Merge branch 'devlink-health'

Eran Ben Elisha says:

====================
Devlink health reporting and recovery system

The health mechanism is targeted for Real Time Alerting, in order to know when
something bad had happened to a PCI device
- Provide alert debug information
- Self healing
- If problem needs vendor support, provide a way to gather all needed debugging
  information.

The main idea is to unify and centralize driver health reports in the
generic devlink instance and allow the user to set different
attributes of the health reporting and recovery procedures.

The devlink health reporter:
Device driver creates a "health reporter" per each error/health type.
Error/Health type can be a known/generic (eg pci error, fw error, rx/tx error)
or unknown (driver specific).
For each registered health reporter a driver can issue error/health reports
asynchronously. All health reports handling is done by devlink.
Device driver can provide specific callbacks for each "health reporter", e.g.
 - Recovery procedures
 - Diagnostics and object dump procedures
 - OOB initial attributes
Different parts of the driver can register different types of health reporters
with different handlers.

Once an error is reported, devlink health will do the following actions:
  * A log is being send to the kernel trace events buffer
  * Health status and statistics are being updated for the reporter instance
  * Object dump is being taken and saved at the reporter instance (as long as
    there is no other dump which is already stored)
  * Auto recovery attempt is being done. Depends on:
    - Auto-recovery configuration
    - Grace period vs. time passed since last recover

The user interface:
User can access/change each reporter attributes and driver specific callbacks
via devlink, e.g per error type (per health reporter)
 - Configure reporter's generic attributes (like: Disable/enable auto recovery)
 - Invoke recovery procedure
 - Run diagnostics
 - Object dump

The devlink health interface (via netlink):
DEVLINK_CMD_HEALTH_REPORTER_GET
  Retrieves status and configuration info per DEV and reporter.
DEVLINK_CMD_HEALTH_REPORTER_SET
  Allows reporter-related configuration setting.
DEVLINK_CMD_HEALTH_REPORTER_RECOVER
  Triggers a reporter's recovery procedure.
DEVLINK_CMD_HEALTH_REPORTER_DIAGNOSE
  Retrieves diagnostics data from a reporter on a device.
DEVLINK_CMD_HEALTH_REPORTER_DUMP_GET
  Retrieves the last stored dump. Devlink health
  saves a single dump. If an dump is not already stored by the devlink
  for this reporter, devlink generates a new dump.
  dump output is defined by the reporter.
DEVLINK_CMD_HEALTH_REPORTER_DUMP_CLEAR
  Clears the last saved dump file for the specified reporter.

                                               netlink
                                      +--------------------------+
                                      |                          |
                                      |            +             |
                                      |            |             |
                                      +--------------------------+
                                                   |request for ops
                                                   |(diagnose,
 mlx5_core                             devlink     |recover,
                                                   |dump)
+--------+                            +--------------------------+
|        |                            |    reporter|             |
|        |                            |  +---------v----------+  |
|        |   ops execution            |  |                    |  |
|     <----------------------------------+                    |  |
|        |                            |  |                    |  |
|        |                            |  + ^------------------+  |
|        |                            |    | request for ops     |
|        |                            |    | (recover, dump)     |
|        |                            |    |                     |
|        |                            |  +-+------------------+  |
|        |     health report          |  | health handler     |  |
|        +------------------------------->                    |  |
|        |                            |  +--------------------+  |
|        |     health reporter create |                          |
|        +---------------------------->                          |
+--------+                            +--------------------------+

In this patchset, mlx5e TX reporter is implemented.

Cmdline format:
    devlink health show [DEV reporter REPORTE_NAME]
    devlink health recover DEV reporter REPORTER_NAME
    devlink health diagnose DEV reporter REPORTER_NAME
    devlink health dump show DEV reporter REPORTER_NAME
    devlink health dump clear DEV reporter REPORTER_NAME
    devlink health set DEV reporter REPORTER_NAME NAME VALUE

Cmdline examples:
$devlink health show
pci/0000:00:09.0:
  name tx
    state healthy #err 1 #recover 0 last_dump_ts N/A
    parameters:
      grace_period 500 auto_recover false

$devlink health diagnose pci/0000:00:09.0 reporter tx -j -p
{
    "SQs": [ {
            "sqn": 138,
            "HW state": 1,
            "stopped": false
        },{
            "sqn": 142,
            "HW state": 1,
            "stopped": false
        } ]
}

$devlink health diagnose pci/0000:00:09.0 reporter tx
SQs:
  sqn: 138 HW state: 1 stopped: false
  sqn: 142 HW state: 1 stopped: false

$devlink health recover pci/0000:00:09 reporter tx

$devlink health set pci/0000:00:09.0 reporter tx grace_period 3500

$devlink health set pci/0000:00:09.0 reporter tx auto_recover false

Changelog:
v4:
- Rebase on latest net-next
- Remove trace_devlink_health signature exposure in case CONFIG_NET_DEVLINK is
  not defined as it shall only be used from devlink.

v3:
- Redesign of devlink <-> driver fmsg API
- Various bug fixes

v2:
- Remove FW* reporters to decrease the amount of patches in the patchset
====================
Signed-off-by: David S. Miller <davem@davemloft.net>

Documentation/networking/devlink-health.txt

+The health mechanism is targeted for Real Time Alerting, in order to know when
+something bad had happened to a PCI device
+- Provide alert debug information
+- Self healing
+- If problem needs vendor support, provide a way to gather all needed debugging
+  information.
+
+The main idea is to unify and centralize driver health reports in the
+generic devlink instance and allow the user to set different
+attributes of the health reporting and recovery procedures.
+
+The devlink health reporter:
+Device driver creates a "health reporter" per each error/health type.
+Error/Health type can be a known/generic (eg pci error, fw error, rx/tx error)
+or unknown (driver specific).
+For each registered health reporter a driver can issue error/health reports
+asynchronously. All health reports handling is done by devlink.
+Device driver can provide specific callbacks for each "health reporter", e.g.
+ - Recovery procedures
+ - Diagnostics and object dump procedures
+ - OOB initial parameters
+Different parts of the driver can register different types of health reporters
+with different handlers.
+
+Once an error is reported, devlink health will do the following actions:
+  * A log is being send to the kernel trace events buffer
+  * Health status and statistics are being updated for the reporter instance
+  * Object dump is being taken and saved at the reporter instance (as long as
+    there is no other dump which is already stored)
+  * Auto recovery attempt is being done. Depends on:
+    - Auto-recovery configuration
+    - Grace period vs. time passed since last recover
+
+The user interface:
+User can access/change each reporter's parameters and driver specific callbacks
+via devlink, e.g per error type (per health reporter)
+ - Configure reporter's generic parameters (like: disable/enable auto recovery)
+ - Invoke recovery procedure
+ - Run diagnostics
+ - Object dump
+
+The devlink health interface (via netlink):
+DEVLINK_CMD_HEALTH_REPORTER_GET
+  Retrieves status and configuration info per DEV and reporter.
+DEVLINK_CMD_HEALTH_REPORTER_SET
+  Allows reporter-related configuration setting.
+DEVLINK_CMD_HEALTH_REPORTER_RECOVER
+  Triggers a reporter's recovery procedure.
+DEVLINK_CMD_HEALTH_REPORTER_DIAGNOSE
+  Retrieves diagnostics data from a reporter on a device.
+DEVLINK_CMD_HEALTH_REPORTER_DUMP_GET
+  Retrieves the last stored dump. Devlink health
+  saves a single dump. If an dump is not already stored by the devlink
+  for this reporter, devlink generates a new dump.
+  dump output is defined by the reporter.
+DEVLINK_CMD_HEALTH_REPORTER_DUMP_CLEAR
+  Clears the last saved dump file for the specified reporter.
+
+
+                                               netlink
+                                      +--------------------------+
+                                      |                          |
+                                      |            +             |
+                                      |            |             |
+                                      +--------------------------+
+                                                   |request for ops
+                                                   |(diagnose,
+ mlx5_core                             devlink     |recover,
+                                                   |dump)
++--------+                            +--------------------------+
+|        |                            |    reporter|             |
+|        |                            |  +---------v----------+  |
+|        |   ops execution            |  |                    |  |
+|     <----------------------------------+                    |  |
+|        |                            |  |                    |  |
+|        |                            |  + ^------------------+  |
+|        |                            |    | request for ops     |
+|        |                            |    | (recover, dump)     |
+|        |                            |    |                     |
+|        |                            |  +-+------------------+  |
+|        |     health report          |  | health handler     |  |
+|        +------------------------------->                    |  |
+|        |                            |  +--------------------+  |
+|        |     health reporter create |                          |
+|        +---------------------------->                          |
++--------+                            +--------------------------+

drivers/net/ethernet/mellanox/mlx5/core/Makefile

@@ -22,7 +22,7 @@ mlx5_core-y :=	main.o cmd.o debugfs.o fw.o eq.o uar.o pagealloc.o \
 #
 mlx5_core-$(CONFIG_MLX5_CORE_EN) += en_main.o en_common.o en_fs.o en_ethtool.o \
 		en_tx.o en_rx.o en_dim.o en_txrx.o en/xdp.o en_stats.o \
-		en_selftest.o en/port.o en/monitor_stats.o
+		en_selftest.o en/port.o en/monitor_stats.o en/reporter_tx.o
 
 #
 # Netdev extra

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

@@ -387,10 +387,7 @@ struct mlx5e_txqsq {
 	struct mlx5e_channel      *channel;
 	int                        txq_ix;
 	u32                        rate_limit;
-	struct mlx5e_txqsq_recover {
-		struct work_struct         recover_work;
-		u64                        last_recover;
-	} recover;
+	struct work_struct         recover_work;
 } ____cacheline_aligned_in_smp;
 
 struct mlx5e_dma_info {
@@ -683,6 +680,13 @@ struct mlx5e_rss_params {
 	u8	hfunc;
 };
 
+struct mlx5e_modify_sq_param {
+	int curr_state;
+	int next_state;
+	int rl_update;
+	int rl_index;
+};
+
 struct mlx5e_priv {
 	/* priv data path fields - start */
 	struct mlx5e_txqsq *txq2sq[MLX5E_MAX_NUM_CHANNELS * MLX5E_MAX_NUM_TC];
@@ -738,6 +742,7 @@ struct mlx5e_priv {
 #ifdef CONFIG_MLX5_EN_TLS
 	struct mlx5e_tls          *tls;
 #endif
+	struct devlink_health_reporter *tx_reporter;
 };
 
 struct mlx5e_profile {
@@ -868,6 +873,11 @@ void mlx5e_set_rq_type(struct mlx5_core_dev *mdev, struct mlx5e_params *params);
 void mlx5e_init_rq_type_params(struct mlx5_core_dev *mdev,
 			       struct mlx5e_params *params);
 
+int mlx5e_modify_sq(struct mlx5_core_dev *mdev, u32 sqn,
+		    struct mlx5e_modify_sq_param *p);
+void mlx5e_activate_txqsq(struct mlx5e_txqsq *sq);
+void mlx5e_tx_disable_queue(struct netdev_queue *txq);
+
 static inline bool mlx5e_tunnel_inner_ft_supported(struct mlx5_core_dev *mdev)
 {
 	return (MLX5_CAP_ETH(mdev, tunnel_stateless_gre) &&

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

+/* SPDX-License-Identifier: GPL-2.0 */
+/* Copyright (c) 2019 Mellanox Technologies. */
+
+#ifndef __MLX5E_EN_REPORTER_H
+#define __MLX5E_EN_REPORTER_H
+
+#include <linux/mlx5/driver.h>
+#include "en.h"
+
+int mlx5e_tx_reporter_create(struct mlx5e_priv *priv);
+void mlx5e_tx_reporter_destroy(struct mlx5e_priv *priv);
+void mlx5e_tx_reporter_err_cqe(struct mlx5e_txqsq *sq);
+int mlx5e_tx_reporter_timeout(struct mlx5e_txqsq *sq);
+
+#endif

drivers/net/ethernet/mellanox/mlx5/core/en/reporter_tx.c

+/* SPDX-License-Identifier: GPL-2.0 */
+/* Copyright (c) 2019 Mellanox Technologies. */
+
+#include <net/devlink.h>
+#include "reporter.h"
+#include "lib/eq.h"
+
+#define MLX5E_TX_REPORTER_PER_SQ_MAX_LEN 256
+
+struct mlx5e_tx_err_ctx {
+	int (*recover)(struct mlx5e_txqsq *sq);
+	struct mlx5e_txqsq *sq;
+};
+
+static int mlx5e_wait_for_sq_flush(struct mlx5e_txqsq *sq)
+{
+	unsigned long exp_time = jiffies + msecs_to_jiffies(2000);
+
+	while (time_before(jiffies, exp_time)) {
+		if (sq->cc == sq->pc)
+			return 0;
+
+		msleep(20);
+	}
+
+	netdev_err(sq->channel->netdev,
+		   "Wait for SQ 0x%x flush timeout (sq cc = 0x%x, sq pc = 0x%x)\n",
+		   sq->sqn, sq->cc, sq->pc);
+
+	return -ETIMEDOUT;
+}
+
+static void mlx5e_reset_txqsq_cc_pc(struct mlx5e_txqsq *sq)
+{
+	WARN_ONCE(sq->cc != sq->pc,
+		  "SQ 0x%x: cc (0x%x) != pc (0x%x)\n",
+		  sq->sqn, sq->cc, sq->pc);
+	sq->cc = 0;
+	sq->dma_fifo_cc = 0;
+	sq->pc = 0;
+}
+
+static int mlx5e_sq_to_ready(struct mlx5e_txqsq *sq, int curr_state)
+{
+	struct mlx5_core_dev *mdev = sq->channel->mdev;
+	struct net_device *dev = sq->channel->netdev;
+	struct mlx5e_modify_sq_param msp = {0};
+	int err;
+
+	msp.curr_state = curr_state;
+	msp.next_state = MLX5_SQC_STATE_RST;
+
+	err = mlx5e_modify_sq(mdev, sq->sqn, &msp);
+	if (err) {
+		netdev_err(dev, "Failed to move sq 0x%x to reset\n", sq->sqn);
+		return err;
+	}
+
+	memset(&msp, 0, sizeof(msp));
+	msp.curr_state = MLX5_SQC_STATE_RST;
+	msp.next_state = MLX5_SQC_STATE_RDY;
+
+	err = mlx5e_modify_sq(mdev, sq->sqn, &msp);
+	if (err) {
+		netdev_err(dev, "Failed to move sq 0x%x to ready\n", sq->sqn);
+		return err;
+	}
+
+	return 0;
+}
+
+static int mlx5e_tx_reporter_err_cqe_recover(struct mlx5e_txqsq *sq)
+{
+	struct mlx5_core_dev *mdev = sq->channel->mdev;
+	struct net_device *dev = sq->channel->netdev;
+	u8 state;
+	int err;
+
+	if (!test_bit(MLX5E_SQ_STATE_RECOVERING, &sq->state))
+		return 0;
+
+	err = mlx5_core_query_sq_state(mdev, sq->sqn, &state);
+	if (err) {
+		netdev_err(dev, "Failed to query SQ 0x%x state. err = %d\n",
+			   sq->sqn, err);
+		return err;
+	}
+
+	if (state != MLX5_SQC_STATE_ERR) {
+		netdev_err(dev, "SQ 0x%x not in ERROR state\n", sq->sqn);
+		return -EINVAL;
+	}
+
+	mlx5e_tx_disable_queue(sq->txq);
+
+	err = mlx5e_wait_for_sq_flush(sq);
+	if (err)
+		return err;
+
+	/* At this point, no new packets will arrive from the stack as TXQ is
+	 * marked with QUEUE_STATE_DRV_XOFF. In addition, NAPI cleared all
+	 * pending WQEs. SQ can safely reset the SQ.
+	 */
+
+	err = mlx5e_sq_to_ready(sq, state);
+	if (err)
+		return err;
+
+	mlx5e_reset_txqsq_cc_pc(sq);
+	sq->stats->recover++;
+	mlx5e_activate_txqsq(sq);
+
+	return 0;
+}
+
+void mlx5e_tx_reporter_err_cqe(struct mlx5e_txqsq *sq)
+{
+	char err_str[MLX5E_TX_REPORTER_PER_SQ_MAX_LEN];
+	struct mlx5e_tx_err_ctx err_ctx = {0};
+
+	err_ctx.sq       = sq;
+	err_ctx.recover  = mlx5e_tx_reporter_err_cqe_recover;
+	sprintf(err_str, "ERR CQE on SQ: 0x%x", sq->sqn);
+
+	devlink_health_report(sq->channel->priv->tx_reporter, err_str,
+			      &err_ctx);
+}
+
+static int mlx5e_tx_reporter_timeout_recover(struct mlx5e_txqsq *sq)
+{
+	struct mlx5_eq_comp *eq = sq->cq.mcq.eq;
+	u32 eqe_count;
+	int ret;
+
+	netdev_err(sq->channel->netdev, "EQ 0x%x: Cons = 0x%x, irqn = 0x%x\n",
+		   eq->core.eqn, eq->core.cons_index, eq->core.irqn);
+
+	eqe_count = mlx5_eq_poll_irq_disabled(eq);
+	ret = eqe_count ? true : false;
+	if (!eqe_count) {
+		clear_bit(MLX5E_SQ_STATE_ENABLED, &sq->state);
+		return ret;
+	}
+
+	netdev_err(sq->channel->netdev, "Recover %d eqes on EQ 0x%x\n",
+		   eqe_count, eq->core.eqn);
+	sq->channel->stats->eq_rearm++;
+	return ret;
+}
+
+int mlx5e_tx_reporter_timeout(struct mlx5e_txqsq *sq)
+{
+	char err_str[MLX5E_TX_REPORTER_PER_SQ_MAX_LEN];
+	struct mlx5e_tx_err_ctx err_ctx;
+
+	err_ctx.sq       = sq;
+	err_ctx.recover  = mlx5e_tx_reporter_timeout_recover;
+	sprintf(err_str,
+		"TX timeout on queue: %d, SQ: 0x%x, CQ: 0x%x, SQ Cons: 0x%x SQ Prod: 0x%x, usecs since last trans: %u\n",
+		sq->channel->ix, sq->sqn, sq->cq.mcq.cqn, sq->cc, sq->pc,
+		jiffies_to_usecs(jiffies - sq->txq->trans_start));
+
+	return devlink_health_report(sq->channel->priv->tx_reporter, err_str,
+				     &err_ctx);
+}
+
+/* state lock cannot be grabbed within this function.
+ * It can cause a dead lock or a read-after-free.
+ */
+int mlx5e_tx_reporter_recover_from_ctx(struct mlx5e_tx_err_ctx *err_ctx)
+{
+	return err_ctx->recover(err_ctx->sq);
+}
+
+static int mlx5e_tx_reporter_recover_all(struct mlx5e_priv *priv)
+{
+	int err;
+
+	rtnl_lock();
+	mutex_lock(&priv->state_lock);
+	mlx5e_close_locked(priv->netdev);
+	err = mlx5e_open_locked(priv->netdev);
+	mutex_unlock(&priv->state_lock);
+	rtnl_unlock();
+
+	return err;
+}
+
+static int mlx5e_tx_reporter_recover(struct devlink_health_reporter *reporter,
+				     void *context)
+{
+	struct mlx5e_priv *priv = devlink_health_reporter_priv(reporter);
+	struct mlx5e_tx_err_ctx *err_ctx = context;
+
+	return err_ctx ? mlx5e_tx_reporter_recover_from_ctx(err_ctx) :
+			 mlx5e_tx_reporter_recover_all(priv);
+}
+
+static int
+mlx5e_tx_reporter_build_diagnose_output(struct devlink_fmsg *fmsg,
+					u32 sqn, u8 state, bool stopped)
+{
+	int err;
+
+	err = devlink_fmsg_obj_nest_start(fmsg);
+	if (err)
+		return err;
+
+	err = devlink_fmsg_u32_pair_put(fmsg, "sqn", sqn);
+	if (err)
+		return err;
+
+	err = devlink_fmsg_u8_pair_put(fmsg, "HW state", state);
+	if (err)
+		return err;
+
+	err = devlink_fmsg_bool_pair_put(fmsg, "stopped", stopped);
+	if (err)
+		return err;
+
+	err = devlink_fmsg_obj_nest_end(fmsg);
+	if (err)
+		return err;
+
+	return 0;
+}
+
+static int mlx5e_tx_reporter_diagnose(struct devlink_health_reporter *reporter,
+				      struct devlink_fmsg *fmsg)
+{
+	struct mlx5e_priv *priv = devlink_health_reporter_priv(reporter);
+	int i, err = 0;
+
+	mutex_lock(&priv->state_lock);
+
+	if (!test_bit(MLX5E_STATE_OPENED, &priv->state))
+		goto unlock;
+
+	err = devlink_fmsg_arr_pair_nest_start(fmsg, "SQs");
+	if (err)
+		goto unlock;
+
+	for (i = 0; i < priv->channels.num * priv->channels.params.num_tc;
+	     i++) {
+		struct mlx5e_txqsq *sq = priv->txq2sq[i];
+		u8 state;
+
+		err = mlx5_core_query_sq_state(priv->mdev, sq->sqn, &state);
+		if (err)
+			break;
+
+		err = mlx5e_tx_reporter_build_diagnose_output(fmsg, sq->sqn,
+							      state,
+							      netif_xmit_stopped(sq->txq));
+		if (err)
+			break;
+	}
+	err = devlink_fmsg_arr_pair_nest_end(fmsg);
+	if (err)
+		goto unlock;
+
+unlock:
+	mutex_unlock(&priv->state_lock);
+	return err;
+}
+
+static const struct devlink_health_reporter_ops mlx5_tx_reporter_ops = {
+		.name = "tx",
+		.recover = mlx5e_tx_reporter_recover,
+		.diagnose = mlx5e_tx_reporter_diagnose,
+};
+
+#define MLX5_REPORTER_TX_GRACEFUL_PERIOD 500
+
+int mlx5e_tx_reporter_create(struct mlx5e_priv *priv)
+{
+	struct mlx5_core_dev *mdev = priv->mdev;
+	struct devlink *devlink = priv_to_devlink(mdev);
+
+	priv->tx_reporter =
+		devlink_health_reporter_create(devlink, &mlx5_tx_reporter_ops,
+					       MLX5_REPORTER_TX_GRACEFUL_PERIOD,
+					       true, priv);
+	if (IS_ERR_OR_NULL(priv->tx_reporter))
+		netdev_warn(priv->netdev,
+			    "Failed to create tx reporter, err = %ld\n",
+			    PTR_ERR(priv->tx_reporter));
+	return PTR_ERR_OR_ZERO(priv->tx_reporter);
+}
+
+void mlx5e_tx_reporter_destroy(struct mlx5e_priv *priv)
+{
+	if (IS_ERR_OR_NULL(priv->tx_reporter))
+		return;
+
+	devlink_health_reporter_destroy(priv->tx_reporter);
+}

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

@@ -51,6 +51,7 @@
 #include "en/xdp.h"
 #include "lib/eq.h"
 #include "en/monitor_stats.h"
+#include "en/reporter.h"
 
 struct mlx5e_rq_param {
 	u32			rqc[MLX5_ST_SZ_DW(rqc)];
@@ -1159,7 +1160,7 @@ static int mlx5e_alloc_txqsq_db(struct mlx5e_txqsq *sq, int numa)
 	return 0;
 }
 
-static void mlx5e_sq_recover(struct work_struct *work);
+static void mlx5e_tx_err_cqe_work(struct work_struct *recover_work);
 static int mlx5e_alloc_txqsq(struct mlx5e_channel *c,
 			     int txq_ix,
 			     struct mlx5e_params *params,
@@ -1181,7 +1182,7 @@ static int mlx5e_alloc_txqsq(struct mlx5e_channel *c,
 	sq->uar_map   = mdev->mlx5e_res.bfreg.map;
 	sq->min_inline_mode = params->tx_min_inline_mode;
 	sq->stats     = &c->priv->channel_stats[c->ix].sq[tc];
-	INIT_WORK(&sq->recover.recover_work, mlx5e_sq_recover);
+	INIT_WORK(&sq->recover_work, mlx5e_tx_err_cqe_work);
 	if (MLX5_IPSEC_DEV(c->priv->mdev))
 		set_bit(MLX5E_SQ_STATE_IPSEC, &sq->state);
 	if (mlx5_accel_is_tls_device(c->priv->mdev))
@@ -1269,15 +1270,8 @@ static int mlx5e_create_sq(struct mlx5_core_dev *mdev,
 	return err;
 }
 
-struct mlx5e_modify_sq_param {
-	int curr_state;
-	int next_state;
-	bool rl_update;
-	int rl_index;
-};
-
-static int mlx5e_modify_sq(struct mlx5_core_dev *mdev, u32 sqn,
-			   struct mlx5e_modify_sq_param *p)
+int mlx5e_modify_sq(struct mlx5_core_dev *mdev, u32 sqn,
+		    struct mlx5e_modify_sq_param *p)
 {
 	void *in;
 	void *sqc;
@@ -1375,17 +1369,7 @@ static int mlx5e_open_txqsq(struct mlx5e_channel *c,
 	return err;
 }
 
-static void mlx5e_reset_txqsq_cc_pc(struct mlx5e_txqsq *sq)
-{
-	WARN_ONCE(sq->cc != sq->pc,
-		  "SQ 0x%x: cc (0x%x) != pc (0x%x)\n",
-		  sq->sqn, sq->cc, sq->pc);
-	sq->cc = 0;
-	sq->dma_fifo_cc = 0;
-	sq->pc = 0;
-}
-
-static void mlx5e_activate_txqsq(struct mlx5e_txqsq *sq)
+void mlx5e_activate_txqsq(struct mlx5e_txqsq *sq)
 {
 	sq->txq = netdev_get_tx_queue(sq->channel->netdev, sq->txq_ix);
 	clear_bit(MLX5E_SQ_STATE_RECOVERING, &sq->state);
@@ -1394,7 +1378,7 @@ static void mlx5e_activate_txqsq(struct mlx5e_txqsq *sq)
 	netif_tx_start_queue(sq->txq);
 }
 
-static inline void netif_tx_disable_queue(struct netdev_queue *txq)
+void mlx5e_tx_disable_queue(struct netdev_queue *txq)
 {
 	__netif_tx_lock_bh(txq);
 	netif_tx_stop_queue(txq);
@@ -1410,7 +1394,7 @@ static void mlx5e_deactivate_txqsq(struct mlx5e_txqsq *sq)
 	/* prevent netif_tx_wake_queue */
 	napi_synchronize(&c->napi);
 
-	netif_tx_disable_queue(sq->txq);
+	mlx5e_tx_disable_queue(sq->txq);
 
 	/* last doorbell out, godspeed .. */
 	if (mlx5e_wqc_has_room_for(wq, sq->cc, sq->pc, 1)) {
@@ -1430,6 +1414,7 @@ static void mlx5e_close_txqsq(struct mlx5e_txqsq *sq)
 	struct mlx5_rate_limit rl = {0};
 
 	cancel_work_sync(&sq->dim.work);
+	cancel_work_sync(&sq->recover_work);
 	mlx5e_destroy_sq(mdev, sq->sqn);
 	if (sq->rate_limit) {
 		rl.rate = sq->rate_limit;
@@ -1439,105 +1424,12 @@ static void mlx5e_close_txqsq(struct mlx5e_txqsq *sq)
 	mlx5e_free_txqsq(sq);
 }
 
-static int mlx5e_wait_for_sq_flush(struct mlx5e_txqsq *sq)
-{
-	unsigned long exp_time = jiffies + msecs_to_jiffies(2000);
-
-	while (time_before(jiffies, exp_time)) {
-		if (sq->cc == sq->pc)
-			return 0;
-
-		msleep(20);
-	}
-
-	netdev_err(sq->channel->netdev,
-		   "Wait for SQ 0x%x flush timeout (sq cc = 0x%x, sq pc = 0x%x)\n",
-		   sq->sqn, sq->cc, sq->pc);
-
-	return -ETIMEDOUT;
-}
-
-static int mlx5e_sq_to_ready(struct mlx5e_txqsq *sq, int curr_state)
-{
-	struct mlx5_core_dev *mdev = sq->channel->mdev;
-	struct net_device *dev = sq->channel->netdev;
-	struct mlx5e_modify_sq_param msp = {0};
-	int err;
-
-	msp.curr_state = curr_state;
-	msp.next_state = MLX5_SQC_STATE_RST;
-
-	err = mlx5e_modify_sq(mdev, sq->sqn, &msp);
-	if (err) {
-		netdev_err(dev, "Failed to move sq 0x%x to reset\n", sq->sqn);
-		return err;
-	}
-
-	memset(&msp, 0, sizeof(msp));
-	msp.curr_state = MLX5_SQC_STATE_RST;
-	msp.next_state = MLX5_SQC_STATE_RDY;
-
-	err = mlx5e_modify_sq(mdev, sq->sqn, &msp);
-	if (err) {
-		netdev_err(dev, "Failed to move sq 0x%x to ready\n", sq->sqn);
-		return err;
-	}
-
-	return 0;
-}
-
-static void mlx5e_sq_recover(struct work_struct *work)
+static void mlx5e_tx_err_cqe_work(struct work_struct *recover_work)
 {
-	struct mlx5e_txqsq_recover *recover =
-		container_of(work, struct mlx5e_txqsq_recover,
-			     recover_work);
-	struct mlx5e_txqsq *sq = container_of(recover, struct mlx5e_txqsq,
-					      recover);
-	struct mlx5_core_dev *mdev = sq->channel->mdev;
-	struct net_device *dev = sq->channel->netdev;
-	u8 state;
-	int err;
-
-	err = mlx5_core_query_sq_state(mdev, sq->sqn, &state);
-	if (err) {
-		netdev_err(dev, "Failed to query SQ 0x%x state. err = %d\n",
-			   sq->sqn, err);
-		return;
-	}
-
-	if (state != MLX5_RQC_STATE_ERR) {
-		netdev_err(dev, "SQ 0x%x not in ERROR state\n", sq->sqn);
-		return;
-	}
-
-	netif_tx_disable_queue(sq->txq);
-
-	if (mlx5e_wait_for_sq_flush(sq))
-		return;
-
-	/* If the interval between two consecutive recovers per SQ is too
-	 * short, don't recover to avoid infinite loop of ERR_CQE -> recover.
-	 * If we reached this state, there is probably a bug that needs to be
-	 * fixed. let's keep the queue close and let tx timeout cleanup.
-	 */
-	if (jiffies_to_msecs(jiffies - recover->last_recover) <
-	    MLX5E_SQ_RECOVER_MIN_INTERVAL) {
-		netdev_err(dev, "Recover SQ 0x%x canceled, too many error CQEs\n",
-			   sq->sqn);
-		return;
-	}
-
-	/* At this point, no new packets will arrive from the stack as TXQ is
-	 * marked with QUEUE_STATE_DRV_XOFF. In addition, NAPI cleared all
-	 * pending WQEs.  SQ can safely reset the SQ.
-	 */
-	if (mlx5e_sq_to_ready(sq, state))
-		return;
+	struct mlx5e_txqsq *sq = container_of(recover_work, struct mlx5e_txqsq,
+					      recover_work);
 
-	mlx5e_reset_txqsq_cc_pc(sq);
-	sq->stats->recover++;
-	recover->last_recover = jiffies;
-	mlx5e_activate_txqsq(sq);
+	mlx5e_tx_reporter_err_cqe(sq);
 }
 
 static int mlx5e_open_icosq(struct mlx5e_channel *c,
@@ -3236,6 +3128,7 @@ static void mlx5e_cleanup_nic_tx(struct mlx5e_priv *priv)
 {
 	int tc;
 
+	mlx5e_tx_reporter_destroy(priv);
 	for (tc = 0; tc < priv->profile->max_tc; tc++)
 		mlx5e_destroy_tis(priv->mdev, priv->tisn[tc]);
 }
@@ -4223,31 +4116,13 @@ netdev_features_t mlx5e_features_check(struct sk_buff *skb,
 	return features;
 }
 
-static bool mlx5e_tx_timeout_eq_recover(struct net_device *dev,
-					struct mlx5e_txqsq *sq)
-{
-	struct mlx5_eq_comp *eq = sq->cq.mcq.eq;
-	u32 eqe_count;
-
-	netdev_err(dev, "EQ 0x%x: Cons = 0x%x, irqn = 0x%x\n",
-		   eq->core.eqn, eq->core.cons_index, eq->core.irqn);
-
-	eqe_count = mlx5_eq_poll_irq_disabled(eq);
-	if (!eqe_count)
-		return false;
-
-	netdev_err(dev, "Recover %d eqes on EQ 0x%x\n", eqe_count, eq->core.eqn);
-	sq->channel->stats->eq_rearm++;
-	return true;
-}
-
 static void mlx5e_tx_timeout_work(struct work_struct *work)
 {
 	struct mlx5e_priv *priv = container_of(work, struct mlx5e_priv,
 					       tx_timeout_work);
-	struct net_device *dev = priv->netdev;
-	bool reopen_channels = false;
-	int i, err;
+	bool report_failed = false;
+	int err;
+	int i;
 
 	rtnl_lock();
 	mutex_lock(&priv->state_lock);
@@ -4256,31 +4131,22 @@ static void mlx5e_tx_timeout_work(struct work_struct *work)
 		goto unlock;
 
 	for (i = 0; i < priv->channels.num * priv->channels.params.num_tc; i++) {
-		struct netdev_queue *dev_queue = netdev_get_tx_queue(dev, i);
+		struct netdev_queue *dev_queue =
+			netdev_get_tx_queue(priv->netdev, i);
 		struct mlx5e_txqsq *sq = priv->txq2sq[i];
 
 		if (!netif_xmit_stopped(dev_queue))
 			continue;
 
-		netdev_err(dev,
-			   "TX timeout on queue: %d, SQ: 0x%x, CQ: 0x%x, SQ Cons: 0x%x SQ Prod: 0x%x, usecs since last trans: %u\n",
-			   i, sq->sqn, sq->cq.mcq.cqn, sq->cc, sq->pc,
-			   jiffies_to_usecs(jiffies - dev_queue->trans_start));
-
-		/* If we recover a lost interrupt, most likely TX timeout will
-		 * be resolved, skip reopening channels
-		 */
-		if (!mlx5e_tx_timeout_eq_recover(dev, sq)) {
-			clear_bit(MLX5E_SQ_STATE_ENABLED, &sq->state);
-			reopen_channels = true;
-		}
+		if (mlx5e_tx_reporter_timeout(sq))
+			report_failed = true;
 	}
 
-	if (!reopen_channels)
+	if (!report_failed)
 		goto unlock;
 
-	mlx5e_close_locked(dev);
-	err = mlx5e_open_locked(dev);
+	mlx5e_close_locked(priv->netdev);
+	err = mlx5e_open_locked(priv->netdev);
 	if (err)
 		netdev_err(priv->netdev,
 			   "mlx5e_open_locked failed recovering from a tx_timeout, err(%d).\n",
@@ -4296,6 +4162,12 @@ static void mlx5e_tx_timeout(struct net_device *dev)
 	struct mlx5e_priv *priv = netdev_priv(dev);
 
 	netdev_err(dev, "TX timeout detected\n");
+
+	if (IS_ERR_OR_NULL(priv->tx_reporter)) {
+		netdev_err_once(priv->netdev, "tx timeout will not be handled, no valid tx reporter\n");
+		return;
+	}
+
 	queue_work(priv->wq, &priv->tx_timeout_work);
 }
 
@@ -4953,6 +4825,7 @@ static int mlx5e_init_nic_tx(struct mlx5e_priv *priv)
 #ifdef CONFIG_MLX5_CORE_EN_DCB
 	mlx5e_dcbnl_initialize(priv);
 #endif
+	mlx5e_tx_reporter_create(priv);
 	return 0;
 }
 

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

@@ -513,8 +513,9 @@ bool mlx5e_poll_tx_cq(struct mlx5e_cq *cq, int napi_budget)
 					      &sq->state)) {
 				mlx5e_dump_error_cqe(sq,
 						     (struct mlx5_err_cqe *)cqe);
-				queue_work(cq->channel->priv->wq,
-					   &sq->recover.recover_work);
+				if (!IS_ERR_OR_NULL(cq->channel->priv->tx_reporter))
+					queue_work(cq->channel->priv->wq,
+						   &sq->recover_work);
 			}
 			stats->cqe_err++;
 		}

include/net/devlink.h

@@ -30,6 +30,7 @@ struct devlink {
 	struct list_head param_list;
 	struct list_head region_list;
 	u32 snapshot_id;
+	struct list_head reporter_list;
 	struct devlink_dpipe_headers *dpipe_headers;
 	const struct devlink_ops *ops;
 	struct device *dev;
@@ -448,6 +449,29 @@ struct devlink_info_req;
 
 typedef void devlink_snapshot_data_dest_t(const void *data);
 
+struct devlink_fmsg;
+struct devlink_health_reporter;
+
+/**
+ * struct devlink_health_reporter_ops - Reporter operations
+ * @name: reporter name
+ * @recover: callback to recover from reported error
+ *           if priv_ctx is NULL, run a full recover
+ * @dump: callback to dump an object
+ *        if priv_ctx is NULL, run a full dump
+ * @diagnose: callback to diagnose the current status
+ */
+
+struct devlink_health_reporter_ops {
+	char *name;
+	int (*recover)(struct devlink_health_reporter *reporter,
+		       void *priv_ctx);
+	int (*dump)(struct devlink_health_reporter *reporter,
+		    struct devlink_fmsg *fmsg, void *priv_ctx);
+	int (*diagnose)(struct devlink_health_reporter *reporter,
+			struct devlink_fmsg *fmsg);
+};
+
 struct devlink_ops {
 	int (*reload)(struct devlink *devlink, struct netlink_ext_ack *extack);
 	int (*port_type_set)(struct devlink_port *devlink_port,
@@ -639,6 +663,50 @@ int devlink_info_version_running_put(struct devlink_info_req *req,
 				     const char *version_name,
 				     const char *version_value);
 
+int devlink_fmsg_obj_nest_start(struct devlink_fmsg *fmsg);
+int devlink_fmsg_obj_nest_end(struct devlink_fmsg *fmsg);
+
+int devlink_fmsg_pair_nest_start(struct devlink_fmsg *fmsg, const char *name);
+int devlink_fmsg_pair_nest_end(struct devlink_fmsg *fmsg);
+
+int devlink_fmsg_arr_pair_nest_start(struct devlink_fmsg *fmsg,
+				     const char *name);
+int devlink_fmsg_arr_pair_nest_end(struct devlink_fmsg *fmsg);
+
+int devlink_fmsg_bool_put(struct devlink_fmsg *fmsg, bool value);
+int devlink_fmsg_u8_put(struct devlink_fmsg *fmsg, u8 value);
+int devlink_fmsg_u32_put(struct devlink_fmsg *fmsg, u32 value);
+int devlink_fmsg_u64_put(struct devlink_fmsg *fmsg, u64 value);
+int devlink_fmsg_string_put(struct devlink_fmsg *fmsg, const char *value);
+int devlink_fmsg_binary_put(struct devlink_fmsg *fmsg, const void *value,
+			    u16 value_len);
+
+int devlink_fmsg_bool_pair_put(struct devlink_fmsg *fmsg, const char *name,
+			       bool value);
+int devlink_fmsg_u8_pair_put(struct devlink_fmsg *fmsg, const char *name,
+			     u8 value);
+int devlink_fmsg_u32_pair_put(struct devlink_fmsg *fmsg, const char *name,
+			      u32 value);
+int devlink_fmsg_u64_pair_put(struct devlink_fmsg *fmsg, const char *name,
+			      u64 value);
+int devlink_fmsg_string_pair_put(struct devlink_fmsg *fmsg, const char *name,
+				 const char *value);
+int devlink_fmsg_binary_pair_put(struct devlink_fmsg *fmsg, const char *name,
+				 const void *value, u16 value_len);
+
+struct devlink_health_reporter *
+devlink_health_reporter_create(struct devlink *devlink,
+			       const struct devlink_health_reporter_ops *ops,
+			       u64 graceful_period, bool auto_recover,
+			       void *priv);
+void
+devlink_health_reporter_destroy(struct devlink_health_reporter *reporter);
+
+void *
+devlink_health_reporter_priv(struct devlink_health_reporter *reporter);
+int devlink_health_report(struct devlink_health_reporter *reporter,
+			  const char *msg, void *priv_ctx);
+
 #else
 
 static inline struct devlink *devlink_alloc(const struct devlink_ops *ops,
@@ -971,6 +1039,149 @@ devlink_info_version_running_put(struct devlink_info_req *req,
 {
 	return 0;
 }
+
+static inline int
+devlink_fmsg_obj_nest_start(struct devlink_fmsg *fmsg)
+{
+	return 0;
+}
+
+static inline int
+devlink_fmsg_obj_nest_end(struct devlink_fmsg *fmsg)
+{
+	return 0;
+}
+
+static inline int
+devlink_fmsg_pair_nest_start(struct devlink_fmsg *fmsg, const char *name)
+{
+	return 0;
+}
+
+static inline int
+devlink_fmsg_pair_nest_end(struct devlink_fmsg *fmsg)
+{
+	return 0;
+}
+
+static inline int
+devlink_fmsg_arr_pair_nest_start(struct devlink_fmsg *fmsg,
+				 const char *name)
+{
+	return 0;
+}
+
+static inline int
+devlink_fmsg_arr_pair_nest_end(struct devlink_fmsg *fmsg)
+{
+	return 0;
+}
+
+static inline int
+devlink_fmsg_bool_put(struct devlink_fmsg *fmsg, bool value)
+{
+	return 0;
+}
+
+static inline int
+devlink_fmsg_u8_put(struct devlink_fmsg *fmsg, u8 value)
+{
+	return 0;
+}
+
+static inline int
+devlink_fmsg_u32_put(struct devlink_fmsg *fmsg, u32 value)
+{
+	return 0;
+}
+
+static inline int
+devlink_fmsg_u64_put(struct devlink_fmsg *fmsg, u64 value)
+{
+	return 0;
+}
+
+static inline int
+devlink_fmsg_string_put(struct devlink_fmsg *fmsg, const char *value)
+{
+	return 0;
+}
+
+static inline int
+devlink_fmsg_binary_put(struct devlink_fmsg *fmsg, const void *value,
+			u16 value_len)
+{
+	return 0;
+}
+
+static inline int
+devlink_fmsg_bool_pair_put(struct devlink_fmsg *fmsg, const char *name,
+			   bool value)
+{
+	return 0;
+}
+
+static inline int
+devlink_fmsg_u8_pair_put(struct devlink_fmsg *fmsg, const char *name,
+			 u8 value)
+{
+	return 0;
+}
+
+static inline int
+devlink_fmsg_u32_pair_put(struct devlink_fmsg *fmsg, const char *name,
+			  u32 value)
+{
+	return 0;
+}
+
+static inline int
+devlink_fmsg_u64_pair_put(struct devlink_fmsg *fmsg, const char *name,
+			  u64 value)
+{
+	return 0;
+}
+
+static inline int
+devlink_fmsg_string_pair_put(struct devlink_fmsg *fmsg, const char *name,
+			     const char *value)
+{
+	return 0;
+}
+
+static inline int
+devlink_fmsg_binary_pair_put(struct devlink_fmsg *fmsg, const char *name,
+			     const void *value, u16 value_len)
+{
+	return 0;
+}
+
+static inline struct devlink_health_reporter *
+devlink_health_reporter_create(struct devlink *devlink,
+			       const struct devlink_health_reporter_ops *ops,
+			       u64 graceful_period, bool auto_recover,
+			       void *priv)
+{
+	return NULL;
+}
+
+static inline void
+devlink_health_reporter_destroy(struct devlink_health_reporter *reporter)
+{
+}
+
+static inline void *
+devlink_health_reporter_priv(struct devlink_health_reporter *reporter)
+{
+	return NULL;
+}
+
+static inline int
+devlink_health_report(struct devlink_health_reporter *reporter,
+		      const char *msg, void *priv_ctx)
+{
+	return 0;
+}
 #endif
 
 #if IS_REACHABLE(CONFIG_NET_DEVLINK)

include/trace/events/devlink.h

@@ -75,6 +75,71 @@ TRACE_EVENT(devlink_hwerr,
 			__get_str(driver_name), __entry->err, __get_str(msg))
 );
 
+/*
+ * Tracepoint for devlink health message:
+ */
+TRACE_EVENT(devlink_health_report,
+	TP_PROTO(const struct devlink *devlink, const char *reporter_name,
+		 const char *msg),
+
+	TP_ARGS(devlink, reporter_name, msg),
+
+	TP_STRUCT__entry(
+		__string(bus_name, devlink->dev->bus->name)
+		__string(dev_name, dev_name(devlink->dev))
+		__string(driver_name, devlink->dev->driver->name)
+		__string(reporter_name, msg)
+		__string(msg, msg)
+	),
+
+	TP_fast_assign(
+		__assign_str(bus_name, devlink->dev->bus->name);
+		__assign_str(dev_name, dev_name(devlink->dev));
+		__assign_str(driver_name, devlink->dev->driver->name);
+		__assign_str(reporter_name, reporter_name);
+		__assign_str(msg, msg);
+	),
+
+	TP_printk("bus_name=%s dev_name=%s driver_name=%s reporter_name=%s: %s",
+		  __get_str(bus_name), __get_str(dev_name),
+		  __get_str(driver_name), __get_str(reporter_name),
+		  __get_str(msg))
+);
+
+/*
+ * Tracepoint for devlink health recover aborted message:
+ */
+TRACE_EVENT(devlink_health_recover_aborted,
+	TP_PROTO(const struct devlink *devlink, const char *reporter_name,
+		 bool health_state, u64 time_since_last_recover),
+
+	TP_ARGS(devlink, reporter_name, health_state, time_since_last_recover),
+
+	TP_STRUCT__entry(
+		__string(bus_name, devlink->dev->bus->name)
+		__string(dev_name, dev_name(devlink->dev))
+		__string(driver_name, devlink->dev->driver->name)
+		__string(reporter_name, reporter_name)
+		__field(bool, health_state)
+		__field(u64, time_since_last_recover)
+	),
+
+	TP_fast_assign(
+		__assign_str(bus_name, devlink->dev->bus->name);
+		__assign_str(dev_name, dev_name(devlink->dev));
+		__assign_str(driver_name, devlink->dev->driver->name);
+		__assign_str(reporter_name, reporter_name);
+		__entry->health_state = health_state;
+		__entry->time_since_last_recover = time_since_last_recover;
+	),
+
+	TP_printk("bus_name=%s dev_name=%s driver_name=%s reporter_name=%s: health_state=%d time_since_last_recover=%llu recover aborted",
+		  __get_str(bus_name), __get_str(dev_name),
+		  __get_str(driver_name), __get_str(reporter_name),
+		  __entry->health_state,
+		  __entry->time_since_last_recover)
+);
+
 #endif /* _TRACE_DEVLINK_H */
 
 /* This part must be outside protection */

include/uapi/linux/devlink.h

@@ -96,6 +96,13 @@ enum devlink_command {
 
 	DEVLINK_CMD_INFO_GET,		/* can dump */
 
+	DEVLINK_CMD_HEALTH_REPORTER_GET,
+	DEVLINK_CMD_HEALTH_REPORTER_SET,
+	DEVLINK_CMD_HEALTH_REPORTER_RECOVER,
+	DEVLINK_CMD_HEALTH_REPORTER_DIAGNOSE,
+	DEVLINK_CMD_HEALTH_REPORTER_DUMP_GET,
+	DEVLINK_CMD_HEALTH_REPORTER_DUMP_CLEAR,
+
 	/* add new commands above here */
 	__DEVLINK_CMD_MAX,
 	DEVLINK_CMD_MAX = __DEVLINK_CMD_MAX - 1
@@ -302,6 +309,23 @@ enum devlink_attr {
 
 	DEVLINK_ATTR_SB_POOL_CELL_SIZE,		/* u32 */
 
+	DEVLINK_ATTR_FMSG,			/* nested */
+	DEVLINK_ATTR_FMSG_OBJ_NEST_START,	/* flag */
+	DEVLINK_ATTR_FMSG_PAIR_NEST_START,	/* flag */
+	DEVLINK_ATTR_FMSG_ARR_NEST_START,	/* flag */
+	DEVLINK_ATTR_FMSG_NEST_END,		/* flag */
+	DEVLINK_ATTR_FMSG_OBJ_NAME,		/* string */
+	DEVLINK_ATTR_FMSG_OBJ_VALUE_TYPE,	/* u8 */
+	DEVLINK_ATTR_FMSG_OBJ_VALUE_DATA,	/* dynamic */
+
+	DEVLINK_ATTR_HEALTH_REPORTER,			/* nested */
+	DEVLINK_ATTR_HEALTH_REPORTER_NAME,		/* string */
+	DEVLINK_ATTR_HEALTH_REPORTER_STATE,		/* u8 */
+	DEVLINK_ATTR_HEALTH_REPORTER_ERR,		/* u64 */
+	DEVLINK_ATTR_HEALTH_REPORTER_RECOVER,		/* u64 */
+	DEVLINK_ATTR_HEALTH_REPORTER_DUMP_TS,		/* u64 */
+	DEVLINK_ATTR_HEALTH_REPORTER_GRACEFUL_PERIOD,	/* u64 */
+	DEVLINK_ATTR_HEALTH_REPORTER_AUTO_RECOVER,	/* u8 */
 	/* add new attributes above here, update the policy in devlink.c */
 
 	__DEVLINK_ATTR_MAX,

net/core/devlink.c

@@ -3879,6 +3879,965 @@ static int devlink_nl_cmd_info_get_dumpit(struct sk_buff *msg,
 	return msg->len;
 }
 
+struct devlink_fmsg_item {
+	struct list_head list;
+	int attrtype;
+	u8 nla_type;
+	u16 len;
+	int value[0];
+};
+
+struct devlink_fmsg {
+	struct list_head item_list;
+};
+
+static struct devlink_fmsg *devlink_fmsg_alloc(void)
+{
+	struct devlink_fmsg *fmsg;
+
+	fmsg = kzalloc(sizeof(*fmsg), GFP_KERNEL);
+	if (!fmsg)
+		return NULL;
+
+	INIT_LIST_HEAD(&fmsg->item_list);
+
+	return fmsg;
+}
+
+static void devlink_fmsg_free(struct devlink_fmsg *fmsg)
+{
+	struct devlink_fmsg_item *item, *tmp;
+
+	list_for_each_entry_safe(item, tmp, &fmsg->item_list, list) {
+		list_del(&item->list);
+		kfree(item);
+	}
+	kfree(fmsg);
+}
+
+static int devlink_fmsg_nest_common(struct devlink_fmsg *fmsg,
+				    int attrtype)
+{
+	struct devlink_fmsg_item *item;
+
+	item = kzalloc(sizeof(*item), GFP_KERNEL);
+	if (!item)
+		return -ENOMEM;
+
+	item->attrtype = attrtype;
+	list_add_tail(&item->list, &fmsg->item_list);
+
+	return 0;
+}
+
+int devlink_fmsg_obj_nest_start(struct devlink_fmsg *fmsg)
+{
+	return devlink_fmsg_nest_common(fmsg, DEVLINK_ATTR_FMSG_OBJ_NEST_START);
+}
+EXPORT_SYMBOL_GPL(devlink_fmsg_obj_nest_start);
+
+static int devlink_fmsg_nest_end(struct devlink_fmsg *fmsg)
+{
+	return devlink_fmsg_nest_common(fmsg, DEVLINK_ATTR_FMSG_NEST_END);
+}
+
+int devlink_fmsg_obj_nest_end(struct devlink_fmsg *fmsg)
+{
+	return devlink_fmsg_nest_end(fmsg);
+}
+EXPORT_SYMBOL_GPL(devlink_fmsg_obj_nest_end);
+
+#define DEVLINK_FMSG_MAX_SIZE (GENLMSG_DEFAULT_SIZE - GENL_HDRLEN - NLA_HDRLEN)
+
+static int devlink_fmsg_put_name(struct devlink_fmsg *fmsg, const char *name)
+{
+	struct devlink_fmsg_item *item;
+
+	if (strlen(name) + 1 > DEVLINK_FMSG_MAX_SIZE)
+		return -EMSGSIZE;
+
+	item = kzalloc(sizeof(*item) + strlen(name) + 1, GFP_KERNEL);
+	if (!item)
+		return -ENOMEM;
+
+	item->nla_type = NLA_NUL_STRING;
+	item->len = strlen(name) + 1;
+	item->attrtype = DEVLINK_ATTR_FMSG_OBJ_NAME;
+	memcpy(&item->value, name, item->len);
+	list_add_tail(&item->list, &fmsg->item_list);
+
+	return 0;
+}
+
+int devlink_fmsg_pair_nest_start(struct devlink_fmsg *fmsg, const char *name)
+{
+	int err;
+
+	err = devlink_fmsg_nest_common(fmsg, DEVLINK_ATTR_FMSG_PAIR_NEST_START);
+	if (err)
+		return err;
+
+	err = devlink_fmsg_put_name(fmsg, name);
+	if (err)
+		return err;
+
+	return 0;
+}
+EXPORT_SYMBOL_GPL(devlink_fmsg_pair_nest_start);
+
+int devlink_fmsg_pair_nest_end(struct devlink_fmsg *fmsg)
+{
+	return devlink_fmsg_nest_end(fmsg);
+}
+EXPORT_SYMBOL_GPL(devlink_fmsg_pair_nest_end);
+
+int devlink_fmsg_arr_pair_nest_start(struct devlink_fmsg *fmsg,
+				     const char *name)
+{
+	int err;
+
+	err = devlink_fmsg_pair_nest_start(fmsg, name);
+	if (err)
+		return err;
+
+	err = devlink_fmsg_nest_common(fmsg, DEVLINK_ATTR_FMSG_ARR_NEST_START);
+	if (err)
+		return err;
+
+	return 0;
+}
+EXPORT_SYMBOL_GPL(devlink_fmsg_arr_pair_nest_start);
+
+int devlink_fmsg_arr_pair_nest_end(struct devlink_fmsg *fmsg)
+{
+	int err;
+
+	err = devlink_fmsg_nest_end(fmsg);
+	if (err)
+		return err;
+
+	err = devlink_fmsg_nest_end(fmsg);
+	if (err)
+		return err;
+
+	return 0;
+}
+EXPORT_SYMBOL_GPL(devlink_fmsg_arr_pair_nest_end);
+
+static int devlink_fmsg_put_value(struct devlink_fmsg *fmsg,
+				  const void *value, u16 value_len,
+				  u8 value_nla_type)
+{
+	struct devlink_fmsg_item *item;
+
+	if (value_len > DEVLINK_FMSG_MAX_SIZE)
+		return -EMSGSIZE;
+
+	item = kzalloc(sizeof(*item) + value_len, GFP_KERNEL);
+	if (!item)
+		return -ENOMEM;
+
+	item->nla_type = value_nla_type;
+	item->len = value_len;
+	item->attrtype = DEVLINK_ATTR_FMSG_OBJ_VALUE_DATA;
+	memcpy(&item->value, value, item->len);
+	list_add_tail(&item->list, &fmsg->item_list);
+
+	return 0;
+}
+
+int devlink_fmsg_bool_put(struct devlink_fmsg *fmsg, bool value)
+{
+	return devlink_fmsg_put_value(fmsg, &value, sizeof(value), NLA_FLAG);
+}
+EXPORT_SYMBOL_GPL(devlink_fmsg_bool_put);
+
+int devlink_fmsg_u8_put(struct devlink_fmsg *fmsg, u8 value)
+{
+	return devlink_fmsg_put_value(fmsg, &value, sizeof(value), NLA_U8);
+}
+EXPORT_SYMBOL_GPL(devlink_fmsg_u8_put);
+
+int devlink_fmsg_u32_put(struct devlink_fmsg *fmsg, u32 value)
+{
+	return devlink_fmsg_put_value(fmsg, &value, sizeof(value), NLA_U32);
+}
+EXPORT_SYMBOL_GPL(devlink_fmsg_u32_put);
+
+int devlink_fmsg_u64_put(struct devlink_fmsg *fmsg, u64 value)
+{
+	return devlink_fmsg_put_value(fmsg, &value, sizeof(value), NLA_U64);
+}
+EXPORT_SYMBOL_GPL(devlink_fmsg_u64_put);
+
+int devlink_fmsg_string_put(struct devlink_fmsg *fmsg, const char *value)
+{
+	return devlink_fmsg_put_value(fmsg, value, strlen(value) + 1,
+				      NLA_NUL_STRING);
+}
+EXPORT_SYMBOL_GPL(devlink_fmsg_string_put);
+
+int devlink_fmsg_binary_put(struct devlink_fmsg *fmsg, const void *value,
+			    u16 value_len)
+{
+	return devlink_fmsg_put_value(fmsg, value, value_len, NLA_BINARY);
+}
+EXPORT_SYMBOL_GPL(devlink_fmsg_binary_put);
+
+int devlink_fmsg_bool_pair_put(struct devlink_fmsg *fmsg, const char *name,
+			       bool value)
+{
+	int err;
+
+	err = devlink_fmsg_pair_nest_start(fmsg, name);
+	if (err)
+		return err;
+
+	err = devlink_fmsg_bool_put(fmsg, value);
+	if (err)
+		return err;
+
+	err = devlink_fmsg_pair_nest_end(fmsg);
+	if (err)
+		return err;
+
+	return 0;
+}
+EXPORT_SYMBOL_GPL(devlink_fmsg_bool_pair_put);
+
+int devlink_fmsg_u8_pair_put(struct devlink_fmsg *fmsg, const char *name,
+			     u8 value)
+{
+	int err;
+
+	err = devlink_fmsg_pair_nest_start(fmsg, name);
+	if (err)
+		return err;
+
+	err = devlink_fmsg_u8_put(fmsg, value);
+	if (err)
+		return err;
+
+	err = devlink_fmsg_pair_nest_end(fmsg);
+	if (err)
+		return err;
+
+	return 0;
+}
+EXPORT_SYMBOL_GPL(devlink_fmsg_u8_pair_put);
+
+int devlink_fmsg_u32_pair_put(struct devlink_fmsg *fmsg, const char *name,
+			      u32 value)
+{
+	int err;
+
+	err = devlink_fmsg_pair_nest_start(fmsg, name);
+	if (err)
+		return err;
+
+	err = devlink_fmsg_u32_put(fmsg, value);
+	if (err)
+		return err;
+
+	err = devlink_fmsg_pair_nest_end(fmsg);
+	if (err)
+		return err;
+
+	return 0;
+}
+EXPORT_SYMBOL_GPL(devlink_fmsg_u32_pair_put);
+
+int devlink_fmsg_u64_pair_put(struct devlink_fmsg *fmsg, const char *name,
+			      u64 value)
+{
+	int err;
+
+	err = devlink_fmsg_pair_nest_start(fmsg, name);
+	if (err)
+		return err;
+
+	err = devlink_fmsg_u64_put(fmsg, value);
+	if (err)
+		return err;
+
+	err = devlink_fmsg_pair_nest_end(fmsg);
+	if (err)
+		return err;
+
+	return 0;
+}
+EXPORT_SYMBOL_GPL(devlink_fmsg_u64_pair_put);
+
+int devlink_fmsg_string_pair_put(struct devlink_fmsg *fmsg, const char *name,
+				 const char *value)
+{
+	int err;
+
+	err = devlink_fmsg_pair_nest_start(fmsg, name);
+	if (err)
+		return err;
+
+	err = devlink_fmsg_string_put(fmsg, value);
+	if (err)
+		return err;
+
+	err = devlink_fmsg_pair_nest_end(fmsg);
+	if (err)
+		return err;
+
+	return 0;
+}
+EXPORT_SYMBOL_GPL(devlink_fmsg_string_pair_put);
+
+int devlink_fmsg_binary_pair_put(struct devlink_fmsg *fmsg, const char *name,
+				 const void *value, u16 value_len)
+{
+	int err;
+
+	err = devlink_fmsg_pair_nest_start(fmsg, name);
+	if (err)
+		return err;
+
+	err = devlink_fmsg_binary_put(fmsg, value, value_len);
+	if (err)
+		return err;
+
+	err = devlink_fmsg_pair_nest_end(fmsg);
+	if (err)
+		return err;
+
+	return 0;
+}
+EXPORT_SYMBOL_GPL(devlink_fmsg_binary_pair_put);
+
+static int
+devlink_fmsg_item_fill_type(struct devlink_fmsg_item *msg, struct sk_buff *skb)
+{
+	switch (msg->nla_type) {
+	case NLA_FLAG:
+	case NLA_U8:
+	case NLA_U32:
+	case NLA_U64:
+	case NLA_NUL_STRING:
+	case NLA_BINARY:
+		return nla_put_u8(skb, DEVLINK_ATTR_FMSG_OBJ_VALUE_TYPE,
+				  msg->nla_type);
+	default:
+		return -EINVAL;
+	}
+}
+
+static int
+devlink_fmsg_item_fill_data(struct devlink_fmsg_item *msg, struct sk_buff *skb)
+{
+	int attrtype = DEVLINK_ATTR_FMSG_OBJ_VALUE_DATA;
+	u8 tmp;
+
+	switch (msg->nla_type) {
+	case NLA_FLAG:
+		/* Always provide flag data, regardless of its value */
+		tmp = *(bool *) msg->value;
+
+		return nla_put_u8(skb, attrtype, tmp);
+	case NLA_U8:
+		return nla_put_u8(skb, attrtype, *(u8 *) msg->value);
+	case NLA_U32:
+		return nla_put_u32(skb, attrtype, *(u32 *) msg->value);
+	case NLA_U64:
+		return nla_put_u64_64bit(skb, attrtype, *(u64 *) msg->value,
+					 DEVLINK_ATTR_PAD);
+	case NLA_NUL_STRING:
+		return nla_put_string(skb, attrtype, (char *) &msg->value);
+	case NLA_BINARY:
+		return nla_put(skb, attrtype, msg->len, (void *) &msg->value);
+	default:
+		return -EINVAL;
+	}
+}
+
+static int
+devlink_fmsg_prepare_skb(struct devlink_fmsg *fmsg, struct sk_buff *skb,
+			 int *start)
+{
+	struct devlink_fmsg_item *item;
+	struct nlattr *fmsg_nlattr;
+	int i = 0;
+	int err;
+
+	fmsg_nlattr = nla_nest_start(skb, DEVLINK_ATTR_FMSG);
+	if (!fmsg_nlattr)
+		return -EMSGSIZE;
+
+	list_for_each_entry(item, &fmsg->item_list, list) {
+		if (i < *start) {
+			i++;
+			continue;
+		}
+
+		switch (item->attrtype) {
+		case DEVLINK_ATTR_FMSG_OBJ_NEST_START:
+		case DEVLINK_ATTR_FMSG_PAIR_NEST_START:
+		case DEVLINK_ATTR_FMSG_ARR_NEST_START:
+		case DEVLINK_ATTR_FMSG_NEST_END:
+			err = nla_put_flag(skb, item->attrtype);
+			break;
+		case DEVLINK_ATTR_FMSG_OBJ_VALUE_DATA:
+			err = devlink_fmsg_item_fill_type(item, skb);
+			if (err)
+				break;
+			err = devlink_fmsg_item_fill_data(item, skb);
+			break;
+		case DEVLINK_ATTR_FMSG_OBJ_NAME:
+			err = nla_put_string(skb, item->attrtype,
+					     (char *) &item->value);
+			break;
+		default:
+			err = -EINVAL;
+			break;
+		}
+		if (!err)
+			*start = ++i;
+		else
+			break;
+	}
+
+	nla_nest_end(skb, fmsg_nlattr);
+	return err;
+}
+
+static int devlink_fmsg_snd(struct devlink_fmsg *fmsg,
+			    struct genl_info *info,
+			    enum devlink_command cmd, int flags)
+{
+	struct nlmsghdr *nlh;
+	struct sk_buff *skb;
+	bool last = false;
+	int index = 0;
+	void *hdr;
+	int err;
+
+	while (!last) {
+		int tmp_index = index;
+
+		skb = genlmsg_new(GENLMSG_DEFAULT_SIZE, GFP_KERNEL);
+		if (!skb)
+			return -ENOMEM;
+
+		hdr = genlmsg_put(skb, info->snd_portid, info->snd_seq,
+				  &devlink_nl_family, flags | NLM_F_MULTI, cmd);
+		if (!hdr) {
+			err = -EMSGSIZE;
+			goto nla_put_failure;
+		}
+
+		err = devlink_fmsg_prepare_skb(fmsg, skb, &index);
+		if (!err)
+			last = true;
+		else if (err != -EMSGSIZE || tmp_index == index)
+			goto nla_put_failure;
+
+		genlmsg_end(skb, hdr);
+		err = genlmsg_reply(skb, info);
+		if (err)
+			return err;
+	}
+
+	skb = genlmsg_new(GENLMSG_DEFAULT_SIZE, GFP_KERNEL);
+	if (!skb)
+		return -ENOMEM;
+	nlh = nlmsg_put(skb, info->snd_portid, info->snd_seq,
+			NLMSG_DONE, 0, flags | NLM_F_MULTI);
+	if (!nlh) {
+		err = -EMSGSIZE;
+		goto nla_put_failure;
+	}
+	err = genlmsg_reply(skb, info);
+	if (err)
+		return err;
+
+	return 0;
+
+nla_put_failure:
+	nlmsg_free(skb);
+	return err;
+}
+
+struct devlink_health_reporter {
+	struct list_head list;
+	void *priv;
+	const struct devlink_health_reporter_ops *ops;
+	struct devlink *devlink;
+	struct devlink_fmsg *dump_fmsg;
+	struct mutex dump_lock; /* lock parallel read/write from dump buffers */
+	u64 graceful_period;
+	bool auto_recover;
+	u8 health_state;
+	u64 dump_ts;
+	u64 error_count;
+	u64 recovery_count;
+	u64 last_recovery_ts;
+};
+
+enum devlink_health_reporter_state {
+	DEVLINK_HEALTH_REPORTER_STATE_HEALTHY,
+	DEVLINK_HEALTH_REPORTER_STATE_ERROR,
+};
+
+void *
+devlink_health_reporter_priv(struct devlink_health_reporter *reporter)
+{
+	return reporter->priv;
+}
+EXPORT_SYMBOL_GPL(devlink_health_reporter_priv);
+
+static struct devlink_health_reporter *
+devlink_health_reporter_find_by_name(struct devlink *devlink,
+				     const char *reporter_name)
+{
+	struct devlink_health_reporter *reporter;
+
+	list_for_each_entry(reporter, &devlink->reporter_list, list)
+		if (!strcmp(reporter->ops->name, reporter_name))
+			return reporter;
+	return NULL;
+}
+
+/**
+ *	devlink_health_reporter_create - create devlink health reporter
+ *
+ *	@devlink: devlink
+ *	@ops: ops
+ *	@graceful_period: to avoid recovery loops, in msecs
+ *	@auto_recover: auto recover when error occurs
+ *	@priv: priv
+ */
+struct devlink_health_reporter *
+devlink_health_reporter_create(struct devlink *devlink,
+			       const struct devlink_health_reporter_ops *ops,
+			       u64 graceful_period, bool auto_recover,
+			       void *priv)
+{
+	struct devlink_health_reporter *reporter;
+
+	mutex_lock(&devlink->lock);
+	if (devlink_health_reporter_find_by_name(devlink, ops->name)) {
+		reporter = ERR_PTR(-EEXIST);
+		goto unlock;
+	}
+
+	if (WARN_ON(auto_recover && !ops->recover) ||
+	    WARN_ON(graceful_period && !ops->recover)) {
+		reporter = ERR_PTR(-EINVAL);
+		goto unlock;
+	}
+
+	reporter = kzalloc(sizeof(*reporter), GFP_KERNEL);
+	if (!reporter) {
+		reporter = ERR_PTR(-ENOMEM);
+		goto unlock;
+	}
+
+	reporter->priv = priv;
+	reporter->ops = ops;
+	reporter->devlink = devlink;
+	reporter->graceful_period = graceful_period;
+	reporter->auto_recover = auto_recover;
+	mutex_init(&reporter->dump_lock);
+	list_add_tail(&reporter->list, &devlink->reporter_list);
+unlock:
+	mutex_unlock(&devlink->lock);
+	return reporter;
+}
+EXPORT_SYMBOL_GPL(devlink_health_reporter_create);
+
+/**
+ *	devlink_health_reporter_destroy - destroy devlink health reporter
+ *
+ *	@reporter: devlink health reporter to destroy
+ */
+void
+devlink_health_reporter_destroy(struct devlink_health_reporter *reporter)
+{
+	mutex_lock(&reporter->devlink->lock);
+	list_del(&reporter->list);
+	mutex_unlock(&reporter->devlink->lock);
+	if (reporter->dump_fmsg)
+		devlink_fmsg_free(reporter->dump_fmsg);
+	kfree(reporter);
+}
+EXPORT_SYMBOL_GPL(devlink_health_reporter_destroy);
+
+static int
+devlink_health_reporter_recover(struct devlink_health_reporter *reporter,
+				void *priv_ctx)
+{
+	int err;
+
+	if (!reporter->ops->recover)
+		return -EOPNOTSUPP;
+
+	err = reporter->ops->recover(reporter, priv_ctx);
+	if (err)
+		return err;
+
+	reporter->recovery_count++;
+	reporter->health_state = DEVLINK_HEALTH_REPORTER_STATE_HEALTHY;
+	reporter->last_recovery_ts = jiffies;
+
+	return 0;
+}
+
+static void
+devlink_health_dump_clear(struct devlink_health_reporter *reporter)
+{
+	if (!reporter->dump_fmsg)
+		return;
+	devlink_fmsg_free(reporter->dump_fmsg);
+	reporter->dump_fmsg = NULL;
+}
+
+static int devlink_health_do_dump(struct devlink_health_reporter *reporter,
+				  void *priv_ctx)
+{
+	int err;
+
+	if (!reporter->ops->dump)
+		return 0;
+
+	if (reporter->dump_fmsg)
+		return 0;
+
+	reporter->dump_fmsg = devlink_fmsg_alloc();
+	if (!reporter->dump_fmsg) {
+		err = -ENOMEM;
+		return err;
+	}
+
+	err = devlink_fmsg_obj_nest_start(reporter->dump_fmsg);
+	if (err)
+		goto dump_err;
+
+	err = reporter->ops->dump(reporter, reporter->dump_fmsg,
+				  priv_ctx);
+	if (err)
+		goto dump_err;
+
+	err = devlink_fmsg_obj_nest_end(reporter->dump_fmsg);
+	if (err)
+		goto dump_err;
+
+	reporter->dump_ts = jiffies;
+
+	return 0;
+
+dump_err:
+	devlink_health_dump_clear(reporter);
+	return err;
+}
+
+int devlink_health_report(struct devlink_health_reporter *reporter,
+			  const char *msg, void *priv_ctx)
+{
+	struct devlink *devlink = reporter->devlink;
+
+	/* write a log message of the current error */
+	WARN_ON(!msg);
+	trace_devlink_health_report(devlink, reporter->ops->name, msg);
+	reporter->error_count++;
+
+	/* abort if the previous error wasn't recovered */
+	if (reporter->auto_recover &&
+	    (reporter->health_state != DEVLINK_HEALTH_REPORTER_STATE_HEALTHY ||
+	     jiffies - reporter->last_recovery_ts <
+	     msecs_to_jiffies(reporter->graceful_period))) {
+		trace_devlink_health_recover_aborted(devlink,
+						     reporter->ops->name,
+						     reporter->health_state,
+						     jiffies -
+						     reporter->last_recovery_ts);
+		return -ECANCELED;
+	}
+
+	reporter->health_state = DEVLINK_HEALTH_REPORTER_STATE_ERROR;
+
+	mutex_lock(&reporter->dump_lock);
+	/* store current dump of current error, for later analysis */
+	devlink_health_do_dump(reporter, priv_ctx);
+	mutex_unlock(&reporter->dump_lock);
+
+	if (reporter->auto_recover)
+		return devlink_health_reporter_recover(reporter, priv_ctx);
+
+	return 0;
+}
+EXPORT_SYMBOL_GPL(devlink_health_report);
+
+static struct devlink_health_reporter *
+devlink_health_reporter_get_from_info(struct devlink *devlink,
+				      struct genl_info *info)
+{
+	char *reporter_name;
+
+	if (!info->attrs[DEVLINK_ATTR_HEALTH_REPORTER_NAME])
+		return NULL;
+
+	reporter_name =
+		nla_data(info->attrs[DEVLINK_ATTR_HEALTH_REPORTER_NAME]);
+	return devlink_health_reporter_find_by_name(devlink, reporter_name);
+}
+
+static int
+devlink_nl_health_reporter_fill(struct sk_buff *msg,
+				struct devlink *devlink,
+				struct devlink_health_reporter *reporter,
+				enum devlink_command cmd, u32 portid,
+				u32 seq, int flags)
+{
+	struct nlattr *reporter_attr;
+	void *hdr;
+
+	hdr = genlmsg_put(msg, portid, seq, &devlink_nl_family, flags, cmd);
+	if (!hdr)
+		return -EMSGSIZE;
+
+	if (devlink_nl_put_handle(msg, devlink))
+		goto genlmsg_cancel;
+
+	reporter_attr = nla_nest_start(msg, DEVLINK_ATTR_HEALTH_REPORTER);
+	if (!reporter_attr)
+		goto genlmsg_cancel;
+	if (nla_put_string(msg, DEVLINK_ATTR_HEALTH_REPORTER_NAME,
+			   reporter->ops->name))
+		goto reporter_nest_cancel;
+	if (nla_put_u8(msg, DEVLINK_ATTR_HEALTH_REPORTER_STATE,
+		       reporter->health_state))
+		goto reporter_nest_cancel;
+	if (nla_put_u64_64bit(msg, DEVLINK_ATTR_HEALTH_REPORTER_ERR,
+			      reporter->error_count, DEVLINK_ATTR_PAD))
+		goto reporter_nest_cancel;
+	if (nla_put_u64_64bit(msg, DEVLINK_ATTR_HEALTH_REPORTER_RECOVER,
+			      reporter->recovery_count, DEVLINK_ATTR_PAD))
+		goto reporter_nest_cancel;
+	if (nla_put_u64_64bit(msg, DEVLINK_ATTR_HEALTH_REPORTER_GRACEFUL_PERIOD,
+			      reporter->graceful_period,
+			      DEVLINK_ATTR_PAD))
+		goto reporter_nest_cancel;
+	if (nla_put_u8(msg, DEVLINK_ATTR_HEALTH_REPORTER_AUTO_RECOVER,
+		       reporter->auto_recover))
+		goto reporter_nest_cancel;
+	if (reporter->dump_fmsg &&
+	    nla_put_u64_64bit(msg, DEVLINK_ATTR_HEALTH_REPORTER_DUMP_TS,
+			      jiffies_to_msecs(reporter->dump_ts),
+			      DEVLINK_ATTR_PAD))
+		goto reporter_nest_cancel;
+
+	nla_nest_end(msg, reporter_attr);
+	genlmsg_end(msg, hdr);
+	return 0;
+
+reporter_nest_cancel:
+	nla_nest_end(msg, reporter_attr);
+genlmsg_cancel:
+	genlmsg_cancel(msg, hdr);
+	return -EMSGSIZE;
+}
+
+static int devlink_nl_cmd_health_reporter_get_doit(struct sk_buff *skb,
+						   struct genl_info *info)
+{
+	struct devlink *devlink = info->user_ptr[0];
+	struct devlink_health_reporter *reporter;
+	struct sk_buff *msg;
+	int err;
+
+	reporter = devlink_health_reporter_get_from_info(devlink, info);
+	if (!reporter)
+		return -EINVAL;
+
+	msg = nlmsg_new(NLMSG_DEFAULT_SIZE, GFP_KERNEL);
+	if (!msg)
+		return -ENOMEM;
+
+	err = devlink_nl_health_reporter_fill(msg, devlink, reporter,
+					      DEVLINK_CMD_HEALTH_REPORTER_GET,
+					      info->snd_portid, info->snd_seq,
+					      0);
+	if (err) {
+		nlmsg_free(msg);
+		return err;
+	}
+
+	return genlmsg_reply(msg, info);
+}
+
+static int
+devlink_nl_cmd_health_reporter_get_dumpit(struct sk_buff *msg,
+					  struct netlink_callback *cb)
+{
+	struct devlink_health_reporter *reporter;
+	struct devlink *devlink;
+	int start = cb->args[0];
+	int idx = 0;
+	int err;
+
+	mutex_lock(&devlink_mutex);
+	list_for_each_entry(devlink, &devlink_list, list) {
+		if (!net_eq(devlink_net(devlink), sock_net(msg->sk)))
+			continue;
+		mutex_lock(&devlink->lock);
+		list_for_each_entry(reporter, &devlink->reporter_list,
+				    list) {
+			if (idx < start) {
+				idx++;
+				continue;
+			}
+			err = devlink_nl_health_reporter_fill(msg, devlink,
+							      reporter,
+							      DEVLINK_CMD_HEALTH_REPORTER_GET,
+							      NETLINK_CB(cb->skb).portid,
+							      cb->nlh->nlmsg_seq,
+							      NLM_F_MULTI);
+			if (err) {
+				mutex_unlock(&devlink->lock);
+				goto out;
+			}
+			idx++;
+		}
+		mutex_unlock(&devlink->lock);
+	}
+out:
+	mutex_unlock(&devlink_mutex);
+
+	cb->args[0] = idx;
+	return msg->len;
+}
+
+static int
+devlink_nl_cmd_health_reporter_set_doit(struct sk_buff *skb,
+					struct genl_info *info)
+{
+	struct devlink *devlink = info->user_ptr[0];
+	struct devlink_health_reporter *reporter;
+
+	reporter = devlink_health_reporter_get_from_info(devlink, info);
+	if (!reporter)
+		return -EINVAL;
+
+	if (!reporter->ops->recover &&
+	    (info->attrs[DEVLINK_ATTR_HEALTH_REPORTER_GRACEFUL_PERIOD] ||
+	     info->attrs[DEVLINK_ATTR_HEALTH_REPORTER_AUTO_RECOVER]))
+		return -EOPNOTSUPP;
+
+	if (info->attrs[DEVLINK_ATTR_HEALTH_REPORTER_GRACEFUL_PERIOD])
+		reporter->graceful_period =
+			nla_get_u64(info->attrs[DEVLINK_ATTR_HEALTH_REPORTER_GRACEFUL_PERIOD]);
+
+	if (info->attrs[DEVLINK_ATTR_HEALTH_REPORTER_AUTO_RECOVER])
+		reporter->auto_recover =
+			nla_get_u8(info->attrs[DEVLINK_ATTR_HEALTH_REPORTER_AUTO_RECOVER]);
+
+	return 0;
+}
+
+static int devlink_nl_cmd_health_reporter_recover_doit(struct sk_buff *skb,
+						       struct genl_info *info)
+{
+	struct devlink *devlink = info->user_ptr[0];
+	struct devlink_health_reporter *reporter;
+
+	reporter = devlink_health_reporter_get_from_info(devlink, info);
+	if (!reporter)
+		return -EINVAL;
+
+	return devlink_health_reporter_recover(reporter, NULL);
+}
+
+static int devlink_nl_cmd_health_reporter_diagnose_doit(struct sk_buff *skb,
+							struct genl_info *info)
+{
+	struct devlink *devlink = info->user_ptr[0];
+	struct devlink_health_reporter *reporter;
+	struct devlink_fmsg *fmsg;
+	int err;
+
+	reporter = devlink_health_reporter_get_from_info(devlink, info);
+	if (!reporter)
+		return -EINVAL;
+
+	if (!reporter->ops->diagnose)
+		return -EOPNOTSUPP;
+
+	fmsg = devlink_fmsg_alloc();
+	if (!fmsg)
+		return -ENOMEM;
+
+	err = devlink_fmsg_obj_nest_start(fmsg);
+	if (err)
+		goto out;
+
+	err = reporter->ops->diagnose(reporter, fmsg);
+	if (err)
+		goto out;
+
+	err = devlink_fmsg_obj_nest_end(fmsg);
+	if (err)
+		goto out;
+
+	err = devlink_fmsg_snd(fmsg, info,
+			       DEVLINK_CMD_HEALTH_REPORTER_DIAGNOSE, 0);
+
+out:
+	devlink_fmsg_free(fmsg);
+	return err;
+}
+
+static int devlink_nl_cmd_health_reporter_dump_get_doit(struct sk_buff *skb,
+							struct genl_info *info)
+{
+	struct devlink *devlink = info->user_ptr[0];
+	struct devlink_health_reporter *reporter;
+	int err;
+
+	reporter = devlink_health_reporter_get_from_info(devlink, info);
+	if (!reporter)
+		return -EINVAL;
+
+	if (!reporter->ops->dump)
+		return -EOPNOTSUPP;
+
+	mutex_lock(&reporter->dump_lock);
+	err = devlink_health_do_dump(reporter, NULL);
+	if (err)
+		goto out;
+
+	err = devlink_fmsg_snd(reporter->dump_fmsg, info,
+			       DEVLINK_CMD_HEALTH_REPORTER_DUMP_GET, 0);
+
+out:
+	mutex_unlock(&reporter->dump_lock);
+	return err;
+}
+
+static int
+devlink_nl_cmd_health_reporter_dump_clear_doit(struct sk_buff *skb,
+					       struct genl_info *info)
+{
+	struct devlink *devlink = info->user_ptr[0];
+	struct devlink_health_reporter *reporter;
+
+	reporter = devlink_health_reporter_get_from_info(devlink, info);
+	if (!reporter)
+		return -EINVAL;
+
+	if (!reporter->ops->dump)
+		return -EOPNOTSUPP;
+
+	mutex_lock(&reporter->dump_lock);
+	devlink_health_dump_clear(reporter);
+	mutex_unlock(&reporter->dump_lock);
+	return 0;
+}
+
 static const struct nla_policy devlink_nl_policy[DEVLINK_ATTR_MAX + 1] = {
 	[DEVLINK_ATTR_BUS_NAME] = { .type = NLA_NUL_STRING },
 	[DEVLINK_ATTR_DEV_NAME] = { .type = NLA_NUL_STRING },
@@ -3904,6 +4863,9 @@ static const struct nla_policy devlink_nl_policy[DEVLINK_ATTR_MAX + 1] = {
 	[DEVLINK_ATTR_PARAM_VALUE_CMODE] = { .type = NLA_U8 },
 	[DEVLINK_ATTR_REGION_NAME] = { .type = NLA_NUL_STRING },
 	[DEVLINK_ATTR_REGION_SNAPSHOT_ID] = { .type = NLA_U32 },
+	[DEVLINK_ATTR_HEALTH_REPORTER_NAME] = { .type = NLA_NUL_STRING },
+	[DEVLINK_ATTR_HEALTH_REPORTER_GRACEFUL_PERIOD] = { .type = NLA_U64 },
+	[DEVLINK_ATTR_HEALTH_REPORTER_AUTO_RECOVER] = { .type = NLA_U8 },
 };
 
 static const struct genl_ops devlink_nl_ops[] = {
@@ -4147,6 +5109,51 @@ static const struct genl_ops devlink_nl_ops[] = {
 		.internal_flags = DEVLINK_NL_FLAG_NEED_DEVLINK,
 		/* can be retrieved by unprivileged users */
 	},
+	{
+		.cmd = DEVLINK_CMD_HEALTH_REPORTER_GET,
+		.doit = devlink_nl_cmd_health_reporter_get_doit,
+		.dumpit = devlink_nl_cmd_health_reporter_get_dumpit,
+		.policy = devlink_nl_policy,
+		.internal_flags = DEVLINK_NL_FLAG_NEED_DEVLINK,
+		/* can be retrieved by unprivileged users */
+	},
+	{
+		.cmd = DEVLINK_CMD_HEALTH_REPORTER_SET,
+		.doit = devlink_nl_cmd_health_reporter_set_doit,
+		.policy = devlink_nl_policy,
+		.flags = GENL_ADMIN_PERM,
+		.internal_flags = DEVLINK_NL_FLAG_NEED_DEVLINK,
+	},
+	{
+		.cmd = DEVLINK_CMD_HEALTH_REPORTER_RECOVER,
+		.doit = devlink_nl_cmd_health_reporter_recover_doit,
+		.policy = devlink_nl_policy,
+		.flags = GENL_ADMIN_PERM,
+		.internal_flags = DEVLINK_NL_FLAG_NEED_DEVLINK,
+	},
+	{
+		.cmd = DEVLINK_CMD_HEALTH_REPORTER_DIAGNOSE,
+		.doit = devlink_nl_cmd_health_reporter_diagnose_doit,
+		.policy = devlink_nl_policy,
+		.flags = GENL_ADMIN_PERM,
+		.internal_flags = DEVLINK_NL_FLAG_NEED_DEVLINK,
+	},
+	{
+		.cmd = DEVLINK_CMD_HEALTH_REPORTER_DUMP_GET,
+		.doit = devlink_nl_cmd_health_reporter_dump_get_doit,
+		.policy = devlink_nl_policy,
+		.flags = GENL_ADMIN_PERM,
+		.internal_flags = DEVLINK_NL_FLAG_NEED_DEVLINK |
+				  DEVLINK_NL_FLAG_NO_LOCK,
+	},
+	{
+		.cmd = DEVLINK_CMD_HEALTH_REPORTER_DUMP_CLEAR,
+		.doit = devlink_nl_cmd_health_reporter_dump_clear_doit,
+		.policy = devlink_nl_policy,
+		.flags = GENL_ADMIN_PERM,
+		.internal_flags = DEVLINK_NL_FLAG_NEED_DEVLINK |
+				  DEVLINK_NL_FLAG_NO_LOCK,
+	},
 };
 
 static struct genl_family devlink_nl_family __ro_after_init = {
@@ -4187,6 +5194,7 @@ struct devlink *devlink_alloc(const struct devlink_ops *ops, size_t priv_size)
 	INIT_LIST_HEAD(&devlink->resource_list);
 	INIT_LIST_HEAD(&devlink->param_list);
 	INIT_LIST_HEAD(&devlink->region_list);
+	INIT_LIST_HEAD(&devlink->reporter_list);
 	mutex_init(&devlink->lock);
 	return devlink;
 }