From 78ccb25861d76a8fc5c678d762180e6918834200 Mon Sep 17 00:00:00 2001
From: Eli Cohen <eli@mellanox.com>
Date: Fri, 25 Sep 2015 10:49:15 +0300
Subject: net/mlx5_core: Fix wrong name in struct

The name refers to syndrome so uset ext_synd instread of ext_sync.

Signed-off-by: Eli Cohen <eli@mellanox.com>
Signed-off-by: Or Gerlitz <ogerlitz@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
---
 include/linux/mlx5/device.h | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

(limited to 'include/linux/mlx5')

diff --git a/include/linux/mlx5/device.h b/include/linux/mlx5/device.h
index 8eb3b19af2a4..41e9f3bd663c 100644
--- a/include/linux/mlx5/device.h
+++ b/include/linux/mlx5/device.h
@@ -440,7 +440,7 @@ struct health_buffer {
 	__be32		rsvd2;
 	u8		irisc_index;
 	u8		synd;
-	__be16		ext_sync;
+	__be16		ext_synd;
 };
 
 struct mlx5_init_seg {
-- 
cgit v1.2.3


From 020446e01eebc9dbe7eda038e570ab9c7ab13586 Mon Sep 17 00:00:00 2001
From: Eli Cohen <eli@mellanox.com>
Date: Thu, 8 Oct 2015 17:13:58 +0300
Subject: net/mlx5_core: Prepare cmd interface to system errors handling

In preparation to handling system errors at the mlx5_core level, change the
interface of cmd_work_handler to accept a 64 bit argument for the vector.

This allows to encode a flag that signifies when the handler is called
as a result of a driver logic that wishes to terminate commands that
the hardware may not be able to terminate. Such command completions
are detected at the handler and proper return status is encoded.

To be able to terminate page handler commands, we make sure to set
the corresponding bit in the bitmask.

Signed-off-by: Eli Cohen <eli@mellanox.com>
Signed-off-by: Or Gerlitz <ogerlitz@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
---
 drivers/net/ethernet/mellanox/mlx5/core/cmd.c | 20 ++++++++++++++++++--
 include/linux/mlx5/driver.h                   |  6 +++++-
 2 files changed, 23 insertions(+), 3 deletions(-)

(limited to 'include/linux/mlx5')

diff --git a/drivers/net/ethernet/mellanox/mlx5/core/cmd.c b/drivers/net/ethernet/mellanox/mlx5/core/cmd.c
index 84838c2f528f..c3e54b7e8780 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/cmd.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/cmd.c
@@ -254,6 +254,10 @@ static void dump_buf(void *buf, int size, int data_only, int offset)
 		pr_debug("\n");
 }
 
+enum {
+	MLX5_DRIVER_STATUS_ABORTED = 0xfe,
+};
+
 const char *mlx5_command_str(int command)
 {
 	switch (command) {
@@ -473,6 +477,7 @@ static void cmd_work_handler(struct work_struct *work)
 	struct mlx5_core_dev *dev = container_of(cmd, struct mlx5_core_dev, cmd);
 	struct mlx5_cmd_layout *lay;
 	struct semaphore *sem;
+	unsigned long flags;
 
 	sem = ent->page_queue ? &cmd->pages_sem : &cmd->sem;
 	down(sem);
@@ -485,6 +490,9 @@ static void cmd_work_handler(struct work_struct *work)
 		}
 	} else {
 		ent->idx = cmd->max_reg_cmds;
+		spin_lock_irqsave(&cmd->alloc_lock, flags);
+		clear_bit(ent->idx, &cmd->bitmask);
+		spin_unlock_irqrestore(&cmd->alloc_lock, flags);
 	}
 
 	ent->token = alloc_token(cmd);
@@ -1081,7 +1089,7 @@ static void free_msg(struct mlx5_core_dev *dev, struct mlx5_cmd_msg *msg)
 	}
 }
 
-void mlx5_cmd_comp_handler(struct mlx5_core_dev *dev, unsigned long vector)
+void mlx5_cmd_comp_handler(struct mlx5_core_dev *dev, u64 vec)
 {
 	struct mlx5_cmd *cmd = &dev->cmd;
 	struct mlx5_cmd_work_ent *ent;
@@ -1092,7 +1100,10 @@ void mlx5_cmd_comp_handler(struct mlx5_core_dev *dev, unsigned long vector)
 	s64 ds;
 	struct mlx5_cmd_stats *stats;
 	unsigned long flags;
+	unsigned long vector;
 
+	/* there can be at most 32 command queues */
+	vector = vec & 0xffffffff;
 	for (i = 0; i < (1 << cmd->log_sz); i++) {
 		if (test_bit(i, &vector)) {
 			struct semaphore *sem;
@@ -1110,11 +1121,16 @@ void mlx5_cmd_comp_handler(struct mlx5_core_dev *dev, unsigned long vector)
 					ent->ret = verify_signature(ent);
 				else
 					ent->ret = 0;
-				ent->status = ent->lay->status_own >> 1;
+				if (vec & MLX5_TRIGGERED_CMD_COMP)
+					ent->status = MLX5_DRIVER_STATUS_ABORTED;
+				else
+					ent->status = ent->lay->status_own >> 1;
+
 				mlx5_core_dbg(dev, "command completed. ret 0x%x, delivery status %s(0x%x)\n",
 					      ent->ret, deliv_status_to_str(ent->status), ent->status);
 			}
 			free_ent(cmd, ent->idx);
+
 			if (ent->callback) {
 				ds = ent->ts2 - ent->ts1;
 				if (ent->op < ARRAY_SIZE(cmd->stats)) {
diff --git a/include/linux/mlx5/driver.h b/include/linux/mlx5/driver.h
index 8b6d6f2154a4..aa899559eec0 100644
--- a/include/linux/mlx5/driver.h
+++ b/include/linux/mlx5/driver.h
@@ -731,7 +731,7 @@ void mlx5_eq_pagefault(struct mlx5_core_dev *dev, struct mlx5_eqe *eqe);
 #endif
 void mlx5_srq_event(struct mlx5_core_dev *dev, u32 srqn, int event_type);
 struct mlx5_core_srq *mlx5_core_get_srq(struct mlx5_core_dev *dev, u32 srqn);
-void mlx5_cmd_comp_handler(struct mlx5_core_dev *dev, unsigned long vector);
+void mlx5_cmd_comp_handler(struct mlx5_core_dev *dev, u64 vec);
 void mlx5_cq_event(struct mlx5_core_dev *dev, u32 cqn, int event_type);
 int mlx5_create_map_eq(struct mlx5_core_dev *dev, struct mlx5_eq *eq, u8 vecidx,
 		       int nent, u64 mask, const char *name, struct mlx5_uar *uar);
@@ -865,4 +865,8 @@ static inline int mlx5_get_gid_table_len(u16 param)
 	return 8 * (1 << param);
 }
 
+enum {
+	MLX5_TRIGGERED_CMD_COMP = (u64)1 << 32,
+};
+
 #endif /* MLX5_DRIVER_H */
-- 
cgit v1.2.3


From ac6ea6e81a80172612e0c9ef93720f371b198918 Mon Sep 17 00:00:00 2001
From: Eli Cohen <eli@mellanox.com>
Date: Thu, 8 Oct 2015 17:14:00 +0300
Subject: net/mlx5_core: Use private health thread for each device

Use a single threaded work queue for each device in the system instead of
using one thread for any device. This is required so we can concurrently
process system error handling for all the devices that need that.

Signed-off-by: Eli Cohen <eli@mellanox.com>
Signed-off-by: Or Gerlitz <ogerlitz@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
---
 drivers/net/ethernet/mellanox/mlx5/core/health.c | 63 +++++++++++-------------
 drivers/net/ethernet/mellanox/mlx5/core/main.c   | 37 +++++++-------
 include/linux/mlx5/driver.h                      |  7 +--
 3 files changed, 52 insertions(+), 55 deletions(-)

(limited to 'include/linux/mlx5')

diff --git a/drivers/net/ethernet/mellanox/mlx5/core/health.c b/drivers/net/ethernet/mellanox/mlx5/core/health.c
index 8770968fff35..9b81e1ceb8de 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/health.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/health.c
@@ -57,31 +57,16 @@ enum {
 	MLX5_HEALTH_SYNDR_HIGH_TEMP		= 0x10
 };
 
-static DEFINE_SPINLOCK(health_lock);
-static LIST_HEAD(health_list);
-static struct work_struct health_work;
-
 static void health_care(struct work_struct *work)
 {
-	struct mlx5_core_health *health, *n;
+	struct mlx5_core_health *health;
 	struct mlx5_core_dev *dev;
 	struct mlx5_priv *priv;
-	LIST_HEAD(tlist);
-
-	spin_lock_irq(&health_lock);
-	list_splice_init(&health_list, &tlist);
-
-	spin_unlock_irq(&health_lock);
 
-	list_for_each_entry_safe(health, n, &tlist, list) {
-		priv = container_of(health, struct mlx5_priv, health);
-		dev = container_of(priv, struct mlx5_core_dev, priv);
-		mlx5_core_warn(dev, "handling bad device here\n");
-		/* nothing yet */
-		spin_lock_irq(&health_lock);
-		list_del_init(&health->list);
-		spin_unlock_irq(&health_lock);
-	}
+	health = container_of(work, struct mlx5_core_health, work);
+	priv = container_of(health, struct mlx5_priv, health);
+	dev = container_of(priv, struct mlx5_core_dev, priv);
+	mlx5_core_warn(dev, "handling bad device here\n");
 }
 
 static const char *hsynd_str(u8 synd)
@@ -168,11 +153,7 @@ static void poll_health(unsigned long data)
 	if (health->miss_counter == MAX_MISSES) {
 		mlx5_core_err(dev, "device's health compromised\n");
 		print_health_info(dev);
-		spin_lock_irq(&health_lock);
-		list_add_tail(&health->list, &health_list);
-		spin_unlock_irq(&health_lock);
-
-		queue_work(mlx5_core_wq, &health_work);
+		queue_work(health->wq, &health->work);
 	} else {
 		get_random_bytes(&next, sizeof(next));
 		next %= HZ;
@@ -185,7 +166,6 @@ void mlx5_start_health_poll(struct mlx5_core_dev *dev)
 {
 	struct mlx5_core_health *health = &dev->priv.health;
 
-	INIT_LIST_HEAD(&health->list);
 	init_timer(&health->timer);
 	health->health = &dev->iseg->health;
 	health->health_counter = &dev->iseg->health_counter;
@@ -201,18 +181,33 @@ void mlx5_stop_health_poll(struct mlx5_core_dev *dev)
 	struct mlx5_core_health *health = &dev->priv.health;
 
 	del_timer_sync(&health->timer);
-
-	spin_lock_irq(&health_lock);
-	if (!list_empty(&health->list))
-		list_del_init(&health->list);
-	spin_unlock_irq(&health_lock);
 }
 
-void mlx5_health_cleanup(void)
+void mlx5_health_cleanup(struct mlx5_core_dev *dev)
 {
+	struct mlx5_core_health *health = &dev->priv.health;
+
+	destroy_workqueue(health->wq);
 }
 
-void  __init mlx5_health_init(void)
+int mlx5_health_init(struct mlx5_core_dev *dev)
 {
-	INIT_WORK(&health_work, health_care);
+	struct mlx5_core_health *health;
+	char *name;
+
+	health = &dev->priv.health;
+	name = kmalloc(64, GFP_KERNEL);
+	if (!name)
+		return -ENOMEM;
+
+	strcpy(name, "mlx5_health");
+	strcat(name, dev_name(&dev->pdev->dev));
+	health->wq = create_singlethread_workqueue(name);
+	kfree(name);
+	if (!health->wq)
+		return -ENOMEM;
+
+	INIT_WORK(&health->work, health_care);
+
+	return 0;
 }
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/main.c b/drivers/net/ethernet/mellanox/mlx5/core/main.c
index 7718f6ac6214..b6edc58766ad 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/main.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/main.c
@@ -62,7 +62,6 @@ static int prof_sel = MLX5_DEFAULT_PROF;
 module_param_named(prof_sel, prof_sel, int, 0444);
 MODULE_PARM_DESC(prof_sel, "profile selector. Valid range 0 - 2");
 
-struct workqueue_struct *mlx5_core_wq;
 static LIST_HEAD(intf_list);
 static LIST_HEAD(dev_list);
 static DEFINE_MUTEX(intf_mutex);
@@ -1046,6 +1045,7 @@ err_pagealloc_cleanup:
 
 static int mlx5_unload_one(struct mlx5_core_dev *dev, struct mlx5_priv *priv)
 {
+	int err;
 
 	mlx5_unregister_device(dev);
 	mlx5_cleanup_mr_table(dev);
@@ -1060,9 +1060,10 @@ static int mlx5_unload_one(struct mlx5_core_dev *dev, struct mlx5_priv *priv)
 	mlx5_eq_cleanup(dev);
 	mlx5_disable_msix(dev);
 	mlx5_stop_health_poll(dev);
-	if (mlx5_cmd_teardown_hca(dev)) {
+	err = mlx5_cmd_teardown_hca(dev);
+	if (err) {
 		dev_err(&dev->pdev->dev, "tear_down_hca failed, skip cleanup\n");
-		return 1;
+		goto out;
 	}
 	mlx5_pagealloc_stop(dev);
 	mlx5_reclaim_startup_pages(dev);
@@ -1070,11 +1071,12 @@ static int mlx5_unload_one(struct mlx5_core_dev *dev, struct mlx5_priv *priv)
 	mlx5_pagealloc_cleanup(dev);
 	mlx5_cmd_cleanup(dev);
 
-	return 0;
+out:
+	return err;
 }
 
 static void mlx5_core_event(struct mlx5_core_dev *dev, enum mlx5_dev_event event,
-			    unsigned long param)
+		     unsigned long param)
 {
 	struct mlx5_priv *priv = &dev->priv;
 	struct mlx5_device_context *dev_ctx;
@@ -1129,14 +1131,22 @@ static int init_one(struct pci_dev *pdev,
 		goto clean_dev;
 	}
 
+	err = mlx5_health_init(dev);
+	if (err) {
+		dev_err(&pdev->dev, "mlx5_health_init failed with error code %d\n", err);
+		goto close_pci;
+	}
+
 	err = mlx5_load_one(dev, priv);
 	if (err) {
 		dev_err(&pdev->dev, "mlx5_load_one failed with error code %d\n", err);
-		goto close_pci;
+		goto clean_health;
 	}
 
 	return 0;
 
+clean_health:
+	mlx5_health_cleanup(dev);
 close_pci:
 	mlx5_pci_close(dev, priv);
 clean_dev:
@@ -1153,8 +1163,10 @@ static void remove_one(struct pci_dev *pdev)
 
 	if (mlx5_unload_one(dev, priv)) {
 		dev_err(&dev->pdev->dev, "mlx5_unload_one failed\n");
+		mlx5_health_cleanup(dev);
 		return;
 	}
+	mlx5_health_cleanup(dev);
 	mlx5_pci_close(dev, priv);
 	pci_set_drvdata(pdev, NULL);
 	kfree(dev);
@@ -1184,16 +1196,10 @@ static int __init init(void)
 	int err;
 
 	mlx5_register_debugfs();
-	mlx5_core_wq = create_singlethread_workqueue("mlx5_core_wq");
-	if (!mlx5_core_wq) {
-		err = -ENOMEM;
-		goto err_debug;
-	}
-	mlx5_health_init();
 
 	err = pci_register_driver(&mlx5_core_driver);
 	if (err)
-		goto err_health;
+		goto err_debug;
 
 #ifdef CONFIG_MLX5_CORE_EN
 	mlx5e_init();
@@ -1201,9 +1207,6 @@ static int __init init(void)
 
 	return 0;
 
-err_health:
-	mlx5_health_cleanup();
-	destroy_workqueue(mlx5_core_wq);
 err_debug:
 	mlx5_unregister_debugfs();
 	return err;
@@ -1215,8 +1218,6 @@ static void __exit cleanup(void)
 	mlx5e_cleanup();
 #endif
 	pci_unregister_driver(&mlx5_core_driver);
-	mlx5_health_cleanup();
-	destroy_workqueue(mlx5_core_wq);
 	mlx5_unregister_debugfs();
 }
 
diff --git a/include/linux/mlx5/driver.h b/include/linux/mlx5/driver.h
index aa899559eec0..41a32873f608 100644
--- a/include/linux/mlx5/driver.h
+++ b/include/linux/mlx5/driver.h
@@ -391,9 +391,10 @@ struct mlx5_core_health {
 	struct health_buffer __iomem   *health;
 	__be32 __iomem		       *health_counter;
 	struct timer_list		timer;
-	struct list_head		list;
 	u32				prev;
 	int				miss_counter;
+	struct workqueue_struct	       *wq;
+	struct work_struct		work;
 };
 
 struct mlx5_cq_table {
@@ -676,8 +677,8 @@ int mlx5_alloc_uuars(struct mlx5_core_dev *dev, struct mlx5_uuar_info *uuari);
 int mlx5_free_uuars(struct mlx5_core_dev *dev, struct mlx5_uuar_info *uuari);
 int mlx5_alloc_map_uar(struct mlx5_core_dev *mdev, struct mlx5_uar *uar);
 void mlx5_unmap_free_uar(struct mlx5_core_dev *mdev, struct mlx5_uar *uar);
-void mlx5_health_cleanup(void);
-void  __init mlx5_health_init(void);
+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);
 void mlx5_stop_health_poll(struct mlx5_core_dev *dev);
 int mlx5_buf_alloc_node(struct mlx5_core_dev *dev, int size,
-- 
cgit v1.2.3


From fd76ee4da55abb21babfc69310d321b9cb9a32e0 Mon Sep 17 00:00:00 2001
From: Eli Cohen <eli@mellanox.com>
Date: Wed, 14 Oct 2015 17:43:45 +0300
Subject: net/mlx5_core: Fix internal error detection conditions

The detection of a fatal condition has been updated to take into account
the state reported by the device or by detecting an all ones read of the
firmware version which indicates that the device is not accessible.

Signed-off-by: Eli Cohen <eli@mellanox.com>
Signed-off-by: Or Gerlitz <ogerlitz@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
---
 drivers/net/ethernet/mellanox/mlx5/core/health.c | 51 ++++++++++++++++++++----
 include/linux/mlx5/driver.h                      |  1 +
 2 files changed, 45 insertions(+), 7 deletions(-)

(limited to 'include/linux/mlx5')

diff --git a/drivers/net/ethernet/mellanox/mlx5/core/health.c b/drivers/net/ethernet/mellanox/mlx5/core/health.c
index 9b81e1ceb8de..f1eb686c45b1 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/health.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/health.c
@@ -57,6 +57,31 @@ enum {
 	MLX5_HEALTH_SYNDR_HIGH_TEMP		= 0x10
 };
 
+enum {
+	MLX5_NIC_IFC_FULL		= 0,
+	MLX5_NIC_IFC_DISABLED		= 1,
+	MLX5_NIC_IFC_NO_DRAM_NIC	= 2
+};
+
+static u8 get_nic_interface(struct mlx5_core_dev *dev)
+{
+	return (ioread32be(&dev->iseg->cmdq_addr_l_sz) >> 8) & 3;
+}
+
+static int in_fatal(struct mlx5_core_dev *dev)
+{
+	struct mlx5_core_health *health = &dev->priv.health;
+	struct health_buffer __iomem *h = health->health;
+
+	if (get_nic_interface(dev) == MLX5_NIC_IFC_DISABLED)
+		return 1;
+
+	if (ioread32be(&h->fw_ver) == 0xffffffff)
+		return 1;
+
+	return 0;
+}
+
 static void health_care(struct work_struct *work)
 {
 	struct mlx5_core_health *health;
@@ -136,11 +161,21 @@ static void print_health_info(struct mlx5_core_dev *dev)
 	dev_err(&dev->pdev->dev, "ext_synd 0x%04x\n", ioread16be(&h->ext_synd));
 }
 
+static unsigned long get_next_poll_jiffies(void)
+{
+	unsigned long next;
+
+	get_random_bytes(&next, sizeof(next));
+	next %= HZ;
+	next += jiffies + MLX5_HEALTH_POLL_INTERVAL;
+
+	return next;
+}
+
 static void poll_health(unsigned long data)
 {
 	struct mlx5_core_dev *dev = (struct mlx5_core_dev *)data;
 	struct mlx5_core_health *health = &dev->priv.health;
-	unsigned long next;
 	u32 count;
 
 	count = ioread32be(health->health_counter);
@@ -151,14 +186,16 @@ static void poll_health(unsigned long data)
 
 	health->prev = count;
 	if (health->miss_counter == MAX_MISSES) {
-		mlx5_core_err(dev, "device's health compromised\n");
+		dev_err(&dev->pdev->dev, "device's health compromised - reached miss count\n");
 		print_health_info(dev);
-		queue_work(health->wq, &health->work);
 	} else {
-		get_random_bytes(&next, sizeof(next));
-		next %= HZ;
-		next += jiffies + MLX5_HEALTH_POLL_INTERVAL;
-		mod_timer(&health->timer, next);
+		mod_timer(&health->timer, get_next_poll_jiffies());
+	}
+
+	if (in_fatal(dev) && !health->sick) {
+		health->sick = true;
+		print_health_info(dev);
+		queue_work(health->wq, &health->work);
 	}
 }
 
diff --git a/include/linux/mlx5/driver.h b/include/linux/mlx5/driver.h
index 41a32873f608..62b7d439813d 100644
--- a/include/linux/mlx5/driver.h
+++ b/include/linux/mlx5/driver.h
@@ -393,6 +393,7 @@ struct mlx5_core_health {
 	struct timer_list		timer;
 	u32				prev;
 	int				miss_counter;
+	bool				sick;
 	struct workqueue_struct	       *wq;
 	struct work_struct		work;
 };
-- 
cgit v1.2.3


From 89d44f0a6c732db23b219be708e2fe1e03ee4842 Mon Sep 17 00:00:00 2001
From: Majd Dibbiny <majd@mellanox.com>
Date: Wed, 14 Oct 2015 17:43:46 +0300
Subject: net/mlx5_core: Add pci error handlers to mlx5_core driver

This patch implement the pci_error_handlers for mlx5_core which allow the
driver to recover from PCI error.

Once an error is detected in the PCI, the mlx5_pci_err_detected is called
and it:
1) Marks the device to be in 'Internal Error' state.
2) Dispatches an event to the mlx5_ib to flush all the outstanding cqes
with error.
3) Returns all the on going commands with error.
4) Unloads the driver.

Afterwards, the FW is reset and mlx5_pci_slot_reset is called and it
enables the device and restore it's pci state.

If the later succeeds, mlx5_pci_resume is called, and it loads the SW
stack.

Signed-off-by: Majd Dibbiny <majd@mellanox.com>
Signed-off-by: Eli Cohen <eli@mellanox.com>
Signed-off-by: Or Gerlitz <ogerlitz@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
---
 drivers/net/ethernet/mellanox/mlx5/core/cmd.c      | 170 +++++++++++++++++++
 drivers/net/ethernet/mellanox/mlx5/core/health.c   |  72 ++++++++
 drivers/net/ethernet/mellanox/mlx5/core/main.c     | 182 ++++++++++++++++++++-
 .../net/ethernet/mellanox/mlx5/core/mlx5_core.h    |   4 +
 .../net/ethernet/mellanox/mlx5/core/pagealloc.c    |  13 +-
 include/linux/mlx5/driver.h                        |  22 +++
 6 files changed, 451 insertions(+), 12 deletions(-)

(limited to 'include/linux/mlx5')

diff --git a/drivers/net/ethernet/mellanox/mlx5/core/cmd.c b/drivers/net/ethernet/mellanox/mlx5/core/cmd.c
index c3e54b7e8780..fabfc9e0a948 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/cmd.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/cmd.c
@@ -256,8 +256,154 @@ static void dump_buf(void *buf, int size, int data_only, int offset)
 
 enum {
 	MLX5_DRIVER_STATUS_ABORTED = 0xfe,
+	MLX5_DRIVER_SYND = 0xbadd00de,
 };
 
+static int mlx5_internal_err_ret_value(struct mlx5_core_dev *dev, u16 op,
+				       u32 *synd, u8 *status)
+{
+	*synd = 0;
+	*status = 0;
+
+	switch (op) {
+	case MLX5_CMD_OP_TEARDOWN_HCA:
+	case MLX5_CMD_OP_DISABLE_HCA:
+	case MLX5_CMD_OP_MANAGE_PAGES:
+	case MLX5_CMD_OP_DESTROY_MKEY:
+	case MLX5_CMD_OP_DESTROY_EQ:
+	case MLX5_CMD_OP_DESTROY_CQ:
+	case MLX5_CMD_OP_DESTROY_QP:
+	case MLX5_CMD_OP_DESTROY_PSV:
+	case MLX5_CMD_OP_DESTROY_SRQ:
+	case MLX5_CMD_OP_DESTROY_XRC_SRQ:
+	case MLX5_CMD_OP_DESTROY_DCT:
+	case MLX5_CMD_OP_DEALLOC_Q_COUNTER:
+	case MLX5_CMD_OP_DEALLOC_PD:
+	case MLX5_CMD_OP_DEALLOC_UAR:
+	case MLX5_CMD_OP_DETTACH_FROM_MCG:
+	case MLX5_CMD_OP_DEALLOC_XRCD:
+	case MLX5_CMD_OP_DEALLOC_TRANSPORT_DOMAIN:
+	case MLX5_CMD_OP_DELETE_VXLAN_UDP_DPORT:
+	case MLX5_CMD_OP_DELETE_L2_TABLE_ENTRY:
+	case MLX5_CMD_OP_DESTROY_TIR:
+	case MLX5_CMD_OP_DESTROY_SQ:
+	case MLX5_CMD_OP_DESTROY_RQ:
+	case MLX5_CMD_OP_DESTROY_RMP:
+	case MLX5_CMD_OP_DESTROY_TIS:
+	case MLX5_CMD_OP_DESTROY_RQT:
+	case MLX5_CMD_OP_DESTROY_FLOW_TABLE:
+	case MLX5_CMD_OP_DESTROY_FLOW_GROUP:
+	case MLX5_CMD_OP_DELETE_FLOW_TABLE_ENTRY:
+		return MLX5_CMD_STAT_OK;
+
+	case MLX5_CMD_OP_QUERY_HCA_CAP:
+	case MLX5_CMD_OP_QUERY_ADAPTER:
+	case MLX5_CMD_OP_INIT_HCA:
+	case MLX5_CMD_OP_ENABLE_HCA:
+	case MLX5_CMD_OP_QUERY_PAGES:
+	case MLX5_CMD_OP_SET_HCA_CAP:
+	case MLX5_CMD_OP_QUERY_ISSI:
+	case MLX5_CMD_OP_SET_ISSI:
+	case MLX5_CMD_OP_CREATE_MKEY:
+	case MLX5_CMD_OP_QUERY_MKEY:
+	case MLX5_CMD_OP_QUERY_SPECIAL_CONTEXTS:
+	case MLX5_CMD_OP_PAGE_FAULT_RESUME:
+	case MLX5_CMD_OP_CREATE_EQ:
+	case MLX5_CMD_OP_QUERY_EQ:
+	case MLX5_CMD_OP_GEN_EQE:
+	case MLX5_CMD_OP_CREATE_CQ:
+	case MLX5_CMD_OP_QUERY_CQ:
+	case MLX5_CMD_OP_MODIFY_CQ:
+	case MLX5_CMD_OP_CREATE_QP:
+	case MLX5_CMD_OP_RST2INIT_QP:
+	case MLX5_CMD_OP_INIT2RTR_QP:
+	case MLX5_CMD_OP_RTR2RTS_QP:
+	case MLX5_CMD_OP_RTS2RTS_QP:
+	case MLX5_CMD_OP_SQERR2RTS_QP:
+	case MLX5_CMD_OP_2ERR_QP:
+	case MLX5_CMD_OP_2RST_QP:
+	case MLX5_CMD_OP_QUERY_QP:
+	case MLX5_CMD_OP_SQD_RTS_QP:
+	case MLX5_CMD_OP_INIT2INIT_QP:
+	case MLX5_CMD_OP_CREATE_PSV:
+	case MLX5_CMD_OP_CREATE_SRQ:
+	case MLX5_CMD_OP_QUERY_SRQ:
+	case MLX5_CMD_OP_ARM_RQ:
+	case MLX5_CMD_OP_CREATE_XRC_SRQ:
+	case MLX5_CMD_OP_QUERY_XRC_SRQ:
+	case MLX5_CMD_OP_ARM_XRC_SRQ:
+	case MLX5_CMD_OP_CREATE_DCT:
+	case MLX5_CMD_OP_DRAIN_DCT:
+	case MLX5_CMD_OP_QUERY_DCT:
+	case MLX5_CMD_OP_ARM_DCT_FOR_KEY_VIOLATION:
+	case MLX5_CMD_OP_QUERY_VPORT_STATE:
+	case MLX5_CMD_OP_MODIFY_VPORT_STATE:
+	case MLX5_CMD_OP_QUERY_ESW_VPORT_CONTEXT:
+	case MLX5_CMD_OP_MODIFY_ESW_VPORT_CONTEXT:
+	case MLX5_CMD_OP_QUERY_NIC_VPORT_CONTEXT:
+	case MLX5_CMD_OP_MODIFY_NIC_VPORT_CONTEXT:
+	case MLX5_CMD_OP_QUERY_ROCE_ADDRESS:
+	case MLX5_CMD_OP_SET_ROCE_ADDRESS:
+	case MLX5_CMD_OP_QUERY_HCA_VPORT_CONTEXT:
+	case MLX5_CMD_OP_MODIFY_HCA_VPORT_CONTEXT:
+	case MLX5_CMD_OP_QUERY_HCA_VPORT_GID:
+	case MLX5_CMD_OP_QUERY_HCA_VPORT_PKEY:
+	case MLX5_CMD_OP_QUERY_VPORT_COUNTER:
+	case MLX5_CMD_OP_ALLOC_Q_COUNTER:
+	case MLX5_CMD_OP_QUERY_Q_COUNTER:
+	case MLX5_CMD_OP_ALLOC_PD:
+	case MLX5_CMD_OP_ALLOC_UAR:
+	case MLX5_CMD_OP_CONFIG_INT_MODERATION:
+	case MLX5_CMD_OP_ACCESS_REG:
+	case MLX5_CMD_OP_ATTACH_TO_MCG:
+	case MLX5_CMD_OP_GET_DROPPED_PACKET_LOG:
+	case MLX5_CMD_OP_MAD_IFC:
+	case MLX5_CMD_OP_QUERY_MAD_DEMUX:
+	case MLX5_CMD_OP_SET_MAD_DEMUX:
+	case MLX5_CMD_OP_NOP:
+	case MLX5_CMD_OP_ALLOC_XRCD:
+	case MLX5_CMD_OP_ALLOC_TRANSPORT_DOMAIN:
+	case MLX5_CMD_OP_QUERY_CONG_STATUS:
+	case MLX5_CMD_OP_MODIFY_CONG_STATUS:
+	case MLX5_CMD_OP_QUERY_CONG_PARAMS:
+	case MLX5_CMD_OP_MODIFY_CONG_PARAMS:
+	case MLX5_CMD_OP_QUERY_CONG_STATISTICS:
+	case MLX5_CMD_OP_ADD_VXLAN_UDP_DPORT:
+	case MLX5_CMD_OP_SET_L2_TABLE_ENTRY:
+	case MLX5_CMD_OP_QUERY_L2_TABLE_ENTRY:
+	case MLX5_CMD_OP_CREATE_TIR:
+	case MLX5_CMD_OP_MODIFY_TIR:
+	case MLX5_CMD_OP_QUERY_TIR:
+	case MLX5_CMD_OP_CREATE_SQ:
+	case MLX5_CMD_OP_MODIFY_SQ:
+	case MLX5_CMD_OP_QUERY_SQ:
+	case MLX5_CMD_OP_CREATE_RQ:
+	case MLX5_CMD_OP_MODIFY_RQ:
+	case MLX5_CMD_OP_QUERY_RQ:
+	case MLX5_CMD_OP_CREATE_RMP:
+	case MLX5_CMD_OP_MODIFY_RMP:
+	case MLX5_CMD_OP_QUERY_RMP:
+	case MLX5_CMD_OP_CREATE_TIS:
+	case MLX5_CMD_OP_MODIFY_TIS:
+	case MLX5_CMD_OP_QUERY_TIS:
+	case MLX5_CMD_OP_CREATE_RQT:
+	case MLX5_CMD_OP_MODIFY_RQT:
+	case MLX5_CMD_OP_QUERY_RQT:
+	case MLX5_CMD_OP_CREATE_FLOW_TABLE:
+	case MLX5_CMD_OP_QUERY_FLOW_TABLE:
+	case MLX5_CMD_OP_CREATE_FLOW_GROUP:
+	case MLX5_CMD_OP_QUERY_FLOW_GROUP:
+	case MLX5_CMD_OP_SET_FLOW_TABLE_ENTRY:
+	case MLX5_CMD_OP_QUERY_FLOW_TABLE_ENTRY:
+		*status = MLX5_DRIVER_STATUS_ABORTED;
+		*synd = MLX5_DRIVER_SYND;
+		return -EIO;
+	default:
+		mlx5_core_err(dev, "Unknown FW command (%d)\n", op);
+		return -EINVAL;
+	}
+}
+
 const char *mlx5_command_str(int command)
 {
 	switch (command) {
@@ -592,6 +738,16 @@ static int wait_func(struct mlx5_core_dev *dev, struct mlx5_cmd_work_ent *ent)
 	return err;
 }
 
+static __be32 *get_synd_ptr(struct mlx5_outbox_hdr *out)
+{
+	return &out->syndrome;
+}
+
+static u8 *get_status_ptr(struct mlx5_outbox_hdr *out)
+{
+	return &out->status;
+}
+
 /*  Notes:
  *    1. Callback functions may not sleep
  *    2. page queue commands do not support asynchrous completion
@@ -1200,6 +1356,11 @@ static struct mlx5_cmd_msg *alloc_msg(struct mlx5_core_dev *dev, int in_size,
 	return msg;
 }
 
+static u16 opcode_from_in(struct mlx5_inbox_hdr *in)
+{
+	return be16_to_cpu(in->opcode);
+}
+
 static int is_manage_pages(struct mlx5_inbox_hdr *in)
 {
 	return be16_to_cpu(in->opcode) == MLX5_CMD_OP_MANAGE_PAGES;
@@ -1214,6 +1375,15 @@ static int cmd_exec(struct mlx5_core_dev *dev, void *in, int in_size, void *out,
 	gfp_t gfp;
 	int err;
 	u8 status = 0;
+	u32 drv_synd;
+
+	if (pci_channel_offline(dev->pdev) ||
+	    dev->state == MLX5_DEVICE_STATE_INTERNAL_ERROR) {
+		err = mlx5_internal_err_ret_value(dev, opcode_from_in(in), &drv_synd, &status);
+		*get_synd_ptr(out) = cpu_to_be32(drv_synd);
+		*get_status_ptr(out) = status;
+		return err;
+	}
 
 	pages_queue = is_manage_pages(in);
 	gfp = callback ? GFP_ATOMIC : GFP_KERNEL;
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/health.c b/drivers/net/ethernet/mellanox/mlx5/core/health.c
index f1eb686c45b1..f5deb642d0d6 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/health.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/health.c
@@ -34,6 +34,7 @@
 #include <linux/module.h>
 #include <linux/random.h>
 #include <linux/vmalloc.h>
+#include <linux/hardirq.h>
 #include <linux/mlx5/driver.h>
 #include <linux/mlx5/cmd.h>
 #include "mlx5_core.h"
@@ -68,6 +69,29 @@ static u8 get_nic_interface(struct mlx5_core_dev *dev)
 	return (ioread32be(&dev->iseg->cmdq_addr_l_sz) >> 8) & 3;
 }
 
+static void trigger_cmd_completions(struct mlx5_core_dev *dev)
+{
+	unsigned long flags;
+	u64 vector;
+
+	/* wait for pending handlers to complete */
+	synchronize_irq(dev->priv.msix_arr[MLX5_EQ_VEC_CMD].vector);
+	spin_lock_irqsave(&dev->cmd.alloc_lock, flags);
+	vector = ~dev->cmd.bitmask & ((1ul << (1 << dev->cmd.log_sz)) - 1);
+	if (!vector)
+		goto no_trig;
+
+	vector |= MLX5_TRIGGERED_CMD_COMP;
+	spin_unlock_irqrestore(&dev->cmd.alloc_lock, flags);
+
+	mlx5_core_dbg(dev, "vector 0x%llx\n", vector);
+	mlx5_cmd_comp_handler(dev, vector);
+	return;
+
+no_trig:
+	spin_unlock_irqrestore(&dev->cmd.alloc_lock, flags);
+}
+
 static int in_fatal(struct mlx5_core_dev *dev)
 {
 	struct mlx5_core_health *health = &dev->priv.health;
@@ -82,6 +106,43 @@ static int in_fatal(struct mlx5_core_dev *dev)
 	return 0;
 }
 
+void mlx5_enter_error_state(struct mlx5_core_dev *dev)
+{
+	if (dev->state == MLX5_DEVICE_STATE_INTERNAL_ERROR)
+		return;
+
+	mlx5_core_err(dev, "start\n");
+	if (pci_channel_offline(dev->pdev) || in_fatal(dev))
+		dev->state = MLX5_DEVICE_STATE_INTERNAL_ERROR;
+
+	mlx5_core_event(dev, MLX5_DEV_EVENT_SYS_ERROR, 0);
+	mlx5_core_err(dev, "end\n");
+}
+
+static void mlx5_handle_bad_state(struct mlx5_core_dev *dev)
+{
+	u8 nic_interface = get_nic_interface(dev);
+
+	switch (nic_interface) {
+	case MLX5_NIC_IFC_FULL:
+		mlx5_core_warn(dev, "Expected to see disabled NIC but it is full driver\n");
+		break;
+
+	case MLX5_NIC_IFC_DISABLED:
+		mlx5_core_warn(dev, "starting teardown\n");
+		break;
+
+	case MLX5_NIC_IFC_NO_DRAM_NIC:
+		mlx5_core_warn(dev, "Expected to see disabled NIC but it is no dram nic\n");
+		break;
+	default:
+		mlx5_core_warn(dev, "Expected to see disabled NIC but it is has invalid value %d\n",
+			       nic_interface);
+	}
+
+	mlx5_disable_device(dev);
+}
+
 static void health_care(struct work_struct *work)
 {
 	struct mlx5_core_health *health;
@@ -92,6 +153,7 @@ static void health_care(struct work_struct *work)
 	priv = container_of(health, struct mlx5_priv, health);
 	dev = container_of(priv, struct mlx5_core_dev, priv);
 	mlx5_core_warn(dev, "handling bad device here\n");
+	mlx5_handle_bad_state(dev);
 }
 
 static const char *hsynd_str(u8 synd)
@@ -147,6 +209,10 @@ static void print_health_info(struct mlx5_core_dev *dev)
 	u32 fw;
 	int i;
 
+	/* If the syndrom is 0, the device is OK and no need to print buffer */
+	if (!ioread8(&h->synd))
+		return;
+
 	for (i = 0; i < ARRAY_SIZE(h->assert_var); i++)
 		dev_err(&dev->pdev->dev, "assert_var[%d] 0x%08x\n", i, ioread32be(h->assert_var + i));
 
@@ -178,6 +244,12 @@ static void poll_health(unsigned long data)
 	struct mlx5_core_health *health = &dev->priv.health;
 	u32 count;
 
+	if (dev->state == MLX5_DEVICE_STATE_INTERNAL_ERROR) {
+		trigger_cmd_completions(dev);
+		mod_timer(&health->timer, get_next_poll_jiffies());
+		return;
+	}
+
 	count = ioread32be(health->health_counter);
 	if (count == health->prev)
 		++health->miss_counter;
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/main.c b/drivers/net/ethernet/mellanox/mlx5/core/main.c
index b6edc58766ad..a103a54d6660 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/main.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/main.c
@@ -45,6 +45,7 @@
 #include <linux/mlx5/srq.h>
 #include <linux/debugfs.h>
 #include <linux/kmod.h>
+#include <linux/delay.h>
 #include <linux/mlx5/mlx5_ifc.h>
 #include "mlx5_core.h"
 
@@ -181,6 +182,34 @@ static int set_dma_caps(struct pci_dev *pdev)
 	return err;
 }
 
+static int mlx5_pci_enable_device(struct mlx5_core_dev *dev)
+{
+	struct pci_dev *pdev = dev->pdev;
+	int err = 0;
+
+	mutex_lock(&dev->pci_status_mutex);
+	if (dev->pci_status == MLX5_PCI_STATUS_DISABLED) {
+		err = pci_enable_device(pdev);
+		if (!err)
+			dev->pci_status = MLX5_PCI_STATUS_ENABLED;
+	}
+	mutex_unlock(&dev->pci_status_mutex);
+
+	return err;
+}
+
+static void mlx5_pci_disable_device(struct mlx5_core_dev *dev)
+{
+	struct pci_dev *pdev = dev->pdev;
+
+	mutex_lock(&dev->pci_status_mutex);
+	if (dev->pci_status == MLX5_PCI_STATUS_ENABLED) {
+		pci_disable_device(pdev);
+		dev->pci_status = MLX5_PCI_STATUS_DISABLED;
+	}
+	mutex_unlock(&dev->pci_status_mutex);
+}
+
 static int request_bar(struct pci_dev *pdev)
 {
 	int err = 0;
@@ -807,7 +836,7 @@ static int mlx5_pci_init(struct mlx5_core_dev *dev, struct mlx5_priv *priv)
 	if (!priv->dbg_root)
 		return -ENOMEM;
 
-	err = pci_enable_device(pdev);
+	err = mlx5_pci_enable_device(dev);
 	if (err) {
 		dev_err(&pdev->dev, "Cannot enable PCI device, aborting\n");
 		goto err_dbg;
@@ -841,7 +870,7 @@ err_clr_master:
 	pci_clear_master(dev->pdev);
 	release_bar(dev->pdev);
 err_disable:
-	pci_disable_device(dev->pdev);
+	mlx5_pci_disable_device(dev);
 
 err_dbg:
 	debugfs_remove(priv->dbg_root);
@@ -853,7 +882,7 @@ static void mlx5_pci_close(struct mlx5_core_dev *dev, struct mlx5_priv *priv)
 	iounmap(dev->iseg);
 	pci_clear_master(dev->pdev);
 	release_bar(dev->pdev);
-	pci_disable_device(dev->pdev);
+	mlx5_pci_disable_device(dev);
 	debugfs_remove(priv->dbg_root);
 }
 
@@ -863,13 +892,25 @@ static int mlx5_load_one(struct mlx5_core_dev *dev, struct mlx5_priv *priv)
 	struct pci_dev *pdev = dev->pdev;
 	int err;
 
+	mutex_lock(&dev->intf_state_mutex);
+	if (dev->interface_state == MLX5_INTERFACE_STATE_UP) {
+		dev_warn(&dev->pdev->dev, "%s: interface is up, NOP\n",
+			 __func__);
+		goto out;
+	}
+
 	dev_info(&pdev->dev, "firmware version: %d.%d.%d\n", fw_rev_maj(dev),
 		 fw_rev_min(dev), fw_rev_sub(dev));
 
+	/* on load removing any previous indication of internal error, device is
+	 * up
+	 */
+	dev->state = MLX5_DEVICE_STATE_UP;
+
 	err = mlx5_cmd_init(dev);
 	if (err) {
 		dev_err(&pdev->dev, "Failed initializing command interface, aborting\n");
-		return err;
+		goto out_err;
 	}
 
 	mlx5_pagealloc_init(dev);
@@ -994,6 +1035,10 @@ static int mlx5_load_one(struct mlx5_core_dev *dev, struct mlx5_priv *priv)
 	if (err)
 		pr_info("failed request module on %s\n", MLX5_IB_MOD);
 
+	dev->interface_state = MLX5_INTERFACE_STATE_UP;
+out:
+	mutex_unlock(&dev->intf_state_mutex);
+
 	return 0;
 
 err_reg_dev:
@@ -1024,7 +1069,7 @@ err_stop_poll:
 	mlx5_stop_health_poll(dev);
 	if (mlx5_cmd_teardown_hca(dev)) {
 		dev_err(&dev->pdev->dev, "tear_down_hca failed, skip cleanup\n");
-		return err;
+		goto out_err;
 	}
 
 err_pagealloc_stop:
@@ -1040,13 +1085,23 @@ err_pagealloc_cleanup:
 	mlx5_pagealloc_cleanup(dev);
 	mlx5_cmd_cleanup(dev);
 
+out_err:
+	dev->state = MLX5_DEVICE_STATE_INTERNAL_ERROR;
+	mutex_unlock(&dev->intf_state_mutex);
+
 	return err;
 }
 
 static int mlx5_unload_one(struct mlx5_core_dev *dev, struct mlx5_priv *priv)
 {
-	int err;
+	int err = 0;
 
+	mutex_lock(&dev->intf_state_mutex);
+	if (dev->interface_state == MLX5_INTERFACE_STATE_DOWN) {
+		dev_warn(&dev->pdev->dev, "%s: interface is down, NOP\n",
+			 __func__);
+		goto out;
+	}
 	mlx5_unregister_device(dev);
 	mlx5_cleanup_mr_table(dev);
 	mlx5_cleanup_srq_table(dev);
@@ -1072,10 +1127,12 @@ static int mlx5_unload_one(struct mlx5_core_dev *dev, struct mlx5_priv *priv)
 	mlx5_cmd_cleanup(dev);
 
 out:
+	dev->interface_state = MLX5_INTERFACE_STATE_DOWN;
+	mutex_unlock(&dev->intf_state_mutex);
 	return err;
 }
 
-static void mlx5_core_event(struct mlx5_core_dev *dev, enum mlx5_dev_event event,
+void mlx5_core_event(struct mlx5_core_dev *dev, enum mlx5_dev_event event,
 		     unsigned long param)
 {
 	struct mlx5_priv *priv = &dev->priv;
@@ -1125,6 +1182,8 @@ static int init_one(struct pci_dev *pdev,
 
 	INIT_LIST_HEAD(&priv->ctx_list);
 	spin_lock_init(&priv->ctx_lock);
+	mutex_init(&dev->pci_status_mutex);
+	mutex_init(&dev->intf_state_mutex);
 	err = mlx5_pci_init(dev, priv);
 	if (err) {
 		dev_err(&pdev->dev, "mlx5_pci_init failed with error code %d\n", err);
@@ -1172,6 +1231,112 @@ static void remove_one(struct pci_dev *pdev)
 	kfree(dev);
 }
 
+static pci_ers_result_t mlx5_pci_err_detected(struct pci_dev *pdev,
+					      pci_channel_state_t state)
+{
+	struct mlx5_core_dev *dev = pci_get_drvdata(pdev);
+	struct mlx5_priv *priv = &dev->priv;
+
+	dev_info(&pdev->dev, "%s was called\n", __func__);
+	mlx5_enter_error_state(dev);
+	mlx5_unload_one(dev, priv);
+	mlx5_pci_disable_device(dev);
+	return state == pci_channel_io_perm_failure ?
+		PCI_ERS_RESULT_DISCONNECT : PCI_ERS_RESULT_NEED_RESET;
+}
+
+static pci_ers_result_t mlx5_pci_slot_reset(struct pci_dev *pdev)
+{
+	struct mlx5_core_dev *dev = pci_get_drvdata(pdev);
+	int err = 0;
+
+	dev_info(&pdev->dev, "%s was called\n", __func__);
+
+	err = mlx5_pci_enable_device(dev);
+	if (err) {
+		dev_err(&pdev->dev, "%s: mlx5_pci_enable_device failed with error code: %d\n"
+			, __func__, err);
+		return PCI_ERS_RESULT_DISCONNECT;
+	}
+	pci_set_master(pdev);
+	pci_set_power_state(pdev, PCI_D0);
+	pci_restore_state(pdev);
+
+	return err ? PCI_ERS_RESULT_DISCONNECT : PCI_ERS_RESULT_RECOVERED;
+}
+
+void mlx5_disable_device(struct mlx5_core_dev *dev)
+{
+	mlx5_pci_err_detected(dev->pdev, 0);
+}
+
+/* wait for the device to show vital signs. For now we check
+ * that we can read the device ID and that the health buffer
+ * shows a non zero value which is different than 0xffffffff
+ */
+static void wait_vital(struct pci_dev *pdev)
+{
+	struct mlx5_core_dev *dev = pci_get_drvdata(pdev);
+	struct mlx5_core_health *health = &dev->priv.health;
+	const int niter = 100;
+	u32 count;
+	u16 did;
+	int i;
+
+	/* Wait for firmware to be ready after reset */
+	msleep(1000);
+	for (i = 0; i < niter; i++) {
+		if (pci_read_config_word(pdev, 2, &did)) {
+			dev_warn(&pdev->dev, "failed reading config word\n");
+			break;
+		}
+		if (did == pdev->device) {
+			dev_info(&pdev->dev, "device ID correctly read after %d iterations\n", i);
+			break;
+		}
+		msleep(50);
+	}
+	if (i == niter)
+		dev_warn(&pdev->dev, "%s-%d: could not read device ID\n", __func__, __LINE__);
+
+	for (i = 0; i < niter; i++) {
+		count = ioread32be(health->health_counter);
+		if (count && count != 0xffffffff) {
+			dev_info(&pdev->dev, "Counter value 0x%x after %d iterations\n", count, i);
+			break;
+		}
+		msleep(50);
+	}
+
+	if (i == niter)
+		dev_warn(&pdev->dev, "%s-%d: could not read device ID\n", __func__, __LINE__);
+}
+
+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__);
+
+	pci_save_state(pdev);
+	wait_vital(pdev);
+
+	err = mlx5_load_one(dev, priv);
+	if (err)
+		dev_err(&pdev->dev, "%s: mlx5_load_one failed with error code: %d\n"
+			, __func__, err);
+	else
+		dev_info(&pdev->dev, "%s: device recovered\n", __func__);
+}
+
+static const struct pci_error_handlers mlx5_err_handler = {
+	.error_detected = mlx5_pci_err_detected,
+	.slot_reset	= mlx5_pci_slot_reset,
+	.resume		= mlx5_pci_resume
+};
+
 static const struct pci_device_id mlx5_core_pci_table[] = {
 	{ PCI_VDEVICE(MELLANOX, 0x1011) }, /* Connect-IB */
 	{ PCI_VDEVICE(MELLANOX, 0x1012) }, /* Connect-IB VF */
@@ -1188,7 +1353,8 @@ static struct pci_driver mlx5_core_driver = {
 	.name           = DRIVER_NAME,
 	.id_table       = mlx5_core_pci_table,
 	.probe          = init_one,
-	.remove         = remove_one
+	.remove         = remove_one,
+	.err_handler	= &mlx5_err_handler
 };
 
 static int __init init(void)
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/mlx5_core.h b/drivers/net/ethernet/mellanox/mlx5/core/mlx5_core.h
index 30c0be721b08..cee5b7a839bc 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/mlx5_core.h
+++ b/drivers/net/ethernet/mellanox/mlx5/core/mlx5_core.h
@@ -86,6 +86,10 @@ int mlx5_query_hca_caps(struct mlx5_core_dev *dev);
 int mlx5_query_board_id(struct mlx5_core_dev *dev);
 int mlx5_cmd_init_hca(struct mlx5_core_dev *dev);
 int mlx5_cmd_teardown_hca(struct mlx5_core_dev *dev);
+void mlx5_core_event(struct mlx5_core_dev *dev, enum mlx5_dev_event event,
+		     unsigned long param);
+void mlx5_enter_error_state(struct mlx5_core_dev *dev);
+void mlx5_disable_device(struct mlx5_core_dev *dev);
 
 void mlx5e_init(void);
 void mlx5e_cleanup(void);
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/pagealloc.c b/drivers/net/ethernet/mellanox/mlx5/core/pagealloc.c
index 76432a510ac2..1cda5d268ec9 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/pagealloc.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/pagealloc.c
@@ -493,15 +493,20 @@ int mlx5_reclaim_startup_pages(struct mlx5_core_dev *dev)
 	struct fw_page *fwp;
 	struct rb_node *p;
 	int nclaimed = 0;
-	int err;
+	int err = 0;
 
 	do {
 		p = rb_first(&dev->priv.page_root);
 		if (p) {
 			fwp = rb_entry(p, struct fw_page, rb_node);
-			err = reclaim_pages(dev, fwp->func_id,
-					    optimal_reclaimed_pages(),
-					    &nclaimed);
+			if (dev->state == MLX5_DEVICE_STATE_INTERNAL_ERROR) {
+				free_4k(dev, fwp->addr);
+				nclaimed = 1;
+			} else {
+				err = reclaim_pages(dev, fwp->func_id,
+						    optimal_reclaimed_pages(),
+						    &nclaimed);
+			}
 			if (err) {
 				mlx5_core_warn(dev, "failed reclaiming pages (%d)\n",
 					       err);
diff --git a/include/linux/mlx5/driver.h b/include/linux/mlx5/driver.h
index 62b7d439813d..9aba8d5139fa 100644
--- a/include/linux/mlx5/driver.h
+++ b/include/linux/mlx5/driver.h
@@ -487,8 +487,26 @@ struct mlx5_priv {
 	spinlock_t              ctx_lock;
 };
 
+enum mlx5_device_state {
+	MLX5_DEVICE_STATE_UP,
+	MLX5_DEVICE_STATE_INTERNAL_ERROR,
+};
+
+enum mlx5_interface_state {
+	MLX5_INTERFACE_STATE_DOWN,
+	MLX5_INTERFACE_STATE_UP,
+};
+
+enum mlx5_pci_status {
+	MLX5_PCI_STATUS_DISABLED,
+	MLX5_PCI_STATUS_ENABLED,
+};
+
 struct mlx5_core_dev {
 	struct pci_dev	       *pdev;
+	/* sync pci state */
+	struct mutex		pci_status_mutex;
+	enum mlx5_pci_status	pci_status;
 	u8			rev_id;
 	char			board_id[MLX5_BOARD_ID_LEN];
 	struct mlx5_cmd		cmd;
@@ -497,6 +515,10 @@ struct mlx5_core_dev {
 	u32 hca_caps_max[MLX5_CAP_NUM][MLX5_UN_SZ_DW(hca_cap_union)];
 	phys_addr_t		iseg_base;
 	struct mlx5_init_seg __iomem *iseg;
+	enum mlx5_device_state	state;
+	/* sync interface state */
+	struct mutex		intf_state_mutex;
+	enum mlx5_interface_state interface_state;
 	void			(*event) (struct mlx5_core_dev *dev,
 					  enum mlx5_dev_event event,
 					  unsigned long param);
-- 
cgit v1.2.3


From e3297246c2c8cf8548ba722da3e3a8104cdcd035 Mon Sep 17 00:00:00 2001
From: Eli Cohen <eli@mellanox.com>
Date: Wed, 14 Oct 2015 17:43:47 +0300
Subject: net/mlx5_core: Wait for FW readiness on startup

On device initialization, wait till firmware indicates that that it is done
with initialization before proceeding to initialize the device.

Also update initialization segment layout to match driver/firmware
interface definitions.

Signed-off-by: Eli Cohen <eli@mellanox.com>
Signed-off-by: Or Gerlitz <ogerlitz@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
---
 drivers/net/ethernet/mellanox/mlx5/core/main.c | 27 ++++++++++++++++++++++++++
 include/linux/mlx5/device.h                    |  3 ++-
 include/linux/mlx5/driver.h                    |  5 +++++
 3 files changed, 34 insertions(+), 1 deletion(-)

(limited to 'include/linux/mlx5')

diff --git a/drivers/net/ethernet/mellanox/mlx5/core/main.c b/drivers/net/ethernet/mellanox/mlx5/core/main.c
index a103a54d6660..2388aec208fa 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/main.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/main.c
@@ -39,6 +39,7 @@
 #include <linux/slab.h>
 #include <linux/io-mapping.h>
 #include <linux/interrupt.h>
+#include <linux/delay.h>
 #include <linux/mlx5/driver.h>
 #include <linux/mlx5/cq.h>
 #include <linux/mlx5/qp.h>
@@ -152,6 +153,25 @@ static struct mlx5_profile profile[] = {
 	},
 };
 
+#define FW_INIT_TIMEOUT_MILI	2000
+#define FW_INIT_WAIT_MS		2
+
+static int wait_fw_init(struct mlx5_core_dev *dev, u32 max_wait_mili)
+{
+	unsigned long end = jiffies + msecs_to_jiffies(max_wait_mili);
+	int err = 0;
+
+	while (fw_initializing(dev)) {
+		if (time_after(jiffies, end)) {
+			err = -EBUSY;
+			break;
+		}
+		msleep(FW_INIT_WAIT_MS);
+	}
+
+	return err;
+}
+
 static int set_dma_caps(struct pci_dev *pdev)
 {
 	int err;
@@ -913,6 +933,13 @@ static int mlx5_load_one(struct mlx5_core_dev *dev, struct mlx5_priv *priv)
 		goto out_err;
 	}
 
+	err = wait_fw_init(dev, FW_INIT_TIMEOUT_MILI);
+	if (err) {
+		dev_err(&dev->pdev->dev, "Firmware over %d MS in initializing state, aborting\n",
+			FW_INIT_TIMEOUT_MILI);
+		goto out_err;
+	}
+
 	mlx5_pagealloc_init(dev);
 
 	err = mlx5_core_enable_hca(dev);
diff --git a/include/linux/mlx5/device.h b/include/linux/mlx5/device.h
index 2a0b95662548..0b473cbfa7ef 100644
--- a/include/linux/mlx5/device.h
+++ b/include/linux/mlx5/device.h
@@ -439,7 +439,8 @@ struct mlx5_init_seg {
 	__be32			cmdq_addr_h;
 	__be32			cmdq_addr_l_sz;
 	__be32			cmd_dbell;
-	__be32			rsvd1[121];
+	__be32			rsvd1[120];
+	__be32			initializing;
 	struct health_buffer	health;
 	__be32			rsvd2[884];
 	__be32			health_counter;
diff --git a/include/linux/mlx5/driver.h b/include/linux/mlx5/driver.h
index 9aba8d5139fa..5c857f2a20d7 100644
--- a/include/linux/mlx5/driver.h
+++ b/include/linux/mlx5/driver.h
@@ -826,6 +826,11 @@ void mlx5_core_put_rsc(struct mlx5_core_rsc_common *common);
 int mlx5_query_odp_caps(struct mlx5_core_dev *dev,
 			struct mlx5_odp_caps *odp_caps);
 
+static inline int fw_initializing(struct mlx5_core_dev *dev)
+{
+	return ioread32be(&dev->iseg->initializing) >> 31;
+}
+
 static inline u32 mlx5_mkey_to_idx(u32 mkey)
 {
 	return mkey >> 8;
-- 
cgit v1.2.3