1 files changed, 496 insertions, 73 deletions

diff --git a/drivers/net/ethernet/mellanox/mlx5/core/health.c b/drivers/net/ethernet/mellanox/mlx5/core/health.c
index a2656f4008d9..2fe6923f7ce0 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/health.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/health.c
@@ -40,6 +40,8 @@
 #include "mlx5_core.h"
 #include "lib/eq.h"
 #include "lib/mlx5.h"
+#include "lib/pci_vsc.h"
+#include "diag/fw_tracer.h"
 
 enum {
 	MLX5_HEALTH_POLL_INTERVAL	= 2 * HZ,
@@ -62,12 +64,20 @@ enum {
 
 enum {
 	MLX5_DROP_NEW_HEALTH_WORK,
-	MLX5_DROP_NEW_RECOVERY_WORK,
+};
+
+enum  {
+	MLX5_SENSOR_NO_ERR		= 0,
+	MLX5_SENSOR_PCI_COMM_ERR	= 1,
+	MLX5_SENSOR_PCI_ERR		= 2,
+	MLX5_SENSOR_NIC_DISABLED	= 3,
+	MLX5_SENSOR_NIC_SW_RESET	= 4,
+	MLX5_SENSOR_FW_SYND_RFR		= 5,
 };
 
 u8 mlx5_get_nic_state(struct mlx5_core_dev *dev)
 {
-	return (ioread32be(&dev->iseg->cmdq_addr_l_sz) >> 8) & 3;
+	return (ioread32be(&dev->iseg->cmdq_addr_l_sz) >> 8) & 7;
 }
 
 void mlx5_set_nic_state(struct mlx5_core_dev *dev, u8 state)
@@ -80,18 +90,105 @@ void mlx5_set_nic_state(struct mlx5_core_dev *dev, u8 state)
 		    &dev->iseg->cmdq_addr_l_sz);
 }
 
-static int in_fatal(struct mlx5_core_dev *dev)
+static bool sensor_pci_not_working(struct mlx5_core_dev *dev)
 {
 	struct mlx5_core_health *health = &dev->priv.health;
 	struct health_buffer __iomem *h = health->health;
 
+	/* Offline PCI reads return 0xffffffff */
+	return (ioread32be(&h->fw_ver) == 0xffffffff);
+}
+
+static bool sensor_fw_synd_rfr(struct mlx5_core_dev *dev)
+{
+	struct mlx5_core_health *health = &dev->priv.health;
+	struct health_buffer __iomem *h = health->health;
+	u32 rfr = ioread32be(&h->rfr) >> MLX5_RFR_OFFSET;
+	u8 synd = ioread8(&h->synd);
+
+	if (rfr && synd)
+		mlx5_core_dbg(dev, "FW requests reset, synd: %d\n", synd);
+	return rfr && synd;
+}
+
+static u32 check_fatal_sensors(struct mlx5_core_dev *dev)
+{
+	if (sensor_pci_not_working(dev))
+		return MLX5_SENSOR_PCI_COMM_ERR;
+	if (pci_channel_offline(dev->pdev))
+		return MLX5_SENSOR_PCI_ERR;
 	if (mlx5_get_nic_state(dev) == MLX5_NIC_IFC_DISABLED)
-		return 1;
+		return MLX5_SENSOR_NIC_DISABLED;
+	if (mlx5_get_nic_state(dev) == MLX5_NIC_IFC_SW_RESET)
+		return MLX5_SENSOR_NIC_SW_RESET;
+	if (sensor_fw_synd_rfr(dev))
+		return MLX5_SENSOR_FW_SYND_RFR;
 
-	if (ioread32be(&h->fw_ver) == 0xffffffff)
-		return 1;
+	return MLX5_SENSOR_NO_ERR;
+}
 
-	return 0;
+static int lock_sem_sw_reset(struct mlx5_core_dev *dev, bool lock)
+{
+	enum mlx5_vsc_state state;
+	int ret;
+
+	if (!mlx5_core_is_pf(dev))
+		return -EBUSY;
+
+	/* Try to lock GW access, this stage doesn't return
+	 * EBUSY because locked GW does not mean that other PF
+	 * already started the reset.
+	 */
+	ret = mlx5_vsc_gw_lock(dev);
+	if (ret == -EBUSY)
+		return -EINVAL;
+	if (ret)
+		return ret;
+
+	state = lock ? MLX5_VSC_LOCK : MLX5_VSC_UNLOCK;
+	/* At this stage, if the return status == EBUSY, then we know
+	 * for sure that another PF started the reset, so don't allow
+	 * another reset.
+	 */
+	ret = mlx5_vsc_sem_set_space(dev, MLX5_SEMAPHORE_SW_RESET, state);
+	if (ret)
+		mlx5_core_warn(dev, "Failed to lock SW reset semaphore\n");
+
+	/* Unlock GW access */
+	mlx5_vsc_gw_unlock(dev);
+
+	return ret;
+}
+
+static bool reset_fw_if_needed(struct mlx5_core_dev *dev)
+{
+	bool supported = (ioread32be(&dev->iseg->initializing) >>
+			  MLX5_FW_RESET_SUPPORTED_OFFSET) & 1;
+	u32 fatal_error;
+
+	if (!supported)
+		return false;
+
+	/* The reset only needs to be issued by one PF. The health buffer is
+	 * shared between all functions, and will be cleared during a reset.
+	 * Check again to avoid a redundant 2nd reset. If the fatal erros was
+	 * PCI related a reset won't help.
+	 */
+	fatal_error = check_fatal_sensors(dev);
+	if (fatal_error == MLX5_SENSOR_PCI_COMM_ERR ||
+	    fatal_error == MLX5_SENSOR_NIC_DISABLED ||
+	    fatal_error == MLX5_SENSOR_NIC_SW_RESET) {
+		mlx5_core_warn(dev, "Not issuing FW reset. Either it's already done or won't help.");
+		return false;
+	}
+
+	mlx5_core_warn(dev, "Issuing FW Reset\n");
+	/* Write the NIC interface field to initiate the reset, the command
+	 * interface address also resides here, don't overwrite it.
+	 */
+	mlx5_set_nic_state(dev, MLX5_NIC_IFC_SW_RESET);
+
+	return true;
 }
 
 void mlx5_enter_error_state(struct mlx5_core_dev *dev, bool force)
@@ -99,14 +196,65 @@ void mlx5_enter_error_state(struct mlx5_core_dev *dev, bool force)
 	mutex_lock(&dev->intf_state_mutex);
 	if (dev->state == MLX5_DEVICE_STATE_INTERNAL_ERROR)
 		goto unlock;
+	if (dev->state == MLX5_DEVICE_STATE_UNINITIALIZED) {
+		dev->state = MLX5_DEVICE_STATE_INTERNAL_ERROR;
+		goto unlock;
+	}
 
-	mlx5_core_err(dev, "start\n");
-	if (pci_channel_offline(dev->pdev) || in_fatal(dev) || force) {
+	if (check_fatal_sensors(dev) || force) {
 		dev->state = MLX5_DEVICE_STATE_INTERNAL_ERROR;
 		mlx5_cmd_flush(dev);
 	}
 
 	mlx5_notifier_call_chain(dev->priv.events, MLX5_DEV_EVENT_SYS_ERROR, (void *)1);
+unlock:
+	mutex_unlock(&dev->intf_state_mutex);
+}
+
+#define MLX5_CRDUMP_WAIT_MS	60000
+#define MLX5_FW_RESET_WAIT_MS	1000
+void mlx5_error_sw_reset(struct mlx5_core_dev *dev)
+{
+	unsigned long end, delay_ms = MLX5_FW_RESET_WAIT_MS;
+	int lock = -EBUSY;
+
+	mutex_lock(&dev->intf_state_mutex);
+	if (dev->state != MLX5_DEVICE_STATE_INTERNAL_ERROR)
+		goto unlock;
+
+	mlx5_core_err(dev, "start\n");
+
+	if (check_fatal_sensors(dev) == MLX5_SENSOR_FW_SYND_RFR) {
+		/* Get cr-dump and reset FW semaphore */
+		lock = lock_sem_sw_reset(dev, true);
+
+		if (lock == -EBUSY) {
+			delay_ms = MLX5_CRDUMP_WAIT_MS;
+			goto recover_from_sw_reset;
+		}
+		/* Execute SW reset */
+		reset_fw_if_needed(dev);
+	}
+
+recover_from_sw_reset:
+	/* Recover from SW reset */
+	end = jiffies + msecs_to_jiffies(delay_ms);
+	do {
+		if (mlx5_get_nic_state(dev) == MLX5_NIC_IFC_DISABLED)
+			break;
+
+		cond_resched();
+	} while (!time_after(jiffies, end));
+
+	if (mlx5_get_nic_state(dev) != MLX5_NIC_IFC_DISABLED) {
+		dev_err(&dev->pdev->dev, "NIC IFC still %d after %lums.\n",
+			mlx5_get_nic_state(dev), delay_ms);
+	}
+
+	/* Release FW semaphore if you are the lock owner */
+	if (!lock)
+		lock_sem_sw_reset(dev, false);
+
 	mlx5_core_err(dev, "end\n");
 
 unlock:
@@ -129,6 +277,20 @@ static void mlx5_handle_bad_state(struct mlx5_core_dev *dev)
 	case MLX5_NIC_IFC_NO_DRAM_NIC:
 		mlx5_core_warn(dev, "Expected to see disabled NIC but it is no dram nic\n");
 		break;
+
+	case MLX5_NIC_IFC_SW_RESET:
+		/* The IFC mode field is 3 bits, so it will read 0x7 in 2 cases:
+		 * 1. PCI has been disabled (ie. PCI-AER, PF driver unloaded
+		 *    and this is a VF), this is not recoverable by SW reset.
+		 *    Logging of this is handled elsewhere.
+		 * 2. FW reset has been issued by another function, driver can
+		 *    be reloaded to recover after the mode switches to
+		 *    MLX5_NIC_IFC_DISABLED.
+		 */
+		if (dev->priv.health.fatal_error != MLX5_SENSOR_PCI_COMM_ERR)
+			mlx5_core_warn(dev, "NIC SW reset in progress\n");
+		break;
+
 	default:
 		mlx5_core_warn(dev, "Expected to see disabled NIC but it is has invalid value %d\n",
 			       nic_interface);
@@ -137,52 +299,32 @@ static void mlx5_handle_bad_state(struct mlx5_core_dev *dev)
 	mlx5_disable_device(dev);
 }
 
-static void health_recover(struct work_struct *work)
-{
-	struct mlx5_core_health *health;
-	struct delayed_work *dwork;
-	struct mlx5_core_dev *dev;
-	struct mlx5_priv *priv;
-	u8 nic_state;
-
-	dwork = container_of(work, struct delayed_work, work);
-	health = container_of(dwork, struct mlx5_core_health, recover_work);
-	priv = container_of(health, struct mlx5_priv, health);
-	dev = container_of(priv, struct mlx5_core_dev, priv);
-
-	nic_state = mlx5_get_nic_state(dev);
-	if (nic_state == MLX5_NIC_IFC_INVALID) {
-		mlx5_core_err(dev, "health recovery flow aborted since the nic state is invalid\n");
-		return;
-	}
-
-	mlx5_core_err(dev, "starting health recovery flow\n");
-	mlx5_recover_device(dev);
-}
-
 /* How much time to wait until health resetting the driver (in msecs) */
-#define MLX5_RECOVERY_DELAY_MSECS 60000
-static void health_care(struct work_struct *work)
+#define MLX5_RECOVERY_WAIT_MSECS 60000
+static int mlx5_health_try_recover(struct mlx5_core_dev *dev)
 {
-	unsigned long recover_delay = msecs_to_jiffies(MLX5_RECOVERY_DELAY_MSECS);
-	struct mlx5_core_health *health;
-	struct mlx5_core_dev *dev;
-	struct mlx5_priv *priv;
-	unsigned long flags;
+	unsigned long end;
 
-	health = container_of(work, struct mlx5_core_health, work);
-	priv = container_of(health, struct mlx5_priv, health);
-	dev = container_of(priv, struct mlx5_core_dev, priv);
 	mlx5_core_warn(dev, "handling bad device here\n");
 	mlx5_handle_bad_state(dev);
+	end = jiffies + msecs_to_jiffies(MLX5_RECOVERY_WAIT_MSECS);
+	while (sensor_pci_not_working(dev)) {
+		if (time_after(jiffies, end)) {
+			mlx5_core_err(dev,
+				      "health recovery flow aborted, PCI reads still not working\n");
+			return -EIO;
+		}
+		msleep(100);
+	}
 
-	spin_lock_irqsave(&health->wq_lock, flags);
-	if (!test_bit(MLX5_DROP_NEW_RECOVERY_WORK, &health->flags))
-		schedule_delayed_work(&health->recover_work, recover_delay);
-	else
-		mlx5_core_err(dev,
-			      "new health works are not permitted at this stage\n");
-	spin_unlock_irqrestore(&health->wq_lock, flags);
+	mlx5_core_err(dev, "starting health recovery flow\n");
+	mlx5_recover_device(dev);
+	if (!test_bit(MLX5_INTERFACE_STATE_UP, &dev->intf_state) ||
+	    check_fatal_sensors(dev)) {
+		mlx5_core_err(dev, "health recovery failed\n");
+		return -EIO;
+	}
+	return 0;
 }
 
 static const char *hsynd_str(u8 synd)
@@ -246,6 +388,282 @@ static void print_health_info(struct mlx5_core_dev *dev)
 	mlx5_core_err(dev, "raw fw_ver 0x%08x\n", fw);
 }
 
+static int
+mlx5_fw_reporter_diagnose(struct devlink_health_reporter *reporter,
+			  struct devlink_fmsg *fmsg)
+{
+	struct mlx5_core_dev *dev = devlink_health_reporter_priv(reporter);
+	struct mlx5_core_health *health = &dev->priv.health;
+	struct health_buffer __iomem *h = health->health;
+	u8 synd;
+	int err;
+
+	synd = ioread8(&h->synd);
+	err = devlink_fmsg_u8_pair_put(fmsg, "Syndrome", synd);
+	if (err || !synd)
+		return err;
+	return devlink_fmsg_string_pair_put(fmsg, "Description", hsynd_str(synd));
+}
+
+struct mlx5_fw_reporter_ctx {
+	u8 err_synd;
+	int miss_counter;
+};
+
+static int
+mlx5_fw_reporter_ctx_pairs_put(struct devlink_fmsg *fmsg,
+			       struct mlx5_fw_reporter_ctx *fw_reporter_ctx)
+{
+	int err;
+
+	err = devlink_fmsg_u8_pair_put(fmsg, "syndrome",
+				       fw_reporter_ctx->err_synd);
+	if (err)
+		return err;
+	err = devlink_fmsg_u32_pair_put(fmsg, "fw_miss_counter",
+					fw_reporter_ctx->miss_counter);
+	if (err)
+		return err;
+	return 0;
+}
+
+static int
+mlx5_fw_reporter_heath_buffer_data_put(struct mlx5_core_dev *dev,
+				       struct devlink_fmsg *fmsg)
+{
+	struct mlx5_core_health *health = &dev->priv.health;
+	struct health_buffer __iomem *h = health->health;
+	int err;
+	int i;
+
+	if (!ioread8(&h->synd))
+		return 0;
+
+	err = devlink_fmsg_pair_nest_start(fmsg, "health buffer");
+	if (err)
+		return err;
+	err = devlink_fmsg_obj_nest_start(fmsg);
+	if (err)
+		return err;
+	err = devlink_fmsg_arr_pair_nest_start(fmsg, "assert_var");
+	if (err)
+		return err;
+
+	for (i = 0; i < ARRAY_SIZE(h->assert_var); i++) {
+		err = devlink_fmsg_u32_put(fmsg, ioread32be(h->assert_var + i));
+		if (err)
+			return err;
+	}
+	err = devlink_fmsg_arr_pair_nest_end(fmsg);
+	if (err)
+		return err;
+	err = devlink_fmsg_u32_pair_put(fmsg, "assert_exit_ptr",
+					ioread32be(&h->assert_exit_ptr));
+	if (err)
+		return err;
+	err = devlink_fmsg_u32_pair_put(fmsg, "assert_callra",
+					ioread32be(&h->assert_callra));
+	if (err)
+		return err;
+	err = devlink_fmsg_u32_pair_put(fmsg, "hw_id", ioread32be(&h->hw_id));
+	if (err)
+		return err;
+	err = devlink_fmsg_u8_pair_put(fmsg, "irisc_index",
+				       ioread8(&h->irisc_index));
+	if (err)
+		return err;
+	err = devlink_fmsg_u8_pair_put(fmsg, "synd", ioread8(&h->synd));
+	if (err)
+		return err;
+	err = devlink_fmsg_u32_pair_put(fmsg, "ext_synd",
+					ioread16be(&h->ext_synd));
+	if (err)
+		return err;
+	err = devlink_fmsg_u32_pair_put(fmsg, "raw_fw_ver",
+					ioread32be(&h->fw_ver));
+	if (err)
+		return err;
+	err = devlink_fmsg_obj_nest_end(fmsg);
+	if (err)
+		return err;
+	return devlink_fmsg_pair_nest_end(fmsg);
+}
+
+static int
+mlx5_fw_reporter_dump(struct devlink_health_reporter *reporter,
+		      struct devlink_fmsg *fmsg, void *priv_ctx)
+{
+	struct mlx5_core_dev *dev = devlink_health_reporter_priv(reporter);
+	int err;
+
+	err = mlx5_fw_tracer_trigger_core_dump_general(dev);
+	if (err)
+		return err;
+
+	if (priv_ctx) {
+		struct mlx5_fw_reporter_ctx *fw_reporter_ctx = priv_ctx;
+
+		err = mlx5_fw_reporter_ctx_pairs_put(fmsg, fw_reporter_ctx);
+		if (err)
+			return err;
+	}
+
+	err = mlx5_fw_reporter_heath_buffer_data_put(dev, fmsg);
+	if (err)
+		return err;
+	return mlx5_fw_tracer_get_saved_traces_objects(dev->tracer, fmsg);
+}
+
+static void mlx5_fw_reporter_err_work(struct work_struct *work)
+{
+	struct mlx5_fw_reporter_ctx fw_reporter_ctx;
+	struct mlx5_core_health *health;
+
+	health = container_of(work, struct mlx5_core_health, report_work);
+
+	if (IS_ERR_OR_NULL(health->fw_reporter))
+		return;
+
+	fw_reporter_ctx.err_synd = health->synd;
+	fw_reporter_ctx.miss_counter = health->miss_counter;
+	if (fw_reporter_ctx.err_synd) {
+		devlink_health_report(health->fw_reporter,
+				      "FW syndrom reported", &fw_reporter_ctx);
+		return;
+	}
+	if (fw_reporter_ctx.miss_counter)
+		devlink_health_report(health->fw_reporter,
+				      "FW miss counter reported",
+				      &fw_reporter_ctx);
+}
+
+static const struct devlink_health_reporter_ops mlx5_fw_reporter_ops = {
+		.name = "fw",
+		.diagnose = mlx5_fw_reporter_diagnose,
+		.dump = mlx5_fw_reporter_dump,
+};
+
+static int
+mlx5_fw_fatal_reporter_recover(struct devlink_health_reporter *reporter,
+			       void *priv_ctx)
+{
+	struct mlx5_core_dev *dev = devlink_health_reporter_priv(reporter);
+
+	return mlx5_health_try_recover(dev);
+}
+
+#define MLX5_CR_DUMP_CHUNK_SIZE 256
+static int
+mlx5_fw_fatal_reporter_dump(struct devlink_health_reporter *reporter,
+			    struct devlink_fmsg *fmsg, void *priv_ctx)
+{
+	struct mlx5_core_dev *dev = devlink_health_reporter_priv(reporter);
+	u32 crdump_size = dev->priv.health.crdump_size;
+	u32 *cr_data;
+	u32 data_size;
+	u32 offset;
+	int err;
+
+	if (!mlx5_core_is_pf(dev))
+		return -EPERM;
+
+	cr_data = kvmalloc(crdump_size, GFP_KERNEL);
+	if (!cr_data)
+		return -ENOMEM;
+	err = mlx5_crdump_collect(dev, cr_data);
+	if (err)
+		return err;
+
+	if (priv_ctx) {
+		struct mlx5_fw_reporter_ctx *fw_reporter_ctx = priv_ctx;
+
+		err = mlx5_fw_reporter_ctx_pairs_put(fmsg, fw_reporter_ctx);
+		if (err)
+			goto free_data;
+	}
+
+	err = devlink_fmsg_arr_pair_nest_start(fmsg, "crdump_data");
+	if (err)
+		goto free_data;
+	for (offset = 0; offset < crdump_size; offset += data_size) {
+		if (crdump_size - offset < MLX5_CR_DUMP_CHUNK_SIZE)
+			data_size = crdump_size - offset;
+		else
+			data_size = MLX5_CR_DUMP_CHUNK_SIZE;
+		err = devlink_fmsg_binary_put(fmsg, cr_data, data_size);
+		if (err)
+			goto free_data;
+	}
+	err = devlink_fmsg_arr_pair_nest_end(fmsg);
+
+free_data:
+	kfree(cr_data);
+	return err;
+}
+
+static void mlx5_fw_fatal_reporter_err_work(struct work_struct *work)
+{
+	struct mlx5_fw_reporter_ctx fw_reporter_ctx;
+	struct mlx5_core_health *health;
+	struct mlx5_core_dev *dev;
+	struct mlx5_priv *priv;
+
+	health = container_of(work, struct mlx5_core_health, fatal_report_work);
+	priv = container_of(health, struct mlx5_priv, health);
+	dev = container_of(priv, struct mlx5_core_dev, priv);
+
+	mlx5_enter_error_state(dev, false);
+	if (IS_ERR_OR_NULL(health->fw_fatal_reporter)) {
+		if (mlx5_health_try_recover(dev))
+			mlx5_core_err(dev, "health recovery failed\n");
+		return;
+	}
+	fw_reporter_ctx.err_synd = health->synd;
+	fw_reporter_ctx.miss_counter = health->miss_counter;
+	devlink_health_report(health->fw_fatal_reporter,
+			      "FW fatal error reported", &fw_reporter_ctx);
+}
+
+static const struct devlink_health_reporter_ops mlx5_fw_fatal_reporter_ops = {
+		.name = "fw_fatal",
+		.recover = mlx5_fw_fatal_reporter_recover,
+		.dump = mlx5_fw_fatal_reporter_dump,
+};
+
+#define MLX5_REPORTER_FW_GRACEFUL_PERIOD 1200000
+static void mlx5_fw_reporters_create(struct mlx5_core_dev *dev)
+{
+	struct mlx5_core_health *health = &dev->priv.health;
+	struct devlink *devlink = priv_to_devlink(dev);
+
+	health->fw_reporter =
+		devlink_health_reporter_create(devlink, &mlx5_fw_reporter_ops,
+					       0, false, dev);
+	if (IS_ERR(health->fw_reporter))
+		mlx5_core_warn(dev, "Failed to create fw reporter, err = %ld\n",
+			       PTR_ERR(health->fw_reporter));
+
+	health->fw_fatal_reporter =
+		devlink_health_reporter_create(devlink,
+					       &mlx5_fw_fatal_reporter_ops,
+					       MLX5_REPORTER_FW_GRACEFUL_PERIOD,
+					       true, dev);
+	if (IS_ERR(health->fw_fatal_reporter))
+		mlx5_core_warn(dev, "Failed to create fw fatal reporter, err = %ld\n",
+			       PTR_ERR(health->fw_fatal_reporter));
+}
+
+static void mlx5_fw_reporters_destroy(struct mlx5_core_dev *dev)
+{
+	struct mlx5_core_health *health = &dev->priv.health;
+
+	if (!IS_ERR_OR_NULL(health->fw_reporter))
+		devlink_health_reporter_destroy(health->fw_reporter);
+
+	if (!IS_ERR_OR_NULL(health->fw_fatal_reporter))
+		devlink_health_reporter_destroy(health->fw_fatal_reporter);
+}
+
 static unsigned long get_next_poll_jiffies(void)
 {
 	unsigned long next;
@@ -264,7 +682,7 @@ void mlx5_trigger_health_work(struct mlx5_core_dev *dev)
 
 	spin_lock_irqsave(&health->wq_lock, flags);
 	if (!test_bit(MLX5_DROP_NEW_HEALTH_WORK, &health->flags))
-		queue_work(health->wq, &health->work);
+		queue_work(health->wq, &health->fatal_report_work);
 	else
 		mlx5_core_err(dev, "new health works are not permitted at this stage\n");
 	spin_unlock_irqrestore(&health->wq_lock, flags);
@@ -274,6 +692,9 @@ static void poll_health(struct timer_list *t)
 {
 	struct mlx5_core_dev *dev = from_timer(dev, t, priv.health.timer);
 	struct mlx5_core_health *health = &dev->priv.health;
+	struct health_buffer __iomem *h = health->health;
+	u32 fatal_error;
+	u8 prev_synd;
 	u32 count;
 
 	if (dev->state == MLX5_DEVICE_STATE_INTERNAL_ERROR)
@@ -289,10 +710,19 @@ static void poll_health(struct timer_list *t)
 	if (health->miss_counter == MAX_MISSES) {
 		mlx5_core_err(dev, "device's health compromised - reached miss count\n");
 		print_health_info(dev);
+		queue_work(health->wq, &health->report_work);
 	}
 
-	if (in_fatal(dev) && !health->sick) {
-		health->sick = true;
+	prev_synd = health->synd;
+	health->synd = ioread8(&h->synd);
+	if (health->synd && health->synd != prev_synd)
+		queue_work(health->wq, &health->report_work);
+
+	fatal_error = check_fatal_sensors(dev);
+
+	if (fatal_error && !health->fatal_error) {
+		mlx5_core_err(dev, "Fatal error %u detected\n", fatal_error);
+		dev->priv.health.fatal_error = fatal_error;
 		print_health_info(dev);
 		mlx5_trigger_health_work(dev);
 	}
@@ -306,9 +736,8 @@ void mlx5_start_health_poll(struct mlx5_core_dev *dev)
 	struct mlx5_core_health *health = &dev->priv.health;
 
 	timer_setup(&health->timer, poll_health, 0);
-	health->sick = 0;
+	health->fatal_error = MLX5_SENSOR_NO_ERR;
 	clear_bit(MLX5_DROP_NEW_HEALTH_WORK, &health->flags);
-	clear_bit(MLX5_DROP_NEW_RECOVERY_WORK, &health->flags);
 	health->health = &dev->iseg->health;
 	health->health_counter = &dev->iseg->health_counter;
 
@@ -324,7 +753,6 @@ void mlx5_stop_health_poll(struct mlx5_core_dev *dev, bool disable_health)
 	if (disable_health) {
 		spin_lock_irqsave(&health->wq_lock, flags);
 		set_bit(MLX5_DROP_NEW_HEALTH_WORK, &health->flags);
-		set_bit(MLX5_DROP_NEW_RECOVERY_WORK, &health->flags);
 		spin_unlock_irqrestore(&health->wq_lock, flags);
 	}
 
@@ -338,21 +766,9 @@ void mlx5_drain_health_wq(struct mlx5_core_dev *dev)
 
 	spin_lock_irqsave(&health->wq_lock, flags);
 	set_bit(MLX5_DROP_NEW_HEALTH_WORK, &health->flags);
-	set_bit(MLX5_DROP_NEW_RECOVERY_WORK, &health->flags);
 	spin_unlock_irqrestore(&health->wq_lock, flags);
-	cancel_delayed_work_sync(&health->recover_work);
-	cancel_work_sync(&health->work);
-}
-
-void mlx5_drain_health_recovery(struct mlx5_core_dev *dev)
-{
-	struct mlx5_core_health *health = &dev->priv.health;
-	unsigned long flags;
-
-	spin_lock_irqsave(&health->wq_lock, flags);
-	set_bit(MLX5_DROP_NEW_RECOVERY_WORK, &health->flags);
-	spin_unlock_irqrestore(&health->wq_lock, flags);
-	cancel_delayed_work_sync(&dev->priv.health.recover_work);
+	cancel_work_sync(&health->report_work);
+	cancel_work_sync(&health->fatal_report_work);
 }
 
 void mlx5_health_flush(struct mlx5_core_dev *dev)
@@ -367,6 +783,7 @@ void mlx5_health_cleanup(struct mlx5_core_dev *dev)
 	struct mlx5_core_health *health = &dev->priv.health;
 
 	destroy_workqueue(health->wq);
+	mlx5_fw_reporters_destroy(dev);
 }
 
 int mlx5_health_init(struct mlx5_core_dev *dev)
@@ -374,20 +791,26 @@ int mlx5_health_init(struct mlx5_core_dev *dev)
 	struct mlx5_core_health *health;
 	char *name;
 
+	mlx5_fw_reporters_create(dev);
+
 	health = &dev->priv.health;
 	name = kmalloc(64, GFP_KERNEL);
 	if (!name)
-		return -ENOMEM;
+		goto out_err;
 
 	strcpy(name, "mlx5_health");
 	strcat(name, dev_name(dev->device));
 	health->wq = create_singlethread_workqueue(name);
 	kfree(name);
 	if (!health->wq)
-		return -ENOMEM;
+		goto out_err;
 	spin_lock_init(&health->wq_lock);
-	INIT_WORK(&health->work, health_care);
-	INIT_DELAYED_WORK(&health->recover_work, health_recover);
+	INIT_WORK(&health->fatal_report_work, mlx5_fw_fatal_reporter_err_work);
+	INIT_WORK(&health->report_work, mlx5_fw_reporter_err_work);
 
 	return 0;
+
+out_err:
+	mlx5_fw_reporters_destroy(dev);
+	return -ENOMEM;
 }