mlx4_en: Fix cleanup if workqueue create in mlx4_en_add() fails
authorRoland Dreier <rdreier@cisco.com>
Mon, 20 Apr 2009 18:32:08 +0000 (18:32 +0000)
committerDavid S. Miller <davem@davemloft.net>
Tue, 21 Apr 2009 08:47:08 +0000 (01:47 -0700)
If creating a workqueue fails, don't jump to the error path where that
same workqueue is destroyed, since destroy_workqueue() can't handle a
NULL pointer.

This was spotted by the Coverity checker (CID 2617).

Signed-off-by: Roland Dreier <rolandd@cisco.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
drivers/net/mlx4/en_main.c

index eda72dd2120f5a3af3291fabc3923ea5c5b94210..510633fd57f613906660f848b879b03d72fccd5c 100644 (file)
@@ -181,7 +181,7 @@ static void *mlx4_en_add(struct mlx4_dev *dev)
        mdev->workqueue = create_singlethread_workqueue("mlx4_en");
        if (!mdev->workqueue) {
                err = -ENOMEM;
-               goto err_close_nic;
+               goto err_mr;
        }
 
        /* At this stage all non-port specific tasks are complete:
@@ -214,9 +214,8 @@ err_free_netdev:
        flush_workqueue(mdev->workqueue);
 
        /* Stop event queue before we drop down to release shared SW state */
-
-err_close_nic:
        destroy_workqueue(mdev->workqueue);
+
 err_mr:
        mlx4_mr_free(dev, &mdev->mr);
 err_uar: