Merge branch 'mlx5-next' into rdma.git for-next

From
git://git.kernel.org/pub/scm/linux/kernel/git/mellanox/linux

Required for dependencies on the next series

* branch 'mlx5-next':
  net/mlx5: E-Switch, add a new prio to be used by the RDMA side
  net/mlx5: E-Switch, don't use hardcoded values for FDB prios
  net/mlx5: Fix false compilation warning
  net/mlx5: Expose MPEIN (Management PCIE INfo) register layout
  net/mlx5: Add rate limit print macros
  net/mlx5: Add explicit bar address field
  net/mlx5: Replace dev_err/warn/info by mlx5_core_err/warn/info
  net/mlx5: Use dev->priv.name instead of dev_name
  net/mlx5: Make mlx5_core messages independent from mdev->pdev
  net/mlx5: Break load_one into three stages
  net/mlx5: Function setup/teardown procedures
  net/mlx5: Move health and page alloc init to mdev_init
  net/mlx5: Split mdev init and pci init
  net/mlx5: Remove redundant init functions parameter
  net/mlx5: Remove spinlock support from mlx5_write64
  net/mlx5: Remove unused MLX5_*_DOORBELL_LOCK macros
Signed-off-by: Jason Gunthorpe <jgg@mellanox.com>

drivers/infiniband/hw/mlx5/cmd.c

@@ -148,7 +148,7 @@ int mlx5_cmd_alloc_memic(struct mlx5_memic *memic, phys_addr_t *addr,
 			return ret;
 		}
 
-		*addr = pci_resource_start(dev->pdev, 0) +
+		*addr = dev->bar_addr +
 			MLX5_GET64(alloc_memic_out, out, memic_start_addr);
 
 		return 0;
@@ -167,7 +167,7 @@ int mlx5_cmd_dealloc_memic(struct mlx5_memic *memic, u64 addr, u64 length)
 	u64 start_page_idx;
 	int err;
 
-	addr -= pci_resource_start(dev->pdev, 0);
+	addr -= dev->bar_addr;
 	start_page_idx = (addr - hw_start_addr) >> PAGE_SHIFT;
 
 	MLX5_SET(dealloc_memic_in, in, opcode, MLX5_CMD_OP_DEALLOC_MEMIC);

drivers/infiniband/hw/mlx5/main.c

@@ -2009,7 +2009,7 @@ static phys_addr_t uar_index2pfn(struct mlx5_ib_dev *dev,
 
 	fw_uars_per_page = MLX5_CAP_GEN(dev->mdev, uar_4k) ? MLX5_UARS_IN_PAGE : 1;
 
-	return (pci_resource_start(dev->mdev->pdev, 0) >> PAGE_SHIFT) + uar_idx / fw_uars_per_page;
+	return (dev->mdev->bar_addr >> PAGE_SHIFT) + uar_idx / fw_uars_per_page;
 }
 
 static int get_command(unsigned long offset)
@@ -2199,7 +2199,7 @@ static int dm_mmap(struct ib_ucontext *context, struct vm_area_struct *vma)
 	    page_idx + npages)
 		return -EINVAL;
 
-	pfn = ((pci_resource_start(dev->mdev->pdev, 0) +
+	pfn = ((dev->mdev->bar_addr +
 	      MLX5_CAP64_DEV_MEM(dev->mdev, memic_bar_start_addr)) >>
 	      PAGE_SHIFT) +
 	      page_idx;
@@ -2283,7 +2283,7 @@ struct ib_dm *mlx5_ib_alloc_dm(struct ib_device *ibdev,
 		goto err_free;
 
 	start_offset = memic_addr & ~PAGE_MASK;
-	page_idx = (memic_addr - pci_resource_start(memic->dev->pdev, 0) -
+	page_idx = (memic_addr - memic->dev->bar_addr -
 		    MLX5_CAP64_DEV_MEM(memic->dev, memic_bar_start_addr)) >>
 		    PAGE_SHIFT;
 
@@ -2326,7 +2326,7 @@ int mlx5_ib_dealloc_dm(struct ib_dm *ibdm, struct uverbs_attr_bundle *attrs)
 	if (ret)
 		return ret;
 
-	page_idx = (dm->dev_addr - pci_resource_start(memic->dev->pdev, 0) -
+	page_idx = (dm->dev_addr - memic->dev->bar_addr -
 		    MLX5_CAP64_DEV_MEM(memic->dev, memic_bar_start_addr)) >>
 		    PAGE_SHIFT;
 	bitmap_clear(rdma_udata_to_drv_context(

drivers/infiniband/hw/mlx5/mr.c

@@ -1194,8 +1194,7 @@ static struct ib_mr *mlx5_ib_get_memic_mr(struct ib_pd *pd, u64 memic_addr,
 	MLX5_SET64(mkc, mkc, len, length);
 	MLX5_SET(mkc, mkc, pd, to_mpd(pd)->pdn);
 	MLX5_SET(mkc, mkc, qpn, 0xffffff);
-	MLX5_SET64(mkc, mkc, start_addr,
-		   memic_addr - pci_resource_start(dev->mdev->pdev, 0));
+	MLX5_SET64(mkc, mkc, start_addr, memic_addr - dev->mdev->bar_addr);
 
 	err = mlx5_core_create_mkey(mdev, &mr->mmkey, in, inlen);
 	if (err)

drivers/infiniband/hw/mlx5/qp.c

@@ -5126,7 +5126,7 @@ static int _mlx5_ib_post_send(struct ib_qp *ibqp, const struct ib_send_wr *wr,
 		wmb();
 
 		/* currently we support only regular doorbells */
-		mlx5_write64((__be32 *)ctrl, bf->bfreg->map + bf->offset, NULL);
+		mlx5_write64((__be32 *)ctrl, bf->bfreg->map + bf->offset);
 		/* Make sure doorbells don't leak out of SQ spinlock
 		 * and reach the HCA out of order.
 		 */

drivers/net/ethernet/mellanox/mlx5/core/cmd.c

@@ -1347,7 +1347,7 @@ 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));
+		 dev->priv.name);
 }
 
 static void clean_debug_files(struct mlx5_core_dev *dev)
@@ -1902,7 +1902,7 @@ int mlx5_cmd_init(struct mlx5_core_dev *dev)
 	memset(cmd, 0, sizeof(*cmd));
 	cmd_if_rev = cmdif_rev(dev);
 	if (cmd_if_rev != CMD_IF_REV) {
-		dev_err(&dev->pdev->dev,
+		mlx5_core_err(dev,
 			      "Driver cmdif rev(%d) differs from firmware's(%d)\n",
 			      CMD_IF_REV, cmd_if_rev);
 		return -EINVAL;
@@ -1921,14 +1921,14 @@ int mlx5_cmd_init(struct mlx5_core_dev *dev)
 	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",
+		mlx5_core_err(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");
+		mlx5_core_err(dev, "command queue size overflow\n");
 		err = -EINVAL;
 		goto err_free_page;
 	}
@@ -1939,7 +1939,7 @@ int mlx5_cmd_init(struct mlx5_core_dev *dev)
 
 	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",
+		mlx5_core_err(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;
@@ -1956,7 +1956,7 @@ int mlx5_cmd_init(struct mlx5_core_dev *dev)
 	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");
+		mlx5_core_err(dev, "invalid command queue address\n");
 		err = -ENOMEM;
 		goto err_free_page;
 	}
@@ -1976,7 +1976,7 @@ int mlx5_cmd_init(struct mlx5_core_dev *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");
+		mlx5_core_err(dev, "failed to create command workqueue\n");
 		err = -ENOMEM;
 		goto err_cache;
 	}

drivers/net/ethernet/mellanox/mlx5/core/diag/fw_tracer_tracepoint.h

@@ -47,7 +47,7 @@ TRACE_EVENT(mlx5_fw,
 	TP_ARGS(tracer, trace_timestamp, lost, event_id, msg),
 
 	TP_STRUCT__entry(
-		__string(dev_name, dev_name(&tracer->dev->pdev->dev))
+		__string(dev_name, tracer->dev->priv.name)
 		__field(u64, trace_timestamp)
 		__field(bool, lost)
 		__field(u8, event_id)
@@ -55,7 +55,7 @@ TRACE_EVENT(mlx5_fw,
 	),
 
 	TP_fast_assign(
-		__assign_str(dev_name, dev_name(&tracer->dev->pdev->dev));
+		__assign_str(dev_name, tracer->dev->priv.name);
 		__entry->trace_timestamp = trace_timestamp;
 		__entry->lost = lost;
 		__entry->event_id = event_id;

drivers/net/ethernet/mellanox/mlx5/core/en.h

@@ -929,7 +929,7 @@ void mlx5e_notify_hw(struct mlx5_wq_cyc *wq, u16 pc,
 	 */
 	wmb();
 
-	mlx5_write64((__be32 *)ctrl, uar_map, NULL);
+	mlx5_write64((__be32 *)ctrl, uar_map);
 }
 
 static inline void mlx5e_cq_arm(struct mlx5e_cq *cq)

drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c

@@ -43,11 +43,6 @@
 #include "ecpf.h"
 #include "lib/eq.h"
 
-enum {
-	FDB_FAST_PATH = 0,
-	FDB_SLOW_PATH
-};
-
 /* There are two match-all miss flows, one for unicast dst mac and
  * one for multicast.
  */

drivers/net/ethernet/mellanox/mlx5/core/fpga/conn.c

@@ -135,7 +135,7 @@ static void mlx5_fpga_conn_notify_hw(struct mlx5_fpga_conn *conn, void *wqe)
 	*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);
+	mlx5_write64(wqe, conn->fdev->conn_res.uar->map + MLX5_BF_OFFSET);
 }
 
 static void mlx5_fpga_conn_post_send(struct mlx5_fpga_conn *conn,

drivers/net/ethernet/mellanox/mlx5/core/fpga/core.h

@@ -37,6 +37,7 @@
 
 #include <linux/mlx5/eq.h>
 
+#include "mlx5_core.h"
 #include "lib/eq.h"
 #include "fpga/cmd.h"
 
@@ -62,26 +63,26 @@ struct mlx5_fpga_device {
 };
 
 #define mlx5_fpga_dbg(__adev, format, ...) \
-	dev_dbg(&(__adev)->mdev->pdev->dev, "FPGA: %s:%d:(pid %d): " format, \
+	mlx5_core_dbg((__adev)->mdev, "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, \
+	mlx5_core_err((__adev)->mdev, "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, \
+	mlx5_core_warn((__adev)->mdev, "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: " \
+	mlx5_core_err_rl((__adev)->mdev, "FPGA: %s:%d: " \
 			 format, __func__, __LINE__, ##__VA_ARGS__)
 
 #define mlx5_fpga_notice(__adev, format, ...) \
-	dev_notice(&(__adev)->mdev->pdev->dev, "FPGA: " format, ##__VA_ARGS__)
+	mlx5_core_info((__adev)->mdev, "FPGA: " format, ##__VA_ARGS__)
 
 #define mlx5_fpga_info(__adev, format, ...) \
-	dev_info(&(__adev)->mdev->pdev->dev, "FPGA: " format, ##__VA_ARGS__)
+	mlx5_core_info((__adev)->mdev, "FPGA: " format, ##__VA_ARGS__)
 
 int mlx5_fpga_init(struct mlx5_core_dev *mdev);
 void mlx5_fpga_cleanup(struct mlx5_core_dev *mdev);

drivers/net/ethernet/mellanox/mlx5/core/fs_core.c

@@ -819,7 +819,7 @@ static int update_root_ft_create(struct mlx5_flow_table *ft, struct fs_prio
 	struct mlx5_flow_root_namespace *root = find_root(&prio->node);
 	struct mlx5_ft_underlay_qp *uqp;
 	int min_level = INT_MAX;
-	int err;
+	int err = 0;
 	u32 qpn;
 
 	if (root->root_ft)
@@ -2516,8 +2516,16 @@ static int init_fdb_root_ns(struct mlx5_flow_steering *steering)
 	if (!steering->fdb_sub_ns)
 		return -ENOMEM;
 
+	maj_prio = fs_create_prio(&steering->fdb_root_ns->ns, FDB_BYPASS_PATH,
+				  1);
+	if (IS_ERR(maj_prio)) {
+		err = PTR_ERR(maj_prio);
+		goto out_err;
+	}
+
 	levels = 2 * FDB_MAX_PRIO * (FDB_MAX_CHAIN + 1);
-	maj_prio = fs_create_prio_chained(&steering->fdb_root_ns->ns, 0,
+	maj_prio = fs_create_prio_chained(&steering->fdb_root_ns->ns,
+					  FDB_FAST_PATH,
 					  levels);
 	if (IS_ERR(maj_prio)) {
 		err = PTR_ERR(maj_prio);
@@ -2542,7 +2550,7 @@ static int init_fdb_root_ns(struct mlx5_flow_steering *steering)
 		steering->fdb_sub_ns[chain] = ns;
 	}
 
-	maj_prio = fs_create_prio(&steering->fdb_root_ns->ns, 1, 1);
+	maj_prio = fs_create_prio(&steering->fdb_root_ns->ns, FDB_SLOW_PATH, 1);
 	if (IS_ERR(maj_prio)) {
 		err = PTR_ERR(maj_prio);
 		goto out_err;

drivers/net/ethernet/mellanox/mlx5/core/health.c

@@ -152,11 +152,11 @@ static void health_recover(struct work_struct *work)
 
 	nic_state = mlx5_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");
+		mlx5_core_err(dev, "health recovery flow aborted since the nic state is invalid\n");
 		return;
 	}
 
-	dev_err(&dev->pdev->dev, "starting health recovery flow\n");
+	mlx5_core_err(dev, "starting health recovery flow\n");
 	mlx5_recover_device(dev);
 }
 
@@ -180,7 +180,7 @@ static void health_care(struct work_struct *work)
 	if (!test_bit(MLX5_DROP_NEW_RECOVERY_WORK, &health->flags))
 		schedule_delayed_work(&health->recover_work, recover_delay);
 	else
-		dev_err(&dev->pdev->dev,
+		mlx5_core_err(dev,
 			      "new health works are not permitted at this stage\n");
 	spin_unlock_irqrestore(&health->wq_lock, flags);
 }
@@ -228,18 +228,22 @@ static void print_health_info(struct mlx5_core_dev *dev)
 		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));
+		mlx5_core_err(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));
+	mlx5_core_err(dev, "assert_exit_ptr 0x%08x\n",
+		      ioread32be(&h->assert_exit_ptr));
+	mlx5_core_err(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));
+	mlx5_core_err(dev, "fw_ver %s\n", fw_str);
+	mlx5_core_err(dev, "hw_id 0x%08x\n", ioread32be(&h->hw_id));
+	mlx5_core_err(dev, "irisc_index %d\n", ioread8(&h->irisc_index));
+	mlx5_core_err(dev, "synd 0x%x: %s\n", ioread8(&h->synd),
+		      hsynd_str(ioread8(&h->synd)));
+	mlx5_core_err(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);
+	mlx5_core_err(dev, "raw fw_ver 0x%08x\n", fw);
 }
 
 static unsigned long get_next_poll_jiffies(void)
@@ -262,8 +266,7 @@ void mlx5_trigger_health_work(struct mlx5_core_dev *dev)
 	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");
+		mlx5_core_err(dev, "new health works are not permitted at this stage\n");
 	spin_unlock_irqrestore(&health->wq_lock, flags);
 }
 
@@ -284,7 +287,7 @@ static void poll_health(struct timer_list *t)
 
 	health->prev = count;
 	if (health->miss_counter == MAX_MISSES) {
-		dev_err(&dev->pdev->dev, "device's health compromised - reached miss count\n");
+		mlx5_core_err(dev, "device's health compromised - reached miss count\n");
 		print_health_info(dev);
 	}
 
@@ -352,6 +355,13 @@ void mlx5_drain_health_recovery(struct mlx5_core_dev *dev)
 	cancel_delayed_work_sync(&dev->priv.health.recover_work);
 }
 
+void mlx5_health_flush(struct mlx5_core_dev *dev)
+{
+	struct mlx5_core_health *health = &dev->priv.health;
+
+	flush_workqueue(health->wq);
+}
+
 void mlx5_health_cleanup(struct mlx5_core_dev *dev)
 {
 	struct mlx5_core_health *health = &dev->priv.health;
@@ -370,7 +380,7 @@ int mlx5_health_init(struct mlx5_core_dev *dev)
 		return -ENOMEM;
 
 	strcpy(name, "mlx5_health");
-	strcat(name, dev_name(&dev->pdev->dev));
+	strcat(name, dev->priv.name);
 	health->wq = create_singlethread_workqueue(name);
 	kfree(name);
 	if (!health->wq)

drivers/net/ethernet/mellanox/mlx5/core/main.c

@@ -587,24 +587,23 @@ static int handle_hca_cap(struct mlx5_core_dev *dev)
 
 static int set_hca_cap(struct mlx5_core_dev *dev)
 {
-	struct pci_dev *pdev = dev->pdev;
 	int err;
 
 	err = handle_hca_cap(dev);
 	if (err) {
-		dev_err(&pdev->dev, "handle_hca_cap failed\n");
+		mlx5_core_err(dev, "handle_hca_cap failed\n");
 		goto out;
 	}
 
 	err = handle_hca_cap_atomic(dev);
 	if (err) {
-		dev_err(&pdev->dev, "handle_hca_cap_atomic failed\n");
+		mlx5_core_err(dev, "handle_hca_cap_atomic failed\n");
 		goto out;
 	}
 
 	err = handle_hca_cap_odp(dev);
 	if (err) {
-		dev_err(&pdev->dev, "handle_hca_cap_odp failed\n");
+		mlx5_core_err(dev, "handle_hca_cap_odp failed\n");
 		goto out;
 	}
 
@@ -736,36 +735,29 @@ static int mlx5_core_set_issi(struct mlx5_core_dev *dev)
 	return -EOPNOTSUPP;
 }
 
-static int mlx5_pci_init(struct mlx5_core_dev *dev, struct mlx5_priv *priv)
+static int mlx5_pci_init(struct mlx5_core_dev *dev, struct pci_dev *pdev,
+			 const struct pci_device_id *id)
 {
-	struct pci_dev *pdev = dev->pdev;
+	struct mlx5_priv *priv = &dev->priv;
 	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);
+	dev->pdev = pdev;
+	priv->pci_dev_data = id->driver_data;
 
-	mutex_init(&priv->alloc_mutex);
+	pci_set_drvdata(dev->pdev, dev);
 
+	dev->bar_addr = pci_resource_start(pdev, 0);
 	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;
+		mlx5_core_err(dev, "Cannot enable PCI device, aborting\n");
+		return err;
 	}
 
 	err = request_bar(pdev);
 	if (err) {
-		dev_err(&pdev->dev, "error requesting BARs, aborting\n");
+		mlx5_core_err(dev, "error requesting BARs, aborting\n");
 		goto err_disable;
 	}
 
@@ -773,7 +765,7 @@ static int mlx5_pci_init(struct mlx5_core_dev *dev, struct mlx5_priv *priv)
 
 	err = set_dma_caps(pdev);
 	if (err) {
-		dev_err(&pdev->dev, "Failed setting DMA capabilities mask, aborting\n");
+		mlx5_core_err(dev, "Failed setting DMA capabilities mask, aborting\n");
 		goto err_clr_master;
 	}
 
@@ -782,11 +774,11 @@ static int mlx5_pci_init(struct mlx5_core_dev *dev, struct mlx5_priv *priv)
 	    pci_enable_atomic_ops_to_root(pdev, PCI_EXP_DEVCAP2_ATOMIC_COMP128))
 		mlx5_core_dbg(dev, "Enabling pci atomics failed\n");
 
-	dev->iseg_base = pci_resource_start(dev->pdev, 0);
+	dev->iseg_base = dev->bar_addr;
 	dev->iseg = ioremap(dev->iseg_base, sizeof(*dev->iseg));
 	if (!dev->iseg) {
 		err = -ENOMEM;
-		dev_err(&pdev->dev, "Failed mapping initialization segment, aborting\n");
+		mlx5_core_err(dev, "Failed mapping initialization segment, aborting\n");
 		goto err_clr_master;
 	}
 
@@ -797,52 +789,47 @@ static int mlx5_pci_init(struct mlx5_core_dev *dev, struct mlx5_priv *priv)
 	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)
+static void mlx5_pci_close(struct mlx5_core_dev *dev)
 {
 	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)
+static int mlx5_init_once(struct mlx5_core_dev *dev)
 {
-	struct pci_dev *pdev = dev->pdev;
 	int err;
 
-	priv->devcom = mlx5_devcom_register_device(dev);
-	if (IS_ERR(priv->devcom))
-		dev_err(&pdev->dev, "failed to register with devcom (0x%p)\n",
-			priv->devcom);
+	dev->priv.devcom = mlx5_devcom_register_device(dev);
+	if (IS_ERR(dev->priv.devcom))
+		mlx5_core_err(dev, "failed to register with devcom (0x%p)\n",
+			      dev->priv.devcom);
 
 	err = mlx5_query_board_id(dev);
 	if (err) {
-		dev_err(&pdev->dev, "query board id failed\n");
+		mlx5_core_err(dev, "query board id failed\n");
 		goto err_devcom;
 	}
 
 	err = mlx5_eq_table_init(dev);
 	if (err) {
-		dev_err(&pdev->dev, "failed to initialize eq\n");
+		mlx5_core_err(dev, "failed to initialize eq\n");
 		goto err_devcom;
 	}
 
 	err = mlx5_events_init(dev);
 	if (err) {
-		dev_err(&pdev->dev, "failed to initialize events\n");
+		mlx5_core_err(dev, "failed to initialize events\n");
 		goto err_eq_cleanup;
 	}
 
 	err = mlx5_cq_debugfs_init(dev);
 	if (err) {
-		dev_err(&pdev->dev, "failed to initialize cq debugfs\n");
+		mlx5_core_err(dev, "failed to initialize cq debugfs\n");
 		goto err_events_cleanup;
 	}
 
@@ -858,31 +845,31 @@ static int mlx5_init_once(struct mlx5_core_dev *dev, struct mlx5_priv *priv)
 
 	err = mlx5_init_rl_table(dev);
 	if (err) {
-		dev_err(&pdev->dev, "Failed to init rate limiting\n");
+		mlx5_core_err(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);
+		mlx5_core_err(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);
+		mlx5_core_err(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);
+		mlx5_core_err(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);
+		mlx5_core_err(dev, "Failed to init fpga device %d\n", err);
 		goto err_sriov_cleanup;
 	}
 
@@ -932,93 +919,78 @@ static void mlx5_cleanup_once(struct mlx5_core_dev *dev)
 	mlx5_devcom_unregister_device(dev->priv.devcom);
 }
 
-static int mlx5_load_one(struct mlx5_core_dev *dev, struct mlx5_priv *priv,
-			 bool boot)
+static int mlx5_function_setup(struct mlx5_core_dev *dev, bool boot)
 {
-	struct pci_dev *pdev = dev->pdev;
 	int err;
 
-	dev->caps.embedded_cpu = mlx5_read_embedded_cpu(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 up, NOP\n",
-			 __func__);
-		goto out;
-	}
-
-	dev_info(&pdev->dev, "firmware version: %d.%d.%d\n", fw_rev_maj(dev),
+	mlx5_core_info(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",
+		mlx5_core_err(dev, "Firmware over %d MS in pre-initializing state, aborting\n",
 			      FW_PRE_INIT_TIMEOUT_MILI);
-		goto out_err;
+		return err;
 	}
 
 	err = mlx5_cmd_init(dev);
 	if (err) {
-		dev_err(&pdev->dev, "Failed initializing command interface, aborting\n");
-		goto out_err;
+		mlx5_core_err(dev, "Failed initializing command interface, aborting\n");
+		return 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",
+		mlx5_core_err(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");
+		mlx5_core_err(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");
+		mlx5_core_err(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");
+		mlx5_core_err(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");
+		mlx5_core_err(dev, "set_hca_ctrl failed\n");
 		goto reclaim_boot_pages;
 	}
 
 	err = set_hca_cap(dev);
 	if (err) {
-		dev_err(&pdev->dev, "set_hca_cap failed\n");
+		mlx5_core_err(dev, "set_hca_cap 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");
+		mlx5_core_err(dev, "failed to allocate init pages\n");
 		goto reclaim_boot_pages;
 	}
 
 	err = mlx5_cmd_init_hca(dev, sw_owner_id);
 	if (err) {
-		dev_err(&pdev->dev, "init hca failed\n");
+		mlx5_core_err(dev, "init hca failed\n");
 		goto reclaim_boot_pages;
 	}
 
@@ -1028,23 +1000,50 @@ static int mlx5_load_one(struct mlx5_core_dev *dev, struct mlx5_priv *priv,
 
 	err = mlx5_query_hca_caps(dev);
 	if (err) {
-		dev_err(&pdev->dev, "query hca failed\n");
-		goto err_stop_poll;
+		mlx5_core_err(dev, "query hca failed\n");
+		goto stop_health;
 	}
 
-	if (boot) {
-		err = mlx5_init_once(dev, priv);
+	return 0;
+
+stop_health:
+	mlx5_stop_health_poll(dev, boot);
+reclaim_boot_pages:
+	mlx5_reclaim_startup_pages(dev);
+err_disable_hca:
+	mlx5_core_disable_hca(dev, 0);
+err_cmd_cleanup:
+	mlx5_cmd_cleanup(dev);
+
+	return err;
+}
+
+static int mlx5_function_teardown(struct mlx5_core_dev *dev, bool boot)
+{
+	int err;
+
+	mlx5_stop_health_poll(dev, boot);
+	err = mlx5_cmd_teardown_hca(dev);
 	if (err) {
-			dev_err(&pdev->dev, "sw objs init failed\n");
-			goto err_stop_poll;
-		}
+		mlx5_core_err(dev, "tear_down_hca failed, skip cleanup\n");
+		return err;
 	}
+	mlx5_reclaim_startup_pages(dev);
+	mlx5_core_disable_hca(dev, 0);
+	mlx5_cmd_cleanup(dev);
+
+	return 0;
+}
+
+static int mlx5_load(struct mlx5_core_dev *dev)
+{
+	int err;
 
 	dev->priv.uar = mlx5_get_uars_page(dev);
 	if (IS_ERR(dev->priv.uar)) {
-		dev_err(&pdev->dev, "Failed allocating uar, aborting\n");
+		mlx5_core_err(dev, "Failed allocating uar, aborting\n");
 		err = PTR_ERR(dev->priv.uar);
-		goto err_get_uars;
+		return err;
 	}
 
 	mlx5_events_start(dev);
@@ -1052,132 +1051,155 @@ static int mlx5_load_one(struct mlx5_core_dev *dev, struct mlx5_priv *priv,
 
 	err = mlx5_eq_table_create(dev);
 	if (err) {
-		dev_err(&pdev->dev, "Failed to create EQs\n");
+		mlx5_core_err(dev, "Failed to create EQs\n");
 		goto err_eq_table;
 	}
 
 	err = mlx5_fw_tracer_init(dev->tracer);
 	if (err) {
-		dev_err(&pdev->dev, "Failed to init FW tracer\n");
+		mlx5_core_err(dev, "Failed to init FW tracer\n");
 		goto err_fw_tracer;
 	}
 
 	err = mlx5_fpga_device_start(dev);
 	if (err) {
-		dev_err(&pdev->dev, "fpga device start failed %d\n", err);
+		mlx5_core_err(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);
+		mlx5_core_err(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);
+		mlx5_core_err(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");
+		mlx5_core_err(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");
+		mlx5_core_err(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);
+		mlx5_core_err(dev, "sriov init failed %d\n", err);
 		goto err_sriov;
 	}
 
 	err = mlx5_ec_init(dev);
 	if (err) {
-		dev_err(&pdev->dev, "Failed to init embedded CPU\n");
+		mlx5_core_err(dev, "Failed to init embedded CPU\n");
 		goto err_ec;
 	}
 
-	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_ec_cleanup(dev);
-
 err_ec:
 	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_fw_tracer_cleanup(dev->tracer);
-
 err_fw_tracer:
 	mlx5_eq_table_destroy(dev);
-
 err_eq_table:
 	mlx5_pagealloc_stop(dev);
 	mlx5_events_stop(dev);
-	mlx5_put_uars_page(dev, priv->uar);
+	mlx5_put_uars_page(dev, dev->priv.uar);
+	return err;
+}
+
+static void mlx5_unload(struct mlx5_core_dev *dev)
+{
+	mlx5_ec_cleanup(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_fw_tracer_cleanup(dev->tracer);
+	mlx5_eq_table_destroy(dev);
+	mlx5_pagealloc_stop(dev);
+	mlx5_events_stop(dev);
+	mlx5_put_uars_page(dev, dev->priv.uar);
+}
 
-err_get_uars:
-	if (boot)
-		mlx5_cleanup_once(dev);
+static int mlx5_load_one(struct mlx5_core_dev *dev, bool boot)
+{
+	int err = 0;
 
-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;
+	dev->caps.embedded_cpu = mlx5_read_embedded_cpu(dev);
+	mutex_lock(&dev->intf_state_mutex);
+	if (test_bit(MLX5_INTERFACE_STATE_UP, &dev->intf_state)) {
+		mlx5_core_warn(dev, "interface is up, NOP\n");
+		goto out;
 	}
+	/* remove any previous indication of internal error */
+	dev->state = MLX5_DEVICE_STATE_UP;
 
-reclaim_boot_pages:
-	mlx5_reclaim_startup_pages(dev);
+	err = mlx5_function_setup(dev, boot);
+	if (err)
+		goto out;
 
-err_disable_hca:
-	mlx5_core_disable_hca(dev, 0);
+	if (boot) {
+		err = mlx5_init_once(dev);
+		if (err) {
+			mlx5_core_err(dev, "sw objs init failed\n");
+			goto function_teardown;
+		}
+	}
 
-err_cmd_cleanup:
-	mlx5_cmd_cleanup(dev);
+	err = mlx5_load(dev);
+	if (err)
+		goto err_load;
 
-out_err:
+	if (mlx5_device_registered(dev)) {
+		mlx5_attach_device(dev);
+	} else {
+		err = mlx5_register_device(dev);
+		if (err) {
+			mlx5_core_err(dev, "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 err;
+
+err_reg_dev:
+	mlx5_unload(dev);
+err_load:
+	if (boot)
+		mlx5_cleanup_once(dev);
+function_teardown:
+	mlx5_function_teardown(dev, boot);
 	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)
+static int mlx5_unload_one(struct mlx5_core_dev *dev, bool cleanup)
 {
 	int err = 0;
 
@@ -1186,7 +1208,7 @@ static int mlx5_unload_one(struct mlx5_core_dev *dev, struct mlx5_priv *priv,
 
 	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",
+		mlx5_core_warn(dev, "%s: interface is down, NOP\n",
 			       __func__);
 		if (cleanup)
 			mlx5_cleanup_once(dev);
@@ -1198,30 +1220,12 @@ static int mlx5_unload_one(struct mlx5_core_dev *dev, struct mlx5_priv *priv,
 	if (mlx5_device_registered(dev))
 		mlx5_detach_device(dev);
 
-	mlx5_ec_cleanup(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_fw_tracer_cleanup(dev->tracer);
-	mlx5_eq_table_destroy(dev);
-	mlx5_pagealloc_stop(dev);
-	mlx5_events_stop(dev);
-	mlx5_put_uars_page(dev, priv->uar);
+	mlx5_unload(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_reclaim_startup_pages(dev);
-	mlx5_core_disable_hca(dev, 0);
-	mlx5_cmd_cleanup(dev);
 
+	mlx5_function_teardown(dev, cleanup);
 out:
 	mutex_unlock(&dev->intf_state_mutex);
 	return err;
@@ -1238,29 +1242,15 @@ static const struct devlink_ops mlx5_devlink_ops = {
 #endif
 };
 
-#define MLX5_IB_MOD "mlx5_ib"
-static int init_one(struct pci_dev *pdev,
-		    const struct pci_device_id *id)
+static int mlx5_mdev_init(struct mlx5_core_dev *dev, int profile_idx, const char *name)
 {
-	struct mlx5_core_dev *dev;
-	struct devlink *devlink;
-	struct mlx5_priv *priv;
+	struct mlx5_priv *priv = &dev->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);
+	strncpy(priv->name, name, MLX5_MAX_NAME_LEN);
+	priv->name[MLX5_MAX_NAME_LEN - 1] = 0;
 
-	dev->pdev = pdev;
-	dev->profile = &profile[prof_sel];
+	dev->profile = &profile[profile_idx];
 
 	INIT_LIST_HEAD(&priv->ctx_list);
 	spin_lock_init(&priv->ctx_lock);
@@ -1272,25 +1262,72 @@ static int init_one(struct pci_dev *pdev,
 	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_dev;
+	mutex_init(&priv->alloc_mutex);
+	mutex_init(&priv->pgdir_mutex);
+	INIT_LIST_HEAD(&priv->pgdir_list);
+	spin_lock_init(&priv->mkey_lock);
+
+	priv->dbg_root = debugfs_create_dir(name, mlx5_debugfs_root);
+	if (!priv->dbg_root) {
+		pr_err("mlx5_core: %s error, Cannot create debugfs dir, aborting\n", name);
+		return -ENOMEM;
 	}
 
 	err = mlx5_health_init(dev);
-	if (err) {
-		dev_err(&pdev->dev, "mlx5_health_init failed with error code %d\n", err);
-		goto close_pci;
-	}
+	if (err)
+		goto err_health_init;
 
 	err = mlx5_pagealloc_init(dev);
 	if (err)
 		goto err_pagealloc_init;
 
-	err = mlx5_load_one(dev, priv, true);
+	return 0;
+
+err_pagealloc_init:
+	mlx5_health_cleanup(dev);
+err_health_init:
+	debugfs_remove(dev->priv.dbg_root);
+
+	return err;
+}
+
+static void mlx5_mdev_uninit(struct mlx5_core_dev *dev)
+{
+	mlx5_pagealloc_cleanup(dev);
+	mlx5_health_cleanup(dev);
+	debugfs_remove_recursive(dev->priv.dbg_root);
+}
+
+#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;
+	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);
+
+	err = mlx5_mdev_init(dev, prof_sel, dev_name(&pdev->dev));
+	if (err)
+		goto mdev_init_err;
+
+	err = mlx5_pci_init(dev, pdev, id);
+	if (err) {
+		mlx5_core_err(dev, "mlx5_pci_init failed with error code %d\n",
+			      err);
+		goto pci_init_err;
+	}
+
+	err = mlx5_load_one(dev, true);
 	if (err) {
-		dev_err(&pdev->dev, "mlx5_load_one failed with error code %d\n", err);
+		mlx5_core_err(dev, "mlx5_load_one failed with error code %d\n",
+			      err);
 		goto err_load_one;
 	}
 
@@ -1304,14 +1341,13 @@ static int init_one(struct pci_dev *pdev,
 	return 0;
 
 clean_load:
-	mlx5_unload_one(dev, priv, true);
+	mlx5_unload_one(dev, true);
+
 err_load_one:
-	mlx5_pagealloc_cleanup(dev);
-err_pagealloc_init:
-	mlx5_health_cleanup(dev);
-close_pci:
-	mlx5_pci_close(dev, priv);
-clean_dev:
+	mlx5_pci_close(dev);
+pci_init_err:
+	mlx5_mdev_uninit(dev);
+mdev_init_err:
 	devlink_free(devlink);
 
 	return err;
@@ -1321,20 +1357,18 @@ 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);
+	if (mlx5_unload_one(dev, true)) {
+		mlx5_core_err(dev, "mlx5_unload_one failed\n");
+		mlx5_health_flush(dev);
 		return;
 	}
 
-	mlx5_pagealloc_cleanup(dev);
-	mlx5_health_cleanup(dev);
-	mlx5_pci_close(dev, priv);
+	mlx5_pci_close(dev);
+	mlx5_mdev_uninit(dev);
 	devlink_free(devlink);
 }
 
@@ -1342,12 +1376,11 @@ 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_core_info(dev, "%s was called\n", __func__);
 
 	mlx5_enter_error_state(dev, false);
-	mlx5_unload_one(dev, priv, false);
+	mlx5_unload_one(dev, false);
 	/* In case of kernel call drain the health wq */
 	if (state) {
 		mlx5_drain_health_wq(dev);
@@ -1374,7 +1407,9 @@ static int wait_vital(struct pci_dev *pdev)
 		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);
+				mlx5_core_info(dev,
+					       "wait vital counter value 0x%x after %d iterations\n",
+					       count, i);
 				return 0;
 			}
 			last_count = count;
@@ -1390,12 +1425,12 @@ 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__);
+	mlx5_core_info(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);
+		mlx5_core_err(dev, "%s: mlx5_pci_enable_device failed with error code: %d\n",
+			      __func__, err);
 		return PCI_ERS_RESULT_DISCONNECT;
 	}
 
@@ -1404,7 +1439,7 @@ static pci_ers_result_t mlx5_pci_slot_reset(struct pci_dev *pdev)
 	pci_save_state(pdev);
 
 	if (wait_vital(pdev)) {
-		dev_err(&pdev->dev, "%s: wait_vital timed out\n", __func__);
+		mlx5_core_err(dev, "%s: wait_vital timed out\n", __func__);
 		return PCI_ERS_RESULT_DISCONNECT;
 	}
 
@@ -1414,17 +1449,16 @@ static pci_ers_result_t mlx5_pci_slot_reset(struct pci_dev *pdev)
 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__);
+	mlx5_core_info(dev, "%s was called\n", __func__);
 
-	err = mlx5_load_one(dev, priv, false);
+	err = mlx5_load_one(dev, false);
 	if (err)
-		dev_err(&pdev->dev, "%s: mlx5_load_one failed with error code: %d\n"
-			, __func__, err);
+		mlx5_core_err(dev, "%s: mlx5_load_one failed with error code: %d\n",
+			      __func__, err);
 	else
-		dev_info(&pdev->dev, "%s: device recovered\n", __func__);
+		mlx5_core_info(dev, "%s: device recovered\n", __func__);
 }
 
 static const struct pci_error_handlers mlx5_err_handler = {
@@ -1486,13 +1520,12 @@ static int mlx5_try_fast_unload(struct mlx5_core_dev *dev)
 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");
+	mlx5_core_info(dev, "Shutdown was called\n");
 	err = mlx5_try_fast_unload(dev);
 	if (err)
-		mlx5_unload_one(dev, priv, false);
+		mlx5_unload_one(dev, false);
 	mlx5_pci_disable_device(dev);
 }
 

drivers/net/ethernet/mellanox/mlx5/core/mlx5_core.h

@@ -48,12 +48,12 @@
 extern uint mlx5_core_debug_mask;
 
 #define mlx5_core_dbg(__dev, format, ...)				\
-	dev_dbg(&(__dev)->pdev->dev, "%s:%d:(pid %d): " format,		\
+	pr_debug("%s:%s:%d:(pid %d): " format, (__dev)->priv.name,      \
 		 __func__, __LINE__, current->pid,			\
 		 ##__VA_ARGS__)
 
 #define mlx5_core_dbg_once(__dev, format, ...)				\
-	dev_dbg_once(&(__dev)->pdev->dev, "%s:%d:(pid %d): " format,	\
+	pr_debug_once("%s:%s:%d:(pid %d): " format, (__dev)->priv.name, \
 		     __func__, __LINE__, current->pid,			\
 		     ##__VA_ARGS__)
 
@@ -64,28 +64,37 @@ do {									\
 } while (0)
 
 #define mlx5_core_err(__dev, format, ...)				\
-	dev_err(&(__dev)->pdev->dev, "%s:%d:(pid %d): " format,	\
+	pr_err("%s:%s:%d:(pid %d): " format, (__dev)->priv.name,        \
 		__func__, __LINE__, current->pid,	\
 	       ##__VA_ARGS__)
 
 #define mlx5_core_err_rl(__dev, format, ...)				     \
-	dev_err_ratelimited(&(__dev)->pdev->dev,			\
-			   "%s:%d:(pid %d): " format,			\
+	pr_err_ratelimited("%s:%s:%d:(pid %d): " format, (__dev)->priv.name, \
 			   __func__, __LINE__, current->pid,		     \
 			   ##__VA_ARGS__)
 
 #define mlx5_core_warn(__dev, format, ...)				\
-	dev_warn(&(__dev)->pdev->dev, "%s:%d:(pid %d): " format,	\
+	pr_warn("%s:%s:%d:(pid %d): " format, (__dev)->priv.name,       \
 		 __func__, __LINE__, current->pid,			\
 		##__VA_ARGS__)
 
 #define mlx5_core_warn_once(__dev, format, ...)				\
-	dev_warn_once(&(__dev)->pdev->dev, "%s:%d:(pid %d): " format,	\
+	pr_warn_once("%s:%s:%d:(pid %d): " format, (__dev)->priv.name,  \
+		      __func__, __LINE__, current->pid,			\
+		      ##__VA_ARGS__)
+
+#define mlx5_core_warn_rl(__dev, format, ...)				      \
+	pr_warn_ratelimited("%s:%s:%d:(pid %d): " format, (__dev)->priv.name, \
 			   __func__, __LINE__, current->pid,		      \
 			   ##__VA_ARGS__)
 
 #define mlx5_core_info(__dev, format, ...)				\
-	dev_info(&(__dev)->pdev->dev, format, ##__VA_ARGS__)
+	pr_info("%s " format, (__dev)->priv.name, ##__VA_ARGS__)
+
+#define mlx5_core_info_rl(__dev, format, ...)				      \
+	pr_info_ratelimited("%s:%s:%d:(pid %d): " format, (__dev)->priv.name, \
+			   __func__, __LINE__, current->pid,		      \
+			   ##__VA_ARGS__)
 
 enum {
 	MLX5_CMD_DATA, /* print command payload only */

drivers/net/ethernet/mellanox/mlx5/core/uar.c

@@ -79,7 +79,7 @@ static u64 uar2pfn(struct mlx5_core_dev *mdev, u32 index)
 	else
 		system_page_index = index;
 
-	return (pci_resource_start(mdev->pdev, 0) >> PAGE_SHIFT) + system_page_index;
+	return (mdev->bar_addr >> PAGE_SHIFT) + system_page_index;
 }
 
 static void up_rel_func(struct kref *kref)

include/linux/mlx5/cq.h

@@ -170,7 +170,7 @@ static inline void mlx5_cq_arm(struct mlx5_core_cq *cq, u32 cmd,
 	doorbell[0] = cpu_to_be32(sn << 28 | cmd | ci);
 	doorbell[1] = cpu_to_be32(cq->cqn);
 
-	mlx5_write64(doorbell, uar_page + MLX5_CQ_DOORBELL, NULL);
+	mlx5_write64(doorbell, uar_page + MLX5_CQ_DOORBELL);
 }
 
 static inline void mlx5_cq_hold(struct mlx5_core_cq *cq)

include/linux/mlx5/doorbell.h

@@ -36,46 +36,25 @@
 #define MLX5_BF_OFFSET	      0x800
 #define MLX5_CQ_DOORBELL      0x20
 
-#if BITS_PER_LONG == 64
 /* Assume that we can just write a 64-bit doorbell atomically.  s390
  * actually doesn't have writeq() but S/390 systems don't even have
  * PCI so we won't worry about it.
+ *
+ * Note that the write is not atomic on 32-bit systems! In contrast to 64-bit
+ * ones, it requires proper locking. mlx5_write64 doesn't do any locking, so use
+ * it at your own discretion, protected by some kind of lock on 32 bits.
+ *
+ * TODO: use write{q,l}_relaxed()
  */
 
-#define MLX5_DECLARE_DOORBELL_LOCK(name)
-#define MLX5_INIT_DOORBELL_LOCK(ptr)    do { } while (0)
-#define MLX5_GET_DOORBELL_LOCK(ptr)      (NULL)
-
-static inline void mlx5_write64(__be32 val[2], void __iomem *dest,
-				spinlock_t *doorbell_lock)
+static inline void mlx5_write64(__be32 val[2], void __iomem *dest)
 {
+#if BITS_PER_LONG == 64
 	__raw_writeq(*(u64 *)val, dest);
-}
-
 #else
-
-/* Just fall back to a spinlock to protect the doorbell if
- * BITS_PER_LONG is 32 -- there's no portable way to do atomic 64-bit
- * MMIO writes.
- */
-
-#define MLX5_DECLARE_DOORBELL_LOCK(name) spinlock_t name;
-#define MLX5_INIT_DOORBELL_LOCK(ptr)     spin_lock_init(ptr)
-#define MLX5_GET_DOORBELL_LOCK(ptr)      (ptr)
-
-static inline void mlx5_write64(__be32 val[2], void __iomem *dest,
-				spinlock_t *doorbell_lock)
-{
-	unsigned long flags;
-
-	if (doorbell_lock)
-		spin_lock_irqsave(doorbell_lock, flags);
 	__raw_writel((__force u32) val[0], dest);
 	__raw_writel((__force u32) val[1], dest + 4);
-	if (doorbell_lock)
-		spin_unlock_irqrestore(doorbell_lock, flags);
-}
-
 #endif
+}
 
 #endif /* MLX5_DOORBELL_H */

include/linux/mlx5/driver.h

@@ -133,6 +133,7 @@ enum {
 	MLX5_REG_MTRC_CONF	 = 0x9041,
 	MLX5_REG_MTRC_STDB	 = 0x9042,
 	MLX5_REG_MTRC_CTRL	 = 0x9043,
+	MLX5_REG_MPEIN		 = 0x9050,
 	MLX5_REG_MPCNT		 = 0x9051,
 	MLX5_REG_MTPPS		 = 0x9053,
 	MLX5_REG_MTPPSE		 = 0x9054,
@@ -660,6 +661,7 @@ struct mlx5_core_dev {
 	u64			sys_image_guid;
 	phys_addr_t		iseg_base;
 	struct mlx5_init_seg __iomem *iseg;
+	phys_addr_t             bar_addr;
 	enum mlx5_device_state	state;
 	/* sync interface state */
 	struct mutex		intf_state_mutex;
@@ -885,6 +887,7 @@ void mlx5_cmd_mbox_status(void *out, u8 *status, u32 *syndrome);
 int mlx5_core_get_caps(struct mlx5_core_dev *dev, enum mlx5_cap_type cap_type);
 int mlx5_cmd_alloc_uar(struct mlx5_core_dev *dev, u32 *uarn);
 int mlx5_cmd_free_uar(struct mlx5_core_dev *dev, u32 uarn);
+void mlx5_health_flush(struct mlx5_core_dev *dev);
 void mlx5_health_cleanup(struct mlx5_core_dev *dev);
 int mlx5_health_init(struct mlx5_core_dev *dev);
 void mlx5_start_health_poll(struct mlx5_core_dev *dev);

include/linux/mlx5/fs.h

@@ -75,6 +75,12 @@ enum mlx5_flow_namespace_type {
 	MLX5_FLOW_NAMESPACE_EGRESS,
 };
 
+enum {
+	FDB_BYPASS_PATH,
+	FDB_FAST_PATH,
+	FDB_SLOW_PATH,
+};
+
 struct mlx5_flow_table;
 struct mlx5_flow_group;
 struct mlx5_flow_namespace;

include/linux/mlx5/mlx5_ifc.h

@@ -8026,6 +8026,52 @@ struct mlx5_ifc_ppcnt_reg_bits {
 	union mlx5_ifc_eth_cntrs_grp_data_layout_auto_bits counter_set;
 };
 
+struct mlx5_ifc_mpein_reg_bits {
+	u8         reserved_at_0[0x2];
+	u8         depth[0x6];
+	u8         pcie_index[0x8];
+	u8         node[0x8];
+	u8         reserved_at_18[0x8];
+
+	u8         capability_mask[0x20];
+
+	u8         reserved_at_40[0x8];
+	u8         link_width_enabled[0x8];
+	u8         link_speed_enabled[0x10];
+
+	u8         lane0_physical_position[0x8];
+	u8         link_width_active[0x8];
+	u8         link_speed_active[0x10];
+
+	u8         num_of_pfs[0x10];
+	u8         num_of_vfs[0x10];
+
+	u8         bdf0[0x10];
+	u8         reserved_at_b0[0x10];
+
+	u8         max_read_request_size[0x4];
+	u8         max_payload_size[0x4];
+	u8         reserved_at_c8[0x5];
+	u8         pwr_status[0x3];
+	u8         port_type[0x4];
+	u8         reserved_at_d4[0xb];
+	u8         lane_reversal[0x1];
+
+	u8         reserved_at_e0[0x14];
+	u8         pci_power[0xc];
+
+	u8         reserved_at_100[0x20];
+
+	u8         device_status[0x10];
+	u8         port_state[0x8];
+	u8         reserved_at_138[0x8];
+
+	u8         reserved_at_140[0x10];
+	u8         receiver_detect_result[0x10];
+
+	u8         reserved_at_160[0x20];
+};
+
 struct mlx5_ifc_mpcnt_reg_bits {
 	u8         reserved_at_0[0x8];
 	u8         pcie_index[0x8];
@@ -8345,7 +8391,9 @@ struct mlx5_ifc_pcam_reg_bits {
 };
 
 struct mlx5_ifc_mcam_enhanced_features_bits {
-	u8         reserved_at_0[0x74];
+	u8         reserved_at_0[0x6e];
+	u8         pci_status_and_power[0x1];
+	u8         reserved_at_6f[0x5];
 	u8         mark_tx_action_cnp[0x1];
 	u8         mark_tx_action_cqe[0x1];
 	u8         dynamic_tx_overflow[0x1];
@@ -8953,6 +9001,7 @@ union mlx5_ifc_ports_control_registers_document_bits {
 	struct mlx5_ifc_pmtu_reg_bits pmtu_reg;
 	struct mlx5_ifc_ppad_reg_bits ppad_reg;
 	struct mlx5_ifc_ppcnt_reg_bits ppcnt_reg;
+	struct mlx5_ifc_mpein_reg_bits mpein_reg;
 	struct mlx5_ifc_mpcnt_reg_bits mpcnt_reg;
 	struct mlx5_ifc_pplm_reg_bits pplm_reg;
 	struct mlx5_ifc_pplr_reg_bits pplr_reg;