diff options
| -rw-r--r-- | drivers/infiniband/hw/mlx5/main.c | 75 | ||||
| -rw-r--r-- | drivers/infiniband/hw/mlx5/mlx5_ib.h | 2 |
2 files changed, 68 insertions, 9 deletions
diff --git a/drivers/infiniband/hw/mlx5/main.c b/drivers/infiniband/hw/mlx5/main.c index e81080622283..e83a5f12e6bc 100644 --- a/drivers/infiniband/hw/mlx5/main.c +++ b/drivers/infiniband/hw/mlx5/main.c @@ -3009,7 +3009,6 @@ static void mlx5_ib_handle_event(struct work_struct *_work) container_of(_work, struct mlx5_ib_event_work, work); struct mlx5_ib_dev *ibdev; struct ib_event ibev; - bool fatal = false; if (work->is_slave) { ibdev = mlx5_ib_get_ibdev_from_mpi(work->mpi); @@ -3020,12 +3019,6 @@ static void mlx5_ib_handle_event(struct work_struct *_work) } switch (work->event) { - case MLX5_DEV_EVENT_SYS_ERROR: - ibev.event = IB_EVENT_DEVICE_FATAL; - mlx5_ib_handle_internal_error(ibdev); - ibev.element.port_num = (u8)(unsigned long)work->param; - fatal = true; - break; case MLX5_EVENT_TYPE_PORT_CHANGE: if (handle_port_change(ibdev, work->param, &ibev)) goto out; @@ -3047,8 +3040,6 @@ static void mlx5_ib_handle_event(struct work_struct *_work) if (ibdev->ib_active) ib_dispatch_event(&ibev); - if (fatal) - ibdev->ib_active = false; out: kfree(work); } @@ -3092,6 +3083,66 @@ static int mlx5_ib_event_slave_port(struct notifier_block *nb, return NOTIFY_OK; } +static void mlx5_ib_handle_sys_error_event(struct work_struct *_work) +{ + struct mlx5_ib_event_work *work = + container_of(_work, struct mlx5_ib_event_work, work); + struct mlx5_ib_dev *ibdev = work->dev; + struct ib_event ibev; + + ibev.event = IB_EVENT_DEVICE_FATAL; + mlx5_ib_handle_internal_error(ibdev); + ibev.element.port_num = (u8)(unsigned long)work->param; + ibev.device = &ibdev->ib_dev; + + if (!rdma_is_port_valid(&ibdev->ib_dev, ibev.element.port_num)) { + mlx5_ib_warn(ibdev, "warning: event on port %d\n", ibev.element.port_num); + goto out; + } + + if (ibdev->ib_active) + ib_dispatch_event(&ibev); + + ibdev->ib_active = false; +out: + kfree(work); +} + +static int mlx5_ib_sys_error_event(struct notifier_block *nb, + unsigned long event, void *param) +{ + struct mlx5_ib_event_work *work; + + if (event != MLX5_DEV_EVENT_SYS_ERROR) + return NOTIFY_DONE; + + work = kmalloc(sizeof(*work), GFP_ATOMIC); + if (!work) + return NOTIFY_DONE; + + INIT_WORK(&work->work, mlx5_ib_handle_sys_error_event); + work->dev = container_of(nb, struct mlx5_ib_dev, sys_error_events); + work->is_slave = false; + work->param = param; + work->event = event; + + queue_work(mlx5_ib_event_wq, &work->work); + + return NOTIFY_OK; +} + +static int mlx5_ib_stage_sys_error_notifier_init(struct mlx5_ib_dev *dev) +{ + dev->sys_error_events.notifier_call = mlx5_ib_sys_error_event; + mlx5_notifier_register(dev->mdev, &dev->sys_error_events); + return 0; +} + +static void mlx5_ib_stage_sys_error_notifier_cleanup(struct mlx5_ib_dev *dev) +{ + mlx5_notifier_unregister(dev->mdev, &dev->sys_error_events); +} + static int mlx5_ib_get_plane_num(struct mlx5_core_dev *mdev, u8 *num_plane) { struct mlx5_hca_vport_context vport_ctx; @@ -4943,6 +4994,9 @@ static const struct mlx5_ib_profile pf_profile = { STAGE_CREATE(MLX5_IB_STAGE_WHITELIST_UID, mlx5_ib_devx_init, mlx5_ib_devx_cleanup), + STAGE_CREATE(MLX5_IB_STAGE_SYS_ERROR_NOTIFIER, + mlx5_ib_stage_sys_error_notifier_init, + mlx5_ib_stage_sys_error_notifier_cleanup), STAGE_CREATE(MLX5_IB_STAGE_IB_REG, mlx5_ib_stage_ib_reg_init, mlx5_ib_stage_ib_reg_cleanup), @@ -5000,6 +5054,9 @@ const struct mlx5_ib_profile raw_eth_profile = { STAGE_CREATE(MLX5_IB_STAGE_WHITELIST_UID, mlx5_ib_devx_init, mlx5_ib_devx_cleanup), + STAGE_CREATE(MLX5_IB_STAGE_SYS_ERROR_NOTIFIER, + mlx5_ib_stage_sys_error_notifier_init, + mlx5_ib_stage_sys_error_notifier_cleanup), STAGE_CREATE(MLX5_IB_STAGE_IB_REG, mlx5_ib_stage_ib_reg_init, mlx5_ib_stage_ib_reg_cleanup), diff --git a/drivers/infiniband/hw/mlx5/mlx5_ib.h b/drivers/infiniband/hw/mlx5/mlx5_ib.h index cc6b3b6c713c..4f4114d95130 100644 --- a/drivers/infiniband/hw/mlx5/mlx5_ib.h +++ b/drivers/infiniband/hw/mlx5/mlx5_ib.h @@ -1007,6 +1007,7 @@ enum mlx5_ib_stages { MLX5_IB_STAGE_BFREG, MLX5_IB_STAGE_PRE_IB_REG_UMR, MLX5_IB_STAGE_WHITELIST_UID, + MLX5_IB_STAGE_SYS_ERROR_NOTIFIER, MLX5_IB_STAGE_IB_REG, MLX5_IB_STAGE_DEVICE_NOTIFIER, MLX5_IB_STAGE_POST_IB_REG_UMR, @@ -1165,6 +1166,7 @@ struct mlx5_ib_dev { /* protect accessing data_direct_dev */ struct mutex data_direct_lock; struct notifier_block mdev_events; + struct notifier_block sys_error_events; struct notifier_block lag_events; int num_ports; /* serialize update of capability mask |