/cmd.h>
+#include "mlx5_core.h"
+
+enum {
+	MLX5_PAGES_CANT_GIVE	= 0,
+	MLX5_PAGES_GIVE		= 1,
+	MLX5_PAGES_TAKE		= 2
+};
+
+struct mlx5_pages_req {
+	struct mlx5_core_dev *dev;
+	u16	func_id;
+	s32	npages;
+	struct work_struct work;
+};
+
+struct fw_page {
+	struct rb_node		rb_node;
+	u64			addr;
+	struct page	       *page;
+	u16			func_id;
+	unsigned long		bitmask;
+	struct list_head	list;
+	unsigned		free_count;
+};
+
+enum {
+	MAX_RECLAIM_TIME_MSECS	= 5000,
+	MAX_RECLAIM_VFS_PAGES_TIME_MSECS = 2 * 1000 * 60,
+};
+
+enum {
+	MLX5_MAX_RECLAIM_TIME_MILI	= 5000,
+	MLX5_NUM_4K_IN_PAGE		= PAGE_SIZE / MLX5_ADAPTER_PAGE_SIZE,
+};
+
+static int insert_page(struct mlx5_core_dev *dev, u64 addr, struct page *page, u16 func_id)
+{
+	struct rb_root *root = &dev->priv.page_root;
+	struct rb_node **new = &root->rb_node;
+	struct rb_node *parent = NULL;
+	struct fw_page *nfp;
+	struct fw_page *tfp;
+	int i;
+
+	while (*new) {
+		parent = *new;
+		tfp = rb_entry(parent, struct fw_page, rb_node);
+		if (tfp->addr < addr)
+			new = &parent->rb_left;
+		else if (tfp->addr > addr)
+			new = &parent->rb_right;
+		else
+			return -EEXIST;
+	}
+
+	nfp = kzalloc(sizeof(*nfp), GFP_KERNEL);
+	if (!nfp)
+		return -ENOMEM;
+
+	nfp->addr = addr;
+	nfp->page = page;
+	nfp->func_id = func_id;
+	nfp->free_count = MLX5_NUM_4K_IN_PAGE;
+	for (i = 0; i < MLX5_NUM_4K_IN_PAGE; i++)
+		set_bit(i, &nfp->bitmask);
+
+	rb_link_node(&nfp->rb_node, parent, new);
+	rb_insert_color(&nfp->rb_node, root);
+	list_add(&nfp->list, &dev->priv.free_list);
+
+	return 0;
+}
+
+static struct fw_page *find_fw_page(struct mlx5_core_dev *dev, u64 addr)
+{
+	struct rb_root *root = &dev->priv.page_root;
+	struct rb_node *tmp = root->rb_node;
+	struct fw_page *result = NULL;
+	struct fw_page *tfp;
+
+	while (tmp) {
+		tfp = rb_entry(tmp, struct fw_page, rb_node);
+		if (tfp->addr < addr) {
+			tmp = tmp->rb_left;
+		} else if (tfp->addr > addr) {
+			tmp = tmp->rb_right;
+		} else {
+			result = tfp;
+			break;
+		}
+	}
+
+	return result;
+}
+
+static int mlx5_cmd_query_pages(struct mlx5_core_dev *dev, u16 *func_id,
+				s32 *npages, int boot)
+{
+	u32 out[MLX5_ST_SZ_DW(query_pages_out)] = {0};
+	u32 in[MLX5_ST_SZ_DW(query_pages_in)]   = {0};
+	int err;
+
+	MLX5_SET(query_pages_in, in, opcode, MLX5_CMD_OP_QUERY_PAGES);
+	MLX5_SET(query_pages_in, in, op_mod, boot ?
+		 MLX5_QUERY_PAGES_IN_OP_MOD_BOOT_PAGES :
+		 MLX5_QUERY_PAGES_IN_OP_MOD_INIT_PAGES);
+
+	err = mlx5_cmd_exec(dev, in, sizeof(in), out, sizeof(out));
+	if (err)
+		return err;
+
+	*npages = MLX5_GET(query_pages_out, out, num_pages);
+	*func_id = MLX5_GET(query_pages_out, out, function_id);
+
+	return err;
+}
+
+static int alloc_4k(struct mlx5_core_dev *dev, u64 *addr)
+{
+	struct fw_page *fp;
+	unsigned n;
+
+	if (list_empty(&dev->priv.free_list))
+		return -ENOMEM;
+
+	fp = list_entry(dev->priv.free_list.next, struct fw_page, list);
+	n = find_first_bit(&fp->bitmask, 8 * sizeof(fp->bitmask));
+	if (n >= MLX5_NUM_4K_IN_PAGE) {
+		mlx5_core_warn(dev, "alloc 4k bug\n");
+		return -ENOENT;
+	}
+	clear_bit(n, &fp->bitmask);
+	fp->free_count--;
+	if (!fp->free_count)
+		list_del(&fp->list);
+
+	*addr = fp->addr + n * MLX5_ADAPTER_PAGE_SIZE;
+
+	return 0;
+}
+
+#define MLX5_U64_4K_PAGE_MASK ((~(u64)0U) << PAGE_SHIFT)
+
+static void free_4k(struct mlx5_core_dev *dev, u64 addr)
+{
+	struct fw_page *fwp;
+	int n;
+
+	fwp = find_fw_page(dev, addr & MLX5_U64_4K_PAGE_MASK);
+	if (!fwp) {
+		mlx5_core_warn(dev, "page not found\n");
+		return;
+	}
+
+	n = (addr & ~MLX5_U64_4K_PAGE_MASK) >> MLX5_ADAPTER_PAGE_SHIFT;
+	fwp->free_count++;
+	set_bit(n, &fwp->bitmask);
+	if (fwp->free_count == MLX5_NUM_4K_IN_PAGE) {
+		rb_erase(&fwp->rb_node, &dev->priv.page_root);
+		if (fwp->free_count != 1)
+			list_del(&fwp->list);
+		dma_unmap_page(&dev->pdev->dev, addr & MLX5_U64_4K_PAGE_MASK,
+			       PAGE_SIZE, DMA_BIDIRECTIONAL);
+		__free_page(fwp->page);
+		kfree(fwp);
+	} else if (fwp->free_count == 1) {
+		list_add(&fwp->list, &dev->priv.free_list);
+	}
+}
+
+static int alloc_system_page(struct mlx5_core_dev *dev, u16 func_id)
+{
+	struct page *page;
+	u64 zero_addr = 1;
+	u64 addr;
+	int err;
+	int nid = dev_to_node(&dev->pdev->dev);
+
+	page = alloc_pages_node(nid, GFP_HIGHUSER, 0);
+	if (!page) {
+		mlx5_core_warn(dev, "failed to allocate page\n");
+		return -ENOMEM;
+	}
+map:
+	addr = dma_map_page(&dev->pdev->dev, page, 0,
+			    PAGE_SIZE, DMA_BIDIRECTIONAL);
+	if (dma_mapping_error(&dev->pdev->dev, addr)) {
+		mlx5_core_warn(dev, "failed dma mapping page\n");
+		err = -ENOMEM;
+		goto err_mapping;
+	}
+
+	/* Firmware doesn't support page with physical address 0 */
+	if (addr == 0) {
+		zero_addr = addr;
+		goto map;
+	}
+
+	err = insert_page(dev, addr, page, func_id);
+	if (err) {
+		mlx5_core_err(dev, "failed to track allocated page\n");
+		dma_unmap_page(&dev->pdev->dev, addr, PAGE_SIZE,
+			       DMA_BIDIRECTIONAL);
+	}
+
+err_mapping:
+	if (err)
+		__free_page(page);
+
+	if (zero_addr == 0)
+		dma_unmap_page(&dev->pdev->dev, zero_addr, PAGE_SIZE,
+			       DMA_BIDIRECTIONAL);
+
+	return err;
+}
+
+static void page_notify_fail(struct mlx5_core_dev *dev, u16 func_id)
+{
+	u32 out[MLX5_ST_SZ_DW(manage_pages_out)] = {0};
+	u32 in[MLX5_ST_SZ_DW(manage_pages_in)]   = {0};
+	int err;
+
+	MLX5_SET(manage_pages_in, in, opcode, MLX5_CMD_OP_MANAGE_PAGES);
+	MLX5_SET(manage_pages_in, in, op_mod, MLX5_PAGES_CANT_GIVE);
+	MLX5_SET(manage_pages_in, in, function_id, func_id);
+
+	err = mlx5_cmd_exec(dev, in, sizeof(in), out, sizeof(out));
+	if (err)
+		mlx5_core_warn(dev, "page notify failed func_id(%d) err(%d)\n",
+			       func_id, err);
+}
+
+static int give_pages(struct mlx5_core_dev *dev, u16 func_id, int npages,
+		      int notify_fail)
+{
+	u32 out[MLX5_ST_SZ_DW(manage_pages_out)] = {0};
+	int inlen = MLX5_ST_SZ_BYTES(manage_pages_in);
+	u64 addr;
+	int err;
+	u32 *in;
+	int i;
+
+	inlen += npages * MLX5_FLD_SZ_BYTES(manage_pages_in, pas[0]);
+	in = kvzalloc(inlen, GFP_KERNEL);
+	if (!in) {
+		err = -ENOMEM;
+		mlx5_core_warn(dev, "vzalloc failed %d\n", inlen);
+		goto out_free;
+	}
+
+	for (i = 0; i < npages; i++) {
+retry:
+		err = alloc_4k(dev, &addr);
+		if (err) {
+			if (err == -ENOMEM)
+				err = alloc_system_page(dev, func_id);
+			if (err)
+				goto out_4k;
+
+			goto retry;
+		}
+		MLX5_ARRAY_SET64(manage_pages_in, in, pas, i, addr);
+	}
+
+	MLX5_SET(manage_pages_in, in, opcode, MLX5_CMD_OP_MANAGE_PAGES);
+	MLX5_SET(manage_pages_in, in, op_mod, MLX5_PAGES_GIVE);
+	MLX5_SET(manage_pages_in, in, function_id, func_id);
+	MLX5_SET(manage_pages_in, in, input_num_entries, npages);
+
+	err = mlx5_cmd_exec(dev, in, inlen, out, sizeof(out));
+	if (err) {
+		mlx5_core_warn(dev, "func_id 0x%x, npages %d, err %d\n",
+			       func_id, npages, err);
+		goto out_4k;
+	}
+
+	dev->priv.fw_pages += npages;
+	if (func_id)
+		dev->priv.vfs_pages += npages;
+
+	mlx5_core_dbg(dev, "err %d\n", err);
+
+	kvfree(in);
+	return 0;
+
+out_4k:
+	for (i--; i >= 0; i--)
+		free_4k(dev, MLX5_GET64(manage_pages_in, in, pas[i]));
+out_free:
+	kvfree(in);
+	if (notify_fail)
+		page_notify_fail(dev, func_id);
+	return err;
+}
+
+static u32 fwp_fill_manage_pages_out(struct fw_page *fwp, u32 *out, u32 index,
+				     u32 npages)
+{
+	u32 pages_set = 0;
+	unsigned int n;
+
+	for_each_clear_bit(n, &fwp->bitmask, MLX5_NUM_4K_IN_PAGE) {
+		MLX5_ARRAY_SET64(manage_pages_out, out, pas, index + pages_set,
+				 fwp->addr + (n * MLX5_ADAPTER_PAGE_SIZE));
+		pages_set++;
+
+		if (!--npages)
+			break;
+	}
+
+	return pages_set;
+}
+
+static int reclaim_pages_cmd(struct mlx5_core_dev *dev,
+			     u32 *in, int in_size, u32 *out, int out_size)
+{
+	struct fw_page *fwp;
+	struct rb_node *p;
+	u32 func_id;
+	u32 npages;
+	u32 i = 0;
+
+	if (dev->state != MLX5_DEVICE_STATE_INTERNAL_ERROR)
+		return mlx5_cmd_exec(dev, in, in_size, out, out_size);
+
+	/* No hard feelings, we want our pages back! */
+	npages = MLX5_GET(manage_pages_in, in, input_num_entries);
+	func_id = MLX5_GET(manage_pages_in, in, function_id);
+
+	p = rb_first(&dev->priv.page_root);
+	while (p && i < npages) {
+		fwp = rb_entry(p, struct fw_page, rb_node);
+		p = rb_next(p);
+		if (fwp->func_id != func_id)
+			continue;
+
+		i += fwp_fill_manage_pages_out(fwp, out, i, npages - i);
+	}
+
+	MLX5_SET(manage_pages_out, out, output_num_entries, i);
+	return 0;
+}
+
+static int reclaim_pages(struct mlx5_core_dev *dev, u32 func_id, int npages,
+			 int *nclaimed)
+{
+	int outlen = MLX5_ST_SZ_BYTES(manage_pages_out);
+	u32 in[MLX5_ST_SZ_DW(manage_pages_in)] = {0};
+	int num_claimed;
+	u32 *out;
+	int err;
+	int i;
+
+	if (nclaimed)
+		*nclaimed = 0;
+
+	outlen += npages * MLX5_FLD_SZ_BYTES(manage_pages_out, pas[0]);
+	out = kvzalloc(outlen, GFP_KERNEL);
+	if (!out)
+		return -ENOMEM;
+
+	MLX5_SET(manage_pages_in, in, opcode, MLX5_CMD_OP_MANAGE_PAGES);
+	MLX5_SET(manage_pages_in, in, op_mod, MLX5_PAGES_TAKE);
+	MLX5_SET(manage_pages_in, in, function_id, func_id);
+	MLX5_SET(manage_pages_in, in, input_num_entries, npages);
+
+	mlx5_core_dbg(dev, "npages %d, outlen %d\n", npages, outlen);
+	err = reclaim_pages_cmd(dev, in, sizeof(in), out, outlen);
+	if (err) {
+		mlx5_core_err(dev, "failed reclaiming pages: err %d\n", err);
+		goto out_free;
+	}
+
+	num_claimed = MLX5_GET(manage_pages_out, out, output_num_entries);
+	if (num_claimed > npages) {
+		mlx5_core_warn(dev, "fw returned %d, driver asked %d => corruption\n",
+			       num_claimed, npages);
+		err = -EINVAL;
+		goto out_free;
+	}
+
+	for (i = 0; i < num_claimed; i++)
+		free_4k(dev, MLX5_GET64(manage_pages_out, out, pas[i]));
+
+	if (nclaimed)
+		*nclaimed = num_claimed;
+
+	dev->priv.fw_pages -= num_claimed;
+	if (func_id)
+		dev->priv.vfs_pages -= num_claimed;
+
+out_free:
+	kvfree(out);
+	return err;
+}
+
+static void pages_work_handler(struct work_struct *work)
+{
+	struct mlx5_pages_req *req = container_of(work, struct mlx5_pages_req, work);
+	struct mlx5_core_dev *dev = req->dev;
+	int err = 0;
+
+	if (req->npages < 0)
+		err = reclaim_pages(dev, req->func_id, -1 * req->npages, NULL);
+	else if (req->npages > 0)
+		err = give_pages(dev, req->func_id, req->npages, 1);
+
+	if (err)
+		mlx5_core_warn(dev, "%s fail %d\n",
+			       req->npages < 0 ? "reclaim" : "give", err);
+
+	kfree(req);
+}
+
+void mlx5_core_req_pages_handler(struct mlx5_core_dev *dev, u16 func_id,
+				 s32 npages)
+{
+	struct mlx5_pages_req *req;
+
+	req = kzalloc(sizeof(*req), GFP_ATOMIC);
+	if (!req) {
+		mlx5_core_warn(dev, "failed to allocate pages request\n");
+		return;
+	}
+
+	req->dev = dev;
+	req->func_id = func_id;
+	req->npages = npages;
+	INIT_WORK(&req->work, pages_work_handler);
+	queue_work(dev->priv.pg_wq, &req->work);
+}
+
+int mlx5_satisfy_startup_pages(struct mlx5_core_dev *dev, int boot)
+{
+	u16 uninitialized_var(func_id);
+	s32 uninitialized_var(npages);
+	int err;
+
+	err = mlx5_cmd_query_pages(dev, &func_id, &npages, boot);
+	if (err)
+		return err;
+
+	mlx5_core_dbg(dev, "requested %d %s pages for func_id 0x%x\n",
+		      npages, boot ? "boot" : "init", func_id);
+
+	return give_pages(dev, func_id, npages, 0);
+}
+
+enum {
+	MLX5_BLKS_FOR_RECLAIM_PAGES = 12
+};
+
+static int optimal_reclaimed_pages(void)
+{
+	struct mlx5_cmd_prot_block *block;
+	struct mlx5_cmd_layout *lay;
+	int ret;
+
+	ret = (sizeof(lay->out) + MLX5_BLKS_FOR_RECLAIM_PAGES * sizeof(block->data) -
+	       MLX5_ST_SZ_BYTES(manage_pages_out)) /
+	       MLX5_FLD_SZ_BYTES(manage_pages_out, pas[0]);
+
+	return ret;
+}
+
+int mlx5_reclaim_startup_pages(struct mlx5_core_dev *dev)
+{
+	unsigned long end = jiffies + msecs_to_jiffies(MAX_RECLAIM_TIME_MSECS);
+	struct fw_page *fwp;
+	struct rb_node *p;
+	int nclaimed = 0;
+	int err = 0;
+
+	do {
+		p = rb_first(&dev->priv.page_root);
+		if (p) {
+			fwp = rb_entry(p, struct fw_page, rb_node);
+			err = reclaim_pages(dev, fwp->func_id,
+					    optimal_reclaimed_pages(),
+					    &nclaimed);
+
+			if (err) {
+				mlx5_core_warn(dev, "failed reclaiming pages (%d)\n",
+					       err);
+				return err;
+			}
+			if (nclaimed)
+				end = jiffies + msecs_to_jiffies(MAX_RECLAIM_TIME_MSECS);
+		}
+		if (time_after(jiffies, end)) {
+			mlx5_core_warn(dev, "FW did not return all pages. giving up...\n");
+			break;
+		}
+	} while (p);
+
+	WARN(dev->priv.fw_pages,
+	     "FW pages counter is %d after reclaiming all pages\n",
+	     dev->priv.fw_pages);
+	WARN(dev->priv.vfs_pages,
+	     "VFs FW pages counter is %d after reclaiming all pages\n",
+	     dev->priv.vfs_pages);
+
+	return 0;
+}
+
+void mlx5_pagealloc_init(struct mlx5_core_dev *dev)
+{
+	dev->priv.page_root = RB_ROOT;
+	INIT_LIST_HEAD(&dev->priv.free_list);
+}
+
+void mlx5_pagealloc_cleanup(struct mlx5_core_dev *dev)
+{
+	/* nothing */
+}
+
+int mlx5_pagealloc_start(struct mlx5_core_dev *dev)
+{
+	dev->priv.pg_wq = create_singlethread_workqueue("mlx5_page_allocator");
+	if (!dev->priv.pg_wq)
+		return -ENOMEM;
+
+	return 0;
+}
+
+void mlx5_pagealloc_stop(struct mlx5_core_dev *dev)
+{
+	destroy_workqueue(dev->priv.pg_wq);
+}
+
+int mlx5_wait_for_vf_pages(struct mlx5_core_dev *dev)
+{
+	unsigned long end = jiffies + msecs_to_jiffies(MAX_RECLAIM_VFS_PAGES_TIME_MSECS);
+	int prev_vfs_pages = dev->priv.vfs_pages;
+
+	/* In case of internal error we will free the pages manually later */
+	if (dev->state == MLX5_DEVICE_STATE_INTERNAL_ERROR) {
+		mlx5_core_warn(dev, "Skipping wait for vf pages stage");
+		return 0;
+	}
+
+	mlx5_core_dbg(dev, "Waiting for %d pages from %s\n", prev_vfs_pages,
+		      dev->priv.name);
+	while (dev->priv.vfs_pages) {
+		if (time_after(jiffies, end)) {
+			mlx5_core_warn(dev, "aborting while there are %d pending pages\n", dev->priv.vfs_pages);
+			return -ETIMEDOUT;
+		}
+		if (dev->priv.vfs_pages < prev_vfs_pages) {
+			end = jiffies + msecs_to_jiffies(MAX_RECLAIM_VFS_PAGES_TIME_MSECS);
+			prev_vfs_pages = dev->priv.vfs_pages;
+		}
+		msleep(50);
+	}
+
+	mlx5_core_dbg(dev, "All pages received from %s\n", dev->priv.name);
+	return 0;
+}
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/pd.c b/drivers/net/ethernet/mellanox/mlx5/core/pd.c
new file mode 100644
index 000000000..bd830d8d6
--- /dev/null
+++ b/drivers/net/ethernet/mellanox/mlx5/core/pd.c
@@ -0,0 +1,62 @@
+/*
+ * Copyright (c) 2013-2015, 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/module.h>
+#include <linux/mlx5/driver.h>
+#include <linux/mlx5/cmd.h>
+#include "mlx5_core.h"
+
+int mlx5_core_alloc_pd(struct mlx5_core_dev *dev, u32 *pdn)
+{
+	u32 out[MLX5_ST_SZ_DW(alloc_pd_out)] = {0};
+	u32 in[MLX5_ST_SZ_DW(alloc_pd_in)]   = {0};
+	int err;
+
+	MLX5_SET(alloc_pd_in, in, opcode, MLX5_CMD_OP_ALLOC_PD);
+	err = mlx5_cmd_exec(dev, in, sizeof(in), out, sizeof(out));
+	if (!err)
+		*pdn = MLX5_GET(alloc_pd_out, out, pd);
+	return err;
+}
+EXPORT_SYMBOL(mlx5_core_alloc_pd);
+
+int mlx5_core_dealloc_pd(struct mlx5_core_dev *dev, u32 pdn)
+{
+	u32 out[MLX5_ST_SZ_DW(dealloc_pd_out)] = {0};
+	u32 in[MLX5_ST_SZ_DW(dealloc_pd_in)]   = {0};
+
+	MLX5_SET(dealloc_pd_in, in, opcode, MLX5_CMD_OP_DEALLOC_PD);
+	MLX5_SET(dealloc_pd_in, in, pd, pdn);
+	return mlx5_cmd_exec(dev, in, sizeof(in), out, sizeof(out));
+}
+EXPORT_SYMBOL(mlx5_core_dealloc_pd);
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/port.c b/drivers/net/ethernet/mellanox/mlx5/core/port.c
new file mode 100644
index 000000000..09b6b1bfb
--- /dev/null
+++ b/drivers/net/ethernet/mellanox/mlx5/core/port.c
@@ -0,0 +1,1116 @@
+/*
+ * Copyright (c) 2013-2015, 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/module.h>
+#include <linux/mlx5/driver.h>
+#include <linux/mlx5/port.h>
+#include <linux/mlx5/cmd.h>
+#include "mlx5_core.h"
+
+int mlx5_core_access_reg(struct mlx5_core_dev *dev, void *data_in,
+			 int size_in, void *data_out, int size_out,
+			 u16 reg_id, int arg, int write)
+{
+	int outlen = MLX5_ST_SZ_BYTES(access_register_out) + size_out;
+	int inlen = MLX5_ST_SZ_BYTES(access_register_in) + size_in;
+	int err = -ENOMEM;
+	u32 *out = NULL;
+	u32 *in = NULL;
+	void *data;
+
+	in = kvzalloc(inlen, GFP_KERNEL);
+	out = kvzalloc(outlen, GFP_KERNEL);
+	if (!in || !out)
+		goto out;
+
+	data = MLX5_ADDR_OF(access_register_in, in, register_data);
+	memcpy(data, data_in, size_in);
+
+	MLX5_SET(access_register_in, in, opcode, MLX5_CMD_OP_ACCESS_REG);
+	MLX5_SET(access_register_in, in, op_mod, !write);
+	MLX5_SET(access_register_in, in, argument, arg);
+	MLX5_SET(access_register_in, in, register_id, reg_id);
+
+	err = mlx5_cmd_exec(dev, in, inlen, out, outlen);
+	if (err)
+		goto out;
+
+	data = MLX5_ADDR_OF(access_register_out, out, register_data);
+	memcpy(data_out, data, size_out);
+
+out:
+	kvfree(out);
+	kvfree(in);
+	return err;
+}
+EXPORT_SYMBOL_GPL(mlx5_core_access_reg);
+
+int mlx5_query_pcam_reg(struct mlx5_core_dev *dev, u32 *pcam, u8 feature_group,
+			u8 access_reg_group)
+{
+	u32 in[MLX5_ST_SZ_DW(pcam_reg)] = {0};
+	int sz = MLX5_ST_SZ_BYTES(pcam_reg);
+
+	MLX5_SET(pcam_reg, in, feature_group, feature_group);
+	MLX5_SET(pcam_reg, in, access_reg_group, access_reg_group);
+
+	return mlx5_core_access_reg(dev, in, sz, pcam, sz, MLX5_REG_PCAM, 0, 0);
+}
+
+int mlx5_query_mcam_reg(struct mlx5_core_dev *dev, u32 *mcam, u8 feature_group,
+			u8 access_reg_group)
+{
+	u32 in[MLX5_ST_SZ_DW(mcam_reg)] = {0};
+	int sz = MLX5_ST_SZ_BYTES(mcam_reg);
+
+	MLX5_SET(mcam_reg, in, feature_group, feature_group);
+	MLX5_SET(mcam_reg, in, access_reg_group, access_reg_group);
+
+	return mlx5_core_access_reg(dev, in, sz, mcam, sz, MLX5_REG_MCAM, 0, 0);
+}
+
+int mlx5_query_qcam_reg(struct mlx5_core_dev *mdev, u32 *qcam,
+			u8 feature_group, u8 access_reg_group)
+{
+	u32 in[MLX5_ST_SZ_DW(qcam_reg)] = {};
+	int sz = MLX5_ST_SZ_BYTES(qcam_reg);
+
+	MLX5_SET(qcam_reg, in, feature_group, feature_group);
+	MLX5_SET(qcam_reg, in, access_reg_group, access_reg_group);
+
+	return mlx5_core_access_reg(mdev, in, sz, qcam, sz, MLX5_REG_QCAM, 0, 0);
+}
+
+struct mlx5_reg_pcap {
+	u8			rsvd0;
+	u8			port_num;
+	u8			rsvd1[2];
+	__be32			caps_127_96;
+	__be32			caps_95_64;
+	__be32			caps_63_32;
+	__be32			caps_31_0;
+};
+
+int mlx5_set_port_caps(struct mlx5_core_dev *dev, u8 port_num, u32 caps)
+{
+	struct mlx5_reg_pcap in;
+	struct mlx5_reg_pcap out;
+
+	memset(&in, 0, sizeof(in));
+	in.caps_127_96 = cpu_to_be32(caps);
+	in.port_num = port_num;
+
+	return mlx5_core_access_reg(dev, &in, sizeof(in), &out,
+				    sizeof(out), MLX5_REG_PCAP, 0, 1);
+}
+EXPORT_SYMBOL_GPL(mlx5_set_port_caps);
+
+int mlx5_query_port_ptys(struct mlx5_core_dev *dev, u32 *ptys,
+			 int ptys_size, int proto_mask, u8 local_port)
+{
+	u32 in[MLX5_ST_SZ_DW(ptys_reg)] = {0};
+
+	MLX5_SET(ptys_reg, in, local_port, local_port);
+	MLX5_SET(ptys_reg, in, proto_mask, proto_mask);
+	return mlx5_core_access_reg(dev, in, sizeof(in), ptys,
+				    ptys_size, MLX5_REG_PTYS, 0, 0);
+}
+EXPORT_SYMBOL_GPL(mlx5_query_port_ptys);
+
+int mlx5_set_port_beacon(struct mlx5_core_dev *dev, u16 beacon_duration)
+{
+	u32 in[MLX5_ST_SZ_DW(mlcr_reg)]  = {0};
+	u32 out[MLX5_ST_SZ_DW(mlcr_reg)];
+
+	MLX5_SET(mlcr_reg, in, local_port, 1);
+	MLX5_SET(mlcr_reg, in, beacon_duration, beacon_duration);
+	return mlx5_core_access_reg(dev, in, sizeof(in), out,
+				    sizeof(out), MLX5_REG_MLCR, 0, 1);
+}
+
+int mlx5_query_port_proto_cap(struct mlx5_core_dev *dev,
+			      u32 *proto_cap, int proto_mask)
+{
+	u32 out[MLX5_ST_SZ_DW(ptys_reg)];
+	int err;
+
+	err = mlx5_query_port_ptys(dev, out, sizeof(out), proto_mask, 1);
+	if (err)
+		return err;
+
+	if (proto_mask == MLX5_PTYS_EN)
+		*proto_cap = MLX5_GET(ptys_reg, out, eth_proto_capability);
+	else
+		*proto_cap = MLX5_GET(ptys_reg, out, ib_proto_capability);
+
+	return 0;
+}
+EXPORT_SYMBOL_GPL(mlx5_query_port_proto_cap);
+
+int mlx5_query_port_proto_admin(struct mlx5_core_dev *dev,
+				u32 *proto_admin, int proto_mask)
+{
+	u32 out[MLX5_ST_SZ_DW(ptys_reg)];
+	int err;
+
+	err = mlx5_query_port_ptys(dev, out, sizeof(out), proto_mask, 1);
+	if (err)
+		return err;
+
+	if (proto_mask == MLX5_PTYS_EN)
+		*proto_admin = MLX5_GET(ptys_reg, out, eth_proto_admin);
+	else
+		*proto_admin = MLX5_GET(ptys_reg, out, ib_proto_admin);
+
+	return 0;
+}
+EXPORT_SYMBOL_GPL(mlx5_query_port_proto_admin);
+
+int mlx5_query_port_link_width_oper(struct mlx5_core_dev *dev,
+				    u8 *link_width_oper, u8 local_port)
+{
+	u32 out[MLX5_ST_SZ_DW(ptys_reg)];
+	int err;
+
+	err = mlx5_query_port_ptys(dev, out, sizeof(out), MLX5_PTYS_IB, local_port);
+	if (err)
+		return err;
+
+	*link_width_oper = MLX5_GET(ptys_reg, out, ib_link_width_oper);
+
+	return 0;
+}
+EXPORT_SYMBOL_GPL(mlx5_query_port_link_width_oper);
+
+int mlx5_query_port_eth_proto_oper(struct mlx5_core_dev *dev,
+				   u32 *proto_oper, u8 local_port)
+{
+	u32 out[MLX5_ST_SZ_DW(ptys_reg)];
+	int err;
+
+	err = mlx5_query_port_ptys(dev, out, sizeof(out), MLX5_PTYS_EN,
+				   local_port);
+	if (err)
+		return err;
+
+	*proto_oper = MLX5_GET(ptys_reg, out, eth_proto_oper);
+
+	return 0;
+}
+EXPORT_SYMBOL(mlx5_query_port_eth_proto_oper);
+
+int mlx5_query_port_ib_proto_oper(struct mlx5_core_dev *dev,
+				  u8 *proto_oper, u8 local_port)
+{
+	u32 out[MLX5_ST_SZ_DW(ptys_reg)];
+	int err;
+
+	err = mlx5_query_port_ptys(dev, out, sizeof(out), MLX5_PTYS_IB,
+				   local_port);
+	if (err)
+		return err;
+
+	*proto_oper = MLX5_GET(ptys_reg, out, ib_proto_oper);
+
+	return 0;
+}
+EXPORT_SYMBOL(mlx5_query_port_ib_proto_oper);
+
+int mlx5_set_port_ptys(struct mlx5_core_dev *dev, bool an_disable,
+		       u32 proto_admin, int proto_mask)
+{
+	u32 out[MLX5_ST_SZ_DW(ptys_reg)];
+	u32 in[MLX5_ST_SZ_DW(ptys_reg)];
+	u8 an_disable_admin;
+	u8 an_disable_cap;
+	u8 an_status;
+
+	mlx5_query_port_autoneg(dev, proto_mask, &an_status,
+				&an_disable_cap, &an_disable_admin);
+	if (!an_disable_cap && an_disable)
+		return -EPERM;
+
+	memset(in, 0, sizeof(in));
+
+	MLX5_SET(ptys_reg, in, local_port, 1);
+	MLX5_SET(ptys_reg, in, an_disable_admin, an_disable);
+	MLX5_SET(ptys_reg, in, proto_mask, proto_mask);
+	if (proto_mask == MLX5_PTYS_EN)
+		MLX5_SET(ptys_reg, in, eth_proto_admin, proto_admin);
+	else
+		MLX5_SET(ptys_reg, in, ib_proto_admin, proto_admin);
+
+	return mlx5_core_access_reg(dev, in, sizeof(in), out,
+				    sizeof(out), MLX5_REG_PTYS, 0, 1);
+}
+EXPORT_SYMBOL_GPL(mlx5_set_port_ptys);
+
+/* This function should be used after setting a port register only */
+void mlx5_toggle_port_link(struct mlx5_core_dev *dev)
+{
+	enum mlx5_port_status ps;
+
+	mlx5_query_port_admin_status(dev, &ps);
+	mlx5_set_port_admin_status(dev, MLX5_PORT_DOWN);
+	if (ps == MLX5_PORT_UP)
+		mlx5_set_port_admin_status(dev, MLX5_PORT_UP);
+}
+EXPORT_SYMBOL_GPL(mlx5_toggle_port_link);
+
+int mlx5_set_port_admin_status(struct mlx5_core_dev *dev,
+			       enum mlx5_port_status status)
+{
+	u32 in[MLX5_ST_SZ_DW(paos_reg)] = {0};
+	u32 out[MLX5_ST_SZ_DW(paos_reg)];
+
+	MLX5_SET(paos_reg, in, local_port, 1);
+	MLX5_SET(paos_reg, in, admin_status, status);
+	MLX5_SET(paos_reg, in, ase, 1);
+	return mlx5_core_access_reg(dev, in, sizeof(in), out,
+				    sizeof(out), MLX5_REG_PAOS, 0, 1);
+}
+EXPORT_SYMBOL_GPL(mlx5_set_port_admin_status);
+
+int mlx5_query_port_admin_status(struct mlx5_core_dev *dev,
+				 enum mlx5_port_status *status)
+{
+	u32 in[MLX5_ST_SZ_DW(paos_reg)] = {0};
+	u32 out[MLX5_ST_SZ_DW(paos_reg)];
+	int err;
+
+	MLX5_SET(paos_reg, in, local_port, 1);
+	err = mlx5_core_access_reg(dev, in, sizeof(in), out,
+				   sizeof(out), MLX5_REG_PAOS, 0, 0);
+	if (err)
+		return err;
+	*status = MLX5_GET(paos_reg, out, admin_status);
+	return 0;
+}
+EXPORT_SYMBOL_GPL(mlx5_query_port_admin_status);
+
+static void mlx5_query_port_mtu(struct mlx5_core_dev *dev, u16 *admin_mtu,
+				u16 *max_mtu, u16 *oper_mtu, u8 port)
+{
+	u32 in[MLX5_ST_SZ_DW(pmtu_reg)] = {0};
+	u32 out[MLX5_ST_SZ_DW(pmtu_reg)];
+
+	MLX5_SET(pmtu_reg, in, local_port, port);
+	mlx5_core_access_reg(dev, in, sizeof(in), out,
+			     sizeof(out), MLX5_REG_PMTU, 0, 0);
+
+	if (max_mtu)
+		*max_mtu  = MLX5_GET(pmtu_reg, out, max_mtu);
+	if (oper_mtu)
+		*oper_mtu = MLX5_GET(pmtu_reg, out, oper_mtu);
+	if (admin_mtu)
+		*admin_mtu = MLX5_GET(pmtu_reg, out, admin_mtu);
+}
+
+int mlx5_set_port_mtu(struct mlx5_core_dev *dev, u16 mtu, u8 port)
+{
+	u32 in[MLX5_ST_SZ_DW(pmtu_reg)] = {0};
+	u32 out[MLX5_ST_SZ_DW(pmtu_reg)];
+
+	MLX5_SET(pmtu_reg, in, admin_mtu, mtu);
+	MLX5_SET(pmtu_reg, in, local_port, port);
+	return mlx5_core_access_reg(dev, in, sizeof(in), out,
+				   sizeof(out), MLX5_REG_PMTU, 0, 1);
+}
+EXPORT_SYMBOL_GPL(mlx5_set_port_mtu);
+
+void mlx5_query_port_max_mtu(struct mlx5_core_dev *dev, u16 *max_mtu,
+			     u8 port)
+{
+	mlx5_query_port_mtu(dev, NULL, max_mtu, NULL, port);
+}
+EXPORT_SYMBOL_GPL(mlx5_query_port_max_mtu);
+
+void mlx5_query_port_oper_mtu(struct mlx5_core_dev *dev, u16 *oper_mtu,
+			      u8 port)
+{
+	mlx5_query_port_mtu(dev, NULL, NULL, oper_mtu, port);
+}
+EXPORT_SYMBOL_GPL(mlx5_query_port_oper_mtu);
+
+static int mlx5_query_module_num(struct mlx5_core_dev *dev, int *module_num)
+{
+	u32 in[MLX5_ST_SZ_DW(pmlp_reg)] = {0};
+	u32 out[MLX5_ST_SZ_DW(pmlp_reg)];
+	int module_mapping;
+	int err;
+
+	MLX5_SET(pmlp_reg, in, local_port, 1);
+	err = mlx5_core_access_reg(dev, in, sizeof(in), out, sizeof(out),
+				   MLX5_REG_PMLP, 0, 0);
+	if (err)
+		return err;
+
+	module_mapping = MLX5_GET(pmlp_reg, out, lane0_module_mapping);
+	*module_num = module_mapping & MLX5_EEPROM_IDENTIFIER_BYTE_MASK;
+
+	return 0;
+}
+
+int mlx5_query_module_eeprom(struct mlx5_core_dev *dev,
+			     u16 offset, u16 size, u8 *data)
+{
+	u32 out[MLX5_ST_SZ_DW(mcia_reg)];
+	u32 in[MLX5_ST_SZ_DW(mcia_reg)];
+	int module_num;
+	u16 i2c_addr;
+	int status;
+	int err;
+	void *ptr = MLX5_ADDR_OF(mcia_reg, out, dword_0);
+
+	err = mlx5_query_module_num(dev, &module_num);
+	if (err)
+		return err;
+
+	memset(in, 0, sizeof(in));
+	size = min_t(int, size, MLX5_EEPROM_MAX_BYTES);
+
+	if (offset < MLX5_EEPROM_PAGE_LENGTH &&
+	    offset + size > MLX5_EEPROM_PAGE_LENGTH)
+		/* Cross pages read, read until offset 256 in low page */
+		size -= offset + size - MLX5_EEPROM_PAGE_LENGTH;
+
+	i2c_addr = MLX5_I2C_ADDR_LOW;
+
+	MLX5_SET(mcia_reg, in, l, 0);
+	MLX5_SET(mcia_reg, in, module, module_num);
+	MLX5_SET(mcia_reg, in, i2c_device_address, i2c_addr);
+	MLX5_SET(mcia_reg, in, page_number, 0);
+	MLX5_SET(mcia_reg, in, device_address, offset);
+	MLX5_SET(mcia_reg, in, size, size);
+
+	err = mlx5_core_access_reg(dev, in, sizeof(in), out,
+				   sizeof(out), MLX5_REG_MCIA, 0, 0);
+	if (err)
+		return err;
+
+	status = MLX5_GET(mcia_reg, out, status);
+	if (status) {
+		mlx5_core_err(dev, "query_mcia_reg failed: status: 0x%x\n",
+			      status);
+		return -EIO;
+	}
+
+	memcpy(data, ptr, size);
+
+	return size;
+}
+EXPORT_SYMBOL_GPL(mlx5_query_module_eeprom);
+
+static int mlx5_query_port_pvlc(struct mlx5_core_dev *dev, u32 *pvlc,
+				int pvlc_size,  u8 local_port)
+{
+	u32 in[MLX5_ST_SZ_DW(pvlc_reg)] = {0};
+
+	MLX5_SET(pvlc_reg, in, local_port, local_port);
+	return mlx5_core_access_reg(dev, in, sizeof(in), pvlc,
+				    pvlc_size, MLX5_REG_PVLC, 0, 0);
+}
+
+int mlx5_query_port_vl_hw_cap(struct mlx5_core_dev *dev,
+			      u8 *vl_hw_cap, u8 local_port)
+{
+	u32 out[MLX5_ST_SZ_DW(pvlc_reg)];
+	int err;
+
+	err = mlx5_query_port_pvlc(dev, out, sizeof(out), local_port);
+	if (err)
+		return err;
+
+	*vl_hw_cap = MLX5_GET(pvlc_reg, out, vl_hw_cap);
+
+	return 0;
+}
+EXPORT_SYMBOL_GPL(mlx5_query_port_vl_hw_cap);
+
+int mlx5_core_query_ib_ppcnt(struct mlx5_core_dev *dev,
+			     u8 port_num, void *out, size_t sz)
+{
+	u32 *in;
+	int err;
+
+	in  = kvzalloc(sz, GFP_KERNEL);
+	if (!in) {
+		err = -ENOMEM;
+		return err;
+	}
+
+	MLX5_SET(ppcnt_reg, in, local_port, port_num);
+
+	MLX5_SET(ppcnt_reg, in, grp, MLX5_INFINIBAND_PORT_COUNTERS_GROUP);
+	err = mlx5_core_access_reg(dev, in, sz, out,
+				   sz, MLX5_REG_PPCNT, 0, 0);
+
+	kvfree(in);
+	return err;
+}
+EXPORT_SYMBOL_GPL(mlx5_core_query_ib_ppcnt);
+
+static int mlx5_query_pfcc_reg(struct mlx5_core_dev *dev, u32 *out,
+			       u32 out_size)
+{
+	u32 in[MLX5_ST_SZ_DW(pfcc_reg)] = {0};
+
+	MLX5_SET(pfcc_reg, in, local_port, 1);
+
+	return mlx5_core_access_reg(dev, in, sizeof(in), out,
+				    out_size, MLX5_REG_PFCC, 0, 0);
+}
+
+int mlx5_set_port_pause(struct mlx5_core_dev *dev, u32 rx_pause, u32 tx_pause)
+{
+	u32 in[MLX5_ST_SZ_DW(pfcc_reg)] = {0};
+	u32 out[MLX5_ST_SZ_DW(pfcc_reg)];
+
+	MLX5_SET(pfcc_reg, in, local_port, 1);
+	MLX5_SET(pfcc_reg, in, pptx, tx_pause);
+	MLX5_SET(pfcc_reg, in, pprx, rx_pause);
+
+	return mlx5_core_access_reg(dev, in, sizeof(in), out,
+				    sizeof(out), MLX5_REG_PFCC, 0, 1);
+}
+EXPORT_SYMBOL_GPL(mlx5_set_port_pause);
+
+int mlx5_query_port_pause(struct mlx5_core_dev *dev,
+			  u32 *rx_pause, u32 *tx_pause)
+{
+	u32 out[MLX5_ST_SZ_DW(pfcc_reg)];
+	int err;
+
+	err = mlx5_query_pfcc_reg(dev, out, sizeof(out));
+	if (err)
+		return err;
+
+	if (rx_pause)
+		*rx_pause = MLX5_GET(pfcc_reg, out, pprx);
+
+	if (tx_pause)
+		*tx_pause = MLX5_GET(pfcc_reg, out, pptx);
+
+	return 0;
+}
+EXPORT_SYMBOL_GPL(mlx5_query_port_pause);
+
+int mlx5_set_port_stall_watermark(struct mlx5_core_dev *dev,
+				  u16 stall_critical_watermark,
+				  u16 stall_minor_watermark)
+{
+	u32 in[MLX5_ST_SZ_DW(pfcc_reg)] = {0};
+	u32 out[MLX5_ST_SZ_DW(pfcc_reg)];
+
+	MLX5_SET(pfcc_reg, in, local_port, 1);
+	MLX5_SET(pfcc_reg, in, pptx_mask_n, 1);
+	MLX5_SET(pfcc_reg, in, pprx_mask_n, 1);
+	MLX5_SET(pfcc_reg, in, ppan_mask_n, 1);
+	MLX5_SET(pfcc_reg, in, critical_stall_mask, 1);
+	MLX5_SET(pfcc_reg, in, minor_stall_mask, 1);
+	MLX5_SET(pfcc_reg, in, device_stall_critical_watermark,
+		 stall_critical_watermark);
+	MLX5_SET(pfcc_reg, in, device_stall_minor_watermark, stall_minor_watermark);
+
+	return mlx5_core_access_reg(dev, in, sizeof(in), out,
+				    sizeof(out), MLX5_REG_PFCC, 0, 1);
+}
+
+int mlx5_query_port_stall_watermark(struct mlx5_core_dev *dev,
+				    u16 *stall_critical_watermark,
+				    u16 *stall_minor_watermark)
+{
+	u32 out[MLX5_ST_SZ_DW(pfcc_reg)];
+	int err;
+
+	err = mlx5_query_pfcc_reg(dev, out, sizeof(out));
+	if (err)
+		return err;
+
+	if (stall_critical_watermark)
+		*stall_critical_watermark = MLX5_GET(pfcc_reg, out,
+						     device_stall_critical_watermark);
+
+	if (stall_minor_watermark)
+		*stall_minor_watermark = MLX5_GET(pfcc_reg, out,
+						  device_stall_minor_watermark);
+
+	return 0;
+}
+
+int mlx5_set_port_pfc(struct mlx5_core_dev *dev, u8 pfc_en_tx, u8 pfc_en_rx)
+{
+	u32 in[MLX5_ST_SZ_DW(pfcc_reg)] = {0};
+	u32 out[MLX5_ST_SZ_DW(pfcc_reg)];
+
+	MLX5_SET(pfcc_reg, in, local_port, 1);
+	MLX5_SET(pfcc_reg, in, pfctx, pfc_en_tx);
+	MLX5_SET(pfcc_reg, in, pfcrx, pfc_en_rx);
+	MLX5_SET_TO_ONES(pfcc_reg, in, prio_mask_tx);
+	MLX5_SET_TO_ONES(pfcc_reg, in, prio_mask_rx);
+
+	return mlx5_core_access_reg(dev, in, sizeof(in), out,
+				    sizeof(out), MLX5_REG_PFCC, 0, 1);
+}
+EXPORT_SYMBOL_GPL(mlx5_set_port_pfc);
+
+int mlx5_query_port_pfc(struct mlx5_core_dev *dev, u8 *pfc_en_tx, u8 *pfc_en_rx)
+{
+	u32 out[MLX5_ST_SZ_DW(pfcc_reg)];
+	int err;
+
+	err = mlx5_query_pfcc_reg(dev, out, sizeof(out));
+	if (err)
+		return err;
+
+	if (pfc_en_tx)
+		*pfc_en_tx = MLX5_GET(pfcc_reg, out, pfctx);
+
+	if (pfc_en_rx)
+		*pfc_en_rx = MLX5_GET(pfcc_reg, out, pfcrx);
+
+	return 0;
+}
+EXPORT_SYMBOL_GPL(mlx5_query_port_pfc);
+
+void mlx5_query_port_autoneg(struct mlx5_core_dev *dev, int proto_mask,
+			     u8 *an_status,
+			     u8 *an_disable_cap, u8 *an_disable_admin)
+{
+	u32 out[MLX5_ST_SZ_DW(ptys_reg)];
+
+	*an_status = 0;
+	*an_disable_cap = 0;
+	*an_disable_admin = 0;
+
+	if (mlx5_query_port_ptys(dev, out, sizeof(out), proto_mask, 1))
+		return;
+
+	*an_status = MLX5_GET(ptys_reg, out, an_status);
+	*an_disable_cap = MLX5_GET(ptys_reg, out, an_disable_cap);
+	*an_disable_admin = MLX5_GET(ptys_reg, out, an_disable_admin);
+}
+EXPORT_SYMBOL_GPL(mlx5_query_port_autoneg);
+
+int mlx5_max_tc(struct mlx5_core_dev *mdev)
+{
+	u8 num_tc = MLX5_CAP_GEN(mdev, max_tc) ? : 8;
+
+	return num_tc - 1;
+}
+
+int mlx5_query_port_dcbx_param(struct mlx5_core_dev *mdev, u32 *out)
+{
+	u32 in[MLX5_ST_SZ_DW(dcbx_param)] = {0};
+
+	MLX5_SET(dcbx_param, in, port_number, 1);
+
+	return  mlx5_core_access_reg(mdev, in, sizeof(in), out,
+				    sizeof(in), MLX5_REG_DCBX_PARAM, 0, 0);
+}
+
+int mlx5_set_port_dcbx_param(struct mlx5_core_dev *mdev, u32 *in)
+{
+	u32 out[MLX5_ST_SZ_DW(dcbx_param)];
+
+	MLX5_SET(dcbx_param, in, port_number, 1);
+
+	return mlx5_core_access_reg(mdev, in, sizeof(out), out,
+				    sizeof(out), MLX5_REG_DCBX_PARAM, 0, 1);
+}
+
+int mlx5_set_port_prio_tc(struct mlx5_core_dev *mdev, u8 *prio_tc)
+{
+	u32 in[MLX5_ST_SZ_DW(qtct_reg)] = {0};
+	u32 out[MLX5_ST_SZ_DW(qtct_reg)];
+	int err;
+	int i;
+
+	for (i = 0; i < 8; i++) {
+		if (prio_tc[i] > mlx5_max_tc(mdev))
+			return -EINVAL;
+
+		MLX5_SET(qtct_reg, in, prio, i);
+		MLX5_SET(qtct_reg, in, tclass, prio_tc[i]);
+
+		err = mlx5_core_access_reg(mdev, in, sizeof(in), out,
+					   sizeof(out), MLX5_REG_QTCT, 0, 1);
+		if (err)
+			return err;
+	}
+
+	return 0;
+}
+EXPORT_SYMBOL_GPL(mlx5_set_port_prio_tc);
+
+int mlx5_query_port_prio_tc(struct mlx5_core_dev *mdev,
+			    u8 prio, u8 *tc)
+{
+	u32 in[MLX5_ST_SZ_DW(qtct_reg)];
+	u32 out[MLX5_ST_SZ_DW(qtct_reg)];
+	int err;
+
+	memset(in, 0, sizeof(in));
+	memset(out, 0, sizeof(out));
+
+	MLX5_SET(qtct_reg, in, port_number, 1);
+	MLX5_SET(qtct_reg, in, prio, prio);
+
+	err = mlx5_core_access_reg(mdev, in, sizeof(in), out,
+				   sizeof(out), MLX5_REG_QTCT, 0, 0);
+	if (!err)
+		*tc = MLX5_GET(qtct_reg, out, tclass);
+
+	return err;
+}
+EXPORT_SYMBOL_GPL(mlx5_query_port_prio_tc);
+
+static int mlx5_set_port_qetcr_reg(struct mlx5_core_dev *mdev, u32 *in,
+				   int inlen)
+{
+	u32 out[MLX5_ST_SZ_DW(qetc_reg)];
+
+	if (!MLX5_CAP_GEN(mdev, ets))
+		return -EOPNOTSUPP;
+
+	return mlx5_core_access_reg(mdev, in, inlen, out, sizeof(out),
+				    MLX5_REG_QETCR, 0, 1);
+}
+
+static int mlx5_query_port_qetcr_reg(struct mlx5_core_dev *mdev, u32 *out,
+				     int outlen)
+{
+	u32 in[MLX5_ST_SZ_DW(qetc_reg)];
+
+	if (!MLX5_CAP_GEN(mdev, ets))
+		return -EOPNOTSUPP;
+
+	memset(in, 0, sizeof(in));
+	return mlx5_core_access_reg(mdev, in, sizeof(in), out, outlen,
+				    MLX5_REG_QETCR, 0, 0);
+}
+
+int mlx5_set_port_tc_group(struct mlx5_core_dev *mdev, u8 *tc_group)
+{
+	u32 in[MLX5_ST_SZ_DW(qetc_reg)] = {0};
+	int i;
+
+	for (i = 0; i <= mlx5_max_tc(mdev); i++) {
+		MLX5_SET(qetc_reg, in, tc_configuration[i].g, 1);
+		MLX5_SET(qetc_reg, in, tc_configuration[i].group, tc_group[i]);
+	}
+
+	return mlx5_set_port_qetcr_reg(mdev, in, sizeof(in));
+}
+EXPORT_SYMBOL_GPL(mlx5_set_port_tc_group);
+
+int mlx5_query_port_tc_group(struct mlx5_core_dev *mdev,
+			     u8 tc, u8 *tc_group)
+{
+	u32 out[MLX5_ST_SZ_DW(qetc_reg)];
+	void *ets_tcn_conf;
+	int err;
+
+	err = mlx5_query_port_qetcr_reg(mdev, out, sizeof(out));
+	if (err)
+		return err;
+
+	ets_tcn_conf = MLX5_ADDR_OF(qetc_reg, out,
+				    tc_configuration[tc]);
+
+	*tc_group = MLX5_GET(ets_tcn_config_reg, ets_tcn_conf,
+			     group);
+
+	return 0;
+}
+EXPORT_SYMBOL_GPL(mlx5_query_port_tc_group);
+
+int mlx5_set_port_tc_bw_alloc(struct mlx5_core_dev *mdev, u8 *tc_bw)
+{
+	u32 in[MLX5_ST_SZ_DW(qetc_reg)] = {0};
+	int i;
+
+	for (i = 0; i <= mlx5_max_tc(mdev); i++) {
+		MLX5_SET(qetc_reg, in, tc_configuration[i].b, 1);
+		MLX5_SET(qetc_reg, in, tc_configuration[i].bw_allocation, tc_bw[i]);
+	}
+
+	return mlx5_set_port_qetcr_reg(mdev, in, sizeof(in));
+}
+EXPORT_SYMBOL_GPL(mlx5_set_port_tc_bw_alloc);
+
+int mlx5_query_port_tc_bw_alloc(struct mlx5_core_dev *mdev,
+				u8 tc, u8 *bw_pct)
+{
+	u32 out[MLX5_ST_SZ_DW(qetc_reg)];
+	void *ets_tcn_conf;
+	int err;
+
+	err = mlx5_query_port_qetcr_reg(mdev, out, sizeof(out));
+	if (err)
+		return err;
+
+	ets_tcn_conf = MLX5_ADDR_OF(qetc_reg, out,
+				    tc_configuration[tc]);
+
+	*bw_pct = MLX5_GET(ets_tcn_config_reg, ets_tcn_conf,
+			   bw_allocation);
+
+	return 0;
+}
+EXPORT_SYMBOL_GPL(mlx5_query_port_tc_bw_alloc);
+
+int mlx5_modify_port_ets_rate_limit(struct mlx5_core_dev *mdev,
+				    u8 *max_bw_value,
+				    u8 *max_bw_units)
+{
+	u32 in[MLX5_ST_SZ_DW(qetc_reg)] = {0};
+	void *ets_tcn_conf;
+	int i;
+
+	MLX5_SET(qetc_reg, in, port_number, 1);
+
+	for (i = 0; i <= mlx5_max_tc(mdev); i++) {
+		ets_tcn_conf = MLX5_ADDR_OF(qetc_reg, in, tc_configuration[i]);
+
+		MLX5_SET(ets_tcn_config_reg, ets_tcn_conf, r, 1);
+		MLX5_SET(ets_tcn_config_reg, ets_tcn_conf, max_bw_units,
+			 max_bw_units[i]);
+		MLX5_SET(ets_tcn_config_reg, ets_tcn_conf, max_bw_value,
+			 max_bw_value[i]);
+	}
+
+	return mlx5_set_port_qetcr_reg(mdev, in, sizeof(in));
+}
+EXPORT_SYMBOL_GPL(mlx5_modify_port_ets_rate_limit);
+
+int mlx5_query_port_ets_rate_limit(struct mlx5_core_dev *mdev,
+				   u8 *max_bw_value,
+				   u8 *max_bw_units)
+{
+	u32 out[MLX5_ST_SZ_DW(qetc_reg)];
+	void *ets_tcn_conf;
+	int err;
+	int i;
+
+	err = mlx5_query_port_qetcr_reg(mdev, out, sizeof(out));
+	if (err)
+		return err;
+
+	for (i = 0; i <= mlx5_max_tc(mdev); i++) {
+		ets_tcn_conf = MLX5_ADDR_OF(qetc_reg, out, tc_configuration[i]);
+
+		max_bw_value[i] = MLX5_GET(ets_tcn_config_reg, ets_tcn_conf,
+					   max_bw_value);
+		max_bw_units[i] = MLX5_GET(ets_tcn_config_reg, ets_tcn_conf,
+					   max_bw_units);
+	}
+
+	return 0;
+}
+EXPORT_SYMBOL_GPL(mlx5_query_port_ets_rate_limit);
+
+int mlx5_set_port_wol(struct mlx5_core_dev *mdev, u8 wol_mode)
+{
+	u32 in[MLX5_ST_SZ_DW(set_wol_rol_in)]   = {0};
+	u32 out[MLX5_ST_SZ_DW(set_wol_rol_out)] = {0};
+
+	MLX5_SET(set_wol_rol_in, in, opcode, MLX5_CMD_OP_SET_WOL_ROL);
+	MLX5_SET(set_wol_rol_in, in, wol_mode_valid, 1);
+	MLX5_SET(set_wol_rol_in, in, wol_mode, wol_mode);
+	return mlx5_cmd_exec(mdev, in, sizeof(in), out, sizeof(out));
+}
+EXPORT_SYMBOL_GPL(mlx5_set_port_wol);
+
+int mlx5_query_port_wol(struct mlx5_core_dev *mdev, u8 *wol_mode)
+{
+	u32 in[MLX5_ST_SZ_DW(query_wol_rol_in)]   = {0};
+	u32 out[MLX5_ST_SZ_DW(query_wol_rol_out)] = {0};
+	int err;
+
+	MLX5_SET(query_wol_rol_in, in, opcode, MLX5_CMD_OP_QUERY_WOL_ROL);
+	err = mlx5_cmd_exec(mdev, in, sizeof(in), out, sizeof(out));
+	if (!err)
+		*wol_mode = MLX5_GET(query_wol_rol_out, out, wol_mode);
+
+	return err;
+}
+EXPORT_SYMBOL_GPL(mlx5_query_port_wol);
+
+static int mlx5_query_ports_check(struct mlx5_core_dev *mdev, u32 *out,
+				  int outlen)
+{
+	u32 in[MLX5_ST_SZ_DW(pcmr_reg)] = {0};
+
+	MLX5_SET(pcmr_reg, in, local_port, 1);
+	return mlx5_core_access_reg(mdev, in, sizeof(in), out,
+				    outlen, MLX5_REG_PCMR, 0, 0);
+}
+
+static int mlx5_set_ports_check(struct mlx5_core_dev *mdev, u32 *in, int inlen)
+{
+	u32 out[MLX5_ST_SZ_DW(pcmr_reg)];
+
+	return mlx5_core_access_reg(mdev, in, inlen, out,
+				    sizeof(out), MLX5_REG_PCMR, 0, 1);
+}
+
+int mlx5_set_port_fcs(struct mlx5_core_dev *mdev, u8 enable)
+{
+	u32 in[MLX5_ST_SZ_DW(pcmr_reg)] = {0};
+
+	MLX5_SET(pcmr_reg, in, local_port, 1);
+	MLX5_SET(pcmr_reg, in, fcs_chk, enable);
+	return mlx5_set_ports_check(mdev, in, sizeof(in));
+}
+
+void mlx5_query_port_fcs(struct mlx5_core_dev *mdev, bool *supported,
+			 bool *enabled)
+{
+	u32 out[MLX5_ST_SZ_DW(pcmr_reg)];
+	/* Default values for FW which do not support MLX5_REG_PCMR */
+	*supported = false;
+	*enabled = true;
+
+	if (!MLX5_CAP_GEN(mdev, ports_check))
+		return;
+
+	if (mlx5_query_ports_check(mdev, out, sizeof(out)))
+		return;
+
+	*supported = !!(MLX5_GET(pcmr_reg, out, fcs_cap));
+	*enabled = !!(MLX5_GET(pcmr_reg, out, fcs_chk));
+}
+
+static const char *mlx5_pme_status[MLX5_MODULE_STATUS_NUM] = {
+	"Cable plugged",   /* MLX5_MODULE_STATUS_PLUGGED    = 0x1 */
+	"Cable unplugged", /* MLX5_MODULE_STATUS_UNPLUGGED  = 0x2 */
+	"Cable error",     /* MLX5_MODULE_STATUS_ERROR      = 0x3 */
+};
+
+static const char *mlx5_pme_error[MLX5_MODULE_EVENT_ERROR_NUM] = {
+	"Power budget exceeded",
+	"Long Range for non MLNX cable",
+	"Bus stuck(I2C or data shorted)",
+	"No EEPROM/retry timeout",
+	"Enforce part number list",
+	"Unknown identifier",
+	"High Temperature",
+	"Bad or shorted cable/module",
+	"Unknown status",
+};
+
+void mlx5_port_module_event(struct mlx5_core_dev *dev, struct mlx5_eqe *eqe)
+{
+	enum port_module_event_status_type module_status;
+	enum port_module_event_error_type error_type;
+	struct mlx5_eqe_port_module *module_event_eqe;
+	struct mlx5_priv *priv = &dev->priv;
+	u8 module_num;
+
+	module_event_eqe = &eqe->data.port_module;
+	module_num = module_event_eqe->module;
+	module_status = module_event_eqe->module_status &
+			PORT_MODULE_EVENT_MODULE_STATUS_MASK;
+	error_type = module_event_eqe->error_type &
+		     PORT_MODULE_EVENT_ERROR_TYPE_MASK;
+
+	if (module_status < MLX5_MODULE_STATUS_ERROR) {
+		priv->pme_stats.status_counters[module_status - 1]++;
+	} else if (module_status == MLX5_MODULE_STATUS_ERROR) {
+		if (error_type >= MLX5_MODULE_EVENT_ERROR_UNKNOWN)
+			/* Unknown error type */
+			error_type = MLX5_MODULE_EVENT_ERROR_UNKNOWN;
+		priv->pme_stats.error_counters[error_type]++;
+	}
+
+	if (!printk_ratelimit())
+		return;
+
+	if (module_status < MLX5_MODULE_STATUS_ERROR)
+		mlx5_core_info(dev,
+			       "Port module event: module %u, %s\n",
+			       module_num, mlx5_pme_status[module_status - 1]);
+
+	else if (module_status == MLX5_MODULE_STATUS_ERROR)
+		mlx5_core_info(dev,
+			       "Port module event[error]: module %u, %s, %s\n",
+			       module_num, mlx5_pme_status[module_status - 1],
+			       mlx5_pme_error[error_type]);
+}
+
+int mlx5_query_mtpps(struct mlx5_core_dev *mdev, u32 *mtpps, u32 mtpps_size)
+{
+	u32 in[MLX5_ST_SZ_DW(mtpps_reg)] = {0};
+
+	return mlx5_core_access_reg(mdev, in, sizeof(in), mtpps,
+				    mtpps_size, MLX5_REG_MTPPS, 0, 0);
+}
+
+int mlx5_set_mtpps(struct mlx5_core_dev *mdev, u32 *mtpps, u32 mtpps_size)
+{
+	u32 out[MLX5_ST_SZ_DW(mtpps_reg)] = {0};
+
+	return mlx5_core_access_reg(mdev, mtpps, mtpps_size, out,
+				    sizeof(out), MLX5_REG_MTPPS, 0, 1);
+}
+
+int mlx5_query_mtppse(struct mlx5_core_dev *mdev, u8 pin, u8 *arm, u8 *mode)
+{
+	u32 out[MLX5_ST_SZ_DW(mtppse_reg)] = {0};
+	u32 in[MLX5_ST_SZ_DW(mtppse_reg)] = {0};
+	int err = 0;
+
+	MLX5_SET(mtppse_reg, in, pin, pin);
+
+	err = mlx5_core_access_reg(mdev, in, sizeof(in), out,
+				   sizeof(out), MLX5_REG_MTPPSE, 0, 0);
+	if (err)
+		return err;
+
+	*arm = MLX5_GET(mtppse_reg, in, event_arm);
+	*mode = MLX5_GET(mtppse_reg, in, event_generation_mode);
+
+	return err;
+}
+
+int mlx5_set_mtppse(struct mlx5_core_dev *mdev, u8 pin, u8 arm, u8 mode)
+{
+	u32 out[MLX5_ST_SZ_DW(mtppse_reg)] = {0};
+	u32 in[MLX5_ST_SZ_DW(mtppse_reg)] = {0};
+
+	MLX5_SET(mtppse_reg, in, pin, pin);
+	MLX5_SET(mtppse_reg, in, event_arm, arm);
+	MLX5_SET(mtppse_reg, in, event_generation_mode, mode);
+
+	return mlx5_core_access_reg(mdev, in, sizeof(in), out,
+				    sizeof(out), MLX5_REG_MTPPSE, 0, 1);
+}
+
+int mlx5_set_trust_state(struct mlx5_core_dev *mdev, u8 trust_state)
+{
+	u32 out[MLX5_ST_SZ_DW(qpts_reg)] = {};
+	u32 in[MLX5_ST_SZ_DW(qpts_reg)] = {};
+	int err;
+
+	MLX5_SET(qpts_reg, in, local_port, 1);
+	MLX5_SET(qpts_reg, in, trust_state, trust_state);
+
+	err = mlx5_core_access_reg(mdev, in, sizeof(in), out,
+				   sizeof(out), MLX5_REG_QPTS, 0, 1);
+	return err;
+}
+
+int mlx5_query_trust_state(struct mlx5_core_dev *mdev, u8 *trust_state)
+{
+	u32 out[MLX5_ST_SZ_DW(qpts_reg)] = {};
+	u32 in[MLX5_ST_SZ_DW(qpts_reg)] = {};
+	int err;
+
+	MLX5_SET(qpts_reg, in, local_port, 1);
+
+	err = mlx5_core_access_reg(mdev, in, sizeof(in), out,
+				   sizeof(out), MLX5_REG_QPTS, 0, 0);
+	if (!err)
+		*trust_state = MLX5_GET(qpts_reg, out, trust_state);
+
+	return err;
+}
+
+int mlx5_set_dscp2prio(struct mlx5_core_dev *mdev, u8 dscp, u8 prio)
+{
+	int sz = MLX5_ST_SZ_BYTES(qpdpm_reg);
+	void *qpdpm_dscp;
+	void *out;
+	void *in;
+	int err;
+
+	in = kzalloc(sz, GFP_KERNEL);
+	out = kzalloc(sz, GFP_KERNEL);
+	if (!in || !out) {
+		err = -ENOMEM;
+		goto out;
+	}
+
+	MLX5_SET(qpdpm_reg, in, local_port, 1);
+	err = mlx5_core_access_reg(mdev, in, sz, out, sz, MLX5_REG_QPDPM, 0, 0);
+	if (err)
+		goto out;
+
+	memcpy(in, out, sz);
+	MLX5_SET(qpdpm_reg, in, local_port, 1);
+
+	/* Update the corresponding dscp entry */
+	qpdpm_dscp = MLX5_ADDR_OF(qpdpm_reg, in, dscp[dscp]);
+	MLX5_SET16(qpdpm_dscp_reg, qpdpm_dscp, prio, prio);
+	MLX5_SET16(qpdpm_dscp_reg, qpdpm_dscp, e, 1);
+	err = mlx5_core_access_reg(mdev, in, sz, out, sz, MLX5_REG_QPDPM, 0, 1);
+
+out:
+	kfree(in);
+	kfree(out);
+	return err;
+}
+
+/* dscp2prio[i]: priority that dscp i mapped to */
+#define MLX5E_SUPPORTED_DSCP 64
+int mlx5_query_dscp2prio(struct mlx5_core_dev *mdev, u8 *dscp2prio)
+{
+	int sz = MLX5_ST_SZ_BYTES(qpdpm_reg);
+	void *qpdpm_dscp;
+	void *out;
+	void *in;
+	int err;
+	int i;
+
+	in = kzalloc(sz, GFP_KERNEL);
+	out = kzalloc(sz, GFP_KERNEL);
+	if (!in || !out) {
+		err = -ENOMEM;
+		goto out;
+	}
+
+	MLX5_SET(qpdpm_reg, in, local_port, 1);
+	err = mlx5_core_access_reg(mdev, in, sz, out, sz, MLX5_REG_QPDPM, 0, 0);
+	if (err)
+		goto out;
+
+	for (i = 0; i < (MLX5E_SUPPORTED_DSCP); i++) {
+		qpdpm_dscp = MLX5_ADDR_OF(qpdpm_reg, out, dscp[i]);
+		dscp2prio[i] = MLX5_GET16(qpdpm_dscp_reg, qpdpm_dscp, prio);
+	}
+
+out:
+	kfree(in);
+	kfree(out);
+	return err;
+}
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/qp.c b/drivers/net/ethernet/mellanox/mlx5/core/qp.c
new file mode 100644
index 000000000..479ac21cd
--- /dev/null
+++ b/drivers/net/ethernet/mellanox/mlx5/core/qp.c
@@ -0,0 +1,636 @@
+/*
+ * Copyright (c) 2013-2015, 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/gfp.h>
+#include <linux/export.h>
+#include <linux/mlx5/cmd.h>
+#include <linux/mlx5/qp.h>
+#include <linux/mlx5/driver.h>
+#include <linux/mlx5/transobj.h>
+
+#include "mlx5_core.h"
+
+static struct mlx5_core_rsc_common *mlx5_get_rsc(struct mlx5_core_dev *dev,
+						 u32 rsn)
+{
+	struct mlx5_qp_table *table = &dev->priv.qp_table;
+	struct mlx5_core_rsc_common *common;
+	unsigned long flags;
+
+	spin_lock_irqsave(&table->lock, flags);
+
+	common = radix_tree_lookup(&table->tree, rsn);
+	if (common)
+		atomic_inc(&common->refcount);
+
+	spin_unlock_irqrestore(&table->lock, flags);
+
+	if (!common) {
+		mlx5_core_warn(dev, "Async event for bogus resource 0x%x\n",
+			       rsn);
+		return NULL;
+	}
+	return common;
+}
+
+void mlx5_core_put_rsc(struct mlx5_core_rsc_common *common)
+{
+	if (atomic_dec_and_test(&common->refcount))
+		complete(&common->free);
+}
+
+static u64 qp_allowed_event_types(void)
+{
+	u64 mask;
+
+	mask = BIT(MLX5_EVENT_TYPE_PATH_MIG) |
+	       BIT(MLX5_EVENT_TYPE_COMM_EST) |
+	       BIT(MLX5_EVENT_TYPE_SQ_DRAINED) |
+	       BIT(MLX5_EVENT_TYPE_SRQ_LAST_WQE) |
+	       BIT(MLX5_EVENT_TYPE_WQ_CATAS_ERROR) |
+	       BIT(MLX5_EVENT_TYPE_PATH_MIG_FAILED) |
+	       BIT(MLX5_EVENT_TYPE_WQ_INVAL_REQ_ERROR) |
+	       BIT(MLX5_EVENT_TYPE_WQ_ACCESS_ERROR);
+
+	return mask;
+}
+
+static u64 rq_allowed_event_types(void)
+{
+	u64 mask;
+
+	mask = BIT(MLX5_EVENT_TYPE_SRQ_LAST_WQE) |
+	       BIT(MLX5_EVENT_TYPE_WQ_CATAS_ERROR);
+
+	return mask;
+}
+
+static u64 sq_allowed_event_types(void)
+{
+	return BIT(MLX5_EVENT_TYPE_WQ_CATAS_ERROR);
+}
+
+static u64 dct_allowed_event_types(void)
+{
+	return BIT(MLX5_EVENT_TYPE_DCT_DRAINED);
+}
+
+static bool is_event_type_allowed(int rsc_type, int event_type)
+{
+	switch (rsc_type) {
+	case MLX5_EVENT_QUEUE_TYPE_QP:
+		return BIT(event_type) & qp_allowed_event_types();
+	case MLX5_EVENT_QUEUE_TYPE_RQ:
+		return BIT(event_type) & rq_allowed_event_types();
+	case MLX5_EVENT_QUEUE_TYPE_SQ:
+		return BIT(event_type) & sq_allowed_event_types();
+	case MLX5_EVENT_QUEUE_TYPE_DCT:
+		return BIT(event_type) & dct_allowed_event_types();
+	default:
+		WARN(1, "Event arrived for unknown resource type");
+		return false;
+	}
+}
+
+void mlx5_rsc_event(struct mlx5_core_dev *dev, u32 rsn, int event_type)
+{
+	struct mlx5_core_rsc_common *common = mlx5_get_rsc(dev, rsn);
+	struct mlx5_core_dct *dct;
+	struct mlx5_core_qp *qp;
+
+	if (!common)
+		return;
+
+	if (!is_event_type_allowed((rsn >> MLX5_USER_INDEX_LEN), event_type)) {
+		mlx5_core_warn(dev, "event 0x%.2x is not allowed on resource 0x%.8x\n",
+			       event_type, rsn);
+		goto out;
+	}
+
+	switch (common->res) {
+	case MLX5_RES_QP:
+	case MLX5_RES_RQ:
+	case MLX5_RES_SQ:
+		qp = (struct mlx5_core_qp *)common;
+		qp->event(qp, event_type);
+		break;
+	case MLX5_RES_DCT:
+		dct = (struct mlx5_core_dct *)common;
+		if (event_type == MLX5_EVENT_TYPE_DCT_DRAINED)
+			complete(&dct->drained);
+		break;
+	default:
+		mlx5_core_warn(dev, "invalid resource type for 0x%x\n", rsn);
+	}
+out:
+	mlx5_core_put_rsc(common);
+}
+
+static int create_resource_common(struct mlx5_core_dev *dev,
+				  struct mlx5_core_qp *qp,
+				  int rsc_type)
+{
+	struct mlx5_qp_table *table = &dev->priv.qp_table;
+	int err;
+
+	qp->common.res = rsc_type;
+	spin_lock_irq(&table->lock);
+	err = radix_tree_insert(&table->tree,
+				qp->qpn | (rsc_type << MLX5_USER_INDEX_LEN),
+				qp);
+	spin_unlock_irq(&table->lock);
+	if (err)
+		return err;
+
+	atomic_set(&qp->common.refcount, 1);
+	init_completion(&qp->common.free);
+	qp->pid = current->pid;
+
+	return 0;
+}
+
+static void destroy_resource_common(struct mlx5_core_dev *dev,
+				    struct mlx5_core_qp *qp)
+{
+	struct mlx5_qp_table *table = &dev->priv.qp_table;
+	unsigned long flags;
+
+	spin_lock_irqsave(&table->lock, flags);
+	radix_tree_delete(&table->tree,
+			  qp->qpn | (qp->common.res << MLX5_USER_INDEX_LEN));
+	spin_unlock_irqrestore(&table->lock, flags);
+	mlx5_core_put_rsc((struct mlx5_core_rsc_common *)qp);
+	wait_for_completion(&qp->common.free);
+}
+
+int mlx5_core_create_dct(struct mlx5_core_dev *dev,
+			 struct mlx5_core_dct *dct,
+			 u32 *in, int inlen)
+{
+	u32 out[MLX5_ST_SZ_DW(create_dct_out)]   = {0};
+	u32 din[MLX5_ST_SZ_DW(destroy_dct_in)]   = {0};
+	u32 dout[MLX5_ST_SZ_DW(destroy_dct_out)] = {0};
+	struct mlx5_core_qp *qp = &dct->mqp;
+	int err;
+
+	init_completion(&dct->drained);
+	MLX5_SET(create_dct_in, in, opcode, MLX5_CMD_OP_CREATE_DCT);
+
+	err = mlx5_cmd_exec(dev, in, inlen, &out, sizeof(out));
+	if (err) {
+		mlx5_core_warn(dev, "create DCT failed, ret %d\n", err);
+		return err;
+	}
+
+	qp->qpn = MLX5_GET(create_dct_out, out, dctn);
+	err = create_resource_common(dev, qp, MLX5_RES_DCT);
+	if (err)
+		goto err_cmd;
+
+	return 0;
+err_cmd:
+	MLX5_SET(destroy_dct_in, din, opcode, MLX5_CMD_OP_DESTROY_DCT);
+	MLX5_SET(destroy_dct_in, din, dctn, qp->qpn);
+	mlx5_cmd_exec(dev, (void *)&in, sizeof(din),
+		      (void *)&out, sizeof(dout));
+	return err;
+}
+EXPORT_SYMBOL_GPL(mlx5_core_create_dct);
+
+int mlx5_core_create_qp(struct mlx5_core_dev *dev,
+			struct mlx5_core_qp *qp,
+			u32 *in, int inlen)
+{
+	u32 out[MLX5_ST_SZ_DW(create_qp_out)] = {0};
+	u32 dout[MLX5_ST_SZ_DW(destroy_qp_out)];
+	u32 din[MLX5_ST_SZ_DW(destroy_qp_in)];
+	int err;
+
+	MLX5_SET(create_qp_in, in, opcode, MLX5_CMD_OP_CREATE_QP);
+
+	err = mlx5_cmd_exec(dev, in, inlen, out, sizeof(out));
+	if (err)
+		return err;
+
+	qp->qpn = MLX5_GET(create_qp_out, out, qpn);
+	mlx5_core_dbg(dev, "qpn = 0x%x\n", qp->qpn);
+
+	err = create_resource_common(dev, qp, MLX5_RES_QP);
+	if (err)
+		goto err_cmd;
+
+	err = mlx5_debug_qp_add(dev, qp);
+	if (err)
+		mlx5_core_dbg(dev, "failed adding QP 0x%x to debug file system\n",
+			      qp->qpn);
+
+	atomic_inc(&dev->num_qps);
+
+	return 0;
+
+err_cmd:
+	memset(din, 0, sizeof(din));
+	memset(dout, 0, sizeof(dout));
+	MLX5_SET(destroy_qp_in, din, opcode, MLX5_CMD_OP_DESTROY_QP);
+	MLX5_SET(destroy_qp_in, din, qpn, qp->qpn);
+	mlx5_cmd_exec(dev, din, sizeof(din), dout, sizeof(dout));
+	return err;
+}
+EXPORT_SYMBOL_GPL(mlx5_core_create_qp);
+
+static int mlx5_core_drain_dct(struct mlx5_core_dev *dev,
+			       struct mlx5_core_dct *dct)
+{
+	u32 out[MLX5_ST_SZ_DW(drain_dct_out)] = {0};
+	u32 in[MLX5_ST_SZ_DW(drain_dct_in)]   = {0};
+	struct mlx5_core_qp *qp = &dct->mqp;
+
+	MLX5_SET(drain_dct_in, in, opcode, MLX5_CMD_OP_DRAIN_DCT);
+	MLX5_SET(drain_dct_in, in, dctn, qp->qpn);
+	return mlx5_cmd_exec(dev, (void *)&in, sizeof(in),
+			     (void *)&out, sizeof(out));
+}
+
+int mlx5_core_destroy_dct(struct mlx5_core_dev *dev,
+			  struct mlx5_core_dct *dct)
+{
+	u32 out[MLX5_ST_SZ_DW(destroy_dct_out)] = {0};
+	u32 in[MLX5_ST_SZ_DW(destroy_dct_in)]   = {0};
+	struct mlx5_core_qp *qp = &dct->mqp;
+	int err;
+
+	err = mlx5_core_drain_dct(dev, dct);
+	if (err) {
+		if (dev->state == MLX5_DEVICE_STATE_INTERNAL_ERROR) {
+			goto destroy;
+		} else {
+			mlx5_core_warn(dev, "failed drain DCT 0x%x with error 0x%x\n", qp->qpn, err);
+			return err;
+		}
+	}
+	wait_for_completion(&dct->drained);
+destroy:
+	destroy_resource_common(dev, &dct->mqp);
+	MLX5_SET(destroy_dct_in, in, opcode, MLX5_CMD_OP_DESTROY_DCT);
+	MLX5_SET(destroy_dct_in, in, dctn, qp->qpn);
+	err = mlx5_cmd_exec(dev, (void *)&in, sizeof(in),
+			    (void *)&out, sizeof(out));
+	return err;
+}
+EXPORT_SYMBOL_GPL(mlx5_core_destroy_dct);
+
+int mlx5_core_destroy_qp(struct mlx5_core_dev *dev,
+			 struct mlx5_core_qp *qp)
+{
+	u32 out[MLX5_ST_SZ_DW(destroy_qp_out)] = {0};
+	u32 in[MLX5_ST_SZ_DW(destroy_qp_in)]   = {0};
+	int err;
+
+	mlx5_debug_qp_remove(dev, qp);
+
+	destroy_resource_common(dev, qp);
+
+	MLX5_SET(destroy_qp_in, in, opcode, MLX5_CMD_OP_DESTROY_QP);
+	MLX5_SET(destroy_qp_in, in, qpn, qp->qpn);
+	err = mlx5_cmd_exec(dev, in, sizeof(in), out, sizeof(out));
+	if (err)
+		return err;
+
+	atomic_dec(&dev->num_qps);
+	return 0;
+}
+EXPORT_SYMBOL_GPL(mlx5_core_destroy_qp);
+
+int mlx5_core_set_delay_drop(struct mlx5_core_dev *dev,
+			     u32 timeout_usec)
+{
+	u32 out[MLX5_ST_SZ_DW(set_delay_drop_params_out)] = {0};
+	u32 in[MLX5_ST_SZ_DW(set_delay_drop_params_in)]   = {0};
+
+	MLX5_SET(set_delay_drop_params_in, in, opcode,
+		 MLX5_CMD_OP_SET_DELAY_DROP_PARAMS);
+	MLX5_SET(set_delay_drop_params_in, in, delay_drop_timeout,
+		 timeout_usec / 100);
+	return mlx5_cmd_exec(dev, in, sizeof(in), out, sizeof(out));
+}
+EXPORT_SYMBOL_GPL(mlx5_core_set_delay_drop);
+
+struct mbox_info {
+	u32 *in;
+	u32 *out;
+	int inlen;
+	int outlen;
+};
+
+static int mbox_alloc(struct mbox_info *mbox, int inlen, int outlen)
+{
+	mbox->inlen  = inlen;
+	mbox->outlen = outlen;
+	mbox->in = kzalloc(mbox->inlen, GFP_KERNEL);
+	mbox->out = kzalloc(mbox->outlen, GFP_KERNEL);
+	if (!mbox->in || !mbox->out) {
+		kfree(mbox->in);
+		kfree(mbox->out);
+		return -ENOMEM;
+	}
+
+	return 0;
+}
+
+static void mbox_free(struct mbox_info *mbox)
+{
+	kfree(mbox->in);
+	kfree(mbox->out);
+}
+
+static int modify_qp_mbox_alloc(struct mlx5_core_dev *dev, u16 opcode, int qpn,
+				u32 opt_param_mask, void *qpc,
+				struct mbox_info *mbox)
+{
+	mbox->out = NULL;
+	mbox->in = NULL;
+
+#define MBOX_ALLOC(mbox, typ)  \
+	mbox_alloc(mbox, MLX5_ST_SZ_BYTES(typ##_in), MLX5_ST_SZ_BYTES(typ##_out))
+
+#define MOD_QP_IN_SET(typ, in, _opcode, _qpn) \
+	MLX5_SET(typ##_in, in, opcode, _opcode); \
+	MLX5_SET(typ##_in, in, qpn, _qpn)
+
+#define MOD_QP_IN_SET_QPC(typ, in, _opcode, _qpn, _opt_p, _qpc) \
+	MOD_QP_IN_SET(typ, in, _opcode, _qpn); \
+	MLX5_SET(typ##_in, in, opt_param_mask, _opt_p); \
+	memcpy(MLX5_ADDR_OF(typ##_in, in, qpc), _qpc, MLX5_ST_SZ_BYTES(qpc))
+
+	switch (opcode) {
+	/* 2RST & 2ERR */
+	case MLX5_CMD_OP_2RST_QP:
+		if (MBOX_ALLOC(mbox, qp_2rst))
+			return -ENOMEM;
+		MOD_QP_IN_SET(qp_2rst, mbox->in, opcode, qpn);
+		break;
+	case MLX5_CMD_OP_2ERR_QP:
+		if (MBOX_ALLOC(mbox, qp_2err))
+			return -ENOMEM;
+		MOD_QP_IN_SET(qp_2err, mbox->in, opcode, qpn);
+		break;
+
+	/* MODIFY with QPC */
+	case MLX5_CMD_OP_RST2INIT_QP:
+		if (MBOX_ALLOC(mbox, rst2init_qp))
+			return -ENOMEM;
+		MOD_QP_IN_SET_QPC(rst2init_qp, mbox->in, opcode, qpn,
+				  opt_param_mask, qpc);
+		break;
+	case MLX5_CMD_OP_INIT2RTR_QP:
+		if (MBOX_ALLOC(mbox, init2rtr_qp))
+			return -ENOMEM;
+		MOD_QP_IN_SET_QPC(init2rtr_qp, mbox->in, opcode, qpn,
+				  opt_param_mask, qpc);
+		break;
+	case MLX5_CMD_OP_RTR2RTS_QP:
+		if (MBOX_ALLOC(mbox, rtr2rts_qp))
+			return -ENOMEM;
+		MOD_QP_IN_SET_QPC(rtr2rts_qp, mbox->in, opcode, qpn,
+				  opt_param_mask, qpc);
+		break;
+	case MLX5_CMD_OP_RTS2RTS_QP:
+		if (MBOX_ALLOC(mbox, rts2rts_qp))
+			return -ENOMEM;
+		MOD_QP_IN_SET_QPC(rts2rts_qp, mbox->in, opcode, qpn,
+				  opt_param_mask, qpc);
+		break;
+	case MLX5_CMD_OP_SQERR2RTS_QP:
+		if (MBOX_ALLOC(mbox, sqerr2rts_qp))
+			return -ENOMEM;
+		MOD_QP_IN_SET_QPC(sqerr2rts_qp, mbox->in, opcode, qpn,
+				  opt_param_mask, qpc);
+		break;
+	case MLX5_CMD_OP_INIT2INIT_QP:
+		if (MBOX_ALLOC(mbox, init2init_qp))
+			return -ENOMEM;
+		MOD_QP_IN_SET_QPC(init2init_qp, mbox->in, opcode, qpn,
+				  opt_param_mask, qpc);
+		break;
+	default:
+		mlx5_core_err(dev, "Unknown transition for modify QP: OP(0x%x) QPN(0x%x)\n",
+			      opcode, qpn);
+		return -EINVAL;
+	}
+	return 0;
+}
+
+int mlx5_core_qp_modify(struct mlx5_core_dev *dev, u16 opcode,
+			u32 opt_param_mask, void *qpc,
+			struct mlx5_core_qp *qp)
+{
+	struct mbox_info mbox;
+	int err;
+
+	err = modify_qp_mbox_alloc(dev, opcode, qp->qpn,
+				   opt_param_mask, qpc, &mbox);
+	if (err)
+		return err;
+
+	err = mlx5_cmd_exec(dev, mbox.in, mbox.inlen, mbox.out, mbox.outlen);
+	mbox_free(&mbox);
+	return err;
+}
+EXPORT_SYMBOL_GPL(mlx5_core_qp_modify);
+
+void mlx5_init_qp_table(struct mlx5_core_dev *dev)
+{
+	struct mlx5_qp_table *table = &dev->priv.qp_table;
+
+	memset(table, 0, sizeof(*table));
+	spin_lock_init(&table->lock);
+	INIT_RADIX_TREE(&table->tree, GFP_ATOMIC);
+	mlx5_qp_debugfs_init(dev);
+}
+
+void mlx5_cleanup_qp_table(struct mlx5_core_dev *dev)
+{
+	mlx5_qp_debugfs_cleanup(dev);
+}
+
+int mlx5_core_qp_query(struct mlx5_core_dev *dev, struct mlx5_core_qp *qp,
+		       u32 *out, int outlen)
+{
+	u32 in[MLX5_ST_SZ_DW(query_qp_in)] = {0};
+
+	MLX5_SET(query_qp_in, in, opcode, MLX5_CMD_OP_QUERY_QP);
+	MLX5_SET(query_qp_in, in, qpn, qp->qpn);
+	return mlx5_cmd_exec(dev, in, sizeof(in), out, outlen);
+}
+EXPORT_SYMBOL_GPL(mlx5_core_qp_query);
+
+int mlx5_core_dct_query(struct mlx5_core_dev *dev, struct mlx5_core_dct *dct,
+			u32 *out, int outlen)
+{
+	u32 in[MLX5_ST_SZ_DW(query_dct_in)] = {0};
+	struct mlx5_core_qp *qp = &dct->mqp;
+
+	MLX5_SET(query_dct_in, in, opcode, MLX5_CMD_OP_QUERY_DCT);
+	MLX5_SET(query_dct_in, in, dctn, qp->qpn);
+
+	return mlx5_cmd_exec(dev, (void *)&in, sizeof(in),
+			     (void *)out, outlen);
+}
+EXPORT_SYMBOL_GPL(mlx5_core_dct_query);
+
+int mlx5_core_xrcd_alloc(struct mlx5_core_dev *dev, u32 *xrcdn)
+{
+	u32 out[MLX5_ST_SZ_DW(alloc_xrcd_out)] = {0};
+	u32 in[MLX5_ST_SZ_DW(alloc_xrcd_in)]   = {0};
+	int err;
+
+	MLX5_SET(alloc_xrcd_in, in, opcode, MLX5_CMD_OP_ALLOC_XRCD);
+	err = mlx5_cmd_exec(dev, in, sizeof(in), out, sizeof(out));
+	if (!err)
+		*xrcdn = MLX5_GET(alloc_xrcd_out, out, xrcd);
+	return err;
+}
+EXPORT_SYMBOL_GPL(mlx5_core_xrcd_alloc);
+
+int mlx5_core_xrcd_dealloc(struct mlx5_core_dev *dev, u32 xrcdn)
+{
+	u32 out[MLX5_ST_SZ_DW(dealloc_xrcd_out)] = {0};
+	u32 in[MLX5_ST_SZ_DW(dealloc_xrcd_in)]   = {0};
+
+	MLX5_SET(dealloc_xrcd_in, in, opcode, MLX5_CMD_OP_DEALLOC_XRCD);
+	MLX5_SET(dealloc_xrcd_in, in, xrcd, xrcdn);
+	return mlx5_cmd_exec(dev, in, sizeof(in), out, sizeof(out));
+}
+EXPORT_SYMBOL_GPL(mlx5_core_xrcd_dealloc);
+
+int mlx5_core_create_rq_tracked(struct mlx5_core_dev *dev, u32 *in, int inlen,
+				struct mlx5_core_qp *rq)
+{
+	int err;
+	u32 rqn;
+
+	err = mlx5_core_create_rq(dev, in, inlen, &rqn);
+	if (err)
+		return err;
+
+	rq->qpn = rqn;
+	err = create_resource_common(dev, rq, MLX5_RES_RQ);
+	if (err)
+		goto err_destroy_rq;
+
+	return 0;
+
+err_destroy_rq:
+	mlx5_core_destroy_rq(dev, rq->qpn);
+
+	return err;
+}
+EXPORT_SYMBOL(mlx5_core_create_rq_tracked);
+
+void mlx5_core_destroy_rq_tracked(struct mlx5_core_dev *dev,
+				  struct mlx5_core_qp *rq)
+{
+	destroy_resource_common(dev, rq);
+	mlx5_core_destroy_rq(dev, rq->qpn);
+}
+EXPORT_SYMBOL(mlx5_core_destroy_rq_tracked);
+
+int mlx5_core_create_sq_tracked(struct mlx5_core_dev *dev, u32 *in, int inlen,
+				struct mlx5_core_qp *sq)
+{
+	int err;
+	u32 sqn;
+
+	err = mlx5_core_create_sq(dev, in, inlen, &sqn);
+	if (err)
+		return err;
+
+	sq->qpn = sqn;
+	err = create_resource_common(dev, sq, MLX5_RES_SQ);
+	if (err)
+		goto err_destroy_sq;
+
+	return 0;
+
+err_destroy_sq:
+	mlx5_core_destroy_sq(dev, sq->qpn);
+
+	return err;
+}
+EXPORT_SYMBOL(mlx5_core_create_sq_tracked);
+
+void mlx5_core_destroy_sq_tracked(struct mlx5_core_dev *dev,
+				  struct mlx5_core_qp *sq)
+{
+	destroy_resource_common(dev, sq);
+	mlx5_core_destroy_sq(dev, sq->qpn);
+}
+EXPORT_SYMBOL(mlx5_core_destroy_sq_tracked);
+
+int mlx5_core_alloc_q_counter(struct mlx5_core_dev *dev, u16 *counter_id)
+{
+	u32 in[MLX5_ST_SZ_DW(alloc_q_counter_in)]   = {0};
+	u32 out[MLX5_ST_SZ_DW(alloc_q_counter_out)] = {0};
+	int err;
+
+	MLX5_SET(alloc_q_counter_in, in, opcode, MLX5_CMD_OP_ALLOC_Q_COUNTER);
+	err = mlx5_cmd_exec(dev, in, sizeof(in), out, sizeof(out));
+	if (!err)
+		*counter_id = MLX5_GET(alloc_q_counter_out, out,
+				       counter_set_id);
+	return err;
+}
+EXPORT_SYMBOL_GPL(mlx5_core_alloc_q_counter);
+
+int mlx5_core_dealloc_q_counter(struct mlx5_core_dev *dev, u16 counter_id)
+{
+	u32 in[MLX5_ST_SZ_DW(dealloc_q_counter_in)]   = {0};
+	u32 out[MLX5_ST_SZ_DW(dealloc_q_counter_out)] = {0};
+
+	MLX5_SET(dealloc_q_counter_in, in, opcode,
+		 MLX5_CMD_OP_DEALLOC_Q_COUNTER);
+	MLX5_SET(dealloc_q_counter_in, in, counter_set_id, counter_id);
+	return mlx5_cmd_exec(dev, in, sizeof(in), out, sizeof(out));
+}
+EXPORT_SYMBOL_GPL(mlx5_core_dealloc_q_counter);
+
+int mlx5_core_query_q_counter(struct mlx5_core_dev *dev, u16 counter_id,
+			      int reset, void *out, int out_size)
+{
+	u32 in[MLX5_ST_SZ_DW(query_q_counter_in)] = {0};
+
+	MLX5_SET(query_q_counter_in, in, opcode, MLX5_CMD_OP_QUERY_Q_COUNTER);
+	MLX5_SET(query_q_counter_in, in, clear, reset);
+	MLX5_SET(query_q_counter_in, in, counter_set_id, counter_id);
+	return mlx5_cmd_exec(dev, in, sizeof(in), out, out_size);
+}
+EXPORT_SYMBOL_GPL(mlx5_core_query_q_counter);
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/rl.c b/drivers/net/ethernet/mellanox/mlx5/core/rl.c
new file mode 100644
index 000000000..bc86dffdc
--- /dev/null
+++ b/drivers/net/ethernet/mellanox/mlx5/core/rl.c
@@ -0,0 +1,288 @@
+/*
+ * Copyright (c) 2013-2016, 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/module.h>
+#include <linux/mlx5/driver.h>
+#include <linux/mlx5/cmd.h>
+#include "mlx5_core.h"
+
+/* Scheduling element fw management */
+int mlx5_create_scheduling_element_cmd(struct mlx5_core_dev *dev, u8 hierarchy,
+				       void *ctx, u32 *element_id)
+{
+	u32 in[MLX5_ST_SZ_DW(create_scheduling_element_in)]  = {0};
+	u32 out[MLX5_ST_SZ_DW(create_scheduling_element_in)] = {0};
+	void *schedc;
+	int err;
+
+	schedc = MLX5_ADDR_OF(create_scheduling_element_in, in,
+			      scheduling_context);
+	MLX5_SET(create_scheduling_element_in, in, opcode,
+		 MLX5_CMD_OP_CREATE_SCHEDULING_ELEMENT);
+	MLX5_SET(create_scheduling_element_in, in, scheduling_hierarchy,
+		 hierarchy);
+	memcpy(schedc, ctx, MLX5_ST_SZ_BYTES(scheduling_context));
+
+	err = mlx5_cmd_exec(dev, in, sizeof(in), out, sizeof(out));
+	if (err)
+		return err;
+
+	*element_id = MLX5_GET(create_scheduling_element_out, out,
+			       scheduling_element_id);
+	return 0;
+}
+
+int mlx5_modify_scheduling_element_cmd(struct mlx5_core_dev *dev, u8 hierarchy,
+				       void *ctx, u32 element_id,
+				       u32 modify_bitmask)
+{
+	u32 in[MLX5_ST_SZ_DW(modify_scheduling_element_in)]  = {0};
+	u32 out[MLX5_ST_SZ_DW(modify_scheduling_element_in)] = {0};
+	void *schedc;
+
+	schedc = MLX5_ADDR_OF(modify_scheduling_element_in, in,
+			      scheduling_context);
+	MLX5_SET(modify_scheduling_element_in, in, opcode,
+		 MLX5_CMD_OP_MODIFY_SCHEDULING_ELEMENT);
+	MLX5_SET(modify_scheduling_element_in, in, scheduling_element_id,
+		 element_id);
+	MLX5_SET(modify_scheduling_element_in, in, modify_bitmask,
+		 modify_bitmask);
+	MLX5_SET(modify_scheduling_element_in, in, scheduling_hierarchy,
+		 hierarchy);
+	memcpy(schedc, ctx, MLX5_ST_SZ_BYTES(scheduling_context));
+
+	return mlx5_cmd_exec(dev, in, sizeof(in), out, sizeof(out));
+}
+
+int mlx5_destroy_scheduling_element_cmd(struct mlx5_core_dev *dev, u8 hierarchy,
+					u32 element_id)
+{
+	u32 in[MLX5_ST_SZ_DW(destroy_scheduling_element_in)]  = {0};
+	u32 out[MLX5_ST_SZ_DW(destroy_scheduling_element_in)] = {0};
+
+	MLX5_SET(destroy_scheduling_element_in, in, opcode,
+		 MLX5_CMD_OP_DESTROY_SCHEDULING_ELEMENT);
+	MLX5_SET(destroy_scheduling_element_in, in, scheduling_element_id,
+		 element_id);
+	MLX5_SET(destroy_scheduling_element_in, in, scheduling_hierarchy,
+		 hierarchy);
+
+	return mlx5_cmd_exec(dev, in, sizeof(in), out, sizeof(out));
+}
+
+/* Finds an entry where we can register the given rate
+ * If the rate already exists, return the entry where it is registered,
+ * otherwise return the first available entry.
+ * If the table is full, return NULL
+ */
+static struct mlx5_rl_entry *find_rl_entry(struct mlx5_rl_table *table,
+					   struct mlx5_rate_limit *rl)
+{
+	struct mlx5_rl_entry *ret_entry = NULL;
+	bool empty_found = false;
+	int i;
+
+	for (i = 0; i < table->max_size; i++) {
+		if (mlx5_rl_are_equal(&table->rl_entry[i].rl, rl))
+			return &table->rl_entry[i];
+		if (!empty_found && !table->rl_entry[i].rl.rate) {
+			empty_found = true;
+			ret_entry = &table->rl_entry[i];
+		}
+	}
+
+	return ret_entry;
+}
+
+static int mlx5_set_pp_rate_limit_cmd(struct mlx5_core_dev *dev,
+				      u16 index,
+				      struct mlx5_rate_limit *rl)
+{
+	u32 in[MLX5_ST_SZ_DW(set_pp_rate_limit_in)]   = {0};
+	u32 out[MLX5_ST_SZ_DW(set_pp_rate_limit_out)] = {0};
+
+	MLX5_SET(set_pp_rate_limit_in, in, opcode,
+		 MLX5_CMD_OP_SET_PP_RATE_LIMIT);
+	MLX5_SET(set_pp_rate_limit_in, in, rate_limit_index, index);
+	MLX5_SET(set_pp_rate_limit_in, in, rate_limit, rl->rate);
+	MLX5_SET(set_pp_rate_limit_in, in, burst_upper_bound, rl->max_burst_sz);
+	MLX5_SET(set_pp_rate_limit_in, in, typical_packet_size, rl->typical_pkt_sz);
+	return mlx5_cmd_exec(dev, in, sizeof(in), out, sizeof(out));
+}
+
+bool mlx5_rl_is_in_range(struct mlx5_core_dev *dev, u32 rate)
+{
+	struct mlx5_rl_table *table = &dev->priv.rl_table;
+
+	return (rate <= table->max_rate && rate >= table->min_rate);
+}
+EXPORT_SYMBOL(mlx5_rl_is_in_range);
+
+bool mlx5_rl_are_equal(struct mlx5_rate_limit *rl_0,
+		       struct mlx5_rate_limit *rl_1)
+{
+	return ((rl_0->rate == rl_1->rate) &&
+		(rl_0->max_burst_sz == rl_1->max_burst_sz) &&
+		(rl_0->typical_pkt_sz == rl_1->typical_pkt_sz));
+}
+EXPORT_SYMBOL(mlx5_rl_are_equal);
+
+int mlx5_rl_add_rate(struct mlx5_core_dev *dev, u16 *index,
+		     struct mlx5_rate_limit *rl)
+{
+	struct mlx5_rl_table *table = &dev->priv.rl_table;
+	struct mlx5_rl_entry *entry;
+	int err = 0;
+
+	mutex_lock(&table->rl_lock);
+
+	if (!rl->rate || !mlx5_rl_is_in_range(dev, rl->rate)) {
+		mlx5_core_err(dev, "Invalid rate: %u, should be %u to %u\n",
+			      rl->rate, table->min_rate, table->max_rate);
+		err = -EINVAL;
+		goto out;
+	}
+
+	entry = find_rl_entry(table, rl);
+	if (!entry) {
+		mlx5_core_err(dev, "Max number of %u rates reached\n",
+			      table->max_size);
+		err = -ENOSPC;
+		goto out;
+	}
+	if (entry->refcount) {
+		/* rate already configured */
+		entry->refcount++;
+	} else {
+		/* new rate limit */
+		err = mlx5_set_pp_rate_limit_cmd(dev, entry->index, rl);
+		if (err) {
+			mlx5_core_err(dev, "Failed configuring rate limit(err %d): \
+				      rate %u, max_burst_sz %u, typical_pkt_sz %u\n",
+				      err, rl->rate, rl->max_burst_sz,
+				      rl->typical_pkt_sz);
+			goto out;
+		}
+		entry->rl = *rl;
+		entry->refcount = 1;
+	}
+	*index = entry->index;
+
+out:
+	mutex_unlock(&table->rl_lock);
+	return err;
+}
+EXPORT_SYMBOL(mlx5_rl_add_rate);
+
+void mlx5_rl_remove_rate(struct mlx5_core_dev *dev, struct mlx5_rate_limit *rl)
+{
+	struct mlx5_rl_table *table = &dev->priv.rl_table;
+	struct mlx5_rl_entry *entry = NULL;
+	struct mlx5_rate_limit reset_rl = {0};
+
+	/* 0 is a reserved value for unlimited rate */
+	if (rl->rate == 0)
+		return;
+
+	mutex_lock(&table->rl_lock);
+	entry = find_rl_entry(table, rl);
+	if (!entry || !entry->refcount) {
+		mlx5_core_warn(dev, "Rate %u, max_burst_sz %u typical_pkt_sz %u \
+			       are not configured\n",
+			       rl->rate, rl->max_burst_sz, rl->typical_pkt_sz);
+		goto out;
+	}
+
+	entry->refcount--;
+	if (!entry->refcount) {
+		/* need to remove rate */
+		mlx5_set_pp_rate_limit_cmd(dev, entry->index, &reset_rl);
+		entry->rl = reset_rl;
+	}
+
+out:
+	mutex_unlock(&table->rl_lock);
+}
+EXPORT_SYMBOL(mlx5_rl_remove_rate);
+
+int mlx5_init_rl_table(struct mlx5_core_dev *dev)
+{
+	struct mlx5_rl_table *table = &dev->priv.rl_table;
+	int i;
+
+	mutex_init(&table->rl_lock);
+	if (!MLX5_CAP_GEN(dev, qos) || !MLX5_CAP_QOS(dev, packet_pacing)) {
+		table->max_size = 0;
+		return 0;
+	}
+
+	/* First entry is reserved for unlimited rate */
+	table->max_size = MLX5_CAP_QOS(dev, packet_pacing_rate_table_size) - 1;
+	table->max_rate = MLX5_CAP_QOS(dev, packet_pacing_max_rate);
+	table->min_rate = MLX5_CAP_QOS(dev, packet_pacing_min_rate);
+
+	table->rl_entry = kcalloc(table->max_size, sizeof(struct mlx5_rl_entry),
+				  GFP_KERNEL);
+	if (!table->rl_entry)
+		return -ENOMEM;
+
+	/* The index represents the index in HW rate limit table
+	 * Index 0 is reserved for unlimited rate
+	 */
+	for (i = 0; i < table->max_size; i++)
+		table->rl_entry[i].index = i + 1;
+
+	/* Index 0 is reserved */
+	mlx5_core_info(dev, "Rate limit: %u rates are supported, range: %uMbps to %uMbps\n",
+		       table->max_size,
+		       table->min_rate >> 10,
+		       table->max_rate >> 10);
+
+	return 0;
+}
+
+void mlx5_cleanup_rl_table(struct mlx5_core_dev *dev)
+{
+	struct mlx5_rl_table *table = &dev->priv.rl_table;
+	struct mlx5_rate_limit rl = {0};
+	int i;
+
+	/* Clear all configured rates */
+	for (i = 0; i < table->max_size; i++)
+		if (table->rl_entry[i].rl.rate)
+			mlx5_set_pp_rate_limit_cmd(dev, table->rl_entry[i].index,
+						   &rl);
+
+	kfree(dev->priv.rl_table.rl_entry);
+}
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/sriov.c b/drivers/net/ethernet/mellanox/mlx5/core/sriov.c
new file mode 100644
index 000000000..a0674962f
--- /dev/null
+++ b/drivers/net/ethernet/mellanox/mlx5/core/sriov.c
@@ -0,0 +1,282 @@
+/*
+ * Copyright (c) 2014, Mellanox Technologies 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/pci.h>
+#include <linux/mlx5/driver.h>
+#include <linux/mlx5/vport.h>
+#include "mlx5_core.h"
+#include "eswitch.h"
+
+bool mlx5_sriov_is_enabled(struct mlx5_core_dev *dev)
+{
+	struct mlx5_core_sriov *sriov = &dev->priv.sriov;
+
+	return !!sriov->num_vfs;
+}
+
+static int sriov_restore_guids(struct mlx5_core_dev *dev, int vf)
+{
+	struct mlx5_core_sriov *sriov = &dev->priv.sriov;
+	struct mlx5_hca_vport_context *in;
+	int err = 0;
+
+	/* Restore sriov guid and policy settings */
+	if (sriov->vfs_ctx[vf].node_guid ||
+	    sriov->vfs_ctx[vf].port_guid ||
+	    sriov->vfs_ctx[vf].policy != MLX5_POLICY_INVALID) {
+		in = kzalloc(sizeof(*in), GFP_KERNEL);
+		if (!in)
+			return -ENOMEM;
+
+		in->node_guid = sriov->vfs_ctx[vf].node_guid;
+		in->port_guid = sriov->vfs_ctx[vf].port_guid;
+		in->policy = sriov->vfs_ctx[vf].policy;
+		in->field_select =
+			!!(in->port_guid) * MLX5_HCA_VPORT_SEL_PORT_GUID |
+			!!(in->node_guid) * MLX5_HCA_VPORT_SEL_NODE_GUID |
+			!!(in->policy) * MLX5_HCA_VPORT_SEL_STATE_POLICY;
+
+		err = mlx5_core_modify_hca_vport_context(dev, 1, 1, vf + 1, in);
+		if (err)
+			mlx5_core_warn(dev, "modify vport context failed, unable to restore VF %d settings\n", vf);
+
+		kfree(in);
+	}
+
+	return err;
+}
+
+static int mlx5_device_enable_sriov(struct mlx5_core_dev *dev, int num_vfs)
+{
+	struct mlx5_core_sriov *sriov = &dev->priv.sriov;
+	int err;
+	int vf;
+
+	if (sriov->enabled_vfs) {
+		mlx5_core_warn(dev,
+			       "failed to enable SRIOV on device, already enabled with %d vfs\n",
+			       sriov->enabled_vfs);
+		return -EBUSY;
+	}
+
+	if (!MLX5_ESWITCH_MANAGER(dev))
+		goto enable_vfs_hca;
+
+	err = mlx5_eswitch_enable_sriov(dev->priv.eswitch, num_vfs, SRIOV_LEGACY);
+	if (err) {
+		mlx5_core_warn(dev,
+			       "failed to enable eswitch SRIOV (%d)\n", err);
+		return err;
+	}
+
+enable_vfs_hca:
+	for (vf = 0; vf < num_vfs; vf++) {
+		err = mlx5_core_enable_hca(dev, vf + 1);
+		if (err) {
+			mlx5_core_warn(dev, "failed to enable VF %d (%d)\n", vf, err);
+			continue;
+		}
+		sriov->vfs_ctx[vf].enabled = 1;
+		sriov->enabled_vfs++;
+		if (MLX5_CAP_GEN(dev, port_type) == MLX5_CAP_PORT_TYPE_IB) {
+			err = sriov_restore_guids(dev, vf);
+			if (err) {
+				mlx5_core_warn(dev,
+					       "failed to restore VF %d settings, err %d\n",
+					       vf, err);
+				continue;
+			}
+		}
+		mlx5_core_dbg(dev, "successfully enabled VF* %d\n", vf);
+	}
+
+	return 0;
+}
+
+static void mlx5_device_disable_sriov(struct mlx5_core_dev *dev)
+{
+	struct mlx5_core_sriov *sriov = &dev->priv.sriov;
+	int err;
+	int vf;
+
+	if (!sriov->enabled_vfs)
+		goto out;
+
+	for (vf = 0; vf < sriov->num_vfs; vf++) {
+		if (!sriov->vfs_ctx[vf].enabled)
+			continue;
+		err = mlx5_core_disable_hca(dev, vf + 1);
+		if (err) {
+			mlx5_core_warn(dev, "failed to disable VF %d\n", vf);
+			continue;
+		}
+		sriov->vfs_ctx[vf].enabled = 0;
+		sriov->enabled_vfs--;
+	}
+
+out:
+	if (MLX5_ESWITCH_MANAGER(dev))
+		mlx5_eswitch_disable_sriov(dev->priv.eswitch);
+
+	if (mlx5_wait_for_vf_pages(dev))
+		mlx5_core_warn(dev, "timeout reclaiming VFs pages\n");
+}
+
+static int mlx5_pci_enable_sriov(struct pci_dev *pdev, int num_vfs)
+{
+	struct mlx5_core_dev *dev  = pci_get_drvdata(pdev);
+	int err = 0;
+
+	if (pci_num_vf(pdev)) {
+		mlx5_core_warn(dev, "Unable to enable pci sriov, already enabled\n");
+		return -EBUSY;
+	}
+
+	err = pci_enable_sriov(pdev, num_vfs);
+	if (err)
+		mlx5_core_warn(dev, "pci_enable_sriov failed : %d\n", err);
+
+	return err;
+}
+
+static void mlx5_pci_disable_sriov(struct pci_dev *pdev)
+{
+	pci_disable_sriov(pdev);
+}
+
+static int mlx5_sriov_enable(struct pci_dev *pdev, int num_vfs)
+{
+	struct mlx5_core_dev *dev  = pci_get_drvdata(pdev);
+	struct mlx5_core_sriov *sriov = &dev->priv.sriov;
+	int err = 0;
+
+	err = mlx5_device_enable_sriov(dev, num_vfs);
+	if (err) {
+		mlx5_core_warn(dev, "mlx5_device_enable_sriov failed : %d\n", err);
+		return err;
+	}
+
+	err = mlx5_pci_enable_sriov(pdev, num_vfs);
+	if (err) {
+		mlx5_core_warn(dev, "mlx5_pci_enable_sriov failed : %d\n", err);
+		mlx5_device_disable_sriov(dev);
+		return err;
+	}
+
+	sriov->num_vfs = num_vfs;
+
+	return 0;
+}
+
+static void mlx5_sriov_disable(struct pci_dev *pdev)
+{
+	struct mlx5_core_dev *dev  = pci_get_drvdata(pdev);
+	struct mlx5_core_sriov *sriov = &dev->priv.sriov;
+
+	mlx5_pci_disable_sriov(pdev);
+	mlx5_device_disable_sriov(dev);
+	sriov->num_vfs = 0;
+}
+
+int mlx5_core_sriov_configure(struct pci_dev *pdev, int num_vfs)
+{
+	struct mlx5_core_dev *dev  = pci_get_drvdata(pdev);
+	int err = 0;
+
+	mlx5_core_dbg(dev, "requested num_vfs %d\n", num_vfs);
+	if (!mlx5_core_is_pf(dev))
+		return -EPERM;
+
+	if (num_vfs) {
+		int ret;
+
+		ret = mlx5_lag_forbid(dev);
+		if (ret && (ret != -ENODEV))
+			return ret;
+	}
+
+	if (num_vfs) {
+		err = mlx5_sriov_enable(pdev, num_vfs);
+	} else {
+		mlx5_sriov_disable(pdev);
+		mlx5_lag_allow(dev);
+	}
+
+	return err ? err : num_vfs;
+}
+
+int mlx5_sriov_attach(struct mlx5_core_dev *dev)
+{
+	struct mlx5_core_sriov *sriov = &dev->priv.sriov;
+
+	if (!mlx5_core_is_pf(dev) || !sriov->num_vfs)
+		return 0;
+
+	/* If sriov VFs exist in PCI level, enable them in device level */
+	return mlx5_device_enable_sriov(dev, sriov->num_vfs);
+}
+
+void mlx5_sriov_detach(struct mlx5_core_dev *dev)
+{
+	if (!mlx5_core_is_pf(dev))
+		return;
+
+	mlx5_device_disable_sriov(dev);
+}
+
+int mlx5_sriov_init(struct mlx5_core_dev *dev)
+{
+	struct mlx5_core_sriov *sriov = &dev->priv.sriov;
+	struct pci_dev *pdev = dev->pdev;
+	int total_vfs;
+
+	if (!mlx5_core_is_pf(dev))
+		return 0;
+
+	total_vfs = pci_sriov_get_totalvfs(pdev);
+	sriov->num_vfs = pci_num_vf(pdev);
+	sriov->vfs_ctx = kcalloc(total_vfs, sizeof(*sriov->vfs_ctx), GFP_KERNEL);
+	if (!sriov->vfs_ctx)
+		return -ENOMEM;
+
+	return 0;
+}
+
+void mlx5_sriov_cleanup(struct mlx5_core_dev *dev)
+{
+	struct mlx5_core_sriov *sriov = &dev->priv.sriov;
+
+	if (!mlx5_core_is_pf(dev))
+		return;
+
+	kfree(sriov->vfs_ctx);
+}
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/srq.c b/drivers/net/ethernet/mellanox/mlx5/core/srq.c
new file mode 100644
index 000000000..23cc337a9
--- /dev/null
+++ b/drivers/net/ethernet/mellanox/mlx5/core/srq.c
@@ -0,0 +1,692 @@
+/*
+ * Copyright (c) 2013-2015, 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/module.h>
+#include <linux/mlx5/driver.h>
+#include <linux/mlx5/cmd.h>
+#include <linux/mlx5/srq.h>
+#include <rdma/ib_verbs.h>
+#include "mlx5_core.h"
+#include <linux/mlx5/transobj.h>
+
+void mlx5_srq_event(struct mlx5_core_dev *dev, u32 srqn, int event_type)
+{
+	struct mlx5_srq_table *table = &dev->priv.srq_table;
+	struct mlx5_core_srq *srq;
+
+	spin_lock(&table->lock);
+
+	srq = radix_tree_lookup(&table->tree, srqn);
+	if (srq)
+		atomic_inc(&srq->refcount);
+
+	spin_unlock(&table->lock);
+
+	if (!srq) {
+		mlx5_core_warn(dev, "Async event for bogus SRQ 0x%08x\n", srqn);
+		return;
+	}
+
+	srq->event(srq, event_type);
+
+	if (atomic_dec_and_test(&srq->refcount))
+		complete(&srq->free);
+}
+
+static int get_pas_size(struct mlx5_srq_attr *in)
+{
+	u32 log_page_size = in->log_page_size + 12;
+	u32 log_srq_size  = in->log_size;
+	u32 log_rq_stride = in->wqe_shift;
+	u32 page_offset   = in->page_offset;
+	u32 po_quanta	  = 1 << (log_page_size - 6);
+	u32 rq_sz	  = 1 << (log_srq_size + 4 + log_rq_stride);
+	u32 page_size	  = 1 << log_page_size;
+	u32 rq_sz_po      = rq_sz + (page_offset * po_quanta);
+	u32 rq_num_pas	  = (rq_sz_po + page_size - 1) / page_size;
+
+	return rq_num_pas * sizeof(u64);
+}
+
+static void set_wq(void *wq, struct mlx5_srq_attr *in)
+{
+	MLX5_SET(wq,   wq, wq_signature,  !!(in->flags
+		 & MLX5_SRQ_FLAG_WQ_SIG));
+	MLX5_SET(wq,   wq, log_wq_pg_sz,  in->log_page_size);
+	MLX5_SET(wq,   wq, log_wq_stride, in->wqe_shift + 4);
+	MLX5_SET(wq,   wq, log_wq_sz,     in->log_size);
+	MLX5_SET(wq,   wq, page_offset,   in->page_offset);
+	MLX5_SET(wq,   wq, lwm,		  in->lwm);
+	MLX5_SET(wq,   wq, pd,		  in->pd);
+	MLX5_SET64(wq, wq, dbr_addr,	  in->db_record);
+}
+
+static void set_srqc(void *srqc, struct mlx5_srq_attr *in)
+{
+	MLX5_SET(srqc,   srqc, wq_signature,  !!(in->flags
+		 & MLX5_SRQ_FLAG_WQ_SIG));
+	MLX5_SET(srqc,   srqc, log_page_size, in->log_page_size);
+	MLX5_SET(srqc,   srqc, log_rq_stride, in->wqe_shift);
+	MLX5_SET(srqc,   srqc, log_srq_size,  in->log_size);
+	MLX5_SET(srqc,   srqc, page_offset,   in->page_offset);
+	MLX5_SET(srqc,	 srqc, lwm,	      in->lwm);
+	MLX5_SET(srqc,	 srqc, pd,	      in->pd);
+	MLX5_SET64(srqc, srqc, dbr_addr,      in->db_record);
+	MLX5_SET(srqc,	 srqc, xrcd,	      in->xrcd);
+	MLX5_SET(srqc,	 srqc, cqn,	      in->cqn);
+}
+
+static void get_wq(void *wq, struct mlx5_srq_attr *in)
+{
+	if (MLX5_GET(wq, wq, wq_signature))
+		in->flags &= MLX5_SRQ_FLAG_WQ_SIG;
+	in->log_page_size = MLX5_GET(wq,   wq, log_wq_pg_sz);
+	in->wqe_shift	  = MLX5_GET(wq,   wq, log_wq_stride) - 4;
+	in->log_size	  = MLX5_GET(wq,   wq, log_wq_sz);
+	in->page_offset   = MLX5_GET(wq,   wq, page_offset);
+	in->lwm		  = MLX5_GET(wq,   wq, lwm);
+	in->pd		  = MLX5_GET(wq,   wq, pd);
+	in->db_record	  = MLX5_GET64(wq, wq, dbr_addr);
+}
+
+static void get_srqc(void *srqc, struct mlx5_srq_attr *in)
+{
+	if (MLX5_GET(srqc, srqc, wq_signature))
+		in->flags &= MLX5_SRQ_FLAG_WQ_SIG;
+	in->log_page_size = MLX5_GET(srqc,   srqc, log_page_size);
+	in->wqe_shift	  = MLX5_GET(srqc,   srqc, log_rq_stride);
+	in->log_size	  = MLX5_GET(srqc,   srqc, log_srq_size);
+	in->page_offset   = MLX5_GET(srqc,   srqc, page_offset);
+	in->lwm		  = MLX5_GET(srqc,   srqc, lwm);
+	in->pd		  = MLX5_GET(srqc,   srqc, pd);
+	in->db_record	  = MLX5_GET64(srqc, srqc, dbr_addr);
+}
+
+struct mlx5_core_srq *mlx5_core_get_srq(struct mlx5_core_dev *dev, u32 srqn)
+{
+	struct mlx5_srq_table *table = &dev->priv.srq_table;
+	struct mlx5_core_srq *srq;
+
+	spin_lock(&table->lock);
+
+	srq = radix_tree_lookup(&table->tree, srqn);
+	if (srq)
+		atomic_inc(&srq->refcount);
+
+	spin_unlock(&table->lock);
+
+	return srq;
+}
+EXPORT_SYMBOL(mlx5_core_get_srq);
+
+static int create_srq_cmd(struct mlx5_core_dev *dev, struct mlx5_core_srq *srq,
+			  struct mlx5_srq_attr *in)
+{
+	u32 create_out[MLX5_ST_SZ_DW(create_srq_out)] = {0};
+	void *create_in;
+	void *srqc;
+	void *pas;
+	int pas_size;
+	int inlen;
+	int err;
+
+	pas_size  = get_pas_size(in);
+	inlen	  = MLX5_ST_SZ_BYTES(create_srq_in) + pas_size;
+	create_in = kvzalloc(inlen, GFP_KERNEL);
+	if (!create_in)
+		return -ENOMEM;
+
+	srqc = MLX5_ADDR_OF(create_srq_in, create_in, srq_context_entry);
+	pas = MLX5_ADDR_OF(create_srq_in, create_in, pas);
+
+	set_srqc(srqc, in);
+	memcpy(pas, in->pas, pas_size);
+
+	MLX5_SET(create_srq_in, create_in, opcode,
+		 MLX5_CMD_OP_CREATE_SRQ);
+
+	err = mlx5_cmd_exec(dev, create_in, inlen, create_out,
+			    sizeof(create_out));
+	kvfree(create_in);
+	if (!err)
+		srq->srqn = MLX5_GET(create_srq_out, create_out, srqn);
+
+	return err;
+}
+
+static int destroy_srq_cmd(struct mlx5_core_dev *dev,
+			   struct mlx5_core_srq *srq)
+{
+	u32 srq_in[MLX5_ST_SZ_DW(destroy_srq_in)] = {0};
+	u32 srq_out[MLX5_ST_SZ_DW(destroy_srq_out)] = {0};
+
+	MLX5_SET(destroy_srq_in, srq_in, opcode,
+		 MLX5_CMD_OP_DESTROY_SRQ);
+	MLX5_SET(destroy_srq_in, srq_in, srqn, srq->srqn);
+
+	return mlx5_cmd_exec(dev, srq_in, sizeof(srq_in),
+			     srq_out, sizeof(srq_out));
+}
+
+static int arm_srq_cmd(struct mlx5_core_dev *dev, struct mlx5_core_srq *srq,
+		       u16 lwm, int is_srq)
+{
+	u32 srq_in[MLX5_ST_SZ_DW(arm_rq_in)] = {0};
+	u32 srq_out[MLX5_ST_SZ_DW(arm_rq_out)] = {0};
+
+	MLX5_SET(arm_rq_in, srq_in, opcode, MLX5_CMD_OP_ARM_RQ);
+	MLX5_SET(arm_rq_in, srq_in, op_mod, MLX5_ARM_RQ_IN_OP_MOD_SRQ);
+	MLX5_SET(arm_rq_in, srq_in, srq_number, srq->srqn);
+	MLX5_SET(arm_rq_in, srq_in, lwm,      lwm);
+
+	return  mlx5_cmd_exec(dev, srq_in, sizeof(srq_in),
+			      srq_out, sizeof(srq_out));
+}
+
+static int query_srq_cmd(struct mlx5_core_dev *dev, struct mlx5_core_srq *srq,
+			 struct mlx5_srq_attr *out)
+{
+	u32 srq_in[MLX5_ST_SZ_DW(query_srq_in)] = {0};
+	u32 *srq_out;
+	void *srqc;
+	int err;
+
+	srq_out = kvzalloc(MLX5_ST_SZ_BYTES(query_srq_out), GFP_KERNEL);
+	if (!srq_out)
+		return -ENOMEM;
+
+	MLX5_SET(query_srq_in, srq_in, opcode,
+		 MLX5_CMD_OP_QUERY_SRQ);
+	MLX5_SET(query_srq_in, srq_in, srqn, srq->srqn);
+	err =  mlx5_cmd_exec(dev, srq_in, sizeof(srq_in),
+			     srq_out, MLX5_ST_SZ_BYTES(query_srq_out));
+	if (err)
+		goto out;
+
+	srqc = MLX5_ADDR_OF(query_srq_out, srq_out, srq_context_entry);
+	get_srqc(srqc, out);
+	if (MLX5_GET(srqc, srqc, state) != MLX5_SRQC_STATE_GOOD)
+		out->flags |= MLX5_SRQ_FLAG_ERR;
+out:
+	kvfree(srq_out);
+	return err;
+}
+
+static int create_xrc_srq_cmd(struct mlx5_core_dev *dev,
+			      struct mlx5_core_srq *srq,
+			      struct mlx5_srq_attr *in)
+{
+	u32 create_out[MLX5_ST_SZ_DW(create_xrc_srq_out)];
+	void *create_in;
+	void *xrc_srqc;
+	void *pas;
+	int pas_size;
+	int inlen;
+	int err;
+
+	pas_size  = get_pas_size(in);
+	inlen	  = MLX5_ST_SZ_BYTES(create_xrc_srq_in) + pas_size;
+	create_in = kvzalloc(inlen, GFP_KERNEL);
+	if (!create_in)
+		return -ENOMEM;
+
+	xrc_srqc = MLX5_ADDR_OF(create_xrc_srq_in, create_in,
+				xrc_srq_context_entry);
+	pas	 = MLX5_ADDR_OF(create_xrc_srq_in, create_in, pas);
+
+	set_srqc(xrc_srqc, in);
+	MLX5_SET(xrc_srqc, xrc_srqc, user_index, in->user_index);
+	memcpy(pas, in->pas, pas_size);
+	MLX5_SET(create_xrc_srq_in, create_in, opcode,
+		 MLX5_CMD_OP_CREATE_XRC_SRQ);
+
+	memset(create_out, 0, sizeof(create_out));
+	err = mlx5_cmd_exec(dev, create_in, inlen, create_out,
+			    sizeof(create_out));
+	if (err)
+		goto out;
+
+	srq->srqn = MLX5_GET(create_xrc_srq_out, create_out, xrc_srqn);
+out:
+	kvfree(create_in);
+	return err;
+}
+
+static int destroy_xrc_srq_cmd(struct mlx5_core_dev *dev,
+			       struct mlx5_core_srq *srq)
+{
+	u32 xrcsrq_in[MLX5_ST_SZ_DW(destroy_xrc_srq_in)]   = {0};
+	u32 xrcsrq_out[MLX5_ST_SZ_DW(destroy_xrc_srq_out)] = {0};
+
+	MLX5_SET(destroy_xrc_srq_in, xrcsrq_in, opcode,
+		 MLX5_CMD_OP_DESTROY_XRC_SRQ);
+	MLX5_SET(destroy_xrc_srq_in, xrcsrq_in, xrc_srqn, srq->srqn);
+
+	return mlx5_cmd_exec(dev, xrcsrq_in, sizeof(xrcsrq_in),
+			     xrcsrq_out, sizeof(xrcsrq_out));
+}
+
+static int arm_xrc_srq_cmd(struct mlx5_core_dev *dev,
+			   struct mlx5_core_srq *srq, u16 lwm)
+{
+	u32 xrcsrq_in[MLX5_ST_SZ_DW(arm_xrc_srq_in)]   = {0};
+	u32 xrcsrq_out[MLX5_ST_SZ_DW(arm_xrc_srq_out)] = {0};
+
+	MLX5_SET(arm_xrc_srq_in, xrcsrq_in, opcode,   MLX5_CMD_OP_ARM_XRC_SRQ);
+	MLX5_SET(arm_xrc_srq_in, xrcsrq_in, op_mod,   MLX5_ARM_XRC_SRQ_IN_OP_MOD_XRC_SRQ);
+	MLX5_SET(arm_xrc_srq_in, xrcsrq_in, xrc_srqn, srq->srqn);
+	MLX5_SET(arm_xrc_srq_in, xrcsrq_in, lwm,      lwm);
+
+	return  mlx5_cmd_exec(dev, xrcsrq_in, sizeof(xrcsrq_in),
+			      xrcsrq_out, sizeof(xrcsrq_out));
+}
+
+static int query_xrc_srq_cmd(struct mlx5_core_dev *dev,
+			     struct mlx5_core_srq *srq,
+			     struct mlx5_srq_attr *out)
+{
+	u32 xrcsrq_in[MLX5_ST_SZ_DW(query_xrc_srq_in)];
+	u32 *xrcsrq_out;
+	void *xrc_srqc;
+	int err;
+
+	xrcsrq_out = kvzalloc(MLX5_ST_SZ_BYTES(query_xrc_srq_out), GFP_KERNEL);
+	if (!xrcsrq_out)
+		return -ENOMEM;
+	memset(xrcsrq_in, 0, sizeof(xrcsrq_in));
+
+	MLX5_SET(query_xrc_srq_in, xrcsrq_in, opcode,
+		 MLX5_CMD_OP_QUERY_XRC_SRQ);
+	MLX5_SET(query_xrc_srq_in, xrcsrq_in, xrc_srqn, srq->srqn);
+
+	err =  mlx5_cmd_exec(dev, xrcsrq_in, sizeof(xrcsrq_in), xrcsrq_out,
+			     MLX5_ST_SZ_BYTES(query_xrc_srq_out));
+	if (err)
+		goto out;
+
+	xrc_srqc = MLX5_ADDR_OF(query_xrc_srq_out, xrcsrq_out,
+				xrc_srq_context_entry);
+	get_srqc(xrc_srqc, out);
+	if (MLX5_GET(xrc_srqc, xrc_srqc, state) != MLX5_XRC_SRQC_STATE_GOOD)
+		out->flags |= MLX5_SRQ_FLAG_ERR;
+
+out:
+	kvfree(xrcsrq_out);
+	return err;
+}
+
+static int create_rmp_cmd(struct mlx5_core_dev *dev, struct mlx5_core_srq *srq,
+			  struct mlx5_srq_attr *in)
+{
+	void *create_in;
+	void *rmpc;
+	void *wq;
+	int pas_size;
+	int inlen;
+	int err;
+
+	pas_size = get_pas_size(in);
+	inlen = MLX5_ST_SZ_BYTES(create_rmp_in) + pas_size;
+	create_in = kvzalloc(inlen, GFP_KERNEL);
+	if (!create_in)
+		return -ENOMEM;
+
+	rmpc = MLX5_ADDR_OF(create_rmp_in, create_in, ctx);
+	wq = MLX5_ADDR_OF(rmpc, rmpc, wq);
+
+	MLX5_SET(rmpc, rmpc, state, MLX5_RMPC_STATE_RDY);
+	set_wq(wq, in);
+	memcpy(MLX5_ADDR_OF(rmpc, rmpc, wq.pas), in->pas, pas_size);
+
+	err = mlx5_core_create_rmp(dev, create_in, inlen, &srq->srqn);
+
+	kvfree(create_in);
+	return err;
+}
+
+static int destroy_rmp_cmd(struct mlx5_core_dev *dev,
+			   struct mlx5_core_srq *srq)
+{
+	return mlx5_core_destroy_rmp(dev, srq->srqn);
+}
+
+static int arm_rmp_cmd(struct mlx5_core_dev *dev,
+		       struct mlx5_core_srq *srq,
+		       u16 lwm)
+{
+	void *in;
+	void *rmpc;
+	void *wq;
+	void *bitmask;
+	int err;
+
+	in = kvzalloc(MLX5_ST_SZ_BYTES(modify_rmp_in), GFP_KERNEL);
+	if (!in)
+		return -ENOMEM;
+
+	rmpc =	  MLX5_ADDR_OF(modify_rmp_in,   in,   ctx);
+	bitmask = MLX5_ADDR_OF(modify_rmp_in,   in,   bitmask);
+	wq   =	  MLX5_ADDR_OF(rmpc,	        rmpc, wq);
+
+	MLX5_SET(modify_rmp_in, in,	 rmp_state, MLX5_RMPC_STATE_RDY);
+	MLX5_SET(modify_rmp_in, in,	 rmpn,      srq->srqn);
+	MLX5_SET(wq,		wq,	 lwm,	    lwm);
+	MLX5_SET(rmp_bitmask,	bitmask, lwm,	    1);
+	MLX5_SET(rmpc, rmpc, state, MLX5_RMPC_STATE_RDY);
+
+	err = mlx5_core_modify_rmp(dev, in, MLX5_ST_SZ_BYTES(modify_rmp_in));
+
+	kvfree(in);
+	return err;
+}
+
+static int query_rmp_cmd(struct mlx5_core_dev *dev, struct mlx5_core_srq *srq,
+			 struct mlx5_srq_attr *out)
+{
+	u32 *rmp_out;
+	void *rmpc;
+	int err;
+
+	rmp_out =  kvzalloc(MLX5_ST_SZ_BYTES(query_rmp_out), GFP_KERNEL);
+	if (!rmp_out)
+		return -ENOMEM;
+
+	err = mlx5_core_query_rmp(dev, srq->srqn, rmp_out);
+	if (err)
+		goto out;
+
+	rmpc = MLX5_ADDR_OF(query_rmp_out, rmp_out, rmp_context);
+	get_wq(MLX5_ADDR_OF(rmpc, rmpc, wq), out);
+	if (MLX5_GET(rmpc, rmpc, state) != MLX5_RMPC_STATE_RDY)
+		out->flags |= MLX5_SRQ_FLAG_ERR;
+
+out:
+	kvfree(rmp_out);
+	return err;
+}
+
+static int create_xrq_cmd(struct mlx5_core_dev *dev, struct mlx5_core_srq *srq,
+			  struct mlx5_srq_attr *in)
+{
+	u32 create_out[MLX5_ST_SZ_DW(create_xrq_out)] = {0};
+	void *create_in;
+	void *xrqc;
+	void *wq;
+	int pas_size;
+	int inlen;
+	int err;
+
+	pas_size = get_pas_size(in);
+	inlen = MLX5_ST_SZ_BYTES(create_xrq_in) + pas_size;
+	create_in = kvzalloc(inlen, GFP_KERNEL);
+	if (!create_in)
+		return -ENOMEM;
+
+	xrqc = MLX5_ADDR_OF(create_xrq_in, create_in, xrq_context);
+	wq = MLX5_ADDR_OF(xrqc, xrqc, wq);
+
+	set_wq(wq, in);
+	memcpy(MLX5_ADDR_OF(xrqc, xrqc, wq.pas), in->pas, pas_size);
+
+	if (in->type == IB_SRQT_TM) {
+		MLX5_SET(xrqc, xrqc, topology, MLX5_XRQC_TOPOLOGY_TAG_MATCHING);
+		if (in->flags & MLX5_SRQ_FLAG_RNDV)
+			MLX5_SET(xrqc, xrqc, offload, MLX5_XRQC_OFFLOAD_RNDV);
+		MLX5_SET(xrqc, xrqc,
+			 tag_matching_topology_context.log_matching_list_sz,
+			 in->tm_log_list_size);
+	}
+	MLX5_SET(xrqc, xrqc, user_index, in->user_index);
+	MLX5_SET(xrqc, xrqc, cqn, in->cqn);
+	MLX5_SET(create_xrq_in, create_in, opcode, MLX5_CMD_OP_CREATE_XRQ);
+	err = mlx5_cmd_exec(dev, create_in, inlen, create_out,
+			    sizeof(create_out));
+	kvfree(create_in);
+	if (!err)
+		srq->srqn = MLX5_GET(create_xrq_out, create_out, xrqn);
+
+	return err;
+}
+
+static int destroy_xrq_cmd(struct mlx5_core_dev *dev, struct mlx5_core_srq *srq)
+{
+	u32 in[MLX5_ST_SZ_DW(destroy_xrq_in)] = {0};
+	u32 out[MLX5_ST_SZ_DW(destroy_xrq_out)] = {0};
+
+	MLX5_SET(destroy_xrq_in, in, opcode, MLX5_CMD_OP_DESTROY_XRQ);
+	MLX5_SET(destroy_xrq_in, in, xrqn,   srq->srqn);
+
+	return mlx5_cmd_exec(dev, in, sizeof(in), out, sizeof(out));
+}
+
+static int arm_xrq_cmd(struct mlx5_core_dev *dev,
+		       struct mlx5_core_srq *srq,
+		       u16 lwm)
+{
+	u32 out[MLX5_ST_SZ_DW(arm_rq_out)] = {0};
+	u32 in[MLX5_ST_SZ_DW(arm_rq_in)] = {0};
+
+	MLX5_SET(arm_rq_in, in, opcode,     MLX5_CMD_OP_ARM_RQ);
+	MLX5_SET(arm_rq_in, in, op_mod,     MLX5_ARM_RQ_IN_OP_MOD_XRQ);
+	MLX5_SET(arm_rq_in, in, srq_number, srq->srqn);
+	MLX5_SET(arm_rq_in, in, lwm,	    lwm);
+
+	return mlx5_cmd_exec(dev, in, sizeof(in), out, sizeof(out));
+}
+
+static int query_xrq_cmd(struct mlx5_core_dev *dev, struct mlx5_core_srq *srq,
+			 struct mlx5_srq_attr *out)
+{
+	u32 in[MLX5_ST_SZ_DW(query_xrq_in)] = {0};
+	u32 *xrq_out;
+	int outlen = MLX5_ST_SZ_BYTES(query_xrq_out);
+	void *xrqc;
+	int err;
+
+	xrq_out = kvzalloc(outlen, GFP_KERNEL);
+	if (!xrq_out)
+		return -ENOMEM;
+
+	MLX5_SET(query_xrq_in, in, opcode, MLX5_CMD_OP_QUERY_XRQ);
+	MLX5_SET(query_xrq_in, in, xrqn, srq->srqn);
+
+	err = mlx5_cmd_exec(dev, in, sizeof(in), xrq_out, outlen);
+	if (err)
+		goto out;
+
+	xrqc = MLX5_ADDR_OF(query_xrq_out, xrq_out, xrq_context);
+	get_wq(MLX5_ADDR_OF(xrqc, xrqc, wq), out);
+	if (MLX5_GET(xrqc, xrqc, state) != MLX5_XRQC_STATE_GOOD)
+		out->flags |= MLX5_SRQ_FLAG_ERR;
+	out->tm_next_tag =
+		MLX5_GET(xrqc, xrqc,
+			 tag_matching_topology_context.append_next_index);
+	out->tm_hw_phase_cnt =
+		MLX5_GET(xrqc, xrqc,
+			 tag_matching_topology_context.hw_phase_cnt);
+	out->tm_sw_phase_cnt =
+		MLX5_GET(xrqc, xrqc,
+			 tag_matching_topology_context.sw_phase_cnt);
+
+out:
+	kvfree(xrq_out);
+	return err;
+}
+
+static int create_srq_split(struct mlx5_core_dev *dev,
+			    struct mlx5_core_srq *srq,
+			    struct mlx5_srq_attr *in)
+{
+	if (!dev->issi)
+		return create_srq_cmd(dev, srq, in);
+	switch (srq->common.res) {
+	case MLX5_RES_XSRQ:
+		return create_xrc_srq_cmd(dev, srq, in);
+	case MLX5_RES_XRQ:
+		return create_xrq_cmd(dev, srq, in);
+	default:
+		return create_rmp_cmd(dev, srq, in);
+	}
+}
+
+static int destroy_srq_split(struct mlx5_core_dev *dev,
+			     struct mlx5_core_srq *srq)
+{
+	if (!dev->issi)
+		return destroy_srq_cmd(dev, srq);
+	switch (srq->common.res) {
+	case MLX5_RES_XSRQ:
+		return destroy_xrc_srq_cmd(dev, srq);
+	case MLX5_RES_XRQ:
+		return destroy_xrq_cmd(dev, srq);
+	default:
+		return destroy_rmp_cmd(dev, srq);
+	}
+}
+
+int mlx5_core_create_srq(struct mlx5_core_dev *dev, struct mlx5_core_srq *srq,
+			 struct mlx5_srq_attr *in)
+{
+	int err;
+	struct mlx5_srq_table *table = &dev->priv.srq_table;
+
+	switch (in->type) {
+	case IB_SRQT_XRC:
+		srq->common.res = MLX5_RES_XSRQ;
+		break;
+	case IB_SRQT_TM:
+		srq->common.res = MLX5_RES_XRQ;
+		break;
+	default:
+		srq->common.res = MLX5_RES_SRQ;
+	}
+
+	err = create_srq_split(dev, srq, in);
+	if (err)
+		return err;
+
+	atomic_set(&srq->refcount, 1);
+	init_completion(&srq->free);
+
+	spin_lock_irq(&table->lock);
+	err = radix_tree_insert(&table->tree, srq->srqn, srq);
+	spin_unlock_irq(&table->lock);
+	if (err) {
+		mlx5_core_warn(dev, "err %d, srqn 0x%x\n", err, srq->srqn);
+		goto err_destroy_srq_split;
+	}
+
+	return 0;
+
+err_destroy_srq_split:
+	destroy_srq_split(dev, srq);
+
+	return err;
+}
+EXPORT_SYMBOL(mlx5_core_create_srq);
+
+int mlx5_core_destroy_srq(struct mlx5_core_dev *dev, struct mlx5_core_srq *srq)
+{
+	struct mlx5_srq_table *table = &dev->priv.srq_table;
+	struct mlx5_core_srq *tmp;
+	int err;
+
+	spin_lock_irq(&table->lock);
+	tmp = radix_tree_delete(&table->tree, srq->srqn);
+	spin_unlock_irq(&table->lock);
+	if (!tmp) {
+		mlx5_core_warn(dev, "srq 0x%x not found in tree\n", srq->srqn);
+		return -EINVAL;
+	}
+	if (tmp != srq) {
+		mlx5_core_warn(dev, "corruption on srqn 0x%x\n", srq->srqn);
+		return -EINVAL;
+	}
+
+	err = destroy_srq_split(dev, srq);
+	if (err)
+		return err;
+
+	if (atomic_dec_and_test(&srq->refcount))
+		complete(&srq->free);
+	wait_for_completion(&srq->free);
+
+	return 0;
+}
+EXPORT_SYMBOL(mlx5_core_destroy_srq);
+
+int mlx5_core_query_srq(struct mlx5_core_dev *dev, struct mlx5_core_srq *srq,
+			struct mlx5_srq_attr *out)
+{
+	if (!dev->issi)
+		return query_srq_cmd(dev, srq, out);
+	switch (srq->common.res) {
+	case MLX5_RES_XSRQ:
+		return query_xrc_srq_cmd(dev, srq, out);
+	case MLX5_RES_XRQ:
+		return query_xrq_cmd(dev, srq, out);
+	default:
+		return query_rmp_cmd(dev, srq, out);
+	}
+}
+EXPORT_SYMBOL(mlx5_core_query_srq);
+
+int mlx5_core_arm_srq(struct mlx5_core_dev *dev, struct mlx5_core_srq *srq,
+		      u16 lwm, int is_srq)
+{
+	if (!dev->issi)
+		return arm_srq_cmd(dev, srq, lwm, is_srq);
+	switch (srq->common.res) {
+	case MLX5_RES_XSRQ:
+		return arm_xrc_srq_cmd(dev, srq, lwm);
+	case MLX5_RES_XRQ:
+		return arm_xrq_cmd(dev, srq, lwm);
+	default:
+		return arm_rmp_cmd(dev, srq, lwm);
+	}
+}
+EXPORT_SYMBOL(mlx5_core_arm_srq);
+
+void mlx5_init_srq_table(struct mlx5_core_dev *dev)
+{
+	struct mlx5_srq_table *table = &dev->priv.srq_table;
+
+	memset(table, 0, sizeof(*table));
+	spin_lock_init(&table->lock);
+	INIT_RADIX_TREE(&table->tree, GFP_ATOMIC);
+}
+
+void mlx5_cleanup_srq_table(struct mlx5_core_dev *dev)
+{
+	/* nothing */
+}
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/transobj.c b/drivers/net/ethernet/mellanox/mlx5/core/transobj.c
new file mode 100644
index 000000000..a1ee9a8a7
--- /dev/null
+++ b/drivers/net/ethernet/mellanox/mlx5/core/transobj.c
@@ -0,0 +1,621 @@
+/*
+ * Copyright (c) 2013-2015, Mellanox Technologies, Ltd.  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/mlx5/driver.h>
+#include "mlx5_core.h"
+#include <linux/mlx5/transobj.h>
+
+int mlx5_core_alloc_transport_domain(struct mlx5_core_dev *dev, u32 *tdn)
+{
+	u32 in[MLX5_ST_SZ_DW(alloc_transport_domain_in)]   = {0};
+	u32 out[MLX5_ST_SZ_DW(alloc_transport_domain_out)] = {0};
+	int err;
+
+	MLX5_SET(alloc_transport_domain_in, in, opcode,
+		 MLX5_CMD_OP_ALLOC_TRANSPORT_DOMAIN);
+
+	err = mlx5_cmd_exec(dev, in, sizeof(in), out, sizeof(out));
+	if (!err)
+		*tdn = MLX5_GET(alloc_transport_domain_out, out,
+				transport_domain);
+
+	return err;
+}
+EXPORT_SYMBOL(mlx5_core_alloc_transport_domain);
+
+void mlx5_core_dealloc_transport_domain(struct mlx5_core_dev *dev, u32 tdn)
+{
+	u32 in[MLX5_ST_SZ_DW(dealloc_transport_domain_in)]   = {0};
+	u32 out[MLX5_ST_SZ_DW(dealloc_transport_domain_out)] = {0};
+
+	MLX5_SET(dealloc_transport_domain_in, in, opcode,
+		 MLX5_CMD_OP_DEALLOC_TRANSPORT_DOMAIN);
+	MLX5_SET(dealloc_transport_domain_in, in, transport_domain, tdn);
+	mlx5_cmd_exec(dev, in, sizeof(in), out, sizeof(out));
+}
+EXPORT_SYMBOL(mlx5_core_dealloc_transport_domain);
+
+int mlx5_core_create_rq(struct mlx5_core_dev *dev, u32 *in, int inlen, u32 *rqn)
+{
+	u32 out[MLX5_ST_SZ_DW(create_rq_out)] = {0};
+	int err;
+
+	MLX5_SET(create_rq_in, in, opcode, MLX5_CMD_OP_CREATE_RQ);
+	err = mlx5_cmd_exec(dev, in, inlen, out, sizeof(out));
+	if (!err)
+		*rqn = MLX5_GET(create_rq_out, out, rqn);
+
+	return err;
+}
+EXPORT_SYMBOL(mlx5_core_create_rq);
+
+int mlx5_core_modify_rq(struct mlx5_core_dev *dev, u32 rqn, u32 *in, int inlen)
+{
+	u32 out[MLX5_ST_SZ_DW(modify_rq_out)];
+
+	MLX5_SET(modify_rq_in, in, rqn, rqn);
+	MLX5_SET(modify_rq_in, in, opcode, MLX5_CMD_OP_MODIFY_RQ);
+
+	memset(out, 0, sizeof(out));
+	return mlx5_cmd_exec(dev, in, inlen, out, sizeof(out));
+}
+EXPORT_SYMBOL(mlx5_core_modify_rq);
+
+void mlx5_core_destroy_rq(struct mlx5_core_dev *dev, u32 rqn)
+{
+	u32 in[MLX5_ST_SZ_DW(destroy_rq_in)]   = {0};
+	u32 out[MLX5_ST_SZ_DW(destroy_rq_out)] = {0};
+
+	MLX5_SET(destroy_rq_in, in, opcode, MLX5_CMD_OP_DESTROY_RQ);
+	MLX5_SET(destroy_rq_in, in, rqn, rqn);
+	mlx5_cmd_exec(dev, in, sizeof(in), out, sizeof(out));
+}
+EXPORT_SYMBOL(mlx5_core_destroy_rq);
+
+int mlx5_core_query_rq(struct mlx5_core_dev *dev, u32 rqn, u32 *out)
+{
+	u32 in[MLX5_ST_SZ_DW(query_rq_in)] = {0};
+	int outlen = MLX5_ST_SZ_BYTES(query_rq_out);
+
+	MLX5_SET(query_rq_in, in, opcode, MLX5_CMD_OP_QUERY_RQ);
+	MLX5_SET(query_rq_in, in, rqn, rqn);
+
+	return mlx5_cmd_exec(dev, in, sizeof(in), out, outlen);
+}
+EXPORT_SYMBOL(mlx5_core_query_rq);
+
+int mlx5_core_create_sq(struct mlx5_core_dev *dev, u32 *in, int inlen, u32 *sqn)
+{
+	u32 out[MLX5_ST_SZ_DW(create_sq_out)] = {0};
+	int err;
+
+	MLX5_SET(create_sq_in, in, opcode, MLX5_CMD_OP_CREATE_SQ);
+	err = mlx5_cmd_exec(dev, in, inlen, out, sizeof(out));
+	if (!err)
+		*sqn = MLX5_GET(create_sq_out, out, sqn);
+
+	return err;
+}
+
+int mlx5_core_modify_sq(struct mlx5_core_dev *dev, u32 sqn, u32 *in, int inlen)
+{
+	u32 out[MLX5_ST_SZ_DW(modify_sq_out)] = {0};
+
+	MLX5_SET(modify_sq_in, in, sqn, sqn);
+	MLX5_SET(modify_sq_in, in, opcode, MLX5_CMD_OP_MODIFY_SQ);
+	return mlx5_cmd_exec(dev, in, inlen, out, sizeof(out));
+}
+EXPORT_SYMBOL(mlx5_core_modify_sq);
+
+void mlx5_core_destroy_sq(struct mlx5_core_dev *dev, u32 sqn)
+{
+	u32 in[MLX5_ST_SZ_DW(destroy_sq_in)]   = {0};
+	u32 out[MLX5_ST_SZ_DW(destroy_sq_out)] = {0};
+
+	MLX5_SET(destroy_sq_in, in, opcode, MLX5_CMD_OP_DESTROY_SQ);
+	MLX5_SET(destroy_sq_in, in, sqn, sqn);
+	mlx5_cmd_exec(dev, in, sizeof(in), out, sizeof(out));
+}
+
+int mlx5_core_query_sq(struct mlx5_core_dev *dev, u32 sqn, u32 *out)
+{
+	u32 in[MLX5_ST_SZ_DW(query_sq_in)] = {0};
+	int outlen = MLX5_ST_SZ_BYTES(query_sq_out);
+
+	MLX5_SET(query_sq_in, in, opcode, MLX5_CMD_OP_QUERY_SQ);
+	MLX5_SET(query_sq_in, in, sqn, sqn);
+	return mlx5_cmd_exec(dev, in, sizeof(in), out, outlen);
+}
+EXPORT_SYMBOL(mlx5_core_query_sq);
+
+int mlx5_core_query_sq_state(struct mlx5_core_dev *dev, u32 sqn, u8 *state)
+{
+	void *out;
+	void *sqc;
+	int inlen;
+	int err;
+
+	inlen = MLX5_ST_SZ_BYTES(query_sq_out);
+	out = kvzalloc(inlen, GFP_KERNEL);
+	if (!out)
+		return -ENOMEM;
+
+	err = mlx5_core_query_sq(dev, sqn, out);
+	if (err)
+		goto out;
+
+	sqc = MLX5_ADDR_OF(query_sq_out, out, sq_context);
+	*state = MLX5_GET(sqc, sqc, state);
+
+out:
+	kvfree(out);
+	return err;
+}
+EXPORT_SYMBOL_GPL(mlx5_core_query_sq_state);
+
+int mlx5_core_create_tir(struct mlx5_core_dev *dev, u32 *in, int inlen,
+			 u32 *tirn)
+{
+	u32 out[MLX5_ST_SZ_DW(create_tir_out)] = {0};
+	int err;
+
+	MLX5_SET(create_tir_in, in, opcode, MLX5_CMD_OP_CREATE_TIR);
+
+	memset(out, 0, sizeof(out));
+	err = mlx5_cmd_exec(dev, in, inlen, out, sizeof(out));
+	if (!err)
+		*tirn = MLX5_GET(create_tir_out, out, tirn);
+
+	return err;
+}
+EXPORT_SYMBOL(mlx5_core_create_tir);
+
+int mlx5_core_modify_tir(struct mlx5_core_dev *dev, u32 tirn, u32 *in,
+			 int inlen)
+{
+	u32 out[MLX5_ST_SZ_DW(modify_tir_out)] = {0};
+
+	MLX5_SET(modify_tir_in, in, tirn, tirn);
+	MLX5_SET(modify_tir_in, in, opcode, MLX5_CMD_OP_MODIFY_TIR);
+	return mlx5_cmd_exec(dev, in, inlen, out, sizeof(out));
+}
+
+void mlx5_core_destroy_tir(struct mlx5_core_dev *dev, u32 tirn)
+{
+	u32 in[MLX5_ST_SZ_DW(destroy_tir_in)]   = {0};
+	u32 out[MLX5_ST_SZ_DW(destroy_tir_out)] = {0};
+
+	MLX5_SET(destroy_tir_in, in, opcode, MLX5_CMD_OP_DESTROY_TIR);
+	MLX5_SET(destroy_tir_in, in, tirn, tirn);
+	mlx5_cmd_exec(dev, in, sizeof(in), out, sizeof(out));
+}
+EXPORT_SYMBOL(mlx5_core_destroy_tir);
+
+int mlx5_core_create_tis(struct mlx5_core_dev *dev, u32 *in, int inlen,
+			 u32 *tisn)
+{
+	u32 out[MLX5_ST_SZ_DW(create_tis_out)] = {0};
+	int err;
+
+	MLX5_SET(create_tis_in, in, opcode, MLX5_CMD_OP_CREATE_TIS);
+	err = mlx5_cmd_exec(dev, in, inlen, out, sizeof(out));
+	if (!err)
+		*tisn = MLX5_GET(create_tis_out, out, tisn);
+
+	return err;
+}
+EXPORT_SYMBOL(mlx5_core_create_tis);
+
+int mlx5_core_modify_tis(struct mlx5_core_dev *dev, u32 tisn, u32 *in,
+			 int inlen)
+{
+	u32 out[MLX5_ST_SZ_DW(modify_tis_out)] = {0};
+
+	MLX5_SET(modify_tis_in, in, tisn, tisn);
+	MLX5_SET(modify_tis_in, in, opcode, MLX5_CMD_OP_MODIFY_TIS);
+
+	return mlx5_cmd_exec(dev, in, inlen, out, sizeof(out));
+}
+EXPORT_SYMBOL(mlx5_core_modify_tis);
+
+void mlx5_core_destroy_tis(struct mlx5_core_dev *dev, u32 tisn)
+{
+	u32 in[MLX5_ST_SZ_DW(destroy_tis_in)]   = {0};
+	u32 out[MLX5_ST_SZ_DW(destroy_tis_out)] = {0};
+
+	MLX5_SET(destroy_tis_in, in, opcode, MLX5_CMD_OP_DESTROY_TIS);
+	MLX5_SET(destroy_tis_in, in, tisn, tisn);
+	mlx5_cmd_exec(dev, in, sizeof(in), out, sizeof(out));
+}
+EXPORT_SYMBOL(mlx5_core_destroy_tis);
+
+int mlx5_core_create_rmp(struct mlx5_core_dev *dev, u32 *in, int inlen,
+			 u32 *rmpn)
+{
+	u32 out[MLX5_ST_SZ_DW(create_rmp_out)] = {0};
+	int err;
+
+	MLX5_SET(create_rmp_in, in, opcode, MLX5_CMD_OP_CREATE_RMP);
+	err = mlx5_cmd_exec(dev, in, inlen, out, sizeof(out));
+	if (!err)
+		*rmpn = MLX5_GET(create_rmp_out, out, rmpn);
+
+	return err;
+}
+
+int mlx5_core_modify_rmp(struct mlx5_core_dev *dev, u32 *in, int inlen)
+{
+	u32 out[MLX5_ST_SZ_DW(modify_rmp_out)] = {0};
+
+	MLX5_SET(modify_rmp_in, in, opcode, MLX5_CMD_OP_MODIFY_RMP);
+	return mlx5_cmd_exec(dev, in, inlen, out, sizeof(out));
+}
+
+int mlx5_core_destroy_rmp(struct mlx5_core_dev *dev, u32 rmpn)
+{
+	u32 in[MLX5_ST_SZ_DW(destroy_rmp_in)]   = {0};
+	u32 out[MLX5_ST_SZ_DW(destroy_rmp_out)] = {0};
+
+	MLX5_SET(destroy_rmp_in, in, opcode, MLX5_CMD_OP_DESTROY_RMP);
+	MLX5_SET(destroy_rmp_in, in, rmpn, rmpn);
+	return mlx5_cmd_exec(dev, in, sizeof(in), out,
+					  sizeof(out));
+}
+
+int mlx5_core_query_rmp(struct mlx5_core_dev *dev, u32 rmpn, u32 *out)
+{
+	u32 in[MLX5_ST_SZ_DW(query_rmp_in)] = {0};
+	int outlen = MLX5_ST_SZ_BYTES(query_rmp_out);
+
+	MLX5_SET(query_rmp_in, in, opcode, MLX5_CMD_OP_QUERY_RMP);
+	MLX5_SET(query_rmp_in, in, rmpn,   rmpn);
+	return mlx5_cmd_exec(dev, in, sizeof(in), out, outlen);
+}
+
+int mlx5_core_arm_rmp(struct mlx5_core_dev *dev, u32 rmpn, u16 lwm)
+{
+	void *in;
+	void *rmpc;
+	void *wq;
+	void *bitmask;
+	int  err;
+
+	in = kvzalloc(MLX5_ST_SZ_BYTES(modify_rmp_in), GFP_KERNEL);
+	if (!in)
+		return -ENOMEM;
+
+	rmpc    = MLX5_ADDR_OF(modify_rmp_in,   in,   ctx);
+	bitmask = MLX5_ADDR_OF(modify_rmp_in,   in,   bitmask);
+	wq      = MLX5_ADDR_OF(rmpc,	        rmpc, wq);
+
+	MLX5_SET(modify_rmp_in, in,	 rmp_state, MLX5_RMPC_STATE_RDY);
+	MLX5_SET(modify_rmp_in, in,	 rmpn,      rmpn);
+	MLX5_SET(wq,		wq,	 lwm,	    lwm);
+	MLX5_SET(rmp_bitmask,	bitmask, lwm,	    1);
+	MLX5_SET(rmpc,		rmpc,	 state,	    MLX5_RMPC_STATE_RDY);
+
+	err =  mlx5_core_modify_rmp(dev, in, MLX5_ST_SZ_BYTES(modify_rmp_in));
+
+	kvfree(in);
+
+	return err;
+}
+
+int mlx5_core_create_xsrq(struct mlx5_core_dev *dev, u32 *in, int inlen,
+			  u32 *xsrqn)
+{
+	u32 out[MLX5_ST_SZ_DW(create_xrc_srq_out)] = {0};
+	int err;
+
+	MLX5_SET(create_xrc_srq_in, in, opcode,     MLX5_CMD_OP_CREATE_XRC_SRQ);
+	err = mlx5_cmd_exec(dev, in, inlen, out, sizeof(out));
+	if (!err)
+		*xsrqn = MLX5_GET(create_xrc_srq_out, out, xrc_srqn);
+
+	return err;
+}
+
+int mlx5_core_destroy_xsrq(struct mlx5_core_dev *dev, u32 xsrqn)
+{
+	u32 in[MLX5_ST_SZ_DW(destroy_xrc_srq_in)]   = {0};
+	u32 out[MLX5_ST_SZ_DW(destroy_xrc_srq_out)] = {0};
+
+	MLX5_SET(destroy_xrc_srq_in, in, opcode,   MLX5_CMD_OP_DESTROY_XRC_SRQ);
+	MLX5_SET(destroy_xrc_srq_in, in, xrc_srqn, xsrqn);
+	return mlx5_cmd_exec(dev, in, sizeof(in), out, sizeof(out));
+}
+
+int mlx5_core_arm_xsrq(struct mlx5_core_dev *dev, u32 xsrqn, u16 lwm)
+{
+	u32 in[MLX5_ST_SZ_DW(arm_xrc_srq_in)]   = {0};
+	u32 out[MLX5_ST_SZ_DW(arm_xrc_srq_out)] = {0};
+
+	MLX5_SET(arm_xrc_srq_in, in, opcode,   MLX5_CMD_OP_ARM_XRC_SRQ);
+	MLX5_SET(arm_xrc_srq_in, in, xrc_srqn, xsrqn);
+	MLX5_SET(arm_xrc_srq_in, in, lwm,      lwm);
+	MLX5_SET(arm_xrc_srq_in, in, op_mod,
+		 MLX5_ARM_XRC_SRQ_IN_OP_MOD_XRC_SRQ);
+	return mlx5_cmd_exec(dev, in, sizeof(in), out, sizeof(out));
+}
+
+int mlx5_core_create_rqt(struct mlx5_core_dev *dev, u32 *in, int inlen,
+			 u32 *rqtn)
+{
+	u32 out[MLX5_ST_SZ_DW(create_rqt_out)] = {0};
+	int err;
+
+	MLX5_SET(create_rqt_in, in, opcode, MLX5_CMD_OP_CREATE_RQT);
+	err = mlx5_cmd_exec(dev, in, inlen, out, sizeof(out));
+	if (!err)
+		*rqtn = MLX5_GET(create_rqt_out, out, rqtn);
+
+	return err;
+}
+EXPORT_SYMBOL(mlx5_core_create_rqt);
+
+int mlx5_core_modify_rqt(struct mlx5_core_dev *dev, u32 rqtn, u32 *in,
+			 int inlen)
+{
+	u32 out[MLX5_ST_SZ_DW(modify_rqt_out)] = {0};
+
+	MLX5_SET(modify_rqt_in, in, rqtn, rqtn);
+	MLX5_SET(modify_rqt_in, in, opcode, MLX5_CMD_OP_MODIFY_RQT);
+	return mlx5_cmd_exec(dev, in, inlen, out, sizeof(out));
+}
+
+void mlx5_core_destroy_rqt(struct mlx5_core_dev *dev, u32 rqtn)
+{
+	u32 in[MLX5_ST_SZ_DW(destroy_rqt_in)]   = {0};
+	u32 out[MLX5_ST_SZ_DW(destroy_rqt_out)] = {0};
+
+	MLX5_SET(destroy_rqt_in, in, opcode, MLX5_CMD_OP_DESTROY_RQT);
+	MLX5_SET(destroy_rqt_in, in, rqtn, rqtn);
+	mlx5_cmd_exec(dev, in, sizeof(in), out, sizeof(out));
+}
+EXPORT_SYMBOL(mlx5_core_destroy_rqt);
+
+static int mlx5_hairpin_create_rq(struct mlx5_core_dev *mdev,
+				  struct mlx5_hairpin_params *params, u32 *rqn)
+{
+	u32 in[MLX5_ST_SZ_DW(create_rq_in)] = {0};
+	void *rqc, *wq;
+
+	rqc = MLX5_ADDR_OF(create_rq_in, in, ctx);
+	wq  = MLX5_ADDR_OF(rqc, rqc, wq);
+
+	MLX5_SET(rqc, rqc, hairpin, 1);
+	MLX5_SET(rqc, rqc, state, MLX5_RQC_STATE_RST);
+	MLX5_SET(rqc, rqc, counter_set_id, params->q_counter);
+
+	MLX5_SET(wq, wq, log_hairpin_data_sz, params->log_data_size);
+	MLX5_SET(wq, wq, log_hairpin_num_packets, params->log_num_packets);
+
+	return mlx5_core_create_rq(mdev, in, MLX5_ST_SZ_BYTES(create_rq_in), rqn);
+}
+
+static int mlx5_hairpin_create_sq(struct mlx5_core_dev *mdev,
+				  struct mlx5_hairpin_params *params, u32 *sqn)
+{
+	u32 in[MLX5_ST_SZ_DW(create_sq_in)] = {0};
+	void *sqc, *wq;
+
+	sqc = MLX5_ADDR_OF(create_sq_in, in, ctx);
+	wq  = MLX5_ADDR_OF(sqc, sqc, wq);
+
+	MLX5_SET(sqc, sqc, hairpin, 1);
+	MLX5_SET(sqc, sqc, state, MLX5_SQC_STATE_RST);
+
+	MLX5_SET(wq, wq, log_hairpin_data_sz, params->log_data_size);
+	MLX5_SET(wq, wq, log_hairpin_num_packets, params->log_num_packets);
+
+	return mlx5_core_create_sq(mdev, in, MLX5_ST_SZ_BYTES(create_sq_in), sqn);
+}
+
+static int mlx5_hairpin_create_queues(struct mlx5_hairpin *hp,
+				      struct mlx5_hairpin_params *params)
+{
+	int i, j, err;
+
+	for (i = 0; i < hp->num_channels; i++) {
+		err = mlx5_hairpin_create_rq(hp->func_mdev, params, &hp->rqn[i]);
+		if (err)
+			goto out_err_rq;
+	}
+
+	for (i = 0; i < hp->num_channels; i++) {
+		err = mlx5_hairpin_create_sq(hp->peer_mdev, params, &hp->sqn[i]);
+		if (err)
+			goto out_err_sq;
+	}
+
+	return 0;
+
+out_err_sq:
+	for (j = 0; j < i; j++)
+		mlx5_core_destroy_sq(hp->peer_mdev, hp->sqn[j]);
+	i = hp->num_channels;
+out_err_rq:
+	for (j = 0; j < i; j++)
+		mlx5_core_destroy_rq(hp->func_mdev, hp->rqn[j]);
+	return err;
+}
+
+static void mlx5_hairpin_destroy_queues(struct mlx5_hairpin *hp)
+{
+	int i;
+
+	for (i = 0; i < hp->num_channels; i++) {
+		mlx5_core_destroy_rq(hp->func_mdev, hp->rqn[i]);
+		if (!hp->peer_gone)
+			mlx5_core_destroy_sq(hp->peer_mdev, hp->sqn[i]);
+	}
+}
+
+static int mlx5_hairpin_modify_rq(struct mlx5_core_dev *func_mdev, u32 rqn,
+				  int curr_state, int next_state,
+				  u16 peer_vhca, u32 peer_sq)
+{
+	u32 in[MLX5_ST_SZ_DW(modify_rq_in)] = {0};
+	void *rqc;
+
+	rqc = MLX5_ADDR_OF(modify_rq_in, in, ctx);
+
+	if (next_state == MLX5_RQC_STATE_RDY) {
+		MLX5_SET(rqc, rqc, hairpin_peer_sq, peer_sq);
+		MLX5_SET(rqc, rqc, hairpin_peer_vhca, peer_vhca);
+	}
+
+	MLX5_SET(modify_rq_in, in, rq_state, curr_state);
+	MLX5_SET(rqc, rqc, state, next_state);
+
+	return mlx5_core_modify_rq(func_mdev, rqn,
+				   in, MLX5_ST_SZ_BYTES(modify_rq_in));
+}
+
+static int mlx5_hairpin_modify_sq(struct mlx5_core_dev *peer_mdev, u32 sqn,
+				  int curr_state, int next_state,
+				  u16 peer_vhca, u32 peer_rq)
+{
+	u32 in[MLX5_ST_SZ_DW(modify_sq_in)] = {0};
+	void *sqc;
+
+	sqc = MLX5_ADDR_OF(modify_sq_in, in, ctx);
+
+	if (next_state == MLX5_SQC_STATE_RDY) {
+		MLX5_SET(sqc, sqc, hairpin_peer_rq, peer_rq);
+		MLX5_SET(sqc, sqc, hairpin_peer_vhca, peer_vhca);
+	}
+
+	MLX5_SET(modify_sq_in, in, sq_state, curr_state);
+	MLX5_SET(sqc, sqc, state, next_state);
+
+	return mlx5_core_modify_sq(peer_mdev, sqn,
+				   in, MLX5_ST_SZ_BYTES(modify_sq_in));
+}
+
+static int mlx5_hairpin_pair_queues(struct mlx5_hairpin *hp)
+{
+	int i, j, err;
+
+	/* set peer SQs */
+	for (i = 0; i < hp->num_channels; i++) {
+		err = mlx5_hairpin_modify_sq(hp->peer_mdev, hp->sqn[i],
+					     MLX5_SQC_STATE_RST, MLX5_SQC_STATE_RDY,
+					     MLX5_CAP_GEN(hp->func_mdev, vhca_id), hp->rqn[i]);
+		if (err)
+			goto err_modify_sq;
+	}
+
+	/* set func RQs */
+	for (i = 0; i < hp->num_channels; i++) {
+		err = mlx5_hairpin_modify_rq(hp->func_mdev, hp->rqn[i],
+					     MLX5_RQC_STATE_RST, MLX5_RQC_STATE_RDY,
+					     MLX5_CAP_GEN(hp->peer_mdev, vhca_id), hp->sqn[i]);
+		if (err)
+			goto err_modify_rq;
+	}
+
+	return 0;
+
+err_modify_rq:
+	for (j = 0; j < i; j++)
+		mlx5_hairpin_modify_rq(hp->func_mdev, hp->rqn[j], MLX5_RQC_STATE_RDY,
+				       MLX5_RQC_STATE_RST, 0, 0);
+	i = hp->num_channels;
+err_modify_sq:
+	for (j = 0; j < i; j++)
+		mlx5_hairpin_modify_sq(hp->peer_mdev, hp->sqn[j], MLX5_SQC_STATE_RDY,
+				       MLX5_SQC_STATE_RST, 0, 0);
+	return err;
+}
+
+static void mlx5_hairpin_unpair_queues(struct mlx5_hairpin *hp)
+{
+	int i;
+
+	/* unset func RQs */
+	for (i = 0; i < hp->num_channels; i++)
+		mlx5_hairpin_modify_rq(hp->func_mdev, hp->rqn[i], MLX5_RQC_STATE_RDY,
+				       MLX5_RQC_STATE_RST, 0, 0);
+
+	/* unset peer SQs */
+	if (hp->peer_gone)
+		return;
+	for (i = 0; i < hp->num_channels; i++)
+		mlx5_hairpin_modify_sq(hp->peer_mdev, hp->sqn[i], MLX5_SQC_STATE_RDY,
+				       MLX5_SQC_STATE_RST, 0, 0);
+}
+
+struct mlx5_hairpin *
+mlx5_core_hairpin_create(struct mlx5_core_dev *func_mdev,
+			 struct mlx5_core_dev *peer_mdev,
+			 struct mlx5_hairpin_params *params)
+{
+	struct mlx5_hairpin *hp;
+	int size, err;
+
+	size = sizeof(*hp) + params->num_channels * 2 * sizeof(u32);
+	hp = kzalloc(size, GFP_KERNEL);
+	if (!hp)
+		return ERR_PTR(-ENOMEM);
+
+	hp->func_mdev = func_mdev;
+	hp->peer_mdev = peer_mdev;
+	hp->num_channels = params->num_channels;
+
+	hp->rqn = (void *)hp + sizeof(*hp);
+	hp->sqn = hp->rqn + params->num_channels;
+
+	/* alloc and pair func --> peer hairpin */
+	err = mlx5_hairpin_create_queues(hp, params);
+	if (err)
+		goto err_create_queues;
+
+	err = mlx5_hairpin_pair_queues(hp);
+	if (err)
+		goto err_pair_queues;
+
+	return hp;
+
+err_pair_queues:
+	mlx5_hairpin_destroy_queues(hp);
+err_create_queues:
+	kfree(hp);
+	return ERR_PTR(err);
+}
+
+void mlx5_core_hairpin_destroy(struct mlx5_hairpin *hp)
+{
+	mlx5_hairpin_unpair_queues(hp);
+	mlx5_hairpin_destroy_queues(hp);
+	kfree(hp);
+}
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/uar.c b/drivers/net/ethernet/mellanox/mlx5/core/uar.c
new file mode 100644
index 000000000..8b97066dd
--- /dev/null
+++ b/drivers/net/ethernet/mellanox/mlx5/core/uar.c
@@ -0,0 +1,329 @@
+/*
+ * Copyright (c) 2013-2015, 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/module.h>
+#include <linux/io-mapping.h>
+#include <linux/mlx5/driver.h>
+#include <linux/mlx5/cmd.h>
+#include "mlx5_core.h"
+
+int mlx5_cmd_alloc_uar(struct mlx5_core_dev *dev, u32 *uarn)
+{
+	u32 out[MLX5_ST_SZ_DW(alloc_uar_out)] = {0};
+	u32 in[MLX5_ST_SZ_DW(alloc_uar_in)]   = {0};
+	int err;
+
+	MLX5_SET(alloc_uar_in, in, opcode, MLX5_CMD_OP_ALLOC_UAR);
+	err = mlx5_cmd_exec(dev, in, sizeof(in), out, sizeof(out));
+	if (!err)
+		*uarn = MLX5_GET(alloc_uar_out, out, uar);
+	return err;
+}
+EXPORT_SYMBOL(mlx5_cmd_alloc_uar);
+
+int mlx5_cmd_free_uar(struct mlx5_core_dev *dev, u32 uarn)
+{
+	u32 out[MLX5_ST_SZ_DW(dealloc_uar_out)] = {0};
+	u32 in[MLX5_ST_SZ_DW(dealloc_uar_in)]   = {0};
+
+	MLX5_SET(dealloc_uar_in, in, opcode, MLX5_CMD_OP_DEALLOC_UAR);
+	MLX5_SET(dealloc_uar_in, in, uar, uarn);
+	return mlx5_cmd_exec(dev, in, sizeof(in), out, sizeof(out));
+}
+EXPORT_SYMBOL(mlx5_cmd_free_uar);
+
+static int uars_per_sys_page(struct mlx5_core_dev *mdev)
+{
+	if (MLX5_CAP_GEN(mdev, uar_4k))
+		return MLX5_CAP_GEN(mdev, num_of_uars_per_page);
+
+	return 1;
+}
+
+static u64 uar2pfn(struct mlx5_core_dev *mdev, u32 index)
+{
+	u32 system_page_index;
+
+	if (MLX5_CAP_GEN(mdev, uar_4k))
+		system_page_index = index >> (PAGE_SHIFT - MLX5_ADAPTER_PAGE_SHIFT);
+	else
+		system_page_index = index;
+
+	return (pci_resource_start(mdev->pdev, 0) >> PAGE_SHIFT) + system_page_index;
+}
+
+static void up_rel_func(struct kref *kref)
+{
+	struct mlx5_uars_page *up = container_of(kref, struct mlx5_uars_page, ref_count);
+
+	list_del(&up->list);
+	iounmap(up->map);
+	if (mlx5_cmd_free_uar(up->mdev, up->index))
+		mlx5_core_warn(up->mdev, "failed to free uar index %d\n", up->index);
+	kfree(up->reg_bitmap);
+	kfree(up->fp_bitmap);
+	kfree(up);
+}
+
+static struct mlx5_uars_page *alloc_uars_page(struct mlx5_core_dev *mdev,
+					      bool map_wc)
+{
+	struct mlx5_uars_page *up;
+	int err = -ENOMEM;
+	phys_addr_t pfn;
+	int bfregs;
+	int i;
+
+	bfregs = uars_per_sys_page(mdev) * MLX5_BFREGS_PER_UAR;
+	up = kzalloc(sizeof(*up), GFP_KERNEL);
+	if (!up)
+		return ERR_PTR(err);
+
+	up->mdev = mdev;
+	up->reg_bitmap = kcalloc(BITS_TO_LONGS(bfregs), sizeof(unsigned long), GFP_KERNEL);
+	if (!up->reg_bitmap)
+		goto error1;
+
+	up->fp_bitmap = kcalloc(BITS_TO_LONGS(bfregs), sizeof(unsigned long), GFP_KERNEL);
+	if (!up->fp_bitmap)
+		goto error1;
+
+	for (i = 0; i < bfregs; i++)
+		if ((i % MLX5_BFREGS_PER_UAR) < MLX5_NON_FP_BFREGS_PER_UAR)
+			set_bit(i, up->reg_bitmap);
+		else
+			set_bit(i, up->fp_bitmap);
+
+	up->bfregs = bfregs;
+	up->fp_avail = bfregs * MLX5_FP_BFREGS_PER_UAR / MLX5_BFREGS_PER_UAR;
+	up->reg_avail = bfregs * MLX5_NON_FP_BFREGS_PER_UAR / MLX5_BFREGS_PER_UAR;
+
+	err = mlx5_cmd_alloc_uar(mdev, &up->index);
+	if (err) {
+		mlx5_core_warn(mdev, "mlx5_cmd_alloc_uar() failed, %d\n", err);
+		goto error1;
+	}
+
+	pfn = uar2pfn(mdev, up->index);
+	if (map_wc) {
+		up->map = ioremap_wc(pfn << PAGE_SHIFT, PAGE_SIZE);
+		if (!up->map) {
+			err = -EAGAIN;
+			goto error2;
+		}
+	} else {
+		up->map = ioremap(pfn << PAGE_SHIFT, PAGE_SIZE);
+		if (!up->map) {
+			err = -ENOMEM;
+			goto error2;
+		}
+	}
+	kref_init(&up->ref_count);
+	mlx5_core_dbg(mdev, "allocated UAR page: index %d, total bfregs %d\n",
+		      up->index, up->bfregs);
+	return up;
+
+error2:
+	if (mlx5_cmd_free_uar(mdev, up->index))
+		mlx5_core_warn(mdev, "failed to free uar index %d\n", up->index);
+error1:
+	kfree(up->fp_bitmap);
+	kfree(up->reg_bitmap);
+	kfree(up);
+	return ERR_PTR(err);
+}
+
+struct mlx5_uars_page *mlx5_get_uars_page(struct mlx5_core_dev *mdev)
+{
+	struct mlx5_uars_page *ret;
+
+	mutex_lock(&mdev->priv.bfregs.reg_head.lock);
+	if (!list_empty(&mdev->priv.bfregs.reg_head.list)) {
+		ret = list_first_entry(&mdev->priv.bfregs.reg_head.list,
+				       struct mlx5_uars_page, list);
+		kref_get(&ret->ref_count);
+		goto out;
+	}
+	ret = alloc_uars_page(mdev, false);
+	if (IS_ERR(ret))
+		goto out;
+	list_add(&ret->list, &mdev->priv.bfregs.reg_head.list);
+out:
+	mutex_unlock(&mdev->priv.bfregs.reg_head.lock);
+
+	return ret;
+}
+EXPORT_SYMBOL(mlx5_get_uars_page);
+
+void mlx5_put_uars_page(struct mlx5_core_dev *mdev, struct mlx5_uars_page *up)
+{
+	mutex_lock(&mdev->priv.bfregs.reg_head.lock);
+	kref_put(&up->ref_count, up_rel_func);
+	mutex_unlock(&mdev->priv.bfregs.reg_head.lock);
+}
+EXPORT_SYMBOL(mlx5_put_uars_page);
+
+static unsigned long map_offset(struct mlx5_core_dev *mdev, int dbi)
+{
+	/* return the offset in bytes from the start of the page to the
+	 * blue flame area of the UAR
+	 */
+	return dbi / MLX5_BFREGS_PER_UAR * MLX5_ADAPTER_PAGE_SIZE +
+	       (dbi % MLX5_BFREGS_PER_UAR) *
+	       (1 << MLX5_CAP_GEN(mdev, log_bf_reg_size)) + MLX5_BF_OFFSET;
+}
+
+static int alloc_bfreg(struct mlx5_core_dev *mdev, struct mlx5_sq_bfreg *bfreg,
+		       bool map_wc, bool fast_path)
+{
+	struct mlx5_bfreg_data *bfregs;
+	struct mlx5_uars_page *up;
+	struct list_head *head;
+	unsigned long *bitmap;
+	unsigned int *avail;
+	struct mutex *lock;  /* pointer to right mutex */
+	int dbi;
+
+	bfregs = &mdev->priv.bfregs;
+	if (map_wc) {
+		head = &bfregs->wc_head.list;
+		lock = &bfregs->wc_head.lock;
+	} else {
+		head = &bfregs->reg_head.list;
+		lock = &bfregs->reg_head.lock;
+	}
+	mutex_lock(lock);
+	if (list_empty(head)) {
+		up = alloc_uars_page(mdev, map_wc);
+		if (IS_ERR(up)) {
+			mutex_unlock(lock);
+			return PTR_ERR(up);
+		}
+		list_add(&up->list, head);
+	} else {
+		up = list_entry(head->next, struct mlx5_uars_page, list);
+		kref_get(&up->ref_count);
+	}
+	if (fast_path) {
+		bitmap = up->fp_bitmap;
+		avail = &up->fp_avail;
+	} else {
+		bitmap = up->reg_bitmap;
+		avail = &up->reg_avail;
+	}
+	dbi = find_first_bit(bitmap, up->bfregs);
+	clear_bit(dbi, bitmap);
+	(*avail)--;
+	if (!(*avail))
+		list_del(&up->list);
+
+	bfreg->map = up->map + map_offset(mdev, dbi);
+	bfreg->up = up;
+	bfreg->wc = map_wc;
+	bfreg->index = up->index + dbi / MLX5_BFREGS_PER_UAR;
+	mutex_unlock(lock);
+
+	return 0;
+}
+
+int mlx5_alloc_bfreg(struct mlx5_core_dev *mdev, struct mlx5_sq_bfreg *bfreg,
+		     bool map_wc, bool fast_path)
+{
+	int err;
+
+	err = alloc_bfreg(mdev, bfreg, map_wc, fast_path);
+	if (!err)
+		return 0;
+
+	if (err == -EAGAIN && map_wc)
+		return alloc_bfreg(mdev, bfreg, false, fast_path);
+
+	return err;
+}
+EXPORT_SYMBOL(mlx5_alloc_bfreg);
+
+static unsigned int addr_to_dbi_in_syspage(struct mlx5_core_dev *dev,
+					   struct mlx5_uars_page *up,
+					   struct mlx5_sq_bfreg *bfreg)
+{
+	unsigned int uar_idx;
+	unsigned int bfreg_idx;
+	unsigned int bf_reg_size;
+
+	bf_reg_size = 1 << MLX5_CAP_GEN(dev, log_bf_reg_size);
+
+	uar_idx = (bfreg->map - up->map) >> MLX5_ADAPTER_PAGE_SHIFT;
+	bfreg_idx = (((uintptr_t)bfreg->map % MLX5_ADAPTER_PAGE_SIZE) - MLX5_BF_OFFSET) / bf_reg_size;
+
+	return uar_idx * MLX5_BFREGS_PER_UAR + bfreg_idx;
+}
+
+void mlx5_free_bfreg(struct mlx5_core_dev *mdev, struct mlx5_sq_bfreg *bfreg)
+{
+	struct mlx5_bfreg_data *bfregs;
+	struct mlx5_uars_page *up;
+	struct mutex *lock; /* pointer to right mutex */
+	unsigned int dbi;
+	bool fp;
+	unsigned int *avail;
+	unsigned long *bitmap;
+	struct list_head *head;
+
+	bfregs = &mdev->priv.bfregs;
+	if (bfreg->wc) {
+		head = &bfregs->wc_head.list;
+		lock = &bfregs->wc_head.lock;
+	} else {
+		head = &bfregs->reg_head.list;
+		lock = &bfregs->reg_head.lock;
+	}
+	up = bfreg->up;
+	dbi = addr_to_dbi_in_syspage(mdev, up, bfreg);
+	fp = (dbi % MLX5_BFREGS_PER_UAR) >= MLX5_NON_FP_BFREGS_PER_UAR;
+	if (fp) {
+		avail = &up->fp_avail;
+		bitmap = up->fp_bitmap;
+	} else {
+		avail = &up->reg_avail;
+		bitmap = up->reg_bitmap;
+	}
+	mutex_lock(lock);
+	(*avail)++;
+	set_bit(dbi, bitmap);
+	if (*avail == 1)
+		list_add_tail(&up->list, head);
+
+	kref_put(&up->ref_count, up_rel_func);
+	mutex_unlock(lock);
+}
+EXPORT_SYMBOL(mlx5_free_bfreg);
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/vport.c b/drivers/net/ethernet/mellanox/mlx5/core/vport.c
new file mode 100644
index 000000000..b02af317c
--- /dev/null
+++ b/drivers/net/ethernet/mellanox/mlx5/core/vport.c
@@ -0,0 +1,1203 @@
+/*
+ * Copyright (c) 2013-2015, Mellanox Technologies, Ltd.  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 <linux/etherdevice.h>
+#include <linux/mlx5/driver.h>
+#include <linux/mlx5/vport.h>
+#include "mlx5_core.h"
+
+/* Mutex to hold while enabling or disabling RoCE */
+static DEFINE_MUTEX(mlx5_roce_en_lock);
+
+static int _mlx5_query_vport_state(struct mlx5_core_dev *mdev, u8 opmod,
+				   u16 vport, u32 *out, int outlen)
+{
+	u32 in[MLX5_ST_SZ_DW(query_vport_state_in)] = {0};
+
+	MLX5_SET(query_vport_state_in, in, opcode,
+		 MLX5_CMD_OP_QUERY_VPORT_STATE);
+	MLX5_SET(query_vport_state_in, in, op_mod, opmod);
+	MLX5_SET(query_vport_state_in, in, vport_number, vport);
+	if (vport)
+		MLX5_SET(query_vport_state_in, in, other_vport, 1);
+
+	return mlx5_cmd_exec(mdev, in, sizeof(in), out, outlen);
+}
+
+u8 mlx5_query_vport_state(struct mlx5_core_dev *mdev, u8 opmod, u16 vport)
+{
+	u32 out[MLX5_ST_SZ_DW(query_vport_state_out)] = {0};
+
+	_mlx5_query_vport_state(mdev, opmod, vport, out, sizeof(out));
+
+	return MLX5_GET(query_vport_state_out, out, state);
+}
+
+int mlx5_modify_vport_admin_state(struct mlx5_core_dev *mdev, u8 opmod,
+				  u16 vport, u8 state)
+{
+	u32 in[MLX5_ST_SZ_DW(modify_vport_state_in)]   = {0};
+	u32 out[MLX5_ST_SZ_DW(modify_vport_state_out)] = {0};
+
+	MLX5_SET(modify_vport_state_in, in, opcode,
+		 MLX5_CMD_OP_MODIFY_VPORT_STATE);
+	MLX5_SET(modify_vport_state_in, in, op_mod, opmod);
+	MLX5_SET(modify_vport_state_in, in, vport_number, vport);
+	if (vport)
+		MLX5_SET(modify_vport_state_in, in, other_vport, 1);
+	MLX5_SET(modify_vport_state_in, in, admin_state, state);
+
+	return mlx5_cmd_exec(mdev, in, sizeof(in), out, sizeof(out));
+}
+
+static int mlx5_query_nic_vport_context(struct mlx5_core_dev *mdev, u16 vport,
+					u32 *out, int outlen)
+{
+	u32 in[MLX5_ST_SZ_DW(query_nic_vport_context_in)] = {0};
+
+	MLX5_SET(query_nic_vport_context_in, in, opcode,
+		 MLX5_CMD_OP_QUERY_NIC_VPORT_CONTEXT);
+	MLX5_SET(query_nic_vport_context_in, in, vport_number, vport);
+	if (vport)
+		MLX5_SET(query_nic_vport_context_in, in, other_vport, 1);
+
+	return mlx5_cmd_exec(mdev, in, sizeof(in), out, outlen);
+}
+
+static int mlx5_modify_nic_vport_context(struct mlx5_core_dev *mdev, void *in,
+					 int inlen)
+{
+	u32 out[MLX5_ST_SZ_DW(modify_nic_vport_context_out)] = {0};
+
+	MLX5_SET(modify_nic_vport_context_in, in, opcode,
+		 MLX5_CMD_OP_MODIFY_NIC_VPORT_CONTEXT);
+	return mlx5_cmd_exec(mdev, in, inlen, out, sizeof(out));
+}
+
+int mlx5_query_nic_vport_min_inline(struct mlx5_core_dev *mdev,
+				    u16 vport, u8 *min_inline)
+{
+	u32 out[MLX5_ST_SZ_DW(query_nic_vport_context_out)] = {0};
+	int err;
+
+	err = mlx5_query_nic_vport_context(mdev, vport, out, sizeof(out));
+	if (!err)
+		*min_inline = MLX5_GET(query_nic_vport_context_out, out,
+				       nic_vport_context.min_wqe_inline_mode);
+	return err;
+}
+EXPORT_SYMBOL_GPL(mlx5_query_nic_vport_min_inline);
+
+void mlx5_query_min_inline(struct mlx5_core_dev *mdev,
+			   u8 *min_inline_mode)
+{
+	switch (MLX5_CAP_ETH(mdev, wqe_inline_mode)) {
+	case MLX5_CAP_INLINE_MODE_L2:
+		*min_inline_mode = MLX5_INLINE_MODE_L2;
+		break;
+	case MLX5_CAP_INLINE_MODE_VPORT_CONTEXT:
+		mlx5_query_nic_vport_min_inline(mdev, 0, min_inline_mode);
+		break;
+	case MLX5_CAP_INLINE_MODE_NOT_REQUIRED:
+		*min_inline_mode = MLX5_INLINE_MODE_NONE;
+		break;
+	}
+}
+EXPORT_SYMBOL_GPL(mlx5_query_min_inline);
+
+int mlx5_modify_nic_vport_min_inline(struct mlx5_core_dev *mdev,
+				     u16 vport, u8 min_inline)
+{
+	u32 in[MLX5_ST_SZ_DW(modify_nic_vport_context_in)] = {0};
+	int inlen = MLX5_ST_SZ_BYTES(modify_nic_vport_context_in);
+	void *nic_vport_ctx;
+
+	MLX5_SET(modify_nic_vport_context_in, in,
+		 field_select.min_inline, 1);
+	MLX5_SET(modify_nic_vport_context_in, in, vport_number, vport);
+	MLX5_SET(modify_nic_vport_context_in, in, other_vport, 1);
+
+	nic_vport_ctx = MLX5_ADDR_OF(modify_nic_vport_context_in,
+				     in, nic_vport_context);
+	MLX5_SET(nic_vport_context, nic_vport_ctx,
+		 min_wqe_inline_mode, min_inline);
+
+	return mlx5_modify_nic_vport_context(mdev, in, inlen);
+}
+
+int mlx5_query_nic_vport_mac_address(struct mlx5_core_dev *mdev,
+				     u16 vport, u8 *addr)
+{
+	u32 *out;
+	int outlen = MLX5_ST_SZ_BYTES(query_nic_vport_context_out);
+	u8 *out_addr;
+	int err;
+
+	out = kvzalloc(outlen, GFP_KERNEL);
+	if (!out)
+		return -ENOMEM;
+
+	out_addr = MLX5_ADDR_OF(query_nic_vport_context_out, out,
+				nic_vport_context.permanent_address);
+
+	err = mlx5_query_nic_vport_context(mdev, vport, out, outlen);
+	if (!err)
+		ether_addr_copy(addr, &out_addr[2]);
+
+	kvfree(out);
+	return err;
+}
+EXPORT_SYMBOL_GPL(mlx5_query_nic_vport_mac_address);
+
+int mlx5_modify_nic_vport_mac_address(struct mlx5_core_dev *mdev,
+				      u16 vport, u8 *addr)
+{
+	void *in;
+	int inlen = MLX5_ST_SZ_BYTES(modify_nic_vport_context_in);
+	int err;
+	void *nic_vport_ctx;
+	u8 *perm_mac;
+
+	in = kvzalloc(inlen, GFP_KERNEL);
+	if (!in)
+		return -ENOMEM;
+
+	MLX5_SET(modify_nic_vport_context_in, in,
+		 field_select.permanent_address, 1);
+	MLX5_SET(modify_nic_vport_context_in, in, vport_number, vport);
+
+	if (vport)
+		MLX5_SET(modify_nic_vport_context_in, in, other_vport, 1);
+
+	nic_vport_ctx = MLX5_ADDR_OF(modify_nic_vport_context_in,
+				     in, nic_vport_context);
+	perm_mac = MLX5_ADDR_OF(nic_vport_context, nic_vport_ctx,
+				permanent_address);
+
+	ether_addr_copy(&perm_mac[2], addr);
+
+	err = mlx5_modify_nic_vport_context(mdev, in, inlen);
+
+	kvfree(in);
+
+	return err;
+}
+EXPORT_SYMBOL_GPL(mlx5_modify_nic_vport_mac_address);
+
+int mlx5_query_nic_vport_mtu(struct mlx5_core_dev *mdev, u16 *mtu)
+{
+	int outlen = MLX5_ST_SZ_BYTES(query_nic_vport_context_out);
+	u32 *out;
+	int err;
+
+	out = kvzalloc(outlen, GFP_KERNEL);
+	if (!out)
+		return -ENOMEM;
+
+	err = mlx5_query_nic_vport_context(mdev, 0, out, outlen);
+	if (!err)
+		*mtu = MLX5_GET(query_nic_vport_context_out, out,
+				nic_vport_context.mtu);
+
+	kvfree(out);
+	return err;
+}
+EXPORT_SYMBOL_GPL(mlx5_query_nic_vport_mtu);
+
+int mlx5_modify_nic_vport_mtu(struct mlx5_core_dev *mdev, u16 mtu)
+{
+	int inlen = MLX5_ST_SZ_BYTES(modify_nic_vport_context_in);
+	void *in;
+	int err;
+
+	in = kvzalloc(inlen, GFP_KERNEL);
+	if (!in)
+		return -ENOMEM;
+
+	MLX5_SET(modify_nic_vport_context_in, in, field_select.mtu, 1);
+	MLX5_SET(modify_nic_vport_context_in, in, nic_vport_context.mtu, mtu);
+
+	err = mlx5_modify_nic_vport_context(mdev, in, inlen);
+
+	kvfree(in);
+	return err;
+}
+EXPORT_SYMBOL_GPL(mlx5_modify_nic_vport_mtu);
+
+int mlx5_query_nic_vport_mac_list(struct mlx5_core_dev *dev,
+				  u32 vport,
+				  enum mlx5_list_type list_type,
+				  u8 addr_list[][ETH_ALEN],
+				  int *list_size)
+{
+	u32 in[MLX5_ST_SZ_DW(query_nic_vport_context_in)] = {0};
+	void *nic_vport_ctx;
+	int max_list_size;
+	int req_list_size;
+	int out_sz;
+	void *out;
+	int err;
+	int i;
+
+	req_list_size = *list_size;
+
+	max_list_size = list_type == MLX5_NVPRT_LIST_TYPE_UC ?
+		1 << MLX5_CAP_GEN(dev, log_max_current_uc_list) :
+		1 << MLX5_CAP_GEN(dev, log_max_current_mc_list);
+
+	if (req_list_size > max_list_size) {
+		mlx5_core_warn(dev, "Requested list size (%d) > (%d) max_list_size\n",
+			       req_list_size, max_list_size);
+		req_list_size = max_list_size;
+	}
+
+	out_sz = MLX5_ST_SZ_BYTES(modify_nic_vport_context_in) +
+			req_list_size * MLX5_ST_SZ_BYTES(mac_address_layout);
+
+	out = kzalloc(out_sz, GFP_KERNEL);
+	if (!out)
+		return -ENOMEM;
+
+	MLX5_SET(query_nic_vport_context_in, in, opcode,
+		 MLX5_CMD_OP_QUERY_NIC_VPORT_CONTEXT);
+	MLX5_SET(query_nic_vport_context_in, in, allowed_list_type, list_type);
+	MLX5_SET(query_nic_vport_context_in, in, vport_number, vport);
+
+	if (vport)
+		MLX5_SET(query_nic_vport_context_in, in, other_vport, 1);
+
+	err = mlx5_cmd_exec(dev, in, sizeof(in), out, out_sz);
+	if (err)
+		goto out;
+
+	nic_vport_ctx = MLX5_ADDR_OF(query_nic_vport_context_out, out,
+				     nic_vport_context);
+	req_list_size = MLX5_GET(nic_vport_context, nic_vport_ctx,
+				 allowed_list_size);
+
+	*list_size = req_list_size;
+	for (i = 0; i < req_list_size; i++) {
+		u8 *mac_addr = MLX5_ADDR_OF(nic_vport_context,
+					nic_vport_ctx,
+					current_uc_mac_address[i]) + 2;
+		ether_addr_copy(addr_list[i], mac_addr);
+	}
+out:
+	kfree(out);
+	return err;
+}
+EXPORT_SYMBOL_GPL(mlx5_query_nic_vport_mac_list);
+
+int mlx5_modify_nic_vport_mac_list(struct mlx5_core_dev *dev,
+				   enum mlx5_list_type list_type,
+				   u8 addr_list[][ETH_ALEN],
+				   int list_size)
+{
+	u32 out[MLX5_ST_SZ_DW(modify_nic_vport_context_out)];
+	void *nic_vport_ctx;
+	int max_list_size;
+	int in_sz;
+	void *in;
+	int err;
+	int i;
+
+	max_list_size = list_type == MLX5_NVPRT_LIST_TYPE_UC ?
+		 1 << MLX5_CAP_GEN(dev, log_max_current_uc_list) :
+		 1 << MLX5_CAP_GEN(dev, log_max_current_mc_list);
+
+	if (list_size > max_list_size)
+		return -ENOSPC;
+
+	in_sz = MLX5_ST_SZ_BYTES(modify_nic_vport_context_in) +
+		list_size * MLX5_ST_SZ_BYTES(mac_address_layout);
+
+	memset(out, 0, sizeof(out));
+	in = kzalloc(in_sz, GFP_KERNEL);
+	if (!in)
+		return -ENOMEM;
+
+	MLX5_SET(modify_nic_vport_context_in, in, opcode,
+		 MLX5_CMD_OP_MODIFY_NIC_VPORT_CONTEXT);
+	MLX5_SET(modify_nic_vport_context_in, in,
+		 field_select.addresses_list, 1);
+
+	nic_vport_ctx = MLX5_ADDR_OF(modify_nic_vport_context_in, in,
+				     nic_vport_context);
+
+	MLX5_SET(nic_vport_context, nic_vport_ctx,
+		 allowed_list_type, list_type);
+	MLX5_SET(nic_vport_context, nic_vport_ctx,
+		 allowed_list_size, list_size);
+
+	for (i = 0; i < list_size; i++) {
+		u8 *curr_mac = MLX5_ADDR_OF(nic_vport_context,
+					    nic_vport_ctx,
+					    current_uc_mac_address[i]) + 2;
+		ether_addr_copy(curr_mac, addr_list[i]);
+	}
+
+	err = mlx5_cmd_exec(dev, in, in_sz, out, sizeof(out));
+	kfree(in);
+	return err;
+}
+EXPORT_SYMBOL_GPL(mlx5_modify_nic_vport_mac_list);
+
+int mlx5_query_nic_vport_vlans(struct mlx5_core_dev *dev,
+			       u32 vport,
+			       u16 vlans[],
+			       int *size)
+{
+	u32 in[MLX5_ST_SZ_DW(query_nic_vport_context_in)];
+	void *nic_vport_ctx;
+	int req_list_size;
+	int max_list_size;
+	int out_sz;
+	void *out;
+	int err;
+	int i;
+
+	req_list_size = *size;
+	max_list_size = 1 << MLX5_CAP_GEN(dev, log_max_vlan_list);
+	if (req_list_size > max_list_size) {
+		mlx5_core_warn(dev, "Requested list size (%d) > (%d) max list size\n",
+			       req_list_size, max_list_size);
+		req_list_size = max_list_size;
+	}
+
+	out_sz = MLX5_ST_SZ_BYTES(modify_nic_vport_context_in) +
+			req_list_size * MLX5_ST_SZ_BYTES(vlan_layout);
+
+	memset(in, 0, sizeof(in));
+	out = kzalloc(out_sz, GFP_KERNEL);
+	if (!out)
+		return -ENOMEM;
+
+	MLX5_SET(query_nic_vport_context_in, in, opcode,
+		 MLX5_CMD_OP_QUERY_NIC_VPORT_CONTEXT);
+	MLX5_SET(query_nic_vport_context_in, in, allowed_list_type,
+		 MLX5_NVPRT_LIST_TYPE_VLAN);
+	MLX5_SET(query_nic_vport_context_in, in, vport_number, vport);
+
+	if (vport)
+		MLX5_SET(query_nic_vport_context_in, in, other_vport, 1);
+
+	err = mlx5_cmd_exec(dev, in, sizeof(in), out, out_sz);
+	if (err)
+		goto out;
+
+	nic_vport_ctx = MLX5_ADDR_OF(query_nic_vport_context_out, out,
+				     nic_vport_context);
+	req_list_size = MLX5_GET(nic_vport_context, nic_vport_ctx,
+				 allowed_list_size);
+
+	*size = req_list_size;
+	for (i = 0; i < req_list_size; i++) {
+		void *vlan_addr = MLX5_ADDR_OF(nic_vport_context,
+					       nic_vport_ctx,
+					       current_uc_mac_address[i]);
+		vlans[i] = MLX5_GET(vlan_layout, vlan_addr, vlan);
+	}
+out:
+	kfree(out);
+	return err;
+}
+EXPORT_SYMBOL_GPL(mlx5_query_nic_vport_vlans);
+
+int mlx5_modify_nic_vport_vlans(struct mlx5_core_dev *dev,
+				u16 vlans[],
+				int list_size)
+{
+	u32 out[MLX5_ST_SZ_DW(modify_nic_vport_context_out)];
+	void *nic_vport_ctx;
+	int max_list_size;
+	int in_sz;
+	void *in;
+	int err;
+	int i;
+
+	max_list_size = 1 << MLX5_CAP_GEN(dev, log_max_vlan_list);
+
+	if (list_size > max_list_size)
+		return -ENOSPC;
+
+	in_sz = MLX5_ST_SZ_BYTES(modify_nic_vport_context_in) +
+		list_size * MLX5_ST_SZ_BYTES(vlan_layout);
+
+	memset(out, 0, sizeof(out));
+	in = kzalloc(in_sz, GFP_KERNEL);
+	if (!in)
+		return -ENOMEM;
+
+	MLX5_SET(modify_nic_vport_context_in, in, opcode,
+		 MLX5_CMD_OP_MODIFY_NIC_VPORT_CONTEXT);
+	MLX5_SET(modify_nic_vport_context_in, in,
+		 field_select.addresses_list, 1);
+
+	nic_vport_ctx = MLX5_ADDR_OF(modify_nic_vport_context_in, in,
+				     nic_vport_context);
+
+	MLX5_SET(nic_vport_context, nic_vport_ctx,
+		 allowed_list_type, MLX5_NVPRT_LIST_TYPE_VLAN);
+	MLX5_SET(nic_vport_context, nic_vport_ctx,
+		 allowed_list_size, list_size);
+
+	for (i = 0; i < list_size; i++) {
+		void *vlan_addr = MLX5_ADDR_OF(nic_vport_context,
+					       nic_vport_ctx,
+					       current_uc_mac_address[i]);
+		MLX5_SET(vlan_layout, vlan_addr, vlan, vlans[i]);
+	}
+
+	err = mlx5_cmd_exec(dev, in, in_sz, out, sizeof(out));
+	kfree(in);
+	return err;
+}
+EXPORT_SYMBOL_GPL(mlx5_modify_nic_vport_vlans);
+
+int mlx5_query_nic_vport_system_image_guid(struct mlx5_core_dev *mdev,
+					   u64 *system_image_guid)
+{
+	u32 *out;
+	int outlen = MLX5_ST_SZ_BYTES(query_nic_vport_context_out);
+
+	out = kvzalloc(outlen, GFP_KERNEL);
+	if (!out)
+		return -ENOMEM;
+
+	mlx5_query_nic_vport_context(mdev, 0, out, outlen);
+
+	*system_image_guid = MLX5_GET64(query_nic_vport_context_out, out,
+					nic_vport_context.system_image_guid);
+
+	kvfree(out);
+
+	return 0;
+}
+EXPORT_SYMBOL_GPL(mlx5_query_nic_vport_system_image_guid);
+
+int mlx5_query_nic_vport_node_guid(struct mlx5_core_dev *mdev, u64 *node_guid)
+{
+	u32 *out;
+	int outlen = MLX5_ST_SZ_BYTES(query_nic_vport_context_out);
+
+	out = kvzalloc(outlen, GFP_KERNEL);
+	if (!out)
+		return -ENOMEM;
+
+	mlx5_query_nic_vport_context(mdev, 0, out, outlen);
+
+	*node_guid = MLX5_GET64(query_nic_vport_context_out, out,
+				nic_vport_context.node_guid);
+
+	kvfree(out);
+
+	return 0;
+}
+EXPORT_SYMBOL_GPL(mlx5_query_nic_vport_node_guid);
+
+int mlx5_modify_nic_vport_node_guid(struct mlx5_core_dev *mdev,
+				    u32 vport, u64 node_guid)
+{
+	int inlen = MLX5_ST_SZ_BYTES(modify_nic_vport_context_in);
+	void *nic_vport_context;
+	void *in;
+	int err;
+
+	if (!vport)
+		return -EINVAL;
+	if (!MLX5_CAP_GEN(mdev, vport_group_manager))
+		return -EACCES;
+
+	in = kvzalloc(inlen, GFP_KERNEL);
+	if (!in)
+		return -ENOMEM;
+
+	MLX5_SET(modify_nic_vport_context_in, in,
+		 field_select.node_guid, 1);
+	MLX5_SET(modify_nic_vport_context_in, in, vport_number, vport);
+	MLX5_SET(modify_nic_vport_context_in, in, other_vport, !!vport);
+
+	nic_vport_context = MLX5_ADDR_OF(modify_nic_vport_context_in,
+					 in, nic_vport_context);
+	MLX5_SET64(nic_vport_context, nic_vport_context, node_guid, node_guid);
+
+	err = mlx5_modify_nic_vport_context(mdev, in, inlen);
+
+	kvfree(in);
+
+	return err;
+}
+
+int mlx5_query_nic_vport_qkey_viol_cntr(struct mlx5_core_dev *mdev,
+					u16 *qkey_viol_cntr)
+{
+	u32 *out;
+	int outlen = MLX5_ST_SZ_BYTES(query_nic_vport_context_out);
+
+	out = kvzalloc(outlen, GFP_KERNEL);
+	if (!out)
+		return -ENOMEM;
+
+	mlx5_query_nic_vport_context(mdev, 0, out, outlen);
+
+	*qkey_viol_cntr = MLX5_GET(query_nic_vport_context_out, out,
+				   nic_vport_context.qkey_violation_counter);
+
+	kvfree(out);
+
+	return 0;
+}
+EXPORT_SYMBOL_GPL(mlx5_query_nic_vport_qkey_viol_cntr);
+
+int mlx5_query_hca_vport_gid(struct mlx5_core_dev *dev, u8 other_vport,
+			     u8 port_num, u16  vf_num, u16 gid_index,
+			     union ib_gid *gid)
+{
+	int in_sz = MLX5_ST_SZ_BYTES(query_hca_vport_gid_in);
+	int out_sz = MLX5_ST_SZ_BYTES(query_hca_vport_gid_out);
+	int is_group_manager;
+	void *out = NULL;
+	void *in = NULL;
+	union ib_gid *tmp;
+	int tbsz;
+	int nout;
+	int err;
+
+	is_group_manager = MLX5_CAP_GEN(dev, vport_group_manager);
+	tbsz = mlx5_get_gid_table_len(MLX5_CAP_GEN(dev, gid_table_size));
+	mlx5_core_dbg(dev, "vf_num %d, index %d, gid_table_size %d\n",
+		      vf_num, gid_index, tbsz);
+
+	if (gid_index > tbsz && gid_index != 0xffff)
+		return -EINVAL;
+
+	if (gid_index == 0xffff)
+		nout = tbsz;
+	else
+		nout = 1;
+
+	out_sz += nout * sizeof(*gid);
+
+	in = kzalloc(in_sz, GFP_KERNEL);
+	out = kzalloc(out_sz, GFP_KERNEL);
+	if (!in || !out) {
+		err = -ENOMEM;
+		goto out;
+	}
+
+	MLX5_SET(query_hca_vport_gid_in, in, opcode, MLX5_CMD_OP_QUERY_HCA_VPORT_GID);
+	if (other_vport) {
+		if (is_group_manager) {
+			MLX5_SET(query_hca_vport_gid_in, in, vport_number, vf_num);
+			MLX5_SET(query_hca_vport_gid_in, in, other_vport, 1);
+		} else {
+			err = -EPERM;
+			goto out;
+		}
+	}
+	MLX5_SET(query_hca_vport_gid_in, in, gid_index, gid_index);
+
+	if (MLX5_CAP_GEN(dev, num_ports) == 2)
+		MLX5_SET(query_hca_vport_gid_in, in, port_num, port_num);
+
+	err = mlx5_cmd_exec(dev, in, in_sz, out, out_sz);
+	if (err)
+		goto out;
+
+	tmp = out + MLX5_ST_SZ_BYTES(query_hca_vport_gid_out);
+	gid->global.subnet_prefix = tmp->global.subnet_prefix;
+	gid->global.interface_id = tmp->global.interface_id;
+
+out:
+	kfree(in);
+	kfree(out);
+	return err;
+}
+EXPORT_SYMBOL_GPL(mlx5_query_hca_vport_gid);
+
+int mlx5_query_hca_vport_pkey(struct mlx5_core_dev *dev, u8 other_vport,
+			      u8 port_num, u16 vf_num, u16 pkey_index,
+			      u16 *pkey)
+{
+	int in_sz = MLX5_ST_SZ_BYTES(query_hca_vport_pkey_in);
+	int out_sz = MLX5_ST_SZ_BYTES(query_hca_vport_pkey_out);
+	int is_group_manager;
+	void *out = NULL;
+	void *in = NULL;
+	void *pkarr;
+	int nout;
+	int tbsz;
+	int err;
+	int i;
+
+	is_group_manager = MLX5_CAP_GEN(dev, vport_group_manager);
+
+	tbsz = mlx5_to_sw_pkey_sz(MLX5_CAP_GEN(dev, pkey_table_size));
+	if (pkey_index > tbsz && pkey_index != 0xffff)
+		return -EINVAL;
+
+	if (pkey_index == 0xffff)
+		nout = tbsz;
+	else
+		nout = 1;
+
+	out_sz += nout * MLX5_ST_SZ_BYTES(pkey);
+
+	in = kzalloc(in_sz, GFP_KERNEL);
+	out = kzalloc(out_sz, GFP_KERNEL);
+	if (!in || !out) {
+		err = -ENOMEM;
+		goto out;
+	}
+
+	MLX5_SET(query_hca_vport_pkey_in, in, opcode, MLX5_CMD_OP_QUERY_HCA_VPORT_PKEY);
+	if (other_vport) {
+		if (is_group_manager) {
+			MLX5_SET(query_hca_vport_pkey_in, in, vport_number, vf_num);
+			MLX5_SET(query_hca_vport_pkey_in, in, other_vport, 1);
+		} else {
+			err = -EPERM;
+			goto out;
+		}
+	}
+	MLX5_SET(query_hca_vport_pkey_in, in, pkey_index, pkey_index);
+
+	if (MLX5_CAP_GEN(dev, num_ports) == 2)
+		MLX5_SET(query_hca_vport_pkey_in, in, port_num, port_num);
+
+	err = mlx5_cmd_exec(dev, in, in_sz, out, out_sz);
+	if (err)
+		goto out;
+
+	pkarr = MLX5_ADDR_OF(query_hca_vport_pkey_out, out, pkey);
+	for (i = 0; i < nout; i++, pkey++, pkarr += MLX5_ST_SZ_BYTES(pkey))
+		*pkey = MLX5_GET_PR(pkey, pkarr, pkey);
+
+out:
+	kfree(in);
+	kfree(out);
+	return err;
+}
+EXPORT_SYMBOL_GPL(mlx5_query_hca_vport_pkey);
+
+int mlx5_query_hca_vport_context(struct mlx5_core_dev *dev,
+				 u8 other_vport, u8 port_num,
+				 u16 vf_num,
+				 struct mlx5_hca_vport_context *rep)
+{
+	int out_sz = MLX5_ST_SZ_BYTES(query_hca_vport_context_out);
+	int in[MLX5_ST_SZ_DW(query_hca_vport_context_in)] = {0};
+	int is_group_manager;
+	void *out;
+	void *ctx;
+	int err;
+
+	is_group_manager = MLX5_CAP_GEN(dev, vport_group_manager);
+
+	out = kzalloc(out_sz, GFP_KERNEL);
+	if (!out)
+		return -ENOMEM;
+
+	MLX5_SET(query_hca_vport_context_in, in, opcode, MLX5_CMD_OP_QUERY_HCA_VPORT_CONTEXT);
+
+	if (other_vport) {
+		if (is_group_manager) {
+			MLX5_SET(query_hca_vport_context_in, in, other_vport, 1);
+			MLX5_SET(query_hca_vport_context_in, in, vport_number, vf_num);
+		} else {
+			err = -EPERM;
+			goto ex;
+		}
+	}
+
+	if (MLX5_CAP_GEN(dev, num_ports) == 2)
+		MLX5_SET(query_hca_vport_context_in, in, port_num, port_num);
+
+	err = mlx5_cmd_exec(dev, in, sizeof(in), out,  out_sz);
+	if (err)
+		goto ex;
+
+	ctx = MLX5_ADDR_OF(query_hca_vport_context_out, out, hca_vport_context);
+	rep->field_select = MLX5_GET_PR(hca_vport_context, ctx, field_select);
+	rep->sm_virt_aware = MLX5_GET_PR(hca_vport_context, ctx, sm_virt_aware);
+	rep->has_smi = MLX5_GET_PR(hca_vport_context, ctx, has_smi);
+	rep->has_raw = MLX5_GET_PR(hca_vport_context, ctx, has_raw);
+	rep->policy = MLX5_GET_PR(hca_vport_context, ctx, vport_state_policy);
+	rep->phys_state = MLX5_GET_PR(hca_vport_context, ctx,
+				      port_physical_state);
+	rep->vport_state = MLX5_GET_PR(hca_vport_context, ctx, vport_state);
+	rep->port_physical_state = MLX5_GET_PR(hca_vport_context, ctx,
+					       port_physical_state);
+	rep->port_guid = MLX5_GET64_PR(hca_vport_context, ctx, port_guid);
+	rep->node_guid = MLX5_GET64_PR(hca_vport_context, ctx, node_guid);
+	rep->cap_mask1 = MLX5_GET_PR(hca_vport_context, ctx, cap_mask1);
+	rep->cap_mask1_perm = MLX5_GET_PR(hca_vport_context, ctx,
+					  cap_mask1_field_select);
+	rep->cap_mask2 = MLX5_GET_PR(hca_vport_context, ctx, cap_mask2);
+	rep->cap_mask2_perm = MLX5_GET_PR(hca_vport_context, ctx,
+					  cap_mask2_field_select);
+	rep->lid = MLX5_GET_PR(hca_vport_context, ctx, lid);
+	rep->init_type_reply = MLX5_GET_PR(hca_vport_context, ctx,
+					   init_type_reply);
+	rep->lmc = MLX5_GET_PR(hca_vport_context, ctx, lmc);
+	rep->subnet_timeout = MLX5_GET_PR(hca_vport_context, ctx,
+					  subnet_timeout);
+	rep->sm_lid = MLX5_GET_PR(hca_vport_context, ctx, sm_lid);
+	rep->sm_sl = MLX5_GET_PR(hca_vport_context, ctx, sm_sl);
+	rep->qkey_violation_counter = MLX5_GET_PR(hca_vport_context, ctx,
+						  qkey_violation_counter);
+	rep->pkey_violation_counter = MLX5_GET_PR(hca_vport_context, ctx,
+						  pkey_violation_counter);
+	rep->grh_required = MLX5_GET_PR(hca_vport_context, ctx, grh_required);
+	rep->sys_image_guid = MLX5_GET64_PR(hca_vport_context, ctx,
+					    system_image_guid);
+
+ex:
+	kfree(out);
+	return err;
+}
+EXPORT_SYMBOL_GPL(mlx5_query_hca_vport_context);
+
+int mlx5_query_hca_vport_system_image_guid(struct mlx5_core_dev *dev,
+					   u64 *sys_image_guid)
+{
+	struct mlx5_hca_vport_context *rep;
+	int err;
+
+	rep = kzalloc(sizeof(*rep), GFP_KERNEL);
+	if (!rep)
+		return -ENOMEM;
+
+	err = mlx5_query_hca_vport_context(dev, 0, 1, 0, rep);
+	if (!err)
+		*sys_image_guid = rep->sys_image_guid;
+
+	kfree(rep);
+	return err;
+}
+EXPORT_SYMBOL_GPL(mlx5_query_hca_vport_system_image_guid);
+
+int mlx5_query_hca_vport_node_guid(struct mlx5_core_dev *dev,
+				   u64 *node_guid)
+{
+	struct mlx5_hca_vport_context *rep;
+	int err;
+
+	rep = kzalloc(sizeof(*rep), GFP_KERNEL);
+	if (!rep)
+		return -ENOMEM;
+
+	err = mlx5_query_hca_vport_context(dev, 0, 1, 0, rep);
+	if (!err)
+		*node_guid = rep->node_guid;
+
+	kfree(rep);
+	return err;
+}
+EXPORT_SYMBOL_GPL(mlx5_query_hca_vport_node_guid);
+
+int mlx5_query_nic_vport_promisc(struct mlx5_core_dev *mdev,
+				 u32 vport,
+				 int *promisc_uc,
+				 int *promisc_mc,
+				 int *promisc_all)
+{
+	u32 *out;
+	int outlen = MLX5_ST_SZ_BYTES(query_nic_vport_context_out);
+	int err;
+
+	out = kzalloc(outlen, GFP_KERNEL);
+	if (!out)
+		return -ENOMEM;
+
+	err = mlx5_query_nic_vport_context(mdev, vport, out, outlen);
+	if (err)
+		goto out;
+
+	*promisc_uc = MLX5_GET(query_nic_vport_context_out, out,
+			       nic_vport_context.promisc_uc);
+	*promisc_mc = MLX5_GET(query_nic_vport_context_out, out,
+			       nic_vport_context.promisc_mc);
+	*promisc_all = MLX5_GET(query_nic_vport_context_out, out,
+				nic_vport_context.promisc_all);
+
+out:
+	kfree(out);
+	return err;
+}
+EXPORT_SYMBOL_GPL(mlx5_query_nic_vport_promisc);
+
+int mlx5_modify_nic_vport_promisc(struct mlx5_core_dev *mdev,
+				  int promisc_uc,
+				  int promisc_mc,
+				  int promisc_all)
+{
+	void *in;
+	int inlen = MLX5_ST_SZ_BYTES(modify_nic_vport_context_in);
+	int err;
+
+	in = kvzalloc(inlen, GFP_KERNEL);
+	if (!in)
+		return -ENOMEM;
+
+	MLX5_SET(modify_nic_vport_context_in, in, field_select.promisc, 1);
+	MLX5_SET(modify_nic_vport_context_in, in,
+		 nic_vport_context.promisc_uc, promisc_uc);
+	MLX5_SET(modify_nic_vport_context_in, in,
+		 nic_vport_context.promisc_mc, promisc_mc);
+	MLX5_SET(modify_nic_vport_context_in, in,
+		 nic_vport_context.promisc_all, promisc_all);
+
+	err = mlx5_modify_nic_vport_context(mdev, in, inlen);
+
+	kvfree(in);
+
+	return err;
+}
+EXPORT_SYMBOL_GPL(mlx5_modify_nic_vport_promisc);
+
+enum {
+	UC_LOCAL_LB,
+	MC_LOCAL_LB
+};
+
+int mlx5_nic_vport_update_local_lb(struct mlx5_core_dev *mdev, bool enable)
+{
+	int inlen = MLX5_ST_SZ_BYTES(modify_nic_vport_context_in);
+	void *in;
+	int err;
+
+	if (!MLX5_CAP_GEN(mdev, disable_local_lb_mc) &&
+	    !MLX5_CAP_GEN(mdev, disable_local_lb_uc))
+		return 0;
+
+	in = kvzalloc(inlen, GFP_KERNEL);
+	if (!in)
+		return -ENOMEM;
+
+	MLX5_SET(modify_nic_vport_context_in, in,
+		 nic_vport_context.disable_mc_local_lb, !enable);
+	MLX5_SET(modify_nic_vport_context_in, in,
+		 nic_vport_context.disable_uc_local_lb, !enable);
+
+	if (MLX5_CAP_GEN(mdev, disable_local_lb_mc))
+		MLX5_SET(modify_nic_vport_context_in, in,
+			 field_select.disable_mc_local_lb, 1);
+
+	if (MLX5_CAP_GEN(mdev, disable_local_lb_uc))
+		MLX5_SET(modify_nic_vport_context_in, in,
+			 field_select.disable_uc_local_lb, 1);
+
+	err = mlx5_modify_nic_vport_context(mdev, in, inlen);
+
+	if (!err)
+		mlx5_core_dbg(mdev, "%s local_lb\n",
+			      enable ? "enable" : "disable");
+
+	kvfree(in);
+	return err;
+}
+EXPORT_SYMBOL_GPL(mlx5_nic_vport_update_local_lb);
+
+int mlx5_nic_vport_query_local_lb(struct mlx5_core_dev *mdev, bool *status)
+{
+	int outlen = MLX5_ST_SZ_BYTES(query_nic_vport_context_out);
+	u32 *out;
+	int value;
+	int err;
+
+	out = kzalloc(outlen, GFP_KERNEL);
+	if (!out)
+		return -ENOMEM;
+
+	err = mlx5_query_nic_vport_context(mdev, 0, out, outlen);
+	if (err)
+		goto out;
+
+	value = MLX5_GET(query_nic_vport_context_out, out,
+			 nic_vport_context.disable_mc_local_lb) << MC_LOCAL_LB;
+
+	value |= MLX5_GET(query_nic_vport_context_out, out,
+			  nic_vport_context.disable_uc_local_lb) << UC_LOCAL_LB;
+
+	*status = !value;
+
+out:
+	kfree(out);
+	return err;
+}
+EXPORT_SYMBOL_GPL(mlx5_nic_vport_query_local_lb);
+
+enum mlx5_vport_roce_state {
+	MLX5_VPORT_ROCE_DISABLED = 0,
+	MLX5_VPORT_ROCE_ENABLED  = 1,
+};
+
+static int mlx5_nic_vport_update_roce_state(struct mlx5_core_dev *mdev,
+					    enum mlx5_vport_roce_state state)
+{
+	void *in;
+	int inlen = MLX5_ST_SZ_BYTES(modify_nic_vport_context_in);
+	int err;
+
+	in = kvzalloc(inlen, GFP_KERNEL);
+	if (!in)
+		return -ENOMEM;
+
+	MLX5_SET(modify_nic_vport_context_in, in, field_select.roce_en, 1);
+	MLX5_SET(modify_nic_vport_context_in, in, nic_vport_context.roce_en,
+		 state);
+
+	err = mlx5_modify_nic_vport_context(mdev, in, inlen);
+
+	kvfree(in);
+
+	return err;
+}
+
+int mlx5_nic_vport_enable_roce(struct mlx5_core_dev *mdev)
+{
+	int err = 0;
+
+	mutex_lock(&mlx5_roce_en_lock);
+	if (!mdev->roce.roce_en)
+		err = mlx5_nic_vport_update_roce_state(mdev, MLX5_VPORT_ROCE_ENABLED);
+
+	if (!err)
+		mdev->roce.roce_en++;
+	mutex_unlock(&mlx5_roce_en_lock);
+
+	return err;
+}
+EXPORT_SYMBOL_GPL(mlx5_nic_vport_enable_roce);
+
+int mlx5_nic_vport_disable_roce(struct mlx5_core_dev *mdev)
+{
+	int err = 0;
+
+	mutex_lock(&mlx5_roce_en_lock);
+	if (mdev->roce.roce_en) {
+		mdev->roce.roce_en--;
+		if (mdev->roce.roce_en == 0)
+			err = mlx5_nic_vport_update_roce_state(mdev, MLX5_VPORT_ROCE_DISABLED);
+
+		if (err)
+			mdev->roce.roce_en++;
+	}
+	mutex_unlock(&mlx5_roce_en_lock);
+	return err;
+}
+EXPORT_SYMBOL_GPL(mlx5_nic_vport_disable_roce);
+
+int mlx5_core_query_vport_counter(struct mlx5_core_dev *dev, u8 other_vport,
+				  int vf, u8 port_num, void *out,
+				  size_t out_sz)
+{
+	int	in_sz = MLX5_ST_SZ_BYTES(query_vport_counter_in);
+	int	is_group_manager;
+	void   *in;
+	int	err;
+
+	is_group_manager = MLX5_CAP_GEN(dev, vport_group_manager);
+	in = kvzalloc(in_sz, GFP_KERNEL);
+	if (!in) {
+		err = -ENOMEM;
+		return err;
+	}
+
+	MLX5_SET(query_vport_counter_in, in, opcode,
+		 MLX5_CMD_OP_QUERY_VPORT_COUNTER);
+	if (other_vport) {
+		if (is_group_manager) {
+			MLX5_SET(query_vport_counter_in, in, other_vport, 1);
+			MLX5_SET(query_vport_counter_in, in, vport_number, vf + 1);
+		} else {
+			err = -EPERM;
+			goto free;
+		}
+	}
+	if (MLX5_CAP_GEN(dev, num_ports) == 2)
+		MLX5_SET(query_vport_counter_in, in, port_num, port_num);
+
+	err = mlx5_cmd_exec(dev, in, in_sz, out,  out_sz);
+free:
+	kvfree(in);
+	return err;
+}
+EXPORT_SYMBOL_GPL(mlx5_core_query_vport_counter);
+
+int mlx5_query_vport_down_stats(struct mlx5_core_dev *mdev, u16 vport,
+				u64 *rx_discard_vport_down,
+				u64 *tx_discard_vport_down)
+{
+	u32 out[MLX5_ST_SZ_DW(query_vnic_env_out)] = {0};
+	u32 in[MLX5_ST_SZ_DW(query_vnic_env_in)] = {0};
+	int err;
+
+	MLX5_SET(query_vnic_env_in, in, opcode,
+		 MLX5_CMD_OP_QUERY_VNIC_ENV);
+	MLX5_SET(query_vnic_env_in, in, op_mod, 0);
+	MLX5_SET(query_vnic_env_in, in, vport_number, vport);
+	if (vport)
+		MLX5_SET(query_vnic_env_in, in, other_vport, 1);
+
+	err = mlx5_cmd_exec(mdev, in, sizeof(in), out, sizeof(out));
+	if (err)
+		return err;
+
+	*rx_discard_vport_down = MLX5_GET64(query_vnic_env_out, out,
+					    vport_env.receive_discard_vport_down);
+	*tx_discard_vport_down = MLX5_GET64(query_vnic_env_out, out,
+					    vport_env.transmit_discard_vport_down);
+	return 0;
+}
+
+int mlx5_core_modify_hca_vport_context(struct mlx5_core_dev *dev,
+				       u8 other_vport, u8 port_num,
+				       int vf,
+				       struct mlx5_hca_vport_context *req)
+{
+	int in_sz = MLX5_ST_SZ_BYTES(modify_hca_vport_context_in);
+	u8 out[MLX5_ST_SZ_BYTES(modify_hca_vport_context_out)];
+	int is_group_manager;
+	void *in;
+	int err;
+	void *ctx;
+
+	mlx5_core_dbg(dev, "vf %d\n", vf);
+	is_group_manager = MLX5_CAP_GEN(dev, vport_group_manager);
+	in = kzalloc(in_sz, GFP_KERNEL);
+	if (!in)
+		return -ENOMEM;
+
+	memset(out, 0, sizeof(out));
+	MLX5_SET(modify_hca_vport_context_in, in, opcode, MLX5_CMD_OP_MODIFY_HCA_VPORT_CONTEXT);
+	if (other_vport) {
+		if (is_group_manager) {
+			MLX5_SET(modify_hca_vport_context_in, in, other_vport, 1);
+			MLX5_SET(modify_hca_vport_context_in, in, vport_number, vf);
+		} else {
+			err = -EPERM;
+			goto ex;
+		}
+	}
+
+	if (MLX5_CAP_GEN(dev, num_ports) > 1)
+		MLX5_SET(modify_hca_vport_context_in, in, port_num, port_num);
+
+	ctx = MLX5_ADDR_OF(modify_hca_vport_context_in, in, hca_vport_context);
+	MLX5_SET(hca_vport_context, ctx, field_select, req->field_select);
+	MLX5_SET(hca_vport_context, ctx, sm_virt_aware, req->sm_virt_aware);
+	MLX5_SET(hca_vport_context, ctx, has_smi, req->has_smi);
+	MLX5_SET(hca_vport_context, ctx, has_raw, req->has_raw);
+	MLX5_SET(hca_vport_context, ctx, vport_state_policy, req->policy);
+	MLX5_SET(hca_vport_context, ctx, port_physical_state, req->phys_state);
+	MLX5_SET(hca_vport_context, ctx, vport_state, req->vport_state);
+	MLX5_SET64(hca_vport_context, ctx, port_guid, req->port_guid);
+	MLX5_SET64(hca_vport_context, ctx, node_guid, req->node_guid);
+	MLX5_SET(hca_vport_context, ctx, cap_mask1, req->cap_mask1);
+	MLX5_SET(hca_vport_context, ctx, cap_mask1_field_select, req->cap_mask1_perm);
+	MLX5_SET(hca_vport_context, ctx, cap_mask2, req->cap_mask2);
+	MLX5_SET(hca_vport_context, ctx, cap_mask2_field_select, req->cap_mask2_perm);
+	MLX5_SET(hca_vport_context, ctx, lid, req->lid);
+	MLX5_SET(hca_vport_context, ctx, init_type_reply, req->init_type_reply);
+	MLX5_SET(hca_vport_context, ctx, lmc, req->lmc);
+	MLX5_SET(hca_vport_context, ctx, subnet_timeout, req->subnet_timeout);
+	MLX5_SET(hca_vport_context, ctx, sm_lid, req->sm_lid);
+	MLX5_SET(hca_vport_context, ctx, sm_sl, req->sm_sl);
+	MLX5_SET(hca_vport_context, ctx, qkey_violation_counter, req->qkey_violation_counter);
+	MLX5_SET(hca_vport_context, ctx, pkey_violation_counter, req->pkey_violation_counter);
+	err = mlx5_cmd_exec(dev, in, in_sz, out, sizeof(out));
+ex:
+	kfree(in);
+	return err;
+}
+EXPORT_SYMBOL_GPL(mlx5_core_modify_hca_vport_context);
+
+int mlx5_nic_vport_affiliate_multiport(struct mlx5_core_dev *master_mdev,
+				       struct mlx5_core_dev *port_mdev)
+{
+	int inlen = MLX5_ST_SZ_BYTES(modify_nic_vport_context_in);
+	void *in;
+	int err;
+
+	in = kvzalloc(inlen, GFP_KERNEL);
+	if (!in)
+		return -ENOMEM;
+
+	err = mlx5_nic_vport_enable_roce(port_mdev);
+	if (err)
+		goto free;
+
+	MLX5_SET(modify_nic_vport_context_in, in, field_select.affiliation, 1);
+	MLX5_SET(modify_nic_vport_context_in, in,
+		 nic_vport_context.affiliated_vhca_id,
+		 MLX5_CAP_GEN(master_mdev, vhca_id));
+	MLX5_SET(modify_nic_vport_context_in, in,
+		 nic_vport_context.affiliation_criteria,
+		 MLX5_CAP_GEN(port_mdev, affiliate_nic_vport_criteria));
+
+	err = mlx5_modify_nic_vport_context(port_mdev, in, inlen);
+	if (err)
+		mlx5_nic_vport_disable_roce(port_mdev);
+
+free:
+	kvfree(in);
+	return err;
+}
+EXPORT_SYMBOL_GPL(mlx5_nic_vport_affiliate_multiport);
+
+int mlx5_nic_vport_unaffiliate_multiport(struct mlx5_core_dev *port_mdev)
+{
+	int inlen = MLX5_ST_SZ_BYTES(modify_nic_vport_context_in);
+	void *in;
+	int err;
+
+	in = kvzalloc(inlen, GFP_KERNEL);
+	if (!in)
+		return -ENOMEM;
+
+	MLX5_SET(modify_nic_vport_context_in, in, field_select.affiliation, 1);
+	MLX5_SET(modify_nic_vport_context_in, in,
+		 nic_vport_context.affiliated_vhca_id, 0);
+	MLX5_SET(modify_nic_vport_context_in, in,
+		 nic_vport_context.affiliation_criteria, 0);
+
+	err = mlx5_modify_nic_vport_context(port_mdev, in, inlen);
+	if (!err)
+		mlx5_nic_vport_disable_roce(port_mdev);
+
+	kvfree(in);
+	return err;
+}
+EXPORT_SYMBOL_GPL(mlx5_nic_vport_unaffiliate_multiport);
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/wq.c b/drivers/net/ethernet/mellanox/mlx5/core/wq.c
new file mode 100644
index 000000000..ddca327e8
--- /dev/null
+++ b/drivers/net/ethernet/mellanox/mlx5/core/wq.c
@@ -0,0 +1,267 @@
+/*
+ * Copyright (c) 2013-2015, Mellanox Technologies, Ltd.  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/mlx5/driver.h>
+#include "wq.h"
+#include "mlx5_core.h"
+
+u32 mlx5_wq_cyc_get_size(struct mlx5_wq_cyc *wq)
+{
+	return (u32)wq->fbc.sz_m1 + 1;
+}
+
+u32 mlx5_cqwq_get_size(struct mlx5_cqwq *wq)
+{
+	return wq->fbc.sz_m1 + 1;
+}
+
+u32 mlx5_wq_ll_get_size(struct mlx5_wq_ll *wq)
+{
+	return (u32)wq->fbc.sz_m1 + 1;
+}
+
+static u32 mlx5_wq_cyc_get_byte_size(struct mlx5_wq_cyc *wq)
+{
+	return mlx5_wq_cyc_get_size(wq) << wq->fbc.log_stride;
+}
+
+static u32 mlx5_wq_qp_get_byte_size(struct mlx5_wq_qp *wq)
+{
+	return mlx5_wq_cyc_get_byte_size(&wq->rq) +
+	       mlx5_wq_cyc_get_byte_size(&wq->sq);
+}
+
+static u32 mlx5_cqwq_get_byte_size(struct mlx5_cqwq *wq)
+{
+	return mlx5_cqwq_get_size(wq) << wq->fbc.log_stride;
+}
+
+static u32 mlx5_wq_ll_get_byte_size(struct mlx5_wq_ll *wq)
+{
+	return mlx5_wq_ll_get_size(wq) << wq->fbc.log_stride;
+}
+
+int mlx5_wq_cyc_create(struct mlx5_core_dev *mdev, struct mlx5_wq_param *param,
+		       void *wqc, struct mlx5_wq_cyc *wq,
+		       struct mlx5_wq_ctrl *wq_ctrl)
+{
+	struct mlx5_frag_buf_ctrl *fbc = &wq->fbc;
+	int err;
+
+	mlx5_fill_fbc(MLX5_GET(wq, wqc, log_wq_stride),
+		      MLX5_GET(wq, wqc, log_wq_sz),
+		      fbc);
+	wq->sz    = wq->fbc.sz_m1 + 1;
+
+	err = mlx5_db_alloc_node(mdev, &wq_ctrl->db, param->db_numa_node);
+	if (err) {
+		mlx5_core_warn(mdev, "mlx5_db_alloc_node() failed, %d\n", err);
+		return err;
+	}
+
+	err = mlx5_frag_buf_alloc_node(mdev, mlx5_wq_cyc_get_byte_size(wq),
+				       &wq_ctrl->buf, param->buf_numa_node);
+	if (err) {
+		mlx5_core_warn(mdev, "mlx5_frag_buf_alloc_node() failed, %d\n", err);
+		goto err_db_free;
+	}
+
+	fbc->frag_buf = wq_ctrl->buf;
+	wq->db  = wq_ctrl->db.db;
+
+	wq_ctrl->mdev = mdev;
+
+	return 0;
+
+err_db_free:
+	mlx5_db_free(mdev, &wq_ctrl->db);
+
+	return err;
+}
+
+static void mlx5_qp_set_frag_buf(struct mlx5_frag_buf *buf,
+				 struct mlx5_wq_qp *qp)
+{
+	struct mlx5_frag_buf_ctrl *sq_fbc;
+	struct mlx5_frag_buf *rqb, *sqb;
+
+	rqb  = &qp->rq.fbc.frag_buf;
+	*rqb = *buf;
+	rqb->size   = mlx5_wq_cyc_get_byte_size(&qp->rq);
+	rqb->npages = DIV_ROUND_UP(rqb->size, PAGE_SIZE);
+
+	sq_fbc = &qp->sq.fbc;
+	sqb    = &sq_fbc->frag_buf;
+	*sqb   = *buf;
+	sqb->size   = mlx5_wq_cyc_get_byte_size(&qp->sq);
+	sqb->npages = DIV_ROUND_UP(sqb->size, PAGE_SIZE);
+	sqb->frags += rqb->npages; /* first part is for the rq */
+	if (sq_fbc->strides_offset)
+		sqb->frags--;
+}
+
+int mlx5_wq_qp_create(struct mlx5_core_dev *mdev, struct mlx5_wq_param *param,
+		      void *qpc, struct mlx5_wq_qp *wq,
+		      struct mlx5_wq_ctrl *wq_ctrl)
+{
+	u16 sq_strides_offset;
+	u32 rq_pg_remainder;
+	int err;
+
+	mlx5_fill_fbc(MLX5_GET(qpc, qpc, log_rq_stride) + 4,
+		      MLX5_GET(qpc, qpc, log_rq_size),
+		      &wq->rq.fbc);
+
+	rq_pg_remainder   = mlx5_wq_cyc_get_byte_size(&wq->rq) % PAGE_SIZE;
+	sq_strides_offset = rq_pg_remainder / MLX5_SEND_WQE_BB;
+
+	mlx5_fill_fbc_offset(ilog2(MLX5_SEND_WQE_BB),
+			     MLX5_GET(qpc, qpc, log_sq_size),
+			     sq_strides_offset,
+			     &wq->sq.fbc);
+
+	err = mlx5_db_alloc_node(mdev, &wq_ctrl->db, param->db_numa_node);
+	if (err) {
+		mlx5_core_warn(mdev, "mlx5_db_alloc_node() failed, %d\n", err);
+		return err;
+	}
+
+	err = mlx5_frag_buf_alloc_node(mdev, mlx5_wq_qp_get_byte_size(wq),
+				       &wq_ctrl->buf, param->buf_numa_node);
+	if (err) {
+		mlx5_core_warn(mdev, "mlx5_frag_buf_alloc_node() failed, %d\n", err);
+		goto err_db_free;
+	}
+
+	mlx5_qp_set_frag_buf(&wq_ctrl->buf, wq);
+
+	wq->rq.db  = &wq_ctrl->db.db[MLX5_RCV_DBR];
+	wq->sq.db  = &wq_ctrl->db.db[MLX5_SND_DBR];
+
+	wq_ctrl->mdev = mdev;
+
+	return 0;
+
+err_db_free:
+	mlx5_db_free(mdev, &wq_ctrl->db);
+
+	return err;
+}
+
+int mlx5_cqwq_create(struct mlx5_core_dev *mdev, struct mlx5_wq_param *param,
+		     void *cqc, struct mlx5_cqwq *wq,
+		     struct mlx5_wq_ctrl *wq_ctrl)
+{
+	int err;
+
+	mlx5_core_init_cq_frag_buf(&wq->fbc, cqc);
+
+	err = mlx5_db_alloc_node(mdev, &wq_ctrl->db, param->db_numa_node);
+	if (err) {
+		mlx5_core_warn(mdev, "mlx5_db_alloc_node() failed, %d\n", err);
+		return err;
+	}
+
+	err = mlx5_frag_buf_alloc_node(mdev, mlx5_cqwq_get_byte_size(wq),
+				       &wq_ctrl->buf,
+				       param->buf_numa_node);
+	if (err) {
+		mlx5_core_warn(mdev, "mlx5_frag_buf_alloc_node() failed, %d\n",
+			       err);
+		goto err_db_free;
+	}
+
+	wq->fbc.frag_buf = wq_ctrl->buf;
+	wq->db  = wq_ctrl->db.db;
+
+	wq_ctrl->mdev = mdev;
+
+	return 0;
+
+err_db_free:
+	mlx5_db_free(mdev, &wq_ctrl->db);
+
+	return err;
+}
+
+int mlx5_wq_ll_create(struct mlx5_core_dev *mdev, struct mlx5_wq_param *param,
+		      void *wqc, struct mlx5_wq_ll *wq,
+		      struct mlx5_wq_ctrl *wq_ctrl)
+{
+	struct mlx5_frag_buf_ctrl *fbc = &wq->fbc;
+	struct mlx5_wqe_srq_next_seg *next_seg;
+	int err;
+	int i;
+
+	mlx5_fill_fbc(MLX5_GET(wq, wqc, log_wq_stride),
+		      MLX5_GET(wq, wqc, log_wq_sz),
+		      fbc);
+
+	err = mlx5_db_alloc_node(mdev, &wq_ctrl->db, param->db_numa_node);
+	if (err) {
+		mlx5_core_warn(mdev, "mlx5_db_alloc_node() failed, %d\n", err);
+		return err;
+	}
+
+	err = mlx5_frag_buf_alloc_node(mdev, mlx5_wq_ll_get_byte_size(wq),
+				       &wq_ctrl->buf, param->buf_numa_node);
+	if (err) {
+		mlx5_core_warn(mdev, "mlx5_frag_buf_alloc_node() failed, %d\n", err);
+		goto err_db_free;
+	}
+
+	wq->fbc.frag_buf = wq_ctrl->buf;
+	wq->db  = wq_ctrl->db.db;
+
+	for (i = 0; i < fbc->sz_m1; i++) {
+		next_seg = mlx5_wq_ll_get_wqe(wq, i);
+		next_seg->next_wqe_index = cpu_to_be16(i + 1);
+	}
+	next_seg = mlx5_wq_ll_get_wqe(wq, i);
+	wq->tail_next = &next_seg->next_wqe_index;
+
+	wq_ctrl->mdev = mdev;
+
+	return 0;
+
+err_db_free:
+	mlx5_db_free(mdev, &wq_ctrl->db);
+
+	return err;
+}
+
+void mlx5_wq_destroy(struct mlx5_wq_ctrl *wq_ctrl)
+{
+	mlx5_frag_buf_free(wq_ctrl->mdev, &wq_ctrl->buf);
+	mlx5_db_free(wq_ctrl->mdev, &wq_ctrl->db);
+}
+
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/wq.h b/drivers/net/ethernet/mellanox/mlx5/core/wq.h
new file mode 100644
index 000000000..b1293d153
--- /dev/null
+++ b/drivers/net/ethernet/mellanox/mlx5/core/wq.h
@@ -0,0 +1,261 @@
+/*
+ * Copyright (c) 2013-2015, Mellanox Technologies, Ltd.  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 __MLX5_WQ_H__
+#define __MLX5_WQ_H__
+
+#include <linux/mlx5/mlx5_ifc.h>
+#include <linux/mlx5/cq.h>
+#include <linux/mlx5/qp.h>
+
+struct mlx5_wq_param {
+	int		buf_numa_node;
+	int		db_numa_node;
+};
+
+struct mlx5_wq_ctrl {
+	struct mlx5_core_dev	*mdev;
+	struct mlx5_frag_buf	buf;
+	struct mlx5_db		db;
+};
+
+struct mlx5_wq_cyc {
+	struct mlx5_frag_buf_ctrl fbc;
+	__be32			*db;
+	u16			sz;
+	u16			wqe_ctr;
+	u16			cur_sz;
+};
+
+struct mlx5_wq_qp {
+	struct mlx5_wq_cyc	rq;
+	struct mlx5_wq_cyc	sq;
+};
+
+struct mlx5_cqwq {
+	struct mlx5_frag_buf_ctrl fbc;
+	__be32			  *db;
+	u32			  cc; /* consumer counter */
+};
+
+struct mlx5_wq_ll {
+	struct mlx5_frag_buf_ctrl fbc;
+	__be32			*db;
+	__be16			*tail_next;
+	u16			head;
+	u16			wqe_ctr;
+	u16			cur_sz;
+};
+
+int mlx5_wq_cyc_create(struct mlx5_core_dev *mdev, struct mlx5_wq_param *param,
+		       void *wqc, struct mlx5_wq_cyc *wq,
+		       struct mlx5_wq_ctrl *wq_ctrl);
+u32 mlx5_wq_cyc_get_size(struct mlx5_wq_cyc *wq);
+
+int mlx5_wq_qp_create(struct mlx5_core_dev *mdev, struct mlx5_wq_param *param,
+		      void *qpc, struct mlx5_wq_qp *wq,
+		      struct mlx5_wq_ctrl *wq_ctrl);
+
+int mlx5_cqwq_create(struct mlx5_core_dev *mdev, struct mlx5_wq_param *param,
+		     void *cqc, struct mlx5_cqwq *wq,
+		     struct mlx5_wq_ctrl *wq_ctrl);
+u32 mlx5_cqwq_get_size(struct mlx5_cqwq *wq);
+
+int mlx5_wq_ll_create(struct mlx5_core_dev *mdev, struct mlx5_wq_param *param,
+		      void *wqc, struct mlx5_wq_ll *wq,
+		      struct mlx5_wq_ctrl *wq_ctrl);
+u32 mlx5_wq_ll_get_size(struct mlx5_wq_ll *wq);
+
+void mlx5_wq_destroy(struct mlx5_wq_ctrl *wq_ctrl);
+
+static inline int mlx5_wq_cyc_is_full(struct mlx5_wq_cyc *wq)
+{
+	return wq->cur_sz == wq->sz;
+}
+
+static inline int mlx5_wq_cyc_missing(struct mlx5_wq_cyc *wq)
+{
+	return wq->sz - wq->cur_sz;
+}
+
+static inline int mlx5_wq_cyc_is_empty(struct mlx5_wq_cyc *wq)
+{
+	return !wq->cur_sz;
+}
+
+static inline void mlx5_wq_cyc_push(struct mlx5_wq_cyc *wq)
+{
+	wq->wqe_ctr++;
+	wq->cur_sz++;
+}
+
+static inline void mlx5_wq_cyc_push_n(struct mlx5_wq_cyc *wq, u8 n)
+{
+	wq->wqe_ctr += n;
+	wq->cur_sz += n;
+}
+
+static inline void mlx5_wq_cyc_pop(struct mlx5_wq_cyc *wq)
+{
+	wq->cur_sz--;
+}
+
+static inline void mlx5_wq_cyc_update_db_record(struct mlx5_wq_cyc *wq)
+{
+	*wq->db = cpu_to_be32(wq->wqe_ctr);
+}
+
+static inline u16 mlx5_wq_cyc_ctr2ix(struct mlx5_wq_cyc *wq, u16 ctr)
+{
+	return ctr & wq->fbc.sz_m1;
+}
+
+static inline u16 mlx5_wq_cyc_get_head(struct mlx5_wq_cyc *wq)
+{
+	return mlx5_wq_cyc_ctr2ix(wq, wq->wqe_ctr);
+}
+
+static inline u16 mlx5_wq_cyc_get_tail(struct mlx5_wq_cyc *wq)
+{
+	return mlx5_wq_cyc_ctr2ix(wq, wq->wqe_ctr - wq->cur_sz);
+}
+
+static inline void *mlx5_wq_cyc_get_wqe(struct mlx5_wq_cyc *wq, u16 ix)
+{
+	return mlx5_frag_buf_get_wqe(&wq->fbc, ix);
+}
+
+static inline u16 mlx5_wq_cyc_get_contig_wqebbs(struct mlx5_wq_cyc *wq, u16 ix)
+{
+	return mlx5_frag_buf_get_idx_last_contig_stride(&wq->fbc, ix) - ix + 1;
+}
+
+static inline int mlx5_wq_cyc_cc_bigger(u16 cc1, u16 cc2)
+{
+	int equal   = (cc1 == cc2);
+	int smaller = 0x8000 & (cc1 - cc2);
+
+	return !equal && !smaller;
+}
+
+static inline u32 mlx5_cqwq_ctr2ix(struct mlx5_cqwq *wq, u32 ctr)
+{
+	return ctr & wq->fbc.sz_m1;
+}
+
+static inline u32 mlx5_cqwq_get_ci(struct mlx5_cqwq *wq)
+{
+	return mlx5_cqwq_ctr2ix(wq, wq->cc);
+}
+
+static inline void *mlx5_cqwq_get_wqe(struct mlx5_cqwq *wq, u32 ix)
+{
+	return mlx5_frag_buf_get_wqe(&wq->fbc, ix);
+}
+
+static inline u32 mlx5_cqwq_get_ctr_wrap_cnt(struct mlx5_cqwq *wq, u32 ctr)
+{
+	return ctr >> wq->fbc.log_sz;
+}
+
+static inline u32 mlx5_cqwq_get_wrap_cnt(struct mlx5_cqwq *wq)
+{
+	return mlx5_cqwq_get_ctr_wrap_cnt(wq, wq->cc);
+}
+
+static inline void mlx5_cqwq_pop(struct mlx5_cqwq *wq)
+{
+	wq->cc++;
+}
+
+static inline void mlx5_cqwq_update_db_record(struct mlx5_cqwq *wq)
+{
+	*wq->db = cpu_to_be32(wq->cc & 0xffffff);
+}
+
+static inline struct mlx5_cqe64 *mlx5_cqwq_get_cqe(struct mlx5_cqwq *wq)
+{
+	u32 ci = mlx5_cqwq_get_ci(wq);
+	struct mlx5_cqe64 *cqe = mlx5_cqwq_get_wqe(wq, ci);
+	u8 cqe_ownership_bit = cqe->op_own & MLX5_CQE_OWNER_MASK;
+	u8 sw_ownership_val = mlx5_cqwq_get_wrap_cnt(wq) & 1;
+
+	if (cqe_ownership_bit != sw_ownership_val)
+		return NULL;
+
+	/* ensure cqe content is read after cqe ownership bit */
+	dma_rmb();
+
+	return cqe;
+}
+
+static inline int mlx5_wq_ll_is_full(struct mlx5_wq_ll *wq)
+{
+	return wq->cur_sz == wq->fbc.sz_m1;
+}
+
+static inline int mlx5_wq_ll_is_empty(struct mlx5_wq_ll *wq)
+{
+	return !wq->cur_sz;
+}
+
+static inline int mlx5_wq_ll_missing(struct mlx5_wq_ll *wq)
+{
+	return wq->fbc.sz_m1 - wq->cur_sz;
+}
+
+static inline void *mlx5_wq_ll_get_wqe(struct mlx5_wq_ll *wq, u16 ix)
+{
+	return mlx5_frag_buf_get_wqe(&wq->fbc, ix);
+}
+
+static inline void mlx5_wq_ll_push(struct mlx5_wq_ll *wq, u16 head_next)
+{
+	wq->head = head_next;
+	wq->wqe_ctr++;
+	wq->cur_sz++;
+}
+
+static inline void mlx5_wq_ll_pop(struct mlx5_wq_ll *wq, __be16 ix,
+				  __be16 *next_tail_next)
+{
+	*wq->tail_next = ix;
+	wq->tail_next = next_tail_next;
+	wq->cur_sz--;
+}
+
+static inline void mlx5_wq_ll_update_db_record(struct mlx5_wq_ll *wq)
+{
+	*wq->db = cpu_to_be32(wq->wqe_ctr);
+}
+
+#endif /* __MLX5_WQ_H__ */