net/mlx5: Fix race between PCI error handlers and health work

[cascardo/linux.git] / drivers / net / ethernet / mellanox / mlx5 / core / health.c

diff --git a/drivers/net/ethernet/mellanox/mlx5/core/health.c b/drivers/net/ethernet/mellanox/mlx5/core/health.c
index 96a5946..2d00022 100644 (file)
--- a/drivers/net/ethernet/mellanox/mlx5/core/health.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/health.c
@@ -64,6 +64,10 @@ enum {
        MLX5_NIC_IFC_NO_DRAM_NIC        = 2
 };
 
+enum {
+       MLX5_DROP_NEW_HEALTH_WORK,
+};
+
 static u8 get_nic_interface(struct mlx5_core_dev *dev)
 {
        return (ioread32be(&dev->iseg->cmdq_addr_l_sz) >> 8) & 3;
@@ -272,7 +276,13 @@ static void poll_health(unsigned long data)
        if (in_fatal(dev) && !health->sick) {
                health->sick = true;
                print_health_info(dev);
-               queue_work(health->wq, &health->work);
+               spin_lock(&health->wq_lock);
+               if (!test_bit(MLX5_DROP_NEW_HEALTH_WORK, &health->flags))
+                       queue_work(health->wq, &health->work);
+               else
+                       dev_err(&dev->pdev->dev,
+                               "new health works are not permitted at this stage\n");
+               spin_unlock(&health->wq_lock);
        }
 }
 
@@ -281,6 +291,8 @@ void mlx5_start_health_poll(struct mlx5_core_dev *dev)
        struct mlx5_core_health *health = &dev->priv.health;
 
        init_timer(&health->timer);
+       health->sick = 0;
+       clear_bit(MLX5_DROP_NEW_HEALTH_WORK, &health->flags);
        health->health = &dev->iseg->health;
        health->health_counter = &dev->iseg->health_counter;
 
@@ -297,6 +309,16 @@ void mlx5_stop_health_poll(struct mlx5_core_dev *dev)
        del_timer_sync(&health->timer);
 }
 
+void mlx5_drain_health_wq(struct mlx5_core_dev *dev)
+{
+       struct mlx5_core_health *health = &dev->priv.health;
+
+       spin_lock(&health->wq_lock);
+       set_bit(MLX5_DROP_NEW_HEALTH_WORK, &health->flags);
+       spin_unlock(&health->wq_lock);
+       cancel_work_sync(&health->work);
+}
+
 void mlx5_health_cleanup(struct mlx5_core_dev *dev)
 {
        struct mlx5_core_health *health = &dev->priv.health;
@@ -320,7 +342,7 @@ int mlx5_health_init(struct mlx5_core_dev *dev)
        kfree(name);
        if (!health->wq)
                return -ENOMEM;
-
+       spin_lock_init(&health->wq_lock);
        INIT_WORK(&health->work, health_care);
 
        return 0;