From c6e77aa9dd82bc18a89bf49418f8f7e961cfccc8 Mon Sep 17 00:00:00 2001 From: Shay Drory Date: Tue, 9 Apr 2024 22:08:10 +0300 Subject: net/mlx5: Register devlink first under devlink lock In case device is having a non fatal FW error during probe, the driver will report the error to user via devlink. This will trigger a WARN_ON, since mlx5 is calling devlink_register() last. In order to avoid the WARN_ON[1], change mlx5 to invoke devl_register() first under devlink lock. [1] WARNING: CPU: 5 PID: 227 at net/devlink/health.c:483 devlink_recover_notify.constprop.0+0xb8/0xc0 CPU: 5 PID: 227 Comm: kworker/u16:3 Not tainted 6.4.0-rc5_for_upstream_min_debug_2023_06_12_12_38 #1 Hardware name: QEMU Standard PC (Q35 + ICH9, 2009), BIOS rel-1.13.0-0-gf21b5a4aeb02-prebuilt.qemu.org 04/01/2014 Workqueue: mlx5_health0000:08:00.0 mlx5_fw_reporter_err_work [mlx5_core] RIP: 0010:devlink_recover_notify.constprop.0+0xb8/0xc0 Call Trace: ? __warn+0x79/0x120 ? devlink_recover_notify.constprop.0+0xb8/0xc0 ? report_bug+0x17c/0x190 ? handle_bug+0x3c/0x60 ? exc_invalid_op+0x14/0x70 ? asm_exc_invalid_op+0x16/0x20 ? devlink_recover_notify.constprop.0+0xb8/0xc0 devlink_health_report+0x4a/0x1c0 mlx5_fw_reporter_err_work+0xa4/0xd0 [mlx5_core] process_one_work+0x1bb/0x3c0 ? process_one_work+0x3c0/0x3c0 worker_thread+0x4d/0x3c0 ? process_one_work+0x3c0/0x3c0 kthread+0xc6/0xf0 ? kthread_complete_and_exit+0x20/0x20 ret_from_fork+0x1f/0x30 Fixes: cf530217408e ("devlink: Notify users when objects are accessible") Signed-off-by: Shay Drory Reviewed-by: Moshe Shemesh Signed-off-by: Saeed Mahameed Signed-off-by: Tariq Toukan Link: https://lore.kernel.org/r/20240409190820.227554-3-tariqt@nvidia.com Signed-off-by: Jakub Kicinski --- drivers/net/ethernet/mellanox/mlx5/core/main.c | 37 ++++++++++++---------- .../ethernet/mellanox/mlx5/core/sf/dev/driver.c | 1 - 2 files changed, 20 insertions(+), 18 deletions(-) diff --git a/drivers/net/ethernet/mellanox/mlx5/core/main.c b/drivers/net/ethernet/mellanox/mlx5/core/main.c index c2593625c09ad..59806553889e9 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/main.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/main.c @@ -1480,6 +1480,14 @@ int mlx5_init_one_devl_locked(struct mlx5_core_dev *dev) if (err) goto err_register; + err = mlx5_crdump_enable(dev); + if (err) + mlx5_core_err(dev, "mlx5_crdump_enable failed with error code %d\n", err); + + err = mlx5_hwmon_dev_register(dev); + if (err) + mlx5_core_err(dev, "mlx5_hwmon_dev_register failed with error code %d\n", err); + mutex_unlock(&dev->intf_state_mutex); return 0; @@ -1505,7 +1513,10 @@ int mlx5_init_one(struct mlx5_core_dev *dev) int err; devl_lock(devlink); + devl_register(devlink); err = mlx5_init_one_devl_locked(dev); + if (err) + devl_unregister(devlink); devl_unlock(devlink); return err; } @@ -1517,6 +1528,8 @@ void mlx5_uninit_one(struct mlx5_core_dev *dev) devl_lock(devlink); mutex_lock(&dev->intf_state_mutex); + mlx5_hwmon_dev_unregister(dev); + mlx5_crdump_disable(dev); mlx5_unregister_device(dev); if (!test_bit(MLX5_INTERFACE_STATE_UP, &dev->intf_state)) { @@ -1534,6 +1547,7 @@ void mlx5_uninit_one(struct mlx5_core_dev *dev) mlx5_function_teardown(dev, true); out: mutex_unlock(&dev->intf_state_mutex); + devl_unregister(devlink); devl_unlock(devlink); } @@ -1680,16 +1694,20 @@ int mlx5_init_one_light(struct mlx5_core_dev *dev) } devl_lock(devlink); + devl_register(devlink); + err = mlx5_devlink_params_register(priv_to_devlink(dev)); - devl_unlock(devlink); if (err) { mlx5_core_warn(dev, "mlx5_devlink_param_reg err = %d\n", err); goto query_hca_caps_err; } + devl_unlock(devlink); return 0; query_hca_caps_err: + devl_unregister(devlink); + devl_unlock(devlink); mlx5_function_disable(dev, true); out: dev->state = MLX5_DEVICE_STATE_INTERNAL_ERROR; @@ -1702,6 +1720,7 @@ void mlx5_uninit_one_light(struct mlx5_core_dev *dev) devl_lock(devlink); mlx5_devlink_params_unregister(priv_to_devlink(dev)); + devl_unregister(devlink); devl_unlock(devlink); if (dev->state != MLX5_DEVICE_STATE_UP) return; @@ -1943,16 +1962,7 @@ static int probe_one(struct pci_dev *pdev, const struct pci_device_id *id) goto err_init_one; } - err = mlx5_crdump_enable(dev); - if (err) - dev_err(&pdev->dev, "mlx5_crdump_enable failed with error code %d\n", err); - - err = mlx5_hwmon_dev_register(dev); - if (err) - mlx5_core_err(dev, "mlx5_hwmon_dev_register failed with error code %d\n", err); - pci_save_state(pdev); - devlink_register(devlink); return 0; err_init_one: @@ -1973,16 +1983,9 @@ static void remove_one(struct pci_dev *pdev) struct devlink *devlink = priv_to_devlink(dev); set_bit(MLX5_BREAK_FW_WAIT, &dev->intf_state); - /* mlx5_drain_fw_reset() and mlx5_drain_health_wq() are using - * devlink notify APIs. - * Hence, we must drain them before unregistering the devlink. - */ mlx5_drain_fw_reset(dev); mlx5_drain_health_wq(dev); - devlink_unregister(devlink); mlx5_sriov_disable(pdev, false); - mlx5_hwmon_dev_unregister(dev); - mlx5_crdump_disable(dev); mlx5_uninit_one(dev); mlx5_pci_close(dev); mlx5_mdev_uninit(dev); diff --git a/drivers/net/ethernet/mellanox/mlx5/core/sf/dev/driver.c b/drivers/net/ethernet/mellanox/mlx5/core/sf/dev/driver.c index bc863e1f062e6..e3bf8c7e4baa6 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/sf/dev/driver.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/sf/dev/driver.c @@ -101,7 +101,6 @@ static void mlx5_sf_dev_remove(struct auxiliary_device *adev) devlink = priv_to_devlink(mdev); set_bit(MLX5_BREAK_FW_WAIT, &mdev->intf_state); mlx5_drain_health_wq(mdev); - devlink_unregister(devlink); if (mlx5_dev_is_lightweight(mdev)) mlx5_uninit_one_light(mdev); else -- cgit 1.2.3-korg