xref: /openbmc/linux/drivers/infiniband/hw/mlx5/main.c (revision add08d765e942eab8eb15a592baeb372a3dd6831)

/*
 * 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/sched.h>
#include <rdma/ib_user_verbs.h>
#include <rdma/ib_addr.h>
#include <rdma/ib_cache.h>
#include <linux/mlx5/vport.h>
#include <rdma/ib_smi.h>
#include <rdma/ib_umem.h>
#include <linux/in.h>
#include <linux/etherdevice.h>
#include <linux/mlx5/fs.h>
#include "user.h"
#include "mlx5_ib.h"

#define DRIVER_NAME "mlx5_ib"
#define DRIVER_VERSION "2.2-1"
#define DRIVER_RELDATE	"Feb 2014"

MODULE_AUTHOR("Eli Cohen <eli@mellanox.com>");
MODULE_DESCRIPTION("Mellanox Connect-IB HCA IB driver");
MODULE_LICENSE("Dual BSD/GPL");
MODULE_VERSION(DRIVER_VERSION);

static int deprecated_prof_sel = 2;
module_param_named(prof_sel, deprecated_prof_sel, int, 0444);
MODULE_PARM_DESC(prof_sel, "profile selector. Deprecated here. Moved to module mlx5_core");

static char mlx5_version[] =
	DRIVER_NAME ": Mellanox Connect-IB Infiniband driver v"
	DRIVER_VERSION " (" DRIVER_RELDATE ")\n";

enum {
	MLX5_ATOMIC_SIZE_QP_8BYTES = 1 << 3,
};

static enum rdma_link_layer
mlx5_port_type_cap_to_rdma_ll(int port_type_cap)
{
	switch (port_type_cap) {
	case MLX5_CAP_PORT_TYPE_IB:
		return IB_LINK_LAYER_INFINIBAND;
	case MLX5_CAP_PORT_TYPE_ETH:
		return IB_LINK_LAYER_ETHERNET;
	default:
		return IB_LINK_LAYER_UNSPECIFIED;
	}
}

static enum rdma_link_layer
mlx5_ib_port_link_layer(struct ib_device *device, u8 port_num)
{
	struct mlx5_ib_dev *dev = to_mdev(device);
	int port_type_cap = MLX5_CAP_GEN(dev->mdev, port_type);

	return mlx5_port_type_cap_to_rdma_ll(port_type_cap);
}

static int mlx5_netdev_event(struct notifier_block *this,
			     unsigned long event, void *ptr)
{
	struct net_device *ndev = netdev_notifier_info_to_dev(ptr);
	struct mlx5_ib_dev *ibdev = container_of(this, struct mlx5_ib_dev,
						 roce.nb);

	if ((event != NETDEV_UNREGISTER) && (event != NETDEV_REGISTER))
		return NOTIFY_DONE;

	write_lock(&ibdev->roce.netdev_lock);
	if (ndev->dev.parent == &ibdev->mdev->pdev->dev)
		ibdev->roce.netdev = (event == NETDEV_UNREGISTER) ? NULL : ndev;
	write_unlock(&ibdev->roce.netdev_lock);

	return NOTIFY_DONE;
}

static struct net_device *mlx5_ib_get_netdev(struct ib_device *device,
					     u8 port_num)
{
	struct mlx5_ib_dev *ibdev = to_mdev(device);
	struct net_device *ndev;

	/* Ensure ndev does not disappear before we invoke dev_hold()
	 */
	read_lock(&ibdev->roce.netdev_lock);
	ndev = ibdev->roce.netdev;
	if (ndev)
		dev_hold(ndev);
	read_unlock(&ibdev->roce.netdev_lock);

	return ndev;
}

static int mlx5_query_port_roce(struct ib_device *device, u8 port_num,
				struct ib_port_attr *props)
{
	struct mlx5_ib_dev *dev = to_mdev(device);
	struct net_device *ndev;
	enum ib_mtu ndev_ib_mtu;
	u16 qkey_viol_cntr;

	memset(props, 0, sizeof(*props));

	props->port_cap_flags  |= IB_PORT_CM_SUP;
	props->port_cap_flags  |= IB_PORT_IP_BASED_GIDS;

	props->gid_tbl_len      = MLX5_CAP_ROCE(dev->mdev,
						roce_address_table_size);
	props->max_mtu          = IB_MTU_4096;
	props->max_msg_sz       = 1 << MLX5_CAP_GEN(dev->mdev, log_max_msg);
	props->pkey_tbl_len     = 1;
	props->state            = IB_PORT_DOWN;
	props->phys_state       = 3;

	mlx5_query_nic_vport_qkey_viol_cntr(dev->mdev, &qkey_viol_cntr);
	props->qkey_viol_cntr = qkey_viol_cntr;

	ndev = mlx5_ib_get_netdev(device, port_num);
	if (!ndev)
		return 0;

	if (netif_running(ndev) && netif_carrier_ok(ndev)) {
		props->state      = IB_PORT_ACTIVE;
		props->phys_state = 5;
	}

	ndev_ib_mtu = iboe_get_mtu(ndev->mtu);

	dev_put(ndev);

	props->active_mtu	= min(props->max_mtu, ndev_ib_mtu);

	props->active_width	= IB_WIDTH_4X;  /* TODO */
	props->active_speed	= IB_SPEED_QDR; /* TODO */

	return 0;
}

static void ib_gid_to_mlx5_roce_addr(const union ib_gid *gid,
				     const struct ib_gid_attr *attr,
				     void *mlx5_addr)
{
#define MLX5_SET_RA(p, f, v) MLX5_SET(roce_addr_layout, p, f, v)
	char *mlx5_addr_l3_addr	= MLX5_ADDR_OF(roce_addr_layout, mlx5_addr,
					       source_l3_address);
	void *mlx5_addr_mac	= MLX5_ADDR_OF(roce_addr_layout, mlx5_addr,
					       source_mac_47_32);

	if (!gid)
		return;

	ether_addr_copy(mlx5_addr_mac, attr->ndev->dev_addr);

	if (is_vlan_dev(attr->ndev)) {
		MLX5_SET_RA(mlx5_addr, vlan_valid, 1);
		MLX5_SET_RA(mlx5_addr, vlan_id, vlan_dev_vlan_id(attr->ndev));
	}

	switch (attr->gid_type) {
	case IB_GID_TYPE_IB:
		MLX5_SET_RA(mlx5_addr, roce_version, MLX5_ROCE_VERSION_1);
		break;
	case IB_GID_TYPE_ROCE_UDP_ENCAP:
		MLX5_SET_RA(mlx5_addr, roce_version, MLX5_ROCE_VERSION_2);
		break;

	default:
		WARN_ON(true);
	}

	if (attr->gid_type != IB_GID_TYPE_IB) {
		if (ipv6_addr_v4mapped((void *)gid))
			MLX5_SET_RA(mlx5_addr, roce_l3_type,
				    MLX5_ROCE_L3_TYPE_IPV4);
		else
			MLX5_SET_RA(mlx5_addr, roce_l3_type,
				    MLX5_ROCE_L3_TYPE_IPV6);
	}

	if ((attr->gid_type == IB_GID_TYPE_IB) ||
	    !ipv6_addr_v4mapped((void *)gid))
		memcpy(mlx5_addr_l3_addr, gid, sizeof(*gid));
	else
		memcpy(&mlx5_addr_l3_addr[12], &gid->raw[12], 4);
}

static int set_roce_addr(struct ib_device *device, u8 port_num,
			 unsigned int index,
			 const union ib_gid *gid,
			 const struct ib_gid_attr *attr)
{
	struct mlx5_ib_dev *dev	= to_mdev(device);
	u32  in[MLX5_ST_SZ_DW(set_roce_address_in)];
	u32 out[MLX5_ST_SZ_DW(set_roce_address_out)];
	void *in_addr = MLX5_ADDR_OF(set_roce_address_in, in, roce_address);
	enum rdma_link_layer ll = mlx5_ib_port_link_layer(device, port_num);

	if (ll != IB_LINK_LAYER_ETHERNET)
		return -EINVAL;

	memset(in, 0, sizeof(in));

	ib_gid_to_mlx5_roce_addr(gid, attr, in_addr);

	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);

	memset(out, 0, sizeof(out));
	return mlx5_cmd_exec(dev->mdev, in, sizeof(in), out, sizeof(out));
}

static int mlx5_ib_add_gid(struct ib_device *device, u8 port_num,
			   unsigned int index, const union ib_gid *gid,
			   const struct ib_gid_attr *attr,
			   __always_unused void **context)
{
	return set_roce_addr(device, port_num, index, gid, attr);
}

static int mlx5_ib_del_gid(struct ib_device *device, u8 port_num,
			   unsigned int index, __always_unused void **context)
{
	return set_roce_addr(device, port_num, index, NULL, NULL);
}

__be16 mlx5_get_roce_udp_sport(struct mlx5_ib_dev *dev, u8 port_num,
			       int index)
{
	struct ib_gid_attr attr;
	union ib_gid gid;

	if (ib_get_cached_gid(&dev->ib_dev, port_num, index, &gid, &attr))
		return 0;

	if (!attr.ndev)
		return 0;

	dev_put(attr.ndev);

	if (attr.gid_type != IB_GID_TYPE_ROCE_UDP_ENCAP)
		return 0;

	return cpu_to_be16(MLX5_CAP_ROCE(dev->mdev, r_roce_min_src_udp_port));
}

static int mlx5_use_mad_ifc(struct mlx5_ib_dev *dev)
{
	return !dev->mdev->issi;
}

enum {
	MLX5_VPORT_ACCESS_METHOD_MAD,
	MLX5_VPORT_ACCESS_METHOD_HCA,
	MLX5_VPORT_ACCESS_METHOD_NIC,
};

static int mlx5_get_vport_access_method(struct ib_device *ibdev)
{
	if (mlx5_use_mad_ifc(to_mdev(ibdev)))
		return MLX5_VPORT_ACCESS_METHOD_MAD;

	if (mlx5_ib_port_link_layer(ibdev, 1) ==
	    IB_LINK_LAYER_ETHERNET)
		return MLX5_VPORT_ACCESS_METHOD_NIC;

	return MLX5_VPORT_ACCESS_METHOD_HCA;
}

static void get_atomic_caps(struct mlx5_ib_dev *dev,
			    struct ib_device_attr *props)
{
	u8 tmp;
	u8 atomic_operations = MLX5_CAP_ATOMIC(dev->mdev, atomic_operations);
	u8 atomic_size_qp = MLX5_CAP_ATOMIC(dev->mdev, atomic_size_qp);
	u8 atomic_req_8B_endianness_mode =
		MLX5_CAP_ATOMIC(dev->mdev, atomic_req_8B_endianess_mode);

	/* Check if HW supports 8 bytes standard atomic operations and capable
	 * of host endianness respond
	 */
	tmp = MLX5_ATOMIC_OPS_CMP_SWAP | MLX5_ATOMIC_OPS_FETCH_ADD;
	if (((atomic_operations & tmp) == tmp) &&
	    (atomic_size_qp & MLX5_ATOMIC_SIZE_QP_8BYTES) &&
	    (atomic_req_8B_endianness_mode)) {
		props->atomic_cap = IB_ATOMIC_HCA;
	} else {
		props->atomic_cap = IB_ATOMIC_NONE;
	}
}

static int mlx5_query_system_image_guid(struct ib_device *ibdev,
					__be64 *sys_image_guid)
{
	struct mlx5_ib_dev *dev = to_mdev(ibdev);
	struct mlx5_core_dev *mdev = dev->mdev;
	u64 tmp;
	int err;

	switch (mlx5_get_vport_access_method(ibdev)) {
	case MLX5_VPORT_ACCESS_METHOD_MAD:
		return mlx5_query_mad_ifc_system_image_guid(ibdev,
							    sys_image_guid);

	case MLX5_VPORT_ACCESS_METHOD_HCA:
		err = mlx5_query_hca_vport_system_image_guid(mdev, &tmp);
		break;

	case MLX5_VPORT_ACCESS_METHOD_NIC:
		err = mlx5_query_nic_vport_system_image_guid(mdev, &tmp);
		break;

	default:
		return -EINVAL;
	}

	if (!err)
		*sys_image_guid = cpu_to_be64(tmp);

	return err;

}

static int mlx5_query_max_pkeys(struct ib_device *ibdev,
				u16 *max_pkeys)
{
	struct mlx5_ib_dev *dev = to_mdev(ibdev);
	struct mlx5_core_dev *mdev = dev->mdev;

	switch (mlx5_get_vport_access_method(ibdev)) {
	case MLX5_VPORT_ACCESS_METHOD_MAD:
		return mlx5_query_mad_ifc_max_pkeys(ibdev, max_pkeys);

	case MLX5_VPORT_ACCESS_METHOD_HCA:
	case MLX5_VPORT_ACCESS_METHOD_NIC:
		*max_pkeys = mlx5_to_sw_pkey_sz(MLX5_CAP_GEN(mdev,
						pkey_table_size));
		return 0;

	default:
		return -EINVAL;
	}
}

static int mlx5_query_vendor_id(struct ib_device *ibdev,
				u32 *vendor_id)
{
	struct mlx5_ib_dev *dev = to_mdev(ibdev);

	switch (mlx5_get_vport_access_method(ibdev)) {
	case MLX5_VPORT_ACCESS_METHOD_MAD:
		return mlx5_query_mad_ifc_vendor_id(ibdev, vendor_id);

	case MLX5_VPORT_ACCESS_METHOD_HCA:
	case MLX5_VPORT_ACCESS_METHOD_NIC:
		return mlx5_core_query_vendor_id(dev->mdev, vendor_id);

	default:
		return -EINVAL;
	}
}

static int mlx5_query_node_guid(struct mlx5_ib_dev *dev,
				__be64 *node_guid)
{
	u64 tmp;
	int err;

	switch (mlx5_get_vport_access_method(&dev->ib_dev)) {
	case MLX5_VPORT_ACCESS_METHOD_MAD:
		return mlx5_query_mad_ifc_node_guid(dev, node_guid);

	case MLX5_VPORT_ACCESS_METHOD_HCA:
		err = mlx5_query_hca_vport_node_guid(dev->mdev, &tmp);
		break;

	case MLX5_VPORT_ACCESS_METHOD_NIC:
		err = mlx5_query_nic_vport_node_guid(dev->mdev, &tmp);
		break;

	default:
		return -EINVAL;
	}

	if (!err)
		*node_guid = cpu_to_be64(tmp);

	return err;
}

struct mlx5_reg_node_desc {
	u8	desc[64];
};

static int mlx5_query_node_desc(struct mlx5_ib_dev *dev, char *node_desc)
{
	struct mlx5_reg_node_desc in;

	if (mlx5_use_mad_ifc(dev))
		return mlx5_query_mad_ifc_node_desc(dev, node_desc);

	memset(&in, 0, sizeof(in));

	return mlx5_core_access_reg(dev->mdev, &in, sizeof(in), node_desc,
				    sizeof(struct mlx5_reg_node_desc),
				    MLX5_REG_NODE_DESC, 0, 0);
}

static int mlx5_ib_query_device(struct ib_device *ibdev,
				struct ib_device_attr *props,
				struct ib_udata *uhw)
{
	struct mlx5_ib_dev *dev = to_mdev(ibdev);
	struct mlx5_core_dev *mdev = dev->mdev;
	int err = -ENOMEM;
	int max_rq_sg;
	int max_sq_sg;
	u64 min_page_size = 1ull << MLX5_CAP_GEN(mdev, log_pg_sz);

	if (uhw->inlen || uhw->outlen)
		return -EINVAL;

	memset(props, 0, sizeof(*props));
	err = mlx5_query_system_image_guid(ibdev,
					   &props->sys_image_guid);
	if (err)
		return err;

	err = mlx5_query_max_pkeys(ibdev, &props->max_pkeys);
	if (err)
		return err;

	err = mlx5_query_vendor_id(ibdev, &props->vendor_id);
	if (err)
		return err;

	props->fw_ver = ((u64)fw_rev_maj(dev->mdev) << 32) |
		(fw_rev_min(dev->mdev) << 16) |
		fw_rev_sub(dev->mdev);
	props->device_cap_flags    = IB_DEVICE_CHANGE_PHY_PORT |
		IB_DEVICE_PORT_ACTIVE_EVENT		|
		IB_DEVICE_SYS_IMAGE_GUID		|
		IB_DEVICE_RC_RNR_NAK_GEN;

	if (MLX5_CAP_GEN(mdev, pkv))
		props->device_cap_flags |= IB_DEVICE_BAD_PKEY_CNTR;
	if (MLX5_CAP_GEN(mdev, qkv))
		props->device_cap_flags |= IB_DEVICE_BAD_QKEY_CNTR;
	if (MLX5_CAP_GEN(mdev, apm))
		props->device_cap_flags |= IB_DEVICE_AUTO_PATH_MIG;
	if (MLX5_CAP_GEN(mdev, xrc))
		props->device_cap_flags |= IB_DEVICE_XRC;
	if (MLX5_CAP_GEN(mdev, imaicl)) {
		props->device_cap_flags |= IB_DEVICE_MEM_WINDOW |
					   IB_DEVICE_MEM_WINDOW_TYPE_2B;
		props->max_mw = 1 << MLX5_CAP_GEN(mdev, log_max_mkey);
	}
	props->device_cap_flags |= IB_DEVICE_MEM_MGT_EXTENSIONS;
	if (MLX5_CAP_GEN(mdev, sho)) {
		props->device_cap_flags |= IB_DEVICE_SIGNATURE_HANDOVER;
		/* At this stage no support for signature handover */
		props->sig_prot_cap = IB_PROT_T10DIF_TYPE_1 |
				      IB_PROT_T10DIF_TYPE_2 |
				      IB_PROT_T10DIF_TYPE_3;
		props->sig_guard_cap = IB_GUARD_T10DIF_CRC |
				       IB_GUARD_T10DIF_CSUM;
	}
	if (MLX5_CAP_GEN(mdev, block_lb_mc))
		props->device_cap_flags |= IB_DEVICE_BLOCK_MULTICAST_LOOPBACK;

	if (MLX5_CAP_GEN(dev->mdev, eth_net_offloads) &&
	    (MLX5_CAP_ETH(dev->mdev, csum_cap)))
			props->device_cap_flags |= IB_DEVICE_RAW_IP_CSUM;

	if (MLX5_CAP_GEN(mdev, ipoib_basic_offloads)) {
		props->device_cap_flags |= IB_DEVICE_UD_IP_CSUM;
		props->device_cap_flags |= IB_DEVICE_UD_TSO;
	}

	props->vendor_part_id	   = mdev->pdev->device;
	props->hw_ver		   = mdev->pdev->revision;

	props->max_mr_size	   = ~0ull;
	props->page_size_cap	   = ~(min_page_size - 1);
	props->max_qp		   = 1 << MLX5_CAP_GEN(mdev, log_max_qp);
	props->max_qp_wr	   = 1 << MLX5_CAP_GEN(mdev, log_max_qp_sz);
	max_rq_sg =  MLX5_CAP_GEN(mdev, max_wqe_sz_rq) /
		     sizeof(struct mlx5_wqe_data_seg);
	max_sq_sg = (MLX5_CAP_GEN(mdev, max_wqe_sz_sq) -
		     sizeof(struct mlx5_wqe_ctrl_seg)) /
		     sizeof(struct mlx5_wqe_data_seg);
	props->max_sge = min(max_rq_sg, max_sq_sg);
	props->max_sge_rd = props->max_sge;
	props->max_cq		   = 1 << MLX5_CAP_GEN(mdev, log_max_cq);
	props->max_cqe = (1 << MLX5_CAP_GEN(mdev, log_max_cq_sz)) - 1;
	props->max_mr		   = 1 << MLX5_CAP_GEN(mdev, log_max_mkey);
	props->max_pd		   = 1 << MLX5_CAP_GEN(mdev, log_max_pd);
	props->max_qp_rd_atom	   = 1 << MLX5_CAP_GEN(mdev, log_max_ra_req_qp);
	props->max_qp_init_rd_atom = 1 << MLX5_CAP_GEN(mdev, log_max_ra_res_qp);
	props->max_srq		   = 1 << MLX5_CAP_GEN(mdev, log_max_srq);
	props->max_srq_wr = (1 << MLX5_CAP_GEN(mdev, log_max_srq_sz)) - 1;
	props->local_ca_ack_delay  = MLX5_CAP_GEN(mdev, local_ca_ack_delay);
	props->max_res_rd_atom	   = props->max_qp_rd_atom * props->max_qp;
	props->max_srq_sge	   = max_rq_sg - 1;
	props->max_fast_reg_page_list_len = (unsigned int)-1;
	get_atomic_caps(dev, props);
	props->masked_atomic_cap   = IB_ATOMIC_NONE;
	props->max_mcast_grp	   = 1 << MLX5_CAP_GEN(mdev, log_max_mcg);
	props->max_mcast_qp_attach = MLX5_CAP_GEN(mdev, max_qp_mcg);
	props->max_total_mcast_qp_attach = props->max_mcast_qp_attach *
					   props->max_mcast_grp;
	props->max_map_per_fmr = INT_MAX; /* no limit in ConnectIB */
	props->hca_core_clock = MLX5_CAP_GEN(mdev, device_frequency_khz);
	props->timestamp_mask = 0x7FFFFFFFFFFFFFFFULL;

#ifdef CONFIG_INFINIBAND_ON_DEMAND_PAGING
	if (MLX5_CAP_GEN(mdev, pg))
		props->device_cap_flags |= IB_DEVICE_ON_DEMAND_PAGING;
	props->odp_caps = dev->odp_caps;
#endif

	if (MLX5_CAP_GEN(mdev, cd))
		props->device_cap_flags |= IB_DEVICE_CROSS_CHANNEL;

	return 0;
}

enum mlx5_ib_width {
	MLX5_IB_WIDTH_1X	= 1 << 0,
	MLX5_IB_WIDTH_2X	= 1 << 1,
	MLX5_IB_WIDTH_4X	= 1 << 2,
	MLX5_IB_WIDTH_8X	= 1 << 3,
	MLX5_IB_WIDTH_12X	= 1 << 4
};

static int translate_active_width(struct ib_device *ibdev, u8 active_width,
				  u8 *ib_width)
{
	struct mlx5_ib_dev *dev = to_mdev(ibdev);
	int err = 0;

	if (active_width & MLX5_IB_WIDTH_1X) {
		*ib_width = IB_WIDTH_1X;
	} else if (active_width & MLX5_IB_WIDTH_2X) {
		mlx5_ib_dbg(dev, "active_width %d is not supported by IB spec\n",
			    (int)active_width);
		err = -EINVAL;
	} else if (active_width & MLX5_IB_WIDTH_4X) {
		*ib_width = IB_WIDTH_4X;
	} else if (active_width & MLX5_IB_WIDTH_8X) {
		*ib_width = IB_WIDTH_8X;
	} else if (active_width & MLX5_IB_WIDTH_12X) {
		*ib_width = IB_WIDTH_12X;
	} else {
		mlx5_ib_dbg(dev, "Invalid active_width %d\n",
			    (int)active_width);
		err = -EINVAL;
	}

	return err;
}

static int mlx5_mtu_to_ib_mtu(int mtu)
{
	switch (mtu) {
	case 256: return 1;
	case 512: return 2;
	case 1024: return 3;
	case 2048: return 4;
	case 4096: return 5;
	default:
		pr_warn("invalid mtu\n");
		return -1;
	}
}

enum ib_max_vl_num {
	__IB_MAX_VL_0		= 1,
	__IB_MAX_VL_0_1		= 2,
	__IB_MAX_VL_0_3		= 3,
	__IB_MAX_VL_0_7		= 4,
	__IB_MAX_VL_0_14	= 5,
};

enum mlx5_vl_hw_cap {
	MLX5_VL_HW_0	= 1,
	MLX5_VL_HW_0_1	= 2,
	MLX5_VL_HW_0_2	= 3,
	MLX5_VL_HW_0_3	= 4,
	MLX5_VL_HW_0_4	= 5,
	MLX5_VL_HW_0_5	= 6,
	MLX5_VL_HW_0_6	= 7,
	MLX5_VL_HW_0_7	= 8,
	MLX5_VL_HW_0_14	= 15
};

static int translate_max_vl_num(struct ib_device *ibdev, u8 vl_hw_cap,
				u8 *max_vl_num)
{
	switch (vl_hw_cap) {
	case MLX5_VL_HW_0:
		*max_vl_num = __IB_MAX_VL_0;
		break;
	case MLX5_VL_HW_0_1:
		*max_vl_num = __IB_MAX_VL_0_1;
		break;
	case MLX5_VL_HW_0_3:
		*max_vl_num = __IB_MAX_VL_0_3;
		break;
	case MLX5_VL_HW_0_7:
		*max_vl_num = __IB_MAX_VL_0_7;
		break;
	case MLX5_VL_HW_0_14:
		*max_vl_num = __IB_MAX_VL_0_14;
		break;

	default:
		return -EINVAL;
	}

	return 0;
}

static int mlx5_query_hca_port(struct ib_device *ibdev, u8 port,
			       struct ib_port_attr *props)
{
	struct mlx5_ib_dev *dev = to_mdev(ibdev);
	struct mlx5_core_dev *mdev = dev->mdev;
	struct mlx5_hca_vport_context *rep;
	int max_mtu;
	int oper_mtu;
	int err;
	u8 ib_link_width_oper;
	u8 vl_hw_cap;

	rep = kzalloc(sizeof(*rep), GFP_KERNEL);
	if (!rep) {
		err = -ENOMEM;
		goto out;
	}

	memset(props, 0, sizeof(*props));

	err = mlx5_query_hca_vport_context(mdev, 0, port, 0, rep);
	if (err)
		goto out;

	props->lid		= rep->lid;
	props->lmc		= rep->lmc;
	props->sm_lid		= rep->sm_lid;
	props->sm_sl		= rep->sm_sl;
	props->state		= rep->vport_state;
	props->phys_state	= rep->port_physical_state;
	props->port_cap_flags	= rep->cap_mask1;
	props->gid_tbl_len	= mlx5_get_gid_table_len(MLX5_CAP_GEN(mdev, gid_table_size));
	props->max_msg_sz	= 1 << MLX5_CAP_GEN(mdev, log_max_msg);
	props->pkey_tbl_len	= mlx5_to_sw_pkey_sz(MLX5_CAP_GEN(mdev, pkey_table_size));
	props->bad_pkey_cntr	= rep->pkey_violation_counter;
	props->qkey_viol_cntr	= rep->qkey_violation_counter;
	props->subnet_timeout	= rep->subnet_timeout;
	props->init_type_reply	= rep->init_type_reply;

	err = mlx5_query_port_link_width_oper(mdev, &ib_link_width_oper, port);
	if (err)
		goto out;

	err = translate_active_width(ibdev, ib_link_width_oper,
				     &props->active_width);
	if (err)
		goto out;
	err = mlx5_query_port_proto_oper(mdev, &props->active_speed, MLX5_PTYS_IB,
					 port);
	if (err)
		goto out;

	mlx5_query_port_max_mtu(mdev, &max_mtu, port);

	props->max_mtu = mlx5_mtu_to_ib_mtu(max_mtu);

	mlx5_query_port_oper_mtu(mdev, &oper_mtu, port);

	props->active_mtu = mlx5_mtu_to_ib_mtu(oper_mtu);

	err = mlx5_query_port_vl_hw_cap(mdev, &vl_hw_cap, port);
	if (err)
		goto out;

	err = translate_max_vl_num(ibdev, vl_hw_cap,
				   &props->max_vl_num);
out:
	kfree(rep);
	return err;
}

int mlx5_ib_query_port(struct ib_device *ibdev, u8 port,
		       struct ib_port_attr *props)
{
	switch (mlx5_get_vport_access_method(ibdev)) {
	case MLX5_VPORT_ACCESS_METHOD_MAD:
		return mlx5_query_mad_ifc_port(ibdev, port, props);

	case MLX5_VPORT_ACCESS_METHOD_HCA:
		return mlx5_query_hca_port(ibdev, port, props);

	case MLX5_VPORT_ACCESS_METHOD_NIC:
		return mlx5_query_port_roce(ibdev, port, props);

	default:
		return -EINVAL;
	}
}

static int mlx5_ib_query_gid(struct ib_device *ibdev, u8 port, int index,
			     union ib_gid *gid)
{
	struct mlx5_ib_dev *dev = to_mdev(ibdev);
	struct mlx5_core_dev *mdev = dev->mdev;

	switch (mlx5_get_vport_access_method(ibdev)) {
	case MLX5_VPORT_ACCESS_METHOD_MAD:
		return mlx5_query_mad_ifc_gids(ibdev, port, index, gid);

	case MLX5_VPORT_ACCESS_METHOD_HCA:
		return mlx5_query_hca_vport_gid(mdev, 0, port, 0, index, gid);

	default:
		return -EINVAL;
	}

}

static int mlx5_ib_query_pkey(struct ib_device *ibdev, u8 port, u16 index,
			      u16 *pkey)
{
	struct mlx5_ib_dev *dev = to_mdev(ibdev);
	struct mlx5_core_dev *mdev = dev->mdev;

	switch (mlx5_get_vport_access_method(ibdev)) {
	case MLX5_VPORT_ACCESS_METHOD_MAD:
		return mlx5_query_mad_ifc_pkey(ibdev, port, index, pkey);

	case MLX5_VPORT_ACCESS_METHOD_HCA:
	case MLX5_VPORT_ACCESS_METHOD_NIC:
		return mlx5_query_hca_vport_pkey(mdev, 0, port,  0, index,
						 pkey);
	default:
		return -EINVAL;
	}
}

static int mlx5_ib_modify_device(struct ib_device *ibdev, int mask,
				 struct ib_device_modify *props)
{
	struct mlx5_ib_dev *dev = to_mdev(ibdev);
	struct mlx5_reg_node_desc in;
	struct mlx5_reg_node_desc out;
	int err;

	if (mask & ~IB_DEVICE_MODIFY_NODE_DESC)
		return -EOPNOTSUPP;

	if (!(mask & IB_DEVICE_MODIFY_NODE_DESC))
		return 0;

	/*
	 * If possible, pass node desc to FW, so it can generate
	 * a 144 trap.  If cmd fails, just ignore.
	 */
	memcpy(&in, props->node_desc, 64);
	err = mlx5_core_access_reg(dev->mdev, &in, sizeof(in), &out,
				   sizeof(out), MLX5_REG_NODE_DESC, 0, 1);
	if (err)
		return err;

	memcpy(ibdev->node_desc, props->node_desc, 64);

	return err;
}

static int mlx5_ib_modify_port(struct ib_device *ibdev, u8 port, int mask,
			       struct ib_port_modify *props)
{
	struct mlx5_ib_dev *dev = to_mdev(ibdev);
	struct ib_port_attr attr;
	u32 tmp;
	int err;

	mutex_lock(&dev->cap_mask_mutex);

	err = mlx5_ib_query_port(ibdev, port, &attr);
	if (err)
		goto out;

	tmp = (attr.port_cap_flags | props->set_port_cap_mask) &
		~props->clr_port_cap_mask;

	err = mlx5_set_port_caps(dev->mdev, port, tmp);

out:
	mutex_unlock(&dev->cap_mask_mutex);
	return err;
}

static struct ib_ucontext *mlx5_ib_alloc_ucontext(struct ib_device *ibdev,
						  struct ib_udata *udata)
{
	struct mlx5_ib_dev *dev = to_mdev(ibdev);
	struct mlx5_ib_alloc_ucontext_req_v2 req = {};
	struct mlx5_ib_alloc_ucontext_resp resp = {};
	struct mlx5_ib_ucontext *context;
	struct mlx5_uuar_info *uuari;
	struct mlx5_uar *uars;
	int gross_uuars;
	int num_uars;
	int ver;
	int uuarn;
	int err;
	int i;
	size_t reqlen;
	size_t min_req_v2 = offsetof(struct mlx5_ib_alloc_ucontext_req_v2,
				     max_cqe_version);

	if (!dev->ib_active)
		return ERR_PTR(-EAGAIN);

	if (udata->inlen < sizeof(struct ib_uverbs_cmd_hdr))
		return ERR_PTR(-EINVAL);

	reqlen = udata->inlen - sizeof(struct ib_uverbs_cmd_hdr);
	if (reqlen == sizeof(struct mlx5_ib_alloc_ucontext_req))
		ver = 0;
	else if (reqlen >= min_req_v2)
		ver = 2;
	else
		return ERR_PTR(-EINVAL);

	err = ib_copy_from_udata(&req, udata, min(reqlen, sizeof(req)));
	if (err)
		return ERR_PTR(err);

	if (req.flags)
		return ERR_PTR(-EINVAL);

	if (req.total_num_uuars > MLX5_MAX_UUARS)
		return ERR_PTR(-ENOMEM);

	if (req.total_num_uuars == 0)
		return ERR_PTR(-EINVAL);

	if (req.comp_mask || req.reserved0 || req.reserved1 || req.reserved2)
		return ERR_PTR(-EOPNOTSUPP);

	if (reqlen > sizeof(req) &&
	    !ib_is_udata_cleared(udata, sizeof(req),
				 reqlen - sizeof(req)))
		return ERR_PTR(-EOPNOTSUPP);

	req.total_num_uuars = ALIGN(req.total_num_uuars,
				    MLX5_NON_FP_BF_REGS_PER_PAGE);
	if (req.num_low_latency_uuars > req.total_num_uuars - 1)
		return ERR_PTR(-EINVAL);

	num_uars = req.total_num_uuars / MLX5_NON_FP_BF_REGS_PER_PAGE;
	gross_uuars = num_uars * MLX5_BF_REGS_PER_PAGE;
	resp.qp_tab_size = 1 << MLX5_CAP_GEN(dev->mdev, log_max_qp);
	resp.bf_reg_size = 1 << MLX5_CAP_GEN(dev->mdev, log_bf_reg_size);
	resp.cache_line_size = L1_CACHE_BYTES;
	resp.max_sq_desc_sz = MLX5_CAP_GEN(dev->mdev, max_wqe_sz_sq);
	resp.max_rq_desc_sz = MLX5_CAP_GEN(dev->mdev, max_wqe_sz_rq);
	resp.max_send_wqebb = 1 << MLX5_CAP_GEN(dev->mdev, log_max_qp_sz);
	resp.max_recv_wr = 1 << MLX5_CAP_GEN(dev->mdev, log_max_qp_sz);
	resp.max_srq_recv_wr = 1 << MLX5_CAP_GEN(dev->mdev, log_max_srq_sz);
	resp.cqe_version = min_t(__u8,
				 (__u8)MLX5_CAP_GEN(dev->mdev, cqe_version),
				 req.max_cqe_version);
	resp.response_length = min(offsetof(typeof(resp), response_length) +
				   sizeof(resp.response_length), udata->outlen);

	context = kzalloc(sizeof(*context), GFP_KERNEL);
	if (!context)
		return ERR_PTR(-ENOMEM);

	uuari = &context->uuari;
	mutex_init(&uuari->lock);
	uars = kcalloc(num_uars, sizeof(*uars), GFP_KERNEL);
	if (!uars) {
		err = -ENOMEM;
		goto out_ctx;
	}

	uuari->bitmap = kcalloc(BITS_TO_LONGS(gross_uuars),
				sizeof(*uuari->bitmap),
				GFP_KERNEL);
	if (!uuari->bitmap) {
		err = -ENOMEM;
		goto out_uar_ctx;
	}
	/*
	 * clear all fast path uuars
	 */
	for (i = 0; i < gross_uuars; i++) {
		uuarn = i & 3;
		if (uuarn == 2 || uuarn == 3)
			set_bit(i, uuari->bitmap);
	}

	uuari->count = kcalloc(gross_uuars, sizeof(*uuari->count), GFP_KERNEL);
	if (!uuari->count) {
		err = -ENOMEM;
		goto out_bitmap;
	}

	for (i = 0; i < num_uars; i++) {
		err = mlx5_cmd_alloc_uar(dev->mdev, &uars[i].index);
		if (err)
			goto out_count;
	}

#ifdef CONFIG_INFINIBAND_ON_DEMAND_PAGING
	context->ibucontext.invalidate_range = &mlx5_ib_invalidate_range;
#endif

	if (MLX5_CAP_GEN(dev->mdev, log_max_transport_domain)) {
		err = mlx5_core_alloc_transport_domain(dev->mdev,
						       &context->tdn);
		if (err)
			goto out_uars;
	}

	INIT_LIST_HEAD(&context->db_page_list);
	mutex_init(&context->db_page_mutex);

	resp.tot_uuars = req.total_num_uuars;
	resp.num_ports = MLX5_CAP_GEN(dev->mdev, num_ports);

	if (field_avail(typeof(resp), cqe_version, udata->outlen))
		resp.response_length += sizeof(resp.cqe_version);

	if (field_avail(typeof(resp), hca_core_clock_offset, udata->outlen)) {
		resp.comp_mask |=
			MLX5_IB_ALLOC_UCONTEXT_RESP_MASK_CORE_CLOCK_OFFSET;
		resp.hca_core_clock_offset =
			offsetof(struct mlx5_init_seg, internal_timer_h) %
			PAGE_SIZE;
		resp.response_length += sizeof(resp.hca_core_clock_offset) +
					sizeof(resp.reserved2) +
					sizeof(resp.reserved3);
	}

	err = ib_copy_to_udata(udata, &resp, resp.response_length);
	if (err)
		goto out_td;

	uuari->ver = ver;
	uuari->num_low_latency_uuars = req.num_low_latency_uuars;
	uuari->uars = uars;
	uuari->num_uars = num_uars;
	context->cqe_version = resp.cqe_version;

	return &context->ibucontext;

out_td:
	if (MLX5_CAP_GEN(dev->mdev, log_max_transport_domain))
		mlx5_core_dealloc_transport_domain(dev->mdev, context->tdn);

out_uars:
	for (i--; i >= 0; i--)
		mlx5_cmd_free_uar(dev->mdev, uars[i].index);
out_count:
	kfree(uuari->count);

out_bitmap:
	kfree(uuari->bitmap);

out_uar_ctx:
	kfree(uars);

out_ctx:
	kfree(context);
	return ERR_PTR(err);
}

static int mlx5_ib_dealloc_ucontext(struct ib_ucontext *ibcontext)
{
	struct mlx5_ib_ucontext *context = to_mucontext(ibcontext);
	struct mlx5_ib_dev *dev = to_mdev(ibcontext->device);
	struct mlx5_uuar_info *uuari = &context->uuari;
	int i;

	if (MLX5_CAP_GEN(dev->mdev, log_max_transport_domain))
		mlx5_core_dealloc_transport_domain(dev->mdev, context->tdn);

	for (i = 0; i < uuari->num_uars; i++) {
		if (mlx5_cmd_free_uar(dev->mdev, uuari->uars[i].index))
			mlx5_ib_warn(dev, "failed to free UAR 0x%x\n", uuari->uars[i].index);
	}

	kfree(uuari->count);
	kfree(uuari->bitmap);
	kfree(uuari->uars);
	kfree(context);

	return 0;
}

static phys_addr_t uar_index2pfn(struct mlx5_ib_dev *dev, int index)
{
	return (pci_resource_start(dev->mdev->pdev, 0) >> PAGE_SHIFT) + index;
}

static int get_command(unsigned long offset)
{
	return (offset >> MLX5_IB_MMAP_CMD_SHIFT) & MLX5_IB_MMAP_CMD_MASK;
}

static int get_arg(unsigned long offset)
{
	return offset & ((1 << MLX5_IB_MMAP_CMD_SHIFT) - 1);
}

static int get_index(unsigned long offset)
{
	return get_arg(offset);
}

static int mlx5_ib_mmap(struct ib_ucontext *ibcontext, struct vm_area_struct *vma)
{
	struct mlx5_ib_ucontext *context = to_mucontext(ibcontext);
	struct mlx5_ib_dev *dev = to_mdev(ibcontext->device);
	struct mlx5_uuar_info *uuari = &context->uuari;
	unsigned long command;
	unsigned long idx;
	phys_addr_t pfn;

	command = get_command(vma->vm_pgoff);
	switch (command) {
	case MLX5_IB_MMAP_REGULAR_PAGE:
		if (vma->vm_end - vma->vm_start != PAGE_SIZE)
			return -EINVAL;

		idx = get_index(vma->vm_pgoff);
		if (idx >= uuari->num_uars)
			return -EINVAL;

		pfn = uar_index2pfn(dev, uuari->uars[idx].index);
		mlx5_ib_dbg(dev, "uar idx 0x%lx, pfn 0x%llx\n", idx,
			    (unsigned long long)pfn);

		vma->vm_page_prot = pgprot_writecombine(vma->vm_page_prot);
		if (io_remap_pfn_range(vma, vma->vm_start, pfn,
				       PAGE_SIZE, vma->vm_page_prot))
			return -EAGAIN;

		mlx5_ib_dbg(dev, "mapped WC at 0x%lx, PA 0x%llx\n",
			    vma->vm_start,
			    (unsigned long long)pfn << PAGE_SHIFT);
		break;

	case MLX5_IB_MMAP_GET_CONTIGUOUS_PAGES:
		return -ENOSYS;

	case MLX5_IB_MMAP_CORE_CLOCK:
		if (vma->vm_end - vma->vm_start != PAGE_SIZE)
			return -EINVAL;

		if (vma->vm_flags & (VM_WRITE | VM_EXEC))
			return -EPERM;

		/* Don't expose to user-space information it shouldn't have */
		if (PAGE_SIZE > 4096)
			return -EOPNOTSUPP;

		vma->vm_page_prot = pgprot_noncached(vma->vm_page_prot);
		pfn = (dev->mdev->iseg_base +
		       offsetof(struct mlx5_init_seg, internal_timer_h)) >>
			PAGE_SHIFT;
		if (io_remap_pfn_range(vma, vma->vm_start, pfn,
				       PAGE_SIZE, vma->vm_page_prot))
			return -EAGAIN;

		mlx5_ib_dbg(dev, "mapped internal timer at 0x%lx, PA 0x%llx\n",
			    vma->vm_start,
			    (unsigned long long)pfn << PAGE_SHIFT);
		break;

	default:
		return -EINVAL;
	}

	return 0;
}

static struct ib_pd *mlx5_ib_alloc_pd(struct ib_device *ibdev,
				      struct ib_ucontext *context,
				      struct ib_udata *udata)
{
	struct mlx5_ib_alloc_pd_resp resp;
	struct mlx5_ib_pd *pd;
	int err;

	pd = kmalloc(sizeof(*pd), GFP_KERNEL);
	if (!pd)
		return ERR_PTR(-ENOMEM);

	err = mlx5_core_alloc_pd(to_mdev(ibdev)->mdev, &pd->pdn);
	if (err) {
		kfree(pd);
		return ERR_PTR(err);
	}

	if (context) {
		resp.pdn = pd->pdn;
		if (ib_copy_to_udata(udata, &resp, sizeof(resp))) {
			mlx5_core_dealloc_pd(to_mdev(ibdev)->mdev, pd->pdn);
			kfree(pd);
			return ERR_PTR(-EFAULT);
		}
	}

	return &pd->ibpd;
}

static int mlx5_ib_dealloc_pd(struct ib_pd *pd)
{
	struct mlx5_ib_dev *mdev = to_mdev(pd->device);
	struct mlx5_ib_pd *mpd = to_mpd(pd);

	mlx5_core_dealloc_pd(mdev->mdev, mpd->pdn);
	kfree(mpd);

	return 0;
}

static bool outer_header_zero(u32 *match_criteria)
{
	int size = MLX5_ST_SZ_BYTES(fte_match_param);
	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 int parse_flow_attr(u32 *match_c, u32 *match_v,
			   union ib_flow_spec *ib_spec)
{
	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);
	switch (ib_spec->type) {
	case IB_FLOW_SPEC_ETH:
		if (ib_spec->size != sizeof(ib_spec->eth))
			return -EINVAL;

		ether_addr_copy(MLX5_ADDR_OF(fte_match_set_lyr_2_4, outer_headers_c,
					     dmac_47_16),
				ib_spec->eth.mask.dst_mac);
		ether_addr_copy(MLX5_ADDR_OF(fte_match_set_lyr_2_4, outer_headers_v,
					     dmac_47_16),
				ib_spec->eth.val.dst_mac);

		if (ib_spec->eth.mask.vlan_tag) {
			MLX5_SET(fte_match_set_lyr_2_4, outer_headers_c,
				 vlan_tag, 1);
			MLX5_SET(fte_match_set_lyr_2_4, outer_headers_v,
				 vlan_tag, 1);

			MLX5_SET(fte_match_set_lyr_2_4, outer_headers_c,
				 first_vid, ntohs(ib_spec->eth.mask.vlan_tag));
			MLX5_SET(fte_match_set_lyr_2_4, outer_headers_v,
				 first_vid, ntohs(ib_spec->eth.val.vlan_tag));

			MLX5_SET(fte_match_set_lyr_2_4, outer_headers_c,
				 first_cfi,
				 ntohs(ib_spec->eth.mask.vlan_tag) >> 12);
			MLX5_SET(fte_match_set_lyr_2_4, outer_headers_v,
				 first_cfi,
				 ntohs(ib_spec->eth.val.vlan_tag) >> 12);

			MLX5_SET(fte_match_set_lyr_2_4, outer_headers_c,
				 first_prio,
				 ntohs(ib_spec->eth.mask.vlan_tag) >> 13);
			MLX5_SET(fte_match_set_lyr_2_4, outer_headers_v,
				 first_prio,
				 ntohs(ib_spec->eth.val.vlan_tag) >> 13);
		}
		MLX5_SET(fte_match_set_lyr_2_4, outer_headers_c,
			 ethertype, ntohs(ib_spec->eth.mask.ether_type));
		MLX5_SET(fte_match_set_lyr_2_4, outer_headers_v,
			 ethertype, ntohs(ib_spec->eth.val.ether_type));
		break;
	case IB_FLOW_SPEC_IPV4:
		if (ib_spec->size != sizeof(ib_spec->ipv4))
			return -EINVAL;

		MLX5_SET(fte_match_set_lyr_2_4, outer_headers_c,
			 ethertype, 0xffff);
		MLX5_SET(fte_match_set_lyr_2_4, outer_headers_v,
			 ethertype, ETH_P_IP);

		memcpy(MLX5_ADDR_OF(fte_match_set_lyr_2_4, outer_headers_c,
				    src_ipv4_src_ipv6.ipv4_layout.ipv4),
		       &ib_spec->ipv4.mask.src_ip,
		       sizeof(ib_spec->ipv4.mask.src_ip));
		memcpy(MLX5_ADDR_OF(fte_match_set_lyr_2_4, outer_headers_v,
				    src_ipv4_src_ipv6.ipv4_layout.ipv4),
		       &ib_spec->ipv4.val.src_ip,
		       sizeof(ib_spec->ipv4.val.src_ip));
		memcpy(MLX5_ADDR_OF(fte_match_set_lyr_2_4, outer_headers_c,
				    dst_ipv4_dst_ipv6.ipv4_layout.ipv4),
		       &ib_spec->ipv4.mask.dst_ip,
		       sizeof(ib_spec->ipv4.mask.dst_ip));
		memcpy(MLX5_ADDR_OF(fte_match_set_lyr_2_4, outer_headers_v,
				    dst_ipv4_dst_ipv6.ipv4_layout.ipv4),
		       &ib_spec->ipv4.val.dst_ip,
		       sizeof(ib_spec->ipv4.val.dst_ip));
		break;
	case IB_FLOW_SPEC_TCP:
		if (ib_spec->size != sizeof(ib_spec->tcp_udp))
			return -EINVAL;

		MLX5_SET(fte_match_set_lyr_2_4, outer_headers_c, ip_protocol,
			 0xff);
		MLX5_SET(fte_match_set_lyr_2_4, outer_headers_v, ip_protocol,
			 IPPROTO_TCP);

		MLX5_SET(fte_match_set_lyr_2_4, outer_headers_c, tcp_sport,
			 ntohs(ib_spec->tcp_udp.mask.src_port));
		MLX5_SET(fte_match_set_lyr_2_4, outer_headers_v, tcp_sport,
			 ntohs(ib_spec->tcp_udp.val.src_port));

		MLX5_SET(fte_match_set_lyr_2_4, outer_headers_c, tcp_dport,
			 ntohs(ib_spec->tcp_udp.mask.dst_port));
		MLX5_SET(fte_match_set_lyr_2_4, outer_headers_v, tcp_dport,
			 ntohs(ib_spec->tcp_udp.val.dst_port));
		break;
	case IB_FLOW_SPEC_UDP:
		if (ib_spec->size != sizeof(ib_spec->tcp_udp))
			return -EINVAL;

		MLX5_SET(fte_match_set_lyr_2_4, outer_headers_c, ip_protocol,
			 0xff);
		MLX5_SET(fte_match_set_lyr_2_4, outer_headers_v, ip_protocol,
			 IPPROTO_UDP);

		MLX5_SET(fte_match_set_lyr_2_4, outer_headers_c, udp_sport,
			 ntohs(ib_spec->tcp_udp.mask.src_port));
		MLX5_SET(fte_match_set_lyr_2_4, outer_headers_v, udp_sport,
			 ntohs(ib_spec->tcp_udp.val.src_port));

		MLX5_SET(fte_match_set_lyr_2_4, outer_headers_c, udp_dport,
			 ntohs(ib_spec->tcp_udp.mask.dst_port));
		MLX5_SET(fte_match_set_lyr_2_4, outer_headers_v, udp_dport,
			 ntohs(ib_spec->tcp_udp.val.dst_port));
		break;
	default:
		return -EINVAL;
	}

	return 0;
}

/* If a flow could catch both multicast and unicast packets,
 * it won't fall into the multicast flow steering table and this rule
 * could steal other multicast packets.
 */
static bool flow_is_multicast_only(struct ib_flow_attr *ib_attr)
{
	struct ib_flow_spec_eth *eth_spec;

	if (ib_attr->type != IB_FLOW_ATTR_NORMAL ||
	    ib_attr->size < sizeof(struct ib_flow_attr) +
	    sizeof(struct ib_flow_spec_eth) ||
	    ib_attr->num_of_specs < 1)
		return false;

	eth_spec = (struct ib_flow_spec_eth *)(ib_attr + 1);
	if (eth_spec->type != IB_FLOW_SPEC_ETH ||
	    eth_spec->size != sizeof(*eth_spec))
		return false;

	return is_multicast_ether_addr(eth_spec->mask.dst_mac) &&
	       is_multicast_ether_addr(eth_spec->val.dst_mac);
}

static bool is_valid_attr(struct ib_flow_attr *flow_attr)
{
	union ib_flow_spec *ib_spec = (union ib_flow_spec *)(flow_attr + 1);
	bool has_ipv4_spec = false;
	bool eth_type_ipv4 = true;
	unsigned int spec_index;

	/* Validate that ethertype is correct */
	for (spec_index = 0; spec_index < flow_attr->num_of_specs; spec_index++) {
		if (ib_spec->type == IB_FLOW_SPEC_ETH &&
		    ib_spec->eth.mask.ether_type) {
			if (!((ib_spec->eth.mask.ether_type == htons(0xffff)) &&
			      ib_spec->eth.val.ether_type == htons(ETH_P_IP)))
				eth_type_ipv4 = false;
		} else if (ib_spec->type == IB_FLOW_SPEC_IPV4) {
			has_ipv4_spec = true;
		}
		ib_spec = (void *)ib_spec + ib_spec->size;
	}
	return !has_ipv4_spec || eth_type_ipv4;
}

static void put_flow_table(struct mlx5_ib_dev *dev,
			   struct mlx5_ib_flow_prio *prio, bool ft_added)
{
	prio->refcount -= !!ft_added;
	if (!prio->refcount) {
		mlx5_destroy_flow_table(prio->flow_table);
		prio->flow_table = NULL;
	}
}

static int mlx5_ib_destroy_flow(struct ib_flow *flow_id)
{
	struct mlx5_ib_dev *dev = to_mdev(flow_id->qp->device);
	struct mlx5_ib_flow_handler *handler = container_of(flow_id,
							  struct mlx5_ib_flow_handler,
							  ibflow);
	struct mlx5_ib_flow_handler *iter, *tmp;

	mutex_lock(&dev->flow_db.lock);

	list_for_each_entry_safe(iter, tmp, &handler->list, list) {
		mlx5_del_flow_rule(iter->rule);
		list_del(&iter->list);
		kfree(iter);
	}

	mlx5_del_flow_rule(handler->rule);
	put_flow_table(dev, &dev->flow_db.prios[handler->prio], true);
	mutex_unlock(&dev->flow_db.lock);

	kfree(handler);

	return 0;
}

#define MLX5_FS_MAX_TYPES	 10
#define MLX5_FS_MAX_ENTRIES	 32000UL
static struct mlx5_ib_flow_prio *get_flow_table(struct mlx5_ib_dev *dev,
						struct ib_flow_attr *flow_attr)
{
	struct mlx5_flow_namespace *ns = NULL;
	struct mlx5_ib_flow_prio *prio;
	struct mlx5_flow_table *ft;
	int num_entries;
	int num_groups;
	int priority;
	int err = 0;

	if (flow_attr->type == IB_FLOW_ATTR_NORMAL) {
		if (flow_is_multicast_only(flow_attr))
			priority = MLX5_IB_FLOW_MCAST_PRIO;
		else
			priority = flow_attr->priority;
		ns = mlx5_get_flow_namespace(dev->mdev,
					     MLX5_FLOW_NAMESPACE_BYPASS);
		num_entries = MLX5_FS_MAX_ENTRIES;
		num_groups = MLX5_FS_MAX_TYPES;
		prio = &dev->flow_db.prios[priority];
	} else if (flow_attr->type == IB_FLOW_ATTR_ALL_DEFAULT ||
		   flow_attr->type == IB_FLOW_ATTR_MC_DEFAULT) {
		ns = mlx5_get_flow_namespace(dev->mdev,
					     MLX5_FLOW_NAMESPACE_LEFTOVERS);
		build_leftovers_ft_param(&priority,
					 &num_entries,
					 &num_groups);
		prio = &dev->flow_db.prios[MLX5_IB_FLOW_LEFTOVERS_PRIO];
	}

	if (!ns)
		return ERR_PTR(-ENOTSUPP);

	ft = prio->flow_table;
	if (!ft) {
		ft = mlx5_create_auto_grouped_flow_table(ns, priority,
							 num_entries,
							 num_groups);

		if (!IS_ERR(ft)) {
			prio->refcount = 0;
			prio->flow_table = ft;
		} else {
			err = PTR_ERR(ft);
		}
	}

	return err ? ERR_PTR(err) : prio;
}

static struct mlx5_ib_flow_handler *create_flow_rule(struct mlx5_ib_dev *dev,
						     struct mlx5_ib_flow_prio *ft_prio,
						     struct ib_flow_attr *flow_attr,
						     struct mlx5_flow_destination *dst)
{
	struct mlx5_flow_table	*ft = ft_prio->flow_table;
	struct mlx5_ib_flow_handler *handler;
	void *ib_flow = flow_attr + 1;
	u8 match_criteria_enable = 0;
	unsigned int spec_index;
	u32 *match_c;
	u32 *match_v;
	int err = 0;

	if (!is_valid_attr(flow_attr))
		return ERR_PTR(-EINVAL);

	match_c = kzalloc(MLX5_ST_SZ_BYTES(fte_match_param), GFP_KERNEL);
	match_v = kzalloc(MLX5_ST_SZ_BYTES(fte_match_param), GFP_KERNEL);
	handler = kzalloc(sizeof(*handler), GFP_KERNEL);
	if (!handler || !match_c || !match_v) {
		err = -ENOMEM;
		goto free;
	}

	INIT_LIST_HEAD(&handler->list);

	for (spec_index = 0; spec_index < flow_attr->num_of_specs; spec_index++) {
		err = parse_flow_attr(match_c, match_v, ib_flow);
		if (err < 0)
			goto free;

		ib_flow += ((union ib_flow_spec *)ib_flow)->size;
	}

	/* Outer header support only */
	match_criteria_enable = (!outer_header_zero(match_c)) << 0;
	handler->rule = mlx5_add_flow_rule(ft, match_criteria_enable,
					   match_c, match_v,
					   MLX5_FLOW_CONTEXT_ACTION_FWD_DEST,
					   MLX5_FS_DEFAULT_FLOW_TAG,
					   dst);

	if (IS_ERR(handler->rule)) {
		err = PTR_ERR(handler->rule);
		goto free;
	}

	handler->prio = ft_prio - dev->flow_db.prios;

	ft_prio->flow_table = ft;
free:
	if (err)
		kfree(handler);
	kfree(match_c);
	kfree(match_v);
	return err ? ERR_PTR(err) : handler;
}

enum {
	LEFTOVERS_MC,
	LEFTOVERS_UC,
};

static struct mlx5_ib_flow_handler *create_leftovers_rule(struct mlx5_ib_dev *dev,
							  struct mlx5_ib_flow_prio *ft_prio,
							  struct ib_flow_attr *flow_attr,
							  struct mlx5_flow_destination *dst)
{
	struct mlx5_ib_flow_handler *handler_ucast = NULL;
	struct mlx5_ib_flow_handler *handler = NULL;

	static struct {
		struct ib_flow_attr	flow_attr;
		struct ib_flow_spec_eth eth_flow;
	} leftovers_specs[] = {
		[LEFTOVERS_MC] = {
			.flow_attr = {
				.num_of_specs = 1,
				.size = sizeof(leftovers_specs[0])
			},
			.eth_flow = {
				.type = IB_FLOW_SPEC_ETH,
				.size = sizeof(struct ib_flow_spec_eth),
				.mask = {.dst_mac = {0x1} },
				.val =  {.dst_mac = {0x1} }
			}
		},
		[LEFTOVERS_UC] = {
			.flow_attr = {
				.num_of_specs = 1,
				.size = sizeof(leftovers_specs[0])
			},
			.eth_flow = {
				.type = IB_FLOW_SPEC_ETH,
				.size = sizeof(struct ib_flow_spec_eth),
				.mask = {.dst_mac = {0x1} },
				.val = {.dst_mac = {} }
			}
		}
	};

	handler = create_flow_rule(dev, ft_prio,
				   &leftovers_specs[LEFTOVERS_MC].flow_attr,
				   dst);
	if (!IS_ERR(handler) &&
	    flow_attr->type == IB_FLOW_ATTR_ALL_DEFAULT) {
		handler_ucast = create_flow_rule(dev, ft_prio,
						 &leftovers_specs[LEFTOVERS_UC].flow_attr,
						 dst);
		if (IS_ERR(handler_ucast)) {
			kfree(handler);
			handler = handler_ucast;
		} else {
			list_add(&handler_ucast->list, &handler->list);
		}
	}

	return handler;
}

static struct ib_flow *mlx5_ib_create_flow(struct ib_qp *qp,
					   struct ib_flow_attr *flow_attr,
					   int domain)
{
	struct mlx5_ib_dev *dev = to_mdev(qp->device);
	struct mlx5_ib_flow_handler *handler = NULL;
	struct mlx5_flow_destination *dst = NULL;
	struct mlx5_ib_flow_prio *ft_prio;
	int err;

	if (flow_attr->priority > MLX5_IB_FLOW_LAST_PRIO)
		return ERR_PTR(-ENOSPC);

	if (domain != IB_FLOW_DOMAIN_USER ||
	    flow_attr->port > MLX5_CAP_GEN(dev->mdev, num_ports) ||
	    flow_attr->flags)
		return ERR_PTR(-EINVAL);

	dst = kzalloc(sizeof(*dst), GFP_KERNEL);
	if (!dst)
		return ERR_PTR(-ENOMEM);

	mutex_lock(&dev->flow_db.lock);

	ft_prio = get_flow_table(dev, flow_attr);
	if (IS_ERR(ft_prio)) {
		err = PTR_ERR(ft_prio);
		goto unlock;
	}

	dst->type = MLX5_FLOW_DESTINATION_TYPE_TIR;
	dst->tir_num = to_mqp(qp)->raw_packet_qp.rq.tirn;

	if (flow_attr->type == IB_FLOW_ATTR_NORMAL) {
		handler = create_flow_rule(dev, ft_prio, flow_attr,
					   dst);
	} else if (flow_attr->type == IB_FLOW_ATTR_ALL_DEFAULT ||
		   flow_attr->type == IB_FLOW_ATTR_MC_DEFAULT) {
		handler = create_leftovers_rule(dev, ft_prio, flow_attr,
						dst);
	} else {
		err = -EINVAL;
		goto destroy_ft;
	}

	if (IS_ERR(handler)) {
		err = PTR_ERR(handler);
		handler = NULL;
		goto destroy_ft;
	}

	ft_prio->refcount++;
	mutex_unlock(&dev->flow_db.lock);
	kfree(dst);

	return &handler->ibflow;

destroy_ft:
	put_flow_table(dev, ft_prio, false);
unlock:
	mutex_unlock(&dev->flow_db.lock);
	kfree(dst);
	kfree(handler);
	return ERR_PTR(err);
}

static int mlx5_ib_mcg_attach(struct ib_qp *ibqp, union ib_gid *gid, u16 lid)
{
	struct mlx5_ib_dev *dev = to_mdev(ibqp->device);
	int err;

	err = mlx5_core_attach_mcg(dev->mdev, gid, ibqp->qp_num);
	if (err)
		mlx5_ib_warn(dev, "failed attaching QPN 0x%x, MGID %pI6\n",
			     ibqp->qp_num, gid->raw);

	return err;
}

static int mlx5_ib_mcg_detach(struct ib_qp *ibqp, union ib_gid *gid, u16 lid)
{
	struct mlx5_ib_dev *dev = to_mdev(ibqp->device);
	int err;

	err = mlx5_core_detach_mcg(dev->mdev, gid, ibqp->qp_num);
	if (err)
		mlx5_ib_warn(dev, "failed detaching QPN 0x%x, MGID %pI6\n",
			     ibqp->qp_num, gid->raw);

	return err;
}

static int init_node_data(struct mlx5_ib_dev *dev)
{
	int err;

	err = mlx5_query_node_desc(dev, dev->ib_dev.node_desc);
	if (err)
		return err;

	dev->mdev->rev_id = dev->mdev->pdev->revision;

	return mlx5_query_node_guid(dev, &dev->ib_dev.node_guid);
}

static ssize_t show_fw_pages(struct device *device, struct device_attribute *attr,
			     char *buf)
{
	struct mlx5_ib_dev *dev =
		container_of(device, struct mlx5_ib_dev, ib_dev.dev);

	return sprintf(buf, "%d\n", dev->mdev->priv.fw_pages);
}

static ssize_t show_reg_pages(struct device *device,
			      struct device_attribute *attr, char *buf)
{
	struct mlx5_ib_dev *dev =
		container_of(device, struct mlx5_ib_dev, ib_dev.dev);

	return sprintf(buf, "%d\n", atomic_read(&dev->mdev->priv.reg_pages));
}

static ssize_t show_hca(struct device *device, struct device_attribute *attr,
			char *buf)
{
	struct mlx5_ib_dev *dev =
		container_of(device, struct mlx5_ib_dev, ib_dev.dev);
	return sprintf(buf, "MT%d\n", dev->mdev->pdev->device);
}

static ssize_t show_fw_ver(struct device *device, struct device_attribute *attr,
			   char *buf)
{
	struct mlx5_ib_dev *dev =
		container_of(device, struct mlx5_ib_dev, ib_dev.dev);
	return sprintf(buf, "%d.%d.%d\n", fw_rev_maj(dev->mdev),
		       fw_rev_min(dev->mdev), fw_rev_sub(dev->mdev));
}

static ssize_t show_rev(struct device *device, struct device_attribute *attr,
			char *buf)
{
	struct mlx5_ib_dev *dev =
		container_of(device, struct mlx5_ib_dev, ib_dev.dev);
	return sprintf(buf, "%x\n", dev->mdev->rev_id);
}

static ssize_t show_board(struct device *device, struct device_attribute *attr,
			  char *buf)
{
	struct mlx5_ib_dev *dev =
		container_of(device, struct mlx5_ib_dev, ib_dev.dev);
	return sprintf(buf, "%.*s\n", MLX5_BOARD_ID_LEN,
		       dev->mdev->board_id);
}

static DEVICE_ATTR(hw_rev,   S_IRUGO, show_rev,    NULL);
static DEVICE_ATTR(fw_ver,   S_IRUGO, show_fw_ver, NULL);
static DEVICE_ATTR(hca_type, S_IRUGO, show_hca,    NULL);
static DEVICE_ATTR(board_id, S_IRUGO, show_board,  NULL);
static DEVICE_ATTR(fw_pages, S_IRUGO, show_fw_pages, NULL);
static DEVICE_ATTR(reg_pages, S_IRUGO, show_reg_pages, NULL);

static struct device_attribute *mlx5_class_attributes[] = {
	&dev_attr_hw_rev,
	&dev_attr_fw_ver,
	&dev_attr_hca_type,
	&dev_attr_board_id,
	&dev_attr_fw_pages,
	&dev_attr_reg_pages,
};

static void pkey_change_handler(struct work_struct *work)
{
	struct mlx5_ib_port_resources *ports =
		container_of(work, struct mlx5_ib_port_resources,
			     pkey_change_work);

	mutex_lock(&ports->devr->mutex);
	mlx5_ib_gsi_pkey_change(ports->gsi);
	mutex_unlock(&ports->devr->mutex);
}

static void mlx5_ib_event(struct mlx5_core_dev *dev, void *context,
			  enum mlx5_dev_event event, unsigned long param)
{
	struct mlx5_ib_dev *ibdev = (struct mlx5_ib_dev *)context;
	struct ib_event ibev;

	u8 port = 0;

	switch (event) {
	case MLX5_DEV_EVENT_SYS_ERROR:
		ibdev->ib_active = false;
		ibev.event = IB_EVENT_DEVICE_FATAL;
		break;

	case MLX5_DEV_EVENT_PORT_UP:
		ibev.event = IB_EVENT_PORT_ACTIVE;
		port = (u8)param;
		break;

	case MLX5_DEV_EVENT_PORT_DOWN:
		ibev.event = IB_EVENT_PORT_ERR;
		port = (u8)param;
		break;

	case MLX5_DEV_EVENT_PORT_INITIALIZED:
		/* not used by ULPs */
		return;

	case MLX5_DEV_EVENT_LID_CHANGE:
		ibev.event = IB_EVENT_LID_CHANGE;
		port = (u8)param;
		break;

	case MLX5_DEV_EVENT_PKEY_CHANGE:
		ibev.event = IB_EVENT_PKEY_CHANGE;
		port = (u8)param;

		schedule_work(&ibdev->devr.ports[port - 1].pkey_change_work);
		break;

	case MLX5_DEV_EVENT_GUID_CHANGE:
		ibev.event = IB_EVENT_GID_CHANGE;
		port = (u8)param;
		break;

	case MLX5_DEV_EVENT_CLIENT_REREG:
		ibev.event = IB_EVENT_CLIENT_REREGISTER;
		port = (u8)param;
		break;
	}

	ibev.device	      = &ibdev->ib_dev;
	ibev.element.port_num = port;

	if (port < 1 || port > ibdev->num_ports) {
		mlx5_ib_warn(ibdev, "warning: event on port %d\n", port);
		return;
	}

	if (ibdev->ib_active)
		ib_dispatch_event(&ibev);
}

static void get_ext_port_caps(struct mlx5_ib_dev *dev)
{
	int port;

	for (port = 1; port <= MLX5_CAP_GEN(dev->mdev, num_ports); port++)
		mlx5_query_ext_port_caps(dev, port);
}

static int get_port_caps(struct mlx5_ib_dev *dev)
{
	struct ib_device_attr *dprops = NULL;
	struct ib_port_attr *pprops = NULL;
	int err = -ENOMEM;
	int port;
	struct ib_udata uhw = {.inlen = 0, .outlen = 0};

	pprops = kmalloc(sizeof(*pprops), GFP_KERNEL);
	if (!pprops)
		goto out;

	dprops = kmalloc(sizeof(*dprops), GFP_KERNEL);
	if (!dprops)
		goto out;

	err = mlx5_ib_query_device(&dev->ib_dev, dprops, &uhw);
	if (err) {
		mlx5_ib_warn(dev, "query_device failed %d\n", err);
		goto out;
	}

	for (port = 1; port <= MLX5_CAP_GEN(dev->mdev, num_ports); port++) {
		err = mlx5_ib_query_port(&dev->ib_dev, port, pprops);
		if (err) {
			mlx5_ib_warn(dev, "query_port %d failed %d\n",
				     port, err);
			break;
		}
		dev->mdev->port_caps[port - 1].pkey_table_len =
						dprops->max_pkeys;
		dev->mdev->port_caps[port - 1].gid_table_len =
						pprops->gid_tbl_len;
		mlx5_ib_dbg(dev, "pkey_table_len %d, gid_table_len %d\n",
			    dprops->max_pkeys, pprops->gid_tbl_len);
	}

out:
	kfree(pprops);
	kfree(dprops);

	return err;
}

static void destroy_umrc_res(struct mlx5_ib_dev *dev)
{
	int err;

	err = mlx5_mr_cache_cleanup(dev);
	if (err)
		mlx5_ib_warn(dev, "mr cache cleanup failed\n");

	mlx5_ib_destroy_qp(dev->umrc.qp);
	ib_free_cq(dev->umrc.cq);
	ib_dealloc_pd(dev->umrc.pd);
}

enum {
	MAX_UMR_WR = 128,
};

static int create_umr_res(struct mlx5_ib_dev *dev)
{
	struct ib_qp_init_attr *init_attr = NULL;
	struct ib_qp_attr *attr = NULL;
	struct ib_pd *pd;
	struct ib_cq *cq;
	struct ib_qp *qp;
	int ret;

	attr = kzalloc(sizeof(*attr), GFP_KERNEL);
	init_attr = kzalloc(sizeof(*init_attr), GFP_KERNEL);
	if (!attr || !init_attr) {
		ret = -ENOMEM;
		goto error_0;
	}

	pd = ib_alloc_pd(&dev->ib_dev);
	if (IS_ERR(pd)) {
		mlx5_ib_dbg(dev, "Couldn't create PD for sync UMR QP\n");
		ret = PTR_ERR(pd);
		goto error_0;
	}

	cq = ib_alloc_cq(&dev->ib_dev, NULL, 128, 0, IB_POLL_SOFTIRQ);
	if (IS_ERR(cq)) {
		mlx5_ib_dbg(dev, "Couldn't create CQ for sync UMR QP\n");
		ret = PTR_ERR(cq);
		goto error_2;
	}

	init_attr->send_cq = cq;
	init_attr->recv_cq = cq;
	init_attr->sq_sig_type = IB_SIGNAL_ALL_WR;
	init_attr->cap.max_send_wr = MAX_UMR_WR;
	init_attr->cap.max_send_sge = 1;
	init_attr->qp_type = MLX5_IB_QPT_REG_UMR;
	init_attr->port_num = 1;
	qp = mlx5_ib_create_qp(pd, init_attr, NULL);
	if (IS_ERR(qp)) {
		mlx5_ib_dbg(dev, "Couldn't create sync UMR QP\n");
		ret = PTR_ERR(qp);
		goto error_3;
	}
	qp->device     = &dev->ib_dev;
	qp->real_qp    = qp;
	qp->uobject    = NULL;
	qp->qp_type    = MLX5_IB_QPT_REG_UMR;

	attr->qp_state = IB_QPS_INIT;
	attr->port_num = 1;
	ret = mlx5_ib_modify_qp(qp, attr, IB_QP_STATE | IB_QP_PKEY_INDEX |
				IB_QP_PORT, NULL);
	if (ret) {
		mlx5_ib_dbg(dev, "Couldn't modify UMR QP\n");
		goto error_4;
	}

	memset(attr, 0, sizeof(*attr));
	attr->qp_state = IB_QPS_RTR;
	attr->path_mtu = IB_MTU_256;

	ret = mlx5_ib_modify_qp(qp, attr, IB_QP_STATE, NULL);
	if (ret) {
		mlx5_ib_dbg(dev, "Couldn't modify umr QP to rtr\n");
		goto error_4;
	}

	memset(attr, 0, sizeof(*attr));
	attr->qp_state = IB_QPS_RTS;
	ret = mlx5_ib_modify_qp(qp, attr, IB_QP_STATE, NULL);
	if (ret) {
		mlx5_ib_dbg(dev, "Couldn't modify umr QP to rts\n");
		goto error_4;
	}

	dev->umrc.qp = qp;
	dev->umrc.cq = cq;
	dev->umrc.pd = pd;

	sema_init(&dev->umrc.sem, MAX_UMR_WR);
	ret = mlx5_mr_cache_init(dev);
	if (ret) {
		mlx5_ib_warn(dev, "mr cache init failed %d\n", ret);
		goto error_4;
	}

	kfree(attr);
	kfree(init_attr);

	return 0;

error_4:
	mlx5_ib_destroy_qp(qp);

error_3:
	ib_free_cq(cq);

error_2:
	ib_dealloc_pd(pd);

error_0:
	kfree(attr);
	kfree(init_attr);
	return ret;
}

static int create_dev_resources(struct mlx5_ib_resources *devr)
{
	struct ib_srq_init_attr attr;
	struct mlx5_ib_dev *dev;
	struct ib_cq_init_attr cq_attr = {.cqe = 1};
	int port;
	int ret = 0;

	dev = container_of(devr, struct mlx5_ib_dev, devr);

	mutex_init(&devr->mutex);

	devr->p0 = mlx5_ib_alloc_pd(&dev->ib_dev, NULL, NULL);
	if (IS_ERR(devr->p0)) {
		ret = PTR_ERR(devr->p0);
		goto error0;
	}
	devr->p0->device  = &dev->ib_dev;
	devr->p0->uobject = NULL;
	atomic_set(&devr->p0->usecnt, 0);

	devr->c0 = mlx5_ib_create_cq(&dev->ib_dev, &cq_attr, NULL, NULL);
	if (IS_ERR(devr->c0)) {
		ret = PTR_ERR(devr->c0);
		goto error1;
	}
	devr->c0->device        = &dev->ib_dev;
	devr->c0->uobject       = NULL;
	devr->c0->comp_handler  = NULL;
	devr->c0->event_handler = NULL;
	devr->c0->cq_context    = NULL;
	atomic_set(&devr->c0->usecnt, 0);

	devr->x0 = mlx5_ib_alloc_xrcd(&dev->ib_dev, NULL, NULL);
	if (IS_ERR(devr->x0)) {
		ret = PTR_ERR(devr->x0);
		goto error2;
	}
	devr->x0->device = &dev->ib_dev;
	devr->x0->inode = NULL;
	atomic_set(&devr->x0->usecnt, 0);
	mutex_init(&devr->x0->tgt_qp_mutex);
	INIT_LIST_HEAD(&devr->x0->tgt_qp_list);

	devr->x1 = mlx5_ib_alloc_xrcd(&dev->ib_dev, NULL, NULL);
	if (IS_ERR(devr->x1)) {
		ret = PTR_ERR(devr->x1);
		goto error3;
	}
	devr->x1->device = &dev->ib_dev;
	devr->x1->inode = NULL;
	atomic_set(&devr->x1->usecnt, 0);
	mutex_init(&devr->x1->tgt_qp_mutex);
	INIT_LIST_HEAD(&devr->x1->tgt_qp_list);

	memset(&attr, 0, sizeof(attr));
	attr.attr.max_sge = 1;
	attr.attr.max_wr = 1;
	attr.srq_type = IB_SRQT_XRC;
	attr.ext.xrc.cq = devr->c0;
	attr.ext.xrc.xrcd = devr->x0;

	devr->s0 = mlx5_ib_create_srq(devr->p0, &attr, NULL);
	if (IS_ERR(devr->s0)) {
		ret = PTR_ERR(devr->s0);
		goto error4;
	}
	devr->s0->device	= &dev->ib_dev;
	devr->s0->pd		= devr->p0;
	devr->s0->uobject       = NULL;
	devr->s0->event_handler = NULL;
	devr->s0->srq_context   = NULL;
	devr->s0->srq_type      = IB_SRQT_XRC;
	devr->s0->ext.xrc.xrcd	= devr->x0;
	devr->s0->ext.xrc.cq	= devr->c0;
	atomic_inc(&devr->s0->ext.xrc.xrcd->usecnt);
	atomic_inc(&devr->s0->ext.xrc.cq->usecnt);
	atomic_inc(&devr->p0->usecnt);
	atomic_set(&devr->s0->usecnt, 0);

	memset(&attr, 0, sizeof(attr));
	attr.attr.max_sge = 1;
	attr.attr.max_wr = 1;
	attr.srq_type = IB_SRQT_BASIC;
	devr->s1 = mlx5_ib_create_srq(devr->p0, &attr, NULL);
	if (IS_ERR(devr->s1)) {
		ret = PTR_ERR(devr->s1);
		goto error5;
	}
	devr->s1->device	= &dev->ib_dev;
	devr->s1->pd		= devr->p0;
	devr->s1->uobject       = NULL;
	devr->s1->event_handler = NULL;
	devr->s1->srq_context   = NULL;
	devr->s1->srq_type      = IB_SRQT_BASIC;
	devr->s1->ext.xrc.cq	= devr->c0;
	atomic_inc(&devr->p0->usecnt);
	atomic_set(&devr->s0->usecnt, 0);

	for (port = 0; port < ARRAY_SIZE(devr->ports); ++port) {
		INIT_WORK(&devr->ports[port].pkey_change_work,
			  pkey_change_handler);
		devr->ports[port].devr = devr;
	}

	return 0;

error5:
	mlx5_ib_destroy_srq(devr->s0);
error4:
	mlx5_ib_dealloc_xrcd(devr->x1);
error3:
	mlx5_ib_dealloc_xrcd(devr->x0);
error2:
	mlx5_ib_destroy_cq(devr->c0);
error1:
	mlx5_ib_dealloc_pd(devr->p0);
error0:
	return ret;
}

static void destroy_dev_resources(struct mlx5_ib_resources *devr)
{
	struct mlx5_ib_dev *dev =
		container_of(devr, struct mlx5_ib_dev, devr);
	int port;

	mlx5_ib_destroy_srq(devr->s1);
	mlx5_ib_destroy_srq(devr->s0);
	mlx5_ib_dealloc_xrcd(devr->x0);
	mlx5_ib_dealloc_xrcd(devr->x1);
	mlx5_ib_destroy_cq(devr->c0);
	mlx5_ib_dealloc_pd(devr->p0);

	/* Make sure no change P_Key work items are still executing */
	for (port = 0; port < dev->num_ports; ++port)
		cancel_work_sync(&devr->ports[port].pkey_change_work);
}

static u32 get_core_cap_flags(struct ib_device *ibdev)
{
	struct mlx5_ib_dev *dev = to_mdev(ibdev);
	enum rdma_link_layer ll = mlx5_ib_port_link_layer(ibdev, 1);
	u8 l3_type_cap = MLX5_CAP_ROCE(dev->mdev, l3_type);
	u8 roce_version_cap = MLX5_CAP_ROCE(dev->mdev, roce_version);
	u32 ret = 0;

	if (ll == IB_LINK_LAYER_INFINIBAND)
		return RDMA_CORE_PORT_IBA_IB;

	if (!(l3_type_cap & MLX5_ROCE_L3_TYPE_IPV4_CAP))
		return 0;

	if (!(l3_type_cap & MLX5_ROCE_L3_TYPE_IPV6_CAP))
		return 0;

	if (roce_version_cap & MLX5_ROCE_VERSION_1_CAP)
		ret |= RDMA_CORE_PORT_IBA_ROCE;

	if (roce_version_cap & MLX5_ROCE_VERSION_2_CAP)
		ret |= RDMA_CORE_PORT_IBA_ROCE_UDP_ENCAP;

	return ret;
}

static int mlx5_port_immutable(struct ib_device *ibdev, u8 port_num,
			       struct ib_port_immutable *immutable)
{
	struct ib_port_attr attr;
	int err;

	err = mlx5_ib_query_port(ibdev, port_num, &attr);
	if (err)
		return err;

	immutable->pkey_tbl_len = attr.pkey_tbl_len;
	immutable->gid_tbl_len = attr.gid_tbl_len;
	immutable->core_cap_flags = get_core_cap_flags(ibdev);
	immutable->max_mad_size = IB_MGMT_MAD_SIZE;

	return 0;
}

static int mlx5_enable_roce(struct mlx5_ib_dev *dev)
{
	int err;

	dev->roce.nb.notifier_call = mlx5_netdev_event;
	err = register_netdevice_notifier(&dev->roce.nb);
	if (err)
		return err;

	err = mlx5_nic_vport_enable_roce(dev->mdev);
	if (err)
		goto err_unregister_netdevice_notifier;

	return 0;

err_unregister_netdevice_notifier:
	unregister_netdevice_notifier(&dev->roce.nb);
	return err;
}

static void mlx5_disable_roce(struct mlx5_ib_dev *dev)
{
	mlx5_nic_vport_disable_roce(dev->mdev);
	unregister_netdevice_notifier(&dev->roce.nb);
}

static void *mlx5_ib_add(struct mlx5_core_dev *mdev)
{
	struct mlx5_ib_dev *dev;
	enum rdma_link_layer ll;
	int port_type_cap;
	int err;
	int i;

	port_type_cap = MLX5_CAP_GEN(mdev, port_type);
	ll = mlx5_port_type_cap_to_rdma_ll(port_type_cap);

	if ((ll == IB_LINK_LAYER_ETHERNET) && !MLX5_CAP_GEN(mdev, roce))
		return NULL;

	printk_once(KERN_INFO "%s", mlx5_version);

	dev = (struct mlx5_ib_dev *)ib_alloc_device(sizeof(*dev));
	if (!dev)
		return NULL;

	dev->mdev = mdev;

	rwlock_init(&dev->roce.netdev_lock);
	err = get_port_caps(dev);
	if (err)
		goto err_dealloc;

	if (mlx5_use_mad_ifc(dev))
		get_ext_port_caps(dev);

	MLX5_INIT_DOORBELL_LOCK(&dev->uar_lock);

	strlcpy(dev->ib_dev.name, "mlx5_%d", IB_DEVICE_NAME_MAX);
	dev->ib_dev.owner		= THIS_MODULE;
	dev->ib_dev.node_type		= RDMA_NODE_IB_CA;
	dev->ib_dev.local_dma_lkey	= 0 /* not supported for now */;
	dev->num_ports		= MLX5_CAP_GEN(mdev, num_ports);
	dev->ib_dev.phys_port_cnt     = dev->num_ports;
	dev->ib_dev.num_comp_vectors    =
		dev->mdev->priv.eq_table.num_comp_vectors;
	dev->ib_dev.dma_device	= &mdev->pdev->dev;

	dev->ib_dev.uverbs_abi_ver	= MLX5_IB_UVERBS_ABI_VERSION;
	dev->ib_dev.uverbs_cmd_mask	=
		(1ull << IB_USER_VERBS_CMD_GET_CONTEXT)		|
		(1ull << IB_USER_VERBS_CMD_QUERY_DEVICE)	|
		(1ull << IB_USER_VERBS_CMD_QUERY_PORT)		|
		(1ull << IB_USER_VERBS_CMD_ALLOC_PD)		|
		(1ull << IB_USER_VERBS_CMD_DEALLOC_PD)		|
		(1ull << IB_USER_VERBS_CMD_REG_MR)		|
		(1ull << IB_USER_VERBS_CMD_REREG_MR)		|
		(1ull << IB_USER_VERBS_CMD_DEREG_MR)		|
		(1ull << IB_USER_VERBS_CMD_CREATE_COMP_CHANNEL)	|
		(1ull << IB_USER_VERBS_CMD_CREATE_CQ)		|
		(1ull << IB_USER_VERBS_CMD_RESIZE_CQ)		|
		(1ull << IB_USER_VERBS_CMD_DESTROY_CQ)		|
		(1ull << IB_USER_VERBS_CMD_CREATE_QP)		|
		(1ull << IB_USER_VERBS_CMD_MODIFY_QP)		|
		(1ull << IB_USER_VERBS_CMD_QUERY_QP)		|
		(1ull << IB_USER_VERBS_CMD_DESTROY_QP)		|
		(1ull << IB_USER_VERBS_CMD_ATTACH_MCAST)	|
		(1ull << IB_USER_VERBS_CMD_DETACH_MCAST)	|
		(1ull << IB_USER_VERBS_CMD_CREATE_SRQ)		|
		(1ull << IB_USER_VERBS_CMD_MODIFY_SRQ)		|
		(1ull << IB_USER_VERBS_CMD_QUERY_SRQ)		|
		(1ull << IB_USER_VERBS_CMD_DESTROY_SRQ)		|
		(1ull << IB_USER_VERBS_CMD_CREATE_XSRQ)		|
		(1ull << IB_USER_VERBS_CMD_OPEN_QP);
	dev->ib_dev.uverbs_ex_cmd_mask =
		(1ull << IB_USER_VERBS_EX_CMD_QUERY_DEVICE)	|
		(1ull << IB_USER_VERBS_EX_CMD_CREATE_CQ)	|
		(1ull << IB_USER_VERBS_EX_CMD_CREATE_QP);

	dev->ib_dev.query_device	= mlx5_ib_query_device;
	dev->ib_dev.query_port		= mlx5_ib_query_port;
	dev->ib_dev.get_link_layer	= mlx5_ib_port_link_layer;
	if (ll == IB_LINK_LAYER_ETHERNET)
		dev->ib_dev.get_netdev	= mlx5_ib_get_netdev;
	dev->ib_dev.query_gid		= mlx5_ib_query_gid;
	dev->ib_dev.add_gid		= mlx5_ib_add_gid;
	dev->ib_dev.del_gid		= mlx5_ib_del_gid;
	dev->ib_dev.query_pkey		= mlx5_ib_query_pkey;
	dev->ib_dev.modify_device	= mlx5_ib_modify_device;
	dev->ib_dev.modify_port		= mlx5_ib_modify_port;
	dev->ib_dev.alloc_ucontext	= mlx5_ib_alloc_ucontext;
	dev->ib_dev.dealloc_ucontext	= mlx5_ib_dealloc_ucontext;
	dev->ib_dev.mmap		= mlx5_ib_mmap;
	dev->ib_dev.alloc_pd		= mlx5_ib_alloc_pd;
	dev->ib_dev.dealloc_pd		= mlx5_ib_dealloc_pd;
	dev->ib_dev.create_ah		= mlx5_ib_create_ah;
	dev->ib_dev.query_ah		= mlx5_ib_query_ah;
	dev->ib_dev.destroy_ah		= mlx5_ib_destroy_ah;
	dev->ib_dev.create_srq		= mlx5_ib_create_srq;
	dev->ib_dev.modify_srq		= mlx5_ib_modify_srq;
	dev->ib_dev.query_srq		= mlx5_ib_query_srq;
	dev->ib_dev.destroy_srq		= mlx5_ib_destroy_srq;
	dev->ib_dev.post_srq_recv	= mlx5_ib_post_srq_recv;
	dev->ib_dev.create_qp		= mlx5_ib_create_qp;
	dev->ib_dev.modify_qp		= mlx5_ib_modify_qp;
	dev->ib_dev.query_qp		= mlx5_ib_query_qp;
	dev->ib_dev.destroy_qp		= mlx5_ib_destroy_qp;
	dev->ib_dev.post_send		= mlx5_ib_post_send;
	dev->ib_dev.post_recv		= mlx5_ib_post_recv;
	dev->ib_dev.create_cq		= mlx5_ib_create_cq;
	dev->ib_dev.modify_cq		= mlx5_ib_modify_cq;
	dev->ib_dev.resize_cq		= mlx5_ib_resize_cq;
	dev->ib_dev.destroy_cq		= mlx5_ib_destroy_cq;
	dev->ib_dev.poll_cq		= mlx5_ib_poll_cq;
	dev->ib_dev.req_notify_cq	= mlx5_ib_arm_cq;
	dev->ib_dev.get_dma_mr		= mlx5_ib_get_dma_mr;
	dev->ib_dev.reg_user_mr		= mlx5_ib_reg_user_mr;
	dev->ib_dev.rereg_user_mr	= mlx5_ib_rereg_user_mr;
	dev->ib_dev.dereg_mr		= mlx5_ib_dereg_mr;
	dev->ib_dev.attach_mcast	= mlx5_ib_mcg_attach;
	dev->ib_dev.detach_mcast	= mlx5_ib_mcg_detach;
	dev->ib_dev.process_mad		= mlx5_ib_process_mad;
	dev->ib_dev.alloc_mr		= mlx5_ib_alloc_mr;
	dev->ib_dev.map_mr_sg		= mlx5_ib_map_mr_sg;
	dev->ib_dev.check_mr_status	= mlx5_ib_check_mr_status;
	dev->ib_dev.get_port_immutable  = mlx5_port_immutable;

	mlx5_ib_internal_fill_odp_caps(dev);

	if (MLX5_CAP_GEN(mdev, imaicl)) {
		dev->ib_dev.alloc_mw		= mlx5_ib_alloc_mw;
		dev->ib_dev.dealloc_mw		= mlx5_ib_dealloc_mw;
		dev->ib_dev.uverbs_cmd_mask |=
			(1ull << IB_USER_VERBS_CMD_ALLOC_MW)	|
			(1ull << IB_USER_VERBS_CMD_DEALLOC_MW);
	}

	if (MLX5_CAP_GEN(mdev, xrc)) {
		dev->ib_dev.alloc_xrcd = mlx5_ib_alloc_xrcd;
		dev->ib_dev.dealloc_xrcd = mlx5_ib_dealloc_xrcd;
		dev->ib_dev.uverbs_cmd_mask |=
			(1ull << IB_USER_VERBS_CMD_OPEN_XRCD) |
			(1ull << IB_USER_VERBS_CMD_CLOSE_XRCD);
	}

	if (mlx5_ib_port_link_layer(&dev->ib_dev, 1) ==
	    IB_LINK_LAYER_ETHERNET) {
		dev->ib_dev.create_flow	= mlx5_ib_create_flow;
		dev->ib_dev.destroy_flow = mlx5_ib_destroy_flow;
		dev->ib_dev.uverbs_ex_cmd_mask |=
			(1ull << IB_USER_VERBS_EX_CMD_CREATE_FLOW) |
			(1ull << IB_USER_VERBS_EX_CMD_DESTROY_FLOW);
	}
	err = init_node_data(dev);
	if (err)
		goto err_dealloc;

	mutex_init(&dev->flow_db.lock);
	mutex_init(&dev->cap_mask_mutex);

	if (ll == IB_LINK_LAYER_ETHERNET) {
		err = mlx5_enable_roce(dev);
		if (err)
			goto err_dealloc;
	}

	err = create_dev_resources(&dev->devr);
	if (err)
		goto err_disable_roce;

	err = mlx5_ib_odp_init_one(dev);
	if (err)
		goto err_rsrc;

	err = ib_register_device(&dev->ib_dev, NULL);
	if (err)
		goto err_odp;

	err = create_umr_res(dev);
	if (err)
		goto err_dev;

	for (i = 0; i < ARRAY_SIZE(mlx5_class_attributes); i++) {
		err = device_create_file(&dev->ib_dev.dev,
					 mlx5_class_attributes[i]);
		if (err)
			goto err_umrc;
	}

	dev->ib_active = true;

	return dev;

err_umrc:
	destroy_umrc_res(dev);

err_dev:
	ib_unregister_device(&dev->ib_dev);

err_odp:
	mlx5_ib_odp_remove_one(dev);

err_rsrc:
	destroy_dev_resources(&dev->devr);

err_disable_roce:
	if (ll == IB_LINK_LAYER_ETHERNET)
		mlx5_disable_roce(dev);

err_dealloc:
	ib_dealloc_device((struct ib_device *)dev);

	return NULL;
}

static void mlx5_ib_remove(struct mlx5_core_dev *mdev, void *context)
{
	struct mlx5_ib_dev *dev = context;
	enum rdma_link_layer ll = mlx5_ib_port_link_layer(&dev->ib_dev, 1);

	ib_unregister_device(&dev->ib_dev);
	destroy_umrc_res(dev);
	mlx5_ib_odp_remove_one(dev);
	destroy_dev_resources(&dev->devr);
	if (ll == IB_LINK_LAYER_ETHERNET)
		mlx5_disable_roce(dev);
	ib_dealloc_device(&dev->ib_dev);
}

static struct mlx5_interface mlx5_ib_interface = {
	.add            = mlx5_ib_add,
	.remove         = mlx5_ib_remove,
	.event          = mlx5_ib_event,
	.protocol	= MLX5_INTERFACE_PROTOCOL_IB,
};

static int __init mlx5_ib_init(void)
{
	int err;

	if (deprecated_prof_sel != 2)
		pr_warn("prof_sel is deprecated for mlx5_ib, set it for mlx5_core\n");

	err = mlx5_ib_odp_init();
	if (err)
		return err;

	err = mlx5_register_interface(&mlx5_ib_interface);
	if (err)
		goto clean_odp;

	return err;

clean_odp:
	mlx5_ib_odp_cleanup();
	return err;
}

static void __exit mlx5_ib_cleanup(void)
{
	mlx5_unregister_interface(&mlx5_ib_interface);
	mlx5_ib_odp_cleanup();
}

module_init(mlx5_ib_init);
module_exit(mlx5_ib_cleanup);