net/mlx4: Lock mlx4 devlink reload callback

Change devlink instance locks in mlx4 driver to have devlink reload
callback locked, while keeping all driver paths which leads to devl_ API
functions called by the mlx4 driver locked.

Signed-off-by: Moshe Shemesh <moshe@nvidia.com>
Reviewed-by: Tariq Toukan <tariqt@nvidia.com>
Reviewed-by: Jiri Pirko <jiri@nvidia.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

2 files changed, 37 insertions, 13 deletions

diff --git a/drivers/net/ethernet/mellanox/mlx4/catas.c b/drivers/net/ethernet/mellanox/mlx4/catas.c
index 5b11557f1ae4..0eb7b83637d8 100644
--- a/drivers/net/ethernet/mellanox/mlx4/catas.c
+++ b/drivers/net/ethernet/mellanox/mlx4/catas.c
@@ -204,9 +204,13 @@ out:
 
 static void mlx4_handle_error_state(struct mlx4_dev_persistent *persist)
 {
+	struct mlx4_dev *dev = persist->dev;
+	struct devlink *devlink;
 	int err = 0;
 
 	mlx4_enter_error_state(persist);
+	devlink = priv_to_devlink(mlx4_priv(dev));
+	devl_lock(devlink);
 	mutex_lock(&persist->interface_state_mutex);
 	if (persist->interface_state & MLX4_INTERFACE_STATE_UP &&
 	    !(persist->interface_state & MLX4_INTERFACE_STATE_DELETION)) {
@@ -215,6 +219,7 @@ static void mlx4_handle_error_state(struct mlx4_dev_persistent *persist)
 			  err);
 	}
 	mutex_unlock(&persist->interface_state_mutex);
+	devl_unlock(devlink);
 }
 
 static void dump_err_buf(struct mlx4_dev *dev)
diff --git a/drivers/net/ethernet/mellanox/mlx4/main.c b/drivers/net/ethernet/mellanox/mlx4/main.c
index 0166d003f22c..2c764d1d897d 100644
--- a/drivers/net/ethernet/mellanox/mlx4/main.c
+++ b/drivers/net/ethernet/mellanox/mlx4/main.c
@@ -3342,6 +3342,7 @@ static int mlx4_load_one(struct pci_dev *pdev, int pci_dev_data,
 	struct mlx4_dev_cap *dev_cap = NULL;
 	int existing_vfs = 0;
 
+	devl_assert_locked(devlink);
 	dev = &priv->dev;
 
 	INIT_LIST_HEAD(&priv->ctx_list);
@@ -3630,7 +3631,6 @@ slave_start:
 		}
 	}
 
-	devl_lock(devlink);
 	for (port = 1; port <= dev->caps.num_ports; port++) {
 		err = mlx4_init_port_info(dev, port);
 		if (err)
@@ -3644,7 +3644,6 @@ slave_start:
 	if (err)
 		goto err_port;
 
-	devl_unlock(devlink);
 	mlx4_request_modules(dev);
 
 	mlx4_sense_init(dev);
@@ -3661,7 +3660,6 @@ slave_start:
 err_port:
 	for (--port; port >= 1; --port)
 		mlx4_cleanup_port_info(&priv->port[port]);
-	devl_unlock(devlink);
 
 	mlx4_cleanup_default_counters(dev);
 	if (!mlx4_is_slave(dev))
@@ -3736,7 +3734,6 @@ static int __mlx4_init_one(struct pci_dev *pdev, int pci_dev_data,
 	int prb_vf[MLX4_MAX_PORTS + 1] = {0, 0, 0};
 	const int param_map[MLX4_MAX_PORTS + 1][MLX4_MAX_PORTS + 1] = {
 		{2, 0, 0}, {0, 1, 2}, {0, 1, 2} };
-	struct devlink *devlink = priv_to_devlink(priv);
 	unsigned total_vfs = 0;
 	unsigned int i;
 
@@ -3849,9 +3846,7 @@ static int __mlx4_init_one(struct pci_dev *pdev, int pci_dev_data,
 		}
 	}
 
-	devl_lock(devlink);
 	err = mlx4_crdump_init(&priv->dev);
-	devl_unlock(devlink);
 	if (err)
 		goto err_release_regions;
 
@@ -3869,9 +3864,7 @@ err_catas:
 	mlx4_catas_end(&priv->dev);
 
 err_crdump:
-	devl_lock(devlink);
 	mlx4_crdump_end(&priv->dev);
-	devl_unlock(devlink);
 
 err_release_regions:
 	pci_release_regions(pdev);
@@ -3965,9 +3958,11 @@ static int mlx4_devlink_reload_down(struct devlink *devlink, bool netns_change,
 		NL_SET_ERR_MSG_MOD(extack, "Namespace change is not supported");
 		return -EOPNOTSUPP;
 	}
+	devl_lock(devlink);
 	if (persist->num_vfs)
 		mlx4_warn(persist->dev, "Reload performed on PF, will cause reset on operating Virtual Functions\n");
 	mlx4_restart_one_down(persist->pdev);
+	devl_unlock(devlink);
 	return 0;
 }
 
@@ -3980,8 +3975,10 @@ static int mlx4_devlink_reload_up(struct devlink *devlink, enum devlink_reload_a
 	struct mlx4_dev_persistent *persist = dev->persist;
 	int err;
 
+	devl_lock(devlink);
 	*actions_performed = BIT(DEVLINK_RELOAD_ACTION_DRIVER_REINIT);
 	err = mlx4_restart_one_up(persist->pdev, true, devlink);
+	devl_unlock(devlink);
 	if (err)
 		mlx4_err(persist->dev, "mlx4_restart_one_up failed, ret=%d\n",
 			 err);
@@ -4008,6 +4005,7 @@ static int mlx4_init_one(struct pci_dev *pdev, const struct pci_device_id *id)
 	devlink = devlink_alloc(&mlx4_devlink_ops, sizeof(*priv), &pdev->dev);
 	if (!devlink)
 		return -ENOMEM;
+	devl_lock(devlink);
 	priv = devlink_priv(devlink);
 
 	dev       = &priv->dev;
@@ -4035,6 +4033,7 @@ static int mlx4_init_one(struct pci_dev *pdev, const struct pci_device_id *id)
 
 	pci_save_state(pdev);
 	devlink_set_features(devlink, DEVLINK_F_RELOAD);
+	devl_unlock(devlink);
 	devlink_register(devlink);
 	return 0;
 
@@ -4044,6 +4043,7 @@ err_params_unregister:
 err_devlink_unregister:
 	kfree(dev->persist);
 err_devlink_free:
+	devl_unlock(devlink);
 	devlink_free(devlink);
 	return ret;
 }
@@ -4069,6 +4069,7 @@ static void mlx4_unload_one(struct pci_dev *pdev)
 	int p, i;
 
 	devlink = priv_to_devlink(priv);
+	devl_assert_locked(devlink);
 	if (priv->removed)
 		return;
 
@@ -4084,12 +4085,10 @@ static void mlx4_unload_one(struct pci_dev *pdev)
 	mlx4_stop_sense(dev);
 	mlx4_unregister_device(dev);
 
-	devl_lock(devlink);
 	for (p = 1; p <= dev->caps.num_ports; p++) {
 		mlx4_cleanup_port_info(&priv->port[p]);
 		mlx4_CLOSE_PORT(dev, p);
 	}
-	devl_unlock(devlink);
 
 	if (mlx4_is_master(dev))
 		mlx4_free_resource_tracker(dev,
@@ -4150,6 +4149,7 @@ static void mlx4_remove_one(struct pci_dev *pdev)
 
 	devlink_unregister(devlink);
 
+	devl_lock(devlink);
 	if (mlx4_is_slave(dev))
 		persist->interface_state |= MLX4_INTERFACE_STATE_NOWAIT;
 
@@ -4174,9 +4174,7 @@ static void mlx4_remove_one(struct pci_dev *pdev)
 	else
 		mlx4_info(dev, "%s: interface is down\n", __func__);
 	mlx4_catas_end(dev);
-	devl_lock(devlink);
 	mlx4_crdump_end(dev);
-	devl_unlock(devlink);
 	if (dev->flags & MLX4_FLAG_SRIOV && !active_vfs) {
 		mlx4_warn(dev, "Disabling SR-IOV\n");
 		pci_disable_sriov(pdev);
@@ -4187,6 +4185,7 @@ static void mlx4_remove_one(struct pci_dev *pdev)
 	devlink_params_unregister(devlink, mlx4_devlink_params,
 				  ARRAY_SIZE(mlx4_devlink_params));
 	kfree(dev->persist);
+	devl_unlock(devlink);
 	devlink_free(devlink);
 }
 
@@ -4307,15 +4306,20 @@ static pci_ers_result_t mlx4_pci_err_detected(struct pci_dev *pdev,
 					      pci_channel_state_t state)
 {
 	struct mlx4_dev_persistent *persist = pci_get_drvdata(pdev);
+	struct mlx4_dev *dev = persist->dev;
+	struct devlink *devlink;
 
 	mlx4_err(persist->dev, "mlx4_pci_err_detected was called\n");
 	mlx4_enter_error_state(persist);
 
+	devlink = priv_to_devlink(mlx4_priv(dev));
+	devl_lock(devlink);
 	mutex_lock(&persist->interface_state_mutex);
 	if (persist->interface_state & MLX4_INTERFACE_STATE_UP)
 		mlx4_unload_one(pdev);
 
 	mutex_unlock(&persist->interface_state_mutex);
+	devl_unlock(devlink);
 	if (state == pci_channel_io_perm_failure)
 		return PCI_ERS_RESULT_DISCONNECT;
 
@@ -4348,6 +4352,7 @@ static void mlx4_pci_resume(struct pci_dev *pdev)
 	struct mlx4_dev	 *dev  = persist->dev;
 	struct mlx4_priv *priv = mlx4_priv(dev);
 	int nvfs[MLX4_MAX_PORTS + 1] = {0, 0, 0};
+	struct devlink *devlink;
 	int total_vfs;
 	int err;
 
@@ -4355,6 +4360,8 @@ static void mlx4_pci_resume(struct pci_dev *pdev)
 	total_vfs = dev->persist->num_vfs;
 	memcpy(nvfs, dev->persist->nvfs, sizeof(dev->persist->nvfs));
 
+	devlink = priv_to_devlink(priv);
+	devl_lock(devlink);
 	mutex_lock(&persist->interface_state_mutex);
 	if (!(persist->interface_state & MLX4_INTERFACE_STATE_UP)) {
 		err = mlx4_load_one(pdev, priv->pci_dev_data, total_vfs, nvfs,
@@ -4373,19 +4380,23 @@ static void mlx4_pci_resume(struct pci_dev *pdev)
 	}
 end:
 	mutex_unlock(&persist->interface_state_mutex);
-
+	devl_unlock(devlink);
 }
 
 static void mlx4_shutdown(struct pci_dev *pdev)
 {
 	struct mlx4_dev_persistent *persist = pci_get_drvdata(pdev);
 	struct mlx4_dev *dev = persist->dev;
+	struct devlink *devlink;
 
 	mlx4_info(persist->dev, "mlx4_shutdown was called\n");
+	devlink = priv_to_devlink(mlx4_priv(dev));
+	devl_lock(devlink);
 	mutex_lock(&persist->interface_state_mutex);
 	if (persist->interface_state & MLX4_INTERFACE_STATE_UP)
 		mlx4_unload_one(pdev);
 	mutex_unlock(&persist->interface_state_mutex);
+	devl_unlock(devlink);
 	mlx4_pci_disable_device(dev);
 }
 
@@ -4400,12 +4411,16 @@ static int __maybe_unused mlx4_suspend(struct device *dev_d)
 	struct pci_dev *pdev = to_pci_dev(dev_d);
 	struct mlx4_dev_persistent *persist = pci_get_drvdata(pdev);
 	struct mlx4_dev	*dev = persist->dev;
+	struct devlink *devlink;
 
 	mlx4_err(dev, "suspend was called\n");
+	devlink = priv_to_devlink(mlx4_priv(dev));
+	devl_lock(devlink);
 	mutex_lock(&persist->interface_state_mutex);
 	if (persist->interface_state & MLX4_INTERFACE_STATE_UP)
 		mlx4_unload_one(pdev);
 	mutex_unlock(&persist->interface_state_mutex);
+	devl_unlock(devlink);
 
 	return 0;
 }
@@ -4417,6 +4432,7 @@ static int __maybe_unused mlx4_resume(struct device *dev_d)
 	struct mlx4_dev	*dev = persist->dev;
 	struct mlx4_priv *priv = mlx4_priv(dev);
 	int nvfs[MLX4_MAX_PORTS + 1] = {0, 0, 0};
+	struct devlink *devlink;
 	int total_vfs;
 	int ret = 0;
 
@@ -4424,6 +4440,8 @@ static int __maybe_unused mlx4_resume(struct device *dev_d)
 	total_vfs = dev->persist->num_vfs;
 	memcpy(nvfs, dev->persist->nvfs, sizeof(dev->persist->nvfs));
 
+	devlink = priv_to_devlink(priv);
+	devl_lock(devlink);
 	mutex_lock(&persist->interface_state_mutex);
 	if (!(persist->interface_state & MLX4_INTERFACE_STATE_UP)) {
 		ret = mlx4_load_one(pdev, priv->pci_dev_data, total_vfs,
@@ -4437,6 +4455,7 @@ static int __maybe_unused mlx4_resume(struct device *dev_d)
 		}
 	}
 	mutex_unlock(&persist->interface_state_mutex);
+	devl_unlock(devlink);
 
 	return ret;
 }